Citation
An ecological analysis of soil and water conservation in hillslope farming systems

Material Information

Title:
An ecological analysis of soil and water conservation in hillslope farming systems Plan Sierra, Dominican Republic
Uncontrolled:
Plan Sierra, Dominican Republic
Creator:
Rocheleau, Dianne E
Publication Date:
Language:
English
Physical Description:
xiii, 420 leaves : ill., maps ; 28 cm.

Subjects

Subjects / Keywords:
Coffee industry ( jstor )
Crops ( jstor )
Eroded soils ( jstor )
Forests ( jstor )
Land use ( jstor )
Rain ( jstor )
Sediments ( jstor )
Soils ( jstor )
Tillage ( jstor )
Watersheds ( jstor )
Dissertations, Academic -- Geography -- UF
Farms, Small -- Economic aspects -- Dominican Republic ( lcsh )
Geography thesis Ph. D
Soil conservation -- Research -- Dominican Republic ( lcsh )
Water conservation -- Research -- Dominican Republic ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1984.
Bibliography:
Bibliography: leaves 394-419.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Dianne E. Rocheleau.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
030502015 ( ALEPH )
11698486 ( OCLC )

Downloads

This item has the following downloads:


Full Text












AN ECOLOGICAL ANALYSIS OF SOIL AND WATER
CONSERVATION IN HILLSLOPE FARMING SYSTEMS:
PLAN SIERRA, DOMINICAN REPUBLIC



















BY

DIANNE E. ROCHELEAU


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA


1984













ACKNOWLEDGMENTS


This work could not have been completed without the kindness and

able assistance of family, friends and professional associates too

numerous to mention. Special thanks are due to Dr. Gustavo Antonini

for his confidence in me, his personal concern and his active support

of my work; to Dr. Katherine Ewel for her guidance and encouragement

throughout my interdisciplinary program at the University of Florida;

to Dr. Helen Safa for her moral support, financial and intellectual

contributions during the last year of dissertation work; to Dr. Robert

Marcus and Dr. James Henry for their careful review and suggestions;

to Dr. H. T. Odum for sparking my imagination; and to Dr. Manuel

Paulet for direction and support in Santo Domingo.

Several offices of the University of Florida contributed

financial and logistic support, including the Graduate School, the

Center for Latin American Studies, the Department of Geography, and

the International Programs Office of the Institute of Food and

Agricultural Sciences. The latter covered research expenses in the

field with fund provided by a Title XII grant from USAID. My research

stipend was provided by a traineeship from the OAS for a period of 18

months. The State Secretariat of Agriculture, principally through

Plan Sierra, provided extensive logistic and financial support for the

field research, as well as employment for a period of six months.

My research benefitted substantially from the professional

dedication and warm friendship of many colleagues at Plan Sierra,








especially Angel Liriano S., Victor Montero, and Geuris Martinez.

Invaluable data and assistance were also provided by the National

Hydrology Institute of the Dominican Republic (INDRHI), the National

Cartographic Institute, the Dominican Electric Company (CDE), and the

Departments of Meteorology and Land and Water, of the State

Secretariat of Agriculture (SEA).

Final data processing, drafting, and typing tasks were completed

with the able assistance of Nelly Mogallon, Kim Feigenbaum, Beth

Higgs, and Pat French.

The warmest appreciation is reserved for those closest to home,

especially Mickie, Nelly, Gustavo, Marie, Mom and Dad. More than any

other, I owe the successful completion of this work to my husband,

Luis, who helped me tap my own energies and gave selflessly of his

time and effort as computer consultant, data processing technician,

editor, critic, and nurturer.


iii













TABLE OF CONTENTS



Page

ACKNOWLEDGMENTS ................................................ ii

LIST OF TABLES ................................................. vi

LIST OF FIGURES.................................................. ix

ABSTRACT....................................... ................ xii

CHAPTER I INTRODUCTION..................................... 1

The Problem.. .................................... 1
The Sierra Region..................o.................. 7
Purpose and Scope of Work............................. 11

CHAPTER II A SELECTIVE REVIEW OF STUDIES APPLICABLE TO THE
PROBLEM ................... ............ ............ 15

Overview ................................. ....... 15
Erosion and Sedimentation Research in Geomorph-
ology, Biogeochemistry and Ecology................. 15
Models of Erosion and Sedimentation.................. 17
A Review of Relevant Findings in Experimental
Watersheds and Erosion Plots.................... 24
Qualitative and Informal Analyses of Land Use
and Erosion in the Caribbean and Similar Environ-
ments............ ......................... .... .... .... .. 44
The Role of Farming Systems Research in Soil and
Water Conservation.............................. 45
Farming Systems, Agroecosystems and Agroforestry
Research ............. ................. ...... ...... 46
Central American Research............................ 51

CHAPTER III METHODOLOGY....................................... 58

The General Approach............................. 58
Materials and Methods.......................... ... 60

CHAPTER IV RESULTS AND DISCUSSION.............................. 101

Regional Profile................................. 101
Study of Large Watersheds..........o............. 130
Study of Small Watersheds......................... 164
Erosion Plot and Household Studies.................. 207














CHAPTER V

APPENDIX A

APPENDIX B

APPENDIX C



APPENDIX D

APPENDIX E



APPENDIX F


APPENDIX

APPENrDIX

APPENDIX

APPENDIX


Application of Erosion and Runoff Coefficients
to the Small Watersheds ........................
Sediment Delivery Ratios .........................

CONCLUSIONS ......................................

COMPARATIVE DATA FROM LITERATURE REVIEW ..........

SURVEYED CROSS SECTIONS OF RIVERS AND STREAMS....

STAGE DISCHARGE CURVES FOR RIVERS, DERIVED BY
INDRHI .................................... .......

MONTHLY RAINFALL FOR STATIONS 3 THROUGH 11 .......

DATA ON SEDIMENT CONCENTRATION, STAGE AND
DISCHARGE FOR MAO AND AMINA RIVERS.....o........

DATA ON SEDIMENT CONCENTRATION, STAGE AND
DISCHARGE FOR SMALL WATERSHEDS ...................

SOIL PROFILE DESCRIPTIONS FOR EROSION PLOT SITES.

FORMS USED FOR INFILTRATION TESTS ................

DERIVATION OF FACTORS FOR USE IN USLE, BY PLOT...

DATA FROM EROSION PLOTS..........................


LITERATURE CITED ...............................................

BIOGRAPHICAL SKETCH ............................................


265
275

278

285

297



302

305



315



343

350

367

369

378

394

420













LIST OF TABLES


TABLE PAGE

1 Relative location and land cover of erosion plots........ 65

2 Equations for system model................................ 123

3 Land use systems in the region............................ 126

4 Rainfall and river discharge in the Amina and Mao
Watersheds ................................................ 144

2
5 R values for regression analyses of subwatershed
rainfall vs. river discharge and sediment load........... 146

6 Summary of regression analyses of river discharge and
sediment concentration.................................... 155

7 Sedimentation transport in Mao and Amina Rivers.......... 158

8 Sedimentation in the Mao River basin estimated from May
1980 measurements......................................... 160

9 Water and sediment yields estimated from 1980 and 1981
data...................................................... 162

10 Characteristics of the small watersheds................... 167

11 Physical characteristics of erosion plot sites........... 179

12 Sediment transport in five watersheds..................... 193

13 Discharge rates measured for low flow conditions in small
watersheds ................................................ 194

14 Flood events yielding peak sediment discharge during the
study period .............................................. 195

15 Analysis of variance of stream discharge and sediment
transport for all streams................................. 200
-I
16 Maximum recorded concentrations (g L ) per flood event
for all streams............................................ 201
-i
17 Average sediment concentration (gm L ) per flood event,
for all streams.. .......................................... 202









PAGE


18 Analysis of variance of stream discharge and sediment
transport comparing streams draining coffee stands and
streams draining food crops and pastures .................. 204

19 Peak discharge ha per flood event for all streams,
results of the a posteriori test of the means............. 205
-1
20 Peak sediment discharge rate ha per flood for all
streams, results of the a posteriori test of the means.... 206

21 Land use, land tenure and production, by household........ 208

22 Total annual storm runoff and soil loss rates, by plot.... 219

23 Relationship of total annual rainfall and storm runoff
in erosion plots........................................... 221

24 Analysis of variance and runoff and sediment losses for
all plots at site Los Montones............................. 229

25 Runoff and sediment losses for plots at the Los Montones
site: Results of the aposteriori tests of the means...... 230

26 Analysis of variance of runoff and sediment losses for
plots grouped by land use at the Los Montones site........ 232

27 Runoff and sediment losses for plots at the Los Montones
site, results of the Duncan Multiple Range Test, by
land use.................................................. 233

28 Analysis of variance of runoff and sediment losses for
all plots at site Pananao.................................. 236

29 Runoff and sediment losses for plots of the Pananao site,
results of the Duncan Multiple Range Test, by land use.... 237

30 Analysis of variance of runoff and sediment losses for
plots grouped by land use at site Pananao................. 238

31 Runoff and sediment losses for plots of the Pananao site,
results of the Duncan Multiple Range Test, by land use.... 239

32 Analysis of variance of runoff and sediment losses for
plots in pasture grouped by site........................... 240

33 Analysis of variance of runoff and sediment losses of
forested plots grouped by site............................. 241


vii









PAGE


34 Analysis of variance of runoff and sediment losses for
plots planted in crops grouped by site..................... 242

35 Runoff and sediment losses for plots in Los Montones and
Pananao sites with crops, results of the Duncan Multiple
Range Test........................................ ......... 243

36 Soil loss coefficients by land use and conservation
practice ......................................... ........ 247

37 Comparison of measured and predicted erosion losses....... 252

38 Comparison of soil loss coefficients derived from
USLE and from empirical data.............................. 255

39 Storm runoff coefficients, total runoff estimates and
runoff/rainfall ratios, by watershed...................... 271

40 Calculation of soil loss coefficients and soil loss
estimates by watershed.................................... 272


viii













LIST OF FIGURES


FIGURE PAGE

1 Systems model of land use and erosion in the Caribbean.. 5

2 Impact area: Plan Sierra. 9

3 Model of applied research process........................ 25

4 Flow chart of research activities........................ 26

5 Input-output diagram for interview notations and
monitoring............................................... 53

6 Organization of research activities...................... 64

7 Diagram of Thiessen polygons superimposed on a map of
the Mao and Amina watershed subdivisions................. 77

8 Sampling sites for large watersheds...................... 78

9 Uppsala-type manual sampler for instantaneous measure-
ment of sediment concentrations in streams.............. .81

10 A. Illustration of velocity-area method. Person "a"
releases float at time t1 and person "b" records time
it takes float to move 2 m. B. Cross-section of
stream showing placement of float to measure velocity
and area of three sections of the stream................ 86

11 Equipment installed in streams to measure sediment
concentrations at different levels of flooding.......... 88

12 Illustration of erosion plot with runoff and sediment
collectors............................................... 93

13 Illustration of alternative erosion plot design with
three subsections .......................................... 95

14 Diagram of sediment and runoff collector indicating the
points at which samples were taken....................... 97

15 Plan Sierra geologic subregions.......................... 103

16 Plan Sierra region with study sites...................... 104









PAGE

17 Plan Sierra life zones................................... 106

18 Plan Sierra land use systems............................. 109

19 System model of the Sierra............................... 120

20 Land use model........................................... 125

21 Monthly discharge of the Mao River....................... 132

22 Monthly discharge of the Amina River..................... 133

23 Monthly rainfall at the San Jose de Las Matas
climatological station, #1............................... 136

24 Monthly rainfall at the Moncion climatological station,
#2 ....................................................... 137

25 Monthly rainfall at the San Jose climatological station,
#1, for two periods...................................... 138

26 Monthly rainfall at Moncion climatological station, #2,
for two periods of record................................ 139

27 Plan Sierra region: Mean annual rainfall for 1967-1979. 142

28 Plan Sierra region: Annual rainfall for 1980........... 143

29 Time series of sediment concentrations during selected
flood events ....... ..................................... .. 153

30 Small watersheds in coffee region........................ 165

31 Land use in the Prieto watershed......................... 168

32 Land use in the larger Prieto watershed.................. 169

33 Land use in the Upper Bajamillo watershed............... 170

34 System model of small watershed in coffee region........ 172

35 Land use model for watershed model of coffee producing
region................................................... 173

36 Land use in Hondo watershed.............................. 184

37 Land use in Pananao watershed............................ 185

38 System model of small watershed in pasture-field crop
association .............................................. 187









PAGE


39 Land use model of small watershed model of pastures-
field crops association.................................. 188

40 Pananao watershed........................................ 190

41 Hondo watershed.......................................... 191

42 Illustration of equivalent sediment discharge according
to distinct regimes...................................... 198

43 Infiltration rates in erosion plots at Carrizal and
Pananao ................................................... 213

44 Infiltration rates in erosion plots at Los Montones..... 214

45 Monthly rainfall, storm runoff, and sediment loss at
Carrizal plots ........................................... 222

46 Monthly rainfall, storm runoff, and sediment loss at
Los Montones plots........................................ 223

47 Monthly rainfall, storm runoff, and sediment loss at
Pananao plots............................................ 225

48 Model of coffee farms: Large and small holdings........ 257

49 Model of dairy and cattle farm, Pananao.................. 260

50 Bitter manioc production on a small holder plot, Pananao 261

51 Model of a mixed production on a well integrated farm... 264

52 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Prieto stream........................ 266

53 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Upper Bajamillo stream.............. 267

54 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Greater Bajamillo stream............ 268

55 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Hondo stream......................... 269

56 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Pananao stream....................... 270














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

AN ECOLOGICAL ANALYSIS OF SOIL AND WATER
CONSERVATION IN HILLSLOPE FARMING SYSTEMS:
PLAN SIERRA, DOMINICAN REPUBLIC

BY

DIANNE E. ROCHELEAU

August 1984

Chairman: G. A. Antonini
Major Department: Geography

The purpose of the study was to develop and test an

interdisciplinary methodology for applied research in soil and water

conservation in hillslope farming systems. The specific objectives

were to collect baseline data on erosion and sedimentation in the

Sierra region of the Dominican Republic and to evaluate soil

conservation technologies and cropping systems recently introduced

into the area.

Erosion, runoff and sedimentation were measured at three scales

2
of analysis: on-farm experimental plots (22 x 2 m ), small watersheds

2 2
(1-30 km ), and large watersheds (300 km ). Erosion losses in the 16
-I -1
plots ranged from 0.05 to 0.10 tons ha yr under pine forest, 1 to
-i -i -I -i
3 tons ha yr in pasture, 0.5 to 3 tons ha yr in coffee
-1 -l
plantations, and 6 to 70 tons ha yr for plots in mixed food crops.

Annual storm runoff varied from 1% of precipitation under forest to

12% in an eroded continuously cropped plot. Infiltration and soil

profile analyses and erosion measurements at the plots showed a


xii









pronounced influence of intensity and longevity of cultivation at the

site. Erosion rates also changed dramatically with the phases of the

cropping cycle on the coffee and food crop plots.

The recently introduced slope modification (hillside ditches) did

not significantly reduce erosion rates at the test plots. The minimum

tillage field trial showed more substantial reductions.

Field data were combined with photogrammetric analyses of land

use to compare erosion and sedimentation rates at the watershed level

for production systems based on coffee and pasture with annual crops,

respectively. The contribution to river sedimentation and flooding

did not differ significantly between the two systems at the watershed

level. In both cases it is the association of annual crops with the

dominant commercial land use that determines the erosion rate. The

planting of coffee for soil conservation was ineffective in most cases

because of high labor and related food crop demands.

Suggested alternatives to coffee include reduction of competition

for land and labor between commercial and subsistence production by

substituting tree crops that meet household demands at the local and

national level. Recommendations for land in pasture and annual crops

are the reduction of tillage in annual crop plots, the mixture of more

food crops into annual cash cropping systems, and the combination of

grazing and tree crops in large holdings.


xiii















CHAPTER I
INTRODUCTION


The Problem



Almost 10 years after the first dramatic success of the Green

Revolution, the technological breakthroughs remain largely

inaccessible to small farmers of the underdeveloped world. Increased

yields have occurred primarily in large scale commercial or state

enterprises (Harris, 1973; Greenland, 1975; Stevens, 1977). Millions

of small farmers who produce commercial and subsistence food crops and

cash crops for export have maintained or increased production only at

great cost to themselves and to the natural environment (Crosson and

Frederick, 1977; Eckholm, 1976).

These farmers have expanded into ever more marginal areas (often

arid, semiarid, or hillslope environments) or have intensified

cultivation of existing plots, often already located in marginally

productive lands (Brush, 1981). The intensification has been for the

most part without benefit of new technology or capital inputs. It is

achieved through increasingly higher inputs of labor, and often

through practices that damage the long-term fertility and stability of

the soil (Geertz, 1972; Lagemann, 1977) and disrupt hydrologic and

geologic cycles in watersheds (Greenland, 1974; Kellman, 1969;

Pereira, 1973; Rapp, 1977).

Given this situation there is an immediate need for research to

adapt technologies to needs of small farmers within the limits of the

available factors of production and environmental constraints (Crosson









et al., 1978; Hildebrand, 1981; Lagemann, 1977; Makhijani and Poole,

1975; Novoa and Posner, 1981). The "external costs" of watershed

degradation must be considered and tested within the context of such

research (Erickson, 1974; Novoa and Posner, 1981). Beyond the need

for an inventory of the current magnitude of the problem, there is a

real need to describe and test the complex functional relationships

between land use, production, erosion and sedimentation under field

conditions, in order to explore viable alternatives to the causes of

the problem.

It is widely recognized that the combined processes of erosion

and sedimentation pose a serious threat to sustained production in

both upper and lower portions of tropical watersheds (Farvar and

Milton, 1973). The decreasing depth and fertility of the soil limit

the productivity of small upland farms (Greenland and Lal, 1977;

Morgan, 1979), while quantity, quality, and regularity of surface

water supplies limit both urban and agricultural development in the

lowlands downstream (McPherson, 1974; Nelson, 1973; Odum and Odum,

1976). This is especially critical in tropical and subtropical humid

montane environments subject to intense population pressure by farmers

using traditional and semitraditional methods (Antonini et al., 1975;

Floyd, 1969; Santos, 1981; Sheng and Michaelson, 1973; Wilson, 1976).

While subsistence farmers, landless laborers and their

traditional technology are often blamed for deforestation and

environmental degradation under such circumstances (Rodriguez, 1980),

they are constrained by limited access to land and lack of

alternatives (Plumwood and Routley, 1982). The settlement of rugged

hillslopes and the near total deforestation of upland watersheds










represent a choice by default rather than a free choice between

rational alternatives for sustained production. Moreover, there is

intense pressure for continuous cropping and/or establishment of

pastures on cleared land. Large local landowners, as well as urban

and foreign markets, play a major role in this process (Amin, 1977;

Cultural Survival Inc., 1982, Hildyard, 1982; Nations and Komer, 1982;

Plumwood and Routley, 1982).

Small farmers and shifting cultivators in such areas often

practice forms of management that can be sustained well at lower

population densities or within more hospitable environments to which

they have no access (Bailey, 1982; Grainger, 1980; Nations and Komer,

1982). Although these small farmers in marginal areas are referred to

by many as "subsistence farmers," they usually produce some surplus

food crops. In some cases this sector is a major source of staple

food production for domestic markets (Brush, 1981; Novoa and Posner,

1981). The same farm families often function as a seasonal labor

force in coffee, lumber, and other cash crop harvests (Beckford, 1972;

Frucht, 1967). These subsistence farmer/farmworker populations in

marginal lands highly susceptible to erosion form an integral part of

the regional economy and ecosystem. As such, the problem must be

treated as a complex phenomenon that not only affects the larger

downstream and lowland production systems, but is partially

conditioned by them.

The question remains as to how production can be maintained or

increased (at a sustained rate) with minimum damage to both portions

of the watershed. The need for an answer to this question is

particularly urgent in the Caribbean because of high population










pressure on hillslope lands (Antonini et al., 1975; Santos, 1981)

coupled with high demand for food crops and relegation of large lowland

tracts to cash crop cultivation (Beckford, 1972; Rankine, 1976; Wilson,

1976).

The resource base in the Caribbean is subject to intensifying

multiple demands by commercial and subsistence sectors for food, cash

crop, wood, fuel and mineral production, as well as protection of the

watershed for downstream development. From a national perspective, the

upland watershed's most important export crop may well be water, needed

for irrigation and hydroelectric projects for downstream development

(Swedforest, 1980). The various types and rates of production demanded

are often competitive in nature, if not mutually exclusive (Crosson and

Frederick, 1977; McPherson, 1974). In many cases the upland regions'

internal situations also clearly indicate the need for change (Brush,

1981; Chaney and Lewis, 1980; Ferreiras, 1979; Hildebrand, 1981; Reiche

and Lee, 1978; Santos, 1981; SEA, 1978).


A Conceptual Model of the Problem


Based upon the information presented above, a model is postulated

that describes the interaction of significant elements in Caribbean

land use systems with regard to the problems of soil erosion and rural

poverty (Fig. 1). The model shows the interaction between population,

land use and the condition of soil, water, and vegetation both within

and outside the study region.

The model diagram follows the format developed by Odum for energy

modelling of ecosystems (Odum, 1971). The tank-shaped symbols (Fig. 1)






































































w
Dr










represent storage of land, soil, biomass, water, and economic assets

within the system. The bullet-shaped symbol indicates a

transformation process (in this case photosynthesis), while the

circles define forcing functions from outside the agroecosystem. The

arrow-shaped devices are workgates, which regulate the interaction

between the various mass and energy flows. Changes in the storage

are determined by the input and output flows, indicated by solid

lines. The imports and exports from the system are indicated

likewise.

In this case the export of soil (soil erosion/sedimentation) is

of special interest, along with changes over time of crops, other

vegetation storage, and-human population. The hypothesis implicitly

stated in the model is that upland land use controls both the export

of soil from the system, and the production of food and income for the

population, both in the uplands and the lowlands.

The problem remains to reconcile internal regional development

goals with national priorities to define desirable changes. At best,

the conflicting needs will be met by careful optimization of land and

water use within entire river basins (Pereira, 1973; White, 1977).

The determination of what is optimal implies a client group within a

defined spatial and temporal context. Decisions must be made to

reconcile differences in "optimal" solutions at local and regional

scales, as well as to adjust short-term and long-term costs and

benefits. Policy makers also need information on the distribution and

nature of costs and benefits among various sectors of the population

in order to formulate "optimal" strategies. Resulting rural

development programs must include short- and long-term incentives at










the local level for the widespread adoption of resource management

practices that will benefit the population of the region as a whole

over the long term (Gladwin, 1981; Hildebrand, 1981; Santos, 1981).

Traditional cost-benefit analysis will not suffice since it

excludes environmental as well as social aspects of the system, assumes

a static system, and has been developed for application over relatively

short time periods (Amin, 1977). The problem requires a more holistic

theoretical and methodological approach that will evaluate

environmental and human concerns on their own terms and within the

total system rather than by econometric criteria.


The Sierra Region


The Sierra is a rugged montane area in central Dominican Republic

that has been subjected to the traditional practice of shifting

cultivation, as well as to extensive exploitation of primary resources

such as timber and mineral deposits. It is a relatively underdeveloped

region within an underdeveloped country where the area under production

2
in the country (27,000 km ) already has surpassed the area of land

2
classified as suitable for agriculture (22,000 km ) (OAS, 1967;

Swedforest, 1980; USAID, 1974). Production increases necessary to meet

the national demands for food and income have come from increased

yields in areas already under production, or from expansion into more

marginal areas such as the Sierra. The latter strategy has dominated

among the poor and landless members of the peasantry (Beckford, 1972;

Antonini et al., 1975), and has been a last resort for the former

employees of mining and sawmill camps and furniture shops, most of










which had closed by 1979 (Ferreiras, 1979). The major alternative,

emigration to the capital city of Santo Domingo, and to New York City,

provided an outlet for a large segment of the population during the

1960s and early 1970s.

The impact area of Plan Sierra (Fig. 2), an integrated rural

development project within the Sierra, offered a unique opportunity to

examine the apparent conflicts between agricultural development and

natural resource conservation (Santos, 1981). Plan Sierra is a joint

venture between the State Secretariat of Agriculture and the private

sector to initiate and coordinate development efforts in the region in

several sectors: agriculture, livestock, credit associations, health

services, transportation, handicrafts, university programs, and

natural resource management (Quezada, 1977; Antonini and York, 1979;

Plan Sierra, 1979). The Plan's objectives are: 1) to improve the

quality of life of the inhabitants of the Sierra; 2) to manage soil,

water and forest resources; and 3) to promote participation by local

people in the development process (Antonini and York, 1979; Santos,

1981).

Several hydroelectric and irrigation projects are planned for the

study area (Jorge, 1981). The impoundments will serve the Cibao

Valley downstream. Sedimentation from this area is already a problem

in the Tavera Reservoir, completed in 1977 (Cepeda, 1980). The

magnitude and distribution of erosion in the uplands, however, has

received little attention until recently (Vasquez, 1980).

Plan Sierra has promoted specific farming practices and changes

in land use that are intended to reduce sediment export. A parallel






9





















I V / ^ -^^^"^^ So-- 0E








iX 2S^\ ^
o eo









E U

0 4-)
\o
1 0
o M
IL,










objective of the program is the development of viable sustained yield

agricultural systems. Existing land use in the region consists of a

combination of forest, coffee, bush, small plots of annual crops, and

pastures. The relative rates of erosion and sediment yields from the

prevailing land use systems in this area have been estimated, but not

measured (Antonini et al., 1975). Neither have the effects of erosion

within the upland system been investigated. In the absence of basic

data Plan Sierra had to experiment with tentative conservation

practices and crop changes to curtail erosion and to maintain or

improve production. The widespread adoption of hillside ditches and

various forms of terracing was strongly encouraged by credit

incentives to large and small landowners alike. The conversion of

pasture, bush, and crop land to coffee also was promoted through

credit incentives and supported by extension and education programs.

The research project described in the following chapters sought

to define the magnitude of the sedimentation and erosion problems in

the region, to measure the variation of erosion rates under existing

land use systems, and to test the effectiveness and feasibility of

specific land use changes and conservation practices promoted by Plan

Sierra. The study was integrated into the Plan Sierra soil, water and

conservation program. Logistic support as well as the active

collaboration of paratechnical, technical and professional staff made

it possible to conduct the study at a regional scale and with an

intensity of effort that would otherwise have proven difficult if not

impossible.










Purpose and Scope of the Work


The purpose of the study was to develop and test an

interdisciplinary methodology that addresses the needs cited above

within the context of the Caribbean. Such a methodology, and the

theoretical framework within which it is developed, should meet the

following criteria:

1. Treat the problem in a holistic way, building upon existing

research in several disciplines, and incorporating both physical and

cultural aspects of erosion and land use problems;

2. Be flexible enough to include analyses within a broad range

of temporal and spatial scales of observation and to allow for

differentiation by client groups;

3. Facilitate the participation of clients in research and

extension programs;

4. Be applicable in areas with limited data bases;

5. Yield practical short-term results at the topographic scale

while working toward more basic solutions over the long-term;

6. Be amenable to integration into rural development programs,

from preliminary research to subsequent extension efforts.


Objectives


The specific objectives of the study included short-term

practical achievements at the local and regional level. The latter

were readily accomplished within the larger task of methodology

development and testing and contributed to the development of the

theoretical aspects of this study. The objectives, stated in

chronological order of completion, were as follows:










1. The study will provide useful information and services for

farmers, rural communities and regional policymakers in the study

area, the Sierra region of the Dominican Republic. This includes the

collection of basic data about the magnitude, distribution, impact and

causes of the problem. (In this case, the basic data will include

knowledge and perceptions of the residents about the study area and

the problems of erosion and sedimentation.) Further, the study will

develop and adapt practical field methods and data analysis techniques

that are suitable for future use by local personnel trained within the

project. The project should become a vehicle for farm extension and

community education in management of natural resources.

2. The study will develop and test a theoretical model of the

relationship between land use, runoff, erosion and sedimentation in

the Sierra. The hypotheses implied in the model will be tested and

the results will contribute to the evaluation and modification of the

model. The refined model will be used as a point of departure for

future simulation and for planning of field trials in the study

region.

3. A general version of the resultant model will be proposed for

application to the problem in the Caribbean. The methodology for

testing the model and applying it to research and extension efforts in

the region will be presented.


The Theoretical Context of the Study


Systems analysis provides a theoretical framework and analytical

tools appropriate to the proposed study and adequate for the complex










interdisciplinary nature of the problem. It allows the researcher to

focus on the interactions and mutual causality so difficult to include

in other analyses and so critical to understanding the links between

land use, poverty, and erosion. The relationship of the upland areas

to the larger watershed and the national context is also readily

included within this perspective by using the concept of nested

hierarchies in systems of various scales (Antonini et al., 1975; Odum

and Odum, 1976). This feature of systems analysis also allows the

inclusion of boundary conditions on subsystems that may be determined

by the larger system or some elements thereof. As such, the

"structural determinants" of land use and production relationships

(Beckford, 1972) can be considered within what has often been viewed

by structural determinists as a noncompatible methodology. The

theoretical assumptions underlying this approach are in fact

consistent with the world-system paradigm in anthropology (Nash,

1981).

Moreover, within the broad framework of systems analysis it is

possible to combine elements of such fruitful and diverse approaches

as the ecosystems-energy analyses developed by Odum and Odum (1976)

and the farming systems research and extension methodology developed

by Hildebrand (1981). Other related lines of research which can be

incorporated include the basic watershed-ecosystems analysis (Hart,

1980), agroecosystems testing and development conducted by Ewel

(1981), farming systems research conducted by Ruthenberg (1976), Flinn

(1980), and Lagemann (1977), and regional characterization of tropical

agroecosystems (Posner et al., 1981).










Systems analysis offers the potential for experimenting with a

variety of "futures possible" through simulation of dynamic nonlinear

models (Ewel et al., 1975; Forrester, 1971; Lagemann, 1977; Meadows et

al., 1972; Odum, 1971). Such analyses allow us to "optimize" over

various time scales and according to different criteria. Ideally,

this provides planners, policymakers, and their clients with

information about the probable outcome of different resource

management alternatives; the net result can be viewed from the

perspective of the farm household to the national level.

The accuracy and quality of the conclusions drawn from the

qualitative or quantitative analysis of such models depend upon the

accurate conceptual design of the model and the quality of the

information used to evaluate its components (Amadeo and Golledge,

1975; Harvey, 1969; Kuhn, 1962). The model, as a selective

simplification of reality, must be constructed by one who is famil-iar

enough with both the system-at-large, and the problem of interest, to.

choose which elements to include, and to describe their relationships

accurately.

A researcher trained in ecosystem analysis and modelling can

perform this combined function by modelling the entire system and by

adding questions of ecological concern and ecological monitoring

methods to interdisciplinary adaptive research in farming systems and

resource management. A general reconnaissance survey of the region,

and interaction with researchers from complementary disciplinary

backgrounds, as well as with project clients, can provide the kind of

information and broad perspective needed to properly define the

system.














CHAPTER II
A SELECTIVE REVIEW OF STUDIES APPLICABLE TO THE PROBLEM


Overview



Various aspects of the relationship between erosion,

sedimentation and land use have been treated by geographers as well as

by researchers in several other fields. Although some geographers

(Antonini et al., 1975; Haggett, 1961; Kellman, 1969; More, 1967;

Morgan, 1979; Pereira, 1973) and other scientists (Geertz, 1972;

Greenland, 1974, 1975; King, 1978; Lagemann, 1977) have spanned both

ends of the research spectrum, the literature is generally divided

along these lines and will be reviewed as such.


Erosion and Sedimentation Research in Geomorphology,
Biogeochemistry and Ecology


The theoretical aspects of erosion and sedimentation mechanics at

the micro-scale are fairly well established and understood and already

have been summarized for general reference use by applied scientists

(Brady, 1974; Chow, 1964; Gregory and Walling, 1973; Morgan, 1979;

Toy, 1977; USDA, 1979). What are less clear, and still require a wide

range of basic and applied, theoretical and empirical research, are

the more complex causes and effects of erosion and sedimentation at a

larger scale, under field conditions.

The study of erosion and sedimentation and the mechanisms that

determine their occurrence at the meso- and regional scales under










field conditions is a proper subfield of physical geography since

geomorphology is a traditional and well-developed avenue of inquiry

within the discipline (Gregory and Walling, 1973; James, 1972; Keller,

1968; Morgan, 1979; Stoddart, 1965; Strahler, 1964). Systems theory

has been applied widely in studies of watersheds and other

geomorphological units by British geographers (Chisholm, 1967;

Chorley, 1962; Kirkby, 1978; Oilier, 1968) as well as by numerous

other geomorphologists (Leopold et al., 1964; Leopold and Langbein,

1962; Toy, 1977) and hydrologists (Chow, 1964; Vemuri and Vemuri,

1970). This general approach offers the advantage of integrating form

and process, by accounting for their interactive relationship

(Chorley, 1962, 1969).

The open systems approach allows the inclusion of the

quantitative hydrology/geomorphology tradition that dates from Horton

(1935, 1938) and continues in the work of Strahler (1964) and other

physical geographers. The use of systems theory in geomorphology also

facilitates the study of the human use of the earth, a longstanding

focus of geographic research (Harvey, 1969; James, 1972; Marsh, 1964;

Thomas, 1974) that could lend itself well to quantification within a

systems framework (Stoddart, 1965).

The application of systems theory to the study of watersheds has

also been tested and developed within ecology in recent studies of

biogeochemistry (Bormann and Likens, 1979; Likens et al., 1977) and

resource management (Cooper, 1971; Hall and Day, 1977; Hopkinson and

Day, 1980; Patton, 1971; Thomas, 1974; Van Dyne, 1969). In such

studies the watershed defines the boundaries of the ecosystem and the










interacting elements whose mutual influences are observed within the

system include physical, chemical, biotic and cultural entities (Odum,

1971, 1982; Odum and Odum, 1976).

The general conceptual framework of systems analysis is well

suited to the study of mutual influence between land type, land use,

plant and animal productivity, and erosion in the upland watersheds of

the Caribbean. Incorporating the broad perspective of the man/land

tradition in geography, the research in open systems geomorphology and

ecosystem analysis also offers valuable methodological examples for

application to interdisciplinary research on the topic.


Models of Erosion and Sedimentation


The range of erosion and sedimentation models in use includes

stochastic and deterministic models, statistical as well as

parametricc" models, and combinations of all of the above. These

models have been developed and applied within scales of analysis

ranging from individual plots to large watersheds.

One school of research focuses on modelling processes and

interactions between the various parts of watersheds as complex

systems (Chorley, 1962; Likens et al., 1977), and the other major

thrust of erosion and sedimentation studies has been to develop

empirical predictive equations (Wischmeier, 1975; Wischmeier and

Smith, 1978). The latter relate land use and management to erosion

loss and sediment yield and have been developed for management

purposes, primarily for use in soil and water conservation programs.










Empirical Models


The best known and most widely used of the empirical models of

erosion loss is the Universal Soil Loss Equation (USLE), an empirical

formula for estimating potential soil loss by sheet and rill erosion

for individual plots (Wischmeier, 1975; Wischmeier and Smith, 1978).

The USLE is used primarily to predict erosion losses based on a

combination of inherent site characteristics and variables subject to

human intervention and management. Tons of soil lost is the dependent

variable. It is determined by a combination of six independent

variables as follows:

A = RKLSCP, where
-i -i
A = soil loss per unit area (tons acre or tons (m) ha ) over

a given time

R = erosive potential of rainfall (based on total energy and

intensity of rainfall)

K = an index of soil erodibility (a measure of soil suscepti-

bility to erosion based on physical properties)

L = length-of-slope factor

S = inclination-of-slope factor

C = vegetative cover factor, integrating type of cover and

management of crops or other vegetation

P = conservation practice, including structural alteration of

site and/or contour planting.

The apparent disadvantages of this model reside mainly in the

costly and time consuming local calibration required for its proper

application. This is particularly important in underdeveloped










countries where research funds and personnel are limited and the

complexity of the rural agricultural landscape in the uplands requires

extensive calibration of the model. Rapid changes in farming

practice, crop types and level of technology in many areas would

require almost a continuous update of the calibration experiments.

Such a program of research not only represents a large investment in

and of itself, it implies a diversion of resources from alternative

avenues of theoretical and applied research directed toward cumulative

growth of knowledge about the processes in question.

Another weakness of the USLE is the failure of the conceptual

framework to account for the difference between land cover and land

use systems. There are no economic or social aspects to the model,-.

yet it is proposed as a practical tool for farm and regional

conservation planning purposes. The premises under which the equation

is applied often can be misleading, resulting in serious errors in

planning and management decisions. Aside from the logical pitfalls

inherent in the use of the model, there is the technical drawback of

its inability to predict the feedback effects of current erosion rates

on future land use and productivity which in turn affect future

erosion rates.

The USLE can be a useful tool for prediction of erosion rates

under specific known conditions, given prior calibration for the full

range of conditions in a region. Used alone, however, it does not

constitute an adequate basis for management decisions at the farm

level, much less for regional planning purposes.

The USLE has also been adapted for prediction of sediment yields

at the watershed scale (McElroy et al., 1976; Onstad and Foster, 1975;










USDA Forest Service, 1980; Williams, 1975) by estimating the sediment

delivery ratio based on drainage area (Holtan and Lopez, 1971; Roehl,

1962). The models derived from the USLE have been used widely in

economic and land use planning studies for evaluation of specific

cropping systems and/or conservation practices (Kling and Olson,

1975). The weaknesses of the approach include the inherent

limitations of the USLE as well as the questionable realism of

sediment routing techniques (Skopp and Daniel, 1978).

One empirical model that does not derive from the USLE is used by

the United States Bureau of Reclamation (Flaxman, 1975; Skopp and

Daniel, 1978; Strand, 1975) as well as by international research

organizations (Rapp, 1977). The technique combines flow duration

curves with sediment concentration, the latter derived from either a

sediment rating curve or a power function. Both flow duration and

sediment rating curves describe empirical relationships that must be

determined on site.

The model is useful for prediction of sediment yield over the

long term, given a continuation of current land use conditions, but is

less amenable to integrated watershed management based on land use and

land treatment programs. The flow duration/rating curve model,

however, could be calibrated to particular watersheds or groups of

watersheds for specific land uses and treatments, as has been the case

in paired watershed studies conducted in experimental catchments.

A number of sophisticated digital computer models of runoff

and/or erosion and sedimentation predict water and sediment discharge

by relating hydrologic and physical characteristics of the source










areas (Chapman and Dunin, 1975; Mein, 1977; Fleming, 1968; Negev,

1967; Ebumive and Todd, 1976; Donigan and Crawford, 1976).


Theoretical Models of Physical Processes


Research conducted by Elwell (1979a) in Zimbabwe resulted in a

simple model similar to the USLE in some aspects, but based on

rational rather than empirical parameters for estimating sheet erosion

from arable land. Rainfall energy measured in 10-day increments

defines erosivity while the protective value of crops and cropping

practices is assessed according to the percentage of seasonal rainfall

energy "i" intercepted by the vegetative canopy and ground cover.

Potential interception is determined in a fashion similar to the leaf

area index measured by ecologists (Odum, 1971).

Application of the model by Elwell (1979a) in Zimbabwe field

trials confirmed the importance of mulches for erosion control and

demonstrated the potential for reducing soil loss by increasing crop

yields. Moreover, important seasonal relationships were identified

between various protective crop covers and erosivity of rainfall in

given areas. This allowed the identification of the crops or crop

combinations that best protected the soil during periods of intense

rainfall.

The model is particularly applicable to the seasonally dry

tropics where single storm events cause a large proportion of the

total annual erosion loss. The practical application of the model is

aided by a description of the techniques for calculation and field

measurement (Elwell, 1979a).










Many of the theoretical models describing erosion and sediment

transport processes are mechanistic approximations of reality

(Bennett, 1974). Most of these constructs take systems theory as a

point of departure, although either combined or dispersed systems may

be postulated. The combined system models assume a uniform watershed,

with average physical features and composite land use, and do not

consider distance between source and outlet. In dispersed systems the

runoff and sediments are generated in spatially distinct source areas

and are then routed spatially and/or chronologically through

successive increments of the watershed (Holtan and Lopez, 1971;

Fleming and Leytham, 1976; Frere, 1978; Renard and Lane, 1975; Skopp

and Daniel, 1978).

The models within the theoretical research tradition generally

offer better conceptual realism than the empirical models (Holtan and

Lopez, 1971) but the theoretical models tend to be much stronger in

the explanation of physical and chemical processes and remain largely

underdeveloped in the treatment of biological and cultural phenomena.

These models are just now beginning to be tested in systems that

include human populations, and management options are being added to

the range of variables to be tested. Meanwhile, the vast majority of

field data collected on erosion, sedimentation and land use has come

from empirical studies designed to calibrate predictive equations that

are, essentially, site-specific black box models of complex systems

and processes (Boughton, 1967; Hayward, 1967; Holtan and Lopez, 1971;

Skopp and Daniels, 1978).










Ecosystem Models of Watershed Processes and Land Use


Two types of ecosystem models that are less frequently applied to

erosion and sedimentation studies warrant special consideration for

their broad applicability and holistic approach. The first is a model

developed to analyze the biogeochemistry of a forest ecosystem at the

Hubbard Brook Experimental Forest in New Hampshire (Likens et al.,

1977). The second is an energy model used to simulate the complex

interaction of population, land use and sedimentation in the Central

Mountains (Cordillera Central) of the Dominican Republic (Antonini et

al., 1975).

The ecosystem analysis studies conducted at Hubbard Brook focus

on quantification of nutrient budgets through monitoring of

meteorological inputs and geologic outputs of nutrients in small

watersheds. The model used for the study places a strong emphasis on

precipitation, runoff, and solute and sediment export from the areas

of interest (Likens et al., 1977). The variables are measured first

under primary undisturbed forest cover, and later under disturbed

conditions. The final product of this analysis is an annual

hydrologic and nutrient budget that includes sediment yields.

Another form of ecosystem analysis is based on the evaluation of

energy pathways within communities of plants and animals, and on the

relationship between those pathways and the physical environment

(Antonini et al., 1975; Odum, 1971). Recent studies using the energy

approach have extended this concept to include the complex transfers

of matter, energy and information in ecosystems that include human

populations. Diverse interactions, ranging from mineral cycling,










photosynthesis, runoff and sedimentation to economic transactions, are

interrelated by the common denominator of energy flow (Antonini et

al., 1975; Odum, 1971; Rappaport, 1971).

Models using the energy flow language developed by Odum have been

applied to analyses of watershed ecosystems in several environments.

The most pertinent case is a model of the interaction between land

use, erosion and sedimentation in the rugged uplands of the Dominican

Republic (Antonini et al., 1975). As illustrated in the diagrams

(Figs. 3 and 4), the models are dynamic non-linear systems models that

can be simulated on an analog or a digital computer by simultaneous

solution of differential equations. The equations describe the change

in the landscape variables over time. The model indicates mutual

causality between variations in land use and population but does not

account for similar relationships between erosion and land use, since

erosion is not considered as a separate process. The model simulation

demonstrates the implications of current and alternative trends in

land use for future conditions of population, land use and reservoir

sedimentation. The two-part model developed in this study provides

the conceptual point of departure for construction of a single energy

flow model of hillslope land use systems, upland erosion and

downstream sedimentation in the Caribbean.


A Review of Relevant Findings in Experimental Watersheds
and Erosion Plots


During the past 10 years the international scientific community

and policymakers at various levels have focused greater attention on

the problems of erosion and sedimentation, particularly in the tropics














I



neati jve I
feedback


P E RC EPTIIAL
EX P N' L Pr' ':ES


I MAC E 0 F
REAL IORPLD
STRUCTURE


unsuccessful


|t

A PRIORI
TODEL
(formal representat.Iion
oF thile izIo()


I
HYPOTHF.SES


EXPERIMENTAI.
DESI;CN-
(definition, classifica-
Lion, mcasuremnent)

DATA

f -


success f I


LAWS AND
THEORY
CONSTRUCTION


posi tive
feedback


EXPLANATION


Fig. 3. Model of applied research process (after Harvey,
1969).


H


VERIFICATION
PROCEDURES
(statistical tests,
etc.)


POLICY


H ilt-1















































Test best alternatives
at other sites


Fig. 4. Flow chart of research activities.










(Brown, 1980; Henkes, 1982). However, a large proportion of the

published literature on the topic refers to studies conducted in the

United States. While the findings and methods are not all directly

applicable to the Caribbean, they supplement the less extensive data

base and cumulative research experience available at present from

Third World and Caribbean sources.


Summary of Recent Research in the United States


Three landmark studies in experimental watersheds have set the

methodological and technical trends for studying the impact of land

use on sediment yields in small watersheds. The Coshocton, Ohio,

watershed studies (Harrold et al., 1962; Mustoneu and McGuiness, 1968)

focused primarily on sedimentation rates under various cropping and

land treatment conditions in agricultural areas. Research conducted

at Coweeta, North Carolina, and Hubbard Brook, New Hampshire,

addressed both methodological and theoretical aspects of watershed

biogeochemistry (including land treatment) in forested ecosystems

(Bormann and Likens, 1979; Douglass and Swank, 1975; Likens et al.,

1977). These experiments served as models for a series of applied

watershed studies recently initiated in forests, agricultural land,

and rangeland.

The more recent research has been conducted primarily under the

auspices of federal environmental legislation that mandates

documentation of non-point sources of pollution, including sediment

discharge (Haith and Dougherty, 1976; USDA Forest Service, 1980;

Jewell and Smith, 1976; Rao, 1980; Reikerk et al., 1978). This










research focuses more on entire watersheds and less on individual

plots, in contrast with prior work conducted by the Soil Conservation

Service at the farm level (Ackerman, 1966). Most of the watershed

experiments combine monitoring of precipitation with recording rain

gauges, continuous monitoring of stream discharge at weirs, sampling

of sediment discharge at weirs, and/or collection of sediments in weir

ponds in watersheds with areas less than 20 km2 (Hewlett et al., 1969;

Ward, 1971).

A major topic of the earlier studies was the role of the

undisturbed forest in regulating the hydrologic cycle and sediment

export (Douglass and Swank, 1975; Helvey, 1967; USDA Forest Service,

1980). Among the more important findings were: the importance of

litter versus canopy in protecting the soil against the erosion

potential of rainfall (Table A-I); the impact of forest vegetation on

stream discharge (Dils, 1957; Johnson and Swank, 1973) (Table A-2);

and the association between undisturbed forest cover and low sediment

concentration in streams (Table A-3).

Streamflow, sediment concentrations and mass transport from

forested watersheds showed dramatic changes after harvesting, various

site clearing and management operations, or conversion to other uses.

Several studies reported heavy increases in suspended sediment and

nitrate concentrations after clearcutting (Bormann and Likens, 1979;

Douglass and Swank, 1975; Hewlett and Nutter, 1969; Likens et al.,

1977; Monk, 1976).

Water yield increments proportional to percent area in cleared

openings were reported for several gauged watersheds (Likens et al.,










1977; Sopper, 1975). This was attributed to increased storm flow as

well as to reduced water consumption by evapotranspiration. Reported

streamflow increases in Georgia, South Carolina, and Oregon ranged

from 40 to 50% (USDA Forest Service, 1980).

The combination of increased sediment concentration and higher

streamflow resulted in dramatic increases in sediment transport after

clearing (Monk, 1976; Sopper, 1975). By contrast, little or no change

was reported in a patch-cut watershed in the Oregon Coast Range, while

much higher figures are reported elsewhere with use of conventional

treatments (Brown, 1982; Monk, 1976).

Results from east Texas forested watersheds identify sediment as

the major pollutant in streamflow. At the same sites, in a study of

harvest practice and related impact on water quality and mass

transport, the highest sediment transport rates occurred in

association with harvesting along streams. The design, construction

and maintenance of roads often are cited as major determinants of

water quality in forested watersheds (Fredriksen, 1970; Pavoni, 1977;

Texas A&M, 1979; Ursic, 1978).

Water yield and mass transport of minerals and sediments from

nonforested watersheds differ markedly from forested sites. A study

of nutrient yields from various categories of land use in the

watersheds of 24 Connecticut lakes estimates contributions of

phosphorus from agricultural and residential-commercial land at 200

and 1100%, respectively, of forest contributions (Hill and Frink,

1978). National averages of erosion rates for various categories of

land use in the United States (Table A-4) agree well with the results










from the Connecticut study. Transport rates from harvested forests

are extremely high, even in comparison to agricultural uses. The high

rates, however, are offset by the fact that forest harvests are

periodic events that only occur once every 15 to 40 years, even in the

fast-growing pine plantations of the Southeast and the Caribbean. By

contrast, many agricultural uses are sustained continuously on a given

parcel of land.

Recent studies in agricultural watersheds in the United States

concentrate on cropping systems and practices in large scale

commercial farms or ranches. Pollution of surface waters by chemical

fertilizers (Haith and Dougherty, 1976) and pesticides often

overshadows sediment pollution as a subject of public concern (USDA

Soil Conservation Service, 1980). Pathogens entering the waterways

from feedlots and grazing lands (Jewell and Smith, 1976) also attract

more attention, although sediment is the major pollutant, by volume,

discharged into surface waters from agricultural lands. Suspended

sediments in streams and rivers carry pathogens as well as chemical

pollutants. Much of the current research, however, emphasizes the

chemical by-products discharged into waterways in solution from

agricultural non-point sources (Rao, 1980).

Studies of erosion in individual plots offer more information on

the variation in erosion rates with changes in cropping systems, farm

management and conservation practices. Data analyses by Wischmeier

and Smith (1978) for cropped and clean-tilled plots corroborate the

conclusions of studies in forest ecosystems. The canopy cover and

ground cover on the site determine how much rainfall energy reaches









the soil surface. As in the forest, both canopy and mulch (comparable

to litter) intercept raindrops, but mulch does this so close to the

surface that the drops regain no fall velocity. Mulches also obstruct

runoff flow and reduce its velocity and sediment transport capacity

(Wischmeier and Smith, 1978). Surface roughness of the soil also

influences the velocity and transport capacity of runoff. Thus,

tillage practices, crop yields and crop rotations, as well as above-

and below-ground architectural characteristics of particular plants,

influence the degree to which given cropping systems reduce erosion

relative to a standard clean-tilled plot (Mannering and Meyer, 1963;

Meyer and Mannering, 1961).

Annual row crops vary widely but erosion rates generally range

between 5 and 50% of those measured in clean-tilled plots. For

example, a field tilled with chisel and disk plows, rotated from wheat

to meadow to corn, with one crop annually, loses an average of 9% of

the total loss for a control plot (Wischmeier and Smith, 1978).

Pasture, rangeland, bush and woodland reduce erosion to between 1

and 10% of clean-till soil losses, and undisturbed forest further

reduces the loss to between 1 and 0.01% of the bare fallow. Losses

under harvested, mechanically prepared woodland sites, however, vary

between 10 and 90% of the clean-till figures. Average annual erosion

losses for cropland in the United States bear out the trends reported

by Wischmeier and Smith (1978) (Table A-5).

The same results have been extrapolated to the watershed level

with the use of sediment delivery ratios (USDA Soil Conservation

Service, 1980; Wischmeier and Smith, 1978). The proportion of total










eroded soil that arrives at a given outlet ranges from 33% in a 1 km2

2 2
area to 18% for a 25 km area to 10% for a 250 km drainage area

(Roehl, 1962).

Studies in watersheds and erosion plots in the United States

cannot be extrapolated directly to tropical and subtropical montane

watersheds. Beyond the difference in the ecosystems themselves there

is the even greater divergence in level of technology and the greater

complexity of land cover associations within upland land use systems

of the tropics. The general relationships established between land

cover, erosion, and sedimentation must be tested further and compared

with results from the Caribbean and similar regions.


Erosion and Sedimentation Research in the Caribbean and Similar
Environments


Although few in number, studies of erosion, sedimentation and

land use have been conducted in the Caribbean in Jamaica, Puerto Rico,

Trinidad-Tobago and the Dominican Republic. Information also can be

found from study areas in New Zealand, Australia, Yugoslavia, Kenya,

Tanzania, Zimbabwe, Malaysia, the Philippines, Costa Rica, Guatemala,

and Colombia. These span a wide spectrum of environmental and

economic conditions, but each has in common some combination of

topographic, climatic, cultural and cropping system characteristics

with the upland forest and farming areas of the Caribbean. The

methods and materials used for watershed monitoring and other data

collecting activities offer proven alternatives to the more

sophisticated, expensive experiments in the United States and other

developed nations (Pereira, 1973).










Watershed and sedimentation studies


Erosion, sedimentation and land use research in tropical

hillslope lands falls into two major categories. The watershed

studies focus primarily on the interaction of climate, topography, and

land use throughout the drainage area in determining river regimes and

erosion and sedimentation rates. By contrast, reservoir sedimentation

studies focus on the identification of sediment source areas as well

as on the immediate protection of reservoir facilities, which often

presupposes an emphasis on the development of physical infrastructure

at various points throughout the watershed.

Representative basins and experimental catchments. Studies

initiated under the auspices of the International Geophysical Year

(IGY) and the International Hydrological Decade account for a large

proportion of the work conducted in the tropics. The representative

basin studies emphasize comparative description of diverse river

schemes and watershed ecosystems on a global basis, while experimental

catchment studies focus more attention on the effects of alternative

land cover and land treatment in a given area.

Among the more important findings to date are the contrasting

characteristics of tropical climate and hydrology when compared to the

more temperate regions. Water balance data, including ratios of

runoff and evapotranspiration to total precipitation, are available in

reports from empirical studies (Golley, 1972; Holdridge, 1967, 1982;

Odum 1970b; Pereira, 1973; Thornthwaite and Mather, 1959). The

seasonal distribution as well as the amount of precipitation varies

substantially from the pattern of temperate areas. Bimodal rainfall









and river discharge distributions are common. The proportion of total

precipitation that leaves the watershed as evapotranspiration is

higher than in temperate areas and the ratio of surface water

discharge to total precipitation generally is lower, except when

deforestation occurs. The overall amount as well as the intensity of

rainfall usually is higher, making the erosive potential greater than

in most areas of temperate or cooler climates. High erosive potential

of climate often coincides with high erodibility of soils in the

seasonally dry tropics (Elwell, 1979a). The international comparative

statistics on river sedimentation bear out the implications of high

erosivity combined with readily eroded soils and high population

densities in such areas (Douglas, 1968; Holeman, 1968; Stoddard,

1965). Measurements of sediment yield in several catchments in

Malaysia demonstrate the relationship of land use to sediment

transport (Table A-6).

In addition to providing baseline information, the representative

basin and experimental catchment research tests various methods,

materials and modelling strategies. Experimental catchment studies in

New Zealand (Campbell, 1962; New Zealand Ministry of Works, 1968a,

1968b, 1968c, 1970) and Australia (Australian Water Resources Council,

1969; Chapman and Dunin, 1975) emphasize methodology and techniques,

from mapping of sediment sources (Mosley, 1980) and evaluation of

suspended sediment data (Campbell, 1962) to calibration of catchment

models (Mein, 1977; Wood and Sutherland, 1970). Multiple catchment

experiments in New Zealand (New Zealand Ministry of Works, 1970)

provide examples for measuring the impact of farming, forestry, and

range management practices on sediment yield in hilly terrain.









Reports from Kenya, Tanzania, and Uganda (Blackie, 1972; Pereira,

1973; Pereira et al., 1962, 1967; Rapp, 1977) present useful data for

comparison with some of the land use systems of the Caribbean,

including coffee and other cash crop plantations, grazing, subsistence

farming and forestry. Paired watershed studies spanning four years or

more demonstrated the impact of both land cover and specific

management practices on runoff and erosion as well as on the harvest

within the watershed. In all of the cases cited, researchers

collected frequent stream discharge and precipitation measurements.

Some cases also include continuous monitoring of the above, as well as

sampling of suspended sediments in streamflow. The results of grazing

and range improvement trials in experimental watersheds in Uganda

include a more than twofold increase in depth of penetration of

rainfall into the soil, and a concurrent reduction in peak streamflows

after restoration of overgrazed grasslands (Pereira et al., 1962). -

Data from experimental sites in Kenya (Blackie, 1972; Pereira, 1973)

document the effects of replacing tall evergreen forest with tea

plantations. The mean water yields over an 11-year period

effectively were equal for the forested control watershed and an

adjacent area planted in tea. The floods resulting from peak storm

events, however, varied substantially. The minimal impact of the tea

plantation reflects in part the stringent conservation measures

employed during its establishment. By contrast, clearing of

indigenous bamboo forest without immediate replacement by tree crops

increased streamflow 16% (Blackie, 1972). In Tanzania, a cleared

forest planted to a maize and vegetable single-crop system yielded a









50% increase in runoff, measured as streamflow (Edwards and Blackie,

1975; Edwards, 1977).

Findings from two other study areas in Tanzania, one a cultivated

montane catchment and the other a series of catchments in semi-arid

savanna, offer comparative data on river regimes as well as on

suspended sediment concentrations under changing land use practices

(Rapp, 1977). The catchments represent a complex mosaic of forest,

farm, pasture and bush, which is comparable to many upland catchments

in the Caribbean. During flood peaks, sediment concentrations ranged

from 2000 to 3500 mg L in the upland areas, and from 15,000 to
-i
75,000 mg L in the semi-arid catchments. Flash floods and high

sediment loads in both the montane and savanna areas were attributed

to land use. A comparison of the relationship between drainage area

and sediment delivery ratio in the United States and Tanzania shows

much less reduction in sediment yield with increased drainage area in

the Tanzanian catchments.

Total streamflow and sediment yield were determined by the use of

flow duration and sediment rating curves along with stream gauge and

sediment concentration data. Sampling was carried out with a home-

made point-integrating hand operated sampler (Nilsson, 1969; Rapp,

1977) and an automatic multi-stage sampler designed for ephemeral

streams. Both the instruments and the methods of analysis used in

this study are feasible for use in the Caribbean.

Reservoir sedimentation. Studies of reservoir sedimentation and

other aspects of regional water utilization and management often

approach the situation as an engineering and economic problem, either









in terms of reservoir design or in subsequent maintenance of the

completed structure and the watershed. Many studies of this type have

been carried out in Latin America (Casco de Aviles, 1979; Crosson and

Frederick, 1977; Farvar and Milton, 1973; Rabinovitch, 1979; United

Nations, 1979) including several studies in the Caribbean (CDE, 1981;

IBRD, 1972; Lahmeyer, 1967; McHenry and Hawks, 1966; Noll, 1953).

Most of these works analyze the feasibility of proposed

impoundments or document sedimentation rates in existing structures.

The major hydroelectric and irrigation projects planned or under

construction in the Caribbean lack empirical data on sediment delivery

by watershed subdivisions. Measures of erosion rates within the

watershed also are seldom included. Soil conservation programs, when

they exist, usually are initiated in response to reservoir

sedimentation problems (Floyd, 1969; Gomez, 1980; Paulet, 1980;

Rocheleau, 1980; Santos, 1981; SEA, 1978; Vasquez, 1980).

Several reservoirs are in danger of filling up in half the time

projected for the useful life of the structure (CDE, 1981; McHenry and

Hawks, 1966; Noll, 1953; INDRHI, 1981), and some reservoirs already

require expensive dredging procedures to continue functioning (CDE,

1981). Serious erosion problems are related to land use in the upper

watersheds as well as in the immediate vicinity of impoundments

(Antonini et al., 1975; Carmona, 1980; Cepeda, 1980; CDE, 1981; de

Leon, 1980).


Erosion rates at the farm and plot level


While experimental watershed research has concentrated primarily

on undisturbed forested areas, the impact of deforestation and









resultant reservoir sedimentation, erosion studies in diverse

hillslope environments throughout the world have documented soil

losses at the scale of individual farms or small plots. These results

are available for varying crop types, rotation and conservation

practices as well as for a wide range of natural environmental

conditions.

Research on erosion rates in the Caribbean and similar regions

consists primarily of experiments in standard runoff and erosion plots

under various land covers and management practices. Methods tested

and adapted in similar environments offer alternatives to the more

expensive and time consuming instrumentation often applied. An

adaptation of the standard erosion plot with sediment and runoff

collectors tested in hillslope experimental fields in Yugoslavia

provides a simple design for applications in other studies (Djorovic,

1977). Dunne (1977) describes simple erosion plot designs and

summarizes several inexpensive techniques for erosion measurement

without the use of standard plot structures. The Gerlach trough

(Gerlach, 1967) is easy to install and to use for soil loss and runoff

measurements (Morgan, 1979). It has been employed successfully in New

Zealand (Soons and Rainer, 1968), the Philippines (Kellman, 1969),

Israel (Yair, 1972), the United Kingdom (Morgan, 1977), and the

Carpathian Mountains (Gerlach, 1967).

Results from erosion studies in the Caribbean and Latin America.

Results from the Caribbean and neighboring Latin American nations

consistently show that erosion rates vary significantly (up to three

orders of magnitude) with changes in vegetation cover and management









practice (Table A-7) (Ahmad and Breckner, 1973; Barnett et al., 1972;

Bertoni, 1966; Marques et al., 1961; Noll, 1953; Rocha, 1977; Sheng,

1973; Sheng and Michaelson, 1973; Smith and Abruna, 1955; Suarez de

Castro, 1952; Suarez de Castro and Rodriguez, 1955, 1962; Vincente-

Chandler, 1976; Uribe, 1966). Clean-tilled fallow consistently

yielded soil losses in the range of 100 to 200 tons ha-1 yr- 1, while

land in annual crops lost approximately 20% of that amount, and

pasture land lost 10% of the cropland losses and about 2% of the

losses sustained in bare fallow plots. Undisturbed forest loses 500

to 1000 times less than the clean-till control plots (Sheng and

Michaelson, 1973). Results from Colombian coffee plantations indicate

fairly low rates of erosion, varying between the ranges common for

forest and field crops, depending on the age of the stand, the method

of establishment, and management practices (Suarez de Castro and

Rodriguez, 1955, 1962). The relative importance of farming practice

is also illustrated by the five- to ten-fold decreases in erosion

reported for various conservation practices tested in hillslope food

crop plots in Jamaica (Sheng and Michaelson, 1973).

Research on erosion potential in the Caribbean is based upon the

factors in the USLE. Studies conducted in Puerto Rico and the

Dominican Republic have defined the erosive potential of rainfall (R)

(Paulet, 1978), the erodibility of soils (K) (Barnett et al., 1972;

Bonnett and Lugo-Lopez, 1950; Lugo-Lopez, 1969) and the cropping

factor (C) (Santana, 1980) for parts of the region, but the

applicability of the USLE to these areas is questionable (Barnett et

al., 1972). More field data collection is needed on erosion and








runoff and their relationships to land use within a variety of land

use and ecosystem types throughout the Caribbean.

Results of erosion studies in similar environments. Basic data

on erosion rates are scarce in Latin America and the Caribbean, in

comparison with the humid tropics of Asia and Africa (Lal, 1977b). A

brief summary of selected erosion studies in these regions provides a

broader frame of reference for work already completed or in progress

in the Caribbean.

Many of the crops, the small farm technology and some elements of

the natural ecosystems of West Africa bear a strong resemblance to

parts of the Caribbean. Reports of experiments conducted by Lal et

al., (1979) and others at the International Institute for Tropical

Agriculture (IITA) in Ibadan, Nigeria, indicate a clear relationship

between vegetation cover and erosion rates, with results approximating

those from the Caribbean. Losses from clean-tilled fallow range from
-l -l -l -I
11 tons ha yr at 1% slope to 230 tons ha yr at 15% slope. On

the average, soil erosion varies much more with slope than does runoff

(Lal, 1977a).

Experiments in Senegal (Charreau, 1972) on gentler slopes in the

savanna demonstrate a similar range of soil loss as from the medium

slopes (10%) studied by Lal (1977b). Runoff in cropped plots exceeds

that in natural bush by a factor of 20 to 35 depending upon the crop,

while soil loss increased 30 to 50 times for the same crops as

compared to natural vegetation (Table A-8). Similar results are

reported for other sites in Ivory Coast and Upper Volta.

Lal (1977b) tested the effectiveness of various types of mulch as

well as several variations of minimum tillage. While mulch had less










effect on runoff, it stopped soil erosion, even on the 15% slopes

(Lal, 1977b). Experiments in Nigeria demonstrated the effectiveness

of alternative methods of field preparation and planting. While

croplands with ridges oriented downslope yielded 28% runoff and 20

tons ha1 of soil loss, alternate tied ridges across the slope reduced
-1 -I
runoff to 13% and soil losses to 6 tons ha yr (Kowal, 1970).

One West African study reported on the continuous measurement of

erosion in the same plots over several years. Lal (1977b) found that

slope effects may be reversed after a few years. After a rapid

initial loss of the topmost layers on steep inclines the erosion rate

decreases, while gentler slopes maintain a more constant erosion rate.

This indicates the importance of documenting the land use history of

hillslope sites so as to account for the influence of past soil loss

and profile modification.

More detailed surveys of the published West African soil erosion

literature have been complied by Lal (1977a), Okigbo (1977), Fournier

(1967), and Jones and Wild (1975). Projects in progress include

minimum tillage and multiple cropping experiments in plots at IITA

(Lal et al., 1979).

There is also substantial similarity between some of the lowland

dry forest and montane ecosystems of East Africa and the Caribbean.

The farming systems have some crops and practices in common, though

fewer than in the case of West Africa.

Erosion plot studies in Uganda yield similar results to the

experiments already cited in West Africa and Latin America (Table A-9)

(Sperow and Keefer, 1975). The major difference is in the magnitude









of total soil loss under bare fallow and annual crops, which is

attributable to lower annual rainfall. The relationship between

vegetation types and soil loss, however, is the same.

Similar experiments in Tanzania and Zimbabwe showed the same

range of soil loss for annual crops, bare fallow, and pasture. Losses
-I -i
under maize in Zimbabwe varied from 4 to 10 tons ha yr (Hudson,

1957). Mosquito netting placed above the soil reduced losses on
-i -i
clean-till plots to 1.2 tons ha yr demonstrating the importance

of interception by canopy and ground cover (Hudson, 1957; Lal, 1977b).

Early erosion studies in Tanzania (Staples, 1939; Rensburg, 1955)

compared sorghum with grass cover and sorghum/grass strip cropping

(Okigbo, 1977; T-rriple, 1972). Soil losses varied between 9 and 116
-i -i
tons ha yr under sorghum, depending upon site and cultivation
-1 -I
practice. Grasslands yielded 1.2 tons ha yr and sorghum strip-
-1 -i
cropped with grass yielded from 4 to 60 tons ha yr .

Findings in Kenya confirm that infiltration is greatly reduced by

grazing (Stephens, 1971; Thomas, 1974) and that while cultivation may

increase initially, the effect is temporary. Similar results have

been observed in Tanzania with a six-fold increase in peak runoff rate

after conversion from forest to annual crops (Wrigley, 1969).

Hutchinson et al. (1958) recorded a ten-fold runoff increase in clean-

tilled land converted from natural grassland. The end result is

higher runoff and erosion rates under both grazing and cultivation

(Ahn, 1977).

A less typical, broader study of erosion in Tanzania demonstrated

the soil conservation potential of many traditional farming methods,

including mulching and intercropping of field crops with bananas as










well as other crop association and rotation schemes. The contrasting

erosion plot sites in the Uluguru highlands and the semi-arid central

plains formed part of the same catchment study mentioned above (Rapp,

1977). Although some of the findings closely resemble the results for

other African and Latin American sites (Temple, 1972), high losses

were recorded for clean-weeded coffee, exceeding losses in nearby

plots'with maize (Anderson, 1962). The somewhat atypical results in

this case demonstrate that perennials do not necessarily conserve soil

better than annuals (Ahn, 1977). A wide range of sediment yields is

also reported for tea plantations in East Africa, with differences

attributed to management variables (Othieno, 1975).

Erosion in tree crop plantations is recognized as a major

contribution to regional sediment yields in Southeast Asia, where long

experience with rubber and tea plantations has demonstrated the wide

variation due to management of canopy and ground cover as well as

tillage practices (Coulter, 1972; Edwards, 1977). While tea

plantations yielded approximately one-seventh the soil loss from bare

fallow (Hasselo and Sikurajapathy, 1965) in two cases in India (Table

A-10), the erosion in tea was double that reported for forested plots
-i -l
in Malaysia and reached rates of 40 tons ha yr in Sri Lanka prior

to implementation of conservation practices (Lal, 1977a, 1977c).

Measurements from several land use and land treatment types in

upland Mindinao in the Philippines (Table A-ll) illustrated the

relationship of both land cover and land use rotations to runoff and

erosion rates (Kellman, 1969). The plantations had relatively little









impact in comparison with logging and farming uses. The results from

long established rice and corn plots suggest cumulative

destabilization of soil structure under permanent cultivation.


Qualitative and Informal Analyses of Land Use and Erosion
in the Caribbean and Similar Environments


Development and technology transfer projects in erosion prone

areas have yielded useful information on land use and erosion

problems. A few examples of interest include: studies of erosion and

overgrazing in the Bolivian highlands (LeBaron et al., 1979); a

summary of the programs of the Yallahs Valley Land Authority in

Jamaica (Floyd, 1969); the progress reports, project summaries and

consultant reports from Plan Sierra in the Dominican Republic

(Antonini and York, 1979; Chaney and Lewis, 1980; Montero, 1979;

Swedforest, 1980), and development agency communiques on resource

management projects in Cajamarca, Peru (Nicholaides and Hildebrand,

1980b), the southwestern slopes of the Dominican Republic, the

interior of Jamaica, and Haiti (Murray, 1977; Zuvekas, 1978).

Published and mimeographed reports of the government agencies charged

with soil conservation in the Dominican Republic (Gomez, 1980; Lopez,

1980; Paulet, 1980; Russo, 1980; Vasquez, 1980), Puerto Rico, Jamaica,

Trinidad, Tobago, and other areas of the Caribbean (Henriquez, 1962)

also provide useful information.

The qualitative information gathered from such sources can help

link land management variables to the onset, severity, and persistence

of various erosion features and sedimentation problems. The

settlement and development history of a region is often indicative of










the changing rates of deforestation and the intensity of cultivation

over time. This provides a basis for relating current erosion rates

to the succession of land use systems in an area. Future trends in

land use and erosion rates can be estimated more realistically with

the aid of this type of background information.


The Role of Farming Systems Research in Soil and
Water Conservation


It is important to view the effects of land use and erosion

problems within the source areas rather than simply calculating the

net export of sediment to downstream areas (Carmona, 1980). The key

to changing the situation is to be found in the internal workings of

the upland land use systems (Quezada, 1977) and in their relationship

to the larger system (Antonini and York, 1979). Any changes in

management of the uplands must take into account the causes of current

practices and the practical feasibility of proposed technological or

land use changes (Morgan, 1979; Russo, 1980).

For example, the proposed solutions to erosion problems may

involve specific technical innovations such as mulching or terracing

(Sheng and Michaelson, 1973), or a change of land use may be

suggested. Terracing and mulching with increased crop cover

alternately have received priority in various conservation projects

with contrasting and somewhat unpredictable results. Reports from

Zimbabwe demonstrate the effectiveness of increased crop cover and

mulches (Ahn, 1977; Hudson, 1957), while climatic and farming system

constraints in Kenya make terracing a more attractive alternative

(Ahn, 1977; Thomas, 1974). In Tanzania, terracing in inappropriate










situations increased the erosion hazard from landslides (Temple and

Rapp, 1972). Both approaches have been tried in the Caribbean (Sheng

and Michaelson, 1973; Wilson, 1976) though the structural

modifications of slope predominate in projects in Jamaica and the

Dominican Republic (Bonilla, 1980; Vasquez, 1980). The advisability

of this approach is questionable.

The evaluation of proposed technological changes must be measured

against the existing practices as well as against the more drastic

option of land use change on a broad scale. For such analyses a

knowledge of erosion, sedimentation and their variation according to

crop and vegetation type will not suffice. The proper choice of

conservation measures depends upon a full understanding of farming and

related land use systems in a region.


Farming Systems, Agroecosystems and Agroforestry Research


Overview


The use of the systems approach in agricultural development

efforts is a relatively recent phenomenon (Dent and Anderson, 1971;

Duckman et al., 1974). Unlike the study of erosion and sedimentation,

the research in this field has been conducted mainly in the tropics.

Ruthenberg (1980) carried out much of the pioneer work in farming

systems, primarily in smallholder farming districts in Kenya, Tanzania

and West Africa. Most of the research in farming systems in Africa

(Ruthenberg, 1980) and Central America has focused on small, limited

resource farmers.

Studies have been conducted by interdisciplinary and often

international teams of agronomists, agricultural economists,










anthropologists and ecologists. The individual farm enterprise and

the farm household have been the preferred units of analysis. Systems

concepts have been consistently employed, although the methodology and

topical emphasis vary with the regional and disciplinary orientation

of the researchers and institutions involved. The methodological and

substantive contributions of this avenue of research provide a solid

point of departure from which to expand the treatment of ecological

aspects of the problem and to extend the scale of analysis beyond the

farm level.


Selected Examples from Africa


Most of the African research in cultivation and grazing systems

has focused on systems that represent transitions from traditional to

more commercialized forms of production. Collecting, or hunting and

gathering, are largely ignored, as is forestry. Cultivation and

grazing systems receive the greatest emphasis. Ruthenberg (1980)

divides cultivation into shifting cultivation, fallow systems

regulated by farming (pasture-crop rotations), permanent upland

cultivation, arable irrigation farming, and perennial crops. Within

grazing systems pastoral nomadism and ranching are considered.

Ruthenberg (1976) and Flinn and Lagemann (1980) analyzed resource

utilization by farmers, their impact on the quality and condition of

the resource base, and the future implications for sustained

production in the area. Carrying capacity was determined by inherent

qualities and current condition of the environment, level and type of

technology, and standard of living (Lagemann, 1977). Within this











context, Lagemann tested and criticized Boserup's theory of

agricultural innovation (Boserup, 1965) with respect to environmental

response to intensification of agriculture. The relationship of

current population densities to environmental carrying capacity under

different production systems was demonstrated by simulation models.

The simulations extrapolate current bush fallow practices (in West

African examples) to predict net environmental degradation,

diminishing net yields and decreasing production per unit labor input,

all due primarily to declining soil fertility.

Patterns of land use, spatial organization of cropping, farm

level resource management and farm level economic analyses are

emphasized within this tradition. Methods for evaluating technical

innovations for low resource farmers have also been presented in

recent studies (Flinn and Lagemann, 1980; Flinn et al., 1980). Many

of the recent studies have been conducted in Nigeria in conjunction

with the International Institute for Tropical Agriculture (IITA).

Farming systems research and extension programs also have been

developed in East Africa (Collinson, 1981).

Studies of shifting cultivation systems in West and Central

Africa indicate a potential for maintaining shifting cultivation

indefinitely at a lowered but sustained rate of production, relative

to undisturbed forest systems. Soil fertility is reduced to

approximately 75% of the value for undisturbed forest soils. The

successful attainment of adequate sustained production hinges on the

rotation cycle, which must vary between 20 to 50 years of forest

fallow per year of cultivation. The implications for carrying











capacity are clear. While the system itself may work, shifting

-2
cultivation cannot support more than about 20 to 50 persons km ,

taking into account the required fallow. Further experimental work at

IITA by Greenland, Lal, and others has explored alternatives to this

system, emphasizing soil management and conservation under bush fallow

and continuous cropping (Lal et al., 1979; Lal, 1977a) and continuous

mixed-cropping systems (Greenland, 1975; Ruthenberg, 1976).

Bioeconomic modelling has been proposed to evaluate alternative soil

conservation practices and cropping systems (Dumsday and Flinn, 1977).

Agroforestry research in Africa has combined experiments with

commercial forestry and subsistence agriculture (King, 1968). The

taungya system features mixed cropping of commercially harvested and

replanted forest tracts, with the tenant farmers caring for the

seedlings and saplings as well as their food crops over a period of

about four years (Dubois, 1979; King, 1978). The field of

agroforestry has further developed to include diagnostic and

experimental work with existing subsistence and commercial production

systems that feature some combination of trees, livestock production

and/or field crops (Brookman, 1976; Douglas and Hart, 1976; Olawoye,

1975; Parry, 1957; Raintree, 1982; Lundgren, 1982). Both cocoa and

oil palm production on small farms have been studied within this

context (Flinn, 1980; Grinnell, 1977; Lagemann, 1977; Letouzay, 1955)

as well as many traditional systems of shifting agriculture that

include management of tree crops (Dubois, 1979; King, 1968). In

general, mixed tree crop/annuals production systems are more diverse

and more stable, both in economic and ecological terms (Lagemann,

1977).










Tree crops can provide fuel wood, high protein forage, lumber,

fiber, food, mulch, and cash crops (Douglas and Hart, 1976). While

establishment of such a stand requires more capital, labor and

management than a plot of annual crops, the products are often of

higher value and can be used on the farm to fill a wide range of

subsistence needs (Lagemann, 1977). Soil fertility and structural

stability are enhanced by partial tree crop cover, providing some of

the ecological benefits of forest fallow without sacrificing economic

production (Nair, 1983). In hillslope environments the combination of

trees and row crops is particularly advantageous since tree crops

offer the soil greater protection from erosion (Douglas and Hart,

1976).

While the study of agroforestry is still relatively new, the

field is developing rapidly, in part as a response to the need to

increase small farm production in marginal lands while maintaining or

rehabilitating watersheds and forest resources (King and Chandler,

1978). Current research at the International Council for Research in

Agroforestry in Kenya (ICRAF) focuses on the elaboration of a

methodology for diagnosis of new ones. Rapid survey diagnostic

techniques and combined research/extension programs are being

developed and tested in field sites that include farm level and

community level work as well as experimental plot studies (Raintree,

1982). The multidisciplinary staff includes foresters, agricultural

economists, agronomists, anthropologists and ecologists, among others.

The international client areas for the methodology being developed by

ICRAF include the hillslope farms of the Caribbean, as well as the










Andean highlands, the Amazonian lowlands, and the African savannas,

and many other fragile and/or marginal environments under cultivation

(King and Chandler, 1978).


Central American Research



Farming systems research in Central America has been conducted at

the Center for Teaching and Research in Tropical Agriculture (CATIE)

in Costa Rica, and the Institute of Agricultural Science and

Technology (ICTA) in Guatemala. Both have sought to serve small

farmers and to modify, rather than replace, traditional systems of

agriculture (including hillslope farming).

Farming systems research conducted at CATIE has emphasized

economic and agronomic description of existing cropping systems

through community and regional level questionnaires and surveys, using

standard sampling procedures (Navarro, 1979). Experiments have

focused on mixed-crop combinations for optimization of production

given the available natural and economic inputs. The general

perspective as well as some specific techniques of ecosystems and

energy analysis have been applied (Hart and Pinchinat, 1980; Moreno,

1977).

Several precedents exist for ecosystem and energy analysis

applied to production systems, including examples from India (Revelle,

1976; Odend'hal, 1972), the United States (Burnett, 1977; Ewel, 1973),

Indonesia (Geertz, 1963, 1972), Guatemala (Carter and Snedaker, 1969),

the Dominican Republic (Werge, 1975) and elsewhere (Lugo and Snedaker,

1971; Odum, 1967). Energy budget analyses more or less independent of









the systems approach also have been conducted by researchers in the

United States (Lockeretz, 1975; Pimentel, 1973), England (Leach,

1976), and Asia (Makhijani and Poole, 1975; Smil, 1979). The energy

studies at CATIE, however, use the systems approach and focus on plant

and animal production subsystems, as opposed to the socioeconomic or

environmental subsystems (Hart, 1980; Holle, 1979). The majority of

the cropping systems research does not address questions of erosion

and watershed protection although some studies have been carried out

on erosion rates under hillslope cropping systems (Bermudez, 1979) and

under agroforestry systems (Apolo, 1979).

The application of energy analysis to individual farms (Hart,

1980) represents a distinct and particularly useful subset of the

general body of research at CATIE. The energy analysis methodology

developed by Odum (1971) and Odum and Odum (1976) has been adapted by

Hart (1980) for use in the rural farming districts of Central America.

This method can be used to identify and describe existing successful

adaptations. It can also be used to highlight the functional

relationships found in the average case, in order to choose the

critical points in the system where specific (and viable) changes will

have a major impact on total production (Fig. 5).

Hart (1980) has elaborated further a field survey methodology

using systems concepts and energy flow diagrams in rapid surveys of

rural farming communities, to characterize existing farming systems in

qualitative and quantitative terms. Still lacking are inclusion of

environmental inputs, outputs and storage, and the extrapolation from

the farm to larger scales of analysis.




















6
.
'. -
a
0
44~4 4 4 4 A


r 'wir^/pooo *aooUri











knia~ar~~ri* 1LH4







-**'T'o /ri^/fTp^~


0
z


-I.


a" aa s


o *. C.
\i a .i *
* *. 0



k f
I 3 S 1 a.






* a c s

I^ .I s "
0 0 <

0 - 3 0j a
< .4 4


A A


4 CO
3 a




S





0 0-


CO 0


U
a o
* U



C-
-4


a~upo.~ ~






twTT/T~~7~r *ri~


daB0 3BT ,poc;



poo.Q ;o *ev-..



o
0
5 .
0 -"

5- a
S -


2i

0 6
-, z


0 0 6 -0
-- 4 -"'
i.~~~~~ &a a. < fi o c

o *~ -1 0
0 0

o .a o -0o* * j 3~ ; ^S3
S ^ -S0 S ^


0
4-)


0
rz
rc









0
.JH


0
4)


m

4J-

0
I
-4


















4J
4J
0
4-J


(H
4l-)


0

4J
*r24


?00,;L










A similar approach has been applied to the environmental

subsystem in a National Science Foundation project in the same area

(Berish, 1982; Brown, 1982; Ewel, 1981). The project focuses on

nutrient cycling in small farm and successional plots, and the

research is directed toward improved management of nutrient cycles in

agroecosystems.

Agroforestry research at CATIE has combined with forestry and

watershed protection research (Budowski, 1977; Combe, 1979). The

major experiments to date emphasize the use of laurel (Cordia

alliodora) and poro (Erythrina poeppigiana) trees to provide shade,

reduce erosion, and produce firewood or lumber within coffee

plantations in montane environments (Beer, 1979; Bermudez, 1979;

Rosero and Gewald, 1979; Russo, 1983). Combinations of grazing with

plantations of alder (Alnus acuminata), Eucalyptus sp. and cypress

(Cypressus lusitanica) also have been tested in multiple field trials

(Combe, 1979; Gonzales et al., 1979). One test of the taungya system

has been conducted, using Gmelina arborea interplanted with maize and

beans (Rosero, 1979). Most of the trials have proven successful when

judged in terms of mixed ecological and economic criteria by

researchers, but widespread adoption by small farmers has not yet

occurred beyond the original centers of innovation.

The planting of nitrogen-fixing leguminous trees in or around

pastures and within coffee and cocoa plantations is common practice in

many parts of Central America, but the density and distribution of the

trees often is very limited. Moreover, the need for the introduction

and/or development of agroforestry systems often is greatest in the










marginal lands where small farmers are dedicated exclusively to

cultivation of annual crops in slash and burn or bush fallow rotations

(Lagemann, 1978).

The farming systems research conducted at the Institute of

Agricultural Science and Technology (ICTA) in Guatemala, concentrated

more on the socioeconomic subsystem and on the integration of research

and extension (Reiche and Lee, 1978). Extensive contact with

hillslope farmers in Guatemala's rugged uplands also produced some

work related to erosion control (Hildebrand, 1981).

The most significant result of the ICTA research was the

development of a practical multidisciplinary approach that merges

research and extension efforts in farming systems. Farming systems

research and extension (FSR/E) integrates farmers, researchers and

extensionists into an effective group to identify and solve problems

and promote dissemination of the -solutions (Hildebrand, 1981; Gladwin,

1981). This is reflected in the rapid survey technique (sondeo)

developed for use by interdisciplinary research and extension teams

with small, low-technology farmers (Swisher et al., 1982).

The survey itself consists of a series of informal interviews

conducted by an interdisciplinary team that includes a social

scientist (usually an anthropologist), an agronomist or related

technical specialist in agriculture, and an economist (Hildebrand,

1981). The approach borrows heavily from traditional anthropological

field methods with reliance on key informants and on corroboration of

information from various sources. The use of open-ended interviews

with chosen informants contrasts with the random sampling of the










population so often used with questionnaires or structured interviews.

The latter, more formal approach is a more common and preferred form

of data gathering in many disciplines. However, it is also expensive

and time-consuming, and it presupposes an accurate population census

or property survey on which to base the sample. Formal surveys and

questionnaires also inherently limit the categories of information to

be treated. Little room is left for the definition in the field of

problems not already recognized prior to the survey. Opportunities to

explore the unique relationship between various aspects of a problem

in an area are also constrained by the format. The major

considerations in using the sondeo technique are quality versus

quantity of information and cost of the survey versus useful

information obtained.

The sondeo provides a practical and effective means of

reconnaissance and data gathering. It also lays the groundwork for

future extension programs in the area. During the intensive one-to-

two-week survey, the knowledge and needs of the farmer are

incorporated into the design (form and content) of subsequent on-farm

research projects. This methodology relies heavily on the judgement

of the research team, the local populace, and individual farmers to

define farming systems and their problems and to choose representative

or exceptional cases for further study, according to the goals of each

project (Hildebrand, 1981).

The experiments themselves are on-farm trials which may or may

not be replicated at experimental stations. The sondeo and the field

trials are supplemented by farm record-keeping. A family member keeps










an account of the farm's inputs and outputs as well as of activities

and movement of materials within the farm itself. These records also

assist in the evaluation of on-farm trial results and provide basic

data for further trials and/or discussions with farmers. The success

of a new technology is judged at least in part by the farmer's

perception of its performance and by its subsequent adoption by him

and other farmers in the area. This indicates t6 some extent the

"fitness" of a technology for the farming system as a whole, at the

farm level (Swisher et al., 1982).

The sondeo as well as the subsequent on-farm trials and farm

record-keeping of the FSR/E approach are readily adaptable to soil and

water conservation research in hillslope environments in the

Caribbean. The major elements lacking are greater attention to

natural resource management at the farm level and a method for

predicting and evaluating the success of a given technology at the

watershed or regional level.















CHAPTER III
METHODOLOGY


The General Approach



The methods reported in the literature review include a variety

of techniques that can be incorporated into a consistent and

appropriate methodology. The end product, however, must be more than

a method or a collection of techniques. The methodology proposed and

tested in this study is an adaptation of the scientific method in

general, and systems analysis in particular, to the interdisciplinary

study and treatment of the problems of land use, erosion, and

sedimentation in the underdeveloped nations of the tropics. The

research approach combined elements of farming system and ecosystem

analysis, drawing most heavily on the work of Odum (1971, 1982),

Antonini et al. (1975), Bormann and Likens (1979), Hildebrand (1981),

and Hart (1980), all described in greater detail in Chapter II.

The research tested specific management-related hypotheses under

field conditions. The cases studied required immediate action based

on tentative solutions from experimental results, prior to extensive

repetition and replication. The general research model included

direct outputs from verification to policy and production sectors and

a feedback from verification to further experimentation (Fig. 3).

The feedbacks in the research program imply an ongoing process of

learning. The time constraints in applied research were handled by

using these feedback loops to continually test and modify the










tentative solutions already proposed. This iterative approach has

already been tested in farming systems research and extension programs

(Hildebrand, 1981).

The usual concept of applied research is one of a finite activity

to be carried out and completed by specialists, after which they will

offer a set of definitive conclusions to be implemented. In this

case, the study area was viewed as a system in a flux, constantly

adjusting to changes in internal and boundary conditions. The object

of study in this case also had a subjective component. The role of

people in determining the direction of ecosystem evolution was taken

into account. Residents of the region contributed to the

investigation as both informants and participants in data gathering,

experimental and verification procedures. The researcher participated

in an on-going experiment in which people living in the area sought

short-term relief and long-term solutions to problems at least

partially perceived and defined by them.

The study was designed to accommodate the distinct priorities and

information needs of local clients, scientists, and the regional

policy sector. The experimental design and data analyses tested

multi-faceted hypotheses concerning technology and land use

alternatives for the region. Each experiment included: a primary

hypothesis as to the biophysical or economic performance of a

particular alternative; a secondary hypothesis concerning how the

proposed change would fit into the existing system; and a third

hypothesis that the system, as such, could and should be sustained,

with modifications.










The primary hypotheses were tested and judged jointly by researchers

and farmers, according to objective, quantitative criteria. The

secondary postulate was tested and judged by the farmer according to

subjective criteria, based on overall practical performance. Researchers

provided a posteriori explanation and interpretation of the farmers'

experience. The value judgement as to the fitness of the existing system

was considered the joint prerogative of the local residents and the

policy sector (clients) while the researcher determined the system's

sustainability by ecological analysis of current trends.


Materials and Methods


The study was conducted at four scales of analysis within a nested

hierarchy of spatial units, including: the Plan Sierra impact area (2500

2 2 2
km ); watersheds of 500 to 100 km ; small watersheds of 1 to 20 km ; and

individual landholdings ranging from 0.5 to 500 ha.

Chronologically the research activities were grouped into three

phases of increasingly finer levels of resolution and greater detail of

analysis. The regional reconnaissance and refinement of problem

definition was followed by detailed characterization of the study sites

at the watershed and plot level. The third phase consisted of monitoring

runoff, erosion, sedimentation, and production under varying land use and

soil conservation practices (Fig. 4).

The conceptual model of the Caribbean region (Fig. 1) formed the

basis for the overall research design, while at each successive level of

resolution a conceptual model was proposed, evaluated through field data

collection, and refined or modified based on empirical evidence and

testing of specific hypotheses inherent in the model.










Regional Reconnaissance


The inventory of existing land use systems and the condition of

soil and water resources within the Plan Sierra impact area provided

the data for refinement of the problem definition and for subsequent

application of the research design within successively smaller units

of analysis.

Most of the reconnaissance activities were completed between

January and March 1980. A regional description and summary of land

use systems was complied from library and field research. A review

and synthesis of maps, aerial photographs, statistical summaries and

literature relating to the study area preceded the field surveys. The

area was stratified into multitopic subregions based on cartographic

analysis of physical, biotic and socioeconomic characteristics mapped

at scales of 1:250,000 and 1:50,000. The major criteria for zonation

were life zones (Holdridge, 1967; OAS, 1967), topographic

characteristics, and land use characteristics, the latter reflecting

population density as well as condition and productivity of the land.

Maps were compiled by Plan Sierra cartographic and project staff from

topographic and thematic maps at 1:250,000 and 1:50,000 (Jennings,

1979a; OAS, 1967; Swedforest, 1980), and from aerial photographs at

1:20,000.

The field survey was similar to the general procedure outlined by

Hildebrand (1981). Several rapid reconnaissance surveys of erosion

features, land cover and land use systems were conducted within the

regional subdivisions outlined above, as part of Plan Sierra program

development in soil and water conservation. Survey teams varied in











composition, but usually included the author and one to four

specialists and paraprofessionals in engineering, forestry, agronomy,

and soil conservation.

The surveys included formal and informal interviews with

residents, as well as field mapping of land cover, land use and

evidence of erosion and sedimentation in the various areas visited.

The selection of sites for more detailed observation reflected a

strong reliance on the knowledge of agronomists, foresters and

conservationists already familiar with the area, as well as the

opinion of residents as to what areas constituted typical or extreme

examples of particular physical and socioeconomic characteristics.

Field survey records included numerical data, maps, and interview with

residents, as well as the impressions and observations of the survey

team.

A synthesis of the cumulative results of prior field

reconnaissance by interdisciplinary teams of consultants (Chaney and

Lewis, 1980; Georges, 1981; Hart, 1981; Montero et al., 1981; Navarro,

1981; Nicholaides and Hildebrand, 1980b; Safa and Gladwin, 1981;

Santos, 1981; Swedforest, 1980) supplements the information gathered

from the author's survey. Written reports and personal communication

from the consultants and visiting scientists contributed to updates of

the regional profile.


Refinement of the Research Design


Based on the reviews of regional information and the completed

field reconnaissance, the conceptual model of the region was modified










to include relationships previously omitted or incorrectly defined.

The research design then was developed to test the major hypotheses

implied in the model. The discharge and sediment yield of two large

watersheds were measured over a 15-month study period. During the

same interval subwatersheds were described and monitored in greater

detail to relate differences in discharge and sediment yield to

varying physical and land use characteristics. Erosion plots

constructed within the subwatersheds provided comparative data on

runoff, erosion and production under different land uses, each with

varying conservation practices.

The 18-month period of study for phases two and three extended

from 1 Apr. 1980 to 30 June 1981. This period included a full

hydrologic year, from 1 Apr. 1980 to 1 Apr. 1981, and also allowed

repetition of sampling and monitoring during the time of peak

rainfall, from April through June.

The spatial and logistical organization of research activities

are illustrated in diagram and tabular form (Fig. 6, Table 1). The

chronological order of analytical and data collection procedures

parallels the general case described in Fig. 4. The choice of study

sites reflects the insights gained from the review and reconnaissance

survey, as well as the information needs of Plan Sierra.


Study Sites


The study sites selected within the Plan Sierra impact region

included two large watersheds (500 to 100 km 2), five small watersheds

(1 to 20 km2) nested within the two larger units, and 16 plots on nine

landholdings situated within three of the small watersheds.



















-f


.t


L.



*-i








= S:


I.. -



A

1k


.7


r~.


:r.





~ S


'I *f.
I-





~f.


I















Cn U)
0 0


S0 0 m 0
U0 0 ( )-
W 00
0 0) 0) 0
U - H 4-1 -Ii


0) )0 -





0
0 0
0 0 0
01 ri 0,



0H d 0 0
I>C 0 f










o 0



0 411 ( 03
-4-I


03





0 + O (:DO 0 0C 0O OO
r-i f0 c I















(3) 4-)
ro 4
0 -') 00)




0 M
0) 0)
0; 44- )4-)
4-I o- 0) i 0
0 0 0 0 0) :00
..-i >-4




0) > d-^ 0)
o u ~~~~ro a nt
04 4 U 44)


0 0U 3
T3 O : 0)(UO +
C4 0) >- o)


>-I 4.44-40)4J 0d 4-) 4)J 4
0 L-i i- tu 0) U) M U) t 0
V0 0 0 0 0-I 0 4-3.H.-4(T
zC 4-) f04 41


0 0

0 I- *0

iCU 0
U r 0 i-CN 44 4-4 ) 4 d 4 W -
0 0)40 4- CO 0 (n z U) 0
r"- a) 0-4 0
r. 4C U 4M 4C4P L
0) U)




- 4-4
4(t 0
$4 4) m I: U, o I,
0 Q0 O0


4- 0 0 *- -
n> -r-I 'O ^ ^ a. -



0) < 0 4-r
.- 00
4- 0 0 000000
(U C ;0 0~ r--l


4 0 0 (V
aJ) zr-.H 0 0 -
0 00 000
r-9 (1) (1 ro ro ra ra ro ra ra ra
43U) rq -H r. z z z z z z r- 0
(3 W 0 0 0 0 0 0 0 0 z
P4 P4 M~ = = = = = = = +-










The criteria for selection of the large watersheds for further

study were as follows:

1. Availability of historical data on climate and river
discharge;

2. Current coverage of the area by rain gauges and water level
recorders;

3. Need for erosion and sedimentation data in the area; and

4. Presence of Plan Sierra personnel and facilities in the
region.

The Plan Sierra region is drained by three major rivers, Mao, Bao, and

Amina. All three watersheds are partially contained within the impact

area and empty into the Yaque del Norte River. Each river is a

current or future source of water for hydroelectric and irrigation

projects planned within the borders of the study area. The study

concentrated on the Mao and Amina Rivers.

Within each of the two large watersheds, subwatersheds of

critical concern were selected based on the land use systems, erosion

features, climate, soil, and topography. Ease of access and presence

of soil conservation personnel also played a part in the decision.

Five small watersheds were chosen to test the following

hypotheses: 1) that land use is the major determinant of erosion and

sedimentation in the region; and 2) that the combination of pasture

and field crops common in the densely populated central region

contributes more heavily to sedimentation than land use practices

based on coffee which is more typical in the highlands. The size of

the watersheds ranged from 1 to 30 km with two replicates in the

2
first size category (1-2 km ) and three in the second group (10-30

km2). The examples represented the smallest scale of analysis in
km ). The examples represented the smallest scale of analysis in










which the land use matrices characteristic of the given subregion

could be measured. In each case a cluster of settlements was

included, along with a combination of subsistence and commercial land

uses. This unit of analysis also was compatible with community level

socioeconomic analyses conducted in the region (Georges, 1982).

The smallest scale of analysis focused on landholdings associated

with individual households. Sixteen experimental plots were installed

on property belonging to nine separate households, all situated within

the watershed study areas mentioned above. As depicted in the

research design (Fig. 6, Table 1), two distinct physical subregions

and four major land use types were represented in the three clusters

of plots. The plots were chosen to reflect average characteristics of

microclimate, slope, soil and land use within each watershed. The

first three variables were held constant, while land use was varied.

Nested within the comparison of land uses was a comparison of soil

conservation practices in plots planted to annual crops and a

comparison of stages of development in coffee and forest plots.


Characterization of Watersheds and Plots Selected for Further Study


Large watersheds


The portions of the watersheds included within the study were

delineated and measured on 1:50,000 scale topographic maps (U.S. Army

Map Service, 1962). Cartographic analysis of drainage networks,

geologic subregions, life zones, topography and land use indicated

areas whose composite characteristics favor high rates of runoff,

erosion and sedimentation.










The choice of sampling sites for more detailed study of discharge

and sedimentation was based on several criteria, including ease of

access, proximity to homes of observers, established hydrologic

monitoring sites, proposed dam sites, and regularity of the

longitudinal and cross-sectional profiles of the channel in the

vicinity of the potential sampling points. Profiles of the river

cross-sections were surveyed. The relative heights of the bridges and

other large structures along the stream bank also were measured to

provide reference points for reporting maximum flood stages. The

profiles are included in Appendix B.


Small watersheds


The small subwatersheds selected for further study were analyzed

by standard cartographic and photogrammetric methods to determine

total area, average slope, and area and distribution of land cover

types. Land use and socioeconomic characteristics of the watersheds

and their settlements were determined by field observation,

interviews, and a review of statistics available at the community

level.

The average slope was determined by the Wentworth (1930) method,

using topographic maps at the 1:50,000 scale. Each watershed was

mapped separately and overlain with a 1-km grid subdivided into four

cells each, with two diagonal cross sections per grid. The slope from

the center toward each corner was determined by the change in

elevation over the four 0.5-km transects. The average slope (%) was

calculated for each small grid cell (0.25 kg2), then for each larger

cell (1 km 2), and finally for the watershed as a whole.









Total area was determined by planimetry. Area and distribution

of land cover types were determined in a three step process, beginning

with the interpretation of black and white aerial photographs at the

1:20,000 scale, taken in January and March 1980. The watershed

boundaries, the stream, its tributaries (if any) and ma3or features

such as roads and paths were delineated on outline maps enlarged to

the same scale, as well as on the aerial photographs. Land cover

subdivisions were outlined on the photographs, based on differences in

color, texture, and pattern. In many cases the fine texture of land

cover subdivisions required aggregation into units of mixed land cover

recognizable by plot or property boundaries. The land cover units

previously outlined were classified according to land use, based on

prior field observations in the study area during the reconnaissance

survey. After transfer to a map of known scale and projection, the

areas of major land use types were measured by planimetry.

The incidence, type and distribution of erosion features in each

small watershed were noted during the initial field visits and during

the 15-month period of study. The number, severity and distribution

of landslips and landslides, gullies, and signs of rill and sheet

erosion were compared between watersheds included in the study as well

as with several other watersheds visited regularly in the course of

related studies. Residents often were questioned about new features,

or noticeable changes in pre-existing gullies, rills and landslides.

Whenever possible, the immediate causes, such as a particular storm

event, agricultural and construction activity, were identified. This

type of qualitative analysis provided a background for further









refinement of the research design (Mosley, 1980) and for more informed

interpretation of subsequent results.

Detailed analyses of channel erosion, deposition, and range of

flood stages were conducted during field inspection of the stream

courses. Variations in depth and areal extent of sediment deposits

gave some indication of sedimentation rates within the basins. The

general form of channel terraces and floodplains, as well as the

height of residual debris lodged in trees and rocks on the stream

banks, provided indicators of peak annual flood stages. The specific

topographic indicators included erosion features on the stream banks

and breaks in slope along the cross-sectional profile. The cross-

sections were surveyed by the stadia transit method (see Appendix B).

Estimates of peak floods for longer periods were obtained from

interviews with elderly residents of the area. Independent

questioning of several persons established the margin of error in the

responses.

The sampling and monitoring locations within each watershed were

chosen according to several criteria:

1. Situation relative to settlements and the land use to be
evaluated;

2. Ease of access;

3. Security of equipment from vandalism

4. Availability of natural footings and fastenings against
flash floods and transported debris; and

5. Regularity of channel longitudinal and cross-sectional
profiles in the vicinity.

Interviews with residents of the communities within the

watersheds were conducted at local meeting places, in the field and in









the homes of agricultural laborers and landholders. The Plan Sierra

soil conservationists living in the area also were interviewed, and

they in turn questioned residents about the settlement history, land

use and farming practice, past and present production levels, and

sources of income. Discussions with two anthropologists conducting

land use and migration research in the region also provided valuable

information and insights into the character of the communities in the

study areas (Pessar, 1981; Georges, 1982). Plan Sierra social

workers, agronomists and foresters familiar with the area of interest

also contributed to the socioeconomic profile of the small watersheds.


Description of individual landholdings


Selection of sites for measurement of runoff and erosion in

experimental plots was based on uniformity and degree of slope, as

well as type of land use, management practice and easy access for

construction and sampling purposes. Wherever possible, replicates of

land use and treatment were established within the same watershed and

also in another watershed to determine the margin of error in

measurement and to compare the relative difference in runoff and

sediment yield under varying conditions of site and land use.

The experimental design further subdivided the categories of

forest, pasture, coffee and annual crops to compare undisturbed and

secondary forest, new and established coffee stands, different types

of annual crops, and use of minimum tillage and hillside ditches for

erosion control in fields planted to annual crops. These subsets of

land use type were tested in paired plots in the same or adjacent land










holdings to guarantee duplication of all other conditions except the

variables of interest. Individual plots were chosen based on field

observation of the above mentioned criteria. The choices were

confirmed after consultation with agronomists, conservation personnel

and residents of the area to determine if the plot in question

constituted a representative example of management relative to the

surrounding watershed.

Formal and informal interviews with owners, residents, neighbors

and local Plan Sierra personnel served to outline the settlement, land

use and production history and the variability of natural conditions

for the individual plots. Detailed information on past and present

crop associations, rotations, yields, labor and material inputs, and

ratios of commercial to subsistence production came from intensive

interviews with the persons directly responsible for management of the

site for a period of 10 years or more. These interviews often spanned

two or three visits by one or more members of the research team. The-

format was open-ended to allow the participants to elaborate on their

experiences.

Farmers were encouraged to discuss their problems with respect to

subsistence and commercial production and to volunteer insights and

judgements as to potential solutions. The women and children working

at each site also were interviewed, usually on separate occasions, to

obtain accounts of their roles in production and natural resource

management as members of the farm household. Their perceptions of

problems and suggestions for changes also were solicited.

The baseline information to be obtained was outlined in diagram

form (Fig. 5). This helped the interviewers to keep track of the









subject matter covered. It also provided a convenient format for

recording and summarizing responses during or following the

discussion. Information noted on the diagrams and tables served to

evaluate the farm level models prior to initiating the erosion plot

experiments and watershed monitoring activities.

All sites were chosen to reflect variation in land use and

treatment, while slope and soil conditions were held constant and as

close as possible to the average for the watershed. Slope

measurements along the downslope transect were made prior to final

siting of all experimental plots.

Soil profile descriptions, characterization of soil samples by

laboratory analysis, and taxonomic classification constituted part of

the site description at each plot. Rectangular trenches at least 1 m

deep, 1 m long and 0.5 m wide were cut for observation and sampling.

Measurement and description of profile stratification, with detailed

description of color, texture, structure, and uniformity, by horizon,

were carried out according to the procedures outlined in the Soil

Survey Manual (USDA, 1951).

Munsell color charts were used for wet and dry color

determinations in the field (Munsell Color Co., 1951). Laboratory

analyses for N, P, K, and organic matter content followed standard

methods for determination of Kjeldahl N and Bray P by colorimetry, and

for determination of K by atomic absorption (USDA, 1975).

The North District Research Laboratory (CENDA) of the State

Secretariat of Agriculture conducted all laboratory tests for the

project, including physical and chemical characterization of soil









samples. Soil classifications were confirmed and refined by soil

survey specialists from the Secretariat's South District Laboratory.

The relative infiltration rates of soils at the various sites

were determined by measurements with ring infiltrometers (Gregory and

Walling, 1973; Wisler and Brater, 1959). The inner ring was cut to a

25 cm diameter and the outer ring measured 40 cm across. After

placement in the ground with a minimum of soil displacement, the outer

ring was filled to form a barrier of saturated soil around the inner

ring which was filled to a 10-cm depth. Throughout the next 4 hours a

nearly constant head of 10 cm was maintained while measurements of

water added were recorded at increasingly longer intervals. The form

used for the field measurements is included in Appendix H.

The frequency, distribution and severity of erosion features on

and around the plot sites were observed and noted prior to

construction of experimental plots. Wherever possible, the

developmental sequence of such features was determined from accounts

by the residents or neighbors and from repeated observation and

photographic records kept over the 15-month study period.


Detailed Analysis and Measurement of Key Parameters


The full characterization of the study areas at all three scales

of analysis served as a point of departure for the third phase of the

study, the measurement of water and soil exports from the individual

plots and from the nested sets of watersheds. Erosion, runoff, and

sediment transport were related to daily and continuous precipitation

measured at 15 stations in and near the Plan Sierra impact area.









In the large watersheds total sediment transport was measured and

sediment yield was calculated in order to estimate the magnitude of

the erosion problem on the watershed, to predict the future

sedimentation rates of the proposed dams, and to compare the losses

per unit area between the study areas and other sites for which

sediment yields have been reported. The measurement in the small

watersheds showed the integrated effects of land use and physical

characteristics in each study site. Subsequent tests of the data by

multiple analysis of variance indicated the relative influence of land

use and physical factors.

Experimental plots were included to demonstrate the impact of

varying specific crop types, land treatments and conservation

practices on erosion and runoff. The plots also provided the erosion

rate data necessary to calculate sediment delivery ratios for the

watersheds. Erosion plots were included because they allow

observation of the problem within the context of individual

landholdings and related households. The experiments were conducted

under the same conditions that limit and influence the management of

individual landholdings. While the most striking effects may be

expressed at the watershed level, the management decisions are made at

the household level. The degree to which such decisions are

constrained by pressures from the larger system does not change the

fact that these decisions directly determine, in turn, the condition

of the larger watershed. Any proposed changes must be tested within a

holistic framework at the level of the land managers.










Precipitation, discharge and sedimentation in the large watersheds


Records from 15 climatological stations collected by SEA's

Department of Meterology, and INDRHI provided daily precipitation

values as well as continuous data on rainfall amount and intensity at

five of the stations in the region. A multiple correlation analysis

of daily data from all 15 sites was performed to check for duplication

and overlap of information and to determine the variability of

rainfall over the study area (SAS, 1979).

The distribution of rainfall over each watershed and its

subdivisions was determined by the Thiessen polygon method (Wisler and

Brater, 1959). The daily rainfall values from the climatological

stations were extrapolated to the surrounding areas, then aggregated

into units more relevant to the study, such as subwatersheds. The

Thiessen polygons defining the area of influence around each station

were delineated and superimposed on a map of the Mao and Amina

watershed divisions, at a scale of 1:250,000 (Fig. 7). The watersheds

were defined by cartographic analysis of topographic maps at the

1:50,000 scale, and Thiessen polygons were constructed according to

standard methods summarized by Kenah (1980). After determining the

proportion of each subwatershed that fell within the polygon assigned

to each station, a weighted average of daily rainfall was calculated

for each subwatershed.

Stage and discharge measurements. On both rivers, hydrometric

stations maintained by INDHRI were located conveniently near the

downstream borders of the Plan Sierra impact area (Fig. 8), and

relatively close to concrete bridges. The river cross-sections were


















U)

04
0)











0


0)
m
4-)



ro














0
r.












M


0
a)






4-
0








U)
0
a)















0

0)







w
-4
0



U)




4 0


4 -H
t02


* *

J 'c-.

0 > x
"1

S 0,
c o -


E .D *
- 0 3
(5 $ cn

*




















- /
- r
- 1


4az


#I--/


-
I-










gauged and calibrated periodically, such that the stage measurements

recorded at 0700 and 1700 hrs daily, as well as maximum flood stage,

could be converted directly to discharge rates based upon a nomograph

constructed by INDRHI hydrologists (see Appendix C). The stage-

discharge graphs relate discharge rates obtained through periodic

field measurements using the area-velocity method (Grover and

Harrington, 1943; Herschy, 1978) to the simultaneous reading of river

stage on a simple gauge.

For purposes of this study the nomographs of each river were

segmented and the equations for each segment were determined by simple

linear regression using the Statpak package of statistical programs

(MUSIC, 1967). The gauge readings recorded for morning, evening, and

flood peaks during the study period were converted to discharge rates

using the appropriate equations derived from the INDRHI nomographs.

Stage measurements made at the bridges by project personnel were

compared to simultaneous measurements at the hydrometric stations. A

simple linear regression equation converted the measurements made at

the bridge at sampling time to a gauge reading for the station. This

procedure replaced independent velocity-area discharge measurements

and allowed the substitution of a simple indirect method for the more

difficult and lengthy procedure used initially.

Comparison of rainfall and discharge during the study period to

the historical period. The total monthly rainfall totals over the

study period were compared to the average monthly rainfall totals over

the full period of record at stations with 11 or more years of data.

The SAS (1979) means program was used to compare the data from each










station for 1980 and 1981 with the prior records. The comparison of

river discharge measurements during the study period and the full

period of record followed the same procedure.

Measurement of suspended sediment concentration. Direct

measurement of sediment concentration required periodic sampling of

river water over the full 15-month study period, spanning three wet

seasons. This allowed sampling under a variety of conditions from low

flow to flash floods. The timing of observations was designed to

sample as wide a range as possible of the variation in discharge and

sediment concentration. In some cases time series samples were drawn

to cover the rise and fall of a particular flood.

The samples were extracted at the bridge site with a modified,

locally built Uppsala-type sampler (Fig. 9), as described by Nilsson

(1969) and Rapp (1977). The finished product resembles the USDA-48

Wading Sampler (USDA, 1979). The diameter of the sample intake nozzle

(0.64 cm) limited the particle size to a maximum of approximately 0.3

cm, which probably did not exclude any suspended sediments. Bedload

was not sampled. The emphasis on suspended sediment load as based on

the high average ratio of suspended sediment to total sediment load

(Gregory and Walling, 1973) and on the even higher ratios reported for

turbulent conditions.

While the sampling sites were not ideal by hydrometric criteria,

the accessibility and safety of bridge crossings outweighed other

factors (Herschy, 1978). Wading for samples from steep banks at river

narrows entails undue risk, particularly in areas subject to flash

flooding. The accessibility of the sampling site also can limit
















Sediment
and water intake


outlet


Nolgene plastic
collecting bottle


Fig. 9. Uppsala-type manual sampler for instantaneous
measurement of sediment concentrations in
streams (Rapp, 1977; Nilsson, 1969; USDA, 1979).










the number of observations. The bridge sites allowed simple grab-

sampling during flood conditions.

Initial results from multiple samples stratified throughout the

river cross-sections showed little variation across the section but

pronounced differences with depth. Subsequent samples were taken in

pairs at 10 and 30 cm below the surface (water depth permitting) at

the center of the section. The samples normally were drawn by

extending the sampler downward from the upstream side of the bridge,

directing the intake into the flow, and extracting the sample just

prior to filling the bottle, to avoid sampling error.

The 1-2 L samples were stored at room temperature at field

headquarters, then shipped periodically to the CENDA laboratory.

Sediment content by dry weight was determined and reported as
-I
concentration (g L ). Analyses were conducted according to standard

methods for the determination of suspended solids in water. After the

sample volume was recorded, the samples were shaken, then poured into

Gooch crucibles lined with pre-weighed Whatman No. 5 fiberglass filter

paper. After the sample was strained into a flask, under a slight

vacuum, the filter and collected sediments were oven-dried at 105C

for 24 hrs, then weighed. The final weight minus the previously

determined paper weight gave the new weight of the sediments for a

given sample. The latter then was divided by the sample volume to

obtain the sediment concentration.

Sediment discharge rates and total sediment transport.

Instantaneous rates of sediment discharge in tons sec were
-1
calculated by multiplying total river discharge (m sec ) by the

sediment concentration (tons m 3):










-l 3 -l -3
Sq (tons sec ) = Q (m sec ) x Cs (tons m )

where Q = river discharge, Cs = concentration of sediment, and Sq =

sediment discharge rate.

Daily sediment discharge for non-flood sampling days was

calculated by averaging morning and evening discharge rates from stage

measurements. This was multiplied by a time conversion factor to

obtain total river discharge for the day. The total discharge times

the concentration approximates total sediment transport past the

sampling point for that day.

For samples drawn during or very close to short-lived peaks, the

flood peak duration was estimated from field records and observations

as well as from reports by the hydrometric station operator and other

nearby residents. River discharge is derived from measurements or

estimates of the flood stage, using the stage-discharge equations

described above. The instantaneous rates of sediment discharge were

calculated by multiplying sediment concentrations times the discharge

at the time of sampling. In cases of time series sampling during

flood events the average discharge rate for each time interval was

converted to a discharge value, then multiplied by the sediment

concentration. The sum of the river discharge and sediment discharge

over the sampling period provided empirical measures of sediment

transport for flood events of a given magnitude.

Analysis of relationships between discharge, sediment

concentration, sediment transport and rainfall. The frequency

distributions of all variables were tested by frequency analysis (SAS,

1979). Based on the results of the preliminary analysis the










relationships between discharge, an independent variable, and sediment

concentration and transport (dependent variables) were tested by

simple linear regression of raw and log-transformed data (SAS, 1979).

The relationships between river stage and sediment transport were

used to estimate sedimentation rates of the dams to be constructed

just downstream of the sampling points in both rivers. The daily

discharge for a full hydrologic year during the period of record was

used to generate daily sediment discharge values based on the

relationship established in the previous analyses. The total was

multiplied by a correction factor (ratio of average annual discharge

to the discharge for the 1980-1981 hydrologic year) to predict the

sediment discharge for an average year, as opposed to the study

period.

The relationship between the amount and distribution of rainfall

on the watershed (independent variable) and the discharge and sediment

concentration in the rivers (dependent variables) was tested by simple

and multiple linear regression of raw and log-transformed data (SAS,

1979). The same procedure was repeated by subwatershed. Critical

areas for further study were singled out by the relative strength of

association between rainfall in each subarea and the subsequent river

flood stages and sediment concentration. This information, combined

with field reconnaissance, contributed to assignment of research

priority by subwatersheds.

The relationships between total daily rainfall on the whole

watershed and discharge and sediment concentration also were tested by

simple linear regression. The total volume of rainfall on the










watershed was compared to total volume of discharge, by month, and the

remainder of rainfall minus discharge was attributed to storage and

evaporation. The results were compared to water balances previously

calculated for weather stations in or near the study area.



Precipitation, discharge, sedimentation, and production in small
watersheds


The proximity of three climatological stations to the respective

study areas allowed direct application of the rainfall data from these

stations. To supplement the existing monitoring network, small

plastic water gauges with 5-cm apertures were mounted at eye level on

wooden supports installed in well-exposed open areas. The amount,

intensity, and duration of rainfall on a daily basis during peak

rainfall periods were recorded.

Discharge and sediment concentration. Discharge measurements

made under relatively low flow conditions followed the procedures

prescribed by the velocity-area method (Herschy, 1978). The surveyed

cross-sections in each stream were segmented and a velocity

measurement was made at the center of each segment.

Surface flow velocity was measured with floats and chronometer as

illustrated in the diagram (Fig. 10). Each velocity measurement

consisted of three to five readings (sec), that were converted to flow
-i
rates (m sec ), then averaged. One complete velocity measurement was

made for the center of each segment, except in cases where the cross-

section was treated as a single segment. The discharge rate (Q) is

equal to the velocity (V) times the cross-sectional area (A), or a

segment thereof (Ax):

Q(m3 sec-1) -1 2
Qmsec ) = V(m sec ) x A(m ).

















































S..... ... ....


















*. . . . . .



. '.. . . .
. . ...

...


J















C


4-4

) 0


0
*rJ -1
'H

4-) 0-



0)
4J E
(10 1

r1
0 U)
r U)



CflC


E Ou
(1)
U)



r4






O

(0 ,-
a1)
z >
00


a) 0

4.J

ra 0
0 r



EW(


CI -4
4-)W




*H) E

o --

WU)
04 )
I) 0)

> '0

O
-1 "


0 U
> ro




r4~.4
0



4.)
En
41C

r-1
H- 0


0
0tI


C


.. .. .. ... .
. .........
. ........
..........
...........
...........
............
............
.. .. .. .. .
............
.............
.............
.............

.............
............
..........

............

..........

..........




Full Text




151
nearly opaque, brown, debris-laden waters of flash flood peaks in May.
The sediment concentrations in both rivers were very regular during
dry weather. Suspended sediment concentrations in Mao during March of
1980 and 1981 ranged from 0.01 to 0.20 g L ^ (see Appendix E).
Concentrations for the Amina under dry weather, low flood conditions
were between 0.02 and 0.10 (see Appendix E). These findings are
consistent with results reported for streams draining forested
watersheds in the southeastern United States, where average sediment
concentrations ranged from 0.02 to 0.20 g L ^ (Duffy et al., 1978;
Reikerk et al., 1979; Switzer and Nelson, 1972).
The maximum concentrations measured for Mao River were 11.04 g
-1 -1
L in May 1980 and 2.9 g L in May 1981. The peak concentration for
the Amina was 3.5 g L \ These concentrations are consistent with
rising stage flood flow sediment concentrations of 2.0 to 3.5 g L ^
reported for the Morogoro River in Tanzania. The Morogoro catchment
is much smaller but is located in an area of similar topography and
land use.
Total sediment transport rather than concentration varies with
discharge (e.g., Likens et al., 1977). Sediment concentration is
especially likely to be uniform in areas where fluctuation in river
stage is gradual and the range of variation is relatively narrow.
However, in areas where river stage (and therefore velocity) varies
abruptly over a wide range, the sediment-carrying capacity of the
river also changes dramatically over a short time (Gregory and
Walling, 1973). If the available sediment from runoff and in channel
deposits is sufficient, then suspended sediment concentrations may


110
Field crops
Production of field crops tends to be limited to low-value crops
such as manioc (yuca), with some mixed subsistence and commercial
production of red beans, corn, cowpeas, peanuts, sisal and tobacco.
Intercropping is common practice and the plots are relatively small
(0.2 to 1.0 ha) with intercropping of two or more crops. Monoculture
is also practiced with tobacco, red beans, and sisal.
Field crops may have several distinct roles in the overall
production strategy of a given farm household. The plot may occupy
the entire landholding of a sedentary subsistence farmer who
supplements family income by raising other crops in more distant
forest lands, by sharecropping, or by seasonal day-labor. Field crops
often are found in areas predominantly covered by forest and pasture.
Shifting cultivation (slash and burn) is still widespread,
particularly in remote areas of higher elevation, where extensive pine
forests are found (Swedforest, 1980; Montero et al., 1981). The
practice also is common in the drier forests at lower elevations. It
deviates from the classic pattern (Greenland, 1974) primarily in its
long fallow (10 to 15 years).
This practice stops primarily because of lack of accessible
forests, in populous zones at 500 to 700 m elevation. The dominant
method more closely resembles the bush-fallow farming systems
described in recent West African research (Ruthenberg, 1976; Lagemann,
1977). Field crops are rotated with pasture and/or a 2-to-10-year
fallow in bush (secondary growth). Plots of this type may be
cultivated by the owner (in a small holding), or by sharecroppers,
renters, and/or hired help in large landholdings.


94
pigeon pea and sweet potato with minimum and normal tillage (numbers
82 and 83, Los Montones). The modified design illustrated in Fig. 13
is a combination of the standard plot and the Gerlach trough
collection device (Morgan, 1979). Due to the modifications both plots
at this site required three collection tanks.
3
The maximum installed storage capacity (0.5 m ) was determined by
cost and maintenance considerations. The use of a slot divisor to
split the sample prior to collection would have been ideal for
experimental purposes, but costly to produce. An increase in storage
3
capacity to 1 m also would have proven costly and difficult to
install.
The plots were installed during March and April 1980 and
collection began at most sites on 1 May. The runoff and sediment
collected in the tanks were measured, sampled and removed as soon as
possible after each rainfall event. _Nevertheless, many composite
results were obtained, as well as some results recorded for individual
events. Overflows were rare, so that the total rainfall and soil loss
measurements were not jeopardized by composite collection of water and
sediment for two or more events.
Procedures for field measurements, sample collection and
laboratory analysis. The procedures for measurement and sampling of
the collected runoff and sediment at all 16 sites followed standard
methods cited in the literature (Dunne, 1977; Djorovic, 1977; USDA,
1979). The total depth from the bottom of the tank to the water
surface (cm) and the depth of sediment deposits (if greater than 5 cm
in depth and uniformly distributed) were measured.
If a deep (>5 cm)














3
represent a choice by default rather than a free choice between
rational alternatives for sustained production. Moreover, there is
intense pressure for continuous cropping and/or establishment of
pastures on cleared land. Large local landowners, as well as urban
and foreign markets, play a major role in this process (Amin, 1977;
Cultural Survival Inc., 1982, Hildyard, 1982; Nations and Komer, 1982;
Plumwood and Routley, 1982).
Small farmers and shifting cultivators in such areas often
practice forms of management that can be sustained well at lower
population densities or within more hospitable environments to which
they have no access (Bailey, 1982; Grainger, 1980; Nations and Komer,
1982). Although these small farmers in marginal areas are referred to
by many as "subsistence farmers," they usually produce some surplus
food crops. In some cases this sector is a major source of staple
food production for domestic markets (Brush, 1981; Novoa and Posner,
1981). The same farm families often function as a seasonal labor
force in coffee, lumber, and other cash crop harvests (Beckford, 1972;
Frucht, 1967). These subsistence farmer/farmworker populations in
marginal lands highly susceptible to erosion form an integral part of
the regional economy and ecosystem. As such, the problem must be
treated as a complex phenomenon that not only affects the larger
downstream and lowland production systems, but is partially
conditioned by them.
The question remains as to how production can be maintained or
increased (at a sustained rate) with minimum damage to both portions
of the watershed. The need for an answer to this question is
particularly urgent in the Caribbean because of high population








273
than 100 times the soil loss from forest lands, 25 times the losses
from pastures and more than 10 times the erosion from coffee
plantations in the same area.
In addition to the influence of land use category, the land use
history and specific cropping practice may ameliorate or exaggerate
the generally high erosion rates attributed to land in annual crops.
The Pananao watershed is an example of an area with a high proportion
of land in annual crops. It also experiences relatively high erosion
rates within the cropland areas. This is probably due to the
intensive cultivation of peanuts during the last 30 years and to the
resultant damage to soil structure.
The analysis of variance for sediment yield per unit area during
floods showed no significant difference between the watersheds in the
coffee region and those in the lower elevations covered in pasture and
croplands. The most plausible explanations are the effects of
sediment discharge ratio and the existence of a uniform erosion rate.
The first explanation seems to hold for the case of the Prieto and the
Upper Bajamillo watersheds compared to Hondo and Pananao. The fact
that all four watersheds yield approximately the same amount of
sediment per unit area during flood events does not by itself indicate
a uniform erosion rate. According to Roehl (1962) the smaller
watersheds should yield approximately 33% of the total amount of
sediment eroded, while the larger watersheds should yield
approximately 10% of the total eroded material.
The comparison of baseflow versus flood flow sediment yields
(Table 39) and the calculation of sediment delivery ratios for


APPENDIX C
STAGE -DISCHARGE CURVES FOR RIVERS, DERIVED BY
INDRHI


248
1.0%, ranging from 0.25 to 0.90%. The values for the study areas in
the Sierra compare well with these examples, at 1.0%.
The increments of runoff over natural vegetation range from 3 to
7 for pastures in the study area, as compared to 12-fold increases
reported for Imperata grassland in Mindanao (Kellman, 1969) and 10-
fold increases reported for grazed savanna in Uganda (Sperow and
Keefer, 1975). Incremental increases for field crops range from 20 to
40 in Senegal (Charreau, 1972; Moutappa, 1973) and Uganda (Sperow and
Keefer, 1975). In Mindanao, Kellman measured widely varying
proportional increases of runoff in field crops over forest values
(Table A-ll). In new fields the runoff increased 4 to 6 times and in
older fields the runoff ranged from 8 to 50 times the amount for
forest.
The runoff rates measured in field crop sites in the study area
showed increases over forest rates ranging from 2 to 12-fold, which is
less than all of the above. However, the runoff as percent of
precipitation very closely approximates the rates reported by Kellman
(1969), with the exception of forest. The 1 to 2% runoff rates for
newly established field crops are repeated at the Los Montones plots
(84 and 85). The trend toward increased runoff with longevity of
field crop production at the site, mentioned earlier, is also
reflected in the Mindanao results. The 12% rate for 12-year-old rice
fields was duplicated in the eroded manioc plot (95) at Pananao. This
plot was the most consistently and intensively cultivated of all the
plots included in the study. The results of both studies indicate the
importance of incorporating land use history into runoff estimates,
whether at the plot or regional level.


28
research focuses more on entire watersheds and less on individual
plots, in contrast with prior work conducted by the Soil Conservation
Service at the farm level (Ackerman, 1966). Most of the watershed
experiments combine monitoring of precipitation with recording rain
gauges, continuous monitoring of stream discharge at weirs, sampling
of sediment discharge at weirs, and/or collection of sediments in weir
2
ponds in watersheds with areas less than 20 km (Hewlett et al., 1969;
Ward, 1971).
A ma]or topic of the earlier studies was the role of the
undisturbed forest in regulating the hydrologic cycle and sediment
export (Douglass and Swank, 1975; Helvey, 1967; USDA Forest Service,
1980). Among the more important findings were: the importance of
litter versus canopy in protecting the soil against the erosion
potential of rainfall (Table A-l); the impact of forest vegetation on
stream discharge (Dils, 1957; Johnson and Swank, 1973) (Table A-2);
and the association between undisturbed forest cover and low sediment
concentration in streams (Table A-3).
Streamflow, sediment concentrations and mass transport from
forested watersheds showed dramatic changes after harvesting, various
site clearing and management operations, or conversion to other uses.
Several studies reported heavy increases in suspended sediment and
nitrate concentrations after clearcutting (Bormann and Likens, 1979;
Douglass and Swank, 1975; Hewlett and Nutter, 1969; Likens et al.,
1977; Monk, 1976).
Water yield increments proportional to percent area in cleared
openings were reported for several gauged watersheds (Likens et al.,




Table 23. Relationship of total annual rainfall and storm runoff in erosion plots.t
Site
Plot
Rainfall
yr 1
3 -i
m ha
yr 2
3
Runoff m
yr 1
ha 1
yr 2
Runoff as
yr 1
% of Precipitation
yr 2
Carrizal
80
25,728.0
27,196.5
402.35
791.46
1.6
2.9
81
II
II
602.43
1038.01
2.3
3.8
Los Montones
82
15,875.0
14,728.0
no data
603.37

4.0
83
II
II
no data
665.95
-
4.5
84
II
II
353.43
327.79
2.2
2.2
85
II
II
285.59
206.89
1.8
1.4
86
II
II
535.17
494.35
3.3
3.4
87
II
II
161.00
172.95
1.0
1.3
88
II
II
146.81
129.76
1.0
0.9
89
II
II
416.60
347.67
2.6
2.4
Pananao
91
12,701.0
12,236.0
801.50
845.81
6.3
6.9
92
II
II
781.22
824.18
6.1
6.7
93
II
II
522.31
712.67
4.1
5.8
94
II
II
518.45
742.41
4.1
6.1
95
II
II
922.30
1434.93
7.3
11.7
96
II
II
98.67
145.10
0. 8
1.2
fAnnual totals are calculated for May 1980 through Apr^l 1981 and from July 1980 through June
1981 to compare the variation in spring rainfall and runoff for both years.












Fig. 53. Evaluated small watershed submodel of land use, erosion, and sedimentation: Upper
Bajamillo stream.
267










49
capacity are clear. While the system itself may work, shifting
_2
cultivation cannot support more than about 20 to 50 persons km ,
taking into account the required fallow. Further experimental work at
IITA by Greenland, Lai, and others has explored alternatives to this
system, emphasizing soil management and conservation under bush fallow
and continuous cropping (Lai et al., 1979; Lai, 1977a) and continuous
mixed-cropping systems (Greenland, 1975; Ruthenberg, 1976).
Bioeconomic modelling has been proposed to evaluate alternative soil
conservation practices and cropping systems (Dumsday and Flinn, 1977).
Agroforestry research in Africa has combined experiments with
commercial forestry and subsistence agriculture (King, 1968). The
taungya system features mixed cropping of commercially harvested and
replanted forest tracts, with the tenant farmers caring for the
seedlings and saplings as well as their food crops over a period of
about four years (Dubois,.1979; King, 1978). The field of
agroforestry has further developed to include diagnostic and
experimental work with existing subsistence and commercial production
systems that feature some combination of trees, livestock production
and/or field crops (Brookman, 1976; Douglas and Hart, 1976; Olawoye,
1975; Parry, 1957; Raintree, 1982; Lundgren, 1982). Both cocoa and
oil palm production on small farms have been studied within this
context (Flinn, 1980; Grinnell, 1977; Lagemann, 1977; Letouzay, 1955)
as well as many traditional systems of shifting agriculture that
include management of tree crops (Dubois, 1979; King, 1968). In
general, mixed tree crop/annuals production systems are more diverse
and more stable, both in economic and ecological terms (Lagemann,
1977).


Table F-2
Station 64, Prieto Stream
Date MAXLEVL
SEDl
SED2
SED3
SED4
SED5
MAXCSCHG
FLDDSCHG
SEDDISl
SEDDIS2
! SEDDIS3 SEDDIS4
SEDDIS5
800516
l
0.10
-
-
-
-
0.223
_
0.0
_
_
_
_
800531
2
0. 29
0.14
-
-
-
0.753
-
0.0
0.1
-
-
_
800612
3
0.45
1.06
0.61
-
-
1.458
-
0.1
0.8
0.8
-
-
800915
3
0.45
1.06
0.61
-
-
1.458
-
0. 1
0.7
0.8
-
-
801010
5
49.76
31.36
3.58
0.84
-
3. 350
5.742
11.0
23.6
5.2
2.2
-
801022
4
94.46
91.70
45.93
13.44
-
2.633
-
21.0
69.0
66.9
35.3
-
801031
1
1.47
-
-
-
-
0.223
-
0.3
-
-
-
-
801107
2
-
1.19
-
-
-
0.753
-
-
0.4
-
-
-
801210
1
0. 14
-
-
-
-
0.223
-
0.0
-
-
-
-
801222
4
3. 30
-
-
1.26
-
2.633
-
0.7
-
-
3.3
-
800109
1
0.19
-
-
-
-
0.223
-
0.0
-
-
-
-
810116
1
0.68
-
-
' -
-
0.223
-
0.1
-
-
-
-
810119
1
0. 35
-
-
-
-
0.223
-
0.0
-
-
-
_
810202
2
0.66
1.38
0.85
-
-
0.753
-
0.1
1.0
1.2
-
-
810204
1
0.68
-
-
-
-
0.223
-
0. 1
-
-
-
-
810223
1
0.24
-
-
-
-
0.223
-
0.0
-
-
-
-
810316
1
0.46
-
-
-
-
0.223
-
0.1
-
-
-
-
810327
1
0.50
-
-
-
-
0.223
-
0.1
-
-
-
-
810429
2
103.20
99.47
0. 28
-
-
0.753
-
23.0
74.9
0.4
-
-
810504
3
0.28
5.46
10.93
-
-
1.458
-
0.0
4.1
15.9
-
-
810509
1
0.15
-
-
-
-
0.223
-
0.0
-
-
-
-
810519
2
-
43.51
-
-
-
0.753
-
-
32.7
-
-
-
810526
2
35.04
5.40
-
-
-
0.753
-
7.8
4.0
-
-
-
810609
2
2.51
13.83

-
0.753
0.5
10.4

. 3
-1
Discharge
rates (m
sec )
for each
level,
corresponding
to concentrations
SEDl
through
SED5, are
as
follows:
1 -
0.223,
2-0.
753, 3 -
1.458,
4-2.
633, 5 -
3.350.

345


382
Plot 84. Mixed Food Crops with Hillside Ditches, Los Montones.
Sample Collection
Sediment Yield
Runoff Rate
Date
(kg ha )
(m ha )
80
05
14
5447.4
26.3842
80
05
31
29579.9
23.9066
80
06
11
10782.4
90.0801
80
06
20
67.0
10.8992
80
08
08
0.4
4.5719
80
09
13
28.2
14.5462
80
09
25
15.6
3.9985
80
10
02
1292.7
88.7756
80
10
30
8.6
2.0483
80
11
06
2.9
2.0483
80
12
02
1.6
1.0194
80
12
10
1.6
2.6276
80
12
23
96.6
3.8924
80
12
29
1.6
1.5101
81
01
08
34.0
2.0483
81
01
15
127.8
44.9961
81
01
19
806.0
5.2796
81
02
04
o

o
rH
12.6875
81
02
23
3.4
3.2434
81
03
16
5.5
6.7727
81
03
18
1.6
1.5101
81
03
25
0.0
0.5859
81
04
03
5.8
5.2796
81
04
09
1.6
3.8924
81
04
14
2.9
5.2796
81
04
27
0.00
5.2796
81
04
28
12.2
41.1991
81
05
05
9.9
41.1991
81
05
06
0.0
3.8924
81
05
13
0.0
7.5552
81
05
28
1.5
5.2796
81
06
10
2.0
6.7727


324
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
801202
0.47
0.47
-
-
-
-
-
11.3
11.3
11.3
-
801203
0.46
0.46
-
-
-
-
-
11.0
11.0
11.0
-
801204
0.46
0.46
-
-
-
-
-
11.0
11.0
11.0
-
801205
0.45
0.45
-
-
-
-
-
10.6
10.6
10.6
-
801206
0.70
0.60
-
120
0.339
0.267
0.30
12.6
16.3
20.0
-
801207
0.81
0.66
-
-
-
-
-
31.9
20.96
26.4
-
801208
0.55
0.52
-
-
-
-
-
14.3
13.2
13.8
-
801209
0.50
0.48
-
-
-
-
-
12.4
11.7
12.1
-
801210
0.47
0.47
-
-
-
-
-
11.3
11.3
11.3
-
801211
0.46
0.45
-
-
-
-
-
11.0
10.6
10.8
-
801212
0.45
0.45
-
141
0.175
-
-
10.6
10.6
10.6
-
801213
0.45
0.45
-
-
-
-
-
10.6
10.6
10.6
-
801214
0.44
0.44
-
-
-
-
-
10.3
10.3
10.3
-
801215
0.48
0.47
-
-
-
-
-
11.7
11.3
11.5
-
801216
0.46
0.46
-
-
-
-
-
11.0
11.0
11.0
-
801217
0.45
0.44
-
-
-
-
-
10.6
10.3
10.5
-
801218
0.44
0.43
1.77
142
0.282
-
-
10.3
9.9
10.1
129.6
801219
1.02
0.70
-
-
-
-
-
51.2
12.6
37.4
-
801220
0.65
0.55
-
133
0.273
0.273
0.27
20.3
14.3
17.3
-
801221
0.50
0.49
-
-
-
-
-
12.4
12.1
12.3
-
801222
0.53
0.48
-
-
-
-
-
13.6
11.7
12.6
-
801223
0.47
0.46
-
-
-
-
-
11.3
11.0
11.2
-
801224
0.44
0.48
-
-
-
-
-
10.3
11.7
11.0
-
801225
0.44
0.43
-
-
-
-
-
10.3
9.9
10.1
-
801226
0.43
0.43
-
-
-
-
-
9.9
9.9
9.9
-
801227
0.43
0.43
-
-
-
-
-
9.9
9.9
9.9
-
801228
0.44
0.42
-
-
-
-
-
10.3
9.6
9.9
-
801229
0.42
0.42
-
142
0.148
0.222
0.19
9.6
9.6
9.6
-
801230
0.42
0.41
-
-
-
-
-
9.6
9.3
9.4
-
801231
0.42
0.41
-
-
-
-
-
9.6
9.3
9.4
-
810101
0.40
0.40
-
-
-
-
-
8.9
8.9
8.9
-
810102
0.40
0.40
-
-
-
-
-
8.9
8.9
8.9
-
810103
0.40
0.39
-
-
-
-
-
8.9
8.6
8.8
-
810104
0.39
0.39
-
-
-
-
-
8.6
8.6
8.6
-
810105
0.39
0.40
-
-
-
-
-
8.6
8.9
8.8
-
810106
1.35
0.90
-
100
-
-
-
91.1
39.6
65.3
-
810107
0.75
0.65
1.15
90
0.450
0.450
0.45
27.2
20.3
23.7
65.5
810108
0.85
0.72
-
114
0.286
0.224
0.26
35.2
25.0
30.1
-
810109
0.61
0.56
-
-
-
-
-
17.8
14.7
16.3
-
810110
0.54
0.55
0.90
134
0.292
-
-
13.9
14.3
14.1
39.6
810111
0.75
0.63
-
128
0.234
0.220
0.23
27.2
19.0
23.1
-
810112
0.68
0.60
-
-
-
-
-
22.2
16.3
19.3
-




397
Chaney, E. M., and M. W. Lewis. 1980. Planning a family food
production program: Some alternatives and suggestions for
Plan Sierra. Report to Plan Sierra, San Jose de las Matas,
Dominican Republic.
Chapman, T. G., and F. Dunin. 1975. Prediction in catchment
hydrology. Aust. Acad. Sci., Canberra.
Charreau, C. 1972. Problemes poses par 1'utilization agricole des
sols tropicaux par des cultures annuelles. Tropical Soil
Research Syme, IITA, Ibadan, Nigeria, May.
Chisholm, M. 1967. General systems theory and geography. Trans.
Inst. Brit. Geog. 42:42-53.
Chorley, R. J. 1962. Geomorphology and general systems theory.
U.S. Geol. Sur. Prof. Paper 500-B, Washington, D.C.
. 1969. The drainage basin as the fundamental geomorphic
unit. In R. J. Chorley (ed.) Water, earth, and man. Methuen,
London.
Chow, V. T. (ed.). 1964. Handbook of applied hydrology: A compen
dium of water-resources technology. McGraw-Hill Book Co., New
York.
Collinson, M. 1981. A low cost approach to understanding small
farmers. Agricultural Administration. 8:6.
Combe, J. 1979. Concepts of agro-forestry research techniques at
CATIE. Proc. Workshop on Agroforestry Systems in Latin
America, Turrialba, Costa Rica, March.
Cooper, C. F. 1971. Ecosystem models in watershed management.
In B. M. Van Dyne (ed.) The ecosystem concept in natural
resources management. Academic Press, New York.
Coulter, J. K. 1972. Soils of Malaysia. Soils and Fert. 35:
475-498.
Crosson, P. R., R. G. Cummings, and K. D. Frederick (eds.) 1978.
Selected water management issues to Latin American agriculture.
Resources for the Future, Baltimore.
, and K. D. Frederick. 1977. The world food situation:
Resource and environmental issues in the developing countries
and the United States. Resources for the Future, Washington,
D.C.






11
Purpose and Scope of the Work
The purpose of the study was to develop and test an
interdisciplinary methodology that addresses the needs cited above
within the context of the Caribbean. Such a methodology, and the
theoretical framework within which it is developed, should meet the
following criteria:
1. Treat the problem in a holistic way, building upon existing
research in several disciplines, and incorporating both physical and
cultural aspects of erosion and land use problems;
2. Be flexible enough to include analyses within a broad range
of temporal and spatial scales of observation and to allow for
differentiation by client groups;
3. Facilitate the participation of clients in research and
extension programs;
4. Be applicable in areas with limited data bases;
5. Yield practical short-term results at the topographic scale
while working toward more basic solutions over the long-term;
6. Be amenable to integration into rural development programs,
from preliminary research to subsequent extension efforts.
Objectives
The specific objectives of the study included short-term
practical achievements at the local and regional level. The latter
were readily accomplished within the larger task of methodology
development and testing and contributed to the development of the
theoretical aspects of this study. The objectives, stated in
chronological order of completion, were as follows:


75
In the large watersheds total sediment transport was measured and
sediment yield was calculated in order to estimate the magnitude of
the erosion problem on the watershed, to predict the future
sedimentation rates of the proposed dams, and to compare the losses
per unit area between the study areas and other sites for which
sediment yields have been reported. The measurement in the small
watersheds showed the integrated effects of land_ use and physical
characteristics in each study site. Subsequent tests of the data by
multiple analysis of variance indicated the relative influence of land
use and physical factors.
Experimental plots were included to demonstrate the impact of
varying specific crop types, land treatments and conservation
practices on erosion and runoff. The plots also provided the erosion
rate data necessary to calculate sediment delivery ratios for the
watersheds. Erosion plots were included because they allow
observation of the problem within the context of individual
landholdings and related households. The experiments were conducted
under the same conditions that limit and influence the management of
individual landholdings. While the most striking effects may be
expressed at the watershed level, the management decisions are made at
the household level. The degree to which such decisions are
constrained by pressures from the larger system does not change the
fact that these decisions directly determine, in turn, the condition
of the larger watershed. Any proposed changes must be tested within a
holistic framework at the level of the land managers.


o
o
OJ
o
rj
r
zr
(O
o
(U
m
1960 P'.INI ill I
1961 Fill 1 NFfil l
MFllN
d
-
cr
Cl_
-i
L.
til
cr
L
ZT
o :
rj -
ro :
Fig. D-3. Monthly rainfall at the Mao climatological station, #5.
i
307




352
Location:
Carrizal
Plot #:
80
Land Use:
Coffee Stand
Land Cover:
Established coffee
with dense shade
Parent Material
: Complex crystallines, meta-volcan
Relief:
Pronounced
Position:
Mid-upper slope
Slope Class:
24-45%
Runof f:
Rapid
Permeability:
Slow
Erosion Class:
None slight
Drainage Class:
Rapid (4)
Soil Moisture:
Uniform, moist
Salts/Alkaline:
None
Stoniness:
Moderate
Soil Profile:
Horizon
A
AC
Depth
Boundaries
0-12 cm
12-20 cm
Color
(dark reddish
browm)
Texture
loam
clay loam
Structure
blocky
blocky
Consistency
-
-
CO 3
no
no
Concretions
no
no
Slickensides
no
no
Mottles
no
no
Roots
yes
yes
litter
C
20-55+ cm
clay loam
blocky
no
no
no
no
yes


374
Location:
Pananao
Land Use:
Sisal and Food
Crops with Hillside
Plot #:
91
A = RKLSCP
R
= 855
K % silt
= 43
% sand
= 43
% O.M.
= 3.02
- K = .24
Structure = 3
Permeability = 2.5
L 22m
= LS
= 8.0
S 34%
C
= 0.2, 0.9
P
li

on
o
1*
A
A
(c=.2)
(p=. 5)
(c=.2)
(c=l)
A
A
(c=.9)
(p=.5)
(c=.9)
(P=l)
= 164.16
t
ha '''yr 1
= 328.32
t
ha '"yr ^
= 738.72
t
, -1 -1
ha yr
=1477.44
t
, -1 -1
ha yr
Additional Information:
% clay = 14.0
% N = 0.15
Location: Pananao
Land Use: Sisal and Food Crops,
Plot #: 92
A = RKLSCP
R
K % silt
% sand
% O.M.
Structure
Permeability
L 22m = LS
S 32%
C
P
47
42
(3.98)4
3
2.5
K
No Conservation
= 855
= .26
= 7.2
= .2, .9
= 1
Practice
A (c=
.2)
= 320.11
t
ha 'yr ^
A (c=
.9)
=1440.50
t
, -1 _1
ha yr
Additional Information:
% clay = 12.0
% N =0.20


CHAPTER V
CONCLUSIONS
The analysis of erosion and sedimentation within a nested
hierarchy of watersheds and land use systems identifies relationships
that are not apparent at the plot or watershed scales. Spatial
analysis of land use associations within watersheds in the Plan Sierra
region indicated parallel proportions of forest and annual crop lands
in the coffee-producing areas and the extensive pasture lands
downstream. The erosion plot data extrapolated to the watershed level
result in almost uniform erosion rates for these two contrasting land _
use systems. The population density and the related demand for food
production are major determinants of erosion rates in the region.
In addition, the sequence of land uses has a strong influence on
erosion rates. The erosion plot data alone indicate a dramatic change
in erosion rates over time, with the highest rates occurring on steep
slopes immediately after deforestation. The threshholds for
destabilization and erosion of the various horizons are evidently
distinct. The change in erosion rates decays after time of clearing
for a given surface soil, but catastrophic events such as complete
removal of the A horizon or the destabilization of the lower horizons
can induce a new cycle of extremely high rates and subsequent
exponential decay. Peak erosion appears to be caused by initial
exposure and disturbance of former forest soils.
278








Ecological
Inputs
Cultural Inputs
N fixation
Lugehd
- lufornalion I'lows
Kitergy/Matarial s Flows
Honey Flows
*
Solar Energy
Ra 1nf al 1
nutranla
Surface VLter
Dutr1ents
w
0round iter
nutrante
w
Soil Foma t Ion
nutriente

i
cl
I
j I
A
LAi
TJ
a.
IHVtffTOHT: CONTEKTS 0/ 6TSTEM
Total floluae Viler
*1*11./ Unit Tl*
at Surface
inla&le: Population*
*nd Bloaaaa
Total Yolune Uitar
Stored for lluuo
Uaa
faga talln:
Iraa of Coiaraga
Population* aJul
Bloauiil
| | Liquid laaata ~~]
j Huaan Population*
Soil Tolu*#/ir#a
Soluat of Nutrient
Crop* In Ground
or In Storag*
(No. Bloaaaa)
firewood and
Charcoal
1;
B

< a
¡
3
d 4
!
3
l
1
J
a
a
Si
Vala of Paraonal *
Bouaahold Property
Valu* of Iaplasanta
* fehlclea
Valu* of Standing
Inf ra* truc tur*
land Valu*
Valu* of lar.) Corar
Valu* of Crop* In
Ground or Stored
Valu*/int. of Poaall
futa
4, o).
Si
i
i,
Brat
Uitr Evaporation
-
Uitrr Tranrplratlon
" w
Uitrr Runoff
Dutrlrnta/fertlll
paatlcldra
pTh6"**
(bacteria, finiere,
paraeltea)

Soli Eroalon
nutrlanta/fartlllief
w
pee ticlda
palbogana
Drtrltua
w
DoUanta
Huaan/Anlsal Wiaie
w
nutr rota
mw
N Voltil1ralion
Surfer
Viter
Oround-
U.t*r
M Drnl trlf 1 oatlon
fool oglcal
Outputa
Cultural Outputa
Fig. 5. Input-output diagram for interview notations and monitoring
(_n
U)


131
mixed forest. Soils on the steep slopes are very shallow. Rainfall
in the upper Mao watershed is higher than in the central Sierra. The
bimodal rainfall distribution peaks in October and to a lesser extent
in May. This regime contrasts with lower elevations where the first
peak occurs in May and the second in September (Jorge, 1970;
Swedforest, 1980).
The drainage network geometry makes the discharge of the Mao
River extremely responsive to rainfall. Flood crests occur less than
24 hours after peak rainfall (Swedforest, 1980). The time of
concentration for each separate tributary is such that storm runoff
from several subwatersheds reaches the Mao almost simultaneously.
This causes extreme flood stages and wide fluctuation in river
discharge.
More moderate rainfall and gentler slope in the smaller Amina
watershed yield slightly lower rates and volume of runoff than the Mao
watershed. Since a small portion of the catchment is influenced by
rainfall in the high Sierra, the river regime closely resembles
typical rainfall patterns of the Plan Sierra impact area. The flood
peaks on the Amina usually occur in May and are less frequent but tend
to be more extreme than the peaks in the Mao River discharge. Average
monthly river discharges are illustrated in Figs. 21 and 22.
The Mao has a greater sediment transport capacity and a greater
erosion potential as indicated by topography and erosivity of rainfall
(Paulet, 1978). Stratification of land use is roughly parallel in
both watersheds (Fig. 18). The major difference is in the proportion
of coffee and forest in the upper reaches. Coffee is far more






337
Table E-2--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
801201
0.59
0.58
-
120
0.170
-
-
3.8
3.5
3.7
801202
0.58
0.57
-
-
-
-
-
3.5
3.2
3.3
-
801203
0.57
0.57
-
-
-
-
-
3.2
3.2
3.2
-
801204
0.57
0.57
1.2
152
0.160
-
-
3.2
3.2
3.2
42.1
801205
0.57
0.65
1.8
-
-
-
-
3.2
5.7
4.4
115.5
801206
1.38
1.05
-
147
0.193
0.212
0.20
59.6
27.9
43.8
-
801207
0.98
0.96
-
-
-
-
-
22.2
20.7
21.5
-
801208
0.94
0.93
-
-
-
-
-
19.2
18.7
19.0
-
801209
0.93
0.92
-
-
-
-
-
18.7
18.0
18.3
-
801210
0.90
0.89
-
-
-
-
-
16.7
16.1
16.4
-
801211
0.87
0.86
-
-
-
-
-
15.0
14.4
14.7
-
801212
0.85
0.83
-
-
-
-
-
13.9
12.8
13.3
-
801213
0.83
0.81
-
-
-
-
-
12.8
11.8
12.3
-
801214
0.77
0.75
-
-
-
-
-
10.0
9.2
9.6
-
801215
0.68
0.65
-
-
-
-
-
6.6
5.7
6.1
-
801216
0.64
0.62
-
-
-
-
-
5.7
4.9
5.3
-
801217
0.60
0.58
-
-
-
-
-
4.1
3.5
3.8
-
801218
0.58
0.57
1.8
-
-
-
-
3.5
3.2
3.3
115.5
801219
1.19
1.17
-
-
-
-
-
41.2
39.5
40.4
-
801220
1.09
1.04
-
147
0.328
0.351
0.34
31.6
27.0
29.3
-
801221
0.99
0.95
-
-
-
-
-
23.0
20.0
21.5
-
801222
0.73
0.72
-
-
-
-
-
8.4
8.0
8.2
-
801223
0.70
0.67
-
180
1.31816.305
8.81
7.3
6.3
6.8
-
801224
0.64
0.62
-
-
-
-
-
5.7
4.9
5.3
-
801225
0.62
0.61
-
-
-
-
-
4.9
4.5
4.7
-
801226
0.60
0.60
-
-
-
-
-
4.1
4.1
4.1
-
801227
0.69
0.68
-
-
-
-
-
6.9
6.6
6.8
-
801228
0.68
0.67
-
-
-
-
-
6.6
6.3
6.5
-
801229
0.65
0.64
-
190
0.349
0.747
0.55
5.7
5.7
5.7
-
801230
0.63
0.63
-
-
-
-
-
5.3
5.3
5.3
-
801231
0.62
0.61
-
-
-
-
-
4.9
4.5
4.7
-
810101
0.61
0.61
-
-
-
-
-
4.5
4.5
4.5
-
810102
0.60
0.60
-
-
-
-
-
4.1
4.1
4.1
-
810103
0.60
0.59
-
-
-
-
-
4.1
3.8
4.0
-
810104
0.58
0.58
-
-
-
-
-
3.5
3.5
3.5
-
810105
0.60
0.98
-
164
0.314
-
-
4.1
22.2
13.2
-
810106
1.27
1.23
-
126
26.089
-
-
48.5
44.8
46.6
-
810107
1.01
0.91
-
-
-
-
-
24.5
17.4
20.9
-
810108
0.85
0.81
-
170
0.214
0.212
0.21
13.9
11.8
12.8
-
810109
0.75
0.72
-
-
-
-
-
9.2
8.0
8.6
-
810110
0.69
0.68
-
-
-
-
-
6.9
6.6
6.8
-
810111
0.81
0.78
-
164
0.869
0.875
0.87
11.8
10.4
11.1
-








187
Fig. 38.
System model of small watershed i,n pasture-field crop association.










317
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800212
0.35
0.34
-
-
-
-
-
7.3
7.0
7.2
-
800213
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800214
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
800215
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
800216
0. 32
0.32
-
-
-
-
-
6.4
6.4
6.4
-
800217
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
-
800218
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
800219
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
800220
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
800221
0.51
0.64
-
-
-
-
-
12.8
19.6
16.2
-
800222
0.47
0.42
-
-
-
-
-
11.3
9.6
10.5
-
800223
0.39
0.37
-
-
-
-
-
8.6
7.9
8.3
800224
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
-
800225
0.34
0.33
-
-
-
-
-
7.0
6.7
6.8
-
800226
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800227
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
800228
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
800229
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
800301
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
800302
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-
800303
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-
800304
0.60
0.65
-
-
-
-
-
16.3
20.3
18.3
-
800305
0.47
0.42
-
-
-
-
-
11.3
9.6
10.5
800306
0.38
0.36
-
-
-
-
-
8.3
7.6
7.9
800307
0.82
0.59
-
-
-
-
-
32.7
15.9
24.3
800308
0.65
0.52
-
-
-
-
-
20.3
13.2
16.7
-
800309
0.44
0.43
-
-
-
-
-
10.3
9.9
10.1
-
800310
0.40
0.38
-
-
-
-
-
8.9
8.3
8.6
800311
0.37
0.36
-
-
-
-
-
7.9
7.6
7.8
800312
0.35
0.35
-
-
-
-
-
7.3
7.3
7.3
-
800313
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
-
800314
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800315
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800316
0.32
0.30
-
-
-
-
-
6.4
5.8
6.1
-
800317
0.30
0.32
-
-
-
-
-
5.8
6.4
6.1
800318
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
800319
0.31
0. 31
-
-
-
-
-
6.1
6.1
6.1
800320
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
800321
0.30
0.30
-
145
0.021
0.021
0.021
5.8
5.8
5.8
800322
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-
800323
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-
800324
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-


20
USDA Forest Service, 1980; Williams, 1975) by estimating the sediment
delivery ratio based on drainage area (Hoitan and Lopez, 1971; Roehl,
1962). The models derived from the USLE have been used widely in
economic and land use planning studies for evaluation of specific
cropping systems and/or conservation practices (Kling and Olson,
1975). The weaknesses of the approach include the inherent
limitations of the OSLE as well as the questionable realism of
sediment routing techniques (Skopp and Daniel, 1978).
One empirical model that does not derive from the USLE is used by
the United States Bureau of Reclamation (Flaxman, 1975; Skopp and
Daniel, 1978; Strand, 1975) as well as by international research
organizations (Rapp, 1977). The technique combines flow duration
curves with sediment concentration, the latter derived from either a
sediment rating curve or a power function. Both flow duration and
sediment rating curves describe empirical relationships that must be
determined on site.
The model is useful for prediction of sediment yield over the
long term, given a continuation of current land use conditions, but is
less amenable to integrated watershed management based on land use and
land treatment programs. The flow duration/rating curve model,
however, could be calibrated to particular watersheds or groups of
watersheds for specific land uses and treatments, as has been the case
in paired watershed studies conducted in experimental catchments.
A number of sophisticated digital computer models of runoff
and/or erosion and sedimentation predict water and sediment discharge
by relating hydrologic and physical characteristics of the source




398
Cultural Survival Inc. 1982. Deforestation: the human costs.
Cultural Survival Quarterly. 6:3-7.
Dasmann, R. F., J. P. Milton, and P. H. Freeman. 1973. Ecological
principles for economic development. John Wiley and Sons, Inc.,
New York.
de la Fuente, S. 1976. Geogrfica Dominicana. Author, Santo
Domingo, Dominican Republic.
de Leon, A. 1980. Obras de conservacin en la Repblica Dominicana,
caso Tavera. Proc. 1st Seminario de Manejo de Cuencas Hydro-
graficas, Santo Domingo, Dominican Republic, May.
Dent, J. B., and J. R. Anderson (eds.). 1971. Systems analysis in
agricultural management. John Wiley and Sons, New York.
Dils, R. E. 1957. A guide to the Coweeta hydrologic laboratory.
Southeastern For. Exp. Stn., Asheville, N. C.
Djorovic, M. 1977. Use of runoff plots to evaluate soil loss.
In S. Kunkle and J. Thanes (eds.) Guidelines for watershed
management, FAO Conservation Guide 1, FAO, Rome.
Donigan, A. S., and N. H. Crawford. 1976. Modelling nonpoint
pollution from the land surface. Ecological Research Series,
Washington, D.C.
Douglas, I. 1968. Erosion of the Sungei Gomback catchment, Selangor
Malaya. J. Trop. Geog. 26:1-16.
Douglas, J. S., and R. A. de J. Hart. 1976. Forest farming.
Watkins, London.
Douglass, J. E., and W. T. Swank. 1975. Effects of management
practices on water quality and quantity: Coweeta Hydrologic
Lab., North Carolina. Tn Municipal Watershed Management
Symp. Proc., USDA For. Serv. Gen., Washington, D.C.
Dubois, J. 1979. Aspects of agroforestry use in Mayombe and lower
Congo (Zaire). Proc. Workshop on Agroforestry Systems in Latin
America, Turrialba, Costa Rica, March.
Duckman, A. N., J. G. W. Jones, and E. H. Rabuts (eds.). 1974.
Food production and consumption: The efficiency of human food
chains and nutrient cycles. American Elsevier Pub. Co., Inc.,
New York.






117
(Georges, 1981). However, many of them find no alternative and depend
on this source of cash income to supplement subsistence production.
The palm-pasture association is widespread in the region,
particularly in the mid-- to low-altitude zones. Erosion features
under this land cover appear to vary little from pasture without
palms. Given the wide spacing of palms within the association, it is
the quality of grass cover that should most affect erosion.
Forests
The Sierra's remaining forests, found primarily in the upper and
lower altitudinal tiers, are poorly managed and dwindling. In the
middle zone, forests are limited to small, isolated plots of pine
trees, successional stands in scrub, or ribbons of riparian forest
along major rivers and streams. Two large stands of second-growth
pine are found in recently acquired state lands spanning the mid- to
upper-altitudinal zones.
Forest resources are both underutilized and overexploited.
Legally, the residents are prohibited from felling any trees without
hard-to-obtain permits. Practically, they are engaged in the "mining"
of remaining forests for lumber, fuel, and expansion of agricultural
lands.
Prohibitive regulation of forest utilization leaves little
incentive for reforestation and offers no opportunity for wise
management based upon technical expertise, local needs, and external
markets for forest products. Subsequent raw material shortages have
contributed to unemployment. Many former artisans and sawmill


89
separate sample container and identified by relative position in the
vertical series as well as by site, date and the estimated peak stage
of the flood. Each sample was assigned to a specific stage of the
rising flood. The samples were sent to the CENDA laboratory for
determination of suspended sediment concentration, as described
previously for river water samples.
Determination of discharge and sediment transport rates from
sampling results. Approximate sediment transport rates (Sa) for each
stage were calculated by multiplying the sediment concentration (Cs)
by an approximation of discharge (0):
-1 -3 3 -1
S(tons sec ) = Cs(tons m ) x Q(m sec ).
Discharge was calculated using the empirically derived stage-discharge
equation for the given site. Stage height was based on the number of
bottles filled in the vertical series on the sampler. Stage was
assumed to be halfway between the intake of the last bottle filled and
the intake for the next bottle up. If the highest bottle filled, the
stage was recorded as greater than or equal to the intake height plus
10 cm.
Total sediment transport (Ts) for individual flood events was
calculated by summing the products of discharge rate (Q) times
sediment concentration (Cs) times the estimated duration of each stage
(t) :
Ts(tons) = Q(m^ sec "*") x Cs(tons m x t (sec) .
Approximate hydrographs of specific events were constructed by
reference to the measured peak height of the flood stage and reports
on duration of the rising and falling stages of the flood. The annual


412
. 1980. El rol del Servicio Nacional de Conservacin de
Suelos. Io Seminario Sobre Mane30 de Cuencas Hidrogrficas,
Santo Domingo, Dominican Republic, May.
Pavoni, J. L. 1977. Handbook of water quality management planning.
Van Nostrand Reinhold Co., New York.
Pereira, H. C. 1973. Land-use and water resources, Cambridge Univer
sity Press, New York.
Pereira, H. D., P. H. Hosegood, and M. Dagg. 1967. Effects of tied
ridges, terraces and grass leys on a laterite soil in Kenya.
Exp. Agrie. 3:89-98.
, J. S. G. McCullock, M. Dagg, 0. Kerfoot, P. H. Hosegood, and
M. A. C. Pratt. 1962. Hydrological effects of changes in land-
use in some East African catchment areas. E. Afr. Agrie. For. J.
27:42-75.
Pessar, P. R. 1981. The role of households in international migra
tion. Conf. on New Directions on Immigration and Ethnicity,
Duke University, Dgrham, N.C., May.
Pimentel, D. 1973. Food production and the energy crisis. Science.
182:443-449.
Plan Sierra. 1979. Informe del primer semestre de operaciones, Plan
Sierra. Secretaria de Estado de Agricultura, San Jos de Las
Matas, Santo Domingo, Dominican Republic.
Plumwood, V., and R. Routley. 1982. World rainforest destruction--
the social factors. The Ecologist 12:4-22.
Posner, J. L., G. Antonini, G. Montanez, R. Cecil, and M. Grigsby.
1981. Un sistema de clasificacin para las areas de ladera y
altiplanos de America Tropical. In A. R. Novoa, and J. L.
Posner (eds.) Agricultura de ladera en American Tropical.
CATIE, Turrialba, Costa Rica.
Quezada, N. A. (ed.). 1977. Desarrollo rural en la Sierra, Repblica
Dominicana. Centro de Investigaciones Econmicas y Alimenticias,
La Herradura, Dominican Republic.
Rabinovitch, J. 1979. Potential conflicts in land and water use in a
tropical river basin: The Guri hydroelectric project, Venezuela.
In Water management and environment in Latin America, United
Nations Economic Commission, Pergamon Press, Oxford, England.


I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.
R. B. Marcus
Professor of Geography
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.
H. I. Safa U
Professor of Latin American
Studies
This dissertation was submitted to the Graduate Faculty of the
Deaprtment of Geography in the College of Liberal Arts and Sciences
and to the Graduate School, and was accepted as partial fulfillment of
the requirements for the degree of Doctor of Philosophy.
August 1984
Dean for Graduate Studies and
Research


Application of Erosion and Runoff Coefficients
to the Small Watersheds 265
Sediment Delivery Ratios 275
CHAPTER V CONCLUSIONS 278
APPENDIX A COMPARATIVE DATA FROM LITERATURE REVIEW 285
APPENDIX B SURVEYED CROSS SECTIONS OF RIVERS AND STREAMS.... 297
APPENDIX C STAGE DISCHARGE CURVES FOR RIVERS, DERIVED BY
INDRHI 302
APPENDIX D MONTHLY RAINFALL FOR STATIONS 3 THROUGH 11 305
APPENDIX E DATA ON SEDIMENT CONCENTRATION, STAGE AND
DISCHARGE FOR MAO AND AMINA RIVERS 315
APPENDIX F DATA ON SEDIMENT CONCENTRATION, STAGE AND
DISCHARGE FOR SMALL WATERSHEDS 343
APPENDIX G SOIL PROFILE DESCRIPTIONS FOR EROSION PLOT SITES. 350
APPENDIX H FORMS USED FOR INFILTRATION TESTS 367
APPENDIX I DERIVATION OF FACTORS FOR USE IN USLE, BY PLOT 369
APPENDIX J DATA FROM EROSION PLOTS 378
LITERATURE CITED 394
BIOGRAPHICAL SKETCH 420
v


161
would reduce the sediment transport to approximately half of the
annual total which is at least plausible for a 50-year flood.
The severity of sedimentation problems in the proposed dam on the
Amina will depend upon the storage capacity of the dam and also on the
occurrence of extreme events such as the one discussed above. The
addition of the latter to the annual total would bring the sediment
transport to 74,900 tons for 1981. The importance of accounting for
such occurrences has been demonstrated in the case of the Valdesia Dam
which was choked with sediment by the runoff from Hurricanes David and
Frederick in 1979 and required dredging to resume operation. The
sediment load in the Tavera Dam also was substantially increased by
sediments transported during the same flood events. A large
proportion of the sediments transported during such events are re
suspended from channel deposits. They represent the cumulative
deposition of prior erosion. For this portion of total sediment yield
the best treatment is a reduction in the amount and discharge rate of
storm runoff and/or the construction of upstream barriers to keep
subwatershed sediment yield from entering the main stream and reaching
the dam. The reduction of the normal sediment yield is better
accomplished by changes in land use and practices upstream, starting
in the watersheds of critical concern identified earlier.
The average sediment yield of the large watersheds gives some
basis for subsequent comparison within and between watersheds. The
sediment yield for Amina is lower than for Mao by a factor of 2.4
(Table 9). The values of 3.1 and 1.3 tons ha 1 yr 1 exceed the
reported yields for forested watersheds in the U.S. (Table A-2) with


Table A-l. Suspended sediment concentrations in streams draining
forested watersheds in the southeastern United States.
Location
Forest type
Concentration
-1
Mg L
South Carolina Piedmontf
Loblolly Pine
20-43
Mississippi Coastal PlainJ
Pine forest
54-269
North Central Florida§
Pine flatwoods
21-81
fSource:
JSource:
§Source:
Switzer and Nelson, 1972.
Duffy et al., 1978.
Reikerk et al., 1979.
285




Table 10. Characteristics of the small watersheds.
Stream
Bajamillo
Bajamillo headwaters
Prieto
Pananao
Hondo
(no. 70)
(no. 71)
(no. 64)
(no. 60)
(no. 67)
Settlements
Rincon de Piedras,
Rincon de Piedras
Carrizal
Pananao
Los Montones/
Las Piedras,
Carrizal
San Jose
Area (ha)
2962.5
95.0
187.5t(490.0)
1312.5
1285.0
Average slope
37%t
43%
48% (42%)
32%
31%
Life zone
Subtropical very
Subtropical very
Subtropical
wet forest
wet forest
very wet forest
Land use system
Coffee-food crops
Coffee-food crops
Coffee-food
Pasture-
Pasture-
crops
food crops
food crops
% Area in:
Coffee
20§
39.9
48.4 (36.6)


Plantains^
5
12.9 ( 5.9)
Field crops
30
15.0
20.4 (27.4)
15.8
15.7
Pasture
30
30. 7
13.8 (19.9)
46.4
41.3
Bush fallow
4
9.9

19.0
18.6
Forest
10
4.0
4.5 (10.2)
18.6
23.2
Roads and
houses
1
0.4
0.5 ( 0.5)
0.5
1.1
Population:#
Residents
125
225 (625)
1100
1250
Seasonal
workers
40
60


fvalues in parentheses refer to larger Prieto watershed and Carrizal community. The first values refer to
the monitored portion only. ^Estimate. §The land use distribution for no. 70 was estimated. All others
were measured. ^Refers to plantains and/or bananas. #Refers to watershed boundaries, not to the political
subdivision. In the upper Bajamillo and Prieto only a part of the community is included.
167




82
the number of observations. The bridge sites allowed simple grab
sampling during flood conditions.
Initial results from multiple samples stratified throughout the
river cross-sections showed little variation across the section but
pronounced differences with depth. Subsequent samples were taken in
pairs at 10 and 30 cm below the surface (water depth permitting) at
the center of the section. The samples normally were drawn by
extending the sampler downward from the upstream side of the bridge,
directing the intake into the flow, and extracting the sample just
prior to filling the bottle, to avoid sampling error.
The 1-2 L samples were stored at room temperature at field
headquarters, then shipped periodically to the CENDA laboratory.
Sediment content by dry weight was determined and reported as
concentration (g L ^). Analyses were conducted according to standard
methods for the determination of suspended solids in water. After the
sample volume was recorded, the samples were shaken, then poured into
Gooch crucibles lined with pre-weighed Whatman No. 5 fiberglass filter
paper. After the sample was strained into a flask, under a slight
vacuum, the filter and collected sediments were oven-dried at 105UC
for 24 hrs, then weighed. The final weight minus the previously
determined paper weight gave the new weight of the sediments for a
given sample. The latter then was divided by the sample volume to
obtain the sediment concentration.
Sediment discharge rates and total sediment transport.
Instantaneous rates of sediment discharge in tons sec ^ were
calculated by multiplying total river discharge (m sec ^) by the
sediment concentration (tons m ):




328
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
810519
1.51
1.38
-
-
-
-
-
114.6
95.3
105.0
-
810520
1.26
1.20
-
-
-
-
-
79.0
71.5
75.3
-
810521
1.16
1.11
-
-
-
-
-
66.7
60.9
63.8
-
810522
1.09
1.02
1.40
-
-
-
-
58.7
51.2
54.9
93.9
810523
1.10
1.17
-
-
-
-
-
59.8
67.9
63.8
-
810524
1.00
0.98
1.82
-
-
-
-
49.2
47.2
48.2
134.7
810525
1.47
1.34
1.74
-
0.336
0.
252
0
.29
108.4
89.7
99.1
126.6
810526
1.31
2.26
2.60
64
2.952
2.
807
2
.88
85.6
262.4
174.0
220.0
810527
1.33
1.18
1.53
-
-
-
-
88.3
69.1
78.7
106.0
810528
1.23
1.64
-
61
1.099
1.
239
1
.17
75.2
135.8
105.5
-
810529
1.35
1.23
-
-
-
-
-
91.1
75.2
83.1
-
810530
1.13
1.08
1.48
-
-
-
-
63.2
57.6
60.4
101.3
810531
1.12
1.07
1.70
-
-
-
-
62.0
56.5
59.3
122.6
810601
1.16
1.08
1.78
99
0.454
0.
363
0
.41
66.7
57.6
62.1
130.6
810602
1.19
1.40
1.64
90
0.169
0.
096
0
.13
70.3
98.1
84.2
116.7
810603
1.16
1.10
1.28
-
-
-
-
66.7
59.8
63.2
83.0
810604
1.12
1.08
-
-
-
-
-
62.0
57.6
59.8
-
810605
1.05
0.98
1.44
-
-
-
-
54.3
47.2
50.7
97.6
810606
1.12
1.35
1.24
-
-
-
-
62.0
91.1
76.6
79.4
810607
1.06
2.30
-
-
-
-
-
55.4
272.1
163.7
-
. 810608
1.16
1.10
1.37
-
-
-
-
66.7
59.8
63.2
91.1
810609
1.17
1.00
-
-
-
-
-
67.9
49.2
58.4
-
810610
0.95
0.96
-
90
0.363
-
-
44.2
45.2
44.7
-
810611
0.90
0.87
-
-
-
-
-
39.6
36.9
38.2
-
810612
0.85
0.84
-
-
-
-
-
35.2
34.3
34.8
-
810613
0.83
0.81
1.36
-
-
-
-
33.5
31.9
32.7
90.2
810614
0.81
0.81
1.50
-
-
-
-
35.2
31.9
33.5
103.2
810615
0.92
0.84
-
-
-
-
-
41.4
34.3
37.9
-
810616
0.85
0.81
1.44
-
-
-
-
35.2
31.9
33.5
97.6
810617
0.96
0.84
1.40
-
-
-
-
45.2
34.3
39.8
93.9
810618
1.00
0.84
-
-
-
-
-
49.2
34.3
41.8
-
810619
0.84
0.80
1.24
-
-
-
-
34.3
31.1
32.7
79.4
810620
0.84
0.81
1.10
-
-
-
-
34.3
31.9
33.1
59.8
810621
0.83
0.80
-
-
-
-
-
33.5
31.1
32.3
-
810622
0.83
0.82
-
-
-
-
-
35.5
32.7
33.1
-
810623
0.75
0.74
1.42
-
-
-
-
27.2
26.5
26.8
95.7
810624
1.04
0.85
1.10
-
-
-
-
53.3
35.2
44.2
59.8
810625
0.85
0.81
1.19
-
-
-
-
35.2
31.9
33.5
70.3
810626
0.89
0.84
1.38
-
-
-
-
38.7
34.3
36.5
92.0
810627
0.90
0.85
1.09
-
-
-
-
39.6
35.2
37.4
58.7
810628
0.86
0.83
1.66
-
-
-
-
36.0
33.5
34.8
118.6
810629
0.98
1.80
1.98
-
-
-
-
47.2
164.5
105.8
151.2
810630
1.15
1.23
1.57
-
-
-
-
65.5
75.2
70.4
109.9


402
Grainger, A. 1980. The state of the world's tropical forests.
The Ecologist. 10:6-54.
Greenland, D. J. 1974. Evolution and development of different types
of shifting cultivation. In Shifting cultivation and soil
conservation in Africa, FAO, Rome.
. 1975. Bringing the green revolution to the shifting
cultivator. Science. 190:841-844.
, and R. Lai (eds.). 1977. Soil conservation and management
in the humid tropics. John Wiley and Sons, New York.
Gregory, K. J., and D. E. Walling. 1973. Drainage basin form and
process. A geomorphological approach. John Wiley and Sons,
New York.
Grinnell, H. R. 1977. A study of agri-silviculture potential in
West Africa. IDRC, Ottawa.
Grover, N. C., and A. W. Harrington. 1943. Streamflow. John Wiley
and Sons, New York.
Haggett, P. 1961. Land use and sediment yield in an old plantation
tract of the Seira do Mar, Brazil. Geog. J. 127:50-62.
Haith, A., and J. V. Dougherty. 1976. Non-point source pollution
from agricultural runoff. Proc. of ASCE Journal of the Environ
mental Engineering Division. 102:EES.
Hall, C., and J. W. Day, Jr. 1977. Systems and models: Terms and
basic principles. In C. Hall and J. W. Day, Jr. (eds.) Ecosystem
modeling in theory and practice: An introduction with case
histories. John Wiley and Sons, New York.
Harris, M. 1973. The withering green revolution. Nat. Hist. 83:
20-23.
Harrold, L. L., D. L. Brakensiek, J. L. McGuinness, C. R. Amerman,
and F. R. Briebelbis. 1962. Influence of land use and treat
ment on the hydrology of small watersheds at Coshocton, Volio,
1938-1957. U.S.D.A. Tech. Bull. 1256, Washington, D.C.
Hart, R. D. 1980. Agroecosistemas: Conceptos bsicos. Centro
Agronmico Tropical de Investigacin y Enseanza, Turrialba,
Costa Rica.


Table A-10. Soil losses in tea plantations compared to alternative land uses (in tons ha
year"l).f
Nation
Land Use
Forest Tea with Cover Crops Tea Vegetables Bare Fallow
Malaysia^
3.2 6.4 9.8
India§
o.4 to 2.1 3.1 to 14.4
Sri Lankafl
20.0 to 35.0 4.0
fSoil loss is converted to tons hal year_l using a soil density of 1.3 g cc-3 if data is reported
in m3.
|Data for Malaysia are reported in Lai (1977b).
§Data for India are reported in Lai (1977b) and Hasselo and Sikurajapathy (1965).
lData for Sri Lanka are reported in Lai (1977b) and Holland and Joachim (1933) .




41
effect on runoff, it stopped soil erosion, even on the 15% slopes
(Lai, 1977b). Experiments in Nigeria demonstrated the effectiveness
of alternative methods of field preparation and planting. While
croplands with ridges oriented downslope yielded 28% runoff and 20
tons ha ^ of soil loss, alternate tied ridges across the slope reduced
runoff to 13% and soil losses to 6 tons ha ^ yr ^ (Kowal, 1970).
One West African study reported on the continuous measurement of
erosion in the same plots over several years. Lai (1977b) found that
slope effects may be reversed after a few years. After a rapid
initial loss of the topmost layers on steep inclines the erosion rate
decreases, while gentler slopes maintain a more constant erosion rate.
This indicates the importance of documenting the land use history of
hillslope sites so as to account for the influence of past soil loss
and profile modification.
More detailed surveys of the published West African soil erosion
literature have been complied by Lai (1977a), Okigbo (1977), Fournier
(1967), and Jones and Wild (1975). Projects in progress include
minimum tillage and multiple cropping experiments in plots at IITA
(Lai et al., 1979).
There is also substantial similarity between some of the lowland
dry forest and montane ecosystems of East Africa and the Caribbean.
The farming systems have some crops and practices in common, though
fewer than in the case of West Africa.
Erosion plot studies in Uganda yield similar results to the
experiments already cited in West Africa and Latin America (Table A-9)
(Sperow and Keefer, 1975). The major difference is in the magnitude


92
were carried out over a much larger area than the erosion plots, and
each landholding was managed by the owners and workers who normally
oversee and perform work on the sites. The erosion plots received no
special treatment.
Plot design. The plots measured 22 x 2 m, which approximates the
Soil Conservation Service plots used in the United States (24 x 1.8 m)
and meets the standard specifications used by international research
organizations (FAO, 1977). In the three forested tracts the plots
were widened to 22 x 3 m in order to accommodate ..tree roots and to
include a more representative sample of the vegetation within the
plot.
The plots and collectors (Fig. 12) were designed after Djorovic
2
(1977) and Dunne (1977). The 44 m rectangular catchments were
oriented parallel to the slope and were bordered by single layer
concrete walls on three sides. The blocks were set into the ground
for secure footing, then sealed with mortar to prevent runoff from
entering or leaving. At the lower boundary of each plot a wooden
plank anchored a sheet metal apron that drained the runoff from the
plot (at soil surface level) to a funnel, through plastic tubing, and
into a modified oil-drum for storage.
Most sites were equipped with two tanks in series, so that
overflow from the first tank drained into the second. One site
(number 95, Pananao) required three tanks due to the high proportion
of runoff and low water storage capacity at the site. A modification
of the design also was required for plots with deep hillside ditches
cut across the slope. This was the case in the two plots planted to


141
Surface II programs using nearest neighbor analysis (Figs. 27 and 28).
The distribution for the study period showed no substantial variation
from the mean annual rainfall map.
Rainfall Distribution and Rainfall/Discharge Relationships
The allocation of rainfall data from the climatological stations
to the subwatersheds (Fig. 7) provided a basis for comparison of total
rainfall, rainfall rates, and ratio of discharge to rainfall volume in
both basins. The monitored portions of basins consisted of the areas
upstream of the hydrometric stations and the future dam sites on both
rivers (Fig. 8). Subsequent references to the watersheds will include
subwatersheds M. through M. for Mao and A_ through A_ for Amina unless
lo 3 5
otherwise stated.
Comparison of the monthly rainfall and discharge for both the Mao
and Amina watersheds (Table 4) showed a relatively high rate of
discharge. Total annual discharge is more than 50% of rainfall
volume. Reports from other tropical and subtropical watersheds
indicate much lower proportions of total rainfall discharged in
streams in forested watersheds (Golley et al., 1975; Odum, 1971;
Pereira, 1973).
The high ratio in this case can be attributed to the interaction
of topography and land use. The extensive deforestation that has
occurred in the past 20 years has exposed the thin soils on steep
slopes to a highly "aggressive" climate (Paulet, 1978) resulting in
erosion and reduced soil moisture storage capacity. Much of the area
(30-40%) is in pastures. Soils have been compacted in many areas by


389
Plot 92. Sisal and Food Crops, Pananao.
Sample Collection
Sediment Yield
, -1 >
Runoff Rate
3 -1
Date
(kg ha )
(m ha )
80
05
05
2537.27
49.59
80
05
12
1280.51
23.68
80
05
16
2881.04
30.73
80
05
18
563.48
13.61
80
05
24
3518.51
49.64
80
05
28
472.57
47.87
80
06
01
1859.98
47.58
80
07
07
1082.18
48.20
80
08
20
1142.60
91.76
80
08
26
1193.76
32.70
80
08
28
17.29
23.68
80
09
08
1086.84
42.45
80
09
29
26.68
23.68
80
10
04
2645.99
22.61.
80
10
10
2073.79
66.03
80
10
15
0.00
0.00
80
10
20
3273.86
9.19
80
10
21
33.67
9.19
80
11
07
54.60
45.00
80
12
10
2.27
1.51
80
12
11
3.76
1.51
81
01
08
63.18
1.51
81
01
12
2130.13
41.20
81
01
28
2.29
4.57
81
02
01
303.38
29.23
81
02
06
11.88
10.90
81
02
21
7.03
13.61
81
04
29
2.71
10.03
81
05
02
4.70
3.24
81
05
06
203.34
92.57
81
05
10
33.41
18.46
81
05
14
27.37
70.86
81
05
24
9.90
45.00
81
05
28
27.53
63.45
81
06
16
1.25
2.05










Location:
Pananao
Plot #:
93
Land Use:
Pasture
Land Cover:
Improved pasture,
overgrazed
Parent Material
: Conglomerate
Relief:
Pronounced
Position:
Mid-slope
Slope Class:
25-45%
Runof f:
Rapid
Permeability:
Moderate
Erosion Class:
Slight to moderate
Drainage Class:
Rapid (4)
Soil Moisture:
Uniform, slightly
moist
Salts/Alkaline:
None
Stoniness:
Slight to moderate
Soil Profile:
Horizon
A
C
Depth
0-24 cm
24-56 cm
Boundaries
sharp
sharp
Color
-
-
Texture
sandy loam
sandy
Structure
granular
granular
Consistency
-
-
C3
no
no
Concretions
no
no
Slickensides
no
no
Mottles
no
no
Roots
yes
yes


380
Plot 82. Pigeon pea, Minimum Tillage, Los Montones.
Sample Collection
Date
Sediment Yield
(kg ha ^)
Runoff Rate
, 3, -1.
(m ha )
80
08
13
114.18
48.134
80
09
13
2339.70
123.597
80
09
25
1259.62
9.730
80
09
25
38.30
9.730
80
10
30
183.85
6.724
80
10
30
0.00
7.883
80
11
06
1820.03
2.039
80
12
02
4.17
9.763
80
12
10
12.53
5.157
80
12
20
5.22
1.605
80
12
22
6.15
7.883
80
12
23
7.33
3.549
81
01
08
27.34
33.384
81
01
13
72.37
5.255
81
01
19
1.26
3.058
81
01
30
7.86
6.145
81
02
05
1.44
18.041
81
02
12
0.19
1.758
81
02
23
31.75
59.452
81
03
11
1.36
4.530
81
03
16
9.44
20.412
81
03
26
7.55
16.573
81
03
27
3.30
6.186
81
04
03
4.41
7.883
81
04
08
4.34
6.186
81
04
09
2.47
6.186
81
04
14
3.31
5.069
81
04
18
0.24
0.586
81
04
28
11.03
23.460
81
05
05
7.75
18.946
81
05
06
6.13
37.469
81
05
12
137.00
65.613
81
05
28
72.00
6.145
81
06
10
224.90
15.237


318
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800325
0.29
0.29
-
-
-
-
-
5.5
5.5
5.5
-
800326
0.29
0.29
-
-
-
-
-
5.5
5.5
5.5
-
800327
0.29
0.29
-
145
0.013
0.013
0.013
5.5
5.5
5.5
-
800328
0.28
0.28
-
-
-
-
-
5.2
5.2
5.2
-
800329
0.28
0.28
-
-
-
-
-
5.2
5.2
5.2
-
800330
0.28
0.28
-
-
-
-
-
5.2
5.2
5.2
-
800331
0.28
0.28
-
-
-
-
-
5.2
5.2
5.2
-
800401
-
-
-
-
-
-
-
-
-
-
-
800402
-
-
-
-
-
-
-
-
-
-
-
800403
-
-
-
-
-
-
-
-
-
-
-
800404
-
-
-
-
-
-
-
-
-
-
-
800405
-
-
-
-
-
-
-
-
-
-
-
800406
-
-
-
-
-
-
-
-
-
-
-
800407
-
-
-
-
-
-
-
-
-
-
-
800408
-
-
-
-
-
-
-
-
-
-
-
800409
-
-
-
-
0.224
-
-
-
-
-
-
800410
-
-
-
-
-
-
-
-
-
-
-
800411
-
-
-
-
-
-
-
-
-
-
-
800412
-
-
-
-
-
-
-
-
-
-
-
800413
-
-
-
-
-
-
-
-
-
-
-
800414
-
-
-
-
-
-
-
-
-
-
-
800415
-
-
-
-
-
-
-
-
-
-
-
800416
-
-
-
-
-
-
-
-
-
-
-
800417
-
-
-
-
-
-
-
-
-
-
-
800418
-
-
-
-
-
-
-
-
-
-
-
800419
-
-
-
-
-
-
-
-
-
-
-
800420
-
-
-
-
-
-
-
-
-
-
-
800421
-
-
-
-
-
-
-
-
-
-
-
800422
-
-
-
-
-
-
-
-
-
-
-
800423
-
-
-
-
-
-
-
-
-
-
-
800424
-
-
-
-
-
-
-
-
-
-
-
800425
-
-
-
-
-
-
-
-
-
-
-
800426
-
-
-
-
-
-
-
-
-
-
-
800427
-
-
-
-
-
-
-
-
-
-
-
800429
-
-
-
-
-
-
-
-
-
-
-
800430
-
-
-
-
-
-
-
-
-
-
-
800501
0.75
0.80
1.98
-
-
-
-
27.2
31.1
29.1
151.2
800502
1.32
1.00
1.41
-
-
-
-
87.0
49.2
68.1
94.8
800503
1.17
1.04
1.63
-
-
-
-
67.9
53.3
60.6
115.7
800504
0.98
0.85
-
-
-
-
-
47.2
35.2
41.2
-
800505
0.77
0.73
1.40
-
-
-
-
28.7
25.7
27.2
93.9


128
to plant biomass, capital inputs (K^g) and animal population. Farmers
sell some food and fuel (K ) to lowland markets. The amount sold
depends on total plant biomass and the price of the products in the
marketplace.
The growth in the number of beef and dairy cattle is determined
by the interaction of the initial herd size (as biomass, A), high
technology inputs for improvement of herd and pasture (K ), and
available biomass for grazing (K ) The energy and material exports
from the animal population include energy expended in self-maintenance
(K^g), slaughter and milking for subsistence consumption (K^g), and
sale of animals, meat and milk to outside markets (K^). Local meat
and milk consumption depends upon the population sizes of both animals
and consumers. The amount exported varies directly with the size of
the herd and the current prices of these products in external markets.
The population (P) of the area depends upon initial population
size (P) and growth rate, consumption of plant and animal biomass
within the region (K^^), and imported subsistence goods (U) (K^^).
Drains on population include energy expended for self-maintenance
(K ), labor (K ), death (K ), and emigration (K ). Seasonal
j> / u jo
migration may bring farm workers into the region from other areas, or
it may be a mechanism for local residents to earn cash income outside
the region (K ).
The imported subsistence goods are purchased with money (M) that
represents the total liquid assets available to the local population.
Income consists of profits from coffee other cash and food
2 jb
crop sales (K^)f animals and animal products (Kg.), and remittances










Form used for field data collection:
Soil type
Date
Stop No.
Classification
| Area
Location
| Elcv.
N. veg. (or crop)
| Climate
Parent material
Physiography
Relief
Slope
Erosion
Drainage
Gr. water
Permeability
Moisture
Salt or alkali
Stoniness
Root distrib.
Remarks
Hori
zon
Depth
Thick
ness
Bound
ary
Color
Tex
ture
Struc
ture
Con
sistence
Reac
tion
Spec.
Feat.
Check. D(ry)
or M(oist)
D
M
D
M
D
M
D
M
D
M
D
M
D
M
D
M
D
M
D
M
D
M
B
USDA, 1951.
350


57
an account of the farm's inputs and outputs as well as of activities
and movement of materials within the farm itself. These records also
assist in the evaluation of on-farm trial results and provide basic
data for further trials and/or discussions with farmers. The success
of a new technology is judged at least in part by the farmer's
perception of its performance and by its subsequent adoption by him
and other farmers in the area. This indicates t some extent the
"fitness" of a technology for the farming system as a whole, at the
farm level (Swisher et al., 1982).
The sondeo as well as the subsequent on-farm trials and farm
record-keeping of the FSR/E approach are readily adaptable to soil and
water conservation research in hillslope environments in the
Caribbean. The major elements lacking are greater attention to
natural resource management at the farm level and a method for
predicting and evaluating the success of a given technology at the
watershed or regional level.


36
50% increase in runoff, measured as streamflow (Edwards and Blackie,
1975; Edwards, 1977).
Findings from two other study areas in Tanzania, one a cultivated
montane catchment and the other a series of catchments in semi-arid
savanna, offer comparative data on river regimes as well as on
suspended sediment concentrations under changing land use practices
(Rapp, 1977). The catchments represent a complex mosaic of forest,
farm, pasture and bush, which is comparable to many upland catchments
in the Caribbean. During flood peaks, sediment concentrations ranged
from 2000 to 3500 mg L ^ in the upland areas, and from 15,000 to
75,000 mg L ^ in the semi-arid catchments. Flash floods and high
sediment loads in both the montane and savanna areas were attributed
to land use. A comparison of the relationship between drainage area
and sediment delivery ratio in the United States and Tanzania shows
much less reduction in sediment yield with increased drainage area in
the Tanzanian catchments.
Total streamflow and sediment yield were determined by the use of
flow duration and sediment rating curves along with stream gauge and
sediment concentration data. Sampling was carried out with a home
made point-integrating hand operated sampler (Nilsson, 1969; Rapp,
1977) and an automatic multi-stage sampler designed for ephemeral
streams. Both the instruments and the methods of analysis used in
this study are feasible for use in the Caribbean.
Reservoir sedimentation. Studies of reservoir sedimentation and
other aspects of regional water utilization and management often
approach the situation as an engineering and economic problem, either






APPENDIX J
DATA FROM EROSION PLOTS






APPENDIX E
DATA ON SEDIMENT CONCENTRATION, STAGE AND
DISCHARGE FOR MAO AND AMINA RIVERS


244
years. Plots 82 and 83 have been under a rotation of subsistence food
crops and pasture for 14 years, and plots 84 and 85 were cleared,
burned and planted just at the onset of the study. The runoff results
thus confirm the tentative conclusions advanced earlier to explain the
infiltration test results. Time under cultivation seems to exert a
strong influence on infiltration and storage capacity.
Erosion. The soil loss groupings were more general with plots
91, 92, and 95 (Pananao) making up the first category and plots 82,
83, 84, and 85 constituting the second group (Table 35). If inherent
site characteristics are the determining factors, the most likely
causes of variation would be climate or soil type. The former
decreases the erosion potential for Pananao versus Los Montones. Soil
type is similar, particularly in texture. The fact that pasture and
forest do not differ significantly between sites also suggests that
inherent site characteristics are not the differentiating variables.
The difference in cropping systems, farming practice and land use
history is a more plausible explanation. In particular, the intensive
cultivation required for peanut crops has had a strong effect on the
structural stability of soils at Pananao. This has been noted by the
farmers, many of whom regard their land as spent because of the period
when they devoted their fields to peanut production (Martinez, 1981).
The analyses discussed above are based on means of log-
transformed data as opposed to totals of raw data. Thus, the
initially high losses at plots 84 and 85 placed them first on the list
for total soil loss, but the sustained higher losses at 91, 92, and 95
placed them in the highest category based on means. The latter is a




410
Navarro, L. A. 1979. Seleccin y caracterizacin de areas como guia
a la investigacin agricola aplicada. Proc. Regional Seminar
on Methods for Development of Alternative Technologies in
Cropping Systems, San Salvador, El Salvador, July.
. 1981. Report presented to Plan Sierra, San Jose de las
Matas, Dominican Republic.
Negev, M. 1967. A sediment model on a digital computer. Rept. 76
Stanford Univ., California.
Nelson, M. 1973. The development of tropical lands: Policy issues
in Latin America. Resources for the Future Inc., The Johns
Hopkins Univ. Press, Baltimore.
New Zealand Ministry of Works. 1968a. Annual hydrological research
report for Moutere, 1. Wellington.
. 1968b. Annual hydrological research report for Otutira, 2.
Wellington.
. 1968c. Annual hydrological research report for Puketurua,
2. Wellington.
. 1970. Representative basins of New Zealand soil and water
divisions. Mise. Pub. 7, Wellington.
Nicholaides, J. J., Ill, and P. Hildebrand. 1980a. Proposed program.
USAID Title XII Soil Management CRSP Steeplands Project,
Washington, D.C.
, and 1980b. Trip report, Dominican Republic and
Peru. USAID Title XII Soil Management CRSP Steeplands Project,
Washington, D.C.
Nilsson, B. 1969. Development of a depth-integrating water sampler.
Dept, of Phys. Geog. Rep. No. 2, Uppsala.
Noll, J. J. 1953. The sitting of Caonillas Reservoir, Puerto Rico.
USDA Soil Conservation TP-119, USGPO, Washington, D.C.
Novoa, A. R., and J. L. Posner. 1981. Agricultura de ladera en
America Tropical. Centro Agronmico Tropical de Investigacin
y Enseanza, CATIE, Turrialba, Costa Rica.
OAS (Organization of American States). 1967. Reconocimiento
y evaluacin de los recursos naturales de la Repblica
Dominicana. Pan American Union, Washington, D.C.




74
samples. Soil classifications were confirmed and refined by soil
survey specialists from the Secretariat's South District Laboratory.
The relative infiltration rates of soils at the various sites
were determined by measurements with ring infiltrometers (Gregory and
Walling, 1973; Wisler and Brater, 1959). The inner ring was cut to a
25 cm diameter and the outer ring measured 40 cm across. After
placement in the ground with a minimum of soil displacement, the outer
ring was filled to form a barrier of saturated soil around the inner
ring which was filled to a 10-cm depth. _Throughout the next 4 hours a
nearly constant head of 10 cm was maintained while measurements of
water added were recorded at increasingly longer intervals. The form
used for the field measurements is included in Appendix H.
The frequency, distribution and severity of erosion features on
and around the plot sites were observed and noted prior to
construction of experimental plots. Wherever possible, the
developmental sequence of such features was determined from accounts
by the residents or neighbors and from repeated observation and
photographic records kept over the 15-month study period.
Detailed Analysis and Measurement of Key Parameters
The full characterization of the study areas at all three scales
of analysis served as a point of departure for the third phase of the
study, the measurement of water and soil exports from the individual
plots and from the nested sets of watersheds. Erosion, runoff, and
sediment transport were related to daily and continuous precipitation
measured at 15 stations in and near the Plan Sierra impact area.


Table 11.Continued.
Plot Number
93
94
95
96
Site
Pananao
Pananao
Pananao
El Rubiof
Slope (%)
35.0
31.0
39.0
30.0
Soil Classification
Troporthent
Troporthent
Troporthent
Troporthent
Soil Texture
Loam
Loam
Sandy loam
Sandy loam
Soil Fertility
% N
0.26
0.17
0.30
0.16
% OM
5.18
3.36
6.02
3.22
Infiltration
24.5
24.5
37.3
2.5
Vegetative Cover
Pasture
Pasture
Manioc
Pine forest
Conservation Practice
fThis nearby site serves as a forest surrogate for Pananao since no suitable forested areas remained in
Pananao.




Table 37. Comparison of measured and predicted erosion losses.
Plot
Number
Land Use
USLE
Prediction!
Measured
Erosion Losses^
tons
ha'1 yr'1
tons
ha-1 yr-1
80
Coffee (old)
37.8
0.5
- 3.3
81
Coffee (new)
1197.0 -
2993.0
1.4
- 2.3
82
Pigeon pea, Sweet potato§
195.6 -
880.1
6.0
- 6.4
83
Pigeon pea, Sweet potato§H
101.6 -
457.4
7.3
- 7.5
84
Yuca, Beans, Com
183.7 -
826.6
2.5
-48.3
85
Yuca, Beans, Com§
91.8 -
413.3
17.2
-70.0
86
Pasture
46.6 -
233.0
0.2
- 0.3
87
Pine Forest
1.3 -
4.5
0.1
88
Pine Forest
0.9 -
3.2
0.1
- 0.2
89
Pasture
37.9 -
189.3
0.2
- 0.3
91
Sisal, Yuca, Sweet Potato**
164.2 -
738.7
13.6
-30.4
92
Sisal, Yuca, Sweet Potato
320.1 -
1440.5
15.5
-28.3
93
Pasture
57.5 -
287. 3
1.1
- 2.0
94
Pasture
43.1 -
215.5
1.4
- 1.8
95
Yuca
273.6 -
1231.2
11.6
-21.0
96
Pine Forest
2.8 -
9. 3
0.1
fThe C factors for field crops and pasture were varied to cover a
conditions. The estimates are based on C factors applied in the
1978) and in the Dominican Republic (Santana, 1980).
listed in Appendix I.
^Annual totals are reported for May 1980 to May 1981 and for July 1980
§Site includes hillside ditches across the slope.
^Minimum tillage practice with grass cover was used.
broad range of plant cover
U.S. (Wischmeier and Smith,
The parameters for the equation are
to July 1981.


176
While the two small upland watersheds in coffee yielded a fairly
regular discharge of clear water upstream of the settlements, the
associated settlements and roads substantially increased flood peaks
and sediment discharge in the lower reaches of these streams. They
also contributed to disturbance of the river regime and the high
sediment load of the lower Bajamillo. This illustrates the difference
between streams draining coffee stands and those draining whole coffee
production areas. In the latter case, coffee production affects
runoff and erosion indirectly through the impact of market roads,
settlements, and related food production plots. As a result, the
Bajamillo resembles streams draining more intensively cultivated
watersheds in pastures and food crops.
While the upper Bajamillo and Prieto channels showed little
evidence of flood peaks beyond 1.5 m above the river bed (see Appendix
F), the channel erosion features and debris in the lower Bajamillo
indicated the recent occurrence of flood stages up to 3 m above the
usual base flow level. Sediment deposits ranging in texture from fine
to coarse sand were up to 10 cm deep in places along the stream
channel and floodplain.
Socioeconomic characteristics
Some aspects of economic interactions in the study area are
clearly expressed in the landscape by land use and land tenure. The
landholdings in Carrizal are dominated by two families that hold
approximately 50% of the land. The family with the largest
landholdings produces coffee as well as staples for the household and




228
Carrizal and Pananao showed significant differences by land use
category, with little significant variation by site (limited to
cropped plots). This supports the hypothesis that land use is the
ma^or determinant of erosion and runoff rates in the area, far
outweighing inherent site characteristics in importance. The
distribution of the specific groupings for erosion and runoff reveals
the underlying rationale for the separate land use coefficients for
erosion and runoff in the regional watershed models (Figs. 19 and 20).
Los Montones and Carrizal"-
The comparison of all plots at Los Montones and Carrizal showed
significant differences between plots for both runoff and soil loss
(Table 24). Duncan's Multiple Range Test (SAS, 1979) identified the
major groups (Table 25).
Runoff. The highest mean runoff was from coffee, with plots 81
2
and 82 forming a distinct subgroup. The second subgroup combines the
newly established coffee (81) with the partly grass covered minimum
tillage plot in pigeon peas (83). The third subgroup includes all
cultivated and pasture plots as distinct from coffee and forest.
Another subgroup combines forest, pasture and the recently cleared and
burned plots, separating them from the coffee and the pigeon pea
''Los Montones and Carrizal are grouped together, based on proximity,
in order to compare the existing coffee area to proposed sites for
establishment of coffee and other tree crops at Los Montones.
2
Data values are all log-transformed for analysis of variance as well
as for the a posteriori tests.








185
Fig. 37. Land use in Pananao watershed


351
Location
Plot
Land Use
Land Cover
Parent Material
Relief
Position
Slope Class
Runoff
Permeability
Erosion Class
Drainage Class
Soil Moisture
Salts/Alkaline
Stoniness
Sample Field Report Sheet
Soil Profile:
Horizon
Depth
Boundaries
Color
Texture*
Structure
Consistency
C3
Concretions
Slickensides
Mottles
Roots
-
*Note: for more
data as reported
accurate determination of texture, see laboratory
in Appendix I; see same for N content.


415
Santana, Q. 1980. Manejo de pastos para conservacin de suelos y
rotacin de cultivos. 1st Seminario de Manejo de Cuencas
Hidrogrficas, Santo Domingo, Dominican Republic, May.
Santos, B. 1981. El Plan Sierra: una experiencia de desarrollo rural
en los montanas de la Repblica Dominicana. In A. R. Novoa
and J. L. Posner (eds.) Agricultura de ladera en America Tropical.
CATIE, Turrialba, Costa Rica.
SAS. 1979. SAS user's guide. SAS Institute, Inc., Cary, N. C.
Schreiber, J. D., P. D. Duffy, and D. C. McClurkin. 1976. Dissolved
nutrient losses in stream runoff from five southern pine water
sheds. J. Environ. Qual. 5:201-205.
Schumm, S. A., and R. F. Hadley. 1961. Progress in the application of
landform analysis studies of semi-arid erosion. Geol. Surv.
Circ. 437, Washington, D. C.
SEA (Secretaria de Estado de Agricultura, Repblica Dominicana).
1978. Plan de Dessarrollo, Sierra. Santo Domingo, Dominican
Republic.
Sharpe, K. 1975. El campesino de la Sierra: El problema de vivir.
Eme-Eme 4:21.
. 1977. Peasant politics: Struggles in a Dominican village.
The Johns Hopkins Univ. Press, Baltimore, Md.
Sheng, T. C. 1973. Watershed management and soil conservation activi
ties in Jamaica: an evaluation report. UNDP/FAO, Kingston,
Jamaica.
, and T. Michaelson. 1973. Runoff and soil loss studies in
yellow yams. Forestry Development and Watershed Management in
Upland Regions. UNDP/FAO, Kingston, Jamaica.
Singer, M. J., and R. H. Rush. 1975. Phosphorus in surface runoff
from a deciduous forest. J. Environ. Qual. 4:307-311.
Skopp, J., and T. C. Daniel. 1978. A review of sediment predictive
techniques as viewed from the perspective of nonpoint pollution
management. Iri Environmental management. Springer-Verlag, New
York.
Smil, V. 1979. Energy flows in the developing world. Amer. Sci.
67:522-530.
Smith, R. W., and F. Abruna. 1955. Soil and water conservation res
earch in Puerto Rico, 1938 to 1947. Puerto Rico Agrie. Exp. Stn.
Bull. 124, San Jose, Puerto Rico.


279
The application of the information from the study to an analysis
of the Plan Sierra conservation programs indicates a need to place
more emphasis on curbing deforestation, which is the highest single
contributor of sediment, per unit area. In already deforested lands,
changes in cropping systems, conservation and integration of cropping
and forestry systems are necessary. The first two options have
received the most emphasis within Plan Sierra, and the third has been
limited primarily to promotion of combined coffee and food crop
production systems.
The soil conservation practices and land use changes advocated by
Plan Sierra have had varying results. The hillside ditches that have
proven so effective in other areas in an experimental setting (Table
A-7) failed to show significant improvement in erosion reduction at
all four plots. The technical and economic feasibility of
constructing and maintaining this type of physical infrastructure in
small farm plots were not included in the experimental analyses that
preceded the transfer and dissemination of this technology in the
Sierra. In many areas of the Sierra, the degree of slope, soil
texture, and soil structure are not well-suited to the successful use
of hillside ditches. The excavated ditches had filled in at most of
the sites prior to completion of the study (less than 15 months). The
seeding of the upslope banks with dense grasses or other cover crops
stabilized some structures in other fields. However, the investment
required for construction and maintenance of the ditches is still
high.
The testing of this technology on existing farm plots showed it
to be inappropriate in most cases for small farmers in the Sierra.




178
reinvestment is usually for home improvements, consumer goods,
vehicles, land, cattle, or establishing a local business. The money,
as a rule, is not channelled into increased coffee or food crop
production. The net result, as documented in Juncalito (located in
the upper Bao watershed), has been a decrease in food production and
an increase in the proportion of land held in pasture (Pessar, 1981).
Purchase of new lands with remittances is more often for speculation
than for increased production.
The effect of this change on the landless and small landholders
is a decrease in local employment opportunities (on food crop plots in
large landholdings), an increase in staple food prices, and an
increase in consumption of foods imported from the lowlands by local
entrepreneurs. This depression of food and forest production in favor
of pasture is reflected in Rincon de Piedras where a large percentage
of the land is in pasture and bush (Table 11). Carrizal has been less
affected because the head of household from the largest landholding
resides more than half-time in the area and continues to manage the
food crop plots as an integral part of his holdings.
Many of the residents work on larger landholdings as
sharecroppers, coffee harvesters, or day laborers. Seasonal migrant
farm workers also come to the area to work at harvest time. They
include residents of lowland areas nearby as well as jobless people
from the cane fields in the more distant lowlands.
Most of the residents produce some staples, especially plantains,
in small plots and depend upon wood and charcoal for cooking, light
and heat. The latter is a minor consideration in the Sierra, but is






7
the local level for the widespread adoption of resource management
practices that will benefit the population of the region as a whole
over the long term (Gladwin, 1981; Hildebrand, 1981; Santos, 1981).
Traditional cost-benefit analysis will not suffice since it
excludes environmental as well as social aspects of the system, assumes
a static system, and has been developed for application over relatively
short time periods (Amin, 1977). The problem requires a more holistic
theoretical and methodological approach that will evaluate
environmental and human concerns on their own terms and within the
total system rather than by econometric criteria.
The Sierra Region
The Sierra is a rugged montane area in central Dominican Republic
that has been subjected to the traditional practice of shifting
cultivation, as well as to extensive exploitation of primary resources
such as timber and mineral deposits. It is a relatively underdeveloped
region within an underdeveloped country where the area under production
2
in the country (27,000 km ) already has surpassed the area of land
2
classified as suitable for agriculture (22,000 km ) (OAS, 1967;
Swedforest, 1980; USAID, 1974). Production increases necessary to meet
the national demands for food and income have come from increased
yields in areas already under production, or from expansion into more
marginal areas such as the Sierra. The latter strategy has dominated
among the poor and landless members of the peasantry (Beckford, 1972;
Antonini et al., 1975), and has been a last resort for the former
employees of mining and sawmill camps and furniture shops, most of








8
which had closed by 1979 (Ferreiras, 1979). The major alternative,
emigration to the capital city of Santo Domingo, and to New York City,
provided an outlet for a large segment of the population during the
1960s and early 1970s.
The impact area of Plan Sierra (Fig. 2), an integrated rural
development project within the Sierra, offered a unique opportunity to
examine the apparent conflicts between agricultural development and
natural resource conservation (Santos, 1981). Plan Sierra is a joint
venture between the State Secretariat of Agriculture and the private
sector to initiate and coordinate development efforts in the region in
several sectors: agriculture, livestock, credit associations, health
services, transportation, handicrafts, university programs, and
natural resource management (Quezada, 1977; Antonini and York, 1979;
Plan Sierra, 1979). The Plan's objectives are: 1) to improve the
quality of life of the inhabitants of the Sierra; 2) to manage soil,
water and forest resources; and 3) to promote participation by local
people in the development process (Antonini and York, 1979; Santos,
1981) .
Several hydroelectric and irrigation projects are planned for the
study area (Jorge, 1981). The impoundments will serve the Cibao
Valley downstream. Sedimentation from this area is already a problem
in the Tavera Reservoir, completed in 1977 (Cepeda, 1980). The
magnitude and distribution of erosion in the uplands, however, has
received little attention until recently (Vasquez, 1980).
Plan Sierra has promoted specific farming practices and changes
in land use that are intended to reduce sediment export. A parallel






387
Plot 89. Pasture, Los Montones
Sample Collection
Date
Sediment Yield
(kg ha ^)
Runoff Rate
, 3, -i,
(m ha )
80
05
14
19.94
44.99
80
05
31
11.40
48.88
80
06
10
118.25
82.43
80
06
20
24.92
41.19
80
08
08
1.33
8.36
80
09
13
68.51
47.57
80
09
25
5.83
3.24
80
09
25
10.21
3.24
80
10
02
0.00
0.00
80
10
17
0.00
47.57
80
10
30
7.78
3.89
80
12
02
0.57
2.62
80
12
10
1.76
2.62
80
12
23
1.64
1.01
80
12
29
1.63
0.58
81
01
08
13.03
7.55
81
01
01
2.59
11.78
81
01
12
7.62
13.60
81
01
13
0.43
1.51
81
01
15
' 0.96
0.58
81
01
14
3.22
5.27
81
01
04
7.34
13.60
81
03
06
1.21
5.27
81
03
16
3.05
14.54
81
03
25
6.44
4.57
81
04
09
1.78
2.62
81
04
09
0.64
3.24
81
04
14
22.31
13.60
81
04
27
1.83
9.18
81
04
28
11.53
41.19
81
05
05
11.01
23.67
81
05
06
4.76
13.60
81
05
12
5.76
33.88
81
05
28
2.34
7.55








113
the Sierra, accounting for more than 5000 ha in 1978 (SEA, 1978).
More than 1900 farmers are engaged in coffee production, and the total
number of persons involved exceeds 10,000 (close to 10% of the
population) if their families are considered. The number of hired
laborers (estimated at 7000) puts the total at over 17,000. (This
estimate is based on figures from SEA and interviews with local
growers). This figure is not surprising since labor accounts for
approximately 80% of normal on-farm operating costs (excluding fixed
costs of land and infrastructure, and marketing costs).
The labor demand is markedly seasonal, and harvest, which is the
peak employment period, lasts two to three months. The laborers often
reside on or near the farm, supplementing their incomes by shifting
and/or bush fallow cultivation or other forms of temporary labor.
Some of the work force also migrates seasonally from the Cibao valley
or from more remote parts of the Sierra. This work force differs from
the usual one which is composed primarily of Dominican men in that it
includes women and children as well as some Haitian workers.
Coffee production also influences the local economy in a less
direct fashion. Traditionally, it has been carried out on medium
sized to large landholdings by families who support and supplement
this activity with adjacent holdings in pasture, field crops, and
forest. Many of the larger holdings are not contiguous; that is, one
household may have scattered large subdivisions of land in various
kinds of production up to two hours distance from the home. The large
farm owners often own and manage other small businesses such as coffee
processing and marketing operations, supply stores, informal credit,


21
areas (Chapman and Dunin, 1975; Mein, 1977; Fleming, 1968; Negev,
1967; Ebumive and Todd, 1976; Donigan and Crawford, 1976).
Theoretical Models of Physical Processes
Research conducted by Elwell (1979a) in Zimbabwe resulted in a
simple model similar to the USLE in some aspects, but based on
rational rather than empirical parametfers for estimating sheet erosion
from arable land. Rainfall energy measured in 10-day increments
defines erosivity while the protective value of crops and cropping
practices is assessed according to the percentage of seasonal rainfall
energy "i" intercepted by the vegetative canopy and ground cover.
Potential interception is determined in a fashion similar to the leaf
area index measured by ecologists (Odum, 1971).
Application of the model by Elwell (1979a) in Zimbabwe field
trials confirmed the importance of mulches for erosion control and
demonstrated the potential for reducing soil loss by increasing crop
yields. Moreover, important seasonal relationships were identified
between various protective crop covers and erosivity of rainfall in
given areas. This allowed the identification of the crops or crop
combinations that best protected the soil during periods of intense
rainfall.
The model is particularly applicable to the seasonally dry
tropics where single storm events cause a large proportion of the
total annual erosion loss. The practical application of the model is
aided by a description of the techniques for calculation and field
measurement (Elwell, 1979a).


145
overgrazing, further reducing infiltration and storage capacity of the
soils. Comparison of monthly rainfall and discharge data reflect the
limited soil moisture storage capacity. Rapid reductions in Amina
river discharge from May to June and in Mao river discharge from June
to July indicate that most of the surface and subsurface runoff
reaches the main rivers soon after major storm events.
Rapid runoff of rainfall represents a loss-of useful water for
upland agriculture and for downstream industrial, domestic and
irrigation uses. This same phenomenon increases the sediment delivery
efficiency of streams and contributes to high flood peaks and
floodplain damage downstream.
Subwatershed Analyses
The subwatersheds within the larger basins do not contribute
equally to river discharge and sediment load. The linear regression
of rainfall rates in each subwatershed on discharge and sediment
transport'- in the respective larger watersheds shows a wide variation
2
in the R value between subwatersheds (Table 5). It is the variation
in relative strength of relationship between subwatersheds that is
most important to note. The proportion of explained variation is low
for all cases, since same-day rainfall on each subwatershed is only
one of many factors that determine river discharge. Comparison of the
2
R values indicates which subareas contribute most to the effects
observed downstream.
All of the variables tested were Poisson distributed and were log-
transformed (In) for the regression analyses.






Fig. B-3. Upper Bajamillo Stream cross section.
Fig. B-4. Prieto Stream cross section.
Fig. B-5. Greater Bajamillo Stream cross section.
299






393
Plot 96. Pine Forest, near Pananao.
Sampli
e Collection
Sediment Yield
Runoff Rate
Date
(kg ha 1)
, 3, -1.
(m ha )
80
08
13
0.12
3.05
80
09
08
1.97
12.30
80
10
03
43.84
23.38
80
10
20
13.56
23.38
80
11
07
1.66
15.78
81
01
16
2.09
12.30
81
01
28
0.27
1.37
81
01
21
0.91
2.59
81
03
27
0.90
4.51
81
04
09
0.72
1.75
81
04
28
2.77
12.30
81
04
30
0.80
4.01
81
05
02
1.27
3.52
81
05
14
3.47
15.78
81
05
26
2.22
9.07


APPENDIX F
DATA ON SEDIMENT CONCENTRATION, STAGE AND
DISCHARGE FOR SMALL WATERSHEDS










79
gauged and calibrated periodically, such that the stage measurements
recorded at 0700 and 1700 hrs daily, as well as maximum flood stage,
could be converted directly to discharge rates based upon a nomograph
constructed by INDRHI hydrologists (see Appendix C). The stage-
discharge graphs relate discharge rates obtained through periodic
field measurements using the area-velocity method (Grover and
Harrington, 1943; Herschy, 1978) to the simultaneous reading of river
stage on a simple gauge.
For purposes of this study the nomographs of each river were
segmented and the equations for each segment were determined by simple
linear regression using the Statpak package of statistical programs
(MUSIC, 1967). The gauge readings recorded for morning, evening, and
flood peaks during the study period were converted to discharge rates
using the appropriate equations derived from the INDRHI nomographs.
Stage measurements made at the bridges by project personnel were
compared to simultaneous measurements at the hydrometric stations. A
simple linear regression equation converted the measurements made at
the bridge at sampling time to a gauge reading for the station. This
procedure replaced independent velocity-area discharge measurements
and allowed the substitution of a simple indirect method for the more
difficult and lengthy procedure used initially.
Comparison of rainfall and discharge during the study period to
the historical period. The total monthly rainfall totals over the
study period were compared to the average monthly rainfall totals over
the full period of record at stations with 11 or more years of data.
The SAS (1979) means program was used to compare the data from each


385
Plot !
37.
Pine Forest,
Los Montones.
Sample Collection
Sediment Yield
-1
Runoff Rate
Date
(kg ha )
(m^ ha )
80
05
14
1.57
7.26
80
05
31
6.88
28.29
80
06
11
5.32
12.97
80
06
20
2.76
15.78
80
08
08
1.92
10.97
80
09
13
7.65
12.97
80
09
13
0.-46
2.59
80
09
25
0.00
0.00
80
10
02
1.03
12.97
80
10
02
2.46
12.97
80
10
17
1.81
9.07
80
10
30
0.93
2.59
80
10
30
0.93
2.59
80
12
02
0.51
1.36
80
12
10
0.16
1.00
80
12
23
0.17
1.00
80
12
29
0.25
1.36
81
01
08
0.70
3.51
81
01
13
0.25
2.59
81
01
15
0.13
0.39
81
01
19
2.98
1.36
81
or
04
2.58
10.33
81
02
23
0.76
3.04
81
03
16
0.38
1.75
81
03
18
1.92
2.16
81
04
03
0.32
1.36
81
04
08
0.17
1.36
81
04
09
0.75
2.59
81
04
14
0.49
2.59
81
04
27
3.89
12.97
81
05
06
1.81
9.07
81
05
12
3.72
11.63
81
05
28
4.28
19.48
81
06
10
1.77
6.12


Legend for Tables F-l through F-5.
MAXLEVL
SED1 5
MAXDSCHG
FLDDSCHG
SEDDIS1 5
= maximum height reached by flood crest as indicated by
number of bottles filled, 1 -> 5 (bottom -> top) on
sampling equipment.
= sediment concentration (g L "S of stream water samples
from rising flood at levels 1-5.
3 -1
= discharge rate (m sec ) at flood peak, (or level 5,
maximum recorded stage).
3 -1
= flood discharge rate (m sec ) calculated for peak
annual flood stage (to be used as a high estimate if
level 5 is passed).
= sediment discharge rate in kg sec ^ at each sampled
stage of the rising flood, levels 1-5.
343


166
to the relative amounts of pasture and cropland, but the percent area
of land in coffee is approximately the same. The physical and land
use characteristics of both sites are summarized and compared with the
other small watersheds in Table 10. The distribution of land use
categories (Figs. 31-33) emphasizes single crops on types of
vegetative cover. However, each mapped unit represents a combination
of uses named for the dominant type in a finely subdivided pattern.
The analysis and calculations of land area (Table 10) accounted for
the proportion of diverse land cover types in each mapped unit.
The larger Bajamillo watershed is included in Table 10, although
the detailed land use analyses were not made at this scale. This
mesoscale watershed includes the association of food crop and pasture
rotations downslope with coffee plantations at higher elevations (Fig.
30). The Bajamillo watershed also contains several settlements and
roads that fall outside the perimeter of the two small watersheds, yet
are important to coffee production within their boundaries.
5
Coffee production in Carrizal and Rincon de Piedras is 1.8 x 10
kg annually. Of this amount, approximately 25% is produced within the
upper Bajamillo (No. 71) and about 30% is grown in the monitored
segment of the Prieto watershed (No. 64). Some of the stands were
established over 60 years ago. The coffee produced in this part of
the Sierra is of the traditional variety, although many growers now
are planting the Brazilian dwarf variety.
Coffee acreage and production are increasing in this area. The
resurgence of coffee production is due to the high international
market prices of 1978 and 1979 and to the extension and credit




40
runoff and their relationships to land use within a variety of land
use and ecosystem types throughout the Caribbean.
Results of erosion studies in similar environments. Basic data
on erosion rates are scarce in Latin America and the Caribbean, in
comparison with the humid tropics of Asia and Africa (Lai, 1977b). A
brief summary of selected erosion studies in these regions provides a
broader frame of reference for work already completed or in progress
in the Caribbean.
Many of the crops, the small farm technology and some elements of
the natural ecosystems of West Africa bear a strong resemblance to
parts of the Caribbean. Reports of experiments conducted by Lai et
al., (1979) and others at the International Institute for Tropical
Agriculture (IITA) in Ibadan, Nigeria, indicate a clear relationship
between vegetation cover and erosion rates, with results approximating
those from the Caribbean. Losses from clean-tilled fallow range from
-1 -1 -1 -1
11 tons ha yr at 1% slope to 230 tons ha yr at 15% slope. On
the average, soil erosion varies much more with slope than does runoff
(Lai, 1977a).
Experiments in Senegal (Charreau, 1972) on gentler slopes in the
savanna demonstrate a similar range of soil loss as from the medium
slopes (10%) studied by Lai (1977b). Runoff in cropped plots exceeds
that in natural bush by a factor of 20 to 35 depending upon the crop,
while soil loss increased 30 to 50 times for the same crops as
compared to natural vegetation (Table A-8). Similar results are
reported for other sites in Ivory Coast and Upper Volta.
Lai (1977b) tested the effectiveness of various types of mulch as
well as several variations of minimum tillage. While mulch had less


84
relationships between discharge, an independent variable, and sediment
concentration and transport (dependent variables) were tested by
simple linear regression of raw and log-transformed data (SAS, 1979).
The relationships between river stage and sediment transport were
used to estimate sedimentation rates of the dams to be constructed
just downstream of the sampling points in both rivers. The daily
discharge for a full hydrologic year during the period of record was
used to generate daily sediment discharge values based on the
relationship established in the previous analyses. The total was
multiplied by a correction factor (ratio of average annual discharge
to the discharge for the 1980-1981 hydrologic year) to predict the
sediment discharge for an average year, as opposed to the study
period.
The relationship between the amount and distribution of rainfall
on the watershed (independent variable) and the discharge and sediment
concentration in the rivers (dependent variables) was tested by simple
and multiple linear regression of raw and log-transformed data (SAS,
1979). The same procedure was repeated by subwatershed. Critical
areas for further study were singled out by the relative strength of
association between rainfall in each subarea and the subsequent river
flood stages and sediment concentration. This information, combined
with field reconnaissance, contributed to assignment of research
priority by subwatersheds.
The relationships between total daily rainfall on the whole
watershed and discharge and sediment concentration also were tested by
simple linear regression. The total volume of rainfall on the


91
Plan Sierra personnel involved in marketing and credit supervision,
and data from a concurrent study by Georges (1982) also were included
in the production profile.
Evaluating the watershed model. The systems models of the
watersheds were evaluated, incorporating the results of the watershed
monitoring and estimates of productivity for the land use units
previously mapped. The meteorologic and geologic inputs and outputs
were calculated directly from empirical data obtained during the study
period. The production estimates for each land use type were based on
a combination of locally recorded yields and estimates of biomass,
production and productivity reported in the literature. Estimates for
bush, dry forest and pine forest were drawn from prior field studies
conducted in nearby areas (Jennings, 1979b; Montero, 1979). Estimates
for coffee crops and pasture were drawn from the literature and
modified to reflect site conditions. Population data were taken from
statistical references and confirmed by a housing count on recent
aerial photographs.
Measurements in small plots
Experimental plots and collectors constructed at 16 sites
captured and retained runoff water and sediments for frequent
measurement and sampling throughout the 15-month period. The
structures were installed within parcels of land maintained under
normal use and management conditions. At the six sites, where new
crops and soil conservation practices were tested, these changes were
already introduced by Plan Sierra. The experiments in these cases


Table F-4. Station 70, Greater Bajamillo Stream
Date
MAXLEVL
SED1
SED2
SED3
SED4
SED5
MAXDSCHG
FLDDSCHG
SEDDIS1
SEDDIS2
SEDDIS3
SEDDIS4
SEDDSI5
800813
2
6.60
4.39
.
6.476
22.7
28.4
800918
4
-
-
-
-
-
15.024
-
_
-
_
_

811001
5
-
-
-
-
-
17.750
57.51
-
-
-
-
-
801010
5
16.42
15.23
200.82
12.56
76.25
17.750
57.51
56.4
98.6
2080.7
188.7
1353.4
801022
5
36.82
7.47
8.04
70.29
17.99
17.750
57.51
126.5
48.3
83.3
1056.0
319.3
801106
3
7.46
1.80
23. 37
-
-
10.361
-
25.6
11.6
252.5
_
_
801222
1
3.58
-
-
-
-
3.436
-
12.3
-
_
_
_
810204
2
4.79
7.79
-
-
-
6.476
-
16.5
50.4
-
_
_
810414
2
-
0. 39
-
-
-
6.476
_
_
2.5
_
_
_
810504
4
0.19
0.42
1.56
3. 18
-
15.024
-
0.7
2.7
16.1
47.7
-
810519
5
10.00
2.94
0.56
34.37
18.80
17.750
57.51
34.4
19.0
5.8
516.3
333.7
810529
5
1.81
58.87
80.56
79.84
5.64
17.750
57.51
6.2
381.2
834.6
1199.5
100.1
810609
2
5.72
4.04


6.476
19.7
26.1
~

3 -1
Discharge rates (m sec )
follows: 1 3.436, 2-6
for each level,
476, 3 10.361
corresponding to
4 15.024, 5 -
concentrations SED1 through SED5,
17.750.
are as
347


Table 11.Continued.
Plot Number
88
Site
Los Montones
Slope (%)
28.0
Soil Classification
Eutropept
Soil Texture
Sandy loam
Soil Feritlity
% N
% OM
0.18
3.50
Infiltration
53.5
Vegetative Cover
Pine forest
Conservation Practice
tAgave spp
89
91
92
Los Montones
Pananao
Pananao
25.5
34.0
32.0
Troporthent
Troporthent
Troporthent
Sandy loam
Loam
Loam
0.15
0.15
0.20
2.94
3.02
3.98
8.63
11.6
7.6
Pasture
Sisalt, Sweet
Sisal
potato
181


33
Watershed and sedimentation studies
Erosion, sedimentation and land use research in tropical
hillslope lands falls into two major categories. The watershed
studies focus primarily on the interaction of climate, topography, and
land use throughout the drainage area in determining river regimes and
erosion and sedimentation rates. By contrast, reservoir sedimentation
studies focus on the identification of sediment source areas as well
as on the immediate protection of reservoir facilities, which often
presupposes an emphasis on the development of physical infrastructure
at various points throughout the watershed.
Representative basins and experimental catchments. Studies
initiated under the auspices of the International Geophysical Year
(IGY) and the International Hydrological Decade account for a large
proportion of the work conducted in the tropics. The representative
basin studies emphasize comparative description of diverse river
schemes and watershed ecosystems on a global basis, while experimental
catchment studies focus more attention on the effects of alternative
land cover and land treatment in a given area.
Among the more important findings to date are the contrasting
characteristics of tropical climate and hydrology when compared to the
more temperate regions. Water balance data, including ratios of
runoff and evapotranspiration to total precipitation, are available in
reports from empirical studies (Golley, 1972; Holdridge, 1967, 1982;
Odum 1970b; Pereira, 1973; Thornthwaite and Mather, 1959). The
seasonal distribution as well as the amount of precipitation varies
substantially from the pattern of temperate areas. Bimodal rainfall


Table A-ll. Variation of runoff and soil loss with land use in upland Mindanao
Vegetation
Runoff
(As percent of
Precipitation)
Soil Loss
During Cropping Period
(kg ha-1 day-1)
Soil
Over
Loss Increments
Primary Forest
Losses
Primary Forest
0.25
0.25
X
1.0
Softwood Trees
0.26
0.36
X
1.5
Imperata Grassland
3.00
0.50
X
2.0
New Abaca Plantation
0.35
0.59
X
2.5
10 Year Old Abaca Plantation 0.64
0.74
X
3.0
Logged Forest
1.73
-

New Corn Field
1.52
3.79
X
115.0
New Rice Field
1.08
1.81
X
70.0
2 Year Old Corn Field
1.78
15.06
X
60.0
12 Year Old Rice Field
11.64
145.14
X
600.0
Source: Kellman, 1969
295




339
Table E-2--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
810224
0.79
0.75
-
-
-
-
-
10.9
9.2
10.0
-
810225
0.68
0.64
-
-
-
-
-
6.6
5.7
6.2
-
810226
0.60
0.60
-
-
-
-
-
4.1
4.1
4.1
-
810227
0.60
0.60
-
-
-
-
-
4.1
4.1
4.1
-
810228
0.60
0.61
-
-
-
-
-
4.1
4.5
4.3
-
810301
0.64
0.64
-
-
-
-
-
5.7
5.7
5.7
-
810302
0.63
0.62
-
-
-
-
-
5.3
4.9
5.1
-
810303
0.62
0.62
-
-
-
-
-
4.9
4.9
4.9
-
810304
0.61
0.60
-
-
-
-
-
4.5
4.1
4.3
-
810305
0.57
0.57
-
-
-
-
-
3.2
3.2
3.2
810306
0.57
0.57
-
-
-
-
-
3.2
3.2
3.2
810307
0.57
0.57
-
-
-
-
-
3.2
3.2
3.2
-
810308
0.56
0.56
-
-
-
-
-
2.9
2.9
2.9
-
810309
0.55
0.55
-
-
-
-
-
2.7
2.7
2.7
-
810310
0.56
0.56
-
-
-
-
-
2.9
2.9
2.9

810311
0.55
0.55
-
-
-
-
-
2.7
2.7
2.7
-
810312
0.55
0.55
-
-
-
-
-
2.7
2.7
2.7
-
810313
0.54
0.54
-
-
-
-
-
2.7
2.7
2.7
-
810314
0.54
0.54
-
150
0.155
-
-
2.4
2.4
2.4
-
810315
0.54
0.58
-
-
-
-
-
2.4
3.5
3.0
810316
0.65
0.62
-
-
-
-
-
5.7
4.9
5.3
810317
0.66
0.64
-
184
0.202
0.202
0.20
6.0
5.7
5.9
-
810318
0.58
0.57
-
187
0.238
0.260
0.25
3.5
3.2
3.3
-
819319
0.54
0.54
-
-
-
-
-
2.4
2.4
2.4
-
810320
0.54
0.54
-
-
-
-
-
2.4
2.4
2.4
-
810321
0.54
0.54
-
187
0.179
0.026
0.10
2.4
2.4
2.4
-
810322
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
810323
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
810324
0.53
0.56
-
-
-
-
-
2.2
2.9
2.6
-
810325
0.58
0.57
-
-
-
-
-
3.5
3.2
3.3
-
810326
0.56
0.58
-
185
0.161
0.190
0.18
2.9
3.5
3.2
-
810327
0.65
1.16
-
145
0.719
0.232
0.48
5.7
38.7
22.2
-
810328
1.02
0.94
2.1
-
-
-
-
25.3
19.3
22.3
169.5
810329
1.49
1.25
-
143
0.471
0.444
0.46
72.2
46.6
59.4
-
810330
0.96
0.85
-
-
-
-
-
20.7
13.9
17.3
-
810331
0.80
0.75
-
-
-
-
-
11.3
9.2
10.2
-
810401
0.72
0.70
-
-
-
-
-
8.0
7.3
7.6
-
810402
0.69
0.69
-
-
-
-
-
6.9
6.9
6.9
-
810403
0.68
0.74
-
-
-
-
-
6.6
8.8
7.7
-
810404
0.71
0.70
-
-
-
-
-
7.6
7.3
7.5
-
810405
0.68
0.67
-
-
-
-
-
6.6
6.3
6.5
-










60
The primary hypotheses were tested and judged jointly by researchers
and farmers, according to objective, quantitative criteria. The
secondary postulate was tested and judged by the farmer according to
subjective criteria, based on overall practical performance. Researchers
provided a posteriori explanation and interpretation of the farmers'
experience. The value judgement as to the fitness of the existing system
was considered the joint prerogative of the local residents and the
policy sector (clients) while the researcher determined the system's
sustainability by ecological analysis of current trends.
Materials and Methods
The study was conducted at four scales of analysis within a nested
hierarchy of spatial units, including: the Plan Sierra impact area (2500
2 2 2
km ); watersheds of 500 to 100 km ; small watersheds of 1 to 20 km ; and
individual landholdings ranging from 0.5 to 500 ha.
Chronologically the research activities were grouped into three
phases of increasingly finer levels of resolution and greater detail of
analysis. The regional reconnaissance and refinement of problem
definition was followed by detailed characterization of the study sites
at the watershed and plot level. The third phase consisted of monitoring
runoff, erosion, sedimentation, and production under varying land use and
soil conservation practices (Fig. 4).
The conceptual model of the Caribbean region (Fig. 1) formed the
basis for the overall research design, while at each successive level of
resolution a conceptual model was proposed, evaluated through field data
collection, and refined or modified based on empirical evidence and
testing of specific hypotheses inherent in the model.








404
Holdridge, L. R. 1967. Life zone ecology. Tropical Science
Center, San Jose, Costa Rica.
. 1982. Ecologia basada en zonas de vida. IICA, San
Jos, Costa Rica.
Holeman, J. N. 1968. The sediment yield of major rivers of the
world. Water Resources Res. 4:737-747.
Holland, T. H., and A. W. R. Joachim. 1933. A soil erosion experi
ment. Trop. Agricst. Ceylon. 80:199-207.
Holl, M. 1979. Un resultado de la investigacin mediante el
enfoque de sistemas: Preparacin de alternativas tecnolgicas
al sistema del agricultor. Paper presented at Regional Seminar
on Methods for Development of Alternative Technologies in
Cropping Systems, San Salvador, El Salvador, July.
Holtan, H. N., and N. C. Lopez. 1971. USDAHL-70 model of watershed
hydrology. USDA-ARS Tech. Bull. 1435, Washington, D.C.
Hopkinson, C. S., and J. W. Day. 1980. Modeling the relationship
between development and storm water and nutrient runoff.
Environmental Management. 4:315-324.
Horten, R. E. 1935. Surface runoff phenomena, Part I--Analysis of
the hydrograph. Horton Hydrological Lab. Pub. 101,
Voorheesville, New York.
. 1938. The interpretation and application of runoff
plot experiments with reference to soil erosion problem. Soil
Sci. Soc. Am. Proc. 3:340-349.
Hudson, N. W. 1957. The design of field experiments on soil erosion.
J. Agrie. Eng. Res. 2:56-57.
Hutchinson, J., M. L. Manning, and H. G. Farbrother. 1958. On the
characterization of tropical rainstorms in relation to runoff
and percolation. Quart. J. Royal Met. Soc. 84:250-258.
IBRD. 1972. Dominican Republic--Yaque del Norte Irrigation
Project. IBRD, Washington, D.C.
INDRHI. 1981. Hydrometric data and stage-discharge curves for
Amina and Mao Rivers. Computed by M. Morel, INDRHI, Santo
Domingo, Dominican Republic.
James, P. E. 1972. All possible worlds. Bobbs-Merri11 Co., Inc.,
New York.
Jennings, P. 1979a. Stencil on a proposal for the reforestation
of some essential watersheds in the Sierra Zone, Dominican
Republic. ISA, Santiago, Dominican Republic.


"






APPENDIX G
SOIL PROFILE DESCRIPTIONS FOR EROSION PLOT SITES








Table 21continued.
Plot Number
86
87
88
89
96
Site
Los Montones
Los Montones
Los Montones
Los Montones
Pananao (El Rubio)
Present Land Use
Idle pasture
Woodland
Forest
Idle pasture
Forest
Past Land Use
Grazing
Forest, selec
Same
Grazing, food
Forest, selectively
tively cut
crops
cut
Time Since Clearing
20 years


20 years

Time Occupied by Owner
20 years
1 year*
5 years**
5 years
recently acquired
by the state
* Same owner as newly cleared food crop plots, 84 and 85.
**Being held for speculation.




Table F-3. Station 67, Hondo Stream
Date
MAXLEVL
SEDl
SED2
SED3
SED4
SED5
MAXDSCHG
FLDDSCHG
SEDDIS1
SEDDIS2
SEDDIS3
SEDDIS4
SEDDIS5
800619
1
0.83
_
_
0.610
0.5
800813
3
264.97
0.12
0.06
-
-
4.054
-
161.6
0.2
0.2
_
801003
4
7.01
-
35.28
-
-
7.856
-
4.2
_
143.0
_
_
801028
4
26.80
5.62
21.50
68. 30
-
7.856
-
16.3
11.2
87.1
536.5
_
810311
5
13.76
4.13
19.68
0.98
0. 18
8.500
47.8
8. 3
8.2
79.7
7.7
1.5
810407
1
0.13
-
-
-
-
0.610
-
0.0
_
810414
1
0.29
-
-
-
-
0.610
_
0. 1
_
_
_
810504
3
24.31
2.04
2.80
-
-
4.054
_
14.8
4.0
11.3
_
810526
1
27.42
-
-
-
-
0.610
_
16.7
_
_
_
810609
4
32.32
33.76
~
7.856
~

64.4
136.8
-
-
Discharge rates (m sec ) for each level, corresponding to concentrations SEDl through SED5, are as
follows: 1 0.610, 2 2.000, 3 4.054, 4 7.856, 5 8.500.


401
Frederiksen, R. L. 1970. Comparative chemical water quality-
natural and disturbed streams following logging and slash
burning. Proc. Symp. Forest Land Uses and Stream Environment,
Oregon State Univ., Corvallis.
Frere, M. H. 1978. Models for predicting water pollution from agri
cultural watersheds. Conf. on Modeling and Simulation of Land,
Air and Water Resource Systems. Int. Fed. Inf. Processes,
Ghent, Belgium.
Frucht, R. 1967. A Caribbean social type: Neither "peasant" nor
"proletarian." Social and Econ. Studies. 16:295-300.
Geertz, C. 1963. Agricultural involution: The process of ecolo
gical change in Indonesia. Univ. of Calif. Press, Berkeley.
. 1972. The wet and the dry: Traditional irrigation in
Bali and Morocco. Human Ecology. 1:23-39.
Georges, E. 1981. Personal communication. San Jose de las Matas,
Dominican Republic.
. 1982. Personal communication. Gainesville, Fla.
Gerlach, T. 1967. Hillslope troughs for measuring sediment movement.
Revue de Geomorph. Dynamique. 17:173.
Gladwin, C. H. 1981. Estrategias de decision de los pequeos
productores in las zonas de ladera y sus implicaciones para el-
diseno de projectos. In A. R. Novoa and J. L. Posner, (eds.)
Agricultural de ladera en America tropical. CATIE/Rocekfeller
Found., Turrialba, Costa Rica.
Golley, F. B. 1972. Energy flux in ecosystems. In J. A. Wiens (ed.)
Ecosystem structure and function. Oregon State Univ. Press,
Corvallis.
, J. T. McGinnis, R. G. Clements, G. I. Child, and M. J.
Deuver. 1975. Mineral cycles in a tropical moist forest
ecosystem. Univ. Georgia Press, Athens.
Gomez, M. 1980. Historia y actividades del Departamento de Tierras
y Aguas. Proc. 1st Seminario de Manejo de Cuencas Hydrograficas,
Santo Domingo, Dominican Republic, May.
Gonzales, M., H. Martinez, and N. Gewald. 1979. Combining grazing
and forestry in the upper Central Valley of Costa Rica: Finca
Esmeraldas. Proc. Workshop on Agroforestry Systems in Latin
America, Turrialba, Costa Rica, March.


246
The coefficients for runoff and soil loss are given in Tables 23
and 36. The results in this case are a measure of annual totals and
also use different rainfall rates, by site, as a basis for comparison
of runoff with rainfall. This accounts for the slight differences
with the analysis of variance results presented earlier. Overall,
this comparison confirms the analysis of variance results and provides
a specific quantitative measure of runoff and erosion effects by
group. The runoff as a percent of precipitation varies from 1.0 to
4.5% at Los Montones and Carrizal. The highest runoff rates are
attributed to the long established cropped fields (82 and 83) and
coffee plots (81 and 82), and the lowest rates are found in forest.
At Pananao the rates are generally higher, except for forest (1.2),
indicating that differences in rainfall distribution do not explain
the higher runoff rates. The cropped plots (91, 92, and 95) have the
highest runoff rates (6.7 to 11.7%) with the,highest rate occurring in
the plot with the longest history of croppings. The runoff
coefficients can be applied, as they are, to separate models of the
small watersheds, or the coefficients for both sites can be combined
in the case of forest and pasture for use in the regional model.
Cropping systems must be differentiated with respect to their present
and historical resemblance to those at Pananao or Los Montones.
Comparison of the results with those of other similar studies in
erosion plots shows a trend in the relationship of land use (present
and past) to runoff rates. Runoff rates under natural savanna in
Uganda (Table A-9), natural bush in Senegal (Table A-8) and primary
forest in upland Mindanao (Table A-ll) are consistently less than


17
interacting elements whose mutual influences are observed within the
system include physical, chemical, biotic and cultural entities (Odum,
1971, 1982; Odum and Odum, 1976).
The general conceptual framework of systems analysis is well
suited to the study of mutual influence between land type, land use,
plant and animal productivity, and erosion in the upland watersheds of
the Caribbean. Incorporating the broad perspective of the man/land
tradition in geography, the research in open systems geomorphology and
ecosystem analysis also offers valuable methodological examples for
application to interdisciplinary research on the topic.
Models of Erosion and Sedimentation
The range of erosion and sedimentation models in use includes
stochastic and deterministic models, statistical as well as
"parametric" models, and combinations of all of the above. These
models have been developed and applied within scales of analysis
ranging from individual plots to large watersheds.
One school of research focuses on modelling processes and
interactions between the various parts of watersheds as complex
systems (Chorley, 1962; Likens et al., 1977), and the other major
thrust of erosion and sedimentation studies has been to develop
empirical predictive equations (Wischmeier, 1975; Wischmeier and
Smith, 1978). The latter relate land use and management to erosion
loss and sediment yield and have been developed for management
purposes, primarily for use in soil and water conservation programs.




320
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800617
0.96
0.93
1.71
-
-
-
-
45.2
42.3
43.8
123.6
800618
1.10
1.06
2.10
100
0.202
0.169
0.
18
59.8
55.4
57.6
164.0
800619
1.13
1.05
-
71
0.170
-
-
63.2
54.3
58.8
-
800620
0.98
0.93
1.75
-
-
-
-
47.2
42.3
44.7
127.6
800621
1.07
0.98
1.62
104
0.150
0.174
0.
16
56.5
47.2
51.8
114.7
800622
1.12
1.00
1.18
-
-
-
-
62.0
49.2
55.6
69.1
800623
0.94
0.94
1.35
104
0.280
0.426
0.
35
43.3
43.3
43.3
69.1
800624
0.99
0.94
-
-
-
-
-
48.1
43.3
45.7
-
800625
0.90
0.87
1.08
-
-
-
-
39.6
36.9
38.2
57.6
800626
0.95
0.88
-
-
-
-
-
44.2
37.8
41.0
-
800627
0.85
0.83
1.14
-
-
-
-
35.2
33.5
34.4
64.3
800628
0.86
0.83
-
-
-
-
36.0
33.5
34.8
-
800629
0.82
0.80
-
-
-
-
-
32.7
31.1
31.9
-
800630
0.78
0.77
1.50
-
-
-
-
29.5
28.7
29.1
103.2
800701
0.88
0.82
1.02
-
-
-
-
37.8
32.7
35.2
51.2
800702
0.88
0.79
-
-
-
-
-
37.8
30.3
34.0
-
800703
0.77
0.76
0.94
-
-
-
-
28.7
28.0
28.3
43.3
800704
0.79
0.76
-
-
-
-
-
30.3
28.0
29.1
-
800705
0.74
0.71
1.20
-
-
-
-
26.5
25.0
25.7
71.5
800706
0.80
0.76
1.00
-
-
-
-
31.1
28.0
29.5
49.2
800707
0.80
0.78
1.22
-
-
-
-
31.1
29.5
30.3
77.7
800708
0.85
0.80
-
-
-
-
-
35.2
31.1
33.1
-
800709
0.76
0.74
-
-
-
-
-
28.0
26.5
27.2
-
800710
0.75
0.73
1.04
-
-
-
-
27.2
25.7
26.5
53.3
800711
0.84
0.85
1.06
-
-
-
-
34.3
35.2
34.8
55.4
800712
0.77
0.73
-
-
-
-
-
28.7
25.7
27.2
-
800713
0.72
0.70
-
-
-
-
-
25.0
23.6
24.3
-
800714
0.69
0.68
-
-
-
-
-
22.9
22.2
22.6
-
800715
0.68
0.67
1.28
-
-
-
-
22.2
21.6
21.9
83.0
800716
0.79
0.73
1.13
-
-
-
-
30.3
25.7
28.0
63.2
800717
0.83
0.76
-
-
-
-
-
33.5
28.0
30.7
-
800718
0.76
0.73
-
-
-
-
-
28.0
25.7
26.8
-
800719
0.72
0.70
1.56
-
-
-
-
25.0
23.6
24.3
108.9
800720
1.04
0.88
1.32
-
-
-
-
53.3
37.8
45.5
86.6
800721
0.98
0.93
1.10
-
-
-
-
47.2
42.3
44.7
59.8
800722
0.88
0.83
1.34
-
-
-
-
37.8
33.5
35.7
88.4
800723
0.90
0.87
1.80
-
-
-
-
39.6
36.9
38.2
132.6
800724
0.95
0.86
-
-
-
-
-
44.2
36.0
40.1
-
800725
0.81
0.78
-
119
0.041
0.042
0.
05
31.9
29.5
30.7
-
800726
0.76
0.74
-
-
-
-
-
28.0
26.5
27.2
-
800727
0.72
0.72
-
-
-
-
-
25.0
25.0
25.0
-
800728
0.71
0.70
-
-
-
-
-
24.3
23.6
24.0
-


201
Table 16. Maximum recorded concentrations (g L ) per flood event,
for all streams.
Duncan groups
Ln (g L
Stream no.
means^
3.39
70
2.76
60
2.29
67
1.45
71
0.58
64
fMeans refer to the means of peak concentrations for all events
at each stream.










359
Location:
Los Montones
Plot #:
89
Land Use:
Pasture
Land Cover:
Sparse natural pasture
Parent material:
Intrusive igneous rock
Relief:
Pronounced
Position:
Mid-to-lower slope
Slope Class:
25-45%
Runof f:
Rapid
Permeability:
Low to moderate
Erosion Class:
Slight
Drainage Class:
Rapid (4)
Soil Moisture:
Uniform, slightly moist
Salts/Alkaline:
None
Stoniness:
Slight
Soil Profile:

Horizon
A
C
Depth
0-15 cm
15-32 cm
Boundaries
regular
irregular
Color
10 YR 4/2
10 YR 5/8
Texture
sandy
clay
Structure
blocky
blocky
Consistency moderately adhesive
-
C3
no
no
Concretions
no
no
Slickensides
no
no
Mottles
no yes (rust &
black; iron & manganese)
Roots
yes
yes








Table 6. Summary of regression analyses of river discharge and sediment concentration.
Y
X
n
2
R
Ft
Amina River
L
n
(average sediment concentration)
River level
41
.21
10.62**
L
n
(average sediment concentration)
L (daily discharge)
n
41
.19
9.24**
L
n
(sediment discharge)
River level
41
.43
29.32***
L
n
(sediment discharge)
(daily discharge)
41
.59
56.05****
L
n
(sediment concentration maximum)
River level
51
.12
6.68*
L
n
(sediment concentration maximum)
(daily discharge)
51
.16
9.06**
Mao
River
L
n
(average sediment concentration)
River level
39
. 38
22.43****
L
n
(average sediment concentration)
L (daily discharge)
n
43
.24
12.92***
L
n
(sediment discharge)
River level
43
.69
93.11****
L
n
(sediment discharge)
(daily discharge)
43
.69
93.11****
L
n
(sediment concentration maximum)
River level
55
.32
24.73****
L
n
(sediment concentration maximum)
(daily discharge)
66
.10
6.97**
^The significance level of the model
0.001, and **** for 0.0001.
is indicated by for 0.
05, **
for 0.01,
*** for
155










71
the homes of agricultural laborers and landholders. The Plan Sierra
soil conservationists living in the area also were interviewed, and
they in turn questioned residents about the settlement history, land
use and farming practice, past and present production levels, and
sources of income. Discussions with two anthropologists conducting
land use and migration research in the region also provided valuable
information and insights into the character of the communities in the
study areas (Pessar, 1981; Georges, 1982). Plan Sierra social
workers, agronomists and foresters familiar with the area of interest
also contributed to the socioeconomic profile of the small watersheds.
Description of individual landholdings
Selection of sites for measurement of runoff and erosion in
experimental plots was based on uniformity and degree of slope, as
well as type of land use, management practice and easy access for
construction and sampling purposes. Wherever possible, replicates of
land use and treatment were established within the same watershed and
also in another watershed to determine the margin of error in
measurement and to compare the relative difference in runoff and
sediment yield under varying conditions of site and land use.
The experimental design further subdivided the categories of
forest, pasture, coffee and annual crops to compare undisturbed and
secondary forest, new and established coffee stands, different types
of annual crops, and use of minimum tillage and hillside ditches for
erosion control in fields planted to annual crops. These subsets of
land use type were tested in paired plots in the same or adjacent land






LIST OF FIGURES
FIGURE PAGE
1 Systems model of land use and erosion in the Caribbean.. 5
2 Impact area: Plan Sierra. 9
3 Model of applied research process 25
4 Flow chart of research activities 26
5 Input-output diagram for interview notations and
monitoring 53
6 Organization of research activities 64
7 Diagram of Thiessen polygons superimposed on a map of
the Mao and Amina watershed subdivisions 77
8 Sampling sites for large watersheds 78
9 Uppsala-type manual sampler for instantaneous measure
ment of sediment concentrations in streams -81
10 A. Illustration of velocity-area method. Person "a"
releases float at time t and person "b" records time
it takes float to move 2 m. B. Cross-section of
stream showing placement of float to measure velocity
and area of three sections of the stream 86
11 Equipment installed in streams to measure sediment
concentrations at different levels of flooding 88
12 Illustration of erosion plot with runoff and sediment
collectors 93
13 Illustration of alternative erosion plot design with
three subsections 95
14 Diagram of sediment and runoff collector indicating the
points at which samples were taken 97
15 Plan Sierra geologic subregions 103
16 Plan Sierra region with study sites 104
IX




162
Table 9. Water and sediment yields estimated from 1980 and 1981
data.
Mao River
Amina River
Total discharge (m yr )
3 -1 -1
Water yield (m ha yr )
Sediment transport (tons yr )
Sediment yield (tons ha 3 yr 3)
818 x 106
13 x 103
197,888.0
3.1
409 x 106
12 x 103
43,363.0
1.3


Table A-2. Mass transport of sediment from forested watersheds.
Location
Loss
Source
. -1 -1
kg ha yr
Arkansas, Ouchita Mountains
15
Ursic, 1978
Florida, Pine Flatwoods
129
Reikerk et al., 1978, 1979
Georgia, Piedmont
(Clearcut and Machine Prepared)
20,000f
Hewlett and Nutter,
1969
Minnesota, Deciduous Forest
37
Singer and Rush, 1975
Mississippi, Pine Forest
225
Duffy et al., 1978;
1976; Ursic, 1978
Schreiber et al. ,
New Hampshire, Hubbard Brook
(Hardwood/Pine)
33
Bormann and Likens,
1979
North Carolina, Coweeta
38-76
Johnson and Swank,
1973; Mitchell
(Hardwood/Pine)
et al., 1976; Monk,
1976
Approximately 397 kg ha cm stormflow leaving watershed.
286






56
population so often used with questionnaires or structured interviews.
The latter, more formal approach is a more common and preferred form
of data gathering in many disciplines. However, it is also expensive
and time-consuming, and it presupposes an accurate population census
or property survey on which to base the sample. Formal surveys and
questionnaires also inherently limit the categories of information to
be treated. Little room is left for the definition in the field of
problems not already recognized prior to the survey. Opportunities to
explore the unique relationship between various aspects of a problem
in an area are also constrained by the format. The major
considerations in using the sondeo technique are quality versus
quantity of information and cost of the survey versus useful
information obtained.
The sondeo provides a practical and effective means of
reconnaissance and data gathering. It also lays the groundwork for
future extension programs in the area. During the intensive one-to-
two-week survey, the knowledge and needs of the farmer are
incorporated into the design (form and content) of subsequent on-farm
research projects. This methodology relies heavily on the judgement
of the research team, the local populace, and individual farmers to
define farming systems and their problems and to choose representative
or exceptional cases for further study, according to the goals of each
project (Hildebrand, 1981).
The experiments themselves are on-farm trials which may or may
not be replicated at experimental stations. The sondeo and the field
trials are supplemented by farm record-keeping. A family member keeps


277
Delayed transport of previously deposited sediments is also a
probable explanation for the high sediment yield in the Mao Basin.
The sediment delivery ratio is estimated at close to 50% if channel
erosion and prior deposition are excluded from consideration. Since
the initial deforestation 20 to 30 years ago, the rate of sediment
deposition in or near the channels may have exceeded the capacity of
the river to remove it at an equivalent rate. The cumulative deposits
would then move downstream principally during major flood events.
Extreme floods, such as those caused by Hurricanes David and Frederick
in 1979, have played a major role in this process. During the floods
a large amount of sediment (new and residual) was transported
downstream, but the aftermath of the storms also left new sediment
deposits and further scoured the stream banks and channel.
In the case of residual sediments, erosion prevention can have
little effect. The construction of sediment detention structures,
however, could reduce the delivery ratio of the eroded materials
already in or near the stream courses. Cleaning and maintenance of
such structures would be far less expensive (in money and energy
terms) than mechanical dredging (or abandonment) of large reservoirs.
In terms of erosion prevention the reduction of forest clearing and
tillage of plots in annual crops will have the greatest impact.




276
These sediment delivery ratios for the study area also may be an
artifact of the specific time interval during which the data were
collected. The streams and rivers of the Sierra may be gradually
transporting a sediment load derived in large part from erosion of
newly cleared and tilled forest lands during the period of intensive
logging and settlement about 30 years ago. Although the data from the
study do not allow further exploration of this explanation, it is a
possibility that should be considered when evaluating the high
sediment delivery ratios.
The need to consider the interaction of land use and river
systems in a dynamic context is demonstrated by the case of Pananao
stream. As depicted in Fig. 19, any given storm may combine newly
eroded soil with resuspended channel deposits in the total suspended
sediment load transported past the sampling point. The sediment
deposits in the stream channel, as well as the condition and
topography of the stream banks already have been affected by prior
land use changes and management practices. The amount of runoff that
reaches the stream is partially conditioned by prior use of the land,
as well as by current use and practice.
Much of the deposited material available for resuspension in
overland flow and streamflow may derive from the initial intensive
clearing and tilling of the Pananao Valley 30 years ago. These
sediments in turn would have contributed to the scouring action of the
streamflow, accelerating channel erosion. Attributing the current
condition and behavior of the stream solely to the existing land use
and cover could yield false conclusions and inappropriate management
decisions.








16
field conditions is a proper subfield of physical geography since
geomorphology is a traditional and well-developed avenue of inquiry
within the discipline (Gregory and Walling, 1973; James, 1972; Keller,
1968; Morgan, 1979; Stoddart, 1965; Strahler, 1964). Systems theory
has been applied widely in studies of watersheds and other
geomorphological units by British geographers (Chisholm, 1967;
Chorley, 1962; Kirkby, 1978; Ollier, 1968) as well as by numerous
other geomorphologists (Leopold et al., 1964; Leopold and Langbein,
1962; Toy, 1977) and hydrologists (Chow, 1964; Vemuri and Vemuri,
1970). This general approach offers the advantage of integrating form
and process, by accounting for their interactive relationship
(Chorley, 1962, 1969).
The open systems approach allows the inclusion of the
quantitative hydrology/geomorphology tradition that dates from Horton
(1935, 1938) and continues in the work of Strahler (1964) and other
physical geographers. The use of systems theory in geomorphology also
facilitates the study of the human use of the earth, a longstanding
focus of geographic research (Harvey, 1969; James, 1972; Marsh, 1964;
Thomas, 1974) that could lend itself well to quantification within a
systems framework (Stoddart, 1965).
The application of systems theory to the study of watersheds has
also been tested and developed within ecology in recent studies of
biogeochemistry (Bormann and Likens, 1979; Likens et al., 1977) and
resource management (Cooper, 1971; Hall and Day, 1977; Hopkinson and
Day, 1980; Patton, 1971; Thomas, 1974; Van Dyne, 1969). In such
studies the watershed defines the boundaries of the ecosystem and the








257
Fig. 48. Model of coffee farms: Large and small holdings.
5.8 RD kg ^ (dry)
10~4 -RD/Kcal




OISCnfiRGE M3 SEC
O
Fig. 21. Monthly discharge of the Mao River.
132




Location: Los
Land Use: Pine
Plot #: 88
A = RKLSCP
R
K % silt
% sand
% O.M.
Structure
- Permeability
L 22 m = LS
S 27.5%
C
P
Montones
forest,
= 26
= 69
= 3.5
= 2
= 1.5
undisturbed
= 700
K = .08
= 5.8
= .003,
= 1
A (c=.01) = 3.248 t ha-lyr-1
A (c=.003) = 0.9744 t ha-lyr-1
Additional Information:
% clay = 5.0
% N = 0.18
Location: Los Montones
Land Use: Pasture
PLot #: 89
A = RKLSCP
R
K % silt
% sand
% O.M.
Structure
Permeability
L 22 m = LS
S 25.5%
C
P
33
66
2.94
3
3.5
= 700
K = .26
= 5.2
= .20,
= 1
A (c=.20) = 189.28 t ha 1yr
A (c=.04) = 37.856 t ha_1yr
Additional Information:
% clay = 6.0
% N = 0.15


186
gone through a 20-year cycle of deforestation, slash and burn
agriculture, bush fallow agriculture and establishment of pasture and
small food crop plots under permanent or almost permanent cultivation
(Table 10). The functional aspects of the current land use system are
illustrated in Figs. 38 and 39.
Biophysical characteristics
The erosion features along the banks of both streams indicated a
relatively wide range of floods stages (see Appendix B). The channels
exhibited a high rate of sediment deposition, ranging from fine and
medium textured sands at Pananao to coarse sand in the Hondo channel
and floodplain. The deposits reached 10 cm in thickness in some
reaches of both streams. Interviews with residents indicated that
both streams rise and fall very rapidly, after spring and autumn
rains. The spring flood events are flash floods that last from 0.5 -
2 hours, leaving large deposits of relatively coarse sediments in the
lower reaches of the channel. Both streams rise to peak annual flood
stages up to 2 m above the normal baseflow, with ten-fold and higher
increments in discharge.
Residents at both sites pointed out that water quality and stage
fluctuation have changed considerably over the years. Most families
in Pananao send the children on burros to fetch water at other,
cleaner streams. They travel up to 5 km from their homes rather than
use the poorer quality water from nearby Pananao stream. When the
valley was first settled the Pananao provided drinking quality water.
The change in quality is probably a combined effect of polluted runoff






160
Table 8. Sedimentation in the Mao River basin estimated from May
1980 measurements.
Annual Average
Base Flow
Flood Stage (>30 cm)*
Discharge
19.4 m^/sec
3
c10 m /sec
3
peak 500 m /sec
sustained for days -
100 mVsec
Sediment
concentration
0.006 g/Z
->2.000 g/Z
~0.022 g/Z
1000->2000 g/Z
Volume of
sediment transport
200,015 t/yr
3330
151,373 t/yr
Sediment concentration
per unit area
3 t/ha/yr
0.05 t/ha/yr
2.26 t/ha/yr
*Refers to flood stages substantially above baseflow levels, as
recorded at INDRHI station.


207
watershed. The role of these coefficients in the watershed models is
illustrated in Figs. 34, 35, 38, and 39. An analysis of the erosion
plot results is a necessary prerequisite to the above analyses at the
watershed level.
Erosion Plot and Household Studies
The erosion plot measurements and household descriptions provided
the basic data necessary to evaluate the watershed models and to
analyze Plan Sierra rural extension policies. The information
obtained at this scale proved essential in two respects. First, the
farm and plot level data provided replicated quantitative descriptions
of specific land use systems, related resource management practices
and the resultant rates of runoff, erosion and sedimentation. This
established the magnitude of the soil erosion and production problems
at the farm level, and the variation of these problems with land use
type and soil conservation practices. Both types of information were
necessary to evaluate Plan Sierra soil conservation and cropping
systems programs. Secondly, the formal and informal interaction with
farm families, landowners and nearby residents over the 15-month study
period provided valuable insights into the persistence of apparently
counterproductive land use practices.
Plot Descriptions
The diverse environmental and socioeconomic characteristics of
the 16 plots and nine households studied are summarized in Tables 11
and 21. The inherent physical characteristics, with the exception of


Table A-7. Erosion rates measured in the Caribbean and Latin America (soil losses in tons ha
yearl).
Land Use
Jamaicat
Puerto
Ricot
Colombia^
Brazil^
Clean-Tilled
Fallow
375 625
253 -
303
225 -
253

Field Crops
*
100 125
15.2 -
36.4
1.0 -
20.0
9.5 -
21.5
Field Crops,
Traditional
134.4


21.4
- 21.5
Field Crops,
on Contours



4.1 -
19.8
Field Crops,
Hillside Ditch and "Hills"
38.8



Field Crops,
Hillside Ditch and Contours
26.6
--


Field Crops,
Bench Terraces
17.5



Field Crops,
with Mulch

1.4


Field Crops, with Mix of Conservation
Practices
9.8 14.8



Pasture

4.4
3.0 -
7.1

Fertilized Pasture
5.0 12.5
1.5

1.2 -
2.7
Coffee, New


1.8 -
24

Coffee, New
with Terraces


0.2
--
Coffee, Old
without Terraces


0.6

Dense Forest
0.5 1.3
i
--
--
--
fsheng, 1973; Sheng and Michaelson, 1973.
jsmith and Abruna, 1955; Vicente-Chandler, 1976.
§Suarez de Castro and Rodriguez, 1955, 1962,
^Bertoni, 1966; Marques et al., 1961.
291









204
Table 18. Analysis of variance of stream discharge and sediment
transport comparing streams draining coffee stands and
streams draining food crops and pastures.
Source
df
Mean Square
F value
Average sediment discharge rate ha during flood peak
Model 1 3.53 0.55
Error 83 6.44
Peak sediment discharge rate ha per flood event
Model 1 6.13 0.79
Error 83 7.72
Peak discharge ha per flood event
Model 1 36.72 30.01****
Error 86 1.22
Average sediment concentration (g L per flood event
Model 1 57.29 17.10****
Error 83 3.35
Maximum sediment concentration (g L per flood event
Model
Error
1
83
66.68
4.04
16.52****


Table 35. Runoff and sediment losses for plots in Los Montones and Pananao sites, with crops,
results of the Duncan Multiple Range Test.
Duncan
groups
Ln (Runoff)
Means
Plot
Number
Site
Duncan
groups
Ln (Sediment)
Means
Plot
Number
Site
3.02
91, 92, 95
Pananao
1
4.75
91
, 92, 95
Pananao
1
2.36
82, 83
Los Montones
3.23
84
, 85
Los Montones
1
1.89
84, 85
Los Montones
2.50
82
, 83
Los Montones
243


340
Table E-2--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
810406
0.63
0.62
-
-
-
-
-
5.3
4.9
5.1
-
810407
0.60
0.60
-
185
0.066
0.176
0.12
4.1
4.1
4.1
-
810408
0.79
0.78
-
-
-
-
-
10.9
10.4
10.6
-
810409
0.76
0.75
-
177
0.030
0.318
0.17
9.6
9.2
9.4
-
810410
0.74
0.74
-
-
-
-
-
8.8
8.8
8.8
-
810411
0.79
0.75
-
-
-
-
-
10.9
9.2
10.0
-
810412
0.73
0.68
-
-
-
-
-
8.4
6.6
7.5
-
810413
0.65
0.64
-
-
-
-
-
5.7
5.7
5.7
-
810414
0.63
0.68
-
-
-
-
-
5.3
6.6
6.0
-
810415
0.67
0.66
-
-
-
-
-
6.3
6.0
6.1
-
810416
0.65
0.65
-
-
-
-
-
5.7
5.7
5.7
-
810417
0.64
0.63
-
-
-
-
-
5.7
5.3
5.5
-
810418
0.65
0.65
-
-
-
-
-
5.7
5.7
5.7
-
810419
0.64
0.64
-
-
-
-
-
5.7
5.7
5.7
-
810420
0.64
0.63
-
-
-
-
-
5.7
5.3
5.5
-
810421
0.64
0.64
-
-
-
-
-
5.7
5.7
5.7
-
810422
0.62
0.61
-
-
-
-
-
4.9
4.5
4.7
-
810423
0.66
0.65
-
-
-
-
-
6.0
5.7
5.8
-
810424
0.63
0.63
-
-
-
-
5.3
5.3
5.3
-
810425
0.62
0.64
-
-
-
-
-
4.9
5.7
5.3
-
810426
0.60
0.60
-
-
-
-
-
4.1
4.1
4.1
-
810427
0.60
0.58
-
-
-
-
-
4.1
3.5
3.8
-
810428
0.17
1.25
5.50
0
1.432
1.549
1.49
0.0
46.6
23.3
1860.2
810429
1.50
1.30
-
-
-
-
-
73.4
51.4
62.4
-
810430
1.40
1.20
-
-
-
-
-
61.8
42.1
52.0
-
810501
1.20
1.30
-
-
-
-
-
42.1
52.4
46.7
-
810502
1.15
1.10
-
150
35.056
1.64718.35
37.7
32.5
35.1
-
810503
1.12
1.25
-
-
-
-
-
34.5
46.6
40.6
-
810504
1.15
1.18
-
126
3.982
-
-
37.7
40.4
39.0
-
810505
1.15
1.19
-
-
-
-
-
36.6
41.2
38.9
-
810506
1.25
1.30
-
-
-
-
-
46.6
51.4
49.0
-
810507
1.20
1.15
-
-
-
-
-
42.1
37.7
39.9
-
810508
1.30
1.25
-
-
-
-
-
51.4
46.6
49.0
-
810509
0.95
0.80
-
-
-
-
-
20.0
11.3
15.7
-
810510
0.86
0.90
-
-
-
-
-
14.4
16.7
15.6
-
810511
1.30
1.25
-
0
1.690
0.279
1.48
51.4
46.6
49.0
-
810512
1.40
1.30
-
-
-
-
-
61.8
51.4
56.6
-
810513
1.25
1.20
-
130
0.068
-
-
46.6
42.1
44.4
-
810514
1.30
1.25
-
-
-
-
-
51.4
46.6
49.0
-
810515
1.30
1.23
-
125
0.010
0.708
0.36
51.4
44.8
48.1
-
810516
1.00
0.90
-
-
-
-
-
23.7
16.7
10.2
-
810517
0.98
0.95
-
-
-
-
-
22.2
20.0
21.1
-




140
Average monthly discharge rates in the Amina and Mao Rivers in
1980 and 1981 reflect the irregularity in the rainfall distribution.
The Mao River deviated from its characteristic October peak in both
years due to the magnitude of the spring floods (Fig. 21). Baseflow
during the January to April dry season was very close to the 1967-79
means, but the 1980 May discharge was more than double the mean
monthly value. The discharge rate fell off sharply after June, but
streamflow remained at levels a little above the norm through October
of the same year.
The same extreme rainfall affected the Amina. The May discharge
rate exceeded the mean value by more than a factor of three (Fig. 22).
As in the case of the Mao, the discharge rate decreased rapidly but
did not reach baseflow levels before the onset of the August rainy
season.
The May rainfall in 1980 and 1981 falls between one and two
standard deviations from the mean for almost every station. The 1981
Moncion value differs by more than two standard deviations from the
mean for 1967-79. The mean for 1931 to 1979 is slightly higher but
the 1980 and 1981 values are still greater than one standard deviation
from the mean.
Results from May for both years must be regarded as exaggerated
relative to average conditions. Rainfall and discharge for the rest
of the study period were well within the average range. Predicted
annual totals and averages based on measurements from this period
should be adjusted downward accordingly.
Rainfall distribution over the region for the study period and
for the 13-year record is illustrated in isohyet maps compiled by


215
two cropped plots were cleared only one year earlier from a section of
the same forest stand. The same contrast in pasture versus forest was
repeated for two paired plots (88 and 89) located nearby. The
reduction in infiltration from forest to pasture was approximately 80%
in both cases. The pastures in both sites had been cleared
approximately 20 years ago and planted to food crops, then to pasture.
They have been grazed intermittently ever since.
A single measurement at the adjacent pigeon pea plots (82 and 83)
yielded a rate midway between the values for pasture and the recently
cleared food crop plots. This field was cleared 14 years ago and has
been cultivated almost continuously with occasional grass fallow. The
coffee plots (80 and J31) had the lowest infiltration rate of all the
soils tested. Prior soil moisture conditions depressed the results
somewhat."' However, the lower rates also reflect the relatively
higher clay content and overall finer texture of the soils at this
site. These soils also are subjected to some compaction by foot
traffic, particularly at harvest time.
At Pananao, the infiltration results are very different both in
magnitude and in relationship to land use. The soils generally are
shallower than at the Los Montones sites, and the texture tends to be
slightly finer (Table 11). The land use history also is distinct
(Table 21).
Plots 91, 92, and 95 were cleared about 30 years ago. The first
has been under continuous cultivation ever since in peanuts and
These were the only plots where the soil was humid prior to testing.






especially Angel Liriano S., Victor Montero, and Geuris Martinez.
Invaluable data and assistance were also provided by the National
Hydrology Institute of the Dominican Republic (INDRHI), the National
Cartographic Institute, the Dominican Electric Company (CDE), and the
Departments of Meteorology and Land and Water, of the State
Secretariat of Agriculture (SEA).
Final data processing, drafting, and typing tasks were completed
with the able assistance of Nelly Mogallon, Kim Feigenbaum, Beth
Higgs, and Pat French.
The warmest appreciation is reserved for those closest to home,
especially Mickie, Nelly, Gustavo, Marie, Mom and Dad. More than any
other, I owe the successful completion of this work to my husband,
Luis, who helped me tap my own energies and gave selflessly of his
time and effort as computer consultant, data processing technician,
editor, critic, and nurturer.
iii


353
Location:
Plot #:
Land Use:
Land Cover:
Carrizal
81
Coffee Stand
Recently planted coffee, bananas, guava,
weeds
Parent Material
Relief:
Position:
Slope Class:
Runof f:
Permeability:
Erosion Class:
Drainage Class:
Soil Moisture:
Salts/Alkaline:
Stoniness:
: Complex crystallines,
Pronounced
Mid-slope
25-45%
Rapid
Slow
Moderate
Rapid (4)
Uniform, moist
None
High
metavolcanic
Soil Profile:
Horizon
A
B
C
Depth
0-10 cm
10-18 cm
18-60+ cm
Boundaries
regular
regular
regular
Color
Texture
(dark reddish
brown)
loam
clay loam
clay loam
Structure
granular
weak blocky
weak blocky
Consistency
-
-
-
C3
no
no
no
Concretions
no
no
no
SIickensides
no
no
no
Mottles
no
yes (red)
no
Roots
yes
yes
yes


Fig. 8
Sampling sites () for large watersheds
'O
oo








174
protected by the combination of litter, live ground cover, coffee
plants and leguminous shade trees. Many of the stands in Carrizal are
old and well established, which explains their forest-like structure.
In the newer stands, however, the red-brown sandy clay soils are
exposed. Visible signs of erosion, such as rills and gullies, are
apparent. This is more pronounced in the numerous plots of field
crops along the road, in the cleared dirt surfaces around the houses,
and on the roads themselves.
Soils generally are very stony .and difficult to break and till.
Reports indicate that fertility is not a problem in coffee stands.
Coffee and other tree crops appear healthy and residents describe
yields as adequate to good by regional standards. Field crops vary
more in terms of yield and condition, both from one plot to another
and from one year to the next. Given the rainfall, solar insolation,
temperature and soil characteristics of the site, plantains and
bananas fare better than annuals and constitute the staple foods in
this area.
Observation of the downstream portions of both small streams
during flood events revealed much higher sediment loads in the runoff
draining the settlement and its access roads than in runoff from the
coffee stands. Gullies along the roadsides drain into the Prieto
stream and the Baiamillo River at crossings along the approach road to
Carrizal (Fig. 30). In both cases, extensive deposits of sediments
matching the roadbed material have built up at the mouths of the road
drains and gullies. These deposits later are resuspended during heavy
floods and contribute to the scouring capacity and the sediment load
of the Bajamillo.






Location:
Pananao
Plot #:
95
Land Use:
Manioc, monocropping
Land Cover:
Manioc, recently
planted, poorly
Parent Material
: Conglomerate
Relief:
Pronounced
Position:
Upper slope
Slope Class:
25-45%
Runof f:
Rapid
Permeability:
Rapid
Erosion Class:
Severe
Drainage Class:
Very rapid (5) (excessive)
Soil Moisture:
Uniform, slightly
moist
Salts/Alkaline:
None
Stoniness:
Moderate
Soil Profile:
Horizon
A
C
Depth
0-12 cm
12-30 cm
Boundaries
irregular
irregular
Color
-
-
Texture
sandy
sandy-clay
Structure
granular
granular
Consistency
-
-
C3
no
no
Concretions
no
no
Slickensides
no
no
Mottles
no
no
Roots
yes*
yes
*fine roots, herbaceous weeds and
manioc crop
developed


PRECIPITATION [N MM
160 240 320 400 460 550 640 720 600
Fig. 24. Monthly rainfall at the Moncion climatological station, #2.
137


355
Location:
Plot #s:
Land Use:
Land Cover:
Parent Material:
Relief:
Position:
Slope Class:
Runof f:
Permeability:
Erosion Class:
Drainage Class:
Soil Moisture:
Salts/Alkaline:
Stoniness:
Los Montones
84, 85
Mixed food crop, recently cleared
beans, maize, manioc
Metasedimentary rocks near intrusive igneous formation
of granite and rhyolite
Pronounced
Mid-slope
25-45%
Moderate
Moderate
Slight-Moderate
Rapid (4)
Uniform, slightly moist
None
Slight
Soil Profile:
Horizon
A
Depth
0-17 cm
Boundaries
irregular
Color
10 YR 4/3
Texture
sandy
Structure
granular
Consistency
-
C3
no
Concretions
no
Slickensides
no
Mottles
no
Roots
yes*
C
17-40 cm
irregular
sandy to clay
granular to blocky
no
no
no
no
yes
*root residuals from prior forest cover, plus roots from crops.


Romfall(mm) SEDIMENT (kg heclor')
226
pananao site
0 SEDIMENT
Q STORM RUNOFF
Fig. 47.Continued
STORM RUNOFF (m heclor )







R= mean annual roinfoll
Fig. 52. Evaluated small watershed submodel of land use, erosion, and sedimentation:
Prieto stream.
266










Legend for Tables E-l and E-2.
LEVLA
= hydrometric station reading, river stage, 7 A.M.
LEVLP
= hydrometric station reading, river stage, 5 P.M.
LEVLM
= hydrometric station reading, stage at flood peak.
LEVLF
= reading of river stage (distance below bridge) at sample time
SEDCA
= sediment concentration (g L "S of first sample.
SEDCB
= sediment concentration (g L of second sample.
SEDC
= average sediment concentration from both samples.
DLEVA
3 -1
= discharge rate (m sec ), derived from 7 A.M. stage.
DLEVP
3 -1
= discharge rate (m sec ), derived from 5 P.M. stage.
DLEVD
= average daily discharge rate (without flood peaks).
DLEVM
3 -1
= discharge rate (m sec ) at flood peak.
315




407
LeBaron, A., L. K. Bond, P. Aitken, and L. Michaelsen. 1979. An
explanation of the Bolivian highlands grazing erosion syndrome.
J. Range Manage. 32:201-208.
Leopold, L. B., and W. B. Langbein. 1962. The concept of entropy in
landscape evolution. U.S. Geol. Survey Prof. Paper 500A.
Washington, D. C.
, M. G. Wolman, and J. P. Miller. 1964. Fluvial processes in
geomorphology. W. H. Freeman, San Francisco.
Letouzay, R. 1955. Les arbres d'ombrage des plantations agricoles
camerounaises. Bois et Foets de Tropiques 42:15-25.
Likens, G. E., F. H. Bormann, R. S. Pierce, J. S. Eaton, N. M.
Johnson. 1977. Biogeochemistry of a forested ecosystem.
Springer-Verlag, New York.
Listin Diario. 1981. C.D.E. Hara compra equipas limpiar embalse
valdesia. Listin Diario, March, Santo Domingo.
Lockeretz, W. (ed.). 1975. Energy and agriculture. Academic Press,
New York.
Lopez, C. 1980. La conservacin de suelos y el desarrollo de activi
dades productivas. In 1st Seminario de Cuencas Hidrogrficos.
Santo Dominigo, Dominican Republic, May.
Lugo, A. E., and S. C. Snedaker. 1971. Readings on ecological
systems: Their function and relation to Man. MSS Educ. Pub.
Co., Inc., New York.
Lugo-Lopez, M. A. 1969. Prediction of the erosiveness of Puerto
Rican soils on a basis of the percentage of silt and clay when
aggregated. J. Agrie. Univ. Puerto Rico. 53:187-190.
Lundgren, B. 1982. The use of agroforestry to improve the producti
vity of converted tropical land. Rpt. to Office of Tech.
Assessment of the U.S. Congress, Washington, D.C.
Makhijani, A., and A. Poole. 1975. Energy and agriculture in the
third world. Ballinger Pub. Co., Cambridge, Mass.
Mannering, J. V., and L. D. Meyer. 1963. The effects of various
rates of surface mulch on infiltration and erosion. Soil Sci.
Soc. Am. Proc. 27:84-86.


Table 36. Soil loss coefficients by land use and conservation practice.!
Los Montones + Carrizal Sites
Forest§ 1 (1)J
Pasture 6 (12)
Coffee (old)
Coffee (new)
Pigeon pea, Sweet
potato^
Pigeon pea, Sweet
potatoH#
Yuca, beans, com
Yuca, beans, com
10
38
182
153
1462
252
(first year|f
second year!!
for both plots)
Forest§§
Pdndnso
1
bl l6S
Pasture
17
Sisal, Sweet
383
(first year
potato
Sisal, Sweet
252
second year
potato
for both plots)
Yuca
213
fCoefficients were calculated relative to soil losses (tons ha-l yr~l)under natural forest cover.
The reference plots for no's 80 through 89 are plots 87 and 8 (mean). Plot 96 is the reference
for no's 91 through 95.
^Numbers in parentheses indicate coefficients that apply to both sites for the regional model.
§Reference for Los Montones and Carrizal sites (80-89).
UHillside ditches are used.
#Minimum tillage is practiced,
ttfirst year, after clearing, burning and tilling.
Second year.
§§Reference for Pananao sites.


400
FAO (Food and Agricultural Organization of the United Nations).
1977. Guidelines for watershed management. FAO Conservation
Guide 1, FAO, Rome.
Farvar, M. T., and J. P. Milton (eds.). 1973. The careless tech
nology: Ecology and international development. Natural
History Press, New York.
Ferreiras, B. A. 1979. Propuesta de investigacin, estrategia de
la poblacin rural de la Sierra para enfrentar la pobreza.
Instituto Superior de Agricultura, La Herredura, Santiago,
Dominican Republic.
Flaxman, E. M. 1975. The use of suspended-sediment load measurements
for evaluation of sediment yield in the west. Ln Present and
prospective technology for predicting sediment yield and
sources. USDA ARS-S-40, Washington, D.C.
Fleming, G. 1968. The Stanford sediment model: I. Translation.
Int. Assoc. Sci. Hyd. Bull. 13:108-125.
, and K. M. Leytham. 1976. The hydrologic and sediment
processes in natural watershed areas. Proc. 3rd Federal
Inter-agency Sedimentation Conf., Washington, D.C.
Flinn, J. C. 1980. Opportunities for economic analysis of component
technology at field sites. Proc. Workshop on the Economics of
Cropping Systems, Manila, Philippines.
, S. Jayasuriya, and C. G. Knight. 1980. Incorporating
multiple objectives in planning models of low-resource farmers.
Aust. J. Agrie. Econ. 24:35-45.
, and J. Lagemann. 1980. Evaluating technical innovations
under low-resource farmer conditions. Exp. Agrie. 16:91-101.
Floyd, B. N. 1969. Agricultural innovation in Jamaica: The
Yallahs Valley Land Authority. Dept, of Geography, Univ. of
West Indies Occasional Pub. No. 4.
Forrester, J. W. 1971. World dynamics. Wright-Alien Press,
Cambridge.
Fournier, F. 1967. Research on soil erosion and soil conservation
in Africa. Afric. Soils. 12:53-96.


329
Table E-2--Amina River.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800101
0.57
0.57
-
-
-
-
-
3.2
3.2
3.2
-
800102
0.56
0.56
-
-
-
-
-
2.9
2.9
2.9
-
800103
0.59
0.58
-
-
-
-
-
3.8
3.5
3.7
-
800104
0.72
0.71
-
-
-
-
-
8.0
7.6
7.8
-
800105
0.60
0.58
-
-
-
-
-
4.1
3.5
3.8
-
800106
0.59
0.59
-
-
-
-
-
3.8
3.8
3.8
-
800107
0.90
0.84
-
-
-
-
-
16.7
13.3
15.0
-
800108
0.82
0.77
-
-
-
-
-
12.3
10.0
11.1
-
800109
0.65
0.63
-
-
-
-
-
5.7
5.3
5.5
-
800110
0.61
0.60
-
-
-
-
-
4.5
4.1
4.3
-
800111
0.60
0.60
-
-
-
-
-
4.1
4.1
4.1
-
800112
0.59
0.59
-
-
-
-
-
3.8
3.8
3.8
-
800113
0.59
0.59
-
-
-
-
-
3.8
3.8
3.8
-
800114
0.58
0.58
-
-
-
-
-
3.5
3.5
3.5
-
800115
0.57
0.57
-
-
-
-
-
3.2
3.2
3.2
-
800116
0.57
0.57
-
-
-
-
-
3.2
3.2
3.2
-
800117
0.56
0.56
-
-
-
-
-
2.9
2.9
2.9
-
800118
0.56
0.56
-
-
-
-
-
2.9
2.9
2.9
-
800119
0.56
0.56
-
-
-
-
-
2.9
2.9
2.9
-
800120
0.55
0.55
-
-
-
-
-
2.7
2.7
2.7
-
800121
0.55
0.55
-
-
-
-
-
2.7
2.7
2.7
-
800122
0.55
0.55
-
-
-
-
-
2.7
2.7
2.7
-
800123
0.54
0.54
-
-
-
-
-
2.4
2.4
2.4
-
800124
0.54
0.54
-
-
-
-
-
2.4
2.4
2.4
-
800125
0.54
0.54
-
-
-
-
-
2.4
2.4
2.4
-
800126
0.55
0.55
-
-
-
-
-
2.7
2.7
2.7
-
800127
0.54
0.54
-
-
-
-
-
2.4
2.4
2.4
-
800128
0.54
0.54
-
-
-
-
-
2.4
2.4
2.4
-
800129
0.54
0.54
-
-
-
-
-
2.4
2.4
2.4
-
800130
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
800131
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
800201
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
800202
0.53
0.53
-
-
-
-
2.2
2.2
2.2
-
800203
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
800204
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
800205
0.54
0.53
-
-
-
-
-
2.4
2.2
2.3
-
800206
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
800207
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
800208
0.55
0.78
-
-
-
-
-
2.7
10.4
6.5
-
800209
0.72
0.69
-
-
-
-
-
8.0
6.9
7.5
-
800210
0.66
0.60
-
-
-
-
-
6.0
4.1
5.1
-
800211
0.57
0.55
-
-
-
-
-
3.2
2.7
2.9
-


341
Table E-2--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
810518
0.80
0.90
-
-
-
-
-
11.3
16.7
14.0
-
810519
0.95
0.93
-
-
-
-
-
20.0
18.7
19.3
-
810520
0.90
0.84
-
-
-
-
-
16.7
13.3
15.0
-
810521
1.04
0.99
-
-
-
-
-
27.0
23.0
25.0
-
810522
0.98
1.18
-
-
-
-
-
22.2
40.4
31.3
-
810523
1.05
1.01
-
150
2.165
2.165
2.17
27.9
24.5
26.2
-
810524
1.30
1.84
2.89
-
-
-
-
51.4
122.0
86.7
375.2
810525
1.43
2.02
1.82
0
2.211
2.211
2.21
65.2
153.9
109.5
353.0
810526
1.34
1.15
-
136
0.733
-
-
55.4
37.7
46.5
-
810527
1.25
1.10
-
-
-
-
-
46.6
32.5
39.6
-
810528
1.13
1.44
-
120
2.963
1.176
2.07
35.5
66.3
50.9
-
810529
1.22
1.04
-
-
-
-
-
43.9
27.0
35.4
-
810530
1.01
0.98
-
-
-
-
-
24.5
22.2
23.4
-
810531
0.93
0.89
-
-
-
-
-
18.7
16.1
17.4
-
810601
0.99
0.96
-
-
-
-
-
23.0
20.7
21.8
-
810612
0.98
1.04
-
140
0.562
0.236
0.40
22.2
27.0
24.6
-
810603
1.00
0.97
-
-
-
-
-
23.7
21.5
22.6
-
810604
0.98
0.90
-
160
0.029
-
-
22.2
16.7
19.5
-
810605
0.87
0.99
-
-
-
-
-
15.0
23.0
19.0
-
810606
0.94
0.89
-
-
-
-
-
19.3
16.1
17.7
-
810607
0.85
0.99
-
-
-
-
-
13.9
23.0
18.4
-
810608
0.92
0.86
-
164
0.670
-
-
18.0
14.4
16.2
-
810609
0.85
0.84
-
-
-
-
-
13.9
13.3
13.6
-
810610
0.80
0.78
-
-
-
-
-
11.3
10.4
10.9
-
810611
0.82
0.81
-
-
-
-
-
12.3
11.8
12.1
-
810612
0.78
0.77
-
-
-
-
-
10.4
10.0
10.2
-
810613
0.76
0.75
-
-
-
-
-
9.6
9.2
9.4
-
810614
0.74
0.73
-
-
-
-
-
8.8
8.4
8.6
-
810615
0.72
0.70
-
-
-
-
-
8.0
7.3
7.6
-
810616
0.69
0.69
-
-
-
-
-
6.9
6.9
6.9
-
810617
0.68
0.67
-
-
-
-
-
6.6
6.3
6.5
-
810618
0.69
0.66
-
-
-
-
-
6.9
6.0
6.5
-
810619
0.68
0.69
-
-
-
-
-
6.6
6.9
6.8
-
810620
0.70
0.69
-
-
-
-
-
7.3
6.9
7.1
-
810621
0.75
0.73
-
-
-
-
-
9.2
8.4
8.8
-
810622
0.70
0.70
-
-
-
-
-
7.3
7.3
7.3
-
810623
0.68
0.68
-
-
-
-
-
6.6
6.6
6.6
-
810624
0.78
0.75
-
-
-
-
-
10.4
9.2
9.8
-
810625
0.74
0.74
-
-
-
-
-
8.8
8.8
8.8
-
810626
0.73
0.71
-
-
-
-
-
8.4
7.6
8.0
-
810627
0.70
0.70
-
-
-
-
-
7.3
7.3
7.3
-
810628
0.69
0.68
-
-
-
-
-
6.9
6.6
6.8
-
810629
0.71
0.82
-
-
-
-
-
7.6
12.3
10.0
-
810630
0.87
0.89
-
-
-
-
-
15.0
16.1
15.6
-


240
Table 32. Analysis of variance of runoff and sediment losses for
plots in pasture grouped by site.
Source
df
Mean Square
F value
Runoff
Model
1
7.13
3.53
Error
133
2.02
Sediment
Model
1
3.97
1.08
'Error
129
3.67


V




102
this relative measure of high rural population density becomes more
apparent when viewed against the backdrop of physical, biotic and
socioeconomic characteristics of the region.
Physical and Biotic Aspects
The landscape is characterized generally by rugged topography,
and pronounced intraregional variations exist in landforms, soils, and
parent material (Fig. 15) (Antonini and York, 1979). Climate also
varies dramatically, with average annual precipitation ranging from
1000 to 1800 mm (Jorge, 1970), influenced primarily by differences in
elevation from 100 to 1800 m (Antonini and York, 1979; Swedforest,
1980).
The drainage density of the area is high, due to the combination
of high rainfall and rugged topography. The project area contains the
upper and middle watersheds of three major rivers (Bao, Mao, and
Amina) which flow into the Yaque del Norte, the country's largest and
most important river (Fig. 16) (de la Fuente, 1976). Numerous
ephemeral and permanent streams of lesser magnitude are encountered
throughout the area. The discharge of both large and small streams
varies markedly, with flash floods occurring during the two rainy
seasons and many stream beds going dry during the midsummer and late
winter dry seasons. Interviews with local residents suggest that
deforestation has substantially exaggerated the normal range of
variation.
The majority of the soils of the region are classified as
Dystropepts or Eustropepts (Nicholaides and Hildebrand, 1980b)








150 240 320 400 480 350 540 "20 500
1931-197*3 IM i 11
196 7-1979 fill III
Fig. 25.
Monthly rainfall at the San Jose climatological station, #1, for
two periods.
138


61
Regional Reconnaissance
The inventory of existing land use systems and the condition of
soil and water resources within the Plan Sierra impact area provided
the data for refinement of the problem definition and for subsequent
application of the research design within successively smaller units
of analysis.
Most of the reconnaissance activities were completed between
January and March 1980. A regional description and summary of land
use systems was complied from library and field research. A review
and synthesis of maps, aerial photographs, statistical summaries and
literature relating to the study area preceded the field surveys. The
area was stratified into multitopic subregions based on cartographic
analysis of physical, biotic and socioeconomic characteristics mapped
at scales of 1:250,000 and 1:50,000. The major criteria for zonation
were life zones (Holdridge, 1967; OAS, 1967), topographic
characteristics, and land use characteristics, the latter reflecting
population density as well as condition and productivity of the land.
Maps were compiled by Plan Sierra cartographic and project staff from
topographic and thematic maps at 1:250,000 and 1:50,000 (Jennings,
1979a; OAS, 1967; Swedforest, 1980), and from aerial photographs at
1:20,000.
The field survey was similar to the general procedure outlined by
Hildebrand (1981). Several rapid reconnaissance surveys of erosion
features, land cover and land use systems were conducted within the
regional subdivisions outlined above, as part of Plan Sierra program
development in soil and water conservation. Survey teams varied in


100
watershed model. The relationship between various land uses at the
watershed level was examined, using the model as a guide. The
repercussions of drastic changes in the proportion of any given land
use within the system were studied. The difficulty of determining
watershed management by a composite of household decisions constrained
by the larger system was illustrated by juxtaposition of the models at
both scales of analysis. A nested model was constructed to illustrate
the relationships of mutual causality between the systems of different
scale within the physical and socioeconomic hierarchy.








360
Location: Pananao
Plot #: 91
Land Use: Perennial and annual crops
Land Cover: Sisal, Sweet Potato, Manioc, with Hillside Ditches
Parent Material: Conglomerates
Relief: Pronounced
Position: Mid-slope
Slope Class: 25-45%
Runoff: Rapid
Permeability: Moderately Rapid
Erosion Class: Moderate to severe
Drainage Class: Rapid (4)
Soil Moisture: Uniform, slightly moist
Salts/Alkaline: None
Stoniness: Moderate to high
Soil Profile:
Horizon
Depth
Boundaries
Color
Texture
Structure
Consistency
c3
Concretions
Slickensides
Mottles
Roots
A
0-10 cm
regular, sharp
5 YR 5/4
sandy, loam
granular
no
no
no
no
yes
C
10-22 cm
sharp
7.5 YR 5/4
sandy
granular
no
no
no
no
no*
*difficult to penetrate for fine roots of crop plants






35
Reports from Kenya, Tanzania, and Uganda (Blackie, 1972; Pereira,
1973; Pereira et al., 1962, 1967; Rapp, 1977) present useful data for
comparison with some of the land use systems of the Caribbean,
including coffee and other cash crop plantations, grazing, subsistence
farming and forestry. Paired watershed studies spanning four years or
more demonstrated the impact of both land cover and specific
management practices on runoff and erosion as well as on the harvest
within the watershed. In all of the cases cited, researchers
collected frequent stream discharge and precipitation measurements.
Some cases also include continuous monitoring of the above, as well as
sampling of suspended sediments in streamflow. The results of grazing
and range improvement trials in experimental watersheds in Uganda
include a more than twofold increase in depth of penetration of
rainfall into the soil, and a concurrent reduction in peak streamflows
after restoration of overgrazed grasslands (Pereira et al., 1962).
Data from experimental sites in Kenya (Blackie, 1972; Pereira, 1973)
document the effects of replacing tall evergreen forest with tea
plantations. The mean water yields over an 11-year period
effectively were equal for the forested control watershed and an
adjacent area planted in tea. The floods resulting from peak storm
events, however, varied substantially. The minimal impact of the tea
plantation reflects in part the stringent conservation measures
employed during its establishment. By contrast, clearing of
indigenous bamboo forest without immediate replacement by tree crops
increased streamflow 16% (Blackie, 1972). In Tanzania, a cleared
forest planted to a maize and vegetable single-crop system yielded a




146
Table 5. R values for regression analyses of subwatershed rainfall
vs. river discharge and sediment load.
Subwatershed
Rainfall
Area
R : Rainfall
vs. Discharge
R^: Rainfall
vs. Sediment Load
_1
mm year
km^
Amina River
A3
1227
70.0
0.09
0.09
A4
1473
110.6
0.06
0.10
A5
1719
156.9
0.10
0.05
MclO
River
M1
1240
95.6
0.02
0.24
M2
1268
126.9
0. 07
0.21
M3
1268
103.1
0.07
0.28
M4
1598
116.9
0.06
0.02
M5
1598
98.1
0.06
0.02
1598
96.9
0.06
0.04




63
to include relationships previously omitted or incorrectly defined.
The research design then was developed to test the major hypotheses
implied in the model. The discharge and sediment yield of two large
watersheds were measured over a 15-month study period. During the
same interval subwatersheds were described and monitored in greater
detail to relate differences in discharge and sediment yield to
varying physical and land use characteristics. Erosion plots
constructed within the subwatersheds provided comparative data on
runoff, erosion and production under different land uses, each with
varying conservation practices.
The 18-month period of study for phases two and three extended
from 1 Apr. 1980 to 30 June 1981. This period included a full
hydrologic year, from 1 Apr. 1980 to 1 Apr. 1981, and also allowed
repetition of sampling and monitoring during the time of peak
rainfall, from April through June.
The spatial and logistical organization of research activities
are illustrated in diagram and tabular form (Fig. 6, Table 1). The
chronological order of analytical and data collection procedures
parallels the general case described in Fig. 4. The choice of study
sites reflects the insights gained from the review and reconnaissance
survey, as well as the information needs of Plan Sierra.
Study Sites
The study sites selected within the Plan Sierra impact region
2
included two large watersheds (500 to 100 km ), five small watersheds
2
(1 to 20 km ) nested within the two larger units, and 16 plots on nine
landholdings situated within three of the small watersheds.


321
Table E-l~-continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800729
0.69
0.68
-
-
-
-
-
22.9
22.2
22.6
-
800730
0.66
0.65
-
-
-
-
-
20.9
20.3
20.6
-
800731
0.65
0.64
-
-
-
-
-
20.6
19.6
19.9
-
800801
0.63
0.62
1.14
-
-
-
-
19.0
18.4
18.7
-
800802
0.77
0.67
0.91
-
-
-
-
28.7
21.6
25.1
40.5
800803
0.75
0.66
-
-
-
-
-
27.2
10.9
24.1
-
800804
0.65
0.64
-
-
-
-
-
20.3
19.6
19.9
-
800805
0.63
0.65
1.12
-
-
-
-
19.0
20.3
10.6
62.0
800806
0.99
0.77
-
171
0.147
-
-
48.1
28.7
38.4
-
800807
0.70
0.67
-
-
-
-
-
23.6
21.6
22.6
-
800808
0.64
0.63
1.75
-
-
-
-
19.6
19.0
19.3
127.6
800809
0.89
0.77
1.02
-
-
-
-
38.7
28.7
33.7
51.2
800810
0.75
0.71
1.54
-
-
-
-
27.2
24.3
25.8
107.0
800811
0.95
0.80
-
-
-
-
-
44.2
31.1
37.7
-
800812
0.76
0.73
-
-
-
-
-
28.0
25.7
26.8
-
800813
0.69
0.67
-
-
-
-
-
22.9
21.6
22.2
-
800814
0.72
0.65
-
-
-
-
-
25.0
20.3
22.6
-
800815
0.65
0.64
1.00
-
-
-
-
20.3
19.6
19.9
49.2
800816
0.77
0.66
-
-
-
-
-
28.7
20.9
24.8
-
800817
0.65
0.64
-
-
-
-
-
20.3
19.6
19.9
-
800818
0.63
0.62
-
-
-
-
-
19.0
18.4
18.7
-
800819
0.60
0.58
0.66
-
-
-
-
16.3
15.5
15.9
20.9
800820
0.63
0.64
1.26
-
-
-
-
19.0
19.6
10.3
81.
8008212
0.73
0.64
0.96
-
-
-
-
25.7
19.6
22.7
45.2
800822
0.68
0.65
0.85
171
0.067
0.052
0.06
22.2
20.3
21.3
35.2
800823
0.67
0.64
0.82
-
-
-
-
21.6
19.6
20.6
32.7
800824
0.68
0.63
0.80
-
-
-
-
22.2
19.0
20.6
31.1
800825
0.67
0.64
0.86
-
-
-
-
21.6
19.6
20.6
36.0
800826
0.68
0.66
1.45
-
-
-
-
22.2
20.9
21.6
98.5
800827
0.90
0.77
1.18
-
-
-
-
39.6
28.7
34.1
69.1
800828
0.83
0.75
-
-
-
-
-
33.5
27.2
30.4
-
800829
0.70
0.68
1.00
-
-
-
-
23.6
22.2
22.9
49.2
800830
0.80
0.75
0.80
-
-
-
-
31.1
27.2
29.1
31.1
800831
0.77
0.73
0.94
-
-
-
-
28.7
25.7
27.2
43.3
800901
0.78
0.73
1.05
-
0.088
0.075
0.08
29.5
25.7
27.6
54.3
800902
0.78
0.73
-
-
-
-
-
29.5
25.7
27.6
-
800903
0.71
0.68
0.92
-
0.051
-
-
24.3
22.2
23.3
41.4
800904
0.75
0.71
-
-
0.027
-
-
27.2
24.3
25.8
-
800905
0.71
0.67
2.66
-
0.144
0.165
0.16
62.0
36.9
49.5
92.0
800907
0.96
0.87
2.16
-
-
-
-
45.2
36.9
41.1
170.4
800908
0.95
0.85
1.14
-
-
-
-
44.2
35.2
39.7
64.3




154
Several combinations of variables were tested. Average daily
discharge rates, morning, evening, and floodpeak stage recordings (at
the hydrometric station), and stage readings from the bridge at time
of sampling were all used as independent variables to explain sediment
transport (tons day 1), average sediment concentration (g L ^) and
maximum sediment concentration (Table 6). The regression of stage
reading at sampling time on average sediment concentration explained
the largest portion of the variance.
The morning and evening stage readings are not adequate
indicators of stage at sampling time. Average daily discharge is
derived from these readings, so it also fails to account for within-
day variation of stage. River stage at sampling time, by contrast,
accounts for 38% of the variation in average sediment concentration
(mean concentration per sample set) in the Mao River, and the same
variable explains 21% of the variation in sediment concentration in
2
the Amina River. Higher R values and F-values for daily sediment
discharge and river stage at time of sampling (Table 6) are
misleading, since sediment transport is derived by multiplying the
average sediment concentration by average daily discharge. The
discharge in turn is derived from morning and evening stage
measurements. The 70 and 43% explained variations in sediment
concentration for Mao and Amina, respectively, are partially artifacts
of autocorrelation. The clearest and most straightforward
relationship that emerges from the analyses is the one first cited
above. The resultant equations for the two rivers are given below.
Amina:
C =
s
e
(-0.0154 x L ) + 1.0218)
(1)


4
pressure on hillslope lands (Antonini et al., 1975; Santos, 1981)
coupled with high demand for food crops and relegation of large lowland
tracts to cash crop cultivation (Beckford, 1972; Rankine, 1976; Wilson,
1976).
The resource base in the Caribbean is subject to intensifying
multiple demands by commercial and subsistence sectors for food, cash
crop, wood, fuel and mineral production, as well as protection of the
watershed for downstream development. From a national perspective, the
upland watershed's most important export crop may well be water, needed
for irrigation and hydroelectric projects for downstream development
(Swedforest, 1980). The various types and rates of production demanded
are often competitive in nature, if not mutually exclusive (Crosson and
Frederick, 1977; McPherson, 1974). In many cases the upland regions'
internal situations also clearly indicate the need for change (Brush,
1981; Chaney and Lewis, 1980; Ferreiras, 1979; Hildebrand, 1981; Reiche
and Lee, 1978; Santos, 1981; SEA, 1978).
A Conceptual Model of the Problem
Based upon the information presented above, a model is postulated
that describes the interaction of significant elements in Caribbean
land use systems with regard to the problems of soil erosion and rural
poverty (Fig. 1). The model shows the interaction between population,
land use and the condition of soil, water, and vegetation both within
and outside the study region.
The model diagram follows the format developed by Odum for energy
modelling of ecosystems (Odum, 1971). The tank-shaped symbols (Fig. 1)






CHAPTER IV
RESULTS AND DISCUSSION
Regional Profile
Erosion and excessive storm runoff are serious problems within
the Sierra. Residents of the area are aware of soil loss and
depletion under shifting cultivation, bush fallow and continuous
cropping on the hillsides. Farmers point to declining yields as
evidence that the soil has "worn out". Many people from the area note
the disruption of river regimes and cite water quality and
availability as problems.
The condition of the land also speaks for itself. Erosion
features of almost every category are seen on the landscape. While
dramatic features, such as deep gullies or denuded hillsides, are not
common, there is widespread evidence of sheet and rill erosion. Many
areas also exhibit impoverished stands of natural vegetation which are
more subtle, secondary results of prior erosion. The varied
expression of the problem within the region reflects the physical and
cultural diversity of the area itself.
2
The Plan Sierra region includes 2500 km situated on the northern
slopes of the Cordillera Central. The population is approximately
120,000 (Chaney and Lewis, 1980) and the average population density is
-2
48 persons km While the density itself is not extreme, the region
has the highest ratio in the country of available labor force per unit
or arable land (SEA, 1978; Ferreiras, 1979). The full significance of
101








Location:
Pananao
Plot #:
94
Land Use:
Pasture
Land Cover:
Improved pasture,
overgrazed
Parent Material
: Conglomerate
Relief:
Pronounced
Position:
Mid-slope
Slope Class:
25-45%
Runoff:
Rapid
Permeability:
Moderate
Erosion Class:
Slight to moderate
Drainage Class:
Rapid (4)
Soil Moisture:
Uniform, slightly
moist
Salts/Alkaline:
None
Stoniness:
Slight to moderate
Soil Profile:
Horizon
A
C
Depth
0-25 cm
25-45 cm
Boundaries
sharp
sharp
Color
-
-
Texture
sandy loam
sandy
Structure
granular
granular
Consistency
no
no
C3
no
no
Concretions
no
no
Slickensides
no
no
Mottles
no
no
Roots
yes
yes








Table 13. Discharge rates measured for low flow conditions in small watersheds
Watershed Area Rainfall Discharge rate
ha
-1
mm yr
total
per i
unit area
Prieto (#64)
187.5
1719
0.08
3
m
-1
sec
4
X
io4
3 ~K ~l
m sec ha
Upper Bajamillo (#71)
95.0
1719
0.04
3
m
-1
sec
4
X
o
I1
3 ~K ~1
m sec ha
Bajamillo (#70)
2962.5
1719
0. 71
3
m
-1
sec
2
X
io4
3 ~K ~1
m sec ha
Pananao (#60)
1312.5
1240
0.12
3
m
-1
sec
1
X
H
O
3 ~K -1
m sec ha
Hondo (#67)
1285.0
1227
0.13
3
m
-1
sec
1
X
O
rl
3 ~K _1
m sec ha




334
Table E-2--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800729
0.66
0.75
-
-
-
-
-
6.0
9.2
7.6
-
800730
0.72
0.71
-
-
-
-
-
8.0
7.6
7.8
-
800731
0.68
0.65
-
-
-
-
-
6.6
5.7
6.1
-
800801
0.65
0.64
-
-
-
-
-
5.7
5.7
5.7
-
800802
0.67
0.69
-
184
0.024
0.025
0.02
6.3
6.9
6.6
-
800803
0.67
0.68
-
-
-
-
-
6.3
6.6
6.5
-
800804
0.65
0.66
-
-
-
-
-
5.7
6.0
5.8
-
800805
0.64
0.63
-
-
-
-
-
5.7
5.3
5.5
-
800806
1.50
1.45
-
100
0.260
0.231
0.25
73.4
67.4
70.4
-
800807
0.95
0.90
-
162
0.113
0.111
0.11
20.0
16.7
18.4
-
800808
0.85
0.86
-
-
-
-
-
13.9
14.4
14.1
-
800809
0.88
0.85
-
-
-
-
-
15.5
13.9
14.7
-
800810
0.83
0.82
-
-
-
-
-
12.8
12.3
12.6
-
800811
0.79
0.76
-
176
0.077
0.061
0.07
10.8
9.6
10.2
-
800812
0.78
0.75
-
-
-
-
-
10.4
9.2
9.8
-
800813
0.77
0.76
-
-
-
-
-
10.0
9.6
9.8
-
800814
0.70
0.75
-
-
-
-
-
7.3
9.2
8.2
-
800815
0.80
0.72
-
-
-
-
-
11.3
8.0
9.7
-
800816
0.68
0.67
-
-
-
-
-
6.6
6.3
6.5
-
800817
0.69
0.70
-
-
-
-
-
6.9
7.3
7.1
-
800818
0.65
0.66
-
-
-
-
-
5.7
6.0
5.8
-
800819
0.62
0.63
-
-
-
-
-
4.9
5.3
5.1
-
800820
0.60
0.65
-
-
-
-
-
4.1
5.7
4.9
-
800821
0.80
0.75
-
-
-
-
-
11.3
9.2
10.2
-
800822
0.72
0.73
-
-
-
-
-
8.0
8.4
8.2
-
800823
0.71
0.70
-
-
-
-
-
7.6
7.3
7.5
-
800824
0.69
0.67
-
-
-
-
-
6.9
6.3
6.6
-
800825
0.75
0.73
-
-
-
-
-
9.2
8.4
8.8
-
800826
0.74
0.70
-
183
0.052
0.057
0.06
8.8
7.3
8.0
-
800827
1.3
0.90
-
-
-
-
-
51.4
16.7
34.1
-
800828
1.20
1.15
-
-
-
-
-
42.1
37.7
39.9
-
800829
0.90
0.87
-
165
0.106
-
-
16.7
15.0
15.9
-
800830
0.85
0.81
-
-
-
-
-
13.9
11.8
12.8
-
800831
0.90
0.88
-
-
-
-
-
16.7
15.5
16.1
-
800901
0.95
0.92
-
160
0.064
0.096
0.08
20.
18.0
19.0
-
800902
0.80
0.78
-
-
-
-
-
11.3
10.4
10.9
-
800903
0.74
0.71
-
-
-
-
-
8.8
7.6
8.2
-
800904
0.70
0.68
-
-
-
-
-
7.3
6.6
6.9
-
800905
0.65
1.34
-
117
0.286
5.322
1.8
5.7
55.4
30.6
-
800906
0.98
0.94
-
-
-
-
-
22.2
19.3
20.8
-
800907
0.90
0.87
-
-
-
-
-
16.7
15.0
15.9
-
800908
0.95
0.89
-
-
-
-
-
20.0
16.1
18.1
-


73
subject matter covered. It also provided a convenient format for
recording and summarizing responses during or following the
discussion. Information noted on the diagrams and tables served to
evaluate the farm level models prior to initiating the erosion plot
experiments and watershed monitoring activities.
All sites were chosen to reflect variation in land use and
treatment, while slope and soil conditions were held constant and as
close as possible to the average for the watershed. Slope
measurements along the downslope transect were made prior to final
siting of all experimental plots.
Soil profile descriptions, characterization of soil samples by
laboratory analysis, and taxonomic classification constituted part of
the site description at each plot. Rectangular trenches at least 1 m
deep, 1 m long and 0.5 m wide were cut for observation and sampling.
Measurement and description of profile stratification, with detailed
description of color, texture, structure, and uniformity, by horizon,
were carried out according to the procedures outlined in the Soil
Survey Manual (USDA, 1951).
Munsell color charts were used for wet and dry color
determinations in the field (Munsell Color Co., 1951). Laboratory
analyses for N, P, K, and organic matter content followed standard
methods for determination of Kjeldahl N and Bray P by colorimetry, and
for determination of K by atomic absorption (USDA, 1975).
The North District Research Laboratory (CENDA) of the State
Secretariat of Agriculture conducted all laboratory tests for the
project, including physical and chemical characterization of soil






386
Plot 88. Pine Forest, Los Montones.
Sample Collection
Date
Sediment Yield
(kg ha ^)
Runoff Rate
, 3 -1,
(m ha )
80
05
14
4.54
14.36
80
05
31
73.67
29.99
80
06
10
81.20
43.78
80
06
20
14.26
5.57
80
08
08
1.39
10.33
80
09
13
3.20
6.68
80
09
25
0.00
0.00
80
10
02
0.00
0.00
80
10
17
0.00
0.00
80
10
30
0.00
0.00
80
11
06
0.52
3.51
80
11
29
0.65
1.36
80
12
10
0.34
0.67
80
12
23
0.78
1.00
80
12
29
0.55
1.36
81
01
08
0.41
2.16
81
01
19
0.73
3.51
81
02
04
4.17
9.07
81
03
16
1.45
9.07
81
03
20
0.43
2.16
81
03
25
1.68
2.16
81
04
03
0.42
1.75
81
04
09
0.29
0.67
81
04
14
1.60
2.59
81
04
27
0.00
3.51
81
04
28
1.10
6.12
81
05
05
2.32
11.63
81
05
06
0.00
4.51
81
05
12
6.98
20.24
81
05
28
4.67
19.48
81
06
09
0.24
6.12


220
in Table 23. The total monthly soil and storm runoff losses from May
1980 to June 1981 are compared with monthly rainfall for all of the
plots in Figs. 45, 46, and 47.
In general the soil and runoff losses reflect the monthly
rainfall distribution, particularly in forest, coffee and pasture.
The cultivated plots, however, show a marked peak in soil loss during
the first two months of data collection (May and June 1980 for most
stations, August and September for plots 82 and 83). Soil loss peaks
in relation to rainfall and runoff then decreases steadily. This
reflects the juxtaposition of the cropping cycle and the rainfall
regime. All of the cultivated plots were monitored from the early
stages of tillage and planting through just over a year of plant
growth, with interim harvests of some short cycle crops and a
continual increase in crop cover (canopy and detrital) by the longer-
lived crops such as manioc and sisal. The dramatic decrease in
erosion rates can be attributed to the combined effects of increased
crop cover and decreased tillage. A major exception to this trend
proves the rule. The highly eroded plot (95) in Pananao planted to
manioc in May 1980 showed little decrease in erosion losses relative
to the other cultivated plots (91 and 92). The latter were planted to
sisal and food crops in the same month, then the sweet potato and some
/
manioc were harvested and grass cover filled in between the sisal and
the remaining manioc plants. By contrast, the manioc on the severely
eroded site did not fare well. The plants did not establish a closed
canopy nor did they produce much litter. The soil was left almost
completely exposed throughout the study period. The limited extent to


157
multiplying the average sediment concentration (C ) given in the
s
3
previous equations (1, 2) by total daily discharge (Q) (m ). The
latter is derived from stage recordings (L ) by the following series
of equations:
If L < 0.54, then Q lo X
^ ^ r ^ ^ ._{C(log,-L ) + 0.8243/0.487}
If L > 0.54 and < 1.21, then Q = 10 ^10 x
x
^ ^ ^ {[(log L ) + 1.28883/0.727}
If L ^ 1.21, then Q = 10 ^10 x
x
for the Mao.
s ~ ^ _{[(log, _L ) + 0.3473/0.201}
If L < 0.64, then Q = 10 y10 x
x
It ^ ,r _{C (log. -L ) + 0.4173/0.303}
If L > 0.64 and < 1.15, then Q = 10 ^10 x
x
T, T >> ,c {C (log L ) + 0.5473/0.402}
If L ^ 1.15, then Q = 10 10 x
X
(5)
(6)
(7)
(8)
(9)
(10)
The estimates of total sediment transport given by the above
equations (5-10) represent conservative predictions (Table 7) because
they do not include the multiplicative effect of high discharge and
high sediment concentrations during short-lived flood events. The
additional estimate of peak one-hour sediment transport during floods
increased the predicted annual sediment yield by greater than 10% in
all cases.
The sedimentation rate of the dam to be built on the Mao River
will exceed the reported rate of 190,000 tons yr 1 for the Tavera Dam.
The useful life of the latter is expected to be severely curtailed by
sedimentation problems. The calculated rate indicates a need for
immediate action to avoid repetition of the Tavera case. The
197,888 tons yr-1 calculated yield for 1980 (without flood peaks)
includes the high rainfall and discharge rates for May of that year.


289
Table A-5. Erosion rates for selected agricultural land uses in the
United States.f
Land Use
Soil Loss
Tons ha ^
All Cropland
10.53
Cultivated Cropland
11.44
Pastureland,
including native pastures
5.83
Rangeland
6.94
Forest Land
2.69
Grazed Forest
Land
9.41
fSheet and rill erosion.
Source: USDA, 1980.


62
composition, but usually included the author and one to four
specialists and paraprofessionals in engineering, forestry, agronomy,
and soil conservation.
The surveys included formal and informal interviews with
residents, as well as field mapping of land cover, land use and
evidence of erosion and sedimentation in the various areas visited.
The selection of sites for more detailed observation reflected a
strong reliance on the knowledge of agronomists, foresters and
conservationists already familiar with the area, as well as the
opinion of residents as to what areas constituted typical or extreme
examples of particular physical and socioeconomic characteristics.
Field survey records included numerical data, maps, and interview with
residents, as well as the impressions and observations of the survey
team.
A synthesis of the cumulative results of prior field
reconnaissance by interdisciplinary teams of consultants (Chaney and
Lewis, 1980; Georges, 1981; Hart, 1981; Montero et al., 1981; Navarro,
1981; Nicholaides and Hildebrand, 1980b; Safa and Gladwin, 1981;
Santos, 1981; Swedforest, 1980) supplements the information gathered
from the author's survey. Written reports and personal communication
from the consultants and visiting scientists contributed to updates of
the regional profile.
Refinement of the Research Design
Based on the reviews of regional information and the completed
field reconnaissance, the conceptual model of the region was modified


172
Fig. 34.
System model of small watershed in coffee region.






118
employees work the land in illegal slash and burn plots. Their jobs
were curtailed by rapid deforestation and subsequent poor management
of the land. At the household level, the current situation creates
worsening shortages of fuel and lumber, which in turn, increase
illegal poaching on existing forests. Thus, erosion, deforestation,
poverty, and unemployment continue to spiral, each reinforcing the
other.
The erosion rate under forest varies with the conditions and
current use of stands. Observation indicates a less effective canopy
and ground cover in the piedmont dry forests than in the denser upland
pine forests (Jennings, 1979b; Jennings and Ferreiras, 1979; Mercedes
Urena, 1980). Many of the dry forests actually were degraded
successional stands in areas formerly covered in humid-subtropical or
transitional forest (mahogany); the successional stands themselves are
results of prior erosion. The current condition and use of the dry
forests encourage the further acceleration of erosion. Numerous land
slips and landslides have occurred on steep limestone slopes and along
the roads that traverse the dry forest areas between San Jose de Las
Matas, Janico, and Santiago (Fig. 17). The same erosion features
appear with less frequency in the upland pine forests.
Thus, many forests have been subjected to fairly severe erosion
during prior deforestation (within the last 20 years). Lumbering,
charcoal production and slash-and-burn agriculture continue to
accelerate erosion beyond the rates one would expect to find in dense
forest.




19
countries where research funds and personnel are limited and the
complexity of the rural agricultural landscape in the uplands requires
extensive calibration of the model. Rapid changes in farming
practice, crop types and level of technology in many areas would
require almost a continuous update of the calibration experiments.
Such a program of research not only represents a large investment in
and of itself, it implies a diversion of resources from alternative
avenues of theoretical and applied research directed toward cumulative
growth of knowledge about the processes in question.
Another weakness of the USLE is the failure of the conceptual
framework to account for the difference between land cover and land
use systems. There are no economic or social aspects to the model,-,
yet it is proposed as a practical tool for farm and regional
conservation planning purposes. The premises under which the equation
is applied often can be misleading, resulting in serious errors in
planning and management decisions. Aside from the logical pitfalls
inherent in the use of the model, there is the technical drawback of
its inability to predict the feedback effects of current erosion rates
on future land use and productivity which in turn affect future
erosion rates.
The USLE can be a useful tool for prediction of erosion rates
under specific known conditions, given prior calibration for the full
range of conditions in a region. Used alone, however, it does not
constitute an adequate basis for management decisions at the farm
level, much less for regional planning purposes.
The USLE has also been adapted for prediction of sediment yields
at the watershed scale (McElroy et al.,
1976; Onstad and Foster, 1975;


338
Table E-2--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
810112
0.85
0.83
-
-
-
-
-
13.9
12.8
13.3
-
810113
0.81
0.79
-
-
-
-
-
11.8
10.9
11.3
-
810114
0.73
0.71
-
-
-
-
-
8.4
7.6
8.0
-
810115
0.74
0.72
-
-
-
-
-
8.8
8.0
8.4
-
810116
0.70
0.72
-
-
-
-
-
7.3
8.0
7.8
-
810117
0.70
0.71
-
-
-
-
-
7.3
7.6
7.5
-
810118
0.73
0.72
-
-
-
-
-
8.4
8.0
8.2
-
810119
0.71
0.71
-
-
-
-
-
7.6
7.6
7.6
-
810120
0.73
0.70
-
-
-
-
-
8.4
7.3
7.8
-
810121
0.69
0.68
-
-
-
-
-
6.9
6.6
6.8
-
810122
0.68
0.66
-
-
-
-
-
6.6
6.0
6.3
-
810123
0.64
0.64
-
-
-
-
-
5.7
5.7
5.7
-
810124
0.62
0.61
-
-
-
-
-
4.9
4.5
4.7
-
810125
0.60
0.60
-
-
-
-
-
4.1
4.1
4.1
-
810126
0.65
0.63
-
-
-
-
-
5.7
5.3
5.5
-
810127
0.54
0.53
-
-
-
-
-
2.4
2.2
2.3
-
810128
0.62
0.61
-
-
-
-
-
4.9
4.5
4.7
-
810129
0.65
0.63
-
165
0.176
0.176
0.18
5.7
5.3
5.5
-
810130
0.61
0.61
-
-
-
-
-
4.5
4.5
4.5
-
810131
0.60
0.60
-
-
-
-
-
4.1
4.1
4.1
-
810201
1.50
1.10
-
135
0.305
-
-
73.4
32.5
53.0
-
810202
0.91
0.89
-
-
-
-
-
17.4
16.1
16.8
-
810203
0.79
0.76
-
-
-
-
-
10.9
9.6
8.4
-
810204
0.74
0.72
-
-
-
-
-
8.8
8.0
8.4
-
810205
0.84
0.83
-
-
-
-
-
13.3
12.8
13.1
-
810206
0.81
0.80
-
-
-
-
-
11.8
11.3
11.1
-
810207
0.80
0.79
-
-
-
-
-
11.3
10.9
8.6
-
810208
0.76
0.71
-
-
-
-
-
9.6
7.6
8.6
-
810209
0.65
0.64
-
-
-
-
-
5.7
5.7
5.7
-
810210
0.63
0.62
-
-
-
-
-
5.3
4.9
5.1
-
810211
0.64
0.62
-
-
-
-
-
5.7
4.9
5.3
-
810212
0.62
0.61
-
-
-
-
-
4.9
4.5
4.7
-
810214
0.59
0.62
-
-
-
-
-
4.1
4.1
4.1
-
810215
0.61
0.61
-
-
-
-
-
4.5
4.5
4.5
-
810216
0.68
0.66
-
-
-
-
-
6.6
6.0
6.3
-
810217
0.68
0.65
-
-
-
-
-
6.6
5.7
6.1
-
810218
0.65
0.64
-
-
-
-
-
5.7
5.7
5.7
-
810219
0.62
0.62
-
-
-
-
-
4.9
4.9
4.9
-
810220
0.60
0.61
-
-
-
-
-
4.1
4.5
4.3
-
810221
0.98
0.94
-
155
0.187
0.182
0.18
22.2
19.3
20.8
-
810222
0.91
0.88
-
-
-
-
-
17.4
15.5
16.5
-
810223
0.85
0.82
-
-
-
-
-
13.9
12.3
13.1
-


CHAPTER I
INTRODUCTION
The Problem
Almost 10 years after the first dramatic success of the Green
Revolution, the technological breakthroughs remain largely
inaccessible to small farmers of the underdeveloped world. Increased
yields have occurred primarily in large scale commercial or state
enterprises (Harris, 1973; Greenland, 1975; Stevens, 1977). Millions
of small farmers who produce commercial and subsistence food crops and
cash crops for export have maintained or increased production only at
great cost to themselves and to the natural environment (Crosson and
Frederick, 1977; Eckholm, 1976).
These farmers have- expanded into ever more marginal areas (often
arid, semiand, or hillslope environments) or have intensified
cultivation of existing plots, often already located in marginally
productive lands (Brush, 1981). The intensification has been for the
most part without benefit of new technology or capital inputs. It is
achieved through increasingly higher inputs of labor, and often
through practices that damage the long-term fertility and stability of
the soil (Geertz, 1972; Lagemann, 1977) and disrupt hydrologic and
geologic cycles in watersheds (Greenland, 1974; Kellman, 1969;
Pereira, 1973; Rapp, 1977).
Given this situation there is an immediate need for research to
adapt technologies to needs of small farmers within the limits of the
available factors of production and environmental constraints (Crosson
1




90
suspended sediment transport for low flow (non-flood) conditions also
was calculated for comparison of sediment transport between
watersheds.
The relationship between discharge and sediment concentration for
all events was tested by simple linear regression analysis (SAS, 1979)
with stage and discharge as the independent variables and sediment
concentration and transport as the dependent variables, respectively,
These analyses followed the procedures for derivation of sediment
rating curves used by the U.S. Bureau of Reclamation (Strand, 1975)
and by Rapp (1977) in Tanzania, as described in Chapter II.
Relationship of discharge and sediment transport to land use and
physical characteristes. The general linear model (GLM) program
package (SAS, 1979) tested the similarity of replicate watersheds and
the significance and degree of difference between watersheds by land
use and other categorical groupings, such as average slope, climate
and size. The similarity of replicates was tested by analysis of
variance, then the significance of differences between watershed
groupings based on land use and physical characteristics was tested.
Summary of production during the study period. The economic
production estimates for the study period were based in part on
extrapolation from intensive interviews with household members
participating in the more detailed farm level experiments. Informal
interviews with other farm families, farmworkers and shopkeepers
contributed to the general profile of economic production. The land
use maps and tables previously described provided a basis for rough
estimates of biological production and productivity. Information from






105
according to the Soil Taxonomy (USDA, 1975), although they vary
substantially in color, texture and chemical composition. As members
of the Inceptisol soil order they are primarily poorly developed
"young" soils, with development limited in most cases by the steep
slopes.
What most soils of this region have in common is their inadequacy
for many forms of agricultural use. They are poorly suited to the
semi-traditional land use systems that currently are found in the
Sierra. These soils are difficult to manage given their shallow
depth, low fertility, weak structure, and high susceptibility to
erosion by overland flow and mass wasting (Antonini et al., 1975;
Swedforest, 1980).
The natural vegetation reflects the combined variation of
geology, climate and soils (Fig. 17), while the actual vegetation is a
result of the natural vegetation further influenced by the intensity
of deforestation during the last 20 years, and by the character and
intensity of current land use. Land cover is a mosaic of forest,
coffee plantations, bush (secondary growth), small plots in mixed
field crops (conucos), and extensive areas in pastures.
Due to the highly "aggressive" climate, rugged topography, and
intrinsic properties of the soils, visible evidence of erosion is
widespread, and recovery from deforestation is slow (Swedforest,
1980). Many large land tracts are covered by eroded and/or compacted
pastures, or by secondary growth woodlands of poor quality.


SEDIMENT CONCENTRATION
153
Fig. 29.
Time series of sediment concentrations during selected
flood events.






Table F-l
Station 60, Pananao Stream
Date
MAXLEVL
SEDl
SED2
SED3
SED4
SED5
MAXDSCHG
FLDDSCHG
SEDDISl
SEDDIS2
SEDDIS3
SEDDIS4
SEDDIS5
800410
5
(g l"1)
91.48
48. 78
8.61
19.08
25.56
(m sec )
4.400
28.08
(kg sec "S
27.9
48.3
17.8
67.5
112.4
800412
1
30.89
-
-
-
-
0.306
-
9.4
-
-
-
-
800501
5
107.19
44.21
33.44
24.81
21.54
4.400
28.08
32.8
43.8
69.2
87.8
94.7
800502
4
99.40
52.71
44.46
31.52
-
3.540
-
30.4
52.2
92.0
111.5
-
800504
1
14.54
-
-
-
-
0. 306
-
4.4
-
-
-
-
800516
5
102.48
33.69
59.24
26.87
4.52
4.400
28.08
31.3
33.4
122.6
95.1
19.8
800524
5
108.54
155.90
5.50
153.57
40.72
4.400
28.08
33.2
154.6
11.3
543.6
179.1
800528
2
36.01
3.70
-
-
-
0.992
-
11.2
3.6
-
-
-
800530
4
8.16
13.50
-
18.16
-
3.540
-
2.5
13.3
-
64.2
-
800601
5
55.35
29.64
32.16
117.15
18.87
4.400
28.08
16.9
29.4
66.5
414.7
83.0
800613
3
43.58
5.40
0.44
-
-
2.070
-
13.3
5.3
0.9
-
-
800721
3
46.89
13.72
4.46
-
-
2.070
-
14.3
13.6
9.2
-
-
800805
2
1.03
0.07
-
-
-
0.992
-
0.3
0.0
-
-
-
800820
1
2.86
-
-
-
-
0. 306
-
0.8
-
-
-
-
800908
2
0. 56
0.08
-
-
-
0.992
-
0.1
0.0
-
-
-
800910
3
-
0.19
0.86
-
-
2.070
-
-
0.1
1.7
-
-
801007
1
12.28
-
-
-
-
0. 306
-
3.7
-
-
-
-
801015
5
47.74
21.09
-
-
-
4.400
28.08
14.6
20.9
-
-
-
801107
3
48. 39
4.05
2.86
-
-
2.070
-
14.8
4.0
5.9
-
-
801220
4
-
0. 30
-
-
-
3.540
-
-
0.3
-
-
-
810106
1
5.68
-
-
-
-
0.306
-
1.7
-
-
-
-
810108
2
24.64
6.46
-
-
-
0.992
-
7.5
6.4
-
-
-
810112
1
18. 39
-
-
-
-
0. 306
-
5.6
-
-
-
-
810115
3
56.61
25.09
~
-
-
2.070
-
17.3
24.8
-
-
-
810118
1
1. 37
-
-
-
-
0. 306
-
0.4
-
-
-
-
810202
1
0.66
-
-
-
* -
0.306
-
0.2
-
-
-
-
810206
5
18. 14
8. 21
3.08
73.69
69.56
4.400
28.08
5.5
8.1
6.3
260.8
306.0
810221
2
9.61
42.47
-
-
-
0.992
-
2.9
42.1
-
-
-
810329
3
31.03
-
2.41
-
-
2.070
-
9.5
-
4.9
-
-
810428
4
20.10
34.07
30.05
9.46
-
3.540
-
6. 1
33.8
62.2
33.4
-
810502
4
36.67
16.08
13. 34
4.65
-
3.540
-
11.2
15.9
27.6
16.4
-
810503
5
40.50
23.44
38. 78
-
-
4.400
28.08
12.3
23.2
80.2
-
-
810505
2
-
0.30
-
-
-
0.992
-
-
0.3
-
-
-
810506
1
29. 38
-
-
-
-
0. 306
-
8.9
-
-
-
-
810510
4
13.15
111.51
28.13
13.91
-
3.540
-
4.0
110.6
58.2
49.2
-
810515
1
12.83
-
-
-
-
0.306
-
3.9
-
-
-
-
810516
2
4.65
1.65
-
-
-
0.992
-
1.4
1.6
-
-
-
810524
2
0.28
2.59
-
-
-
0.992
-
0.0
2.5
-
-
-
810528
4
35.82
7.55
7.25
8.96
3.540
-
10.9
7.4
15.0
31.7
3
Discharge rates (m
-1,
sec )
for each
level,
corresponding
to concentrations
SEDl
through
SED5 are
as follows

o
i
i
306, 2 -
0.992
,3-2
.070, 4
- 3.540
, 5 -
4.400.
344


414
Rocheleau, D. 1980. Diseno y operacin de una red de muestreo y
parcelas experimentales para medir la cantidad de erosion y
sedimentacin bajo varios sistemas del uso de terreno en la
Sierra. 1st Seminario Sobre Manejo de Cuencas Hidrogrficas,
Santo Domingo, Dominican Republic, May.
. 1981. El papel del gegrafo y investigaciones geogr
ficas en proyectos de desarollo rural: El Plan Sierra como
ejemplo. Pan American Conf. of Geographers, Santo Domingo,
Dominican Republic, Aug.
Rodriguez, L. A. 1980. History of the timber industry in Santo
Domingo. ISA, Santiago, Dominican Republic.
Roehl, J. W. 1962. Sediment source areas, delivery ratios and
influencing morphological factors. Int. Assoc. Sci-Hyd. 59:
202-13.
Rosero, P. 1979. Some data on secondary forest managed in Siguiries,
Costa Rica. Proc. Workshop on Agro-forestry Systems in Latin
America. CATIE, Turrialba, Costa Rica, March.
, and N. Gewald. 1979. Growth of laurel (Cordia alliodora)
in coffee and cacao plantations and pastures in the Atlantic
region of Costa Rica. Proc. Workshop on Agro-forestry Systems in
Latin America. CATIE, Turrialba, Costa Rica, March.
Russo, I. 1980. Importancia de un plan de reforestacion en las
cuencas altas. 1st Seminario de Cuencas Hidrogrficas, Santo
Domingo, Dominican Republic, May.
Russo, R. 0. 1983. Efecto de la poda de Erythrina poeppigiana (Wal-
pers) O. F. Cook (Poro) sobre la nodulacion, produccin de biomasa
y contenido de nitrgeno en el suelo, en un sistema agroforestal
"cafe-poro". M. S. Thesis, CATIE, Turrialba, Costa Rica.
Ruthenberg, H. 1976. Farming systems in the tropics. Clarendon
Press, Oxford, England.
. 1980. Farming systems in the tropics. 3rd Ed. Clarendon
Press, Oxford.
Safa, H., and C. Gladwin. 1981. Designing a woman's component for
Plan Sierra. Rpt. to Plan Sierra, San Jose de las Matas,
Dominican Republic.


¡>H oO Q o NI 0< > oJh o( o op oS CU
Table 2. Equations for system model
= Jn K RT
Li
= R K5L2 K6 K12JrTpHL3 K182.
= K,
K^RT
L1
- K D(K OL + K,0)
3 5 6
[ (K OL ) + (K 0)]
5 2 6
= K RT K K RT
1L
1 1
+ K3K(K5OL2 + K60> KnQ
(K OL + K 0)
5 2 6
= K5OL2 + K6 K92 K105
= K16JrOTPHL3 K1?V K19aVLcPH K24bVAP K^VHA K^V
- K21VL0PH K22Y K2 3Y
= K_ _VHA K A K_ VAP K A
27 28 29 30
= K M K VHA K,_JrOTHP
33 26 15
K36UP + K35M
= K24aY + K31A + K41V K32M K34H K42M
= K36bUP + K37VAP K38P K39P q4JrPL3
o
Jr =
0
1 + K OTPHL V
16 3


Table F-5
Station 71, Upper Bajamillo Stream
Date MAXLEVL
SEDl
SED2
SED3
SED4
SED5
MAXDSCHG
FLDDSCHG
SEDDISl
SEDDIS2
SEDDIS3
SEDDIS4
SEDDIS5
800619
5
0. 39
0.19
0.13
0. 10
0. 10
2.360
9.96
0.0
0.1
0.1
0.2
0. 2
800708
l
0.28
-
-
-
-
0.213
-
0.0
-
_
_
_
800915
3
-
-
-
-
-
1.161
_
_
_
_
_
_
801001
4
-
-
-
-
-
1.920
-
_
-
_
_
_
801010
5
13. 36
32. 33
10. 35
5.14
4.54
2.360
9.96
2.8
19.1
12.0
9.8
10.7
801022
5
33.46
-
-
0.41
19.95
2.360
9.96
7.1
-
-
0.7
47.0
801105
1
-
-
-
-
-
0.213
-
_
_
_
_
_
801106
2
0.94
0. 28
-
-
-
0.593
-
0.2
0.1
_
_
_
801202
1
11.07
-
-
-
-
0.213
-
2.3
_
_
__
_
801210
1
10. 71
-
-
-
-
0.213
-
2.2
-
_
_
_
810202
2
3.73
0.82
-
-
-
0.593
-
0.7
0.4
-
_
_
810504
4
1.70
1.41
5.05
2.46
-
1.920
-
0. 3
0.8
5.8
4.7
_
810506
4
7.74
0.40
9. 71
7.42
-
1.920
-
1.6
0.2
11.2
14.2
-
810526
5
22.98
0. 38
1.41
9.00
0. 34
2.360
9.96
4.8
0.2
1.6
17.2
0.8
810609
1
0.65

~
0.213
-
0.1
Discharge
3
-1
rates (m
sec )
for each
level,
corresponding
to concentrations
SEDl
through
SED5, are
as
follows:
1 -
0.213,
2-0.
593, 3 -
1.161,
4 -
1.920, 5 -
2.360.
348




30
from the Connecticut study. Transport rates from harvested forests
are extremely high, even in comparison to agricultural uses. The high
rates, however, are offset by the fact that forest harvests are
periodic events that only occur once every 15 to 40 years, even in the
fast-growing pine plantations of the Southeast and the Caribbean. By
contrast, many agricultural uses are sustained continuously on a given
parcel of land.
Recent studies in agricultural watersheds in the United States
concentrate on cropping systems and practices in large scale
commercial farms or ranches. Pollution of surface waters by chemical
fertilizers (Haith and Dougherty, 1976) and pesticides often
overshadows sediment pollution as a subject of public concern (USDA
Soil Conservation Service, 1980). Pathogens entering the waterways
from feedlots and grazing lands (Jewell and Smith, 1976) also attract
more attention, although sediment is the major pollutant, by volume,
discharged into surface waters from agricultural lands. Suspended
sediments in streams and rivers carry pathogens as well as chemical
pollutants. Much of the current research, however, emphasizes the
chemical by-products discharged into waterways in solution from
agricultural non-point sources (Rao, 1980).
Studies of erosion in individual plots offer more information on
the variation in erosion rates with changes in cropping systems, farm
management and conservation practices. Data analyses by Wischmeier
and Smith (1978) for cropped and clean-tilled plots corroborate the
conclusions of studies in forest ecosystems. The canopy cover and
ground cover on the site determine how much rainfall energy reaches


Location:
Plot #:
Land Use:
Land Cover:
Parent Material:
Relief:
Position:
Slope Class:
Runof f:
Permeability:
Erosion Class:
Drainage Class:
Soil Moisture:
Salts/Alkaline:
Stoniness:
Pananao-El Rubio
96
Pine forest
Pines, 20-40 yrs,
Acid metamorphics
Pronounced
Mid-to-lower slope
25-45%
Rapid
Slow
Slight
Moderate (3) (due to litter)
Uniform, moist
None
Slight
Soil Profile:
Horizon
A
Depth
0-4 cm
Boundaries
irregular
Color
-
Texture
sandy
Structure
granular
Consistency
-
C3
no
Concretions
no
Slickensides
no
Mottles
no
Roots
yes*
selectively cut in past
otherwise rapid
C
4 20 cm
irregular
sandy
granular
no
no
no
no
yes
*roots from trees plus fine roots of herbaceous plants








87
Stage measurements were made at the posts installed as supports for
stationary sediment samplers. An empirical parabolic equation was
derived by plotting field measurements of stage against the discharge
measured at that time. The discharge rate (Q) equals an empirically
derived constant (a) times the square of the stream depth (as
indicated by stage measurements): Q = a(h)2. The discharge rates, by
stage, are listed for each site in Appendix F.
Suspended sediment sampling in the streams required installation
of stationary samplers since flash floods are difficult to predict and
preclude the use of wading samplers. A modified, locally constructed
version of the USDA 59 siphon sampler (Fig. 11) collected samples at
pre-determined stage height intervals during the rising stage of each
flood. The sampling equipment is similar to the "Hayim 7" sampler
used successfully in flash flood sampling in Tanzania and Israel
(Rapp, 1977).
The equipment was located at the center, if possible, or to one
side of the surveyed cross-section in each stream (Fig. 11), depending
upon the requirements for stable installation. All samplers carried
at least five sample bottles arranged in a vertical series, from a
point just above the low flow water surface to approximately 1 m above
that point. The exact range depended upon the indicators of flood
stage along the stream banks. The cross-section profiles with the
position of the sampler and the height of sample intakes are included
in Appendix B.
Project personnel collected the samples as soon as possible after
each flood. The contents of each collection bottle were placed in a








370
Location:
Land Use:
Plot #:
A = RKLSCP
R
K % silt
% sand
% O.M.
Structure
L
S
C
P
Los Montones
Pigeon Pea, Minimum
82
= 29
= 63
=(6.16)4 h K =
= 3
Permeability = 1
22m
46.5%
= LS
tillage
700
.11
13.2
.2, .9
1, .50
A
A
A
A
(c=.2)
-1
-1
(p=l)
= 203.28
t
ha
yr
cm m
. .
II ll
O i
= 101.64
t
ha ^yr ^
(c=.9)
(p=l)
= 914.76
t
ha lyr ^
(c=.9)
(p=.5)
= 457.38
t
, _1 _1
ha yr
Additional Information:
% clay = 8.0
% N = 0.31
Los Montones _
Pigeon pea, Conventional Tillage
83
= 700
= 29
= 63
= (6.16)4 L K = .11
Location:
Land Use:
Plot #:
A = RKLSCP
R
K % silt
% sand
% O.M.
Structure = 3
Permeability = 1
L 22 m = LS =12.7
S 44%
C = 2, .9
P = 1
A (c=.2) = 195.58 t ha Xyr 1
A (c=.9) = 880.11 t ha ^yr ^
Additional Information:
% Clay = 8.0
% N = 0.31




135
effect on the structure of the model but were reflected in the local,
evaluated models for small watersheds and plots.
Comparison of the Precipitation and Discharge During the Study Period
with Period of Record
Rainfall in the Mao and Amina watersheds during the 15-month
study period showed a sharp peak in the spring relative to the mean
for the period of record, as illustrated by data from Moncion and San
Jose de Las Matas (Figs. 23 and 24). Figures D-l through D-9 in
Appendix D compare the mean monthly rainfall from 1967 through 1979
with the values from January 1980 through June 1981 at nine other
stations. A 13-year record was not sufficient by itself as a baseline
for comparison with the study period (USDA, 1979). However, the mean
monthly rainfall for the 13-year period of record compared well with
the 50-year means available for the San Jose de Las Matas and Moncion
climatological stations (Figs. 25 and 26). All monthly means for the
13-year period were less than one standard deviation from the 50-year
means. The close correspondence between the means of the two periods
justified the use of the shorter record available for the other
climatological stations and for the Amina and Mao Rivers.
Rainfall data for both 1980 and 1981 exhibited a typical bimodal
distribution at 10 of the 11 stations (see Appendix D, Figs. 1-9).
The first (spring) peak was high in comparison to both other spring
peaks and the second (fall) peak for the same year. Even at stations
that normally have two rainy seasons of equal magnitude (see Appendix
D, Figs. 8 and 9), the May maxima were sharply accentuated for 1980
and 1981. Aside from this month, the 1980 and 1981 monthly rainfall
values closely approximated the means for 1967 to 1979.




59
tentative solutions already proposed. This iterative approach has
already been tested in farming systems research and extension programs
(Hildebrand, 1981).
The usual concept of applied research is one of a finite activity
to be carried out and completed by specialists, after which they will
offer a set of definitive conclusions to be implemented. In this
case, the study area was viewed as a system in a flux, constantly
adjusting to changes in internal and boundary conditions. The object
of study in this case also had a subjective component. The role of
people in determining the direction of ecosystem evolution was taken
into account. Residents of the region contributed to the
investigation as both informants and participants in data gathering,
experimental and verification procedures. The researcher participated
in an on-going experiment in which people living in the area sought
short-term relief and long-term solutions to problems at least
partially perceived and defined by them.
The study was designed to accommodate the distinct priorities and
information needs of local clients, scientists, and the regional
policy sector. The experimental design and data analyses tested
multi-faceted hypotheses concerning technology and land use
alternatives for the region. Each experiment included: a primary
hypothesis as to the biophysical or economic performance of a
particular alternative; a secondary hypothesis concerning how the
proposed change would fit into the existing system; and a third
hypothesis that the system, as such, could and should be sustained,
with modifications.










Fig. 2. Impact area: Plan Sierra.


197
watershed size categories and other groupings based on site
characteristics and land use (Table 14). The peak floods that
occurred in the Hondo and Pananao streams during the study period
transported more sediment than the amount exported in a full year's
discharge under low flow (non-flood) conditions. In these two cases,
the texture of the eroded material, the rainfall regime, and land use
interact to produce high pulses of sediment transport. This contrasts
with the more even discharge of sediments in the Bajamillo watershed
(Table 12).
The Prieto watershed also exhibited a high peak flood sediment
yield relative to the other watersheds and to its own sediment yield
for low flow conditions. The high sediment export during this flood
can be attributed to the combination of an intense storm in October
1980 (54 mm rainfall) with the recent weeding and tilling of coffee
and food crop plots on the hillslopes upstream.
In spite of the contrast in sediment concentrations during floods
and the pronounced differences in the temporal distribution of
suspended sediment export, the sediment yields per unit area are
relatively uniform for all five watersheds. A graph of two
hypothetical distributions (Fig. 42) illustrates the export of
equivalent amounts of sediment (per unit area) in watersheds
characterized by very distinct regimes of discharge and suspended
sediment export. The areas under the curves are equivalent and
represent suspended sediment export. The first case exhibits the
pulsed export observed at Pananao and Hondo and to a lesser extent at
the Prieto watershed (Tables 12 and 14). The peaks in sediment


98
The total sediment volume was read directly from the nomograph.
Total runoff for each plot was calculated by summing the water volume
for all the tanks on the plot.
Total soil loss was determined by combining the total volume and
sediment volume calculated above with the sediment concentration of
the samples. For the cases in which the tank contents were mixed
prior to sampling, the total soil loss (Ls) from each plot and for
each sample period was determined by the following equation:
Ls(tons) = Cs(tons m x Vt(m^),
where Cs = average sediment concentration of all three samples, and Vt
= total sample volume in tank.
In cases that required separate sampling of the sediment, the
concentration of the sediment sample (Csl) was multiplied by sediment
volume (Vs) and the suspended sediment concentrations were averaged
(Cs2) then multiplied by the corresponding liquid sample volume (VI)-.
The total soil loss was calculated as follows:
-33 -33
Ls(tons) = Cs(tons m ) x Vs(m ) + [Cs2(tons m ) x vl(m )],
where VI = Vt Vs. The program used for both calculations is
included in Appendix E.
Relationships between rainfall, runoff and erosion. Rainfall
data for the plot experiments were take from the same source as the
small watershed study areas in which the plots were situated. The
rainfall totals for the plot sampling intervals were calculated
separately for each case, since the sampling schedule was not uniform
within or between sites. Simple linear regression was used to test
and describe the relationships of total rainfall amount, total runoff,
and soil loss per unit area for each land use and management category


-sJ
Fig. B-2. Amina River cross section.
298


124
The contribution of soil moisture to stream baseflow (K^) is more
constant but proceeds at a much slower rate than storm runoff. The
two combined rates constitute the surface water discharge (K^) from a
given area. During flood events, part of the transported soil (K )
forms deposits (D) in the stream channels; the remainder (K z)
reaches the reservoir (Q). A small proportion of the sediments that
reach the reservoir escapes in suspension (K ) with throughflow
(K ). If the reservoir is full, both water (Kn) and sediment K )
i u y o
bypass or overflow the dam.
With successive storm events, sediment deposits in the upstream
channels may be re-suspended (K^) an<3 transported downstream (K^).
The rates of deposition and re-suspension depend upon particle size
and streamflow velocity. In the model these flows are proportional to
the stream discharge rate (K ), and particle size is assumed to be
constant for the given area.
The sensor on land use (L) represents the influence of composite
land use in the region. Erosion, runoff and production are affected
separately, although the resultant flows interact. For example, a
high rate of erosion releases a large mass of sediment to surface
waters, while a high rate of runoff increases the sediment transport
capacity of the streams. Each land use type has a separate
coefficient for erosion, runoff and production. The composite land
use coefficient for each process (L^, i, L^) expresses the net effect
of combined land use.
The detailed land use model (Fig. 20) illustrates the most common
spatial associations and rotation sequences (Table 3).








34
PAGE
Analysis of variance of runoff and sediment losses for
plots planted in crops grouped by site 242
35 Runoff and sediment losses for plots in Los Montones and
Pananao sites with crops, results of the Duncan Multiple
Range Test 243
36 Soil loss coefficients by land use and conservation
practice 247
37 Comparison of measured and predicted erosion losses 252
38 Comparison of soil loss coefficients derived from
USLE and from empirical data 255
39 Storm runoff coefficients, total runoff estimates and
runoff/rainfall ratios, by watershed 271
40 Calculation of soil loss coefficients and soil loss
estimates by watershed 272
viii




274
baseflow and peak annual flood showed that while the behavior of the
streams may be distinct, the net result (sediment yield) is very
similar. The calculated erosion rate does indicate the lowest soil
loss from the surface of the two small watersheds in the uplands of
the coffee region (Table 40).
The larger Bajamillo, however, exhibits the highest average
sediment yield per unit area for flood events, and closely
approximates the yields for Hondo and Pananao as indicated previously
in the a posteriori analyses. In this case the size of the watershed
relative to Hondo and Pananao rules out the explanation of sediment
delivery ratio. Uniformity of sediment yield at the same or similar
scales suggests a uniform erosion rate. The calculated rates for the
larger Bajamillo illustrate the potential for erosion equivalent to
that at Pananao. The variation of the proportion of land in annual
crops between 25 and 35% changes the composite erosion rate from 5 to
7 tons ha ^ yr ^.
The combination of measured sediment yield and calculated erosion
rates for the five small watersheds disproves the hypothesis that the
coffee region loses less soil per unit than the pasture and croplands
in the drier areas at lower elevations. Rather, it is the proportion
of land in field crops that determines the erosion rate on the
watershed. In spite of the visible differences between the two
landscapes the coffee-field crops and pasture-field crops associations
do not differ substantially in terms of erosion rates or sediment
yields. The population density and resultant demand for food crop
production (and/or annual cash crop production to generate income) are
roughly the same for the two areas.


331
Table E-2--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800325
0. 54
0.54
-
-
-
-
-
2.4
2.4
2.4
800326
0.54
0.54
-
-
-
-
-
2.4
2.4
2.4
-
800327
0.53
0.53
-
191
0.018
0.018
0.02
2.2
2.2
2.2
800328
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
800329
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2

800330
0.53
0.53
-
-
-
-
-
2.2
2.2
2.2
-
800331
0.52
0.52
-
-
-
-
-
2.0
2.0
2.0
-
800401
0.52
0.52
-
-
-
-
-
2.0
2.0
2.0
-
800402
0.52
0.52
-
-
-
-
-
2.0
2.0
2.0
-
800403
0.52
0.52
-
-
-
-
-
2.0
2.0
2.0
800404
0.52
0.52
-
-
-
-
-
2.0
2.0
2.0
-
800405
0.51
0.51
-
-
-
-
-
1.8
1.8
1.8
-
800406
0.51
0.51
-
-
-
-
-
1.8
1.8
1.8

800407
0.52
0.51
-
-
-
-
-
2.0
1.8
1.9
-
800408
0.51
0.51
-
-
-
-
-
1.8
1.8
1.8
-
800409
0.78
0.75
1.94
-
-
-
-
10.4
9.2
9.8
139.2
800410
1.05
0.95
-
-
-
-
-
27.9
20.0
24.0
-
800411
0.80
0.75
1.99
-
-
-
-
11.3
9.2
10.2
148.3
800412
1.07
0.98
-
-
-
-
-
29.7
22.2
25.9
-
800413
0.84
0.80
-
-
-
-
-
13.3
11.3
12.3
-
800414
0.77
. 74
-
-
-
-
-
10.0
11.3
12.3
-
800415
0.68
0.66
-
-
-
-
-
6.6
6.0
6.3
-
800416
0.65
0.65
-
-
-
-
-
5.7
5.7
5.7
-
800417
0.64
0.61
-
-
-
-
-
5.7
4.5
5.1
-
800418
0.70
0.70
-
183
0.062
0.071
0.07
7.3
7.3
7.3
-
800419
0.71
0.76
-
-
-
-
-
7.6
9.6
8.6
-
800420
0.79
0.75
-
-
-
-
-
10.9
9.2
10.0
-
800421
0.72
0.70
-
-
-
-
-
8.0
7.3
7.6
-
800422
0.75
0.73
-
-
-
-
-
9.2
8.4
8.8
-
800423
0.68
0.67
-
-
-
-
-
6.6
6.3
6.5
-
800424
0.65
0.63
-
-
-
-
-
5.7
5.3
5.5
-
800425
0.62
0.60
-
-
-
-
-
4.9
4.1
4.5
-
800426
0.75
0.72
-
-
-
-
-
9.2
8.0
8.6
-
800427
0.70
0.67
-
-
-
-
-
7.3
6.3
6.8
-
800428
1.05
1.12
-
-
-
-
-
27.9
25.3
26.6
56.2
800429
0.95
0.91
-
-
-
-
-
20.0
17.4
18.7
-
800430
0.94
0.90
-
-
-
-
-
19.3
16.7
18.0
-
800501
0.97
1.25
-
-
-
-
-
21.5
46.0
34.0
-
800502
1.50
1.20
3.30
-
-
-
-
73.4
42.1
57.7
522.0
800503
1.50
1.25
3.15
-
-
-
-
73.4
46.6
60.0
464.9
800504
1.08
1.08
1.30
-
-
-
-
30.6
30.6
30.6
51.4
800505
1.00
0.98
-
-
-
-
-
23.7
22.2
23.0
-


Fig. 28. Plan Sierra Region: Annual rainfall for 1980.
i
143






156
where Cis average sediment concentration (g la ^ ), and (cm) is
stage measured at the bridge as distance from the reference point to
the water surface.
Mao:
r (-0.1423 x L ) + 0.1601
= e f (2)
The equations are both significant to the 0.005 level and can be
used to predict sediment concentration and to derive sediment
transport estimates based on the available stage readings for both
rivers. The best predictions will be obtained with continuous stage
recordings from continuous monitors. Such data are available at times
for the Mao River. It is strongly recommended that maintenance of the
continuous monitor be improved at Mao and that a similar instrument be
placed at the Amina station. However, the equation can also apply to
the morning and evening spot readings. The resultant sediment
concentration predictions simply will have the same inherent
limitations as the current discharge calculations routinely based on
the stage readings. In the absence of such continuous data the total
annual sediment transport for the study period was calculated on the
basis of the twice daily stage readings. These were converted to L
equivalent values by the following equations for Mao and Amina,
respectively.
L = (L 2.559)/-0.0150 (Mao) (3)
Jl X
and
Lf = (Lx ~ 2.471 )/-0.0097 (Amina) (4)
where is the stage reading at the gauge for morning, evening or
floodpeak recordings. Daily sediment transport was calculated by


390
Plot 93. Pasture, Pananao.
Sample Collection
Date
Sediment Yield
(kg ha ^)
Runoff Rate
, 3, 1,
(m ha )
80
05
05
140.43
47.58
80
05
12
91.44
4.57
80
05
18
39.01
47.58
80
05
24
1594.29
47.58
80
05
28
0.00
0.00
80
06
01
11.89
47.58
80
07
07
34.73
47.58
80
08
20
1.70
42.45
80
08
26
2.72
8.36
80
08
28
0.02
0.59
80
19
08
2.79
12.69
80
09
29
2.60
23.68
80
10
14
10.54
50.19
80
11
07
6.59
41.20
80
12
10
0.78
3.89
80
12
11
1.23
2.05
81
01
07
0.65
3.24
81
01
12
13.01
25.26
81
01
28
1.35
6.77
81
02
06
10.81
29.23
81
02
21
7.75
29.23
81
03
27
0.57
1.02
81
04
09
0.12
0.59
81
04
28
7.35
13.61
81
04
29
1.96
1.02
81
05
02
0.90
1.02
81
05
06
30.96
91.28
81
05
10
33.22
92.57
81
05
12
3757.90
90.00
81
05
16
0.47
2.05
81
05
24
14.40
45.00
81
05
26
4.63
13.61
81
05
28
6.72
29.23
81
06
16
0.48
5.28




68
The choice of sampling sites for more detailed study of discharge
and sedimentation was based on several criteria, including ease of
access, proximity to homes of observers, established hydrologic
monitoring sites, proposed dam sites, and regularity of the
longitudinal and cross-sectional profiles of the channel in the
vicinity of the potential sampling points. Profiles of the river
cross-sections were surveyed. The relative heights of the bridges and
other large structures along the stream bank also were measured to
provide reference points for reporting maximum flood stages. The
profiles are included in Appendix B.
Small watersheds
The small subwatersheds selected for further study were analyzed
by standard cartographic and photogrammetric methods to determine
total area, average slope, and area and distribution of land cover
types. Land use and socioeconomic characteristics of the watersheds
and their settlements were determined by field observation,
interviews, and a review of statistics available at the community
level.
The average slope was determined by the Wentworth (1930) method,
using topographic maps at the 1:50,000 scale. Each watershed was
mapped separately and overlain with a 1-km grid subdivided into four
cells each, with two diagonal cross sections per grid. The slope from
the center toward each corner was determined by the change in
elevation over the four 0.5-km transects. The average slope (%) was
calculated for each small grid cell (0.25 kg ), then for each larger
2
cell (1 km ), and finally for the watershed as a whole.


83
-1 3 -1 -3
Sq (tons sec ) = Q (m sec ) x Cs (tons m )
where Q = river discharge, Cs = concentration of sediment, and Sq =
sediment discharge rate.
Daily sediment discharge for non-flood sampling days was
calculated by averaging morning and evening discharge rates from stage
measurements. This was multiplied by a time conversion factor to
obtain total river discharge for the day. The total discharge times
the concentration approximates total sediment transport past the
sampling point for that day.
For samples drawn during or very close to short-lived peaks, the
flood peak duration was estimated from field records and observations
as well as from reports by the hydrometric station operator and other
nearby residents. River discharge is derived from measurements or
estimates of the flood stage, using the stage-discharge equations
described above. The instantaneous rates of sediment discharge were
calculated by multiplying sediment concentrations times the discharge
at the time of sampling. In cases of time series sampling during
flood events the average discharge rate for each time interval was
converted to a discharge value, then multiplied by the sediment
concentration. The sum of the river discharge and sediment discharge
over the sampling period provided empirical measures of sediment
transport for flood events of a given magnitude.
Analysis of relationships between discharge, sediment
concentration, sediment transport and rainfall. The frequency
distributions of all variables were tested by frequency analysis (SAS,
1979). Based on the results of the preliminary analysis the


AN ECOLOGICAL ANALYSIS OF SOIL AND WATER
CONSERVATION IN HILLSLOPE FARMING SYSTEMS:
PLAN SIERRA, DOMINICAN REPUBLIC
BY
DIANNE E. ROCHELEAU
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1984

Digitized by the Internet Archive
in 2013
http://archive.org/details/ecologicalanalysOOroch

ACKNOWLEDGMENTS
This work could not have been completed without the kindness and
able assistance of family, friends and professional associates too
numerous to mention. Special thanks are due to Dr. Gustavo Antonini
for his confidence in me, his personal concern and his active support
of my work; to Dr. Katherine Ewel for her guidance and encouragement
throughout my interdisciplinary program at the University of Florida;
to Dr. Helen Safa for her moral support, financial and intellectual
contributions during the last year of dissertation work; to Dr. Robert
Marcus and Dr. James Henry for their careful review and suggestions;
to Dr. H. T. Odum for sparking my imagination; and to Dr. Manuel
Paulet for direction and support in Santo Domingo.
Several offices of the University of Florida contributed
financial and logistic support, including the Graduate School, the
Center for Latin American Studies, the Department of Geography, and
the International Programs Office of the Institute of Food and
Agricultural Sciences. The latter covered research expenses in the
field with fund provided by a Title XII grant from USAID. My research
stipend was provided by a traineeship from the OAS for a period of 18
months. The State Secretariat of Agriculture, principally through
Plan Sierra, provided extensive logistic and financial support for the
field research, as well as employment for a period of six months.
My research benefitted substantially from the professional
dedication and warm friendship of many colleagues at Plan Sierra,
11


especially Angel Liriano S., Victor Montero, and Geuris Martinez.
Invaluable data and assistance were also provided by the National
Hydrology Institute of the Dominican Republic (INDRHI), the National
Cartographic Institute, the Dominican Electric Company (CDE), and the
Departments of Meteorology and Land and Water, of the State
Secretariat of Agriculture (SEA).
Final data processing, drafting, and typing tasks were completed
with the able assistance of Nelly Mogallon, Kim Feigenbaum, Beth
Higgs, and Pat French.
The warmest appreciation is reserved for those closest to home,
especially Mickie, Nelly, Gustavo, Marie, Mom and Dad. More than any
other, I owe the successful completion of this work to my husband,
Luis, who helped me tap my own energies and gave selflessly of his
time and effort as computer consultant, data processing technician,
editor, critic, and nurturer.
iii


TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS ii
LIST OF TABLES vi
LIST OF FIGURES ix
ABSTRACT xii
CHAPTER I INTRODUCTION 1
The Problem 1
The Sierra Region 7
Purpose and Scope of Work 11
CHAPTER II A SELECTIVE REVIEW OF STUDIES APPLICABLE TO THE
PROBLEM 15
Overview 15
Erosion and Sedimentation Research in Geomorph
ology, Biogeochemistry and Ecology 15
Models of Erosion and Sedimentation 17
A Review of Relevant Findings in Experimental
Watersheds and Erosion Plots 24
Qualitative and Informal Analyses of Land Use
and Erosion in the Caribbean and Similar Environ
ments 44
The Role of Farming Systems Research in Soil and
Water Conservation 45
Farming Systems, Agroecosystems and Agroforestry
Research 46
Central American Research 51
CHAPTER III METHODOLOGY 58
The General Approach 58
Materials and Methods 60
CHAPTER IV RESULTS AND DISCUSSION 101
Regional Profile 101
Study of Large Watersheds 130
Study of Small Watersheds 164
Erosion Plot and Household Studies 207
IV


Application of Erosion and Runoff Coefficients
to the Small Watersheds 265
Sediment Delivery Ratios 275
CHAPTER V CONCLUSIONS 278
APPENDIX A COMPARATIVE DATA FROM LITERATURE REVIEW 285
APPENDIX B SURVEYED CROSS SECTIONS OF RIVERS AND STREAMS.... 297
APPENDIX C STAGE DISCHARGE CURVES FOR RIVERS, DERIVED BY
INDRHI 302
APPENDIX D MONTHLY RAINFALL FOR STATIONS 3 THROUGH 11 305
APPENDIX E DATA ON SEDIMENT CONCENTRATION, STAGE AND
DISCHARGE FOR MAO AND AMINA RIVERS 315
APPENDIX F DATA ON SEDIMENT CONCENTRATION, STAGE AND
DISCHARGE FOR SMALL WATERSHEDS 343
APPENDIX G SOIL PROFILE DESCRIPTIONS FOR EROSION PLOT SITES. 350
APPENDIX H FORMS USED FOR INFILTRATION TESTS 367
APPENDIX I DERIVATION OF FACTORS FOR USE IN USLE, BY PLOT 369
APPENDIX J DATA FROM EROSION PLOTS 378
LITERATURE CITED 394
BIOGRAPHICAL SKETCH 420
v


LIST OF TABLES
TABLE PAGE
1 Relative location and land cover of erosion plots 65
2 Equations for system model 123
3 Land use systems in the region 126
4 Rainfall and river discharge in the Amina and Mao
Watersheds 144
2
5 R values for regression analyses of subwatershed
rainfall vs. river discharge and sediment load 146
6 Summary of regression analyses of river discharge and
sediment concentration 155
7 Sedimentation transport in Mao and Amina Rivers 158
8 Sedimentation in the Mao River basin estimated from May
1980 measurements 160
9 Water and sediment yields estimated from 1980 and 1981
data 162
10 Characteristics of the small watersheds 167
11 Physical characteristics of erosion plot sites 179
12 Sediment transport in five watersheds 193
13 Discharge rates measured for low flow conditions in small
watersheds 194
14 Flood events yielding peak sediment discharge during the
study period 195
15 Analysis of variance of stream discharge and sediment
transport for all streams 200
16 Maximum recorded concentrations (g L ^) per flood event
for all streams 201
17 Average sediment concentration (gm L ^) per flood event,
for all streams 202
vx


18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
PAGE
Analysis of variance of stream discharge and sediment
transport comparing streams draining coffee stands and
streams draining food crops and pastures 204
Peak discharge ha ^ per flood event for all streams,
results of the a posteriori test of the means 205
Peak sediment discharge rate ha ^ per flood for all
streams, results of the a posteriori test of the means.... 206
Land use, land tenure and production, by household 208
Total annual storm runoff and soil loss rates, by plot.... 219
Relationship of total annual rainfall and storm runoff
in erosion plots 221
Analysis of variance and runoff and sediment losses for
all plots at site Los Montones 229
Runoff and sediment losses for plots at the Los Montones
site: Results of the aposteriori tests of the means 230
Analysis of variance of runoff and sediment losses for
plots grouped by land use at the Los Montones site 232
Runoff and sediment losses for plots at the Los Montones
site, results of the Duncan Multiple Range Test, by
land use 233
Analysis of variance of runoff and sediment losses for
all plots at site Pananao 236
Runoff and sediment losses for plots of the Pananao site,
results of the Duncan Multiple Range Test, by land use.... 237
Analysis of variance of runoff and sediment losses for
plots grouped by land use at site Pananao 238
Runoff and sediment losses for plots of the Pananao site,
results of the Duncan Multiple Range Test, by land use.... 239
Analysis of variance of runoff and sediment losses for
plots in pasture grouped by site 240
Analysis of variance of runoff and sediment losses of
forested plots grouped by site
vii
241


34
PAGE
Analysis of variance of runoff and sediment losses for
plots planted in crops grouped by site 242
35 Runoff and sediment losses for plots in Los Montones and
Pananao sites with crops, results of the Duncan Multiple
Range Test 243
36 Soil loss coefficients by land use and conservation
practice 247
37 Comparison of measured and predicted erosion losses 252
38 Comparison of soil loss coefficients derived from
USLE and from empirical data 255
39 Storm runoff coefficients, total runoff estimates and
runoff/rainfall ratios, by watershed 271
40 Calculation of soil loss coefficients and soil loss
estimates by watershed 272
viii


LIST OF FIGURES
FIGURE PAGE
1 Systems model of land use and erosion in the Caribbean.. 5
2 Impact area: Plan Sierra. 9
3 Model of applied research process 25
4 Flow chart of research activities 26
5 Input-output diagram for interview notations and
monitoring 53
6 Organization of research activities 64
7 Diagram of Thiessen polygons superimposed on a map of
the Mao and Amina watershed subdivisions 77
8 Sampling sites for large watersheds 78
9 Uppsala-type manual sampler for instantaneous measure
ment of sediment concentrations in streams -81
10 A. Illustration of velocity-area method. Person "a"
releases float at time t and person "b" records time
it takes float to move 2 m. B. Cross-section of
stream showing placement of float to measure velocity
and area of three sections of the stream 86
11 Equipment installed in streams to measure sediment
concentrations at different levels of flooding 88
12 Illustration of erosion plot with runoff and sediment
collectors 93
13 Illustration of alternative erosion plot design with
three subsections 95
14 Diagram of sediment and runoff collector indicating the
points at which samples were taken 97
15 Plan Sierra geologic subregions 103
16 Plan Sierra region with study sites 104
IX


PAGE
17 Plan Sierra life zones 106
18 Plan Sierra land use systems 109
19 System model of the Sierra 120
20 Land use model 125
21 Monthly discharge of the Mao River 132
22 Monthly discharge of the Amina River 133
23 Monthly rainfall at the San Jose de Las Matas
climatological station, #1 136
24 Monthly rainfall at the Moncion climatological station,
#2 137
25 Monthly rainfall at the San Jose climatological station,
#1, for two periods 138
26 Monthly rainfall at Moncion climatological station, #2,
for two periods of record 139
27 Plan Sierra region: Mean annual rainfall for 1967-1979. 142
28 Plan Sierra region: Annual rainfall for 1980 143
29 Time series of sediment concentrations during selected
flood events 153
30 Small watersheds in coffee region 165
31 Land use in the Prieto watershed 168
32 Land use in the larger Prieto watershed 169
33 Land use in the Upper Bajamillo watershed 170
34 System model of small watershed in coffee region 172
35 Land use model for watershed model of coffee producing
region 173
36 Land use in Hondo watershed 184
37 Land use in Pananao watershed 185
38 System model of small watershed in pasture-field crop
association 187
x


PAGE
39 Land use model of small watershed model of pastures-
field crops association 188
40 Pananao watershed 190
41 Hondo watershed 191
42 Illustration of equivalent sediment discharge according
to distinct regimes 198
43 Infiltration rates in erosion plots at Carrizal and
Pananao 213
44 Infiltration rates in erosion plots at Los Montones 214
45 Monthly rainfall, storm runoff, and sediment loss at
Carrizal plots 222
46 Monthly rainfall, storm runoff, and sediment loss at
Los Montones plots 223
47 Monthly rainfall, storm runoff, and sediment loss at
Pananao plots 225
48 Model of coffee farms: Large and small holdings 257
49 Model of dairy and cattle farm, Pananao 260
50 Bitter manioc production on a small holder plot, Pananao 261
51 Model of a mixed production on a well integrated farm... 264
52 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Prieto stream 266
53 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Upper Bajamillo stream 267
54 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Greater Bajamillo stream 268
55 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Hondo stream 269
56 Evaluated small watershed submodel of land use, erosion,
and sedimentation: Pananao stream 270
xi


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
AN ECOLOGICAL ANALYSIS OF SOIL AND WATER
CONSERVATION IN HILLSLOPE FARMING SYSTEMS:
PLAN SIERRA, DOMINICAN REPUBLIC
BY
DIANNE E. ROCHELEAU
August 1984
Chairman: G. A. Antonini
Major Department: Geography
The purpose of the study was to develop and test an
interdisciplinary methodology for applied research in soil and water
conservation in hillslope farming systems. The specific objectives
were to collect baseline data on erosion and sedimentation in the
Sierra region of the Dominican Republic and to evaluate soil
conservation technologies and cropping systems recently introduced
into the area.
Erosion, runoff and sedimentation were measured at three scales
2
of analysis: on-farm experimental plots (22 x 2 m ), small watersheds
2 2
(1-30 km ), and large watersheds (300 km ). Erosion losses in the 16
plots ranged from 0.05 to 0.10 tons ha 1 yr 1 under pine forest, 1 to
-1 -1 -1 -1
3 tons ha yr in pasture, 0.5 to 3 tons ha yr in coffee
plantations, and 6 to 70 tons ha ^ yr ^ for plots in mixed food crops.
Annual storm runoff varied from 1% of precipitation under forest to
12% in an eroded continuously cropped plot. Infiltration and soil
profile analyses and erosion measurements at the plots showed a
Xll


pronounced influence of intensity and longevity of cultivation at the
site. Erosion rates also changed dramatically with the phases of the
cropping cycle on the coffee and food crop plots.
The recently introduced slope modification (hillside ditches) did
not significantly reduce erosion rates at the test plots. The minimum
tillage field trial showed more substantial reductions.
Field data were combined with photogrammetric analyses of land
use to compare erosion and sedimentation rates at the watershed level
for production systems based on coffee and pasture with annual crops,
respectively. The contribution to river sedimentation and flooding
did not differ significantly between the two systems at the watershed
level. In both cases it is the association of annual crops with the
dominant commercial land use that determines the erosion rate. The
planting of coffee for soil conservation was ineffective in most cases
because of high labor and related food crop demands.
Suggested alternatives to coffee include reduction of competition
for land and labor between commercial and subsistence production by
substituting tree crops that meet household demands at the local and
national level. Recommendations for land in pasture and annual crops
are the reduction of tillage in annual crop plots, the mixture of more
food crops into annual cash cropping systems, and the combination of
grazing and tree crops in large holdings.
xm


CHAPTER I
INTRODUCTION
The Problem
Almost 10 years after the first dramatic success of the Green
Revolution, the technological breakthroughs remain largely
inaccessible to small farmers of the underdeveloped world. Increased
yields have occurred primarily in large scale commercial or state
enterprises (Harris, 1973; Greenland, 1975; Stevens, 1977). Millions
of small farmers who produce commercial and subsistence food crops and
cash crops for export have maintained or increased production only at
great cost to themselves and to the natural environment (Crosson and
Frederick, 1977; Eckholm, 1976).
These farmers have- expanded into ever more marginal areas (often
arid, semiand, or hillslope environments) or have intensified
cultivation of existing plots, often already located in marginally
productive lands (Brush, 1981). The intensification has been for the
most part without benefit of new technology or capital inputs. It is
achieved through increasingly higher inputs of labor, and often
through practices that damage the long-term fertility and stability of
the soil (Geertz, 1972; Lagemann, 1977) and disrupt hydrologic and
geologic cycles in watersheds (Greenland, 1974; Kellman, 1969;
Pereira, 1973; Rapp, 1977).
Given this situation there is an immediate need for research to
adapt technologies to needs of small farmers within the limits of the
available factors of production and environmental constraints (Crosson
1


2
et al., 1978; Hildebrand, 1981; Lagemann, 1977; Makhijani and Poole,
1975; Novoa and Posner, 1981). The "external costs" of watershed
degradation must be considered and tested within the context of such
research (Erickson, 1974; Novoa and Posner, 1981). Beyond the need
for an inventory of the current magnitude of the problem, there is a
real need to describe and test the complex functional relationships
between land use, production, erosion and sedimentation under field
conditions, in order to explore viable alternatives to the causes of
the problem.
It is widely recognized that the combined processes of erosion
and sedimentation pose a serious threat to sustained production in
both upper and lower portions of tropical watersheds (Farvar and
Milton, 1973). The decreasing depth and fertility of the soil limit
the productivity of small upland farms (Greenland and Lai, 1977;
Morgan, 1979), while quantity, quality, and regularity of surface
water supplies limit both urban and agricultural development in the
lowlands downstream (McPherson, 1974; Nelson, 1973; Odum and Odum,
1976). This is especially critical in tropical and subtropical humid
montane environments subject to intense population pressure by farmers
using traditional and semitraditional methods (Antonini et al., 1975;
Floyd, 1969; Santos, 1981; Sheng and Michaelson, 1973; Wilson, 1976).
While subsistence farmers, landless laborers and their
traditional technology are often blamed for deforestation and
environmental degradation under such circumstances (Rodriguez, 1980),
they are constrained by limited access to land and lack of
alternatives (Plumwood and Routley, 1982). The settlement of rugged
hillslopes and the near total deforestation of upland watersheds


3
represent a choice by default rather than a free choice between
rational alternatives for sustained production. Moreover, there is
intense pressure for continuous cropping and/or establishment of
pastures on cleared land. Large local landowners, as well as urban
and foreign markets, play a major role in this process (Amin, 1977;
Cultural Survival Inc., 1982, Hildyard, 1982; Nations and Komer, 1982;
Plumwood and Routley, 1982).
Small farmers and shifting cultivators in such areas often
practice forms of management that can be sustained well at lower
population densities or within more hospitable environments to which
they have no access (Bailey, 1982; Grainger, 1980; Nations and Komer,
1982). Although these small farmers in marginal areas are referred to
by many as "subsistence farmers," they usually produce some surplus
food crops. In some cases this sector is a major source of staple
food production for domestic markets (Brush, 1981; Novoa and Posner,
1981). The same farm families often function as a seasonal labor
force in coffee, lumber, and other cash crop harvests (Beckford, 1972;
Frucht, 1967). These subsistence farmer/farmworker populations in
marginal lands highly susceptible to erosion form an integral part of
the regional economy and ecosystem. As such, the problem must be
treated as a complex phenomenon that not only affects the larger
downstream and lowland production systems, but is partially
conditioned by them.
The question remains as to how production can be maintained or
increased (at a sustained rate) with minimum damage to both portions
of the watershed. The need for an answer to this question is
particularly urgent in the Caribbean because of high population


4
pressure on hillslope lands (Antonini et al., 1975; Santos, 1981)
coupled with high demand for food crops and relegation of large lowland
tracts to cash crop cultivation (Beckford, 1972; Rankine, 1976; Wilson,
1976).
The resource base in the Caribbean is subject to intensifying
multiple demands by commercial and subsistence sectors for food, cash
crop, wood, fuel and mineral production, as well as protection of the
watershed for downstream development. From a national perspective, the
upland watershed's most important export crop may well be water, needed
for irrigation and hydroelectric projects for downstream development
(Swedforest, 1980). The various types and rates of production demanded
are often competitive in nature, if not mutually exclusive (Crosson and
Frederick, 1977; McPherson, 1974). In many cases the upland regions'
internal situations also clearly indicate the need for change (Brush,
1981; Chaney and Lewis, 1980; Ferreiras, 1979; Hildebrand, 1981; Reiche
and Lee, 1978; Santos, 1981; SEA, 1978).
A Conceptual Model of the Problem
Based upon the information presented above, a model is postulated
that describes the interaction of significant elements in Caribbean
land use systems with regard to the problems of soil erosion and rural
poverty (Fig. 1). The model shows the interaction between population,
land use and the condition of soil, water, and vegetation both within
and outside the study region.
The model diagram follows the format developed by Odum for energy
modelling of ecosystems (Odum, 1971). The tank-shaped symbols (Fig. 1)


OI
Fig. 1. Systems model of land use and erosion in the Caribbean.


6
represent storages of land, soil, biomass, water, and economic assets
within the system. The bullet-shaped symbol indicates a
transformation process (in this case photosynthesis), while the
circles define forcing functions from outside the agroecosystem. The
arrow-shaped devices are workgates, which regulate the interaction
between the various mass and energy flows. Changes in the storages
are determined by the input and output flows, indicated by solid
lines. The imports and exports from the system are indicated
likewise.
In this case the export of soil (soil erosion/sedimentation) is
of special interest, along with changes over time of crops, other
vegetation storages, and-human population. The hypothesis implicitly
stated in the model is that upland land use controls both the export
of soil from the system, and the production of food and income for the
population, both in the uplands and the lowlands.
The problem remains to reconcile internal regional development
goals with national priorities to define desirable changes. At best,
the conflicting needs will be met by careful optimization of land and
water use within entire river basins (Pereira, 1973; White, 1977).
The determination of what is optimal implies a client group within a
defined spatial and temporal context. Decisions must be made to
reconcile differences in "optimal" solutions at local and regional
scales, as well as to adjust short-term and long-term costs and
benefits. Policy makers also need information on the distribution and
nature of costs and benefits among various sectors of the population
in order to formulate "optimal" strategies. Resulting rural
development programs must include short- and long-term incentives at


7
the local level for the widespread adoption of resource management
practices that will benefit the population of the region as a whole
over the long term (Gladwin, 1981; Hildebrand, 1981; Santos, 1981).
Traditional cost-benefit analysis will not suffice since it
excludes environmental as well as social aspects of the system, assumes
a static system, and has been developed for application over relatively
short time periods (Amin, 1977). The problem requires a more holistic
theoretical and methodological approach that will evaluate
environmental and human concerns on their own terms and within the
total system rather than by econometric criteria.
The Sierra Region
The Sierra is a rugged montane area in central Dominican Republic
that has been subjected to the traditional practice of shifting
cultivation, as well as to extensive exploitation of primary resources
such as timber and mineral deposits. It is a relatively underdeveloped
region within an underdeveloped country where the area under production
2
in the country (27,000 km ) already has surpassed the area of land
2
classified as suitable for agriculture (22,000 km ) (OAS, 1967;
Swedforest, 1980; USAID, 1974). Production increases necessary to meet
the national demands for food and income have come from increased
yields in areas already under production, or from expansion into more
marginal areas such as the Sierra. The latter strategy has dominated
among the poor and landless members of the peasantry (Beckford, 1972;
Antonini et al., 1975), and has been a last resort for the former
employees of mining and sawmill camps and furniture shops, most of


8
which had closed by 1979 (Ferreiras, 1979). The major alternative,
emigration to the capital city of Santo Domingo, and to New York City,
provided an outlet for a large segment of the population during the
1960s and early 1970s.
The impact area of Plan Sierra (Fig. 2), an integrated rural
development project within the Sierra, offered a unique opportunity to
examine the apparent conflicts between agricultural development and
natural resource conservation (Santos, 1981). Plan Sierra is a joint
venture between the State Secretariat of Agriculture and the private
sector to initiate and coordinate development efforts in the region in
several sectors: agriculture, livestock, credit associations, health
services, transportation, handicrafts, university programs, and
natural resource management (Quezada, 1977; Antonini and York, 1979;
Plan Sierra, 1979). The Plan's objectives are: 1) to improve the
quality of life of the inhabitants of the Sierra; 2) to manage soil,
water and forest resources; and 3) to promote participation by local
people in the development process (Antonini and York, 1979; Santos,
1981) .
Several hydroelectric and irrigation projects are planned for the
study area (Jorge, 1981). The impoundments will serve the Cibao
Valley downstream. Sedimentation from this area is already a problem
in the Tavera Reservoir, completed in 1977 (Cepeda, 1980). The
magnitude and distribution of erosion in the uplands, however, has
received little attention until recently (Vasquez, 1980).
Plan Sierra has promoted specific farming practices and changes
in land use that are intended to reduce sediment export. A parallel


Fig. 2. Impact area: Plan Sierra.


10
objective of the program is the development of viable sustained yield
agricultural systems. Existing land use in the region consists of a
combination of forest, coffee, bush, small plots of annual crops, and
pastures. The relative rates of erosion and sediment yields from the
prevailing land use systems in this area have been estimated, but not
measured (Antonini et al., 1975). Neither have the effects of erosion
within the upland system been investigated. In the absence of basic
data Plan Sierra had to experiment with tentative conservation
practices and crop changes to curtail erosion and to maintain or
improve production. The widespread adoption of hillside ditches and
various forms of terracing was strongly encouraged by credit
incentives to large and small landowners alike. The conversion of
pasture, bush, and crop land to coffee also was promoted through
credit incentives and supported by extension and education programs.
The research project described in the following chapters sought
to define the magnitude of the sedimentation and erosion problems in
the region, to measure the variation of erosion rates under existing
land use systems, and to test the effectiveness and feasibility of
specific land use changes and conservation practices promoted by Plan
Sierra. The study was integrated into the Plan Sierra soil, water and
conservation program. Logistic support as well as the active
collaboration of paratechnical, technical and professional staff made
it possible to conduct the study at a regional scale and with an
intensity of effort that would otherwise have proven difficult if not
impossible.


11
Purpose and Scope of the Work
The purpose of the study was to develop and test an
interdisciplinary methodology that addresses the needs cited above
within the context of the Caribbean. Such a methodology, and the
theoretical framework within which it is developed, should meet the
following criteria:
1. Treat the problem in a holistic way, building upon existing
research in several disciplines, and incorporating both physical and
cultural aspects of erosion and land use problems;
2. Be flexible enough to include analyses within a broad range
of temporal and spatial scales of observation and to allow for
differentiation by client groups;
3. Facilitate the participation of clients in research and
extension programs;
4. Be applicable in areas with limited data bases;
5. Yield practical short-term results at the topographic scale
while working toward more basic solutions over the long-term;
6. Be amenable to integration into rural development programs,
from preliminary research to subsequent extension efforts.
Objectives
The specific objectives of the study included short-term
practical achievements at the local and regional level. The latter
were readily accomplished within the larger task of methodology
development and testing and contributed to the development of the
theoretical aspects of this study. The objectives, stated in
chronological order of completion, were as follows:


12
1. The study will provide useful information and services for
farmers, rural communities and regional policymakers in the study
area, the Sierra region of the Dominican Republic. This includes the
collection of basic data about the magnitude, distribution, impact and
causes of the problem. (In this case, the basic data will include
knowledge and perceptions of the residents about the study area and
the problems of erosion and sedimentation.) Further, the study will
develop and adapt practical field methods and data analysis techniques
that are suitable for future use by local personnel trained within the
project. The project should become a vehicle for farm extension and
community education in management of natural resources.
2. The study will develop and test a theoretical model of the
relationship between land use, runoff, erosion and sedimentation in
the Sierra. The hypotheses implied in the model will be tested and
the results will contribute to the evaluation and modification of the
model. The refined model will be used as a point of departure for
future simulation and for planning of field trials in the study
region.
3. A general version of the resultant model will be proposed for
application to the problem in the Caribbean. The methodology for
testing the model and applying it to research and extension efforts in
the region will be presented.
The Theoretical Context of the Study
Systems analysis provides a theoretical framework and analytical
tools appropriate to the proposed study and adequate for the complex


13
interdisciplinary nature of the problem. It allows the researcher to
focus on the interactions and mutual causality so difficult to include
in other analyses and so critical to understanding the links between
land use, poverty, and erosion. The relationship of the upland areas
to the larger watershed and the national context is also readily
included within this perspective by using the concept of nested
hierarchies in systems of various scales (Antonini et al., 1975; Odum
and Odum, 1976). This feature of systems analysis also allows the
inclusion of boundary conditions on subsystems that may be determined
by the larger system or some elements thereof. As such, the
"structural determinants" of land use and production relationships
(Beckford, 1972) can be considered within what has often been viewed
by structural determinists as a noncompatible methodology. The
theoretical assumptions underlying this approach are in fact
consistent with the world-system paradigm in anthropology (Nash,
1981).
Moreover, within the broad framework of systems analysis it is
possible to combine elements of such fruitful and diverse approaches
as the ecosystems-energy analyses developed by Odum and Odum (1976)
and the farming systems research and extension methodology developed
by Hildebrand (1981). Other related lines of research which can be
incorporated include the basic watershed-ecosystems analysis (Hart,
1980), agroecosystems testing and development conducted by Ewel
(1981), farming systems research conducted by Ruthenberg (1976), FIinn
(1980), and Lagemann (1977), and regional characterization of tropical
agroecosystems (Posner et al., 1981).


14
Systems analysis offers the potential for experimenting with a
variety of "futures possible" through simulation of dynamic nonlinear
models (Ewel et al., 1975; Forrester, 1971; Lagemann, 1977; Meadows et
al., 1972; Odum, 1971). Such analyses allow us to "optimize" over
various time scales and according to different criteria. Ideally,
this provides planners, policymakers, and their clients with
information about the probable outcome of different resource
management alternatives; the net result can be viewed from the
perspective of the farm household to the national level.
The accuracy and quality of the conclusions drawn from the
qualitative or quantitative analysis of such models depend upon the
accurate conceptual design of the model and the quality of the
information used to evaluate its components (Amadeo and Golledge,
1975; Harvey, 1969; Kuhn, 1962). The model, as a selective
simplification of reality, must be constructed by one who is famil-iar
enough with both the system-at-large, and the problem of interest, to
choose which elements to include, and to describe their relationships
accurately.
A researcher trained in ecosystem analysis and modelling can
perform this combined function by modelling the entire system and by
adding questions of ecological concern and ecological monitoring
methods to interdisciplinary adaptive research in farming systems and
resource management. A general reconnaissance survey of the region,
and interaction with researchers from complementary disciplinary
backgrounds, as well as with project clients, can provide the kind of
information and broad perspective needed to properly define the
system.


CHAPTER II
A SELECTIVE REVIEW OF STUDIES APPLICABLE TO THE PROBLEM
Overview
Various aspects of the relationship between erosion,
sedimentation and land use have been treated by geographers as well as
by researchers in several other fields. Although some geographers
(Antonini et al., 1975; Haggett, 1961; Kellman, 1969; More, 1967;
Morgan, 1979; Pereira, 1973) and other scientists (Geertz, 1972;
Greenland, 1974, 1975; King, 1978; Lagemann, 1977) have spanned both
ends of the research spectrum, the literature is generally divided
along these lines and will be reviewed as such.
Erosion and Sedimentation Research in Geomorphology,
Biogeochemistry and Ecology
The theoretical aspects of erosion and sedimentation mechanics at
the micro-scale are fairly well established and understood and already
have been summarized for general reference use by applied scientists
(Brady, 1974; Chow, 1964; Gregory and Walling, 1973; Morgan, 1979;
Toy, 1977; USDA, 1979). What are less clear, and still require a wide
range of basic and applied, theoretical and empirical research, are
the more complex causes and effects of erosion and sedimentation at a
larger scale, under field conditions.
The study of erosion and sedimentation and the mechanisms that
determine their occurrence at the meso- and regional scales under
15


16
field conditions is a proper subfield of physical geography since
geomorphology is a traditional and well-developed avenue of inquiry
within the discipline (Gregory and Walling, 1973; James, 1972; Keller,
1968; Morgan, 1979; Stoddart, 1965; Strahler, 1964). Systems theory
has been applied widely in studies of watersheds and other
geomorphological units by British geographers (Chisholm, 1967;
Chorley, 1962; Kirkby, 1978; Ollier, 1968) as well as by numerous
other geomorphologists (Leopold et al., 1964; Leopold and Langbein,
1962; Toy, 1977) and hydrologists (Chow, 1964; Vemuri and Vemuri,
1970). This general approach offers the advantage of integrating form
and process, by accounting for their interactive relationship
(Chorley, 1962, 1969).
The open systems approach allows the inclusion of the
quantitative hydrology/geomorphology tradition that dates from Horton
(1935, 1938) and continues in the work of Strahler (1964) and other
physical geographers. The use of systems theory in geomorphology also
facilitates the study of the human use of the earth, a longstanding
focus of geographic research (Harvey, 1969; James, 1972; Marsh, 1964;
Thomas, 1974) that could lend itself well to quantification within a
systems framework (Stoddart, 1965).
The application of systems theory to the study of watersheds has
also been tested and developed within ecology in recent studies of
biogeochemistry (Bormann and Likens, 1979; Likens et al., 1977) and
resource management (Cooper, 1971; Hall and Day, 1977; Hopkinson and
Day, 1980; Patton, 1971; Thomas, 1974; Van Dyne, 1969). In such
studies the watershed defines the boundaries of the ecosystem and the


17
interacting elements whose mutual influences are observed within the
system include physical, chemical, biotic and cultural entities (Odum,
1971, 1982; Odum and Odum, 1976).
The general conceptual framework of systems analysis is well
suited to the study of mutual influence between land type, land use,
plant and animal productivity, and erosion in the upland watersheds of
the Caribbean. Incorporating the broad perspective of the man/land
tradition in geography, the research in open systems geomorphology and
ecosystem analysis also offers valuable methodological examples for
application to interdisciplinary research on the topic.
Models of Erosion and Sedimentation
The range of erosion and sedimentation models in use includes
stochastic and deterministic models, statistical as well as
"parametric" models, and combinations of all of the above. These
models have been developed and applied within scales of analysis
ranging from individual plots to large watersheds.
One school of research focuses on modelling processes and
interactions between the various parts of watersheds as complex
systems (Chorley, 1962; Likens et al., 1977), and the other major
thrust of erosion and sedimentation studies has been to develop
empirical predictive equations (Wischmeier, 1975; Wischmeier and
Smith, 1978). The latter relate land use and management to erosion
loss and sediment yield and have been developed for management
purposes, primarily for use in soil and water conservation programs.


18
Empirical Models
The best known and most widely used of the empirical models of
erosion loss is the Universal Soil Loss Equation (USLE), an empirical
formula for estimating potential soil loss by sheet and rill erosion
for individual plots (Wischmeier, 1975; Wischmeier and Smith, 1978).
The USLE is used primarily to predict erosion losses based on a
combination of inherent site characteristics and variables subject to
human intervention and management. Tons of soil lost is the dependent
variable. It is determined by a combination of six independent
variables as follows:
A = RKLSCP, where
A = soil loss per unit area (tons acre ^ or tons (m) ha ^) over
a given time
R = erosive potential of rainfall (based on total energy and
intensity of rainfall)
K = an index of soil erodibility (a measure of soil suscepti
bility to erosion based on physical properties)
L = length-of-slope factor
S = inclination-of-slope factor
C = vegetative cover factor, integrating type of cover and
management of crops or other vegetation
P = conservation practice, including structural alteration of
site and/or contour planting.
The apparent disadvantages of this model reside mainly in the
costly and time consuming local calibration required for its proper
application. This is particularly important in underdeveloped


19
countries where research funds and personnel are limited and the
complexity of the rural agricultural landscape in the uplands requires
extensive calibration of the model. Rapid changes in farming
practice, crop types and level of technology in many areas would
require almost a continuous update of the calibration experiments.
Such a program of research not only represents a large investment in
and of itself, it implies a diversion of resources from alternative
avenues of theoretical and applied research directed toward cumulative
growth of knowledge about the processes in question.
Another weakness of the USLE is the failure of the conceptual
framework to account for the difference between land cover and land
use systems. There are no economic or social aspects to the model,-,
yet it is proposed as a practical tool for farm and regional
conservation planning purposes. The premises under which the equation
is applied often can be misleading, resulting in serious errors in
planning and management decisions. Aside from the logical pitfalls
inherent in the use of the model, there is the technical drawback of
its inability to predict the feedback effects of current erosion rates
on future land use and productivity which in turn affect future
erosion rates.
The USLE can be a useful tool for prediction of erosion rates
under specific known conditions, given prior calibration for the full
range of conditions in a region. Used alone, however, it does not
constitute an adequate basis for management decisions at the farm
level, much less for regional planning purposes.
The USLE has also been adapted for prediction of sediment yields
at the watershed scale (McElroy et al.,
1976; Onstad and Foster, 1975;


20
USDA Forest Service, 1980; Williams, 1975) by estimating the sediment
delivery ratio based on drainage area (Hoitan and Lopez, 1971; Roehl,
1962). The models derived from the USLE have been used widely in
economic and land use planning studies for evaluation of specific
cropping systems and/or conservation practices (Kling and Olson,
1975). The weaknesses of the approach include the inherent
limitations of the OSLE as well as the questionable realism of
sediment routing techniques (Skopp and Daniel, 1978).
One empirical model that does not derive from the USLE is used by
the United States Bureau of Reclamation (Flaxman, 1975; Skopp and
Daniel, 1978; Strand, 1975) as well as by international research
organizations (Rapp, 1977). The technique combines flow duration
curves with sediment concentration, the latter derived from either a
sediment rating curve or a power function. Both flow duration and
sediment rating curves describe empirical relationships that must be
determined on site.
The model is useful for prediction of sediment yield over the
long term, given a continuation of current land use conditions, but is
less amenable to integrated watershed management based on land use and
land treatment programs. The flow duration/rating curve model,
however, could be calibrated to particular watersheds or groups of
watersheds for specific land uses and treatments, as has been the case
in paired watershed studies conducted in experimental catchments.
A number of sophisticated digital computer models of runoff
and/or erosion and sedimentation predict water and sediment discharge
by relating hydrologic and physical characteristics of the source


21
areas (Chapman and Dunin, 1975; Mein, 1977; Fleming, 1968; Negev,
1967; Ebumive and Todd, 1976; Donigan and Crawford, 1976).
Theoretical Models of Physical Processes
Research conducted by Elwell (1979a) in Zimbabwe resulted in a
simple model similar to the USLE in some aspects, but based on
rational rather than empirical parametfers for estimating sheet erosion
from arable land. Rainfall energy measured in 10-day increments
defines erosivity while the protective value of crops and cropping
practices is assessed according to the percentage of seasonal rainfall
energy "i" intercepted by the vegetative canopy and ground cover.
Potential interception is determined in a fashion similar to the leaf
area index measured by ecologists (Odum, 1971).
Application of the model by Elwell (1979a) in Zimbabwe field
trials confirmed the importance of mulches for erosion control and
demonstrated the potential for reducing soil loss by increasing crop
yields. Moreover, important seasonal relationships were identified
between various protective crop covers and erosivity of rainfall in
given areas. This allowed the identification of the crops or crop
combinations that best protected the soil during periods of intense
rainfall.
The model is particularly applicable to the seasonally dry
tropics where single storm events cause a large proportion of the
total annual erosion loss. The practical application of the model is
aided by a description of the techniques for calculation and field
measurement (Elwell, 1979a).


22
Many of the theoretical models describing erosion and sediment
transport processes are mechanistic approximations of reality
(Bennett, 1974). Most of these constructs take systems theory as a
point of departure, although either combined or dispersed systems may
be postulated. The combined system models assume a uniform watershed,
with average physical features and composite land use, and do not
consider distance between source and outlet. In dispersed systems the
runoff and sediments are generated in spatially distinct source areas
and are then routed spatially and/or chronologically through
successive increments of the watershed (Holtan and Lopez, 1971;
Fleming and Leytham, 1976; Frere, 1978; Renard and Lane, 1975; Skopp
and Daniel, 1978).
The models within the theoretical research tradition generally
offer better conceptual realism than the empirical models (Holtan and
Lopez, 1971) but the theoretical models tend to be much stronger in
the explanation of physical and chemical processes and remain largely
underdeveloped in the treatment of biological and cultural phenomena.
These models are just now beginning to be tested in systems that
include human populations, and management options are being added to
the range of variables to be tested. Meanwhile, the vast majority of
field data collected on erosion, sedimentation and land use has come
from empirical studies designed to calibrate predictive equations that
are, essentially, site-specific black box models of complex systems
and processes (Boughton, 1967; Hayward, 1967; Holtan and Lopez, 1971;
Skopp and Daniels, 1978).


23
Ecosystem Models of Watershed Processes and Land Use
Two types of ecosystem models that are less frequently applied to
erosion and sedimentation studies warrant special consideration for
their broad applicability and holistic approach. The first is a model
developed to analyze the biogeochemistry of a forest ecosystem at the
Hubbard Brook Experimental Forest in New Hampshire (Likens et al.,
1977). The second is an energy model used to simulate the complex
interaction of population, land use and sedimentation in the Central
Mountains (Cordillera Central) of the Dominican Republic (Antonini et
al., 1975).
The ecosystem analysis studies conducted at Hubbard Brook focus
on quantification of nutrient budgets through monitoring of
meteorological inputs and geologic outputs of nutrients in small
watersheds. The model used for the study places a strong emphasis on
precipitation, runoff, and solute and sediment export from the areas
of interest (Likens et al., 1977). The variables are measured first
under primary undisturbed forest cover, and later under disturbed
conditions. The final product of this analysis is an annual
hydrologic and nutrient budget that includes sediment yields.
Another form of ecosystem analysis is based on the evaluation of
energy pathways within communities of plants and animals, and on the
relationship between those pathways and the physical environment
(Antonini et al., 1975; Odum, 1971). Recent studies using the energy
approach have extended this concept to include the complex transfers
of matter, energy and information in ecosystems that include human
populations. Diverse interactions, ranging from mineral cycling,

i

24
photosynthesis, runoff and sedimentation to economic transactions, are
interrelated by the common denominator of energy flow (Antonini et
al., 1975; Odum, 1971; Rappaport, 1971).
Models using the energy flow language developed by Odum have been
applied to analyses of watershed ecosystems in several environments.
The most pertinent case is a model of the interaction between land
use, erosion and sedimentation in the rugged uplands of the Dominican
Republic (Antonini et al., 1975). As illustrated in the diagrams
(Figs. 3 and 4), the models are dynamic non-linear systems models that
can be simulated on an analog or a digital computer by simultaneous
solution of differential equations. The equations describe the change
in the landscape variables over time. The model indicates mutual
causality between variations in land use and population but does not
account for similar relationships between erosion and land use, since
erosion is not considered as a separate process. The model simulation
demonstrates the implications of current and alternative trends in
land use for future conditions of population, land use and reservoir
sedimentation. The two-part model developed in this study provides
the conceptual point of departure for construction of a single energy
flow model of hillslope land use systems, upland erosion and
downstream sedimentation in the Caribbean.
A Review of Relevant Findings in Experimental Watersheds
and Erosion Plots
During the past 10 years the international scientific community
and policymakers at various levels have focused greater attention on
the problems of erosion and sedimentation, particularly in the tropics


Fig. 3
Model of applied research process (after Harvey
1969).


Fig. 4. Flow chart of research activities


27
(Brown, 1980; Henkes, 1982). However, a large proportion of the
published literature on the topic refers to studies conducted in the
United States. While the findings and methods are not all directly
applicable to the Caribbean, they supplement the less extensive data
base and cumulative research experience available at present from
Third World and Caribbean sources.
Summary of Recent Research in the United States
Three landmark studies in experimental watersheds have set the
methodological and technical trends for studying the impact of land
use on sediment yields in small watersheds. The Coshocton, Ohio,
watershed studies (Harrold et al., 1962; Mustoneu and McGuiness, 1968)
focused primarily on sedimentation rates under various cropping and
land treatment conditions in agricultural areas. Research conducted
at Coweeta, North Carolina, and Hubbard Brook, New Hampshire,
addressed both methodological and theoretical aspects of watershed
biogeochemistry (including land treatment) in forested ecosystems
(Bormann and Likens, 1979; Douglass and Swank, 1975; Likens et al.,
1977). These experiments served as models for a series of applied
watershed studies recently initiated in forests, agricultural land,
and rangeland.
The more recent research has been conducted primarily under the
auspices of federal environmental legislation that mandates
documentation of non-point sources of pollution, including sediment
discharge (Haith and Dougherty, 1976; USDA Forest Service, 1980;
Jewell and Smith, 1976; Rao, 1980; Reikerk et al., 1978). This


28
research focuses more on entire watersheds and less on individual
plots, in contrast with prior work conducted by the Soil Conservation
Service at the farm level (Ackerman, 1966). Most of the watershed
experiments combine monitoring of precipitation with recording rain
gauges, continuous monitoring of stream discharge at weirs, sampling
of sediment discharge at weirs, and/or collection of sediments in weir
2
ponds in watersheds with areas less than 20 km (Hewlett et al., 1969;
Ward, 1971).
A ma]or topic of the earlier studies was the role of the
undisturbed forest in regulating the hydrologic cycle and sediment
export (Douglass and Swank, 1975; Helvey, 1967; USDA Forest Service,
1980). Among the more important findings were: the importance of
litter versus canopy in protecting the soil against the erosion
potential of rainfall (Table A-l); the impact of forest vegetation on
stream discharge (Dils, 1957; Johnson and Swank, 1973) (Table A-2);
and the association between undisturbed forest cover and low sediment
concentration in streams (Table A-3).
Streamflow, sediment concentrations and mass transport from
forested watersheds showed dramatic changes after harvesting, various
site clearing and management operations, or conversion to other uses.
Several studies reported heavy increases in suspended sediment and
nitrate concentrations after clearcutting (Bormann and Likens, 1979;
Douglass and Swank, 1975; Hewlett and Nutter, 1969; Likens et al.,
1977; Monk, 1976).
Water yield increments proportional to percent area in cleared
openings were reported for several gauged watersheds (Likens et al.,


29
1977; Sopper, 1975). This was attributed to increased storm flow as
well as to reduced water consumption by evapotranspiration. Reported
streamflow increases in Georgia, South Carolina, and Oregon ranged
from 40 to 50% (USDA Forest Service, 1980).
The combination of increased sediment concentration and higher
streamflow resulted in dramatic increases in sediment transport after
clearing (Monk, 1976; Sopper, 1975). By contrast, little or no change
was reported in a patch-cut watershed in the Oregon Coast Range, while
much higher figures are reported elsewhere with use of conventional
treatments (Brown, 1982; Monk, 1976).
Results from east Texas forested watersheds identify sediment as
the major pollutant in streamflow. At the same sites, in a study of
harvest practice and related impact on water quality and mass
transport, the highest sediment transport rates occurred in
association with harvesting along streams. The design, construction
and maintenance of roads often are cited as major determinants of
water quality in forested watersheds (Fredriksen, 1970; Pavoni, 1977;
Texas A&M, 1979; Ursic, 1978).
Water yield and mass transport of minerals and sediments from
nonforested watersheds differ markedly from forested sites. A study
of nutrient yields from various categories of land use in the
watersheds of 24 Connecticut lakes estimates contributions of
phosphorus from agricultural and residential-commercial land at 200
and 1100%, respectively, of forest contributions (Hill and Frink,
1978). National averages of erosion rates for various categories of
land use in the United States (Table A-4) agree well with the results


30
from the Connecticut study. Transport rates from harvested forests
are extremely high, even in comparison to agricultural uses. The high
rates, however, are offset by the fact that forest harvests are
periodic events that only occur once every 15 to 40 years, even in the
fast-growing pine plantations of the Southeast and the Caribbean. By
contrast, many agricultural uses are sustained continuously on a given
parcel of land.
Recent studies in agricultural watersheds in the United States
concentrate on cropping systems and practices in large scale
commercial farms or ranches. Pollution of surface waters by chemical
fertilizers (Haith and Dougherty, 1976) and pesticides often
overshadows sediment pollution as a subject of public concern (USDA
Soil Conservation Service, 1980). Pathogens entering the waterways
from feedlots and grazing lands (Jewell and Smith, 1976) also attract
more attention, although sediment is the major pollutant, by volume,
discharged into surface waters from agricultural lands. Suspended
sediments in streams and rivers carry pathogens as well as chemical
pollutants. Much of the current research, however, emphasizes the
chemical by-products discharged into waterways in solution from
agricultural non-point sources (Rao, 1980).
Studies of erosion in individual plots offer more information on
the variation in erosion rates with changes in cropping systems, farm
management and conservation practices. Data analyses by Wischmeier
and Smith (1978) for cropped and clean-tilled plots corroborate the
conclusions of studies in forest ecosystems. The canopy cover and
ground cover on the site determine how much rainfall energy reaches


31
the soil surface. As in the forest, both canopy and mulch (comparable
to litter) intercept raindrops, but mulch does this so close to the
surface that the drops regain no fall velocity. Mulches also obstruct
runoff flow and reduce its velocity and sediment transport capacity
(Wischmeier and Smith, 1978). Surface roughness of the soil also
influences the velocity and transport capacity of runoff. Thus,
tillage practices, crop yields and crop rotations, as well as above-
and below-ground architectural characteristics of particular plants,
influence the degree to which given cropping systems reduce erosion
relative to a standard clean-tilled plot (Mannering and Meyer, 1963;
Meyer and Mannering, 1961).
Annual row crops vary widely but erosion rates generally range
between 5 and 50% of those measured in clean-tilled plots. For
example, a field tilled with chisel and disk plows, rotated from wheat
to meadow to corn, with one crop annually, loses an average of 9% of
the total loss for a control plot (Wischmeier and Smith, 1978).
Pasture, rangeland, bush and woodland reduce erosion to between 1
and 10% of clean-till soil losses, and undisturbed forest further
reduces the loss to between 1 and 0.01% of the bare fallow. Losses
under harvested, mechanically prepared woodland sites, however, vary
between 10 and 90% of the clean-till figures. Average annual erosion
losses for cropland in the United States bear out the trends reported
by Wischmeier and Smith (1978) (Table A-5).
The same results have been extrapolated to the watershed level
with the use of sediment delivery ratios (USDA Soil Conservation
Service, 1980; Wischmeier and Smith, 1978). The proportion of total


32
eroded soil that arrives at a given outlet ranges from 33% in a 1 km
2 2
area to 18% for a 25 km area to 10% for a 250 km drainage area
(Roehl, 1962).
Studies in watersheds and erosion plots in the United States
cannot be extrapolated directly to tropical and subtropical montane
watersheds. Beyond the difference in the ecosystems themselves there
is the even greater divergence in level of technology and the greater
complexity of land cover associations within upland land use systems
of the tropics. The general relationships established between land
cover, erosion, and sedimentation must be tested further and compared
with results from the Caribbean and similar regions.
Erosion and Sedimentation Research in the Caribbean and Similar
Environments
Although few in number, studies of erosion, sedimentation and
land use have been conducted in the Caribbean in Jamaica, Puerto Rico,
Trinidad-Tobago and the Dominican Republic. Information also can be
found from study areas in New Zealand, Australia, Yugoslavia, Kenya,
Tanzania, Zimbabwe, Malaysia, the Philippines, Costa Rica, Guatemala,
and Colombia. These span a wide spectrum of environmental and
economic conditions, but each has in common some combination of
topographic, climatic, cultural and cropping system characteristics
with the upland forest and farming areas of the Caribbean. The
methods and materials used for watershed monitoring and other data
collecting activities offer proven alternatives to the more
sophisticated, expensive experiments in the United States and other
developed nations (Pereira, 1973).


33
Watershed and sedimentation studies
Erosion, sedimentation and land use research in tropical
hillslope lands falls into two major categories. The watershed
studies focus primarily on the interaction of climate, topography, and
land use throughout the drainage area in determining river regimes and
erosion and sedimentation rates. By contrast, reservoir sedimentation
studies focus on the identification of sediment source areas as well
as on the immediate protection of reservoir facilities, which often
presupposes an emphasis on the development of physical infrastructure
at various points throughout the watershed.
Representative basins and experimental catchments. Studies
initiated under the auspices of the International Geophysical Year
(IGY) and the International Hydrological Decade account for a large
proportion of the work conducted in the tropics. The representative
basin studies emphasize comparative description of diverse river
schemes and watershed ecosystems on a global basis, while experimental
catchment studies focus more attention on the effects of alternative
land cover and land treatment in a given area.
Among the more important findings to date are the contrasting
characteristics of tropical climate and hydrology when compared to the
more temperate regions. Water balance data, including ratios of
runoff and evapotranspiration to total precipitation, are available in
reports from empirical studies (Golley, 1972; Holdridge, 1967, 1982;
Odum 1970b; Pereira, 1973; Thornthwaite and Mather, 1959). The
seasonal distribution as well as the amount of precipitation varies
substantially from the pattern of temperate areas. Bimodal rainfall


34
and river discharge distributions are common. The proportion of total
precipitation that leaves the watershed as evapotranspiration is
higher than in temperate areas and the ratio of surface water
discharge to total precipitation generally is lower, except when
deforestation occurs. The overall amount as well as the intensity of
rainfall usually is higher, making the erosive potential greater than
in most areas of temperate or cooler climates. High erosive potential
of climate often coincides with high erodibility of soils in the
seasonally dry tropics (Elwell, 1979a). The international comparative
statistics on river sedimentation bear out the implications of high
erosivity combined with readily eroded soils and high population
densities in such areas (Douglas, 1968; Holeman, 1968; Stoddard,
1965). Measurements of sediment yield in several catchments in
Malaysia demonstrate the relationship of land use to sediment
transport (Table A-6).
In addition to providing baseline information, the representative
basin and experimental catchment research tests various methods,
materials and modelling strategies. Experimental catchment studies in
New Zealand (Campbell, 1962; New Zealand Ministry of Works, 1968a,
1968b, 1968c, 1970) and Australia (Australian Water Resources Council,
1969; Chapman and Dunin, 1975) emphasize methodology and techniques,
from mapping of sediment sources (Mosley, 1980) and evaluation of
suspended sediment data (Campbell, 1962) to calibration of catchment
models (Mein, 1977; Wood and Sutherland, 1970). Multiple catchment
experiments in New Zealand (New Zealand Ministry of Works, 1970)
provide examples for measuring the impact of farming, forestry, and
range management practices on sediment yield in hilly terrain.


35
Reports from Kenya, Tanzania, and Uganda (Blackie, 1972; Pereira,
1973; Pereira et al., 1962, 1967; Rapp, 1977) present useful data for
comparison with some of the land use systems of the Caribbean,
including coffee and other cash crop plantations, grazing, subsistence
farming and forestry. Paired watershed studies spanning four years or
more demonstrated the impact of both land cover and specific
management practices on runoff and erosion as well as on the harvest
within the watershed. In all of the cases cited, researchers
collected frequent stream discharge and precipitation measurements.
Some cases also include continuous monitoring of the above, as well as
sampling of suspended sediments in streamflow. The results of grazing
and range improvement trials in experimental watersheds in Uganda
include a more than twofold increase in depth of penetration of
rainfall into the soil, and a concurrent reduction in peak streamflows
after restoration of overgrazed grasslands (Pereira et al., 1962).
Data from experimental sites in Kenya (Blackie, 1972; Pereira, 1973)
document the effects of replacing tall evergreen forest with tea
plantations. The mean water yields over an 11-year period
effectively were equal for the forested control watershed and an
adjacent area planted in tea. The floods resulting from peak storm
events, however, varied substantially. The minimal impact of the tea
plantation reflects in part the stringent conservation measures
employed during its establishment. By contrast, clearing of
indigenous bamboo forest without immediate replacement by tree crops
increased streamflow 16% (Blackie, 1972). In Tanzania, a cleared
forest planted to a maize and vegetable single-crop system yielded a


36
50% increase in runoff, measured as streamflow (Edwards and Blackie,
1975; Edwards, 1977).
Findings from two other study areas in Tanzania, one a cultivated
montane catchment and the other a series of catchments in semi-arid
savanna, offer comparative data on river regimes as well as on
suspended sediment concentrations under changing land use practices
(Rapp, 1977). The catchments represent a complex mosaic of forest,
farm, pasture and bush, which is comparable to many upland catchments
in the Caribbean. During flood peaks, sediment concentrations ranged
from 2000 to 3500 mg L ^ in the upland areas, and from 15,000 to
75,000 mg L ^ in the semi-arid catchments. Flash floods and high
sediment loads in both the montane and savanna areas were attributed
to land use. A comparison of the relationship between drainage area
and sediment delivery ratio in the United States and Tanzania shows
much less reduction in sediment yield with increased drainage area in
the Tanzanian catchments.
Total streamflow and sediment yield were determined by the use of
flow duration and sediment rating curves along with stream gauge and
sediment concentration data. Sampling was carried out with a home
made point-integrating hand operated sampler (Nilsson, 1969; Rapp,
1977) and an automatic multi-stage sampler designed for ephemeral
streams. Both the instruments and the methods of analysis used in
this study are feasible for use in the Caribbean.
Reservoir sedimentation. Studies of reservoir sedimentation and
other aspects of regional water utilization and management often
approach the situation as an engineering and economic problem, either


37
in terms of reservoir design or in subsequent maintenance of the
completed structure and the watershed. Many studies of this type have
been carried out in Latin America (Casco de Aviles, 1979; Crosson and
Frederic^, 1977; Farvar and Milton, 1973; Rabinovitch, 1979; United
Nations, 1979) including several studies in the Caribbean (CDE, 1981;
IBRD, 1972; Lahmeyer, 1967; McHenry and Hawks, 1966; Noll, 1953).
Most of these works analyze the feasibility of proposed
impoundments or document sedimentation rates in existing structures.
The major hydroelectric and irrigation projects planned or under
construction in the Caribbean lack empirical data on sediment delivery
by watershed subdivisions. Measures of erosion rates within the
watershed also are seldom included. Soil conservation programs, when
they exist, usually are initiated in response to reservoir
sedimentation problems (Floyd, 1969; Gomez, 1980; Paulet, 1980;
Rocheleau, 1980; Santop, 1981; SEA, 1978; Vasquez, 1980).
Several reservoirs are in danger of filling up in half the time
projected for the useful life of the structure (CDE, 1981; McHenry and
Hawks, 1966; Noll, 1953; INDRHI, 1981), and some reservoirs already
require expensive dredging procedures to continue functioning (CDE,
1981) Serious erosion problems are related to land use in the upper
watersheds as well as in the immediate vicinity of impoundments
(Antonini et al., 1975; Carmona, 1980; Cepeda, 1980; CDE, 1981; de
Leon, 1980).
Erosion rates at the farm and plot level
While experimental watershed research has concentrated primarily
on undisturbed forested areas, the impact of deforestation and


38
resultant reservoir sedimentation, erosion studies in diverse
hillslope environments throughout the world have documented soil
losses at the scale of individual farms or small plots. These results
are available for varying crop types, rotation and conservation
practices as well as for a wide range of natural environmental
conditions.
Research on erosion rates in the Caribbean and similar regions
consists primarily of experiments in standard runoff and erosion plots
under various land covers and management practices. Methods tested
and adapted in similar environments offer alternatives to the more
expensive and time consuming instrumentation often applied. An
adaptation of the standard erosion plot with sediment and runoff
collectors tested in hillslope experimental fields in Yugoslavia
provides a simple design for applications in other studies (Djorovic,
1977). Dunne (1977) describes simple erosion plot designs and
summarizes several inexpensive techniques for erosion measurement
without the use of standard plot structures. The Gerlach trough
(Gerlach, 1967) is easy to install and to use for soil loss and runoff
measurements (Morgan, 1979). It has been employed successfully in New
Zealand (Soons and Rainer, 1968), the Philippines (Kellman, 1969),
Israel (Yair, 1972), the United Kingdom (Morgan, 1977), and the
Carpathian Mountains (Gerlach, 1967).
Results from erosion studies in the Caribbean and Latin America.
Results from the Caribbean and neighboring Latin American nations
consistently show that erosion rates vary significantly (up to three
orders of magnitude) with changes in vegetation cover and management


39
practice (Table A-7) (Ahmad and Breckner, 1973; Barnett et al., 1972;
Bertoni, 1966; Marques et al., 1961; Noll, 1953; Rocha, 1977; Sheng,
1973; Sheng and Michaelson, 1973; Smith and Abruna, 1955; Suarez de
Castro, 1952; Suarez de Castro and Rodriguez, 1955, 1962; Vincente-
Chandler, 1976; Uribe, 1966). Clean-tilled fallow consistently
yielded soil losses in the range of 100 to 200 tons ha ^ yr \ while
land in annual crops lost approximately 20% of that amount, and
pasture land lost 10% of the cropland losses and about 2% of the
losses sustained in bare fallow plots. Undisturbed forest loses 500
to 1000 times less than the clean-till control plots (Sheng and
Michaelson, 1973). Results from Colombian coffee plantations indicate
fairly low rates of erosion, varying between the ranges common for
forest and field crops, depending on the age of the stand, the method
of establishment, and management practices (Suarez de Castro and
Rodriguez, 1955, 1962). The relative^ importance of farming practice
is also illustrated by the five- to ten-fold decreases in erosion
reported for various conservation practices tested in hillslope food
crop plots in Jamaica (Sheng and Michaelson, 1973).
Research on erosion potential in the Caribbean is based upon the
factors in the USLE. Studies conducted in Puerto Rico and the
Dominican Republic have defined the erosive potential of rainfall (R)
(Paulet, 1978), the erodibility of soils (K) (Barnett et al., 1972;
Bonnett and Lugo-Lopez, 1950; Lugo-Lopez, 1969) and the cropping
factor (C) (Santana, 1980) for parts of the region, but the
applicability of the USLE to these areas is questionable (Barnett et
al., 1972). More field data collection is needed on erosion and


40
runoff and their relationships to land use within a variety of land
use and ecosystem types throughout the Caribbean.
Results of erosion studies in similar environments. Basic data
on erosion rates are scarce in Latin America and the Caribbean, in
comparison with the humid tropics of Asia and Africa (Lai, 1977b). A
brief summary of selected erosion studies in these regions provides a
broader frame of reference for work already completed or in progress
in the Caribbean.
Many of the crops, the small farm technology and some elements of
the natural ecosystems of West Africa bear a strong resemblance to
parts of the Caribbean. Reports of experiments conducted by Lai et
al., (1979) and others at the International Institute for Tropical
Agriculture (IITA) in Ibadan, Nigeria, indicate a clear relationship
between vegetation cover and erosion rates, with results approximating
those from the Caribbean. Losses from clean-tilled fallow range from
-1 -1 -1 -1
11 tons ha yr at 1% slope to 230 tons ha yr at 15% slope. On
the average, soil erosion varies much more with slope than does runoff
(Lai, 1977a).
Experiments in Senegal (Charreau, 1972) on gentler slopes in the
savanna demonstrate a similar range of soil loss as from the medium
slopes (10%) studied by Lai (1977b). Runoff in cropped plots exceeds
that in natural bush by a factor of 20 to 35 depending upon the crop,
while soil loss increased 30 to 50 times for the same crops as
compared to natural vegetation (Table A-8). Similar results are
reported for other sites in Ivory Coast and Upper Volta.
Lai (1977b) tested the effectiveness of various types of mulch as
well as several variations of minimum tillage. While mulch had less


41
effect on runoff, it stopped soil erosion, even on the 15% slopes
(Lai, 1977b). Experiments in Nigeria demonstrated the effectiveness
of alternative methods of field preparation and planting. While
croplands with ridges oriented downslope yielded 28% runoff and 20
tons ha ^ of soil loss, alternate tied ridges across the slope reduced
runoff to 13% and soil losses to 6 tons ha ^ yr ^ (Kowal, 1970).
One West African study reported on the continuous measurement of
erosion in the same plots over several years. Lai (1977b) found that
slope effects may be reversed after a few years. After a rapid
initial loss of the topmost layers on steep inclines the erosion rate
decreases, while gentler slopes maintain a more constant erosion rate.
This indicates the importance of documenting the land use history of
hillslope sites so as to account for the influence of past soil loss
and profile modification.
More detailed surveys of the published West African soil erosion
literature have been complied by Lai (1977a), Okigbo (1977), Fournier
(1967), and Jones and Wild (1975). Projects in progress include
minimum tillage and multiple cropping experiments in plots at IITA
(Lai et al., 1979).
There is also substantial similarity between some of the lowland
dry forest and montane ecosystems of East Africa and the Caribbean.
The farming systems have some crops and practices in common, though
fewer than in the case of West Africa.
Erosion plot studies in Uganda yield similar results to the
experiments already cited in West Africa and Latin America (Table A-9)
(Sperow and Keefer, 1975). The major difference is in the magnitude


42
of total soil loss under bare fallow and annual crops, which is
attributable to lower annual rainfall. The relationship between
vegetation types and soil loss, however, is the same.
Similar experiments in Tanzania and Zimbabwe showed the same
range of soil loss for annual crops, bare fallow, and pasture. Losses
under maize in Zimbabwe varied from 4 to 10 tons ha ^ yr ^ (Hudson,
1957). Mosquito netting placed above the soil reduced losses on
clean-till plots to 1.2 tons ha ^ yr \ demonstrating the importance
of interception by canopy and ground cover (Hudson, 1957; Lai, 1977b).
Early erosion studies in Tanzania (Staples, 1939; Rensburg, 1955)
compared sorghum with grass cover and sorghum/grass strip cropping
(Okigbo, 1977; Temple, 1972). Soil losses varied between 9 and 116
tons ha ^ yr under sorghum, depending upon site and cultivation
practice. Grasslands yielded 1.2 tons ha 1 yr ^ and sorghum strip-
cropped with grass yielded from 4 to 60 tons ha ^ yr ^.
Findings in Kenya confirm that infiltration is greatly reduced by
grazing (Stephens, 1971; Thomas, 1974) and that while cultivation may
increase initially, the effect is temporary. Similar results have
been observed in Tanzania with a six-fold increase in peak runoff rate
after conversion from forest to annual crops (Wrigley, 1969).
Hutchinson et al. (1958) recorded a ten-fold runoff increase in clean-
tilled land converted from natural grassland. The end result is
higher runoff and erosion rates under both grazing and cultivation
(Ahn, 1977).
A less typical, broader study of erosion in Tanzania demonstrated
the soil conservation potential of many traditional farming methods,
including mulching and intercropping of field crops with bananas as


43
well as other crop association and rotation schemes. The contrasting
erosion plot sites in the Uluguru highlands and the semi-arid central
plains formed part of the same catchment study mentioned above (Rapp,
1977). Although some of the findings closely resemble the results for
other African and Latin American sites (Temple, 1972), high losses
were recorded for clean-weeded coffee, exceeding losses in nearby
plotswith maize (Anderson, 1962). The somewhat atypical results in
this case demonstrate that perennials do not necessarily conserve soil
better than annuals (Ahn, 1977). A wide range of sediment yields is
also reported for tea plantations in East Africa, with differences
attributed to management variables (Othieno, 1975).
Erosion in tree crop plantations is recognized as a ma^or
contribution to regional sediment yields in Southeast Asia, where long
experience with rubber and tea plantations has demonstrated the wide
variation due to management of canopy and ground cover as well as
tillage practices (Coulter, 1972; Edwards, 1977). While tea
plantations yielded approximately one-seventh the soil loss from bare
fallow (Hasselo and Sikurajapathy, 1965) in two cases in India (Table
A-10), the erosion in tea was double that reported for forested plots
-1 -1
m Malaysia and reached rates of 40 tons ha yr in Sri Lanka prior
to implementation of conservation practices (Lai, 1977a, 1977c).
Measurements from several land use and land treatment types in
upland Mindinao in the Philippines (Table A-ll) illustrated the
relationship of both land cover and land use rotations to runoff and
erosion rates (Kellman, 1969). The plantations had relatively little


44
impact in comparison with logging and farming uses. The results from
long established rice and corn plots suggest cumulative
destabilization of soil structure under permanent cultivation.
Qualitative and Informal Analyses of Land Use and Erosion
in the Caribbean and Similar Environments
Development and technology transfer projects in erosion prone
areas have yielded useful information on land use and erosion
problems. A few examples of interest include: studies of erosion and
overgrazing in the Bolivian highlands (LeBaron et al., 1979); a
summary of the programs of the Yallahs Valley Land Authority in
Jamaica (Floyd, 1969); the progress reports, project summaries and
consultant reports from Plan Sierra in the Dominican Republic
(Antonini and York, 1979; Chaney and Lewis, 1980; Montero, 1979;
Swedforest, 1980), and development agency communiques on resource
management projects in Cajamarca, Peru (Nicholaides and Hildebrand,
1980b), the southwestern slopes of the Dominican Republic, the
interior of Jamaica, and Haiti (Murray, 1977; Zuvekas, 1978).
Published and mimeographed reports of the government agencies charged
with soil conservation in the Dominican Republic (Gomez, 1980; Lopez,
1980; Paulet, 1980; Russo, 1980; Vasquez, 1980), Puerto Rico, Jamaica,
Trinidad, Tobago, and other areas of the Caribbean (Henriquez, 1962)
also provide useful information.
The qualitative information gathered from such sources can help
link land management variables to the onset, severity, and persistence
of various erosion features and sedimentation problems. The
settlement and development history of a region is often indicative of


45
the changing rates of deforestation and the intensity of cultivation
over time. This provides a basis for relating current erosion rates
to the succession of land use systems in an area. Future trends in
land use and erosion rates can be estimated more realistically with
the aid of this type of background information.
The Role of Farming Systems Research in Soil and
Water Conservation
It is important to view the effects of land use and erosion
problems within the source areas rather than simply calculating the
net export of sediment to downstream areas (Carmona, 1980). The key
to changing the situation is to be found in the internal workings of
the upland land use systems (Quezada, 1977) and in their relationship
to the larger system (Antonini and York, 1979). Any changes in
management of the uplands must take into account the causes of current
practices and the practical feasibility of proposed technological or
land use changes (Morgan, 1979; Russo, 1980).
For example, the proposed solutions to erosion problems may
involve specific technical innovations such as mulching or terracing
(Sheng and Michaelson, 1973), or a change of land use may be
suggested. Terracing and mulching with increased crop cover
alternately have received priority in various conservation projects
with contrasting and somewhat unpredictable results. Reports from
Zimbabwe demonstrate the effectiveness of increased crop cover and
mulches (Ahn, 1977; Hudson, 1957), while climatic and farming system
constraints in Kenya make terracing a more attractive alternative
(Ahn, 1977; Thomas, 1974). In Tanzania, terracing in inappropriate


46
situations increased the erosion hazard from landslides (Temple and
Rapp, 1972). Both approaches have been tried in the Caribbean (Sheng
and Michaelson, 1973; Wilson, 1976) though the structural
modifications of slope predominate in projects in Jamaica and the
Dominican Republic (Bonilla, 1980; Vasquez, 1980). The advisability
of this approach is questionable.
The evaluation of proposed technological changes must be measured
against the existing practices as well as against the more drastic
option of land use change on a broad scale. For such analyses a
knowledge of erosion, sedimentation and their variation according to
crop and vegetation type will not suffice. The proper choice of
conservation measures depends upon a full understanding of farming and
related land use systems in a region.
Farming Systems, Agroecosystems and Agroforestry Research
Overview
The use of the systems approach in agricultural development
efforts is a relatively recent phenomenon (Dent and Anderson, 1971;
Duckman et al., 1974). Unlike the study of erosion and sedimentation,
the research in this field has been conducted mainly in the tropics.
Ruthenberg (1980) carried out much of the pioneer work in farming
systems, primarily in smallholder farming districts in Kenya, Tanzania
and West Africa. Most of the research in farming systems in Africa
(Ruthenberg, 1980) and Central America has focused on small, limited
resource farmers.
Studies have been conducted by interdisciplinary and often
international teams of agronomists, agricultural economists,


47
anthropologists and ecologists. The individual farm enterprise and
the farm household have been the preferred units of analysis. Systems
concepts have been consistently employed, although the methodology and
topical emphasis vary with the regional and disciplinary orientation
of the researchers and institutions involved. The methodological and
substantive contributions of this avenue of research provide a solid
point of departure from which to expand the treatment of ecological
aspects of the problem and to extend the scale of analysis beyond the
farm level.
Selected Examples from Africa
Most of the African research in cultivation and grazing systems
has focused on systems that represent transitions from traditional to
more commercialized forms of production. Collecting, or hunting and
gathering, are largely ignored, as is forestry. Cultivation and
grazing systems receive the greatest emphasis. Ruthenberg (1980)
divides cultivation into shifting cultivation, fallow systems
regulated by farming (pasture-crop rotations), permanent upland
cultivation, arable irrigation farming, and perennial crops. Within
grazing systems pastoral nomadism and ranching are considered.
Ruthenberg (1976) and Flinn and Lagemann (1980) analyzed resource
utilization by farmers, their impact on the quality and condition of
the resource base, and the future implications for sustained
production in the area. Carrying capacity was determined by inherent
qualities and current condition of the environment, level and type of
technology, and standard of living (Lagemann, 1977). Within this


48
context, Lagemann tested and criticized Boserup's theory of
agricultural innovation (Boserup, 1965) with respect to environmental
response to intensification of agriculture. The relationship of
current population densities to environmental carrying capacity under
different production systems was demonstrated by simulation models.
The simulations extrapolate current bush fallow practices (in West
African examples) to predict net environmental degradation,
diminishing net yields and decreasing production per unit labor input,
all due primarily to declining soil fertility.
Patterns of land use, spatial organization of cropping, farm
level resource management and farm level economic analyses are
emphasized within this tradition. Methods for evaluating technical
innovations for low resource farmers have also been presented in
recent studies (Flinn and Lagemann, 1980; Flinn et al., 1980). Many
of the recent studies have been conducted in Nigeria in conjunction
with the International Institute for Tropical Agriculture (IITA).
Farming systems research and extension programs also have been
developed in East Africa (Collinson, 1981).
Studies of shifting cultivation systems in West and Central
Africa indicate a potential for maintaining shifting cultivation
indefinitely at a lowered but sustained rate of production, relative
to undisturbed forest systems. Soil fertility is reduced to
approximately 75% of the value for undisturbed forest soils. The
successful attainment of adequate sustained production hinges on the
rotation cycle, which must vary between 20 to 50 years of forest
fallow per year of cultivation. The implications for carrying


49
capacity are clear. While the system itself may work, shifting
_2
cultivation cannot support more than about 20 to 50 persons km ,
taking into account the required fallow. Further experimental work at
IITA by Greenland, Lai, and others has explored alternatives to this
system, emphasizing soil management and conservation under bush fallow
and continuous cropping (Lai et al., 1979; Lai, 1977a) and continuous
mixed-cropping systems (Greenland, 1975; Ruthenberg, 1976).
Bioeconomic modelling has been proposed to evaluate alternative soil
conservation practices and cropping systems (Dumsday and Flinn, 1977).
Agroforestry research in Africa has combined experiments with
commercial forestry and subsistence agriculture (King, 1968). The
taungya system features mixed cropping of commercially harvested and
replanted forest tracts, with the tenant farmers caring for the
seedlings and saplings as well as their food crops over a period of
about four years (Dubois,.1979; King, 1978). The field of
agroforestry has further developed to include diagnostic and
experimental work with existing subsistence and commercial production
systems that feature some combination of trees, livestock production
and/or field crops (Brookman, 1976; Douglas and Hart, 1976; Olawoye,
1975; Parry, 1957; Raintree, 1982; Lundgren, 1982). Both cocoa and
oil palm production on small farms have been studied within this
context (Flinn, 1980; Grinnell, 1977; Lagemann, 1977; Letouzay, 1955)
as well as many traditional systems of shifting agriculture that
include management of tree crops (Dubois, 1979; King, 1968). In
general, mixed tree crop/annuals production systems are more diverse
and more stable, both in economic and ecological terms (Lagemann,
1977).


50
Tree crops can provide fuel wood, high protein forage, lumber,
fiber, food, mulch, and cash crops (Douglas and Hart, 1976). While
establishment of such a stand requires more capital, labor and
management than a plot of annual crops, the products are often of
higher value and can be used on the farm to fill a wide range of
subsistence needs (Lagemann, 1977). Soil fertility and structural
stability are enhanced by partial tree crop cover, providing some of
the ecological benefits of forest fallow without sacrificing economic
production (Nair, 1983). In hillslope environments the combination of
trees and row crops is particularly advantageous since tree crops
offer the soil greater protection from erosion (Douglas and Hart,
1976) .
While the study of agroforestry is still relatively new, the
field is developing rapidly, in part as a response to the need to
increase small farm production in marginal lands while maintaining or
rehabilitating watersheds and forest resources (King and Chandler,
1978). Current research at the International Council for Research in
Agroforestry in Kenya (ICRAF) focuses on the elaboration of a
methodology for diagnosis of new ones. Rapid survey diagnostic
techniques and combined research/extension programs are being
developed and tested in field sites that include farm level and
community level work as well as experimental plot studies (Raintree,
1982). The multidisciplinary staff includes foresters, agricultural
economists, agronomists, anthropologists and ecologists, among others.
The international client areas for the methodology being developed by
ICRAF include the hillslope farms of the Caribbean, as well as the


51
Andean highlands, the Amazonian lowlands, and the African savannas,
and many other fragile and/or marginal environments under cultivation
(King and Chandler, 1978).
Central American Research
Farming systems research in Central America has been conducted at
the Center for Teaching and Research in Tropical Agriculture (CATIE)
in Costa Rica, and the Institute of Agricultural Science and
Technology (ICTA) in Guatemala. Both have sought to serve small
farmers and to modify, rather than replace, traditional systems of
agriculture (including hillslope farming).
Farming systems research conducted at CATIE has emphasized
economic and agronomic description of existing cropping systems
through community and regional level questionnaires and surveys, using
standard sampling procedures (Navarro, 1979). Experiments have
focused on mixed-crop combinations for optimization of production
given the available natural and economic inputs. The general
perspective as well as some specific techniques of ecosystems and
energy analysis have been applied (Hart and Pinchmat, 1980; Moreno,
1977).
Several precedents exist for ecosystem and energy analysis
applied to production systems, including examples from India (Revelle,
1976; Odend'hal, 1972), the United States (Burnett, 1977; Ewel, 1973),
Indonesia (Geertz, 1963, 1972), Guatemala (Carter and Snedaker, 1969),
the Dominican Republic (Werge, 1975) and elsewhere (Lugo and Snedaker,
1971; Odum, 1967). Energy budget analyses more or less independent of


52
the systems approach also have been conducted by researchers in the
United States (Lockeretz, 1975; Pimentel, 1973), England (Leach,
1976), and Asia (Makhijani and Poole, 1975; Smil, 1979). The energy
studies at CATIE, however, use the systems approach and focus on plant
and animal production subsystems, as opposed to the socioeconomic or
environmental subsystems (Hart, 1980; Holl, 1979). The majority of
the cropping systems research does not address questions of erosion
and watershed protection although some studies have been carried out
on erosion rates under hillslope cropping systems (Bermudez, 1979) and
under agroforestry systems (Apolo, 1979).
The application of energy analysis to individual farms (Hart,
1980) represents a distinct and particularly useful subset of the
general body of research at CATIE. The energy analysis methodology
developed by Odum (1971) and Odum and Odum (1976) has been adapted by
Hart (1980) for use in the rural farming districts of Central America.
This method can be used to identify and describe existing successful
adaptations. It can also be used to highlight the functional
relationships found in the average case, in order to choose the
critical points in the system where specific (and viable) changes will
have a major impact on total production (Fig. 5).
Hart (1980) has elaborated further a field survey methodology
using systems concepts and energy flow diagrams in rapid surveys of
rural farming communities, to characterize existing farming systems in
qualitative and quantitative terms. Still lacking are inclusion of
environmental inputs, outputs and storages, and the extrapolation from
the farm to larger scales of analysis.


Ecological
Inputs
Cultural Inputs
N fixation
Lugehd
- lufornalion I'lows
Kitergy/Matarial s Flows
Honey Flows
*
Solar Energy
Ra 1nf al 1
nutranla
Surface VLter
Dutr1ents
w
0round iter
nutrante
w
Soil Foma t Ion
nutriente

i
cl
I
j I
A
LAi
TJ
a.
IHVtffTOHT: CONTEKTS 0/ 6TSTEM
Total floluae Viler
*1*11./ Unit Tl*
at Surface
inla&le: Population*
*nd Bloaaaa
Total Yolune Uitar
Stored for lluuo
Uaa
faga talln:
Iraa of Coiaraga
Population* aJul
Bloauiil
| | Liquid laaata ~~]
j Huaan Population*
Soil Tolu*#/ir#a
Soluat of Nutrient
Crop* In Ground
or In Storag*
(No. Bloaaaa)
firewood and
Charcoal
1;
B

< a
¡
3
d 4
!
3
l
1
J
a
a
Si
Vala of Paraonal *
Bouaahold Property
Valu* of Iaplasanta
* fehlclea
Valu* of Standing
Inf ra* truc tur*
land Valu*
Valu* of lar.) Corar
Valu* of Crop* In
Ground or Stored
Valu*/int. of Poaall
futa
4, o).
Si
i
i,
Brat
Uitr Evaporation
-
Uitrr Tranrplratlon
" w
Uitrr Runoff
Dutrlrnta/fertlll
paatlcldra
pTh6"**
(bacteria, finiere,
paraeltea)

Soli Eroalon
nutrlanta/fartlllief
w
pee ticlda
palbogana
Drtrltua
w
DoUanta
Huaan/Anlsal Wiaie
w
nutr rota
mw
N Voltil1ralion
Surfer
Viter
Oround-
U.t*r
M Drnl trlf 1 oatlon
fool oglcal
Outputa
Cultural Outputa
Fig. 5. Input-output diagram for interview notations and monitoring
(_n
U)


54
A similar approach has been applied to the environmental
subsystem in a National Science Foundation project in the same area
(Berish, 1982; Brown, 1982; Ewel, 1981). The project focuses on
nutrient cycling in small farm and successional plots, and the
research is directed toward improved management of nutrient cycles in
agroecosystems.
Agroforestry research at CATIE has combined with forestry and
watershed protection research (Budowski, 1977; Combe, 1979). The
major experiments to date emphasize the use of laurel (Cordia
alliodora) and poro (Erythrina poeppigiana) trees to provide shade,
reduce erosion, and produce firewood or lumber within coffee
plantations in montane environments (Beer, 1979; Bermudez, 1979;
Rosero and Gewald, 1979; Russo, 1983). Combinations of grazing with
plantations of alder (Alnus acuminata), Eucalyptus sp. and cypress
(Cypressus lusitanica) also have been tested in multiple field trials
(Combe, 1979; Gonzales et al., 1979). One test of the taungya system
has been conducted, using Gmelina arbrea interplanted with maize and
beans (Rosero, 1979). Most of the trials have proven successful when
judged in terms of mixed ecological and economic criteria by
researchers, but widespread adoption by small farmers has not yet
occurred beyond the original centers of innovation.
The planting of nitrogen-fixing leguminous trees in or around
pastures and within coffee and cocoa plantations is common practice in
many parts of Central America, but the density and distribution of the
trees often is very limited. Moreover, the need for the introduction
and/or development of agroforestry systems often is greatest in the


55
marginal lands where small farmers are dedicated exclusively to
cultivation of annual crops in slash and burn or bush fallow rotations
(Lagemann, 1978).
The farming systems research conducted at the Institute of
Agricultural Science and Technology (ICTA) in Guatemala, concentrated
more on the socioeconomic subsystem and on the integration of research
and extension (Reiche and Lee, 1978). Extensive contact with
hillslope farmers in Guatemala's rugged uplands also produced some
work related to erosion control (Hildebrand, 1981).
The most significant result of the ICTA research was the
development of a practical multidisciplinary approach that merges
research and extension efforts in farming systems. Farming systems
research and extension (FSR/E) integrates farmers, researchers and
extensionists into an effective group to identify and solve problems
and promote dissemination of the -solutions (Hildebrand, 1981; Gladwin,
1981). This is reflected in the rapid survey technique (sondeo)
developed for use by interdisciplinary research and extension teams
with small, low-technology farmers (Swisher et al., 1982).
The survey itself consists of a series of informal interviews
conducted by an interdisciplinary team that includes a social
scientist (usually an anthropologist), an agronomist or related
technical specialist in agriculture, and an economist (Hildebrand,
1981). The approach borrows heavily from traditional anthropological
field methods with reliance on key informants and on corroboration of
information from various sources. The use of open-ended interviews
with chosen informants contrasts with the random sampling of the


56
population so often used with questionnaires or structured interviews.
The latter, more formal approach is a more common and preferred form
of data gathering in many disciplines. However, it is also expensive
and time-consuming, and it presupposes an accurate population census
or property survey on which to base the sample. Formal surveys and
questionnaires also inherently limit the categories of information to
be treated. Little room is left for the definition in the field of
problems not already recognized prior to the survey. Opportunities to
explore the unique relationship between various aspects of a problem
in an area are also constrained by the format. The major
considerations in using the sondeo technique are quality versus
quantity of information and cost of the survey versus useful
information obtained.
The sondeo provides a practical and effective means of
reconnaissance and data gathering. It also lays the groundwork for
future extension programs in the area. During the intensive one-to-
two-week survey, the knowledge and needs of the farmer are
incorporated into the design (form and content) of subsequent on-farm
research projects. This methodology relies heavily on the judgement
of the research team, the local populace, and individual farmers to
define farming systems and their problems and to choose representative
or exceptional cases for further study, according to the goals of each
project (Hildebrand, 1981).
The experiments themselves are on-farm trials which may or may
not be replicated at experimental stations. The sondeo and the field
trials are supplemented by farm record-keeping. A family member keeps


57
an account of the farm's inputs and outputs as well as of activities
and movement of materials within the farm itself. These records also
assist in the evaluation of on-farm trial results and provide basic
data for further trials and/or discussions with farmers. The success
of a new technology is judged at least in part by the farmer's
perception of its performance and by its subsequent adoption by him
and other farmers in the area. This indicates t some extent the
"fitness" of a technology for the farming system as a whole, at the
farm level (Swisher et al., 1982).
The sondeo as well as the subsequent on-farm trials and farm
record-keeping of the FSR/E approach are readily adaptable to soil and
water conservation research in hillslope environments in the
Caribbean. The major elements lacking are greater attention to
natural resource management at the farm level and a method for
predicting and evaluating the success of a given technology at the
watershed or regional level.


CHAPTER III
METHODOLOGY
The General Approach
The methods reported in the literature review include a variety
of techniques that can be incorporated into a consistent and
appropriate methodology. The end product, however, must be more than
a method or a collection of techniques. The methodology proposed and
tested in this study is an adaptation of the scientific method in
general, and systems analysis in particular, to the interdisciplinary
study and treatment of the problems of land use, erosion, and
sedimentation in the underdeveloped nations of the tropics. The
research approach combined elements of farming system and ecosystem
analysis, drawing most heavily on the work of Odum (1971, 1982),
Antonini et al. (1975), Bormann and Likens (1979), Hildebrand (1981),
and Hart (1980), all described in greater detail in Chapter II.
The research tested specific management-related hypotheses under
field conditions. The cases studied required immediate action based
on tentative solutions from experimental results, prior to extensive
repetition and replication. The general research model included
direct outputs from verification to policy and production sectors and
a feedback from verification to further experimentation (Fig. 3).
The feedbacks in the research program imply an ongoing process of
learning. The time constraints in applied research were handled by
using these feedback loops to continually test and modify the


59
tentative solutions already proposed. This iterative approach has
already been tested in farming systems research and extension programs
(Hildebrand, 1981).
The usual concept of applied research is one of a finite activity
to be carried out and completed by specialists, after which they will
offer a set of definitive conclusions to be implemented. In this
case, the study area was viewed as a system in a flux, constantly
adjusting to changes in internal and boundary conditions. The object
of study in this case also had a subjective component. The role of
people in determining the direction of ecosystem evolution was taken
into account. Residents of the region contributed to the
investigation as both informants and participants in data gathering,
experimental and verification procedures. The researcher participated
in an on-going experiment in which people living in the area sought
short-term relief and long-term solutions to problems at least
partially perceived and defined by them.
The study was designed to accommodate the distinct priorities and
information needs of local clients, scientists, and the regional
policy sector. The experimental design and data analyses tested
multi-faceted hypotheses concerning technology and land use
alternatives for the region. Each experiment included: a primary
hypothesis as to the biophysical or economic performance of a
particular alternative; a secondary hypothesis concerning how the
proposed change would fit into the existing system; and a third
hypothesis that the system, as such, could and should be sustained,
with modifications.


60
The primary hypotheses were tested and judged jointly by researchers
and farmers, according to objective, quantitative criteria. The
secondary postulate was tested and judged by the farmer according to
subjective criteria, based on overall practical performance. Researchers
provided a posteriori explanation and interpretation of the farmers'
experience. The value judgement as to the fitness of the existing system
was considered the joint prerogative of the local residents and the
policy sector (clients) while the researcher determined the system's
sustainability by ecological analysis of current trends.
Materials and Methods
The study was conducted at four scales of analysis within a nested
hierarchy of spatial units, including: the Plan Sierra impact area (2500
2 2 2
km ); watersheds of 500 to 100 km ; small watersheds of 1 to 20 km ; and
individual landholdings ranging from 0.5 to 500 ha.
Chronologically the research activities were grouped into three
phases of increasingly finer levels of resolution and greater detail of
analysis. The regional reconnaissance and refinement of problem
definition was followed by detailed characterization of the study sites
at the watershed and plot level. The third phase consisted of monitoring
runoff, erosion, sedimentation, and production under varying land use and
soil conservation practices (Fig. 4).
The conceptual model of the Caribbean region (Fig. 1) formed the
basis for the overall research design, while at each successive level of
resolution a conceptual model was proposed, evaluated through field data
collection, and refined or modified based on empirical evidence and
testing of specific hypotheses inherent in the model.


61
Regional Reconnaissance
The inventory of existing land use systems and the condition of
soil and water resources within the Plan Sierra impact area provided
the data for refinement of the problem definition and for subsequent
application of the research design within successively smaller units
of analysis.
Most of the reconnaissance activities were completed between
January and March 1980. A regional description and summary of land
use systems was complied from library and field research. A review
and synthesis of maps, aerial photographs, statistical summaries and
literature relating to the study area preceded the field surveys. The
area was stratified into multitopic subregions based on cartographic
analysis of physical, biotic and socioeconomic characteristics mapped
at scales of 1:250,000 and 1:50,000. The major criteria for zonation
were life zones (Holdridge, 1967; OAS, 1967), topographic
characteristics, and land use characteristics, the latter reflecting
population density as well as condition and productivity of the land.
Maps were compiled by Plan Sierra cartographic and project staff from
topographic and thematic maps at 1:250,000 and 1:50,000 (Jennings,
1979a; OAS, 1967; Swedforest, 1980), and from aerial photographs at
1:20,000.
The field survey was similar to the general procedure outlined by
Hildebrand (1981). Several rapid reconnaissance surveys of erosion
features, land cover and land use systems were conducted within the
regional subdivisions outlined above, as part of Plan Sierra program
development in soil and water conservation. Survey teams varied in


62
composition, but usually included the author and one to four
specialists and paraprofessionals in engineering, forestry, agronomy,
and soil conservation.
The surveys included formal and informal interviews with
residents, as well as field mapping of land cover, land use and
evidence of erosion and sedimentation in the various areas visited.
The selection of sites for more detailed observation reflected a
strong reliance on the knowledge of agronomists, foresters and
conservationists already familiar with the area, as well as the
opinion of residents as to what areas constituted typical or extreme
examples of particular physical and socioeconomic characteristics.
Field survey records included numerical data, maps, and interview with
residents, as well as the impressions and observations of the survey
team.
A synthesis of the cumulative results of prior field
reconnaissance by interdisciplinary teams of consultants (Chaney and
Lewis, 1980; Georges, 1981; Hart, 1981; Montero et al., 1981; Navarro,
1981; Nicholaides and Hildebrand, 1980b; Safa and Gladwin, 1981;
Santos, 1981; Swedforest, 1980) supplements the information gathered
from the author's survey. Written reports and personal communication
from the consultants and visiting scientists contributed to updates of
the regional profile.
Refinement of the Research Design
Based on the reviews of regional information and the completed
field reconnaissance, the conceptual model of the region was modified


63
to include relationships previously omitted or incorrectly defined.
The research design then was developed to test the major hypotheses
implied in the model. The discharge and sediment yield of two large
watersheds were measured over a 15-month study period. During the
same interval subwatersheds were described and monitored in greater
detail to relate differences in discharge and sediment yield to
varying physical and land use characteristics. Erosion plots
constructed within the subwatersheds provided comparative data on
runoff, erosion and production under different land uses, each with
varying conservation practices.
The 18-month period of study for phases two and three extended
from 1 Apr. 1980 to 30 June 1981. This period included a full
hydrologic year, from 1 Apr. 1980 to 1 Apr. 1981, and also allowed
repetition of sampling and monitoring during the time of peak
rainfall, from April through June.
The spatial and logistical organization of research activities
are illustrated in diagram and tabular form (Fig. 6, Table 1). The
chronological order of analytical and data collection procedures
parallels the general case described in Fig. 4. The choice of study
sites reflects the insights gained from the review and reconnaissance
survey, as well as the information needs of Plan Sierra.
Study Sites
The study sites selected within the Plan Sierra impact region
2
included two large watersheds (500 to 100 km ), five small watersheds
2
(1 to 20 km ) nested within the two larger units, and 16 plots on nine
landholdings situated within three of the small watersheds.


I1.0 IS
J s i i m.itos o f
Jaluir ami
mi la
V
Il.OTS
1 li|'ll( S
I,A IS! IH ANI> ri/MTUIAl.S
Fig. 6. Organization of research activities
OA


Table 1. Relative location and land cover of erosion plots.
Amina
River Watershed
Mao River Watershed
Subwatershed
Plot No.
Cover
Subwatershed Plot No.
Cover
Prieto
(no. 64)
80
Coffee (established)
Pananao
(no. 60) 91
Sisal, foodcrops
Prieto
(no. 64)
81
Coffee (new)
II
92
Sisal, foodcrops
Hondo
(no. 67)
82
Pigeon peas, sweet
It
93
Pasture
potato
Hondo
(no. 67)
83
II II
II
94
Pasture
Hondo
II
84
Yuca, beans, corn
II
95
Yuca
Hondo
II
85
II II II
f
96
Pine forest
Hondo
II
86
Pasture
Hondo
It
87
Pine forest
Hondo
II
88
Pine forest
Hondo
II
89
Pasture
t]Mo. 96 falls outside of Pananao watershed but serves as a surrogate Pananao forest plot.
cn


66
The criteria for selection of the large watersheds for further
study were as follows:
1. Availability of historical data on climate and river
discharge;
2. Current coverage of the area by rain gauges and water level
recorders;
3. Need for erosion and sedimentation data in the area; and
4. Presence of Plan Sierra personnel and facilities in the
region.
The Plan Sierra region is drained by three major rivers, Mao, Bao, and
Amina. All three watersheds are partially contained within the impact
area and empty into the Yaque del Norte River. Each river is a
current or future source of water for hydroelectric and irrigation
projects planned within the borders of the study area. The study
concentrated on the Mao and Amina Rivers.
Within each of the two large watersheds, subwatersheds of
critical concern were selected based on the land use systems, erosion
features, climate, soil, and topography. Ease of access and presence
of soil conservation personnel also played a part in the decision.
Five small watersheds were chosen to test the following
hypotheses: 1) that land use is the major determinant of erosion and
sedimentation in the region; and 2) that the combination of pasture
and field crops common in the densely populated central region
contributes more heavily to sedimentation than land use practices
based on coffee which is more typical in the highlands. The size of
2
the watersheds ranged from 1 to 30 km with two replicates in the
2
first size category (1-2 km ) and three in the second group (10-30
2
km ). The examples represented the smallest scale of analysis in

a

67
which the land use matrices characteristic of the given subregion
could be measured. In each case a cluster of settlements was
included, along with a combination of subsistence and commercial land
uses. This unit of analysis also was compatible with community level
socioeconomic analyses conducted in the region (Georges, 1982).
The smallest scale of analysis focused on landholdings associated
with individual households. Sixteen experimental plots were installed
on property belonging to nine separate households, all situated within
the watershed study areas mentioned above. As depicted in the
research design (Fig. 6, Table 1), two distinct physical subregions
and four major land use types were represented in the three clusters
of plots. The plots were chosen to reflect average characteristics of
microclimate, slope, soil and land use within each watershed. The
first three variables were held constant, while land use was varied.
Nested within the compari-son of land uses was a comparison of soil
conservation practices in plots planted to annual crops and a
comparison of stages of development in coffee and forest plots.
Characterization of Watersheds and Plots Selected for Further Study
Large watersheds
The portions of the watersheds included within the study were
delineated and measured on 1:50,000 scale topographic maps (U.S. Army
Map Service, 1962). Cartographic analysis of drainage networks,
geologic subregions, life zones, topography and land use indicated
areas whose composite characteristics favor high rates of runoff,
erosion and sedimentation.


68
The choice of sampling sites for more detailed study of discharge
and sedimentation was based on several criteria, including ease of
access, proximity to homes of observers, established hydrologic
monitoring sites, proposed dam sites, and regularity of the
longitudinal and cross-sectional profiles of the channel in the
vicinity of the potential sampling points. Profiles of the river
cross-sections were surveyed. The relative heights of the bridges and
other large structures along the stream bank also were measured to
provide reference points for reporting maximum flood stages. The
profiles are included in Appendix B.
Small watersheds
The small subwatersheds selected for further study were analyzed
by standard cartographic and photogrammetric methods to determine
total area, average slope, and area and distribution of land cover
types. Land use and socioeconomic characteristics of the watersheds
and their settlements were determined by field observation,
interviews, and a review of statistics available at the community
level.
The average slope was determined by the Wentworth (1930) method,
using topographic maps at the 1:50,000 scale. Each watershed was
mapped separately and overlain with a 1-km grid subdivided into four
cells each, with two diagonal cross sections per grid. The slope from
the center toward each corner was determined by the change in
elevation over the four 0.5-km transects. The average slope (%) was
calculated for each small grid cell (0.25 kg ), then for each larger
2
cell (1 km ), and finally for the watershed as a whole.


69
Total area was determined by planimetry. Area and distribution
of land cover types were determined in a three step process, beginning
with the interpretation of black and white aerial photographs at the
1:20,000 scale, taken in January and March 1980. The watershed
boundaries, the stream, its tributaries (if any) and major features
such as roads and paths were delineated on outline maps enlarged to
the same scale, as well as on the aerial photographs. Land cover
subdivisions were outlined on the photographs, based on differences in
color, texture, and pattern. In many cases the fine texture of land
cover subdivisions required aggregation into units of mixed land cover
recognizable by plot or property boundaries. The land cover units
previously outlined were classified according to land use, based on
prior field observations in the study area during the reconnaissance
survey. After transfer to a map of known scale and projection, the
areas of major land use types were measured by planimetry.
The incidence, type and distribution of erosion features in each
small watershed were noted during the initial field visits and during
the 15-month period of study. The number, severity and distribution
of landslips and landslides, gullies, and signs of rill and sheet
erosion were compared between watersheds included in the study as well
as with several other watersheds visited regularly in the course of
related studies. Residents often were questioned about new features,
or noticeable changes in pre-existing gullies, rills and landslides.
Whenever possible, the immediate causes, such as a particular storm
event, agricultural and construction activity, were identified. This
type of qualitative analysis provided a background for further


70
refinement of the research design (Mosley, 1980) and for more informed
interpretation of subsequent results.
Detailed analyses of channel erosion, deposition, and range of
flood stages were conducted during field inspection of the stream
courses. Variations in depth and areal extent of sediment deposits
gave some indication of sedimentation rates within the basins. The
general form of channel terraces and floodplains, as well as the
height of residual debris lodged in trees and rocks on the stream
banks, provided indicators of peak annual flood stages. The specific
topographic indicators included erosion features on the stream banks
and breaks in slope along the cross-sectional profile. The cross-
sections were surveyed by the stadia transit method (see Appendix B).
Estimates of peak floods for longer periods were obtained from
interviews with elderly residents of the area. Independent
questioning of several persons established the margin of error in the
responses.
The sampling and monitoring locations within each watershed were
chosen according to several criteria:
1. Situation relative to settlements and the land use to be
evaluated;
2. Ease of access;
3. Security of equipment from vandalism
4. Availability of natural footings and fastenings against
flash floods and transported debris; and
5. Regularity of channel longitudinal and cross-sectional
profiles in the vicinity.
Interviews with residents of the communities within the
watersheds were conducted at local meeting places, in the field and in


71
the homes of agricultural laborers and landholders. The Plan Sierra
soil conservationists living in the area also were interviewed, and
they in turn questioned residents about the settlement history, land
use and farming practice, past and present production levels, and
sources of income. Discussions with two anthropologists conducting
land use and migration research in the region also provided valuable
information and insights into the character of the communities in the
study areas (Pessar, 1981; Georges, 1982). Plan Sierra social
workers, agronomists and foresters familiar with the area of interest
also contributed to the socioeconomic profile of the small watersheds.
Description of individual landholdings
Selection of sites for measurement of runoff and erosion in
experimental plots was based on uniformity and degree of slope, as
well as type of land use, management practice and easy access for
construction and sampling purposes. Wherever possible, replicates of
land use and treatment were established within the same watershed and
also in another watershed to determine the margin of error in
measurement and to compare the relative difference in runoff and
sediment yield under varying conditions of site and land use.
The experimental design further subdivided the categories of
forest, pasture, coffee and annual crops to compare undisturbed and
secondary forest, new and established coffee stands, different types
of annual crops, and use of minimum tillage and hillside ditches for
erosion control in fields planted to annual crops. These subsets of
land use type were tested in paired plots in the same or adjacent land


72
holdings to guarantee duplication of all other conditions except the
variables of interest. Individual plots were chosen based on field
observation of the above mentioned criteria. The choices were
confirmed after consultation with agronomists, conservation personnel
and residents of the area to determine if the plot in question
constituted a representative example of management relative to the
surrounding watershed.
Formal and informal interviews with owners, residents, neighbors
and local Plan Sierra personnel served to outline the settlement, land
use and production history and the variability of natural conditions
for the individual plots. Detailed information on past and present
crop associations, rotations, yields, labor and material inputs, and
ratios of commercial to subsistence production came from intensive
interviews with the persons directly responsible for management of the
site for a period of 10 years or more. These interviews often spanned
two or three visits by one or more members of the research team. The -
format was open-ended to allow the participants to elaborate on their
experiences.
Farmers were encouraged to discuss their problems with respect to
subsistence and commercial production and to volunteer insights and
judgements as to potential solutions. The women and children working
at each site also were interviewed, usually on separate occasions, to
obtain accounts of their roles in production and natural resource
management as members of the farm household. Their perceptions of
problems and suggestions for changes also were solicited.
The baseline information to be obtained was outlined in diagram
form (Fig. 5). This helped the interviewers to keep track of the

V

73
subject matter covered. It also provided a convenient format for
recording and summarizing responses during or following the
discussion. Information noted on the diagrams and tables served to
evaluate the farm level models prior to initiating the erosion plot
experiments and watershed monitoring activities.
All sites were chosen to reflect variation in land use and
treatment, while slope and soil conditions were held constant and as
close as possible to the average for the watershed. Slope
measurements along the downslope transect were made prior to final
siting of all experimental plots.
Soil profile descriptions, characterization of soil samples by
laboratory analysis, and taxonomic classification constituted part of
the site description at each plot. Rectangular trenches at least 1 m
deep, 1 m long and 0.5 m wide were cut for observation and sampling.
Measurement and description of profile stratification, with detailed
description of color, texture, structure, and uniformity, by horizon,
were carried out according to the procedures outlined in the Soil
Survey Manual (USDA, 1951).
Munsell color charts were used for wet and dry color
determinations in the field (Munsell Color Co., 1951). Laboratory
analyses for N, P, K, and organic matter content followed standard
methods for determination of Kjeldahl N and Bray P by colorimetry, and
for determination of K by atomic absorption (USDA, 1975).
The North District Research Laboratory (CENDA) of the State
Secretariat of Agriculture conducted all laboratory tests for the
project, including physical and chemical characterization of soil


74
samples. Soil classifications were confirmed and refined by soil
survey specialists from the Secretariat's South District Laboratory.
The relative infiltration rates of soils at the various sites
were determined by measurements with ring infiltrometers (Gregory and
Walling, 1973; Wisler and Brater, 1959). The inner ring was cut to a
25 cm diameter and the outer ring measured 40 cm across. After
placement in the ground with a minimum of soil displacement, the outer
ring was filled to form a barrier of saturated soil around the inner
ring which was filled to a 10-cm depth. _Throughout the next 4 hours a
nearly constant head of 10 cm was maintained while measurements of
water added were recorded at increasingly longer intervals. The form
used for the field measurements is included in Appendix H.
The frequency, distribution and severity of erosion features on
and around the plot sites were observed and noted prior to
construction of experimental plots. Wherever possible, the
developmental sequence of such features was determined from accounts
by the residents or neighbors and from repeated observation and
photographic records kept over the 15-month study period.
Detailed Analysis and Measurement of Key Parameters
The full characterization of the study areas at all three scales
of analysis served as a point of departure for the third phase of the
study, the measurement of water and soil exports from the individual
plots and from the nested sets of watersheds. Erosion, runoff, and
sediment transport were related to daily and continuous precipitation
measured at 15 stations in and near the Plan Sierra impact area.


75
In the large watersheds total sediment transport was measured and
sediment yield was calculated in order to estimate the magnitude of
the erosion problem on the watershed, to predict the future
sedimentation rates of the proposed dams, and to compare the losses
per unit area between the study areas and other sites for which
sediment yields have been reported. The measurement in the small
watersheds showed the integrated effects of land_ use and physical
characteristics in each study site. Subsequent tests of the data by
multiple analysis of variance indicated the relative influence of land
use and physical factors.
Experimental plots were included to demonstrate the impact of
varying specific crop types, land treatments and conservation
practices on erosion and runoff. The plots also provided the erosion
rate data necessary to calculate sediment delivery ratios for the
watersheds. Erosion plots were included because they allow
observation of the problem within the context of individual
landholdings and related households. The experiments were conducted
under the same conditions that limit and influence the management of
individual landholdings. While the most striking effects may be
expressed at the watershed level, the management decisions are made at
the household level. The degree to which such decisions are
constrained by pressures from the larger system does not change the
fact that these decisions directly determine, in turn, the condition
of the larger watershed. Any proposed changes must be tested within a
holistic framework at the level of the land managers.


76
Precipitation, discharge and sedimentation in the large watersheds
Records from 15 climatological stations collected by SEA's
Department of Meterology, and INDRHI provided daily precipitation
values as well as continuous data on rainfall amount and intensity at
five of the stations in the region. A multiple correlation analysis
of daily data from all 15 sites was performed to check for duplication
and overlap of information and to determine the variability of
rainfall over the study area (SAS, 1979).
The distribution of rainfall over each watershed and its
subdivisions was determined by the Thiessen polygon method (Wisler and
Brater, 1959). The daily rainfall values from the climatological
stations were extrapolated to the surrounding areas, then aggregated
into units more relevant to the study, such as subwatersheds. The
Thiessen polygons defining the area of influence around each station
were delineated and superimposed on a map of the Mao and Amina
watershed divisions, at a scale of 1:250,000 (Fig. 7). The watersheds
were defined by cartographic analysis of topographic maps at the
1:50,000 scale, and Thiessen polygons were constructed according to
standard methods summarized by Kenah (1980). After determining the
proportion of each subwatershed that fell within the polygon assigned
to each station, a weighted average of daily rainfall was calculated
for each subwatershed.
Stage and discharge measurements. On both rivers, hydrometric
stations maintained by INDHRI were located conveniently near the
downstream borders of the Plan Sierra impact area (Fig. 8), and
relatively close to concrete bridges. The river cross-sections were


Legend
Cllmatologicol stations
Watershed boundaries
Subwotershed boundaries
Thiessen polygon edges
Fig. 7. Diagram of Thiessen polygons superimposed on a map of the Mao and Amina watershed
subdivisions.
'j
-j
(Ji /


Fig. 8
Sampling sites () for large watersheds
'O
oo


79
gauged and calibrated periodically, such that the stage measurements
recorded at 0700 and 1700 hrs daily, as well as maximum flood stage,
could be converted directly to discharge rates based upon a nomograph
constructed by INDRHI hydrologists (see Appendix C). The stage-
discharge graphs relate discharge rates obtained through periodic
field measurements using the area-velocity method (Grover and
Harrington, 1943; Herschy, 1978) to the simultaneous reading of river
stage on a simple gauge.
For purposes of this study the nomographs of each river were
segmented and the equations for each segment were determined by simple
linear regression using the Statpak package of statistical programs
(MUSIC, 1967). The gauge readings recorded for morning, evening, and
flood peaks during the study period were converted to discharge rates
using the appropriate equations derived from the INDRHI nomographs.
Stage measurements made at the bridges by project personnel were
compared to simultaneous measurements at the hydrometric stations. A
simple linear regression equation converted the measurements made at
the bridge at sampling time to a gauge reading for the station. This
procedure replaced independent velocity-area discharge measurements
and allowed the substitution of a simple indirect method for the more
difficult and lengthy procedure used initially.
Comparison of rainfall and discharge during the study period to
the historical period. The total monthly rainfall totals over the
study period were compared to the average monthly rainfall totals over
the full period of record at stations with 11 or more years of data.
The SAS (1979) means program was used to compare the data from each


80
station for 1980 and 1981 with the prior records. The comparison of
river discharge measurements during the study period and the full
period of record followed the same procedure.
Measurement of suspended sediment concentration. Direct
measurement of sediment concentration required periodic sampling of
river water over the full 15-month study period, spanning three wet
seasons. This allowed sampling under a variety of conditions from low
flow to flash floods. The timing of observations was designed to
sample as wide a range as possible of the variation in discharge and
sediment concentration. In some cases time series samples were drawn
to cover the rise and fall of a particular flood.
The samples were extracted at the bridge site with a modified,
locally built Uppsala-type sampler (Fig. 9), as described by Nilsson
(1969) and Rapp (1977). The finished product resembles the USDA-48
Wading Sampler (USDA, 1979). The diameter of the sample intake nozzle
(0.64 cm) limited the particle size to a maximum of approximately 0.3
cm, which probably did not exclude any suspended sediments. Bedload
was not sampled. The emphasis on suspended sediment load as based on
the high average ratio of suspended sediment to total sediment load
(Gregory and Walling, 1973) and on the even higher ratios reported for
turbulent conditions.
While the sampling sites were not ideal by hydrometric criteria,
the accessibility and safety of bridge crossings outweighed other
factors (Herschy, 1978). Wading for samples from steep banks at river
narrows entails undue risk, particularly in areas subject to flash
flooding. The accessibility of the sampling site also can limit


81
Sediment
Fig. 9. Uppsala-type manual sampler for instantaneous
measurement of sediment concentrations in
streams (Rapp, 1977; Nilsson, 1969; USDA, 1979).


82
the number of observations. The bridge sites allowed simple grab
sampling during flood conditions.
Initial results from multiple samples stratified throughout the
river cross-sections showed little variation across the section but
pronounced differences with depth. Subsequent samples were taken in
pairs at 10 and 30 cm below the surface (water depth permitting) at
the center of the section. The samples normally were drawn by
extending the sampler downward from the upstream side of the bridge,
directing the intake into the flow, and extracting the sample just
prior to filling the bottle, to avoid sampling error.
The 1-2 L samples were stored at room temperature at field
headquarters, then shipped periodically to the CENDA laboratory.
Sediment content by dry weight was determined and reported as
concentration (g L ^). Analyses were conducted according to standard
methods for the determination of suspended solids in water. After the
sample volume was recorded, the samples were shaken, then poured into
Gooch crucibles lined with pre-weighed Whatman No. 5 fiberglass filter
paper. After the sample was strained into a flask, under a slight
vacuum, the filter and collected sediments were oven-dried at 105UC
for 24 hrs, then weighed. The final weight minus the previously
determined paper weight gave the new weight of the sediments for a
given sample. The latter then was divided by the sample volume to
obtain the sediment concentration.
Sediment discharge rates and total sediment transport.
Instantaneous rates of sediment discharge in tons sec ^ were
calculated by multiplying total river discharge (m sec ^) by the
sediment concentration (tons m ):


83
-1 3 -1 -3
Sq (tons sec ) = Q (m sec ) x Cs (tons m )
where Q = river discharge, Cs = concentration of sediment, and Sq =
sediment discharge rate.
Daily sediment discharge for non-flood sampling days was
calculated by averaging morning and evening discharge rates from stage
measurements. This was multiplied by a time conversion factor to
obtain total river discharge for the day. The total discharge times
the concentration approximates total sediment transport past the
sampling point for that day.
For samples drawn during or very close to short-lived peaks, the
flood peak duration was estimated from field records and observations
as well as from reports by the hydrometric station operator and other
nearby residents. River discharge is derived from measurements or
estimates of the flood stage, using the stage-discharge equations
described above. The instantaneous rates of sediment discharge were
calculated by multiplying sediment concentrations times the discharge
at the time of sampling. In cases of time series sampling during
flood events the average discharge rate for each time interval was
converted to a discharge value, then multiplied by the sediment
concentration. The sum of the river discharge and sediment discharge
over the sampling period provided empirical measures of sediment
transport for flood events of a given magnitude.
Analysis of relationships between discharge, sediment
concentration, sediment transport and rainfall. The frequency
distributions of all variables were tested by frequency analysis (SAS,
1979). Based on the results of the preliminary analysis the


84
relationships between discharge, an independent variable, and sediment
concentration and transport (dependent variables) were tested by
simple linear regression of raw and log-transformed data (SAS, 1979).
The relationships between river stage and sediment transport were
used to estimate sedimentation rates of the dams to be constructed
just downstream of the sampling points in both rivers. The daily
discharge for a full hydrologic year during the period of record was
used to generate daily sediment discharge values based on the
relationship established in the previous analyses. The total was
multiplied by a correction factor (ratio of average annual discharge
to the discharge for the 1980-1981 hydrologic year) to predict the
sediment discharge for an average year, as opposed to the study
period.
The relationship between the amount and distribution of rainfall
on the watershed (independent variable) and the discharge and sediment
concentration in the rivers (dependent variables) was tested by simple
and multiple linear regression of raw and log-transformed data (SAS,
1979). The same procedure was repeated by subwatershed. Critical
areas for further study were singled out by the relative strength of
association between rainfall in each subarea and the subsequent river
flood stages and sediment concentration. This information, combined
with field reconnaissance, contributed to assignment of research
priority by subwatersheds.
The relationships between total daily rainfall on the whole
watershed and discharge and sediment concentration also were tested by
simple linear regression. The total volume of rainfall on the


85
watershed was compared to total volume of discharge, by month, and the
remainder of rainfall minus discharge was attributed to storage and
evaporation. The results were compared to water balances previously
calculated for weather stations in or near the study area.
Precipitation, discharge, sedimentation, and production in small
watersheds
The proximity of three climatological stations to the respective
study areas allowed direct application of the rainfall data from these
stations. To supplement the existing monitoring network, small
plastic water gauges with 5-cm apertures were mounted at eye level on
wooden supports installed in well-exposed open areas. The amount,
intensity, and duration of rainfall on a daily basis during peak
rainfall periods were recorded.
Discharge and sediment concentration. Discharge measurements
made under relatively low flow conditions followed the procedures
prescribed by the velocity-area method (Herschy, 1978). The surveyed
cross-sections in each stream were segmented and a velocity
measurement was made at the center of each segment.
Surface flow velocity was measured with floats and chronometer as
illustrated in the diagram (Fig. 10). Each velocity measurement
consisted of three to five readings (sec), that were converted to flow
rates (m sec ^), then averaged. One complete velocity measurement was
made for the center of each segment, except in cases where the cross-
section was treated as a single segment. The discharge rate (Q) is
equal to the velocity (V) times the cross-sectional area (A), or a
segment thereof (Ax):
Q(m^sec ^) = V(m sec ^) x A(m^).


o
Floot
"a"
releases float at time t
Fig. 10. A. Illustration of velocity-area method. Person
and person "b" records time it takes float to move 2m. B. Cross-section of
stream showing placement of float to measure velocity and area of three
sections of the stream.
oo
a>


87
Stage measurements were made at the posts installed as supports for
stationary sediment samplers. An empirical parabolic equation was
derived by plotting field measurements of stage against the discharge
measured at that time. The discharge rate (Q) equals an empirically
derived constant (a) times the square of the stream depth (as
indicated by stage measurements): Q = a(h)2. The discharge rates, by
stage, are listed for each site in Appendix F.
Suspended sediment sampling in the streams required installation
of stationary samplers since flash floods are difficult to predict and
preclude the use of wading samplers. A modified, locally constructed
version of the USDA 59 siphon sampler (Fig. 11) collected samples at
pre-determined stage height intervals during the rising stage of each
flood. The sampling equipment is similar to the "Hayim 7" sampler
used successfully in flash flood sampling in Tanzania and Israel
(Rapp, 1977).
The equipment was located at the center, if possible, or to one
side of the surveyed cross-section in each stream (Fig. 11), depending
upon the requirements for stable installation. All samplers carried
at least five sample bottles arranged in a vertical series, from a
point just above the low flow water surface to approximately 1 m above
that point. The exact range depended upon the indicators of flood
stage along the stream banks. The cross-section profiles with the
position of the sampler and the height of sample intakes are included
in Appendix B.
Project personnel collected the samples as soon as possible after
each flood. The contents of each collection bottle were placed in a


CD
00
Fig. 11. Equipment installed in streams to measure sediment concentrations at different levels
of flooding.


89
separate sample container and identified by relative position in the
vertical series as well as by site, date and the estimated peak stage
of the flood. Each sample was assigned to a specific stage of the
rising flood. The samples were sent to the CENDA laboratory for
determination of suspended sediment concentration, as described
previously for river water samples.
Determination of discharge and sediment transport rates from
sampling results. Approximate sediment transport rates (Sa) for each
stage were calculated by multiplying the sediment concentration (Cs)
by an approximation of discharge (0):
-1 -3 3 -1
S(tons sec ) = Cs(tons m ) x Q(m sec ).
Discharge was calculated using the empirically derived stage-discharge
equation for the given site. Stage height was based on the number of
bottles filled in the vertical series on the sampler. Stage was
assumed to be halfway between the intake of the last bottle filled and
the intake for the next bottle up. If the highest bottle filled, the
stage was recorded as greater than or equal to the intake height plus
10 cm.
Total sediment transport (Ts) for individual flood events was
calculated by summing the products of discharge rate (Q) times
sediment concentration (Cs) times the estimated duration of each stage
(t) :
Ts(tons) = Q(m^ sec "*") x Cs(tons m x t (sec) .
Approximate hydrographs of specific events were constructed by
reference to the measured peak height of the flood stage and reports
on duration of the rising and falling stages of the flood. The annual


90
suspended sediment transport for low flow (non-flood) conditions also
was calculated for comparison of sediment transport between
watersheds.
The relationship between discharge and sediment concentration for
all events was tested by simple linear regression analysis (SAS, 1979)
with stage and discharge as the independent variables and sediment
concentration and transport as the dependent variables, respectively,
These analyses followed the procedures for derivation of sediment
rating curves used by the U.S. Bureau of Reclamation (Strand, 1975)
and by Rapp (1977) in Tanzania, as described in Chapter II.
Relationship of discharge and sediment transport to land use and
physical characteristes. The general linear model (GLM) program
package (SAS, 1979) tested the similarity of replicate watersheds and
the significance and degree of difference between watersheds by land
use and other categorical groupings, such as average slope, climate
and size. The similarity of replicates was tested by analysis of
variance, then the significance of differences between watershed
groupings based on land use and physical characteristics was tested.
Summary of production during the study period. The economic
production estimates for the study period were based in part on
extrapolation from intensive interviews with household members
participating in the more detailed farm level experiments. Informal
interviews with other farm families, farmworkers and shopkeepers
contributed to the general profile of economic production. The land
use maps and tables previously described provided a basis for rough
estimates of biological production and productivity. Information from


91
Plan Sierra personnel involved in marketing and credit supervision,
and data from a concurrent study by Georges (1982) also were included
in the production profile.
Evaluating the watershed model. The systems models of the
watersheds were evaluated, incorporating the results of the watershed
monitoring and estimates of productivity for the land use units
previously mapped. The meteorologic and geologic inputs and outputs
were calculated directly from empirical data obtained during the study
period. The production estimates for each land use type were based on
a combination of locally recorded yields and estimates of biomass,
production and productivity reported in the literature. Estimates for
bush, dry forest and pine forest were drawn from prior field studies
conducted in nearby areas (Jennings, 1979b; Montero, 1979). Estimates
for coffee crops and pasture were drawn from the literature and
modified to reflect site conditions. Population data were taken from
statistical references and confirmed by a housing count on recent
aerial photographs.
Measurements in small plots
Experimental plots and collectors constructed at 16 sites
captured and retained runoff water and sediments for frequent
measurement and sampling throughout the 15-month period. The
structures were installed within parcels of land maintained under
normal use and management conditions. At the six sites, where new
crops and soil conservation practices were tested, these changes were
already introduced by Plan Sierra. The experiments in these cases


92
were carried out over a much larger area than the erosion plots, and
each landholding was managed by the owners and workers who normally
oversee and perform work on the sites. The erosion plots received no
special treatment.
Plot design. The plots measured 22 x 2 m, which approximates the
Soil Conservation Service plots used in the United States (24 x 1.8 m)
and meets the standard specifications used by international research
organizations (FAO, 1977). In the three forested tracts the plots
were widened to 22 x 3 m in order to accommodate ..tree roots and to
include a more representative sample of the vegetation within the
plot.
The plots and collectors (Fig. 12) were designed after Djorovic
2
(1977) and Dunne (1977). The 44 m rectangular catchments were
oriented parallel to the slope and were bordered by single layer
concrete walls on three sides. The blocks were set into the ground
for secure footing, then sealed with mortar to prevent runoff from
entering or leaving. At the lower boundary of each plot a wooden
plank anchored a sheet metal apron that drained the runoff from the
plot (at soil surface level) to a funnel, through plastic tubing, and
into a modified oil-drum for storage.
Most sites were equipped with two tanks in series, so that
overflow from the first tank drained into the second. One site
(number 95, Pananao) required three tanks due to the high proportion
of runoff and low water storage capacity at the site. A modification
of the design also was required for plots with deep hillside ditches
cut across the slope. This was the case in the two plots planted to


Cement
blocks-
Fig. 12.
Illustration of erosion plot with runoff and
sediment collectors.


94
pigeon pea and sweet potato with minimum and normal tillage (numbers
82 and 83, Los Montones). The modified design illustrated in Fig. 13
is a combination of the standard plot and the Gerlach trough
collection device (Morgan, 1979). Due to the modifications both plots
at this site required three collection tanks.
3
The maximum installed storage capacity (0.5 m ) was determined by
cost and maintenance considerations. The use of a slot divisor to
split the sample prior to collection would have been ideal for
experimental purposes, but costly to produce. An increase in storage
3
capacity to 1 m also would have proven costly and difficult to
install.
The plots were installed during March and April 1980 and
collection began at most sites on 1 May. The runoff and sediment
collected in the tanks were measured, sampled and removed as soon as
possible after each rainfall event. _Nevertheless, many composite
results were obtained, as well as some results recorded for individual
events. Overflows were rare, so that the total rainfall and soil loss
measurements were not jeopardized by composite collection of water and
sediment for two or more events.
Procedures for field measurements, sample collection and
laboratory analysis. The procedures for measurement and sampling of
the collected runoff and sediment at all 16 sites followed standard
methods cited in the literature (Dunne, 1977; Djorovic, 1977; USDA,
1979). The total depth from the bottom of the tank to the water
surface (cm) and the depth of sediment deposits (if greater than 5 cm
in depth and uniformly distributed) were measured.
If a deep (>5 cm)


95
Fig. 13.
Illustration of alternative erosion plot design with three
subsections.


96
uniform sediment layer was present, one sample of sediment and two
water samples from the center and top of the tank were taken by
inserting closed 1- L bottles, opening them at the desired sampling
depth, filling them slowly, and then capping the bottles while still
in place (Fig. 14 ) .
Where sediment depth was uneven or less than 5 cm, the water and
sediments were mixed vigorously with a paddle to suspend all of the
sediment. Three samples were drawn, one each from the bottom, center,
and top of the mixture. The remainder of the water and sediment was
drained and mopped to leave the tanks clean and dry before the next
storm. Maintenance and equipment checks were performed also after
sampling.
Samples were sent in lots to the laboratory for analysis of total
solids by the same methods described above for river and stream
samples. Samples high in sediments were transferred to containers of
known weight, then were oven-dried and weighed, without filtration.
Determination of runoff, sediment volume and total soil loss.
Water and sediment depth measurements recorded were converted to
volume using a nomograph that relates depth and volume in the
cylindrical tanks, when positioned as in Fig. 14. The water volume
(Vw) was calculated as follows from total volume (Vt) and sediment
volume (Vs):
Vw(1) = Vt(l) (0.5 x Vs)(l).
This assumes that 50% of the volume occupied by the saturated sediment
deposits actually was filled by water.


55 gallon drum
Fig. 14.
Diagram
taken.
of sediment and runoff collector indicating the points at which samples were


98
The total sediment volume was read directly from the nomograph.
Total runoff for each plot was calculated by summing the water volume
for all the tanks on the plot.
Total soil loss was determined by combining the total volume and
sediment volume calculated above with the sediment concentration of
the samples. For the cases in which the tank contents were mixed
prior to sampling, the total soil loss (Ls) from each plot and for
each sample period was determined by the following equation:
Ls(tons) = Cs(tons m x Vt(m^),
where Cs = average sediment concentration of all three samples, and Vt
= total sample volume in tank.
In cases that required separate sampling of the sediment, the
concentration of the sediment sample (Csl) was multiplied by sediment
volume (Vs) and the suspended sediment concentrations were averaged
(Cs2) then multiplied by the corresponding liquid sample volume (VI)-.
The total soil loss was calculated as follows:
-33 -33
Ls(tons) = Cs(tons m ) x Vs(m ) + [Cs2(tons m ) x vl(m )],
where VI = Vt Vs. The program used for both calculations is
included in Appendix E.
Relationships between rainfall, runoff and erosion. Rainfall
data for the plot experiments were take from the same source as the
small watershed study areas in which the plots were situated. The
rainfall totals for the plot sampling intervals were calculated
separately for each case, since the sampling schedule was not uniform
within or between sites. Simple linear regression was used to test
and describe the relationships of total rainfall amount, total runoff,
and soil loss per unit area for each land use and management category


99
tested. Coefficients of erosion and runoff were derived for each land
use and agronomic practice. These factors represent ratios of soil
loss and runoff under a given land use or practice to the values
obtained for forest sites. The significance and relative strength of
the relationship between land use and management and runoff and
erosion losses were tested by analysis of variance (SAS, 1979).
Evaluation of the farm household models. Both the input-output
model and the systems model that were developed and evaluated during
phase two were updated and refined. Modifications were based on the
results of the runoff and erosion experiments and the final interviews
with the residents, including an inventory of the year's production
activities and inputs and outputs at the household level. The changes
in income, labor requirements and ratio of subsistence to commercial
production received special emphasis in evaluating the socioeconomic
aspects of changes in crops, land use, and conservation practices.
The physical impact of the proposed changes was also evaluated using
the empirical data from erosion plot measurements.
The final model was modified to demonstrate clearly the effects
of each proposed solution on the physical, biological, and
socioeconomic elements and their interactions within the system. The
model provided a common ground on which to consider the effectiveness
of proposed changes in environmental terms and their practicality and
acceptability within the conditions under which households must manage
their resources.
After modification, evaluation, and analysis of the household
models, the results and insights from the analysis, along with the
empirical data collected at that level, were extrapolated to the small


100
watershed model. The relationship between various land uses at the
watershed level was examined, using the model as a guide. The
repercussions of drastic changes in the proportion of any given land
use within the system were studied. The difficulty of determining
watershed management by a composite of household decisions constrained
by the larger system was illustrated by juxtaposition of the models at
both scales of analysis. A nested model was constructed to illustrate
the relationships of mutual causality between the systems of different
scale within the physical and socioeconomic hierarchy.


CHAPTER IV
RESULTS AND DISCUSSION
Regional Profile
Erosion and excessive storm runoff are serious problems within
the Sierra. Residents of the area are aware of soil loss and
depletion under shifting cultivation, bush fallow and continuous
cropping on the hillsides. Farmers point to declining yields as
evidence that the soil has "worn out". Many people from the area note
the disruption of river regimes and cite water quality and
availability as problems.
The condition of the land also speaks for itself. Erosion
features of almost every category are seen on the landscape. While
dramatic features, such as deep gullies or denuded hillsides, are not
common, there is widespread evidence of sheet and rill erosion. Many
areas also exhibit impoverished stands of natural vegetation which are
more subtle, secondary results of prior erosion. The varied
expression of the problem within the region reflects the physical and
cultural diversity of the area itself.
2
The Plan Sierra region includes 2500 km situated on the northern
slopes of the Cordillera Central. The population is approximately
120,000 (Chaney and Lewis, 1980) and the average population density is
-2
48 persons km While the density itself is not extreme, the region
has the highest ratio in the country of available labor force per unit
or arable land (SEA, 1978; Ferreiras, 1979). The full significance of
101


102
this relative measure of high rural population density becomes more
apparent when viewed against the backdrop of physical, biotic and
socioeconomic characteristics of the region.
Physical and Biotic Aspects
The landscape is characterized generally by rugged topography,
and pronounced intraregional variations exist in landforms, soils, and
parent material (Fig. 15) (Antonini and York, 1979). Climate also
varies dramatically, with average annual precipitation ranging from
1000 to 1800 mm (Jorge, 1970), influenced primarily by differences in
elevation from 100 to 1800 m (Antonini and York, 1979; Swedforest,
1980).
The drainage density of the area is high, due to the combination
of high rainfall and rugged topography. The project area contains the
upper and middle watersheds of three major rivers (Bao, Mao, and
Amina) which flow into the Yaque del Norte, the country's largest and
most important river (Fig. 16) (de la Fuente, 1976). Numerous
ephemeral and permanent streams of lesser magnitude are encountered
throughout the area. The discharge of both large and small streams
varies markedly, with flash floods occurring during the two rainy
seasons and many stream beds going dry during the midsummer and late
winter dry seasons. Interviews with local residents suggest that
deforestation has substantially exaggerated the normal range of
variation.
The majority of the soils of the region are classified as
Dystropepts or Eustropepts (Nicholaides and Hildebrand, 1980b)


PLAN SIERRA
Geologic Sub-Regions
,/v
' I
v
- eLv:?.)
a ^v,;
' feu'

O
CD
BH1

m
ra
o
RH3
IZI
Legend
Calcareous sands, clayey schists, limestone
Limestone, clay and some conglomerates
Tonalile tiornblende
Basic melamorphic volcanic rocks
Acid melamorphic rocks
Volcanic rocks and undifferentiated sediments
Conglomerates
Limestones
Sands and silty schists
Conglomerates
Limestone, conglomerates.and some volcanic rocks
Alluvium
o
5
i
10 Km3
Based on maps
published byO.A.S. 1967
Data compiled 1965-66
Modified by
Plan Sierra 1981
Fig. 15. Plan Sierra geologic subregions.


K ilomelers
Fig. 16. Plan Sierra region with study sites.


105
according to the Soil Taxonomy (USDA, 1975), although they vary
substantially in color, texture and chemical composition. As members
of the Inceptisol soil order they are primarily poorly developed
"young" soils, with development limited in most cases by the steep
slopes.
What most soils of this region have in common is their inadequacy
for many forms of agricultural use. They are poorly suited to the
semi-traditional land use systems that currently are found in the
Sierra. These soils are difficult to manage given their shallow
depth, low fertility, weak structure, and high susceptibility to
erosion by overland flow and mass wasting (Antonini et al., 1975;
Swedforest, 1980).
The natural vegetation reflects the combined variation of
geology, climate and soils (Fig. 17), while the actual vegetation is a
result of the natural vegetation further influenced by the intensity
of deforestation during the last 20 years, and by the character and
intensity of current land use. Land cover is a mosaic of forest,
coffee plantations, bush (secondary growth), small plots in mixed
field crops (conucos), and extensive areas in pastures.
Due to the highly "aggressive" climate, rugged topography, and
intrinsic properties of the soils, visible evidence of erosion is
widespread, and recovery from deforestation is slow (Swedforest,
1980). Many large land tracts are covered by eroded and/or compacted
pastures, or by secondary growth woodlands of poor quality.


*:: -V/r.{- Mon n inn'^-VvV/r. v:*. '?)s£:*C .-
vgv:V/.Sqn Jpseg^^ ;
^^cle Las Matasfe*^- .: /
i- ,1i;-,v.:.fc&: \v.",,;vv:'<
/ 1 1 V''"* '. ,#'.' *,' t* *i *
!|sl dmM\,,*:V VZ
Vi > M^feftfcUvc r0
I. .(
M
\
Legend
Life Zones:
(ZD bs-S
Subtropical dry forest
(S3 bh-S
Subtropical wet forest
H bmh-S
Subtropical very wet forest
M bh-MB
Lower Montane wet forest
03 bmh-MB
Lower Montane very wet
forest
o
L.
10 Kms
Based on maps
published by O.A.S. 1967
Data compiled 1965-66
Modified by
Plan Sierra 1981
Fig. 17. Plan Sierra life zones.
106


107
Socioeconomic Aspects
The population of the Sierra is characterized by a deeply
entrenched poverty manifested in several forms: high rates of
illiteracy (73% outside of the towns); high unemployment and
underemployment; 88% incidence of parasite infestation (SEA, 1978);
high incidence of malnutrition and other chronic health problems
(Ferreiras, 1979); depressed wage scale ($3-$4 R.D. day ^ for day
labor vs $7 R.D. day ^ in the Cibao Valley); low annual income (65%
-1 -1
earned $2 day or <$500 R.D. year per family in 1975) (Sharpe,
1977); high rates of emigration to Santiago, Santo Domingo and New
York City; and high level of dependency on remittances sent by
relatives working outside the region (Antonini and York, 1979; Pessar,
1981).
Basic services which are taken for granted in more urbanized
parts of the country, such as electricity, all weather roads, and
minimal health care and educational facilities, are limited primarily
to the towns and the areas along the connecting highways (SEA, 1978;
Ferreiras, 1979). Residents of the area, particularly women, cite
lack of fuel (firewood) and potable water as serious problems
(Jennings and Ferreiras, 1979; Safa and Gladwin, 1981).
Land tenure within the region reflects the situation in the
country as a whole. Most land is in medium-sized holdings, while the
remainder is divided into a few large tracts (some of which are state
land), and numerous small holdings (Ferreiras, 1979; Wilson, 1976).
Sharecropping is widespread.
The region is maintained in a subordinate position within the
national economy. The status of the economy, the distribution of


108
land, and the management of the region's natural resources are
inextricably linked. The area is predominantly agricultural and a
large percentage of the population is employed in subsistence
agriculture and/or coffee production (Antonini and York, 1979).
Yields of agricultural products are 21% below the national average
(Ferreiras, 1979), a condition which can be attributed in part to a
"hostile" natural environment with steep slopes, shallow soils of low
native fertility, and variable, unpredictable rainfall. Superimposed
on these conditions are a variety of land use systems which have
resulted in varying degrees of soil erosion, depletion of soil
fertility, and modification of microenvironmental factors such as soil
moisture and temperature.
Farming and Related Production Systems
The agricultural production systems of the Sierra are almost all
agroforestry systems (Fig. 18). They are mixtures of field crops,
commercial tree crops (coffee), pasture and forest (semicommercial,
successional, and "wildlands"). Although these categories are named
for the dominant vegetation, or the products of greatest value, each
refers to a matrix of subsistence and commercial production, including
a variety of food and cash crops. Combinations occur as rotations
over time or as associations (of crops and/or land use) in space.
They vary according to the demands of local, national, and
international markets, the subsistence needs of the farm family, the
availability of land, labor, water, energy and capital resources,
long- and short-term climatic constraints, tradition, and impact of
emigration.


Los
Los Coobas Quemodos
Sonllago
* Rodrguez
Santiago
Legend
Municipalities
Towns/Villages
1 Pine forest/Coffee/
Shifting agriculture
2 Field crops/Pasture/Palms
3 Dry forest/Field crops/
Pasture
10 Kms
Based on maps
published by OAS. 1967
Data compiled 1966
Modified by
Plan Sierra 1981
Fig. 18. Plan Sierra land use systems.
109


110
Field crops
Production of field crops tends to be limited to low-value crops
such as manioc (yuca), with some mixed subsistence and commercial
production of red beans, corn, cowpeas, peanuts, sisal and tobacco.
Intercropping is common practice and the plots are relatively small
(0.2 to 1.0 ha) with intercropping of two or more crops. Monoculture
is also practiced with tobacco, red beans, and sisal.
Field crops may have several distinct roles in the overall
production strategy of a given farm household. The plot may occupy
the entire landholding of a sedentary subsistence farmer who
supplements family income by raising other crops in more distant
forest lands, by sharecropping, or by seasonal day-labor. Field crops
often are found in areas predominantly covered by forest and pasture.
Shifting cultivation (slash and burn) is still widespread,
particularly in remote areas of higher elevation, where extensive pine
forests are found (Swedforest, 1980; Montero et al., 1981). The
practice also is common in the drier forests at lower elevations. It
deviates from the classic pattern (Greenland, 1974) primarily in its
long fallow (10 to 15 years).
This practice stops primarily because of lack of accessible
forests, in populous zones at 500 to 700 m elevation. The dominant
method more closely resembles the bush-fallow farming systems
described in recent West African research (Ruthenberg, 1976; Lagemann,
1977). Field crops are rotated with pasture and/or a 2-to-10-year
fallow in bush (secondary growth). Plots of this type may be
cultivated by the owner (in a small holding), or by sharecroppers,
renters, and/or hired help in large landholdings.


Ill
Field crops often also constitute the first "stage" in the
transition from forest, bush, or unimproved pasture to coffee or
improved pasture. Sharecroppers usually are engaged to clear and
cultivate the land for one or two years with an agreement to plant
pasture when they leave. Coffee plantations also may be started with
intercropping of bananas, plantains, other fruit trees, coffee plants
and some field crops (Martinez, 1979). Erosion features of varying
severity often are present in field crop plots. The most common forms
of erosion are rill and laminar (or sheet) erosion. In some cases it
clearly affects production, particularly in cases where sloping land
has been under continuous or near continuous cultivation for long
periods (up to 30 years). The problem is aggravated by the common
practice of planting row crops vertically on the slope (Antonini and
York, 1979).
Pasture
Large land tracts are maintained in poorly managed pasture by
absentee owners or by large landholders (Antonini and York, 1979;
Pessar, 1981). The limited milk and meat production from these
pastures is not available to most of the local population, and much of
it is marketed outside the region. Production is relatively low (300
kg ha *yr ^) (Antonini et al., 1975), but this form of land use
requires few inputs other than land, animals, and minimal labor and
management. Moreover, the product is always in demand and brings a
good price locally or in Santiago, so marketing is not difficult.
This form of land use has proven to be very efficient for large


112
landholders (Antonini et al., 1975) in terms of return on resources
invested (excluding land). It also has been widely used as a strategy
for holding land in speculation.
Soil degradation is one reason for relatively low production per
unit area. The shallow soils on steep slopes, often overgrazed and
compacted, form winding, irregular, impermeable "terraces" along the
well-worn cattle trails. Overgrazing in clay soil areas underlain by
nonpermeable bedrock also provides the necessary conditions for
landslips (Antonini et al., 1975). Continued grazing after slippage
retards recovery even further. Gully erosion is common.
Tree crops
Tree crops include bananas, plantains, coffee, and a variety of
palm products. Plantains and bananas often are intercropped with
coffee and are staples in most parts of the Sierra. They frequently
are intercropped with field crops and are found in dooryard (or patio)
gardens. The trees provide harvestable produce for about two years,
but the plants may remain in place for some time thereafter. Unlike
most tree crops, bananas and plantains afford little protection to the
soil, at least not under current management practice. Their principal
erosion prevention value lies in their longevity relative to most
field crops.
Coffee
Coffee is the most important tree crop in terms of number of
persons employed. It is one of the most widely distributed crops in


113
the Sierra, accounting for more than 5000 ha in 1978 (SEA, 1978).
More than 1900 farmers are engaged in coffee production, and the total
number of persons involved exceeds 10,000 (close to 10% of the
population) if their families are considered. The number of hired
laborers (estimated at 7000) puts the total at over 17,000. (This
estimate is based on figures from SEA and interviews with local
growers). This figure is not surprising since labor accounts for
approximately 80% of normal on-farm operating costs (excluding fixed
costs of land and infrastructure, and marketing costs).
The labor demand is markedly seasonal, and harvest, which is the
peak employment period, lasts two to three months. The laborers often
reside on or near the farm, supplementing their incomes by shifting
and/or bush fallow cultivation or other forms of temporary labor.
Some of the work force also migrates seasonally from the Cibao valley
or from more remote parts of the Sierra. This work force differs from
the usual one which is composed primarily of Dominican men in that it
includes women and children as well as some Haitian workers.
Coffee production also influences the local economy in a less
direct fashion. Traditionally, it has been carried out on medium
sized to large landholdings by families who support and supplement
this activity with adjacent holdings in pasture, field crops, and
forest. Many of the larger holdings are not contiguous; that is, one
household may have scattered large subdivisions of land in various
kinds of production up to two hours distance from the home. The large
farm owners often own and manage other small businesses such as coffee
processing and marketing operations, supply stores, informal credit,


114
breeding, sale and rental of work animals, and transportation of
people and goods to San Jose, Janico, and Santiago.
Fragmented holdings and diversity of land use also are
traditional strategies of smallholders, but within a different
context. Smallholder coffee production has been limited primarily to
small plots for home or local consumption or it has been a secondary
activity supplementing field crop production or off-farm employment.
The crop also is economically important at the national level,
being the export of second highest value (SEA, 1978). The Sierra is
the home of the Juncalito variety (Coffea arabica var Juncalito), the
best of the traditional varieties (SEA, 1978). There is a long
history of coffee production in some parts of the region. Yields are
low (230 kg ha ^) compared to international averages for commercial
production. However, these yields reflect the use of traditional
methods, without fertilizers, pesticides, technical assistance or
credit. Post-harvest losses due to transportation problems also
decrease the net yields.
Most of the established plantations are of the native variety,
but the Brazilian dwarf variety (var Caturra) has been propagated in
the region with great success in recent years. The coffee is
intercropped with plantains, bananas, fruit trees, and shade trees,
resulting in a diverse, three-tiered "forest" of sorts (Martinez,
1979).
The "imitation" of the adaptive mechanisms present in the natural
forest is striking. The planting of nitrogen-fixing leguminous shade
trees and the use of natural successional species, such as Cecropia


115
spp., facilitate recycling between soil and vegetation and retain more
nutrients within the stand.
Ecologically the coffee plantations are limited mainly to the
part of the region at or above 700 m. The principal coffee producing
areas are characterized by higher and less variable rainfall, lower
temperatures, high relative humidity, and a tendency toward heavier
(clay) soils. Erosion in established plantations tends to be very
low, in some cases approximating the erosion rates under forest
(Rocheleau, 1980). However, the system as a whole has erosion
problems.
The relative advantage for erosion protection is based upon the
length of the production cycle (about 20 years) and the excellent
protective properties of the established plantation, with shade trees
playing a critical role (Suarez de Castro, 1952; Suarez de Castro and
Rodriguez, 1955). The clearing, tilling, and planting required to
establish the plantation may cause extensive losses of soil, equal to
those under field crops, during the initial two years of the cycle.
The renewal of an older plantation also may cause varying degrees of
erosion, depending upon the method employed. In some traditional
operations the renewal is done on a continual basis and the shade
trees are left intact for two or three cycles of coffee planting, thus
reducing total erosion over the long term.
Another aspect will be treated in depth below: the almost
universal association of field crops and pasture with coffee
plantations, both within, adjacent to, and scattered around the main
holdings. Depending upon the functional relationships that exist, and


116
the researchers interpretation, much of the loss from these associated
land uses can be attributed to the coffee production system.
Palms
Palms, a common feature of the agricultural landscape in this
region, are important for wood, fiber, food, and animal feed at the
farm and household levels (Antonini et al., 1975; Safa and Gladwin,
1981). They also serve as important sources of income for rural
women. The fronds of a number of locally abundant palm species serve
as raw material for various forms of cottage industry. The ma]or
commercial uses include: thatch roofing material for tobacco sheds in
the Cibao Valley; woven seats for handcrafted wooden chairs produced
in the region; woven shipping containers for dried tobacco; woven
cargo bags for transport of local produce on burros and mules.
The latter two processes yield extremely low wages: $.50 R.D.
paid per pair of bags, at >12 hours labor invested per pair. In spite
of this, weaving persists because the work can be performed by women
and children in the home, with little or no requirements for
investment of land and capital.
The palms used for fiber often are located in pastures or fallow
land owned by neighboring families or relatives. There is free access
(a clear advantage) but no control over cutting or replacement; this
raw material supply is both free and uncertain. Reports by local
artisans indicate that supplies are dwindling in some areas of the
Sierra.
The weavers often complain about the lack of control over raw
materials and marketing, and most acknowledge the low return on labor


117
(Georges, 1981). However, many of them find no alternative and depend
on this source of cash income to supplement subsistence production.
The palm-pasture association is widespread in the region,
particularly in the mid-- to low-altitude zones. Erosion features
under this land cover appear to vary little from pasture without
palms. Given the wide spacing of palms within the association, it is
the quality of grass cover that should most affect erosion.
Forests
The Sierra's remaining forests, found primarily in the upper and
lower altitudinal tiers, are poorly managed and dwindling. In the
middle zone, forests are limited to small, isolated plots of pine
trees, successional stands in scrub, or ribbons of riparian forest
along major rivers and streams. Two large stands of second-growth
pine are found in recently acquired state lands spanning the mid- to
upper-altitudinal zones.
Forest resources are both underutilized and overexploited.
Legally, the residents are prohibited from felling any trees without
hard-to-obtain permits. Practically, they are engaged in the "mining"
of remaining forests for lumber, fuel, and expansion of agricultural
lands.
Prohibitive regulation of forest utilization leaves little
incentive for reforestation and offers no opportunity for wise
management based upon technical expertise, local needs, and external
markets for forest products. Subsequent raw material shortages have
contributed to unemployment. Many former artisans and sawmill


118
employees work the land in illegal slash and burn plots. Their jobs
were curtailed by rapid deforestation and subsequent poor management
of the land. At the household level, the current situation creates
worsening shortages of fuel and lumber, which in turn, increase
illegal poaching on existing forests. Thus, erosion, deforestation,
poverty, and unemployment continue to spiral, each reinforcing the
other.
The erosion rate under forest varies with the conditions and
current use of stands. Observation indicates a less effective canopy
and ground cover in the piedmont dry forests than in the denser upland
pine forests (Jennings, 1979b; Jennings and Ferreiras, 1979; Mercedes
Urena, 1980). Many of the dry forests actually were degraded
successional stands in areas formerly covered in humid-subtropical or
transitional forest (mahogany); the successional stands themselves are
results of prior erosion. The current condition and use of the dry
forests encourage the further acceleration of erosion. Numerous land
slips and landslides have occurred on steep limestone slopes and along
the roads that traverse the dry forest areas between San Jose de Las
Matas, Janico, and Santiago (Fig. 17). The same erosion features
appear with less frequency in the upland pine forests.
Thus, many forests have been subjected to fairly severe erosion
during prior deforestation (within the last 20 years). Lumbering,
charcoal production and slash-and-burn agriculture continue to
accelerate erosion beyond the rates one would expect to find in dense
forest.


119
Watershed protection
The area also functions as part of a national production system
and plays an important role as part of the watershed for the Cibao
Valley. The Cibao is a major food-producing region, second only to
the Santo Domingo metropolitan area in urban and industrial
development (de la Fuente, 1976). The Cibao yields a high proportion
of national agricultural production: plantains, 53%; manioc, 46%;
rice, 19%; processed tomatoes, 59%; red beans, 15%; cacao, 25%;
coffee, 33%; and tobacco, 90% (Jennings, 1979a; Swedforest, 1980).
At the national level the Sierra is more important as a water
catchment area for downstream irrigation reservoirs, flood control,
and hydropower production than as a direct source of agricultural
products. The present practices of shifting cultivation,
deforestation, and overgrazing of poor pasture lands on steep slopes
result in high sediment transport rates and disturbance of stream
discharge (Swedforest, 1980; Montero et al., 1981; Rocheleau, 1981)
both of which threaten the vital supporting role of the Sierra region
in the development of the national economy.
Model of the Sierra
The information presented above mirrors the conceptual model of
erosion, sedimentation and land use in the Caribbean (Fig. 1). The
Plan Sierra model (Fig. 19) expands the upland portion of the general
model, includes the effects of land use on runoff, erosion and
production, and illustrates the relationships between discharge,
sediment delivery and useful water supply in streams and reservoirs.


Fig. 19. System model of the Sierra.
A
G
H
JN
JO
Jr
L
I
LB
LC
LF
LP
LW
M
O
P
Q
R
S
T
U
V
Z
= animal biomass
= geological substrate
= high technology imports such as fossil fuels,
agricultural chemicals, also includes metals,
building materials
= weathering of substrate into soil
= total solar energy
= measure of solar energy available to
photosynthetic process
= land use
= land use-soil erosion coefficient
= land use-runoff coefficient
= land use-production coefficient
= bushfallow (relative area)
= coffee (relative area)
= field crops (relative area)
= pasture (relative area)
= forest (relative area)
= money (cash storage)
= water storage
= human biomass
= water storage in hypothetical reservoir
at stream outlet
= rainfall
= solar energy
= soil volume
= subsistence imports such as food, fuel,
clothing
= vegetation biomass
= sediment storage in hypothetical reservoir
at stream outlet


121


122
The interaction of population and trade with land use, biological
production, and erosion also is included in the more detailed regional
model.
The model is best understood by following the flow of energy
through the system from left to right. The same model is expressed as
a series of differential equations (Table 2) that can be solved
simultaneously to simulate system behavior.
The main energy sources that power the system are the sun (S) and
rainfall (R). Rainfall and land use (L) directly affect erosion,
runoff and production. Topography and soil type affect these
processes but are treated as constants within the model; they must be
varied for each site.
Rainfall energy and volume influence the erosion rate (K^) and
contribute to storm runoff (K ) and soil moisture storage (0). Soil
D
erosion and storm runoff are depicted as drains on soil (T) and soil
moisture storages, respectively. Soil storage is replenished by
weathering (JN) from parent material and exports soil through erosion.
Soil also provides a medium for plant growth and nutrient transport.
Soil moisture is fed by rainfall and exports water via storm runoff
(K ), percolation and sub-surface gravity flow (K ), evaporation
(K ), and transpiration (K ). Total soil volume is diminished by
1 o 1 z
erosion, which in turn directly influences storm runoff. Storm runoff
begins when the soil moisture storage equals soil volume. Thus,
reduced soil volume means an increase in runoff and a decrease in soil
moisture storage.


¡>H oO Q o NI 0< > oJh o( o op oS CU
Table 2. Equations for system model
= Jn K RT
Li
= R K5L2 K6 K12JrTpHL3 K182.
= K,
K^RT
L1
- K D(K OL + K,0)
3 5 6
[ (K OL ) + (K 0)]
5 2 6
= K RT K K RT
1L
1 1
+ K3K(K5OL2 + K60> KnQ
(K OL + K 0)
5 2 6
= K5OL2 + K6 K92 K105
= K16JrOTPHL3 K1?V K19aVLcPH K24bVAP K^VHA K^V
- K21VL0PH K22Y K2 3Y
= K_ _VHA K A K_ VAP K A
27 28 29 30
= K M K VHA K,_JrOTHP
33 26 15
K36UP + K35M
= K24aY + K31A + K41V K32M K34H K42M
= K36bUP + K37VAP K38P K39P q4JrPL3
o
Jr =
0
1 + K OTPHL V
16 3


124
The contribution of soil moisture to stream baseflow (K^) is more
constant but proceeds at a much slower rate than storm runoff. The
two combined rates constitute the surface water discharge (K^) from a
given area. During flood events, part of the transported soil (K )
forms deposits (D) in the stream channels; the remainder (K z)
reaches the reservoir (Q). A small proportion of the sediments that
reach the reservoir escapes in suspension (K ) with throughflow
(K ). If the reservoir is full, both water (Kn) and sediment K )
i u y o
bypass or overflow the dam.
With successive storm events, sediment deposits in the upstream
channels may be re-suspended (K^) an<3 transported downstream (K^).
The rates of deposition and re-suspension depend upon particle size
and streamflow velocity. In the model these flows are proportional to
the stream discharge rate (K ), and particle size is assumed to be
constant for the given area.
The sensor on land use (L) represents the influence of composite
land use in the region. Erosion, runoff and production are affected
separately, although the resultant flows interact. For example, a
high rate of erosion releases a large mass of sediment to surface
waters, while a high rate of runoff increases the sediment transport
capacity of the streams. Each land use type has a separate
coefficient for erosion, runoff and production. The composite land
use coefficient for each process (L^, i, L^) expresses the net effect
of combined land use.
The detailed land use model (Fig. 20) illustrates the most common
spatial associations and rotation sequences (Table 3).


L = Land area
L-W = Woodland
LF = Land in food and annual cash crops
L-C = Land in coffee
Lp = Land in pasture
l-B = Bushland
Lp = Population
125


Table 3. Land use systems in the region.
Common Rotation Sequencet
Production Systemt
Forest Field Crops (2) Forest (20)
Forest Field Crops (2) Bush (5-10) Field Crops (2) Bush (5-10)
Forest Field Crops (2) Coffee (20+)
Forest Field Crops (2) Pasture (1-10) Field Crops (2)
Forest Field Crops (20+)
Shifting Cultivation (S)
Bush Fallow (S, C)
Coffee Production (C, S)
Pasture (C, S)
Permanent Cultivation (S, C)
1"The duration (in years) of each stage of the rotation appears in parentheses.
Subsistence and commercial production are indicated by (S) and (C), respectively, in order of impor
tance .


127
Annual food crops precede establishment of coffee and pasture. Food
crop harvests defray part of the initial cost to clear, till and weed
the site. This actually represents a single land use system with an
overlapping sequence of multiple uses. This phenomenon is especially
important because of the low erosion and runoff rates under coffee
plantations, juxtaposed with the very high rates commonly attributed
to plots in field crops. With pasture, runoff rates are high but the
same contrast exists between erosion rates for the initial and
established phases of the rotation. Shifting cultivation in the
forest is another example of the spatial and temporal association of
two types of vegetation with markedly different rates of soil loss and
runoff.
The production (K ) of vegetation biomass (V) is powered by
sunlight, with inputs from soil moisture (K^)/ soil (T), labor (K.^)
and high technology or industrial sectors (K14). The land use affects
the process (L^) but does not directly contribute energy or materials.
The drains from the vegetation tank include detrital losses (K ),
transpiration (K ) and harvest (K K K__, K._) by human (P)
1 / 19a 2bb 2b 4(j
and animal (A) populations. Transpiration and detrital losses are
proportional to biomass.
The harvest is a sum of four separate flows. The coffee harvest
depends upon the proportion of total land in coffee (L ), the total
primary production as indicated by biomass (V), labor (K-¡.9b^' and high
technology inputs (K^q). The food and fuel harvests for local
subsistence are a product (K^^) of human population (P) and total
plant biomass (V). Grazing animals take an amount (K ) proportional


128
to plant biomass, capital inputs (K^g) and animal population. Farmers
sell some food and fuel (K ) to lowland markets. The amount sold
depends on total plant biomass and the price of the products in the
marketplace.
The growth in the number of beef and dairy cattle is determined
by the interaction of the initial herd size (as biomass, A), high
technology inputs for improvement of herd and pasture (K ), and
available biomass for grazing (K ) The energy and material exports
from the animal population include energy expended in self-maintenance
(K^g), slaughter and milking for subsistence consumption (K^g), and
sale of animals, meat and milk to outside markets (K^). Local meat
and milk consumption depends upon the population sizes of both animals
and consumers. The amount exported varies directly with the size of
the herd and the current prices of these products in external markets.
The population (P) of the area depends upon initial population
size (P) and growth rate, consumption of plant and animal biomass
within the region (K^^), and imported subsistence goods (U) (K^^).
Drains on population include energy expended for self-maintenance
(K ), labor (K ), death (K ), and emigration (K ). Seasonal
j> / u jo
migration may bring farm workers into the region from other areas, or
it may be a mechanism for local residents to earn cash income outside
the region (K ).
The imported subsistence goods are purchased with money (M) that
represents the total liquid assets available to the local population.
Income consists of profits from coffee other cash and food
2 jb
crop sales (K^)f animals and animal products (Kg.), and remittances


129
and earnings of permanent and seasonal emigrants from the region
(£49). Cash outflows include purchase of subsistence items from
outside the area (K34)# Pechase of fossil fuels and other industrial
inputs such as fertilizers, pesticides, machinery and metal products
(K32), partial export of wages paid to seasonal laborers (K^), and
outside investments (K4 ), usually in Santiago, Santo Domingo, or New
York City.
The stock of imported subsistence goods (U) is a net result of
purchases made from outside markets (K^) and consumption by local
residents (K^^). The fossil fuel stores and other industrial inputs
likewise are dependent upon the rate of purchase from external sources
(K ) and the rates of usage in crop production (K._), cash crop
J j lb
processing (K^), an<^ animal production (K^)-
Regional Subdivisions for Further Study
The study area was subdivided by watershed boundaries in
descending order from the Bao, Amina and Mao Rivers into small
2
watersheds of 1-20 km and then into individual landholdings (Fig.
16). The selection of the small watershed sites indicated a bias
toward the central portion of the Plan Sierra impact area. Each major
watershed also is subdivided by life zone and land use.
The most densely populated and ecologically diverse area within
the Sierra is the center band of the three-tiered life zone and land
use subdivisions (Figs. 17 and 18). It is in this region that changes
in land use and resource management can have the greatest impact.
Deforestation is almost complete and much of the area has


130
been converted to pasture. The areas within this zone that fall
within the higher elevation and rainfall ranges are targeted by Plan
Sierra for promotion of coffee plantations. The drier, more densely
settled and intensively cultivated lower elevation areas are major
targets of soil conservation and annual cropping system programs. Two
major dams will be constructed within or immediately downstream of
this central zone within the next five years (Listin Diario, 1981).
The more detailed descriptive and experimental research focused
on this central subregion. The land use associations based on pasture
and field crops, and coffee, respectively, were singled out for more
in-depth study within the area described above.
Study of Large Watersheds
The Mao and Amina Rivers drain approximately 65% of the Plan
Sierra impact area. The watersheds of both rivers contain a complex
mosaic of topographic, geologic, climatic and land cover subdivisions.
All of these variables tend to follow the altitudinal gradient in
east-west bands across the watersheds. The rivers run from south to
north.
Description of Mao and Amina Watersheds
The major differences between the two rivers are the size and
shape of the watersheds, which influence both the amount and temporal
distribution of discharge. The Mao watershed extends well beyond the
Plan Sierra boundaries into the rugged forested slopes within the
Bermudez National Park. This forest reserve is covered in pine and


131
mixed forest. Soils on the steep slopes are very shallow. Rainfall
in the upper Mao watershed is higher than in the central Sierra. The
bimodal rainfall distribution peaks in October and to a lesser extent
in May. This regime contrasts with lower elevations where the first
peak occurs in May and the second in September (Jorge, 1970;
Swedforest, 1980).
The drainage network geometry makes the discharge of the Mao
River extremely responsive to rainfall. Flood crests occur less than
24 hours after peak rainfall (Swedforest, 1980). The time of
concentration for each separate tributary is such that storm runoff
from several subwatersheds reaches the Mao almost simultaneously.
This causes extreme flood stages and wide fluctuation in river
discharge.
More moderate rainfall and gentler slope in the smaller Amina
watershed yield slightly lower rates and volume of runoff than the Mao
watershed. Since a small portion of the catchment is influenced by
rainfall in the high Sierra, the river regime closely resembles
typical rainfall patterns of the Plan Sierra impact area. The flood
peaks on the Amina usually occur in May and are less frequent but tend
to be more extreme than the peaks in the Mao River discharge. Average
monthly river discharges are illustrated in Figs. 21 and 22.
The Mao has a greater sediment transport capacity and a greater
erosion potential as indicated by topography and erosivity of rainfall
(Paulet, 1978). Stratification of land use is roughly parallel in
both watersheds (Fig. 18). The major difference is in the proportion
of coffee and forest in the upper reaches. Coffee is far more


OISCnfiRGE M3 SEC
O
Fig. 21. Monthly discharge of the Mao River.
132


DISCHARGE IN H3 SEC
Fig. 22. Monthly discharge of the Amina River.
133


134
prevalent in the upper Amina watershed, while forest predominates in
the Mao headwaters.
The large watershed model is basically the same as the Plan
Sierra model (Fig. 19). In effect, the large watersheds are parallel
subsystems that exhibit the same general responses to variations in
rainfall, land use, trade and other variables. The major differences
are in characteristics such as rainfall, slope and soil type, which
are treated as constants within the model. The structure remains the
same. Variation of these constants would cause differences in degree,
not in kind of response by the model.
Other major differences are the particular crops and the specific
rotation sequences within the major land use types previously defined.
The pasture-field crops association is common to the central portion
of both watersheds, while coffee production and forest cover are less
evenly distributed in the respective headwaters. However, both
watershed models have the same structure with respect to land use. It
is the values used in the model that would change.
Specific combinations of annual crops also vary from one
watershed to the next. Sweet and bitter manioc (Manihot spp.), white
sweet potatoes (Ipomea batatas), corn (Zea mays), and pigeon peas
(Cajanus cajanus) are common in the lower and central altitudinal
belts in both watersheds. Red beans (Phaseolus vulgaris) and
plantains (Musa paradisiaca) are more common in the middle to upper
Amina watershed, while peanuts (Arachis hypogaea), tobacco (Nicotina
tabacum) and bitter manioc as a cash crop are more typical of the
drier middle reaches of the Mao basin. These differences had little


135
effect on the structure of the model but were reflected in the local,
evaluated models for small watersheds and plots.
Comparison of the Precipitation and Discharge During the Study Period
with Period of Record
Rainfall in the Mao and Amina watersheds during the 15-month
study period showed a sharp peak in the spring relative to the mean
for the period of record, as illustrated by data from Moncion and San
Jose de Las Matas (Figs. 23 and 24). Figures D-l through D-9 in
Appendix D compare the mean monthly rainfall from 1967 through 1979
with the values from January 1980 through June 1981 at nine other
stations. A 13-year record was not sufficient by itself as a baseline
for comparison with the study period (USDA, 1979). However, the mean
monthly rainfall for the 13-year period of record compared well with
the 50-year means available for the San Jose de Las Matas and Moncion
climatological stations (Figs. 25 and 26). All monthly means for the
13-year period were less than one standard deviation from the 50-year
means. The close correspondence between the means of the two periods
justified the use of the shorter record available for the other
climatological stations and for the Amina and Mao Rivers.
Rainfall data for both 1980 and 1981 exhibited a typical bimodal
distribution at 10 of the 11 stations (see Appendix D, Figs. 1-9).
The first (spring) peak was high in comparison to both other spring
peaks and the second (fall) peak for the same year. Even at stations
that normally have two rainy seasons of equal magnitude (see Appendix
D, Figs. 8 and 9), the May maxima were sharply accentuated for 1980
and 1981. Aside from this month, the 1980 and 1981 monthly rainfall
values closely approximated the means for 1967 to 1979.


o
O n
03
O
(\J
Fig. 23. Monthly rainfall at the San Jose de Las Matas climatological
station, #1.
136


PRECIPITATION [N MM
160 240 320 400 460 550 640 720 600
Fig. 24. Monthly rainfall at the Moncion climatological station, #2.
137


150 240 320 400 480 350 540 "20 500
1931-197*3 IM i 11
196 7-1979 fill III
Fig. 25.
Monthly rainfall at the San Jose climatological station, #1, for
two periods.
138


o
o
o
o
CJ
I
3*
ir
o
(O
II)
z:
o
IZ CD
v- rT
£1
-4 o
l o
h-
i
o.
*- o
O rvj
UJ
o:
ol
1931-1979 nmn
1967-1979 u 111
Fig. 26. Monthly rainfall at Moncion climatological station, #2, for
of record.
jnn
two periods
139


140
Average monthly discharge rates in the Amina and Mao Rivers in
1980 and 1981 reflect the irregularity in the rainfall distribution.
The Mao River deviated from its characteristic October peak in both
years due to the magnitude of the spring floods (Fig. 21). Baseflow
during the January to April dry season was very close to the 1967-79
means, but the 1980 May discharge was more than double the mean
monthly value. The discharge rate fell off sharply after June, but
streamflow remained at levels a little above the norm through October
of the same year.
The same extreme rainfall affected the Amina. The May discharge
rate exceeded the mean value by more than a factor of three (Fig. 22).
As in the case of the Mao, the discharge rate decreased rapidly but
did not reach baseflow levels before the onset of the August rainy
season.
The May rainfall in 1980 and 1981 falls between one and two
standard deviations from the mean for almost every station. The 1981
Moncion value differs by more than two standard deviations from the
mean for 1967-79. The mean for 1931 to 1979 is slightly higher but
the 1980 and 1981 values are still greater than one standard deviation
from the mean.
Results from May for both years must be regarded as exaggerated
relative to average conditions. Rainfall and discharge for the rest
of the study period were well within the average range. Predicted
annual totals and averages based on measurements from this period
should be adjusted downward accordingly.
Rainfall distribution over the region for the study period and
for the 13-year record is illustrated in isohyet maps compiled by


141
Surface II programs using nearest neighbor analysis (Figs. 27 and 28).
The distribution for the study period showed no substantial variation
from the mean annual rainfall map.
Rainfall Distribution and Rainfall/Discharge Relationships
The allocation of rainfall data from the climatological stations
to the subwatersheds (Fig. 7) provided a basis for comparison of total
rainfall, rainfall rates, and ratio of discharge to rainfall volume in
both basins. The monitored portions of basins consisted of the areas
upstream of the hydrometric stations and the future dam sites on both
rivers (Fig. 8). Subsequent references to the watersheds will include
subwatersheds M. through M. for Mao and A_ through A_ for Amina unless
lo 3 5
otherwise stated.
Comparison of the monthly rainfall and discharge for both the Mao
and Amina watersheds (Table 4) showed a relatively high rate of
discharge. Total annual discharge is more than 50% of rainfall
volume. Reports from other tropical and subtropical watersheds
indicate much lower proportions of total rainfall discharged in
streams in forested watersheds (Golley et al., 1975; Odum, 1971;
Pereira, 1973).
The high ratio in this case can be attributed to the interaction
of topography and land use. The extensive deforestation that has
occurred in the past 20 years has exposed the thin soils on steep
slopes to a highly "aggressive" climate (Paulet, 1978) resulting in
erosion and reduced soil moisture storage capacity. Much of the area
(30-40%) is in pastures. Soils have been compacted in many areas by


Fig. 27. Plan Sierra region: Mean annual rainfall for 1967-1979.
142


Fig. 28. Plan Sierra Region: Annual rainfall for 1980.
i
143


144
Table 4. Rainfall and river discharge in the Amina and Mao watersheds.
Mao
River
Amina
River
Month
Year
Rainfall
Volume
(m3 106)
River
Discharge
(m3 106)
Rainfall
Volume
(m3 106)
River
Discharge
(m3 106)
April
1980
261
30
74
23
May
1980
253
147
173
101
June
1980
155
137
41
39
July
1980
47
79
27
17
August
1980
111
64
43
35
September
1980
157
80
81
36
October
1980
149
99
48
45
November
1980
67
47
24
16
December
1980
141
34
50
31
January
1981
85
43
58
23
February
1981
103
35
14
24
March
1981
90
23
46
19
April
1981
122
26
75
27
May
1981
324
248
178
104
June
1981
126
133
41
32
Annual Totals
4/80 3/81
1619
818
679
409
7/80 6/81
1522
911
685
408


145
overgrazing, further reducing infiltration and storage capacity of the
soils. Comparison of monthly rainfall and discharge data reflect the
limited soil moisture storage capacity. Rapid reductions in Amina
river discharge from May to June and in Mao river discharge from June
to July indicate that most of the surface and subsurface runoff
reaches the main rivers soon after major storm events.
Rapid runoff of rainfall represents a loss-of useful water for
upland agriculture and for downstream industrial, domestic and
irrigation uses. This same phenomenon increases the sediment delivery
efficiency of streams and contributes to high flood peaks and
floodplain damage downstream.
Subwatershed Analyses
The subwatersheds within the larger basins do not contribute
equally to river discharge and sediment load. The linear regression
of rainfall rates in each subwatershed on discharge and sediment
transport'- in the respective larger watersheds shows a wide variation
2
in the R value between subwatersheds (Table 5). It is the variation
in relative strength of relationship between subwatersheds that is
most important to note. The proportion of explained variation is low
for all cases, since same-day rainfall on each subwatershed is only
one of many factors that determine river discharge. Comparison of the
2
R values indicates which subareas contribute most to the effects
observed downstream.
All of the variables tested were Poisson distributed and were log-
transformed (In) for the regression analyses.


146
Table 5. R values for regression analyses of subwatershed rainfall
vs. river discharge and sediment load.
Subwatershed
Rainfall
Area
R : Rainfall
vs. Discharge
R^: Rainfall
vs. Sediment Load
_1
mm year
km^
Amina River
A3
1227
70.0
0.09
0.09
A4
1473
110.6
0.06
0.10
A5
1719
156.9
0.10
0.05
MclO
River
M1
1240
95.6
0.02
0.24
M2
1268
126.9
0. 07
0.21
M3
1268
103.1
0.07
0.28
M4
1598
116.9
0.06
0.02
M5
1598
98.1
0.06
0.02
1598
96.9
0.06
0.04


147
Amina
The summary of results in Table 5 shows differential
contributions by subwatershed to river discharge and sediment
transport in the Amina River. The daily rainfall in subwatersheds A^
and A^ consistently show a higher correlation than A<- with average
sediment concentration, maximum sediment concentration, and sediment
2
transport. By contrast, A^ (forested) has the highest R value for
rainfall versus Amina River discharge. This indicates that A
contributes more to discharge and less to sediment load than either A^
or A^. The larger size, rugged topography and higher rainfall in A
(Table 5) could easily account for the stronger relationship to river
discharge. The contributions of A^ and A^ to the sediment load, by
contrast, are best explained by land use. Both areas are largely
deforested. A^ has a high proportion of land in food crops and
pasture land and A^ is planted to coffee in the upper reaches with
extensive tracts of pasture and food crops in the downstream portion.
The lower rainfall and gentler topography in A^ and A^ suggest that
land use is the major factor contributing to variation of sediment
load contribution by subwatershed.
These preliminary analyses point to A^ as the subwatershed that
contributes most to sediment load. This is consistent with the
observations of project staff and survey participants who identified
the Inoa watershed (A ) as the area most disturbed by deforestation
within the larger Amina watershed. An increasingly erratic
fluctuation of river stage and higher sediment loads have been
observed during the last 20 years.


148
Mao
The relationship between Mao River discharge and rainfall on the
respective subwatersheds is relatively uniform (Table 5). While the
upland watersheds (M^, M^, and M^) receive the highest rainfall and
have a higher average slope, daily rainfall in M^ and M^ shows an
equivalent or slightly stronger relationship to river discharge. The
larger size of the M^ and M^ watersheds probably accounts for this.
The results for M^ are distinct in part because it includes the
influence of the Mao (Station 5) rainfall data. This is based on the
Thiessen polygon distribution and reflects a transition from humid to
dry subtropical forest zones.
2
The R values for rainfall versus Mao sediment transport suggest
a high contribution from M^, M^, and M These watersheds have lower
annual rainfall and have higher monthly maxima for May than the upland
areas. The first intense spring rains on freshly cleaved and tilled
soil have the greatest erosion potential. This same area has
extensive tracts of pasture and annual crops for both commercial and
subsistence production. The lower three are covered primarily in a
pasture-palms-feld crop association that includes commercial
production of bitter manioc (Manihot spp.), peanut (Arachis hypogaea),
tobacco (Nicotina tabacum), sisal (Agave spp.) and thatch and baskets
for lowland markets. Land cover in the upper three areas includes
forest, pasture, and subsistence agriculture (usually small slash and
burn plots).
However, the extent to which the difference can be attributed to
land use is limited by another consideration. Sediment delivery ratio


149
decreases with distance, so at least part of the effect can be
attributed to distance. On the other hand, the distance between
, 2
and does not warrant the order of magnitude difference in R
values. Moreover, M the farthest upstream, actually has a higher
value than and M^. The most striking contrast between the two
groups (M M_, M_., and M., Mr, M-) is in land use.
IZ Abb
The M subwatershed shows the greatest relationship between
rainfall and sediment load in spite of its relatively small size. In
addition, this area will contribute sediment most directly to the
proposed dam. For these reasons it will receive priority for more
detailed analyses.
Comparison of land use, rainfall, and sediment delivery set
priorities by subwatershed for further research, extension and land
reclamation projects. Such indicators are useful where sediment
concentration and discharge data are not available for subunits of
larger river basins. Results of these analyses are presented to
identify areas of critical concern. Causes of variation between
subwatersheds are suggested.
In both cases daily rainfall on the separate subwatersheds and on
all of the watersheds combined explains less than 12% of the variation
in daily river discharge (Table 5). These same-day comparisons mainly
show the role of rapid storm runoff from various subareas in
determining the river discharge and sediment load. This simple linear
model of daily variation does not account for prior conditions of
discharge, sediment deposition, or sediment concentration in the
rivers, nor does it deal with gradual runoff by gravity flow from


150
hillslope soil moisture storage. Non-linear models and/or linear
models applied to longer time periods could better explain the
interactions and contributions not directly attributable to rapid
storm runoff. However, the variable contributions of rapid storm
runoff from the subwatersheds was a useful indicator for the purposes
mentioned above.
-Relationships between land use, discharge and sedimentation are
best discussed in more detail, and on a firmer basis, after a review
of the results from studies in small watersheds and erosion plots.
Further discussion of the river studies focuses on the results of
sediment concentration measurements, the relationships of river stage,
discharge and sediment concentration, and the prediction of
sedimentation rates in proposed dams on the Mao and Amina Rivers.
Sediment Concentration and Sediment Rating Curves for Amina and Mao
Rivers
The changing color of river discharge with the source and
magnitude of upstream storm events is observable in both watersheds.
Local residents at river crossings and in river-bank settlements note
the variation in stage and color and comment upon these as upon other
elements of the weather that enter into daily conversation.
Quantitative data from other areas (Gregory and Walling, 1973; Rapp,
1977; USDA, 1979) confirm the observation that sediment concentration
(related to color) changes with river discharge (observable at river
stage). The regularity and quantitative predictors of that variation,
however, are more difficult to discern.
During the study both rivers were observed and sampled regularly
under conditions ranging from clear water and low flow in March to the


151
nearly opaque, brown, debris-laden waters of flash flood peaks in May.
The sediment concentrations in both rivers were very regular during
dry weather. Suspended sediment concentrations in Mao during March of
1980 and 1981 ranged from 0.01 to 0.20 g L ^ (see Appendix E).
Concentrations for the Amina under dry weather, low flood conditions
were between 0.02 and 0.10 (see Appendix E). These findings are
consistent with results reported for streams draining forested
watersheds in the southeastern United States, where average sediment
concentrations ranged from 0.02 to 0.20 g L ^ (Duffy et al., 1978;
Reikerk et al., 1979; Switzer and Nelson, 1972).
The maximum concentrations measured for Mao River were 11.04 g
-1 -1
L in May 1980 and 2.9 g L in May 1981. The peak concentration for
the Amina was 3.5 g L \ These concentrations are consistent with
rising stage flood flow sediment concentrations of 2.0 to 3.5 g L ^
reported for the Morogoro River in Tanzania. The Morogoro catchment
is much smaller but is located in an area of similar topography and
land use.
Total sediment transport rather than concentration varies with
discharge (e.g., Likens et al., 1977). Sediment concentration is
especially likely to be uniform in areas where fluctuation in river
stage is gradual and the range of variation is relatively narrow.
However, in areas where river stage (and therefore velocity) varies
abruptly over a wide range, the sediment-carrying capacity of the
river also changes dramatically over a short time (Gregory and
Walling, 1973). If the available sediment from runoff and in channel
deposits is sufficient, then suspended sediment concentrations may


152
climb rapidly with the rising stage of the river and decrease
accordingly with the falling stage. Moreover, in wide, shallow river
beds the increased turbulence during the initial phase of the rising
stage may bring much of what is normally bedload into pseudo
suspension, moving it more quickly than usual to points downstream.
This portion of the total sediment load, or part of it, also may be
measured as suspended sediment in samples from medium and lower depths
in the cross-section.
Both the Amina and Mao Rivers demonstrated a sharp rise in
sediment concentration with the rising stage, in some cases showing
peak concentration prior to the flood peak (Fig. 29). The latter
phenomenon is probably an example of pseudo-suspension induced by
turbulence, which decreases somewhat as the stage continues to rise.
The rising stage concentration generally is higher than the falling
stage concentration for the same level. This may be explained by the
greater turbulence and velocity of the rising stage (Herschy, 1978)
and by the availability of different types of sediment for suspension
and transport at the onset versus the recession of a flood (Rapp,
1977) .
The derivation of a stage concentration curve for general
application requires a larger number of samples from a broad range of
conditions. Simple linear regression analysis of stage versus the
natural logarithm of sediment concentration for all sample days
provided a quantitative measure of the nature and degree of the
underlying general relationship between stage and concentration.


SEDIMENT CONCENTRATION
153
Fig. 29.
Time series of sediment concentrations during selected
flood events.


154
Several combinations of variables were tested. Average daily
discharge rates, morning, evening, and floodpeak stage recordings (at
the hydrometric station), and stage readings from the bridge at time
of sampling were all used as independent variables to explain sediment
transport (tons day 1), average sediment concentration (g L ^) and
maximum sediment concentration (Table 6). The regression of stage
reading at sampling time on average sediment concentration explained
the largest portion of the variance.
The morning and evening stage readings are not adequate
indicators of stage at sampling time. Average daily discharge is
derived from these readings, so it also fails to account for within-
day variation of stage. River stage at sampling time, by contrast,
accounts for 38% of the variation in average sediment concentration
(mean concentration per sample set) in the Mao River, and the same
variable explains 21% of the variation in sediment concentration in
2
the Amina River. Higher R values and F-values for daily sediment
discharge and river stage at time of sampling (Table 6) are
misleading, since sediment transport is derived by multiplying the
average sediment concentration by average daily discharge. The
discharge in turn is derived from morning and evening stage
measurements. The 70 and 43% explained variations in sediment
concentration for Mao and Amina, respectively, are partially artifacts
of autocorrelation. The clearest and most straightforward
relationship that emerges from the analyses is the one first cited
above. The resultant equations for the two rivers are given below.
Amina:
C =
s
e
(-0.0154 x L ) + 1.0218)
(1)


Table 6. Summary of regression analyses of river discharge and sediment concentration.
Y
X
n
2
R
Ft
Amina River
L
n
(average sediment concentration)
River level
41
.21
10.62**
L
n
(average sediment concentration)
L (daily discharge)
n
41
.19
9.24**
L
n
(sediment discharge)
River level
41
.43
29.32***
L
n
(sediment discharge)
(daily discharge)
41
.59
56.05****
L
n
(sediment concentration maximum)
River level
51
.12
6.68*
L
n
(sediment concentration maximum)
(daily discharge)
51
.16
9.06**
Mao
River
L
n
(average sediment concentration)
River level
39
. 38
22.43****
L
n
(average sediment concentration)
L (daily discharge)
n
43
.24
12.92***
L
n
(sediment discharge)
River level
43
.69
93.11****
L
n
(sediment discharge)
(daily discharge)
43
.69
93.11****
L
n
(sediment concentration maximum)
River level
55
.32
24.73****
L
n
(sediment concentration maximum)
(daily discharge)
66
.10
6.97**
^The significance level of the model
0.001, and **** for 0.0001.
is indicated by for 0.
05, **
for 0.01,
*** for
155


156
where Cis average sediment concentration (g la ^ ), and (cm) is
stage measured at the bridge as distance from the reference point to
the water surface.
Mao:
r (-0.1423 x L ) + 0.1601
= e f (2)
The equations are both significant to the 0.005 level and can be
used to predict sediment concentration and to derive sediment
transport estimates based on the available stage readings for both
rivers. The best predictions will be obtained with continuous stage
recordings from continuous monitors. Such data are available at times
for the Mao River. It is strongly recommended that maintenance of the
continuous monitor be improved at Mao and that a similar instrument be
placed at the Amina station. However, the equation can also apply to
the morning and evening spot readings. The resultant sediment
concentration predictions simply will have the same inherent
limitations as the current discharge calculations routinely based on
the stage readings. In the absence of such continuous data the total
annual sediment transport for the study period was calculated on the
basis of the twice daily stage readings. These were converted to L
equivalent values by the following equations for Mao and Amina,
respectively.
L = (L 2.559)/-0.0150 (Mao) (3)
Jl X
and
Lf = (Lx ~ 2.471 )/-0.0097 (Amina) (4)
where is the stage reading at the gauge for morning, evening or
floodpeak recordings. Daily sediment transport was calculated by


157
multiplying the average sediment concentration (C ) given in the
s
3
previous equations (1, 2) by total daily discharge (Q) (m ). The
latter is derived from stage recordings (L ) by the following series
of equations:
If L < 0.54, then Q lo X
^ ^ r ^ ^ ._{C(log,-L ) + 0.8243/0.487}
If L > 0.54 and < 1.21, then Q = 10 ^10 x
x
^ ^ ^ {[(log L ) + 1.28883/0.727}
If L ^ 1.21, then Q = 10 ^10 x
x
for the Mao.
s ~ ^ _{[(log, _L ) + 0.3473/0.201}
If L < 0.64, then Q = 10 y10 x
x
It ^ ,r _{C (log. -L ) + 0.4173/0.303}
If L > 0.64 and < 1.15, then Q = 10 ^10 x
x
T, T >> ,c {C (log L ) + 0.5473/0.402}
If L ^ 1.15, then Q = 10 10 x
X
(5)
(6)
(7)
(8)
(9)
(10)
The estimates of total sediment transport given by the above
equations (5-10) represent conservative predictions (Table 7) because
they do not include the multiplicative effect of high discharge and
high sediment concentrations during short-lived flood events. The
additional estimate of peak one-hour sediment transport during floods
increased the predicted annual sediment yield by greater than 10% in
all cases.
The sedimentation rate of the dam to be built on the Mao River
will exceed the reported rate of 190,000 tons yr 1 for the Tavera Dam.
The useful life of the latter is expected to be severely curtailed by
sedimentation problems. The calculated rate indicates a need for
immediate action to avoid repetition of the Tavera case. The
197,888 tons yr-1 calculated yield for 1980 (without flood peaks)
includes the high rainfall and discharge rates for May of that year.


Table 7. Sediment transport in Mao and Amina Rivers.t
Month
Year
Mao River
Amina River
Sediment
transport
Sediment transport
during peak hour
of flood
Sediment
transport
Sediment transport
during peak hour
of flood
tons
April
1980
1,772
499
May
1980
45,596
7,971
14,370
13,218
June
1980
40,651
1,684
3,431
0
July
1980
17,477
1,341
1,044
0
August
1980
13,358
1,298
3,745
0
September
1980
18,134
3,442
3,389
0
October
1980
24,402
4,473
5,646
1,205
November
1980
8,966
528
1,029
0
December
1980
5,981
259
2,882
314
January
1981
8,820
108
1,905
0
February
1981
6,958
654
2,185
0
March
1981
3,545
45
1,966
346
April
1981
4,901
509
2,993
836,47lt
May
1981
110,683
15,964
17,853
4,944
June
1981
45,961
2,889
2,422
0
Annual Totals
4/80 3/81
193,888
21,803
43,364
15,236
7/80 6/81
269,186
31,510
47,059
843,280§
fCalculated sediment transport based on previously cited river stage/sediment concentration
regression equations from 41 days of data and over 60 individual data points.
$See textvalue adjusted to 23,000.
§6,809 tons plus single flood.


159
However, even if the peak month was removed the total would exceed
150,000 tons. A preliminary analysis combined the continuous stage
recordings for the Mao in 1977 and the sediment concentration data
from March through May of 1980. The estimate of sediment yield
derived from these indices (Rocheleau, 1981) very closely approximates
the final results of the study. Annual sediment yield was estimated
at 3.0 tons ha ^ for the Mao watershed, with a total sediment load of
200,015 tons yr ^ transported past the sampling site (Table 8).
The Amina River sedimentation rate is approximately 22% of the
rate for the Mao. The monthly sediment transport values generally are
more uniform, but the rare flood peaks in May can radically alter the
totals. Flood peak contributions are more important for the Amina
than for the Mao, but the frequency of occurrence is much less. The
real effect of the extreme events is difficult to predict accurately
because the stage discharge curves and equations (INDRHI, 1981) are
not calibrated for the upper range of the scale. The sediment
transport value for 28 Apr. 1981 is an example of this problem. The
flood was determined to have a return period of at least 50 years,
having exceeded the flood stages from the 1979 hurricanes and every
other event witnessed by residents for the past 60 years. While very
high discharge and sediment concentrations are to be expected for such
an event, the discharge rate of 1800 m^ sec ^ and the sediment
concentration of 135 g L 1 exaggerated the estimate of sediment
transport. A reduction of the sediment concentration by a factor of
approximately 10 (to the maximum of 11 g L 1 recorded during the
3 -1
study) and a decrease in discharge by a factor of 3 (to 600 m sec )


160
Table 8. Sedimentation in the Mao River basin estimated from May
1980 measurements.
Annual Average
Base Flow
Flood Stage (>30 cm)*
Discharge
19.4 m^/sec
3
c10 m /sec
3
peak 500 m /sec
sustained for days -
100 mVsec
Sediment
concentration
0.006 g/Z
->2.000 g/Z
~0.022 g/Z
1000->2000 g/Z
Volume of
sediment transport
200,015 t/yr
3330
151,373 t/yr
Sediment concentration
per unit area
3 t/ha/yr
0.05 t/ha/yr
2.26 t/ha/yr
*Refers to flood stages substantially above baseflow levels, as
recorded at INDRHI station.


161
would reduce the sediment transport to approximately half of the
annual total which is at least plausible for a 50-year flood.
The severity of sedimentation problems in the proposed dam on the
Amina will depend upon the storage capacity of the dam and also on the
occurrence of extreme events such as the one discussed above. The
addition of the latter to the annual total would bring the sediment
transport to 74,900 tons for 1981. The importance of accounting for
such occurrences has been demonstrated in the case of the Valdesia Dam
which was choked with sediment by the runoff from Hurricanes David and
Frederick in 1979 and required dredging to resume operation. The
sediment load in the Tavera Dam also was substantially increased by
sediments transported during the same flood events. A large
proportion of the sediments transported during such events are re
suspended from channel deposits. They represent the cumulative
deposition of prior erosion. For this portion of total sediment yield
the best treatment is a reduction in the amount and discharge rate of
storm runoff and/or the construction of upstream barriers to keep
subwatershed sediment yield from entering the main stream and reaching
the dam. The reduction of the normal sediment yield is better
accomplished by changes in land use and practices upstream, starting
in the watersheds of critical concern identified earlier.
The average sediment yield of the large watersheds gives some
basis for subsequent comparison within and between watersheds. The
sediment yield for Amina is lower than for Mao by a factor of 2.4
(Table 9). The values of 3.1 and 1.3 tons ha 1 yr 1 exceed the
reported yields for forested watersheds in the U.S. (Table A-2) with


162
Table 9. Water and sediment yields estimated from 1980 and 1981
data.
Mao River
Amina River
Total discharge (m yr )
3 -1 -1
Water yield (m ha yr )
Sediment transport (tons yr )
Sediment yield (tons ha 3 yr 3)
818 x 106
13 x 103
197,888.0
3.1
409 x 106
12 x 103
43,363.0
1.3


163
the exception of clearcut machine-prepared forest tracts that yielded
20 tons ha 1 yr ^ (Hewlett and Nutter, 1969). The yields in Mao and
Amina also exceed the average yields for grassland (0.85 ton ha 1
yr "*") in the U.S. and the sediment yields for Malaysian watersheds in
vegetable farms (1.0 ton ha ^ yr 1).
The erosion losses reported for similar land use in Africa (Table
A-9) and in Jamaica and Colombia (Table A-7) range from 5 to 125 tons
ha 1 yr 1 for grasslands and crops (Lai, 1977a; Okigbo, 1977; Temple,
1972; Sheng, 1973) and from 0.6 to 40 tons ha ^ yr ^ in coffee and tea
plantations (Table A-10) (Lai, 1977a; Suarez de Castro and Rodriguez,
1955). A sediment delivery ratio of 10% applied to this range of
erosion rates could easily account for the sediment yields in Amina
and Mao.
The results from the Sierra also compare very well with reports
from Tanzania (Rapp, 1977) where sediment yields consistently exceeded
the rates reported for catchments in areas of comparable topography in
the western U.S. (Schumm and Hadley, 1961). The higher rates (2.0
-1-1 2
tons ha yr for a 650 km watershed) in the Tanzanian examples have
been attributed solely to land use but the explanation of the Mao and
Amina yields, as well as the Tanzanian case, is more likely an
interaction of high rainfall intensity during short peak rainfall
periods and a land use pattern that augments both runoff and erosion
. 1
rates.
The high sediment delivery ratio also could be attributed in part to
the unusually high volume of residual channel deposits resulting from
extreme rates of watershed and channel erosion during hurricanes David
and Frederick in 1979.


164
The high sediment yields are important to the upland system since
they represent net losses from the resource base for forestry and
agricultural production. The further analyses of the problem rest on
a more detailed study of land use, sedimentation and erosion in small
watersheds and plots located within the subwatersheds targeted for
immediate action.
Study of Small Watersheds
2
The small watershed sites range in size from 1 to 30 km and
represent examples of land use systems dominated by coffee and field
crops, and pasture and field crops, respectively. A brief description
of each site is followed by the results of the detailed land use
analysis and field surveys..
Description of Watersheds in Coffee
Two small watersheds in the upper reaches of the Bajamillo River
(Fig. 30) were compared with each other and with the larger Bajamillo
watershed. The detailed land use analysis was confined to the smaller
watersheds, but general field reconnaissance and map interpretation of
land use included the entire Bajamillo watershed.
The upper Baiamillo miniwatershed (Fig. 30, No. 71) was monitored
at Rincon de Piedras, and Prieto Stream (Fig. 30, No. 64) was
monitored at Carrizal. Both settlements are coffee-producing
communities located at 800-900 m elevation on the slopes of Angola
Mountain (Cerro Angola). There is little variation between the two
watersheds in climate, soil and slope. Land use differs with respect


165
LEGEND
" R ivers
0
Plots
A
Watershed Monitors
Watershed
Boundary
T owns
Fig. 30. Small watersheds in coffee region.


166
to the relative amounts of pasture and cropland, but the percent area
of land in coffee is approximately the same. The physical and land
use characteristics of both sites are summarized and compared with the
other small watersheds in Table 10. The distribution of land use
categories (Figs. 31-33) emphasizes single crops on types of
vegetative cover. However, each mapped unit represents a combination
of uses named for the dominant type in a finely subdivided pattern.
The analysis and calculations of land area (Table 10) accounted for
the proportion of diverse land cover types in each mapped unit.
The larger Bajamillo watershed is included in Table 10, although
the detailed land use analyses were not made at this scale. This
mesoscale watershed includes the association of food crop and pasture
rotations downslope with coffee plantations at higher elevations (Fig.
30). The Bajamillo watershed also contains several settlements and
roads that fall outside the perimeter of the two small watersheds, yet
are important to coffee production within their boundaries.
5
Coffee production in Carrizal and Rincon de Piedras is 1.8 x 10
kg annually. Of this amount, approximately 25% is produced within the
upper Bajamillo (No. 71) and about 30% is grown in the monitored
segment of the Prieto watershed (No. 64). Some of the stands were
established over 60 years ago. The coffee produced in this part of
the Sierra is of the traditional variety, although many growers now
are planting the Brazilian dwarf variety.
Coffee acreage and production are increasing in this area. The
resurgence of coffee production is due to the high international
market prices of 1978 and 1979 and to the extension and credit


Table 10. Characteristics of the small watersheds.
Stream
Bajamillo
Bajamillo headwaters
Prieto
Pananao
Hondo
(no. 70)
(no. 71)
(no. 64)
(no. 60)
(no. 67)
Settlements
Rincon de Piedras,
Rincon de Piedras
Carrizal
Pananao
Los Montones/
Las Piedras,
Carrizal
San Jose
Area (ha)
2962.5
95.0
187.5t(490.0)
1312.5
1285.0
Average slope
37%t
43%
48% (42%)
32%
31%
Life zone
Subtropical very
Subtropical very
Subtropical
wet forest
wet forest
very wet forest
Land use system
Coffee-food crops
Coffee-food crops
Coffee-food
Pasture-
Pasture-
crops
food crops
food crops
% Area in:
Coffee
20§
39.9
48.4 (36.6)


Plantains^
5
12.9 ( 5.9)
Field crops
30
15.0
20.4 (27.4)
15.8
15.7
Pasture
30
30. 7
13.8 (19.9)
46.4
41.3
Bush fallow
4
9.9

19.0
18.6
Forest
10
4.0
4.5 (10.2)
18.6
23.2
Roads and
houses
1
0.4
0.5 ( 0.5)
0.5
1.1
Population:#
Residents
125
225 (625)
1100
1250
Seasonal
workers
40
60


fvalues in parentheses refer to larger Prieto watershed and Carrizal community. The first values refer to
the monitored portion only. ^Estimate. §The land use distribution for no. 70 was estimated. All others
were measured. ^Refers to plantains and/or bananas. #Refers to watershed boundaries, not to the political
subdivision. In the upper Bajamillo and Prieto only a part of the community is included.
167


Crops
Forest
0
500 1000 1500 2000
Meters
Fig. 31
Land use in the Prieto watershed


169
O
I
500
1000 1500 2000
Meters
Fig
32
Land use in the larger Prieto watershed


CROPS
as
FOREST
V'V PASTURE
sa
fil
l?&v COFFEE
BUSH FALLOW
0
500
1000
1
. 1

1
Meters
t
Fig. 33. Land use in the Upper Bajamillo watershed.
170


171
programs of Plan Sierra. Nurseries at Rincon de Piedras, Carrizal and
Las Piedras produce over 1,000,000 coffee plants annually for sale to
local growers. While fruit and wood producing trees also are grown,
coffee production has predominated at these nurseries, and coffee
growers have been the major clients of the related special credit
programs.
A major goal of Plan Sierra is to change the current land use
distribution (Table 10). The policy developed for Carrizal, Rincon de
Piedras and similar areas has focused on increasing the proportion of
land area planted to coffee. The latter is motivated by the double
objective of economic development and watershed protection. The
further description and analysis of the two small watersheds and the
larger Bajamillo watershed explore the socioeconomic and ecological
aspects of existing coffee production and related land uses within
these nested study areas. A conceptual model of the land use system
(Figs. 34 and 35) illustrates the functional relationships- between
different types of land cover in the watershed. The feasibility and
advisability of proposed land use changes in these areas are discussed
in this context. Carrizal (Prieto stream) serves as the main example,
while Rincon de Piedras is a replicate case that confirms or
supplements information from Carrizal.
Biophysical characteristics
The watershed appears to be forested from a distance because of
the broadleaf canopy provided by shade trees in the coffee stands. At
closer range, the steep slopes (about 45%) still appear to be well


172
Fig. 34.
System model of small watershed in coffee region.


Cleormg ond Cultivation of Lond
Outside of Wotershed
Ll= land oreo inside watershed
LO= lond orea used outside watershed
LW = woodland
L B = bushlond (secondary growth)
LF= food cropped land
LC= coffee land
LP = posture lond
P= population
Fig. 35. Land use model for watershed model of coffee
producing region.
173


174
protected by the combination of litter, live ground cover, coffee
plants and leguminous shade trees. Many of the stands in Carrizal are
old and well established, which explains their forest-like structure.
In the newer stands, however, the red-brown sandy clay soils are
exposed. Visible signs of erosion, such as rills and gullies, are
apparent. This is more pronounced in the numerous plots of field
crops along the road, in the cleared dirt surfaces around the houses,
and on the roads themselves.
Soils generally are very stony .and difficult to break and till.
Reports indicate that fertility is not a problem in coffee stands.
Coffee and other tree crops appear healthy and residents describe
yields as adequate to good by regional standards. Field crops vary
more in terms of yield and condition, both from one plot to another
and from one year to the next. Given the rainfall, solar insolation,
temperature and soil characteristics of the site, plantains and
bananas fare better than annuals and constitute the staple foods in
this area.
Observation of the downstream portions of both small streams
during flood events revealed much higher sediment loads in the runoff
draining the settlement and its access roads than in runoff from the
coffee stands. Gullies along the roadsides drain into the Prieto
stream and the Baiamillo River at crossings along the approach road to
Carrizal (Fig. 30). In both cases, extensive deposits of sediments
matching the roadbed material have built up at the mouths of the road
drains and gullies. These deposits later are resuspended during heavy
floods and contribute to the scouring capacity and the sediment load
of the Bajamillo.


175
The Prieto stream channel cross-section is narrow and "U" shaped,
and the longitudinal profile is quite steep (about 30-40% slope). The
stream bed is covered with boulders and worn, rounded rocks
(metamorphic) of 5 cm to 0.5 m in diameter. The stream banks
generally are more exposed than in natural riparian forests because
natural ground cover has been removed to the waterline, where
possible, in favor of coffee trees. This contributed to rapid runoff
and increased scouring capacity of the stream.
Inspection of the upstream portion of the Prieto channel showed
some effect from footpaths along the streambanks but very little
deposition was apparent. The Bajamillo upstream of Rincon de Piedras
is similar to Prieto stream. Channel erosion and deposition of
sediments in the streambed are more apparent, but the characteristics
and distribution of streamflow are similar.
The headwaters of both streams normally are clear and are sources
of community drinking water. During the field survey, Rincon de
Piedras residents reported changes in color and sediment load in
Bajamillo stream following clearing, burning, and herbicide
applications in upstream plots recently planted to coffee by one of
the large landholders. The perceived threat to the water supply
provoked a formal protest by the community to the president,
demonstrating awareness ofthe relationships between land use and soil
and water quality within the small watershed unit. The incident also
illustrated the need to combine resource management with agricultural
extension. Widespread application of herbicides was promoted at the
plot level by agricultural extension personnel without regard for
water management criteria at the small watershed scale.


176
While the two small upland watersheds in coffee yielded a fairly
regular discharge of clear water upstream of the settlements, the
associated settlements and roads substantially increased flood peaks
and sediment discharge in the lower reaches of these streams. They
also contributed to disturbance of the river regime and the high
sediment load of the lower Bajamillo. This illustrates the difference
between streams draining coffee stands and those draining whole coffee
production areas. In the latter case, coffee production affects
runoff and erosion indirectly through the impact of market roads,
settlements, and related food production plots. As a result, the
Bajamillo resembles streams draining more intensively cultivated
watersheds in pastures and food crops.
While the upper Bajamillo and Prieto channels showed little
evidence of flood peaks beyond 1.5 m above the river bed (see Appendix
F), the channel erosion features and debris in the lower Bajamillo
indicated the recent occurrence of flood stages up to 3 m above the
usual base flow level. Sediment deposits ranging in texture from fine
to coarse sand were up to 10 cm deep in places along the stream
channel and floodplain.
Socioeconomic characteristics
Some aspects of economic interactions in the study area are
clearly expressed in the landscape by land use and land tenure. The
landholdings in Carrizal are dominated by two families that hold
approximately 50% of the land. The family with the largest
landholdings produces coffee as well as staples for the household and


177
the coffee workers. On the same holdings they also produce food crops
for sale to neighboring families, beef cattle for sale to lowland
markets (Santiago), and teams of oxen for household use and for hire
throughout the larger Bajamillo watershed. Aside from the holdings in
Carrizal, various members of the immediate family own and maintain
coffee, pasture and food crop plots in adjacent watersheds and in the
drier areas at lower elevation just downstream from the Prieto
watershed.
Most of the coffee produced within the watershed and on the first
family's other holdings is processed at its depulping and drying
facility. The beans are sold whole and dried by the sack to
commercial buyers and exporters in the lowland urban markets. The
owners of the driers and trucks have first priority for processing,
transporting and marketing of their own harvests. Then the harvests
of nearby smallholders are purchased according to what the market
demands. At Rincon de Piedras the proportion of small and middle
range landholders contributing to the commercial processing facility
is larger.
In years when the prices are very low the large growers may not
even harvest the crop. If the harvest brings a net loss or even a low
profit, the larger growers sell more of their cattle to maintain cash
flow in the household.
Most of the households of the medium and large growers have
family members in New York. The major landholders have invested in
establishing their emigrant children in New York. The emigrants, in
turn, send remittance money to their families. Subsequent local


178
reinvestment is usually for home improvements, consumer goods,
vehicles, land, cattle, or establishing a local business. The money,
as a rule, is not channelled into increased coffee or food crop
production. The net result, as documented in Juncalito (located in
the upper Bao watershed), has been a decrease in food production and
an increase in the proportion of land held in pasture (Pessar, 1981).
Purchase of new lands with remittances is more often for speculation
than for increased production.
The effect of this change on the landless and small landholders
is a decrease in local employment opportunities (on food crop plots in
large landholdings), an increase in staple food prices, and an
increase in consumption of foods imported from the lowlands by local
entrepreneurs. This depression of food and forest production in favor
of pasture is reflected in Rincon de Piedras where a large percentage
of the land is in pasture and bush (Table 11). Carrizal has been less
affected because the head of household from the largest landholding
resides more than half-time in the area and continues to manage the
food crop plots as an integral part of his holdings.
Many of the residents work on larger landholdings as
sharecroppers, coffee harvesters, or day laborers. Seasonal migrant
farm workers also come to the area to work at harvest time. They
include residents of lowland areas nearby as well as jobless people
from the cane fields in the more distant lowlands.
Most of the residents produce some staples, especially plantains,
in small plots and depend upon wood and charcoal for cooking, light
and heat. The latter is a minor consideration in the Sierra, but is


Table 11. Physical characteristics of erosion
plot sites.
Plot number
80
81
82
83
Site
Carrizal
Carrizal
Los Montones
Los Montones
Slope (%)
35
37.9
44.0
46.5
Soil Classification
Tropudalf
Eutropept
Troporthent
Troporthent
Soil Texture
Soil Fertility
Silty loam
Silty loam
Sandy loam
Sandy loam
% N
0.30
0.09
0.31
0.31
% OM
6.02
1.82
6.16
6.16
Infiltrationf
4.6
3.5
20.7
20.7
Vegetative Cover
Conservation Practice
Established
coffeej
New coffee
Pigeon pea§
Sweet potato^
Hillside ditches
Pigeon pea
Sweet potato
Hillside ditches
Minimum tillage
fAmount in cm infiltrated during first hour. Totals include lateral leakage from soil test column and are
intended as relative measures of infiltration capacity.
|Coffea arabica.
§Cajanus cajanus.
Impomea hypogaea.
179


Table 11.Continued
Plot number
84
85
86
87
Site
Los Montones
Los Montones
Los Montones
Los Montones
Slope (%)
34.3
34.3
25.5
34.0
Soil Classification
Troporthent
Troporthent
Troporthent
Eutropept
Soil Texture
Soil Fertility
Sandy loam
Sandy loam
Loamy sand
Sandy loam
% N
0.29
0.29
0.29
0.29
% OM
5.86
5.86
5.86
5.90
Infiltration
36.6
36.6
10.9
Pasture
48.5
Vegetative Cover
Conservation Practice
t
Manioc, Corn
Red beans^
Hillside ditches
Manioc, Corn
Red beans
Pine forest
fManihot spp.
$Zea mays L.
§Pangola.
flPinus occidentalis.
#Phaseolus vulgaris.


Table 11.Continued.
Plot Number
88
Site
Los Montones
Slope (%)
28.0
Soil Classification
Eutropept
Soil Texture
Sandy loam
Soil Feritlity
% N
% OM
0.18
3.50
Infiltration
53.5
Vegetative Cover
Pine forest
Conservation Practice
tAgave spp
89
91
92
Los Montones
Pananao
Pananao
25.5
34.0
32.0
Troporthent
Troporthent
Troporthent
Sandy loam
Loam
Loam
0.15
0.15
0.20
2.94
3.02
3.98
8.63
11.6
7.6
Pasture
Sisalt, Sweet
Sisal
potato
181


Table 11.Continued.
Plot Number
93
94
95
96
Site
Pananao
Pananao
Pananao
El Rubiof
Slope (%)
35.0
31.0
39.0
30.0
Soil Classification
Troporthent
Troporthent
Troporthent
Troporthent
Soil Texture
Loam
Loam
Sandy loam
Sandy loam
Soil Fertility
% N
0.26
0.17
0.30
0.16
% OM
5.18
3.36
6.02
3.22
Infiltration
24.5
24.5
37.3
2.5
Vegetative Cover
Pasture
Pasture
Manioc
Pine forest
Conservation Practice
fThis nearby site serves as a forest surrogate for Pananao since no suitable forested areas remained in
Pananao.


183
more important in this cool, wet region. Kerosene is used also for
lighting, but more sparingly. Most of the household energy needs are
met by firewood gathered from nearby riparian forests, coffee stands
and more distant patches of pine forest. Charcoal is more expensive
and is used primarily in the larger towns or by the most well-to-do of
the families in areas such as Rincon de Piedras and Carrizal.
3
Assuming an average of 1 m per person per year, the Prieto watershed
must produce 225 m^ of firewood per year.
Electricity is generated by a small waterwheel and generator
installed on Prieto stream. This equipment recently replaced a diesel
generator that became too expensive to run due to high fuel costs.
Power from the Carrizal installation may be extended to other homes
and stores in the near future. However, it currently serves the
residence and store of the major landowner. This unit was the only
functioning hydroelectric installation in the region, but has sparked
the interest of other landholders (individuals and groups) situated in
similar terrain.
Description of Small Watersheds in Pasture and Food Crops
Pananao and Hondo streams drain areas of similar size and land
use within the lower altitudinal strata of the central subregion.
Pananao drains into the Mao River and Hondo feeds the Inoa River,
which empties into the Amina (Fig. 16). The predominant land uses are
pasture and field crop production (Figs. 36 and 37), although the
emphasis is quite different in the two cases. Both areas are more
recently settled than Carrizal and Rincon de Piedras, and both have


Bush follow
Crops
Forest
Posture
Urbon
Fig. 36.
Land use in Hondo watershed.
184


185
Fig. 37. Land use in Pananao watershed


186
gone through a 20-year cycle of deforestation, slash and burn
agriculture, bush fallow agriculture and establishment of pasture and
small food crop plots under permanent or almost permanent cultivation
(Table 10). The functional aspects of the current land use system are
illustrated in Figs. 38 and 39.
Biophysical characteristics
The erosion features along the banks of both streams indicated a
relatively wide range of floods stages (see Appendix B). The channels
exhibited a high rate of sediment deposition, ranging from fine and
medium textured sands at Pananao to coarse sand in the Hondo channel
and floodplain. The deposits reached 10 cm in thickness in some
reaches of both streams. Interviews with residents indicated that
both streams rise and fall very rapidly, after spring and autumn
rains. The spring flood events are flash floods that last from 0.5 -
2 hours, leaving large deposits of relatively coarse sediments in the
lower reaches of the channel. Both streams rise to peak annual flood
stages up to 2 m above the normal baseflow, with ten-fold and higher
increments in discharge.
Residents at both sites pointed out that water quality and stage
fluctuation have changed considerably over the years. Most families
in Pananao send the children on burros to fetch water at other,
cleaner streams. They travel up to 5 km from their homes rather than
use the poorer quality water from nearby Pananao stream. When the
valley was first settled the Pananao provided drinking quality water.
The change in quality is probably a combined effect of polluted runoff


187
Fig. 38.
System model of small watershed i,n pasture-field crop association.


Fig. 39. Land use model of small watershed model of pastures-field crops association.
188


189
from the settlements (Fig. 40) and sediment deposits from deforested
croplands. The same is true of Hondo stream to a lesser extent, since
settlements are clustered on the ridges (Fig. 41).
Socioeconomic characteristics
Both watersheds, to varying degrees, reflect the influence of the
shift from subsistence to commercial production, and show the impact
of emigration and land speculation. The latter is evident in the high
proportion of land in pasture (Table 10, Figs. 36 and 37) held by non
residents or part-time residents with homes in Santo Domingo and New
York. The watershed of Hondo stream is affected primarily by a
combination of subsistence farming, large tracts of overgrazed
pasture, and small plots planted to manioc, beans, corn, and pigeon
peas for sale in local and regional markets. Most of the remaining
forest tracts are owned by large landholders from Santiago who hold
pasture and forest land for speculation and/or as family vacation
homes.
Pananao has been characterized by smallholder commercial
production of cash crops, food crops, dairy products and finished
products for regional and lowland markets (Figs. 37 and 38).
Dependence on the production of tobacco, peanuts, and bitter manioc
has carried the community through boom and bust cycles over the last
30 years. Farmers are plagued by highly unpredictable rainfall and
price fluctuations. Soil depletion and destabilization by tobacco and
peanut cultivation and by deforestation also have contributed to
decreasing yields and declining quality of produce.


190
Fig. 40. Pananao watershed.


Fig. 41. Hondo watershed.
191


192
The large landholders own tracts of land in pasture rather than
coffee. Like Carrizal, much of the money for expansion of previously
medium to large landholdings has come from remittance income and later
from investments in local business (stores) or from diversified local
production initially financed by remittance (Georges, 1981; Martinez,
1981).
The full-time residents at Pananao supplement cash income and
subsistence goods from agricultural production with day labor outside
the community. The large landholders also maintain tracts of owned or
rented pasture in nearby areas with more humid microclimates.
Sediment Transport in Small Watersheds
The concentration of stream discharge sampled in the five
-1 -1
watersheds varied from 0.80 g L to more than 200 g L
Concentrations during spring floods often ranged from 10 to 100 g L .
Similar concentrations were reported by Rapp (1977) for semiarid
catchments in Tanzania. The nature of the sampling apparatus limited
regular samples to flood events. The transport of sediments in these
watersheds is episodic and the samples measured infrequent but
significant contributions to the sediment load in the respective
larger rivers.
The discharge, sediment concentration, sediment transport, water
yield and sediment yield associated with each sampled event in all
five watersheds are given in Appendix F. The discharge, sediment
concentration, sediment transport, water yield and sediment yield for
low flow conditions are peak flood events for each watershed are
listed in Tables 12, 13, and 14.


Table 12. Sediment transport in five watersheds
Average low-
flow sediment
Sediment transport
under low-flow
Estimate of
sediment transport
Watershed
Area
Baseflow
concentration
conditions
Storm runoff
for peak
flood
ha
-1~
m yr
-1
g L
tons yr*1
tons ha 1yr ^
3
m
yr
-1
tons
tons ha
Prieto
187.5
2.52 x106
0.14
353
1.9
1.10
x
10s
286
1.5
(#64)
Upper
5
*
Bajamillo
95.0
1.14 x10
0.15
171
1.8
0.62
X
10
76
0.8
(#71)

Hondo
1285.0
4.10 x106
0.13
533
0.4
4.73
X
io5
875
0.7
(#67)
Pananao
1312.5
3.78 x106
0.14
529
0.4
7.41
X
105
1465
1.1
(#60)
Bajamillo 2962.5
(#70)
22.39 x106
0.20
4478
1.5
17.31 x 105
4090
1.4


Table 13. Discharge rates measured for low flow conditions in small watersheds
Watershed Area Rainfall Discharge rate
ha
-1
mm yr
total
per i
unit area
Prieto (#64)
187.5
1719
0.08
3
m
-1
sec
4
X
io4
3 ~K ~l
m sec ha
Upper Bajamillo (#71)
95.0
1719
0.04
3
m
-1
sec
4
X
o
I1
3 ~K ~1
m sec ha
Bajamillo (#70)
2962.5
1719
0. 71
3
m
-1
sec
2
X
io4
3 ~K ~1
m sec ha
Pananao (#60)
1312.5
1240
0.12
3
m
-1
sec
1
X
H
O
3 ~K -1
m sec ha
Hondo (#67)
1285.0
1227
0.13
3
m
-1
sec
1
X
O
rl
3 ~K _1
m sec ha


Table 14. Flood events yielding peak sediment discharge during the study period.
Watershed
Date
Average sediment
concentration
Discharge rate
Duration
Total suspended
sediment transport
Sediment
yield
-1
3 -1
-1
g L
m sec
hrs
tons
tons ha
Prieto (64)
10-22-80
61.3
2.6
0.5
286.0
1.53
Upper
10-22-80
17.6
2.4
0.5
76.0
0.80
Bajamillo (71)
\
Hondo (67)
10-22-80
30.8
7.9
1.0
874.5
0.68
Pananao (60)
5-24-80
92.7
4.4
1.0
1465.2
1.12
Bajamillo (70)
10-10-80
64.0
17.8
1.0
4090.0
1.38


196
The watersheds covered in pasture, food crops and annual cash
crops showed consistently higher concentrations than the small upland
watersheds in coffee. The amount of discharge in the pasture/cropland
watersheds also was greater, further increasing total sediment
transport. The sediment concentrations in the larger Bajamillo
watershed were in approximately the same range as Hondo and Pananao.
The total discharge and concentration in each case provide an
estimate of the total sediment transport for a given flood event.
However, the relationship of sediment transport and discharge to land
use only can be evaluated in terms of the area of the watershed, and
the contribution per unit area. These values are listed in Table 12
and show a pronounced deviation from the rank ordering of the
watersheds by sediment concentration and volume of flood discharge.
The sediment yield for low flow conditions (Table 12) at four of
the watersheds shows a distance decay function for sediment transport,
based on watershed size. This is typical of the pattern observed by
Roehl (1962) for watersheds in the U.S. The larger Bajamillo stream
deviates from this trend. Its relatively high yield is probably due
to high rainfall and the high proportion of clay particles in the
suspended sediment load. Fine particles remain in suspension for
greater distances from the source than sands. Hondo and Pananao
watersheds are characterized by coarse soils and sediment loads.
The sediment yields for flood events differ substantially from
the general pattern observed for low flow conditions. The ranking of
watersheds by sediment yield during peak annual floods cuts across


197
watershed size categories and other groupings based on site
characteristics and land use (Table 14). The peak floods that
occurred in the Hondo and Pananao streams during the study period
transported more sediment than the amount exported in a full year's
discharge under low flow (non-flood) conditions. In these two cases,
the texture of the eroded material, the rainfall regime, and land use
interact to produce high pulses of sediment transport. This contrasts
with the more even discharge of sediments in the Bajamillo watershed
(Table 12).
The Prieto watershed also exhibited a high peak flood sediment
yield relative to the other watersheds and to its own sediment yield
for low flow conditions. The high sediment export during this flood
can be attributed to the combination of an intense storm in October
1980 (54 mm rainfall) with the recent weeding and tilling of coffee
and food crop plots on the hillslopes upstream.
In spite of the contrast in sediment concentrations during floods
and the pronounced differences in the temporal distribution of
suspended sediment export, the sediment yields per unit area are
relatively uniform for all five watersheds. A graph of two
hypothetical distributions (Fig. 42) illustrates the export of
equivalent amounts of sediment (per unit area) in watersheds
characterized by very distinct regimes of discharge and suspended
sediment export. The areas under the curves are equivalent and
represent suspended sediment export. The first case exhibits the
pulsed export observed at Pananao and Hondo and to a lesser extent at
the Prieto watershed (Tables 12 and 14). The peaks in sediment


DISCHARGE m/unit time
a. more pulsed
distribution
b. more even distribution
TIME
Fig. 42. Illustration of equivalent sediment discharge according to distinct regimes.
198


199
discharge would be characterized by a combination of high discharge
and high sediment concentration. The magnitude of constant export
under low flow conditions would be determined primarily by the volume
of discharge, since the suspended sediment concentrations are low and
relatively uniform for all of the watersheds, outside of flood events.
The availability of sediments for suspension and transport during
low flow and flood conditions is determined by an interaction of
conditions in the watersheds and the stream channels. The
characteristics of overland flow reaching the stream depend upon
current runoff and erosion rates in the watershed. The nature and
condition of the stream channel, including sediment deposits, also are
largely determined by prior runoff and erosion rates. These rates
must be established for each watershed before the discharge and
sediment export regimes can be more fully discussed.
The significance of the differences between watersheds and
watershed groups (by land use and size) was tested by analysis of
variance. The watersheds were compared with respect to sediment
concentration, water yield and sediment yield.
Significant differences were found between sediment
concentrations in the five watersheds (Table 15). The larger
Bajamillo watershed, Pananao and Hondo were included in the higher
sediment concentration category as determined by Duncan's Multiple
Range Test (SAS, 1979) (Tables 16 and 17). Prieto stream and the
upper Bajamillo form the second group. The distribution is at least
partly influenced by size. The same would be true of total discharge


200
Table 15. Analysis of variance of stream discharge and sediment
transport for all streams.
Source
df
Mean Square
F Value
Average sediment discharge rate ha during flood peak
Model 4 14.28 2.33
Error 91 6.13
Peak sediment discharge rate ha per flood event
Model 4 18.60 2.53*
Error 91 7.34
Peak discharge ha ^ per flood event
Model 4 13.24 13.26****
Error 96 9.00
Average sediment concentration (g L ~*~) per flood event
Model 4 19.49 5.66***
Error 91 3.44
Maximum concentration (g L per flood event
Model 4 23.98 5.83***
Error 91 4.11


201
Table 16. Maximum recorded concentrations (g L ) per flood event,
for all streams.
Duncan groups
Ln (g L
Stream no.
means^
3.39
70
2.76
60
2.29
67
1.45
71
0.58
64
fMeans refer to the means of peak concentrations for all events
at each stream.


202
Table 17. Average sediment concentration (g L ) per flood event,
for all streams.
Duncan groups
Ln (g L *)
Stream no.
means'!"
2.80
70
2.32
60
1.86
67
0.99
71
0.36
64
Means refer to the means of peak concentrations for all events
at each stream.


203
and sediment transport. The analyses by land use group show the same
results for sediment concentration (Tables 18, 19, and 20).
The analysis of discharge per unit area also shows significant
differences between the watersheds (Table 15). In this case three
subgroups are formed. The highest discharge per ha is in the upper
Bajamillo. The second group includes the larger Bajamillo, Prieto and
Hondo. The third group combines Hondo and Pananao. This distribution
is best accounted for by an interaction of rainfall distribution,
slope, soil structure and land use. The major difference between
Prieto and the upper Bajamillo is the high percentage of land in
pasture in the latter. This watershed also has a higher percentage of
land in roads and settlement than the Prieto (Figs. 36 and 37).
Sediment discharge per unit area showed no significant difference
at the 0.05 level by simple analysis of variance (Table 15). The a
posteriori analysis indicated that the greatest variation occurred
between the two hierarchical levels within the coffee region, rather
than between the upland coffee region and the lowland pasture-food
crops production system (Table 18). The question remains as to the
extent to which this distribution is due to uniformity of erosion
rates, or to differences in the sediment delivery ratio between the
watersheds planted to coffee and those covered in pasture.
The evaluation of the land use/sediment yield interactions
requires primary data on runoff and erosion rates. The combination of
variable runoff and erosion rates, by land use type, with the detailed
land use maps (Figs. 30, 36, and 37) of the four small watersheds,
will allow the calculation of runoff and erosion coefficients for each


204
Table 18. Analysis of variance of stream discharge and sediment
transport comparing streams draining coffee stands and
streams draining food crops and pastures.
Source
df
Mean Square
F value
Average sediment discharge rate ha during flood peak
Model 1 3.53 0.55
Error 83 6.44
Peak sediment discharge rate ha per flood event
Model 1 6.13 0.79
Error 83 7.72
Peak discharge ha per flood event
Model 1 36.72 30.01****
Error 86 1.22
Average sediment concentration (g L per flood event
Model 1 57.29 17.10****
Error 83 3.35
Maximum sediment concentration (g L per flood event
Model
Error
1
83
66.68
4.04
16.52****


205
Table 19. Peak discharge ha per flood event for all streams:
results of the a posteriori test of the means.
Duncan groups Ln (Peak discharge ha ) Stream no.
means
-4.77 71
-5.62 70
-5.83 64
-6.24 67
-6.84 60


206
Table 20.
Peak sediment discharge rate ha per flood for all
streams: results of the a posteriori test of the means.
Duncan groups Ln (Peak sed. disch. rate ha ) Stream no.
-2.31 70
-3.31 71
-4.00 67
-4.09 60
-5.25 64


207
watershed. The role of these coefficients in the watershed models is
illustrated in Figs. 34, 35, 38, and 39. An analysis of the erosion
plot results is a necessary prerequisite to the above analyses at the
watershed level.
Erosion Plot and Household Studies
The erosion plot measurements and household descriptions provided
the basic data necessary to evaluate the watershed models and to
analyze Plan Sierra rural extension policies. The information
obtained at this scale proved essential in two respects. First, the
farm and plot level data provided replicated quantitative descriptions
of specific land use systems, related resource management practices
and the resultant rates of runoff, erosion and sedimentation. This
established the magnitude of the soil erosion and production problems
at the farm level, and the variation of these problems with land use
type and soil conservation practices. Both types of information were
necessary to evaluate Plan Sierra soil conservation and cropping
systems programs. Secondly, the formal and informal interaction with
farm families, landowners and nearby residents over the 15-month study
period provided valuable insights into the persistence of apparently
counterproductive land use practices.
Plot Descriptions
The diverse environmental and socioeconomic characteristics of
the 16 plots and nine households studied are summarized in Tables 11
and 21. The inherent physical characteristics, with the exception of


Table 21. Land use, land tenure and production, by household.
Plot Number
91
92
95
Site
Pananao
Pananao
Pananao
Area
1.50 ha
1.20 ha
0.75 ha
Present Land Use
Annual cash and food crops:
Sisal, sweet potatoes, peanuts,
plantains
Annual cash and food crops:
Sisal, sweet potatoes, manioc,
squash, peanuts
Annual cash crops: manioc
Past Land Use
Annual cash and food crops:
Peanuts, bitter manioc
Pasture, cash crops (peanuts)
Peanuts and manioc
Number of Crops Per Year
2 annuals, 1 sisal
2 annuals, 1 sisal
2
Time since clearing
30 years
30 years
30 years
Time occupied by owner
Original Yields
(semi-annual)
20
8000 lbs ha ^ for peanuts at
$8.15* 100 lbs-1
6 (inherited)
3600 lbs ha ^ peanuts
7500
100
30
lbs ha peanuts at $8.15
lbs, 27,000 lbs ha-! manioc
5000 units casabe at $20 100
units-1
Current Yields
<4000 lbs ha peanuts at
<2400 lbs ha peanuts at
<6000 lbs ha
manioc, 1000 units
(semi-annual)
$15 100 lbs-1
$15 100 lbs"1
casabe at $50
100 units-!
Number Persons in Household
8 (1 in N.Y.)
7
5
Number Employed off Farm
2 (plus 1 in N.Y.)
4**
2
Man-days Worked off Farm
160 180 yr1
30
variable
Type of Work
day labor in fields
panning for gold
day labor
-1
-1
-1
Wage
$3 to $4 day
varies $0 to $15 day
$3 to $4 day
-1
-1
1
Total Cash Income
$600 yr
$400 yr
$600 yr
***
* All amounts expressed in $ RD unless otherwise indicated.
** Family pans for gold on or near farm property.
***0ther sources of income not declared.


Table 21continued
Plot Number
Site
Area
Present Land Use
Past Land Use
Number Crops Per Year
Time Since Clearing
Time Occupied by Owner
Original Yields
(semi-annual)
Current Yields
(semi-annual)
Number Persons in Household
Number Employed off Farm
Man-days Worked by Employees
Type of Work
Wage
Total Cash Income
Stores and marketing were not
80 and 81
Carrizal
22 ha in watershed (34 ha total)
Coffee, pasture, food crops
> 60 years coffee
2
> 60 years
- 30 years
Food crops: 2500-3000 lbs ha ^ red beans,
3000 lbs ha-l corn
Cash crops: 2000-6000 lbs ha coffee at $60
to $230, 100 lbs-1, 10-15 head
cattle yrl (herd=100), 3000-
4500 lbs at $1 lb~l
8 residents (3 in N.Y.)
3 in N.Y.
-3000
Weeding, tilling, cleaning, spraying, harvesting
$2 box $3 day ^ plus lunch
$10,000-40,000 yr ^ (gross)*
93 and 94
Pananao
26 ha
Pasture, food crops, cash crops
> 50 years pasture
2
> 50 years
8 years
20 sacks tobacco ha ^ at $30 sack ^
3500 lbs peanuts ha-^ at $11.25, 100 lbs
11-15 head cattle yr \ 3300 lbs at $1 lb,
1000-4000 lbs cheese yrl at $1 lb~l
4 residents (4 in N.Y.)
3* $10,000
included in declarations of earnings and were estimated.
209


Table 21continued
Plot Number
82 and 83
84 and 85
Site
Los Montones
Los Montones
Area
1.0 ha
1.3 ha
Present Land Use
Subsistence and commercial production,
Subsistence and commercial production,
mixed food crops (pigeon pea and others)
mixed food crops
Past Land Use
Pine forest clearcut, then pasture rotated
Pine forest (20 year 2nd growth)
with annual cash and food crops*
Number Crops Per Year
2
2
Time Since Clearing
25 years
< 1 year
Time occupied by Owner
24 years
< 1 year
Original Yields
1500 lbs sweet potatoes, 450 lbs rice, 2000
2000 lbs red beans, 300 lbs rice,
2000 j
sold
(semi-annual)
lbs manioc, 500 lbs beans, 15,000 lbs hal
bananas, 1000 lbs manioc (sweet)
peanuts
10,000 bananas at $150, 15,000 lbs
, -1
ha j
subs
bitter manioc at $3 100 lbs"l
Current Yields
200 lbs sweet potatoes, pigeon pea crop lost,
Same as Original Yields
(semi-annual)
Overall, current yields = 20% of original
Number of Persons in Household
4
5
Number Employed off Farm
2
1
Man-days Worked off Farm
> 200
100
Type of Work
day labor in fields
forester
Wage
$3 to $4 day
= $720 yr1
Total Cash Income
$700 yr
$1000-$!300 yr1
*
peanuts, tobacco, mixed tubers, corn, beans.
1
fo
o


Table 21continued.
Plot Number
86
87
88
89
96
Site
Los Montones
Los Montones
Los Montones
Los Montones
Pananao (El Rubio)
Present Land Use
Idle pasture
Woodland
Forest
Idle pasture
Forest
Past Land Use
Grazing
Forest, selec
Same
Grazing, food
Forest, selectively
tively cut
crops
cut
Time Since Clearing
20 years


20 years

Time Occupied by Owner
20 years
1 year*
5 years**
5 years
recently acquired
by the state
* Same owner as newly cleared food crop plots, 84 and 85.
**Being held for speculation.


212
slope, vary little between plots clustered in the same watershed.
Climate and soil type do vary, however, between the clusters of plots,
by watershed.
The variability in slope between plots within subwatersheds
reflected the difficulty of encountering and replicating all of the
relevant conditions in pre-existing farm plots located in close
proximity. The plots that deviated-most from the average slope in
their respective watersheds were 82 and 83. These were included
primarily for comparison of minimum and regular tillage practices at
the same site, as a subset of the larger experiment. The slope is
equivalent for these two plots, allowing comparison between them.
These plots also can be compared with the coffee plots (80 and 81),
which have similar slopes, chosen to approximate the average slope in
the Prieto watershed.
Infiltration capacity of soils
Some of the physical characteristics at the plots sites varied
noticeably with existing or prior land use. Infiltration capacity and
soil profile descriptions reflect the influence of clearing, burning,
and cultivation in addition to the inherent site characteristics.
This is particularly evident in adjacent plots in Los Montones within
the Hondo watershed. Measurements made in adjacent plots in pasture,
mixed food crops, and forest (86, 84, 85, and 87, respectively) show
the effects of overgrazing and soil compaction on infiltration
capacity (Figs. 43 and 44, Table 11). The food crop plots (84, 85)
show a 20% reduction in infiltration compared to the forest plot. The


RATE ( c m/m i n )
Fig. 43. Infiltration rates in erosion plots at Carrizal and Pananao.
213


RATE (cm.min
Fig. 44
Infiltration rates in erosion plots at Los Montones
214


215
two cropped plots were cleared only one year earlier from a section of
the same forest stand. The same contrast in pasture versus forest was
repeated for two paired plots (88 and 89) located nearby. The
reduction in infiltration from forest to pasture was approximately 80%
in both cases. The pastures in both sites had been cleared
approximately 20 years ago and planted to food crops, then to pasture.
They have been grazed intermittently ever since.
A single measurement at the adjacent pigeon pea plots (82 and 83)
yielded a rate midway between the values for pasture and the recently
cleared food crop plots. This field was cleared 14 years ago and has
been cultivated almost continuously with occasional grass fallow. The
coffee plots (80 and J31) had the lowest infiltration rate of all the
soils tested. Prior soil moisture conditions depressed the results
somewhat."' However, the lower rates also reflect the relatively
higher clay content and overall finer texture of the soils at this
site. These soils also are subjected to some compaction by foot
traffic, particularly at harvest time.
At Pananao, the infiltration results are very different both in
magnitude and in relationship to land use. The soils generally are
shallower than at the Los Montones sites, and the texture tends to be
slightly finer (Table 11). The land use history also is distinct
(Table 21).
Plots 91, 92, and 95 were cleared about 30 years ago. The first
has been under continuous cultivation ever since in peanuts and
These were the only plots where the soil was humid prior to testing.


216
manioc. The second was planted to pasture and grazed for 25 years,
then planted to peanuts and cultivated continuously since 1975. Both
plots showed low and relatively similar infiltration rates, with the
second perhaps showing the residual effects of subsoil compaction from
grazing.
The third plot, like the first, has been under nearly continuous
cultivation in a rotation of peanuts and manioc. The site was badly
damaged by rill erosion during an intense storm at the onset of the
study (April 1980). The rills dissecting the soil surface penetrated
to bedrock in many places. Mass movement of structurally unstable
soils upslope from the plot covered over part of the plot during the
same period. The infiltration rate measured under the extreme
conditions prevailing at this site is somewhat misleading. The
infiltrated water was being drained horizontally from the test column
and ran rapidly downslope over the relatively impermeable underlying
bedrock.
The nearby forest site used as a surrogate for Pananao showed the
most anomalous results. The second growth forest cover is somewhat
deceiving. The soil is extremely shallow (see Appendix G), which is a
combined result of inherent site characteristics and prior logging.
Soil profile description and classification
The soils at the Pananao and nearby El Rubio site were classified
as Troporthents. In Los Montones all of the sites except 87 and 88
fell in the same category. The occurrence of Entisols and Inceptisols
on hillslopes is typical of the region. The high incidence of


217
Eutropepts and Dystropepts in this area was noted by Nicholaides and
Hildebrand (Nicholaides and Hildebrand, 1980a). The same soil groups
also occur on the southern slopes of the Central Mountains (Cordillera
Central) and in the dry rugged Linea Noroeste region to the northwest
of the Sierra (OAS, 1967).
This soils group is characterized by lack of profile development,
in this case due to steep slopes. The shallowness and lack of
horizons, however, may be an artifact of land use, at least in the
plots at Los Montones. The soils at the forested sites (87 and 88)
qualify as Inceptisols (Eutropepts), which indicates a greater degree
of zonation and overall profile development. These plots are
approximately 2 km apart and both are adjacent to other plots whose
soils were classified as Entisols. The shallow A horizon probably has
been removed by clearing, burning, and cultivation in the nonforested
plots.
The deeper, finer textured soils of the plots in the upland
coffee region (80 and 81) show better development of soil horizons and
are classified as Tropudalfs and Eutropepts, respectively. The soils
in coffee plots typically are finer textured and deeper than in
surrounding land cover. This reflects a circular self-reinforcing
relationship in which the best soils are used for coffee and the
coffee and shade trees, in turn, provide shade, rain shelter, detritus
and nitrogen, promoting the conservation and even the further
development of the soil.


218
Socioeconomic characteristics and land use
Land use is the variable on which the research design was based,
as discussed previously. Nested within this design was a systematic
(and initially unintentional) bias toward very small farms and poor
families in plots with subsistence food crops and annual crops in
general. Large landholders of much higher economic status usually own
the plots dedicated to pasture and coffee production and undisturbed
forest. For the latter group the land serves as a means of
subsistence and commercial production and provides security. It
represents a form of savings or investment over the long term. For
the smallholders the land is the major source of staple foods and/or
cash income to meet short-term and immediate subsistence needs. The
relationship between land use and socioeconomic characteristics is
apparent in the land use, land tenure and household descriptions for
the nine sites (Table 21).
Summary of Precipitation, Runoff and Erosion in the Plots During the
Study Period
The erosion and runoff rates measured at the study sites indicate
the magnitude and distribution of erosion and excessive runoff
problems within the Sierra. The annual rate of soil loss ranges from
-1 -1
less than 0.1 ton ha under forest cover to 70 tons ha in a
recently cleared and burned plot planted to mixed food crops. The
annual storm runoff and soil loss rates for each plot are listed in
Table 22. The total annual rainfall and storm runoff volumes are
compared and the runoff as percent of rainfall is listed for each plot


219
Table 22. Total
annualt
storm
runofft and
soil loss
rates, by plot.
Plot Number
Year
Storm
1
runoff
Year 2
Year
Soil loss
1 Year 2
3 -1
m ha
-1
yr
_ -1 -1
tons ha yr
80
402
791
3.3
0.2
81
602
1038
2.3
1.4
82
411
603
-
6.4
83
459
666
-
7.5
84
353
328
48.3
2.3
85
286
207
70.0
17.2
86
525
494
0. 3
0.2
87
161
173
<0.1
0.1
88
147
130
0.2
<0.1
89
417
348
0. 3
0.2
91
802
846
30.4
13.6
92
781
824
28.3
15.5
93
522
713
2.0
1.1
94
518
742
1.8
1.4
95
922
1435
21.0
11.6
96
99
145
0.1
0.1
tAnnual losses were calculated for May 1980 to April 1981 (year 1)
and for July 1980 to June 1981 (year 2) to evaluate the relative
importance of the contribution from the first three months of the
cropping cycle.
Storm runoff refers to the immediate surface runoff during and
after rainfall events. It does not include the subsurface water
that eventually reaches streams and contributes to total "runoff"
or discharge from the basin.


220
in Table 23. The total monthly soil and storm runoff losses from May
1980 to June 1981 are compared with monthly rainfall for all of the
plots in Figs. 45, 46, and 47.
In general the soil and runoff losses reflect the monthly
rainfall distribution, particularly in forest, coffee and pasture.
The cultivated plots, however, show a marked peak in soil loss during
the first two months of data collection (May and June 1980 for most
stations, August and September for plots 82 and 83). Soil loss peaks
in relation to rainfall and runoff then decreases steadily. This
reflects the juxtaposition of the cropping cycle and the rainfall
regime. All of the cultivated plots were monitored from the early
stages of tillage and planting through just over a year of plant
growth, with interim harvests of some short cycle crops and a
continual increase in crop cover (canopy and detrital) by the longer-
lived crops such as manioc and sisal. The dramatic decrease in
erosion rates can be attributed to the combined effects of increased
crop cover and decreased tillage. A major exception to this trend
proves the rule. The highly eroded plot (95) in Pananao planted to
manioc in May 1980 showed little decrease in erosion losses relative
to the other cultivated plots (91 and 92). The latter were planted to
sisal and food crops in the same month, then the sweet potato and some
/
manioc were harvested and grass cover filled in between the sisal and
the remaining manioc plants. By contrast, the manioc on the severely
eroded site did not fare well. The plants did not establish a closed
canopy nor did they produce much litter. The soil was left almost
completely exposed throughout the study period. The limited extent to


Table 23. Relationship of total annual rainfall and storm runoff in erosion plots.t
Site
Plot
Rainfall
yr 1
3 -i
m ha
yr 2
3
Runoff m
yr 1
ha 1
yr 2
Runoff as
yr 1
% of Precipitation
yr 2
Carrizal
80
25,728.0
27,196.5
402.35
791.46
1.6
2.9
81
II
II
602.43
1038.01
2.3
3.8
Los Montones
82
15,875.0
14,728.0
no data
603.37

4.0
83
II
II
no data
665.95
-
4.5
84
II
II
353.43
327.79
2.2
2.2
85
II
II
285.59
206.89
1.8
1.4
86
II
II
535.17
494.35
3.3
3.4
87
II
II
161.00
172.95
1.0
1.3
88
II
II
146.81
129.76
1.0
0.9
89
II
II
416.60
347.67
2.6
2.4
Pananao
91
12,701.0
12,236.0
801.50
845.81
6.3
6.9
92
II
II
781.22
824.18
6.1
6.7
93
II
II
522.31
712.67
4.1
5.8
94
II
II
518.45
742.41
4.1
6.1
95
II
II
922.30
1434.93
7.3
11.7
96
II
II
98.67
145.10
0. 8
1.2
fAnnual totals are calculated for May 1980 through Apr^l 1981 and from July 1980 through June
1981 to compare the variation in spring rainfall and runoff for both years.


RAINFALL (mm) SEDIMENT (kg hectare1)
222
PLOT No. 80
\ v-<
463
200
100
0
(HI SEDIMENT
CARRIZAL SITE storm runoff
Fig. 45
Monthly rainfall, storm runoff, and sediment loss
at Carrizal plots.
STORM RUNOFF ( m* hectare')


SEDIMENT (kg hectare')
223
3502? r
10809 -
' V
301 -
4.
> > >
968 -
iooh
O1
I
V >
El
1
327l5r
>
20067 -j j jn
17092
N
100
0
l
II
s>
PLOT 04
JLfa.
OL
1
rJlil
PLOT 85
1
JL
300
200
100
0
300
200
100
0
Fig. 46. Monthly rainfall, storm runoff, and sediment
loss at Los Montones plots.
STORM RUNOFF (m3 hectore )


rainfall (mm) l SEDIMENT (kg hectare
224
-
PLOT 87
-
Lel
p
JW~U
atO J~l-i s
(35
o
3)
Fig. 46.Continued


SEDIMENT (kg hectare
225
600
500
4 0 0
3 0 0
2 00
l 0 0
0
PANANAO SITE
I I STORM RUNOFF
Fig. 47.
Monthly rainfall, storm runoff, and sediment
loss at Pananao plots.
STORM RUNOFF (m hector e


Romfall(mm) SEDIMENT (kg heclor')
226
pananao site
0 SEDIMENT
Q STORM RUNOFF
Fig. 47.Continued
STORM RUNOFF (m heclor )


227
which soil loss was reduced can best be attributed to the cessation of
tillage.
The monthly soil loss at plots 84 and 85 in Los Montones
consistently reflects the cropping cycle and related activities. The
field was partially deforested, then burned, plowed across the slope
with a team of oxen, then planted to beans and corn in late April
1980. After a summer harvest, manioc was established and the plots
were tilled and weeded by hand. The effects of this activity are
mirrored in the September and October 1980 soil losses.
The differences in soil loss during the first and second years of
two-year cycles such as the one described above are illustrated by the
comparison of annual soil loss totals calculated for May 1980 to April
1981 and from July 1980 to June 1981. The inclusion of the first
three months of the cycle changes the totals dramatically. The
results of the study support the hypothesis advanced by many
researchers that crop cover (Elwell, 1979a; Lai et al., 1979) and
tillage (Lai, 1977c; Meyer and Mannering, 1961) are the major
determinants of soil loss in cultivated fields. This implies that
crop cover and tillage are the variables that first should be
manipulated, rather than slope, to reduce erosion.
Runoff consistently parallels monthly rainfall in all plots. The
apparent increase in runoff rates with time at plots 80, 81, and 95 is
due to installation of increased runoff storage capacity at these
sites in June 1980. The data for July 1980 to June 1981 more
accurately show the rainfall-runoff relationship.
Analysis of variance of erosion and runoff for the plots
clustered at each site and then for all of the plots at Los Montones,


Carrizal and Pananao showed significant differences by land use
category, with little significant variation by site (limited to
cropped plots). This supports the hypothesis that land use is the
major determinant of erosion and runoff rates in the area, far
outweighing inherent site characteristics in importance. The
distribution of the specific groupings for erosion and runoff reveals
the underlying rationale for the separate land use coefficients for
erosion and runoff in the regional watershed models (Figs. 19 and 20).
Los Montones and Carrizal'*'
The comparison of all plots at Los Montones and Carrizal showed
significant differences between plots for both runoff and soil loss
(Table 24). Duncan's Multiple Range Test (SAS, 1979) identified the
major groups (Table 25).
Runoff. The highest mean runoff was from coffee, with plots 81
2
and 82 forming a distinct subgroup. The second subgroup combines the
newly established coffee (81) with the partly grass covered minimum
tillage plot in pigeon peas (83). The third subgroup includes all
cultivated and pasture plots as distinct from coffee and forest.
Another subgroup combines forest, pasture and the recently cleared and
burned plots, separating them from the coffee and the pigeon pea
"'"Los Montones and Carrizal are grouped together, based on proximity,
in order to compare the existing coffee area to proposed sites for
establishment of coffee and other tree crops at Los Montones.
2
Data values are all log-transformed for analysis of variance as well
as for the a posteriori tests.


228
Carrizal and Pananao showed significant differences by land use
category, with little significant variation by site (limited to
cropped plots). This supports the hypothesis that land use is the
ma^or determinant of erosion and runoff rates in the area, far
outweighing inherent site characteristics in importance. The
distribution of the specific groupings for erosion and runoff reveals
the underlying rationale for the separate land use coefficients for
erosion and runoff in the regional watershed models (Figs. 19 and 20).
Los Montones and Carrizal"-
The comparison of all plots at Los Montones and Carrizal showed
significant differences between plots for both runoff and soil loss
(Table 24). Duncan's Multiple Range Test (SAS, 1979) identified the
major groups (Table 25).
Runoff. The highest mean runoff was from coffee, with plots 81
2
and 82 forming a distinct subgroup. The second subgroup combines the
newly established coffee (81) with the partly grass covered minimum
tillage plot in pigeon peas (83). The third subgroup includes all
cultivated and pasture plots as distinct from coffee and forest.
Another subgroup combines forest, pasture and the recently cleared and
burned plots, separating them from the coffee and the pigeon pea
''Los Montones and Carrizal are grouped together, based on proximity,
in order to compare the existing coffee area to proposed sites for
establishment of coffee and other tree crops at Los Montones.
2
Data values are all log-transformed for analysis of variance as well
as for the a posteriori tests.


229
Table 24. Analysis of variance of runoff and sediment losses
for
all plots at
site Los
Montones.
Source
df
Mean Square
F value
Ln
(Runo f f)
Model
9
9.88
6.31****
Error
286
1.57
Ln
(Sediment)
Model
9
36.73
8.31****
Error
286
4.46


Table 2 5. Runoff and sediment losses for plots at the Los Montones site:
results of the a posteriori tests of the means.
Duncan
groups
(Runoff)
Means
Plot
Number
Duncan
groups
Ln (Sediment)
Means
Plot
Number
3.47
80
3.56
85
3.09
81
3.04
81
2.49
83
3.00
84
2.23
82
2.66
82
2.12
89
2.64
80
2.09
86
2.32
83
1.90
85
1.53
89
1.88
84
1.52
86
1.53
88
0.53
88
1.46
87
0.01
87
230


231
plots. The subdivision of the plots into coffee, pasture and field
crops, and forest categories (for runoff) is justified by the results
of the a posteriori tests (Duncan's Multiple Range Test).
Erosion. The groupings for soil loss differ markedly from those
for runoff (Table 25). The plots grouped in the highest mean erosion
category include all cultivated and coffee plots. The two forest
plots form the category with the lowest mean erosion. An intermediate
grouping combined the two pasture plots with the pigeon pea plots and
established coffee plot. Based on the analysis, land use can be
grouped into three categories by erosion values: cultivated land,
including coffee; pasture and minimum till plots (with grass); and
forest. Subsequent analyses by land use groups confirmed these
results (Tables 26 and 27).
The findings at the Los Montones sites support the conclusions of
Lai (1976, 1979) and Okigbo (1977) based on erosion plot experiments
at research stations in Nigeria. Lai observed a marked decline in
erosion rates on hillslope plots in mixed food crops over a five-year
period. This may explain some of the apparently anomalous results
obtained concerning the effects of slope. According to the
experiments at IITA, the erosion rates on plots in steeper slopes were
higher at the onset of the study period, but fell below the rates for
plots on gentler slopes as the years passed. The most highly erodible
portion of the surface soil was removed rapidly on steep slopes, but
once this layer or component of the soil was gone the erosion rate
decreased markedly. Much of the soil mantle at the Los Montones sites
appears to have lost the A horizon (Table 11). In forested areas with


23?
Table 26. Analysis of variance of runoff and sediment losses for
plots grouped by land use at the Los Montones site.
Source df Mean Square F value
Ln (Runoff)
Model
3
27.60
Error
224
1.67
Model
3
99.05
Error
210
4.43
16.48****
Ln (Sediment)
22.34****


Table 27. Runoff and sediment losses for plots at the Los Montones site, results of the Duncan
Multiple Range Test, by land use.
Duncan
groups
Ln (Runoff)
Means
Plot
Number
Land use
Duncan
groups
Ln (Sediment)
Means
Plot
Number
Land use
3.24
80,
81
Coffee
3.23
84,
85
Crops
2.10
86,
89
Pasture
2.88
80,
81
Coffee
1.89
84,
85
Crops
1
1.53
86,
98
Pasture
1.49
87,
88
Forest
1
0.23
87,
88
Forest


234
sandy to sandy loam topsoil, the uppermost layers are considered
resistant to erosion in large part because of the organic matter
content and permeability promoted by the vegetative cover. However,
the burning and subsequent tilling of such forest soils reduce the
organic matter to ash, temporarily increasing fertility but also
increasing solubility of nutrients and decreasing the structural
stability and moisture retention capacity of the soil. This makes the
soil extremely vulnerable to erosion and to nutrient depletion
immediately after clearing. Soil tillage with hand implements, draft
animals or machinery (in increasing order of magnitude) also
destabilizes the soil profile structure and exposes loose soil to the
heaviest rains of the year concurrent with spring planting.
The high erosion rates in the newly cleared fields (84 and 85) as
compared to the steeper and visibly eroded site cleared 14 years ago
(82 and 83) make sense in view of this interpretation. The
differences between these fields (84 and 85) and the other cultivated
sites at Los Montones shows the need to consider the land use history
and/or prior condition of the soil and the position of a particular
land cover in the long-term rotation. Any predictive model, whether
theoretical or empirical, should include this consideration in areas
such as the Sierra, where land use change is frequent and exerts a
strong effect on soil erosion. The dynamic aspects of land use and
soil interactions must be incorporated into models of the erosion
process.


235
Pananao
A comparison among plots at Pananao also showed significant
differences for both runoff and soil loss (Table 28). The specific
groupings identified by analysis of variance are discussed below.
Runoff. The same analyses for the plots at Pananao showed no
significant differences in runoff for plots 91 through 95 (Table 29).
Forest, however, differed from the cultivated plots.
Erosion. Analysis of soil loss values yielded a far more
significant and clearly defined grouping (Table 30). Cultivated
plots, pasture and forest, in that order, constituted three mutually
exclusive groups, based on erosion yield (Table 31).
Comparison between Los Montones and Pananao
For pasture and forest, no significant difference was found
between plots at both sites (Tables 32 and 33) for both runoff and
soil loss. Significant differences did appear for cropped plots
(Table 34).
Runoff. The runoff values were grouped into three categories,
wit Pananao sites 91, 92, and 95 forming the highest group. Plots 82
and 83 formed the second (intermediate) group and plots 84 and 85 made
up the last category (Table 35). Since the second and third groups
receive the same rainfall, the climatic differences between the two
sites cannot adequately account for the grouping. One possible
explanation is the time since deforestation and the intensity of
tillage. Plots 91, 92, and 95 have been cleared and maintained in
high tillage annual crops (peanuts) or annual crops and pasture for 30


236
Table 28. Analysis of variance of runoff and sediment losses for
all plots at site Paranao.
Source
df
Mean Square
F value
Ln (Runoff)
Model
5
4.57
2.45*
Error
181
1.87
Ln (Sediment)
Model
5
95.11
17.39****
Error
181
5.47


Table 29. Runoff and sediment losses for plots of the Pananao site, results of the Duncan
Multiple Range Test, by land use.
Duncan
groups
Ln (Runoff)
Means
Plot
Number
Land use
Duncan
groups
Ln (Sediment)
Means
Plot
Number
Land use
3.02
91, 92, 95
Crops
1
4.75
91, 92, 95
Crops
2.56
93, 94
Pasture
1
1.88
93, 94
Pasture
1.91
96
Forest
1
0.50
96
Forest
t
237


238
Table 30. Analysis of variance of runoff and sediment losses for
plots grouped by land use at site Pananao.
Source
df
Mean Square
F value
Ln (Runoff)
Model
2
10.32
5.58**
Error
184
1.85
Ln (Sediment)
Model
2
237.01
43.97****
Error
184
5.39


Table 31. Runoff and sediment for plots of the Pananao site, results of the Duncan Multiple
Range Test, by land use.
Duncan
groups
Ln (Runoff)
Means
Plot
Number
Land use
Duncan
groups
Ln (Sediment)
Means
Plot
Number
Land use
3.02
91, 92, 95
Crops
1
4.75
91, 92, 95
Crops
2.56
93, 94
Pasture
1
1.88
93, 94
Pasture
1.91
96
Forest
1
0. 50
96
Forest
i
239


240
Table 32. Analysis of variance of runoff and sediment losses for
plots in pasture grouped by site.
Source
df
Mean Square
F value
Runoff
Model
1
7.13
3.53
Error
133
2.02
Sediment
Model
1
3.97
1.08
'Error
129
3.67


241
Table 33. Analysis of variance of runoff and sediment losses for
forested plots grouped by site.
Source
df
Mean Square
F value
Ln (Runoff)
Model
1
2.14
1.91
Error
74
1.12
Ln (Sediment)
Model
1
0.86
0.42
Error
72
2.04


242
Table 34. Analysis of variance of runoff and sediment losses for
plots planted in crops grouped by site.
Source
df
Mean Square
F value
Runo f f
Model
2
24.94
15.96****
Error
225
1.56
Sediment
Model
2
109.53
16.67****
Error
213
6.57


Table 35. Runoff and sediment losses for plots in Los Montones and Pananao sites, with crops,
results of the Duncan Multiple Range Test.
Duncan
groups
Ln (Runoff)
Means
Plot
Number
Site
Duncan
groups
Ln (Sediment)
Means
Plot
Number
Site
3.02
91, 92, 95
Pananao
1
4.75
91
, 92, 95
Pananao
1
2.36
82, 83
Los Montones
3.23
84
, 85
Los Montones
1
1.89
84, 85
Los Montones
2.50
82
, 83
Los Montones
243


244
years. Plots 82 and 83 have been under a rotation of subsistence food
crops and pasture for 14 years, and plots 84 and 85 were cleared,
burned and planted just at the onset of the study. The runoff results
thus confirm the tentative conclusions advanced earlier to explain the
infiltration test results. Time under cultivation seems to exert a
strong influence on infiltration and storage capacity.
Erosion. The soil loss groupings were more general with plots
91, 92, and 95 (Pananao) making up the first category and plots 82,
83, 84, and 85 constituting the second group (Table 35). If inherent
site characteristics are the determining factors, the most likely
causes of variation would be climate or soil type. The former
decreases the erosion potential for Pananao versus Los Montones. Soil
type is similar, particularly in texture. The fact that pasture and
forest do not differ significantly between sites also suggests that
inherent site characteristics are not the differentiating variables.
The difference in cropping systems, farming practice and land use
history is a more plausible explanation. In particular, the intensive
cultivation required for peanut crops has had a strong effect on the
structural stability of soils at Pananao. This has been noted by the
farmers, many of whom regard their land as spent because of the period
when they devoted their fields to peanut production (Martinez, 1981).
The analyses discussed above are based on means of log-
transformed data as opposed to totals of raw data. Thus, the
initially high losses at plots 84 and 85 placed them first on the list
for total soil loss, but the sustained higher losses at 91, 92, and 95
placed them in the highest category based on means. The latter is a


245
better indicator of continued losses, beyond the study period, if the
cropping systems are sustained with present or similar cover. One
question raised by the results is whether the losses at plots 84 and
85 for the first part of the cycle would decrease significantly in
future cropping cycles as the topsoil available for erosion decreases.
Runoff and Soil Loss Coefficients
Using the groupings established in the previous analyses,
coefficients of runoff and erosion were calculated for each land use
type. The coefficients compare soil and runoff losses in all other
plots to forest plots. The use of the natural forest cover as a
standard runs counter to the bias in the Universal Soil Loss Equation
and many agronomic studies, which use a clean-tilled bare plot as the
reference point.
Reliman compared runoff and soil loss under several land use and
cropping systems to the primary forest in upland Mindanao (Reliman,
1969). Studies in Uganda compared cropped and grazed savanna to
natural savanna (Sperow and Reefer, 1975). Researchers in Senegal
(Charreau, 1972; Moutappa, 1973) compared runoff and soil losses under
different crops to the rates for natural bush vegetation (Okigbo,
1977). The same approach is standard practice at the watershed level,
where paired watershed studies take the behavior of the watershed
under forest cover as a point of departure for further experiments
with land use and management variables (Hewlett and Nutter, 1969;
Likens et al., 1977).
Runoff for plot analyses refers to storm runoff unless otherwise
stated.


246
The coefficients for runoff and soil loss are given in Tables 23
and 36. The results in this case are a measure of annual totals and
also use different rainfall rates, by site, as a basis for comparison
of runoff with rainfall. This accounts for the slight differences
with the analysis of variance results presented earlier. Overall,
this comparison confirms the analysis of variance results and provides
a specific quantitative measure of runoff and erosion effects by
group. The runoff as a percent of precipitation varies from 1.0 to
4.5% at Los Montones and Carrizal. The highest runoff rates are
attributed to the long established cropped fields (82 and 83) and
coffee plots (81 and 82), and the lowest rates are found in forest.
At Pananao the rates are generally higher, except for forest (1.2),
indicating that differences in rainfall distribution do not explain
the higher runoff rates. The cropped plots (91, 92, and 95) have the
highest runoff rates (6.7 to 11.7%) with the,highest rate occurring in
the plot with the longest history of croppings. The runoff
coefficients can be applied, as they are, to separate models of the
small watersheds, or the coefficients for both sites can be combined
in the case of forest and pasture for use in the regional model.
Cropping systems must be differentiated with respect to their present
and historical resemblance to those at Pananao or Los Montones.
Comparison of the results with those of other similar studies in
erosion plots shows a trend in the relationship of land use (present
and past) to runoff rates. Runoff rates under natural savanna in
Uganda (Table A-9), natural bush in Senegal (Table A-8) and primary
forest in upland Mindanao (Table A-ll) are consistently less than


Table 36. Soil loss coefficients by land use and conservation practice.!
Los Montones + Carrizal Sites
Forest§ 1 (1)J
Pasture 6 (12)
Coffee (old)
Coffee (new)
Pigeon pea, Sweet
potato^
Pigeon pea, Sweet
potatoH#
Yuca, beans, com
Yuca, beans, com
10
38
182
153
1462
252
(first year|f
second year!!
for both plots)
Forest§§
Pdndnso
1
bl l6S
Pasture
17
Sisal, Sweet
383
(first year
potato
Sisal, Sweet
252
second year
potato
for both plots)
Yuca
213
fCoefficients were calculated relative to soil losses (tons ha-l yr~l)under natural forest cover.
The reference plots for no's 80 through 89 are plots 87 and 8 (mean). Plot 96 is the reference
for no's 91 through 95.
^Numbers in parentheses indicate coefficients that apply to both sites for the regional model.
§Reference for Los Montones and Carrizal sites (80-89).
UHillside ditches are used.
#Minimum tillage is practiced,
ttfirst year, after clearing, burning and tilling.
Second year.
§§Reference for Pananao sites.


248
1.0%, ranging from 0.25 to 0.90%. The values for the study areas in
the Sierra compare well with these examples, at 1.0%.
The increments of runoff over natural vegetation range from 3 to
7 for pastures in the study area, as compared to 12-fold increases
reported for Imperata grassland in Mindanao (Kellman, 1969) and 10-
fold increases reported for grazed savanna in Uganda (Sperow and
Keefer, 1975). Incremental increases for field crops range from 20 to
40 in Senegal (Charreau, 1972; Moutappa, 1973) and Uganda (Sperow and
Keefer, 1975). In Mindanao, Kellman measured widely varying
proportional increases of runoff in field crops over forest values
(Table A-ll). In new fields the runoff increased 4 to 6 times and in
older fields the runoff ranged from 8 to 50 times the amount for
forest.
The runoff rates measured in field crop sites in the study area
showed increases over forest rates ranging from 2 to 12-fold, which is
less than all of the above. However, the runoff as percent of
precipitation very closely approximates the rates reported by Kellman
(1969), with the exception of forest. The 1 to 2% runoff rates for
newly established field crops are repeated at the Los Montones plots
(84 and 85). The trend toward increased runoff with longevity of
field crop production at the site, mentioned earlier, is also
reflected in the Mindanao results. The 12% rate for 12-year-old rice
fields was duplicated in the eroded manioc plot (95) at Pananao. This
plot was the most consistently and intensively cultivated of all the
plots included in the study. The results of both studies indicate the
importance of incorporating land use history into runoff estimates,
whether at the plot or regional level.


249
Soil loss coefficients (Table 36) indicate a 12-fold increase
over forest erosion rates in pastures. With land under coffee,
erosion losses increase from 10 to 38 times over forest rates,
depending upon the age and condition of the stand. Mixed crops of
various types show erosion rate increases of 153 to 383 times the
forest losses, with the exception of the newly cleared and burned
plots. In the latter case the increase relative to forest erosion
rates was greater than 1000-fold.
The soil loss increment reported by Kellman (1969) for pasture
(2.0) is well below the value for the study area. However, the
average erosion rates for forest and pasture in the U.S. (Table A-4)
(Pavoni, 1977; USEPA, 1973) show a 10-fold increase in pasture.
Reports from Jamaica also indicate an increment of 10 (Table A-7)
(Sheng and Michaelson, 1973).
Jamaican field crops showed soil loss increases of approximately
100 to 250-fold over forest in the same study. The increment for
vegetable crops in Malaysia (Table A-6) (Morgan, 1979) varied from 30
to 250; plots in field crops in Senegal yielded 35 to 50-fold
increases, and reports from Uganda showed ratios of 260 to 340 for
cowpeas and corn (Table A-9) (Sperow and Keefer, 1975). Kellman
(1969) measured increases ranging from 70 to 600-fold in corn and rice
fields in Mindanao. The values for new corn versus a 2-year-old corn
field show the same trend apparent in the comparison of first and
second year data for plots in manioc, beans and corn (plots 84 and 85)
in Los Montones. The initial erosion rate decreases from the first to
second years.


250
By contrast, the continued repetition of the full cropping cycle
in a given plot yields higher erosion rates over the long term as
indicated by the plots in Pananao, 91, 92, and 95, and the comparison
of "second year" (in the study period) results for plots 91, 92, 95,
82, and 83 with plots 84 and 85 (newly established). In contrast to
the Mindanao results, however, the highest total losses for the first
year of the cropping cycle occur in the newly cleared plots.
The absolute values for soil loss (Table 22) are comparable to
results from studies conducted in Puerto Rico, Colombia, Brazil (Table
A-7) and Uganda (Table A-9). Comparisons with results from Jamaica
(Table A-7), Mindanao (Table A-ll), the United States (Table A-4), Sri
Lanka (Table A-10), and Nigeria (Lai, 1977c), show similar erosion
rates for selected land uses and cropping systems. The erosion rate
for forest (0.1 ton ha ^ yr ^) in the study area compares well with
the studies conducted in Uganda (0.1 ton ha ^ yr ^, natural savanna),
Mindanao (approximately 0.07 ton ha 1 yr ^), Senegal (0.2 ton ha
-1 -1 -1 1
yr natural bush), the U.S. (0.9 ton ha hr forest) and
Malaysia (0.03 ton ha ^ yr \ rainforest) (Morgan, 1979). By
contrast, Sheng and Michaelson report losses of 0.5 to 1.3 tons ha
yr under dense forest cover in Jamaica.
-1 -1
Erosion rates under pasture (0.2 to 2.0 tons ha yr ) agree
well with the rates reported for fertilized pasture in Puerto Rico
(1.5 tons ha 1 yr ^) and Brazil (1.2 to 2.7 tons ha ^ yr ^) (Table A-
7). Results from Mindanao (0.2 ton ha 1 yr ^) (Table A-ll) and the
"''The USEPA (1973) and Pavoni (1977) report national averages which
disagree with estimates published by USDA (1980). The first group
will be cited henceforth, unless otherwise indicated.


251
U.S. (0.9 ton ha ^ yr ^) (Table A-4) also fall in the same range.
Soil losses in Uganda were slightly higher (2.5 to 4.4 tons ha ^ yr 1)
(Table A-9) and erosion rates reported for Jamaica ranged from 5.0 to
12.5 tons ha 1 yr 1 (Table A-7).
The established coffee plots yielded 0.5 to 3.3 tons ha ^ yr \
the latter being influenced somewhat by site disturbance from
construction. The plot newly planted to coffee yielded about 2.0 tons
ha 1 yr ^. Both results compare well with the reported erosion losses
for Colombian plantations at Chinchina (Table A-7), which started at
-1 -1
0.6 tons ha yr for established plantations and ranged from 1.8 to
24.0 tons ha ^ yr ^ for new plots.
Soil losses reported for annuals and other field crops fall
within the wide range of erosion rates reported in the literature for
various cropping systems. The losses in the study area ranged from
-1 -1 -1 -1
6.0 tons ha yr for plots in pigeon peas, to 70.0 tons ha yr
for the newly cleared plots in manioc, beans, and corn (Table 22).
The values found in the literature range from 1.0 to 125.0 tons ha
-1
yr (Table A-7).
Comparison of Measured Erosion Rates with USLE Estimates
The erosion rates measured on the plots were compared with
calculated results using the Universal Soil Loss Equation. The
equation overestimated the soil losses in all cases (Table 37). One
possible explanation is the overestimation of the C factor, which is
the only factor that was not estimated from quantitative field


Table 37. Comparison of measured and predicted erosion losses.
Plot
Number
Land Use
USLE
Prediction!
Measured
Erosion Losses^
tons
ha'1 yr'1
tons
ha-1 yr-1
80
Coffee (old)
37.8
0.5
- 3.3
81
Coffee (new)
1197.0 -
2993.0
1.4
- 2.3
82
Pigeon pea, Sweet potato§
195.6 -
880.1
6.0
- 6.4
83
Pigeon pea, Sweet potato§H
101.6 -
457.4
7.3
- 7.5
84
Yuca, Beans, Com
183.7 -
826.6
2.5
-48.3
85
Yuca, Beans, Com§
91.8 -
413.3
17.2
-70.0
86
Pasture
46.6 -
233.0
0.2
- 0.3
87
Pine Forest
1.3 -
4.5
0.1
88
Pine Forest
0.9 -
3.2
0.1
- 0.2
89
Pasture
37.9 -
189.3
0.2
- 0.3
91
Sisal, Yuca, Sweet Potato**
164.2 -
738.7
13.6
-30.4
92
Sisal, Yuca, Sweet Potato
320.1 -
1440.5
15.5
-28.3
93
Pasture
57.5 -
287. 3
1.1
- 2.0
94
Pasture
43.1 -
215.5
1.4
- 1.8
95
Yuca
273.6 -
1231.2
11.6
-21.0
96
Pine Forest
2.8 -
9. 3
0.1
fThe C factors for field crops and pasture were varied to cover a
conditions. The estimates are based on C factors applied in the
1978) and in the Dominican Republic (Santana, 1980).
listed in Appendix I.
^Annual totals are reported for May 1980 to May 1981 and for July 1980
§Site includes hillside ditches across the slope.
^Minimum tillage practice with grass cover was used.
broad range of plant cover
U.S. (Wischmeier and Smith,
The parameters for the equation are
to July 1981.


253
measurements in the area. However, the equation also overestimated
erosion losses in forest and pasture by a factor of 10 or more in
spite of the fact that the C factors for these land covers are well
established and vary little by site.
The rainfall erosivity index (R) has a fairly sound theoretical
basis. The parameters for this factor are based on a well documented
analysis of empirical data from the study area (Paulet, 1978). The
error is most likely due to overestimation of the effects of slope
(LS) and erroneous assumptions built into the soil erodibility (K)
factor. Similar conclusions have been advanced by researchers in
Puerto Rico (Smith and Abruna, 1955), where the soil erodibility
factor as normally derived failed to account for experimental results.
Results of erosion plot experiments in Africa and Hawaii have
indicated the need for inclusion of clay mineralogy of soils as a
determinant of erosion (Greenland and Lai, 1977). The high proportion
of iron and aluminum compounds in some tropical soils significantly
alters the probability of particle detachment from the soil surface.
The K factor also fails to address the change in soil properties over
time. Based on reports from West Africa (Lai, 1976, 1979), the
Philippines (Kellman, 1969), and the results obtained at the plots in
the Sierra, the K factor for the exposed surface soil changes over
time in response to erosion. The A and B horizons and their
respective subunits cannot be represented adequately given a single K
Estimates were based on values reported in the literature for compa
rable cropping systems. The C factors calculated for mixed crops in
smallholder fields at Mata Grande (Santana, 1980) were applied to
the field crop plots.


254
factor derived from data on the uppermost horizon. An erodibility
factor applicable over the long term would need to incorporate
feedbacks from the erosion process itself.
The applicability of the P factor is also questionable. A
comparison of soil loss coefficients (relative to estimates for
forest) showed a parallel increase with measured erosion rate
increases for all land uses, with the exception of cropping systems .
with soil conservations practices. The P factor overestimated the
incremental reduction in erosion in every case (Table 38).
The predicted effects in soil erosion prevention or decrease are
not even nearly achieved in any of the cases (82, 83, 85, and 91).
When compared to the control plots (84, 92), the plots with hillside
ditches showed no significant improvement. The minimum tillage
experimental plots at 82 and 83 did show some significant difference
as indicated previously by the analysis of variance, but did not
approach the effect predicted by the USLE. This reflects the failure
of such a one-dimensional analysis to account for the feasibility of
proper construction and maintenance. The implementation of this
technology in the study area does in fact stem from the widespread use
of the USLE to predict erosion reductions by alternative conservation
and cropping practices, under conditions where the parameters of the
equation have not been calibrated.
Beyond the problems with local or regional calibration, this
empirical equation cannot by nature adequately represent the
agroecosystems in question well enough to accurately predict their
responses to change, particularly when cultural and economic variables


Table 38. Comparison of soil loss coefficients derived from USLE and from empirical data.f
Coefficients derived from USLE
Coefficients derived from empirical data
Los Montones + Carrizal Sites
Forest§
1
1
Pasture
19-93
6
Coffee (old)
15
10
Coffee (new)
479-1197
38
Pigeon pea, Sweet
potato 1
78-352
182
Pigeon pea, Sweet
potato 1#
41-183
153
Yuca, beans, com
74-331 (all
years)
1462
Yuca, beans, com
37-165 (all
years)
252
(Dt
(12)
(first yeartt
second yeartt
for both plots)
Pananao Sites
Forest§§
1
Pasture
8-42
Sisal, Sweet
potato
66-296
(all
years)
Sisal, Sweet
potato
128-578
(all
years)
Yuca
46-205
1
17
383 (first year
225 second year
for both plots)
213
fCoefficients were calculated relative to soil losses (tons ha~l yr~l)under natural forest
cover. The reference plots for no's 80 through 89 are plots 87 and 88 (mean). Plot 96 is
the reference for no's 91 through 95.
tNumbers in parentheses indicate coefficients that apply to both sites for the regional model.
§Reference for Los Montones and Carrizal sites (80-89).
HHillside ditches are used.
#Minimum tillage is practiced.
ttFlrst year, after clearing, burning and tilling.
^Second year.
§§Reference for Pananao sites.
255


256
are added to the already complex situation of the natural systems.
The use of the USLE to determine soil conservation policies is
predicated upon comparison of proposed alternatives with the worst
possible conditions (tilled bare fallow). The range of alternatives
includes modifications of slope, cropping practice, and conservation
practice. The requirements of the site for sustained production are
not taken into account. The adaptations of natural vegetation to
prevent excessive erosion (<0.2 ton ha 1 yr ^ in most cases cited) are
not considered, nor is the soil loss under a given cropping system
evaluated against this alternative. By contrast, a model of the
natural ecosystem and the complex interactions and feedbacks that
maintain and build the soil would serve as a point of departure for
incorporating some of the protective mechanisms into annual and
perennial cropping systems or alternative land uses. Systems models
of the combined natural and socioeconomic aspects of selected small
farms are presented as an alternative to the empirical approach
embodied in the USLE.
Farm Level Models
The household of the major coffee grower, along with the adjacent
smallholdings of one son, were chosen for detailed analysis to
evaluate the farm level model for coffee production (Fig. 48). The
two related landholdings are typical of a type of coffee production
unit found throughout the Sierra, in Juncalito, Diferencia, Mata
Grande, and Carrizal. This kind of family enterprise is usually
associated with smallholder producers of coffee as well as with


257
Fig. 48. Model of coffee farms: Large and small holdings.
5.8 RD kg ^ (dry)
10~4 -RD/Kcal


258
subsistence farmers in adjacent or nearby holdings. The promotion of
coffee by development programs will either reproduce, modify of clash
with this existing production system. This typical example of the
established system provides the best basis for prediction and
evaluation of the effects of coffee promotion programs, since no
viable alternative is yet available for study.
The amount of food crops produced, relative to the area in coffee
is an important feature of the system (Fig. 48). The proportion of
income from crop, coffee and pasture lands indicates that the food
crops are not a major source of income but function as a support for
the coffee production enterprise. The duplication of coffee and crop
production at a small scale would constitute a high risk enterprise
without the associated pasturelands and cattle that serve as security
against bad harvests or failing market prices.
The.proportion of food and fuel production that goes to maintain
the labor force is another critical flow to consider if this system is
to be promoted at the smallholder level. The smallholder would need
to dedicate land to the production of food crops for the family as
well as for the harvesters and other hired help. Figure 48 indicates
the flows and storages that are integral to the maintenance of the
largeholder coffee farm and that could possibly be duplicated or
replaced by parallel features in coffee production systems based on
smallholder associations like those promoted by Plan Sierra programs.
At present coffee production and marketing receive priority in
existing smallholder associations and in programs to organize
cooperatives.


259
The average erosion and water yield from the complete holdings of
the large coffee grower are substantially higher than for the coffee
stand itself. This, however, still accounts for only a part of the
food production associated with his coffee holdings. The food crop
plots of seasonal laborers who reside in the area also serve to
maintain the coffee production in the large landholdings. The forest
lands provide fuel for the grower's household and more importantly,
for the workers and their families.
The large growers, as the diagrams show (Figs. 34 and 48), manage
entire small watersheds both directly and indirectly through their
management of the production system at the plot level. The transfer
of this production system as a whole to smallholder enterprises would
require a similar coordination of diverse land uses and production
subsystems at the watershed and community levels. One of the
important questions that remains is whether the reproduction of this
system is desirable based on ecological and economic criteria at the
watershed and household levels.
Two landholdings at Pananao were chosen as examples for model
(Figs. 49 and 50) evaluation of pasture and annual crop production
systems. As in Carrizal, the examples show a large and a small
landholder, respectively. In the case of Pananao, the two landholders
are more independent of each other's activities at the household
level. There is some interaction, however, at the community and
watershed level over the long term. The evaluated models illustrate
the two classes of land use and production systems most common in the
watershed.


260
Fig. 49. Model of dairy and cattle farm, Pananao.


261
Fig. 50. Bitter manioc production on a small holder plot, Pananao.


262
The two models contrast more in scale than in structural and
functional attributes. However, the larger scale of operation
possesses certain attributes that are inaccessible at small scales,
such as retailing ones own produce, local investment and
establishment of savings or equivalent security (land and cattle).
These pathways and storages in the pastureland model reflect a
structural constraint imposed by a combination of scale and the
prevailing economic system in which both the large landowner and the
smallholder operate.
Another major constraint between the two scales of operation and
the respective land use systems at Pananao is in the erosion rates.
The smallholder loses a much larger proportion of the soil resource
base with the spring rains each year. The large landholder can afford
to rotate land cover types within a general land use system, and can
afford to use the land much less intensively. The smallholder uses
all or most of his plot to meet subsistence needs directly, or
indirectly through income from cash crops. These relationships are
shown in the contrasting rates of similar processes in the two models.
The intensity and longevity of cultivation on a given plot are
increased by a self-reinforcing mechanism in the case of the small
farmer. The high rates of soil erosion and nutrient loss lower the
yields. This requires cultivation of a larger portion of the plot to
produce an adequate food or cash crop harvest. This, in turn, reduces
the fallow time in the rotation, which increases erosion and soil
depletion over the long term. The spiralling effect leads to
permanent cultivation of the entire holding with increasingly lower


263
yields per unit of land and labor. This is the situation that leads
the smallholder to work part time off the farm. The model of
smallholder production in Pananao shows the process of agricultural
involution observed by Geertz (1963) in Indonesia. It is also similar
to the process modelled by Lagemann (1977) in Nigeria.
The switch to manioc production at the small farm level, simply
because nothing else will grow on the site, is a classic example of
the phenomenon of agricultural involution. The natural system's
positive feedbacks to the soil have been almost completely eliminated
(Fig. 50). The recovery from such a situation requires the design of
farming systems that incorporate such feedbacks to soil nutrient
build-up and that promote development or maintenance of soil profile
stability. Commercial production must be reconciled with the
provision of food, fuel and shelter for local consumption, if the
competition between the two is not to end in an eroded degraded
landscape characterized by chronically low production levels.
The diagram of a hypothetical smallholder plot (Fig. 51)
illustrates an integrated farming system dedicated to a diverse
combination of land uses that produces for both commercial sale and
subsistence needs. The case demonstrates the potential for
reincorporating positive feedback loops into smallholder
agroecosystems. The example is drawn from a functioning farming
system in a plot located near the Pananao Valley. The family produces
wax, honey, and goats for sale in local and regional markets. They
also produce most of the staples (manioc and other tubers) for
household needs. The complex mosaic of fruit, coffee, shade and


264
Fig. 51. Model of a mixed production on a well integrated farm.


265
fencerow trees provides forage for the goats, food and shelter for the
bees, and fruit for the family. They are maintained by this system on
a 0.5-ha parcel supplemented by a sharecropped plot of equal size
(Georges, 1981).
Application of Erosion and Runoff Coefficients to the Small Watersheds
The results from the erosion plot study provide the basic data
for the evaluation of the watershed level land use-erosion submodels.
The combination of the erosion and runoff rates by land use category
with the land use distribution allows the calculation of composite
erosion and runoff rates (L and in Figs. 52-56) by watershed. The
derivation of these rates is summarized in Tables 39 and 40.
The comparison of the calculated runoff rates between watersheds
shows a relatively uniform distribution of storm runoff as a percent
of rainfall. The values range from 3.0% for Hondo watershed to 4.2%
for Pananao. This confirms the results of the analyses of variance
with respect to the relative uniformity of storm discharge per unit
area on all five watersheds. The rank order, however, was reversed by
the calculation of runoff rate relative to rainfall volume at each
sit e.
The erosion rates vary more than the runoff rates but the range
is still surprisingly narrow. The calculated composite rate varies
-I -1 -1
from 3.58 tons ha yr for the Upper Bajamillo to 5.85 tons ha yr
1
for Pananao (Table 40). The overriding factor that determines the
rate of erosion loss for the watershed is the proportion of total land
area in crops, the condition of the croplands and the cropping
practices in the area. The land area in annual crops contributes more


R= mean annual roinfoll
Fig. 52. Evaluated small watershed submodel of land use, erosion, and sedimentation:
Prieto stream.
266


Fig. 53. Evaluated small watershed submodel of land use, erosion, and sedimentation: Upper
Bajamillo stream.
267


Legend
R£ mean annual rainfall
S = soil (mass & volume)
M= soil moisture storage(50% soil
volume)
L= land area
F= % land area in forest
A = % land in crops,bananas, plantains
C= % land in coffee
P= % land in pasture
SUSPENDED SEDIMENT EXPORT
1.5 Ions ho1 yr1 (bo sel low)
1.38 ion* ho (peak)
STREAM DISCHARGE
bosellow
75580 rn ho yr
8148 06 m ho"1 yr1
Fig. 54. Evaluated small watershed submodel of land use, erosion, and sedimentation:
Greater Bajamillo stream.
268


RAIN = MEAN ANNUAL RAINFALL
SOIL = MASS AND VOLUME
Fig. 55. Evaluated small watershed submodel of land use, erosion, and sedimentation: Hondo
stream.
269


Fig. 56.
Evaluated small watershed submodel of land use, erosion, and sedimentation:
Pananao stream.
270


Table 39. Storm runoff coefficients, total runoff estimates'}" and runoff/rainfall ratios by watershed
Storm
% Area in each land use
runoff coefficient
(% rainfall)
Estimated contribution to^
Composite runoff (m ha yr )
Total rainfall estimated runoff
and storm runoff
Watershed #64
48.4
coffee
2. 3%
189.1
Total rainfall = 3,750,000 m3yr 1
Prieto Stream
12.9
new coffee, plantains,
3.0
68.8
Rainfall = 1719 mm yr 3
20.4
bananas
food crops
4.0
137.5
Baseflow = 2,522,800 m3yr 3
Area = 187.5 ha
13.8
pasture
6.02
142.3
(67%) x
4.5
forest
1.0
7.7
Storm runoff = 110,325 m yr
0.5
"urban"
50.0
43.0
(3.4%) 3
588.4 i
3 -1 -1
m ha yr
Total runoff = 2,633,205 m yr
(70.4%)
Watershed #71
39.9
coffee
2.3%
154.7
Total rainfall = 1,633,050 m yr
Upper Bajamillo
15.0
food crops
4.0
103.1
Baseflow = 1,135,296 m^yr-^
Stream
31.0
pasture
6.0
319.7
(70%)
Rainfall = 1719 mm yr
13.9
forest
1.0
22.3
Storm runoff = 62,700
Area = 95 ha
0.7
"urban"
50.0
60.2
(3.8%)
660.0 i
3 -1 -1
m ha yr
Total runoff = 1,197,996
(73.4%)
Watershed #67
15.7
food crops
4.0%
77. 1
Total rainfall = 15,766.950 m3yr
Baseflow = 4,099,680 m yr-'*'
Hondo Stream
41.3
pasture
3.0
152.0
Rainfall = 1227 mm yr
Area = 1285 ha
18.6
23.2
bush fallow
forest
2.0
1.0
45.6
28. 5
(26%) 3 _i
Storm runoff = 473,000 m yr
1.1
"urban
50.0
67.5
(3%)
Total runoff = 4,572,416 m yr
370.69
J -L .
m ha yr
(29%)
Watershed # 60
15.8
food crops
7.0%
137.1
Total rainfall = 16,275,000 m3yr
Pananao Stream
46.4
pasture
5.0
287. 7
Baseflow = 3,784,320 m-^yr-2'
Rainfall = 1240 mm yr
19.0
bush fallow
2.0
47.1
(23%) 3 1
Area = 1312.5 ha
18.6
forest
1.0
23. 1
Stormflow = 690,375 m yr
0. 5
"urban"
50.0
31.0
(4.2%) 3 ^
Total runoff = 4,474,697 m yr
(27.5%)

526.0
3 -1 -1
m ha yr
Watershed # 70
20.0
coffee
2.3%
79. 1
Total rainfall = 50,925,375 m3yr
Greater Bajamillo
Rainfall = 1719 mm yr
25.0
(35) food crops
4.0
171.9
(240.7)
Baseflow = 22,390,560 m^yr--*-
30.0
(20) pasture
6.0
309.4
(206.4)
(44%) 3 ^
Area = 2962.5 ha
14.0
forest
1.0
24. 1
Storm flow = 1,731,463 m yr
1.0
"urban"
50.0
90.0
(3.4%) 3 3
Total runoff = 24,138,628 m yr
(47.4%)
674.41
3 -1 -1
m ha yr
fsee Appendix F for list of calculated flood discharge and low-flow discharge, by watershed, from velocity-
area measurements.
$Coefficients for pasture in watersheds 64, 70, and 71 adjusted up to compensate for lower soil permeability,
as compared to Los Montones sites.
271


Table 40. Calculation of soil loss coefficients'*' and soil loss estimates (tons ha ^yr "S by watershed.
% Area in each x
Erosion coefficient
= Estimated contribution to composite erosion
(tons ha~l yrl) by land use
Estimated
erosion rate,
1 average
by watershed
#64
48. 4
coffee
16 (5)
7.7 (2.4)
Prieto Stream
12.9
plantains/new coffee
38 2
4.9
at Carrizal
20.4
food crops
1693
34.5
13.8
pasture
6
0.8
4.5
forest
1
0. 1
0. 5
"urban"
300
1.5
(4.5)
Composite
49.5 x forest (45.0)
5.0
tons
-* i
ha yr
#71
39.9
coffee
16 (28)
6.4 (11.2)
Upper Bajamillo
15.0
food crops
169
25.4
Stream at
31.0
pasture
6
1.9
Rincon de Piedras
13.9
forest
1
0.1
0. 7
"urban"
300
2.1
(4.1)
-1 1
35.8 (40.6)
3.6
tons
ha '"yr
#67
15.7
food crops
300
47. 10
Hondo Stream
41.3
6
2.48
San Jose
18.6
bush fallow
3
0.56
Los Montones
23. 2
forest
1
0. 23
1.1
"urban"
300
3. 30
-1 -1
53.67
5. 4
tons
ha yr
#60
15.8
food crops
300
47.4
Pananao
46.4
pasture
17
7.89
19.0
bush fallow
8
1.52
18.6
forest
1
0. 19
0. 5
"urban"
300
1.50
-1 -1
58.50
5.9
tons
ha yr
#70
20.0
cof fee
16
3.20
Greater Bajamillo
25.0
(35.0) food crops
169
42.25 (59.20) *
40.0
(20.0) pasture
6
2.40 ( 1.20)
14.0
forest
1
0. 14
1.0
"urban"
300
3.00
(6.7)
-1 -1
50.99 (66.74)
5.1
tons
ha yr
2Expressed as incremental increase over forest values of 0.1 ton ha ^yr ^.
^Numbers in parentheses indicate alternative coefficient for coffee based on age and condition of stands.
Soil loss of 169 tons ha-lyr-^ based on assumption of 50% plantains and banana intercropped with annuals.
272


273
than 100 times the soil loss from forest lands, 25 times the losses
from pastures and more than 10 times the erosion from coffee
plantations in the same area.
In addition to the influence of land use category, the land use
history and specific cropping practice may ameliorate or exaggerate
the generally high erosion rates attributed to land in annual crops.
The Pananao watershed is an example of an area with a high proportion
of land in annual crops. It also experiences relatively high erosion
rates within the cropland areas. This is probably due to the
intensive cultivation of peanuts during the last 30 years and to the
resultant damage to soil structure.
The analysis of variance for sediment yield per unit area during
floods showed no significant difference between the watersheds in the
coffee region and those in the lower elevations covered in pasture and
croplands. The most plausible explanations are the effects of
sediment discharge ratio and the existence of a uniform erosion rate.
The first explanation seems to hold for the case of the Prieto and the
Upper Bajamillo watersheds compared to Hondo and Pananao. The fact
that all four watersheds yield approximately the same amount of
sediment per unit area during flood events does not by itself indicate
a uniform erosion rate. According to Roehl (1962) the smaller
watersheds should yield approximately 33% of the total amount of
sediment eroded, while the larger watersheds should yield
approximately 10% of the total eroded material.
The comparison of baseflow versus flood flow sediment yields
(Table 39) and the calculation of sediment delivery ratios for


274
baseflow and peak annual flood showed that while the behavior of the
streams may be distinct, the net result (sediment yield) is very
similar. The calculated erosion rate does indicate the lowest soil
loss from the surface of the two small watersheds in the uplands of
the coffee region (Table 40).
The larger Bajamillo, however, exhibits the highest average
sediment yield per unit area for flood events, and closely
approximates the yields for Hondo and Pananao as indicated previously
in the a posteriori analyses. In this case the size of the watershed
relative to Hondo and Pananao rules out the explanation of sediment
delivery ratio. Uniformity of sediment yield at the same or similar
scales suggests a uniform erosion rate. The calculated rates for the
larger Bajamillo illustrate the potential for erosion equivalent to
that at Pananao. The variation of the proportion of land in annual
crops between 25 and 35% changes the composite erosion rate from 5 to
7 tons ha ^ yr ^.
The combination of measured sediment yield and calculated erosion
rates for the five small watersheds disproves the hypothesis that the
coffee region loses less soil per unit than the pasture and croplands
in the drier areas at lower elevations. Rather, it is the proportion
of land in field crops that determines the erosion rate on the
watershed. In spite of the visible differences between the two
landscapes the coffee-field crops and pasture-field crops associations
do not differ substantially in terms of erosion rates or sediment
yields. The population density and resultant demand for food crop
production (and/or annual cash crop production to generate income) are
roughly the same for the two areas.


275
The promotion of coffee for watershed protection becomes
contradictory when the erosion from associated food crop production
raises the net yield from the watershed to the same level as that for
Pananao. The basic demands for food, fuel, and shelter are relatively
inelastic for a given population. If the needs are not met by local
products they must be purchased with income from cash crops or wage
labor. In the case of coffee plantations the labor force could not
afford to purchase imported food and fuel with their wages (Fig. 48).
Given the existing functional relationship between food crop and
coffee production the coffee producing system does not offer a net
reduction in erosion unless coffee is replacing annual cash crops.
Even in the latter case it may be ineffective in replacing the
smallholder cash crop plots. It may simply displace them to other
adjacent areas, depending upon whether coffee can meet the income
needs of the smallholder at the scale of the small plot.
Sediment Delivery Ratios
Given the erosion rates measured in the plots, the suspended
sediment transport observed in the small streams, and the sediment
yields derived for the large watersheds, the sediment delivery ratio
2
is approximately 60% for the 10 to 20 km watersheds and 25% for the
2
watersheds greater than 300 km in area. This contrasts sharply with
the 40 and 10% rates for baseflow conditions. It also contrasts with
2
results for the U.S. where the ratio is approximately 33% for a 1-km
watershed (Roehl, 1962). The apparent sediment delivery ratios in the
Sierra study areas more closely resemble the ratios reported for
catchments in Tanzania (Temple, 1972; Rapp, 1977).


276
These sediment delivery ratios for the study area also may be an
artifact of the specific time interval during which the data were
collected. The streams and rivers of the Sierra may be gradually
transporting a sediment load derived in large part from erosion of
newly cleared and tilled forest lands during the period of intensive
logging and settlement about 30 years ago. Although the data from the
study do not allow further exploration of this explanation, it is a
possibility that should be considered when evaluating the high
sediment delivery ratios.
The need to consider the interaction of land use and river
systems in a dynamic context is demonstrated by the case of Pananao
stream. As depicted in Fig. 19, any given storm may combine newly
eroded soil with resuspended channel deposits in the total suspended
sediment load transported past the sampling point. The sediment
deposits in the stream channel, as well as the condition and
topography of the stream banks already have been affected by prior
land use changes and management practices. The amount of runoff that
reaches the stream is partially conditioned by prior use of the land,
as well as by current use and practice.
Much of the deposited material available for resuspension in
overland flow and streamflow may derive from the initial intensive
clearing and tilling of the Pananao Valley 30 years ago. These
sediments in turn would have contributed to the scouring action of the
streamflow, accelerating channel erosion. Attributing the current
condition and behavior of the stream solely to the existing land use
and cover could yield false conclusions and inappropriate management
decisions.


277
Delayed transport of previously deposited sediments is also a
probable explanation for the high sediment yield in the Mao Basin.
The sediment delivery ratio is estimated at close to 50% if channel
erosion and prior deposition are excluded from consideration. Since
the initial deforestation 20 to 30 years ago, the rate of sediment
deposition in or near the channels may have exceeded the capacity of
the river to remove it at an equivalent rate. The cumulative deposits
would then move downstream principally during major flood events.
Extreme floods, such as those caused by Hurricanes David and Frederick
in 1979, have played a major role in this process. During the floods
a large amount of sediment (new and residual) was transported
downstream, but the aftermath of the storms also left new sediment
deposits and further scoured the stream banks and channel.
In the case of residual sediments, erosion prevention can have
little effect. The construction of sediment detention structures,
however, could reduce the delivery ratio of the eroded materials
already in or near the stream courses. Cleaning and maintenance of
such structures would be far less expensive (in money and energy
terms) than mechanical dredging (or abandonment) of large reservoirs.
In terms of erosion prevention the reduction of forest clearing and
tillage of plots in annual crops will have the greatest impact.


CHAPTER V
CONCLUSIONS
The analysis of erosion and sedimentation within a nested
hierarchy of watersheds and land use systems identifies relationships
that are not apparent at the plot or watershed scales. Spatial
analysis of land use associations within watersheds in the Plan Sierra
region indicated parallel proportions of forest and annual crop lands
in the coffee-producing areas and the extensive pasture lands
downstream. The erosion plot data extrapolated to the watershed level
result in almost uniform erosion rates for these two contrasting land _
use systems. The population density and the related demand for food
production are major determinants of erosion rates in the region.
In addition, the sequence of land uses has a strong influence on
erosion rates. The erosion plot data alone indicate a dramatic change
in erosion rates over time, with the highest rates occurring on steep
slopes immediately after deforestation. The threshholds for
destabilization and erosion of the various horizons are evidently
distinct. The change in erosion rates decays after time of clearing
for a given surface soil, but catastrophic events such as complete
removal of the A horizon or the destabilization of the lower horizons
can induce a new cycle of extremely high rates and subsequent
exponential decay. Peak erosion appears to be caused by initial
exposure and disturbance of former forest soils.
278


279
The application of the information from the study to an analysis
of the Plan Sierra conservation programs indicates a need to place
more emphasis on curbing deforestation, which is the highest single
contributor of sediment, per unit area. In already deforested lands,
changes in cropping systems, conservation and integration of cropping
and forestry systems are necessary. The first two options have
received the most emphasis within Plan Sierra, and the third has been
limited primarily to promotion of combined coffee and food crop
production systems.
The soil conservation practices and land use changes advocated by
Plan Sierra have had varying results. The hillside ditches that have
proven so effective in other areas in an experimental setting (Table
A-7) failed to show significant improvement in erosion reduction at
all four plots. The technical and economic feasibility of
constructing and maintaining this type of physical infrastructure in
small farm plots were not included in the experimental analyses that
preceded the transfer and dissemination of this technology in the
Sierra. In many areas of the Sierra, the degree of slope, soil
texture, and soil structure are not well-suited to the successful use
of hillside ditches. The excavated ditches had filled in at most of
the sites prior to completion of the study (less than 15 months). The
seeding of the upslope banks with dense grasses or other cover crops
stabilized some structures in other fields. However, the investment
required for construction and maintenance of the ditches is still
high.
The testing of this technology on existing farm plots showed it
to be inappropriate in most cases for small farmers in the Sierra.


280
According to the double criteria mentioned previously (Chapter III)
for judging the utility of a given technology, both technical
performance and goodness-of-fit within the system were considered.
This practice fails on both counts, as judged by researchers and the
farmers, respectively.
Plan Sierra has promoted the planting of coffee by farmers
willing and able to do so, with or without special credit assistance.
The result is that many of the larger growers are expanding their
coffee plantations and are improving the production of existing
stands. Many smallholders also are expanding existing coffee acreage
and many are planting former food crop plots and pastures to coffee.
The full impact of a cumulative increase in coffee acreage and a
decrease in food crop production has not yet been felt. The local
demand for food crops will increase first as a result of the demands
of the growing seasonal labor force needed to harvest the coffee.
Secondly, the conversion to coffee of small plots formerly in field
crops will decrease local subsistence production and food crops grown
for sale in local markets. This will increase pressure on adjacent
areas to produce food crops for sale to these coffee-producing areas.
The other alternative is for coffee acreage to be expanded at the
expense of forest and pasture land in the coffee region, thus
decreasing the alternative areas available for rotation with field
crops. This will further degrade the soils in the surviving food crop
plots, at the same time increasing erosion in the specific plots
planted to coffee from less erosion-prone cover such as grass or
forest.


281
Based on results from erosion plots and small watersheds, the
planting of coffee as an indirect conservation measure cannot be
justified. Based on the models of farm level production, this
strategy is also lacking as a development policy aimed at the small
farmers and the landless.
The same concept of agroforestry ecosystems based on sequential
production of food crops and permanent tree crops could be applied in
the Sierra with the substitution of food, fuel, forage or wood-
producing trees in place of coffee. Whereas coffee increases the
total population pressure on the land by drawing large numbers of
seasonal workers into the area, the production of a tree cash crop
with a less pronounced peak labor demand per area would not require
the same proportion of food to cash crop land.
For example, the production of fast-growing leguminous trees for
wood, forage, fuel or even pulp production would constitute a more
viable long-term alternative than coffee production, for two reasons.
The utility of the products would strengthen both the local and
national economy. The integrated production of small animals
(confined pigs or goats) with tree crops would improve local diet as
well as the income of smallholders. On the other hand, the spread of
coffee production would decrease the amount of food and fuel available
per household and would increase erosion and sediment export at the
watershed level.
Aside from the sediments generated by present erosion, there is a
need to deal with the residual products of past and apparently more
severe erosion. The construction of small dams in the uplands would


282
provide a mechanism for retention of sediments prior to their arrival
at the large dam sites. Such structures also could provide energy,
water supply, and irrigation for upland farms and settlements. The
integrated planning of such sedimentation control projects and farming
systems programs would protect the large dams downstream and improve
production and public services in the uplands.
The magnitude of the sedimentation problem, particularly in the
Mao watershed, indicates a need for simultaneous action to reduce the
sediment delivery at the proposed dam site, and to improve watershed
management in the upland source areas. While the sedimentation rates
are more severe than the average erosion rates, relative to reports
from other areas, the watershed average masks a very skewed
distribution of the erosion problem. The occurrence of severe
degradation and depletion of soils on smallholder plots has
selectively undermined the resource base for subsistence production.
Moreover, the current erosion rates in many fields do not reflect the
degree of damage that has been sustained at the sites during prior
periods of heavy initial soil loss.
The smallholders occupying such sites need both technical and
financial aid to rehabilitate these areas or to relocate. Technical
aid may suffice in holdings not already severely degraded. A combined
farming systems and conservation program is suggested to design
sustainable production systems at the small watershed level with
community groups and farmers' associations. The initial field trials
should include crops that maximize ground cover throughout the
cropping cycle and tree crops that offer a saleable product within 4


283
or 5 years. The integration of animal production also should be given
priority. The farming system should be extrapolated to the watershed
and regional level through models and/or cartographic analysis before
the field trials and demonstrations are finalized.


APPENDIX A
COMPARATIVE DATA FROM LITERATURE REVIEW


Table A-l. Suspended sediment concentrations in streams draining
forested watersheds in the southeastern United States.
Location
Forest type
Concentration
-1
Mg L
South Carolina Piedmontf
Loblolly Pine
20-43
Mississippi Coastal PlainJ
Pine forest
54-269
North Central Florida§
Pine flatwoods
21-81
fSource:
JSource:
§Source:
Switzer and Nelson, 1972.
Duffy et al., 1978.
Reikerk et al., 1979.
285


Table A-2. Mass transport of sediment from forested watersheds.
Location
Loss
Source
. -1 -1
kg ha yr
Arkansas, Ouchita Mountains
15
Ursic, 1978
Florida, Pine Flatwoods
129
Reikerk et al., 1978, 1979
Georgia, Piedmont
(Clearcut and Machine Prepared)
20,000f
Hewlett and Nutter,
1969
Minnesota, Deciduous Forest
37
Singer and Rush, 1975
Mississippi, Pine Forest
225
Duffy et al., 1978;
1976; Ursic, 1978
Schreiber et al. ,
New Hampshire, Hubbard Brook
(Hardwood/Pine)
33
Bormann and Likens,
1979
North Carolina, Coweeta
38-76
Johnson and Swank,
1973; Mitchell
(Hardwood/Pine)
et al., 1976; Monk,
1976
Approximately 397 kg ha cm stormflow leaving watershed.
286


237
Table A-3. Variation of sediment transport by forest type and land
use: Coweeta, North Carolina.
Vegetation
Annual Sediment Yield
kg ha ^
Mature Hardwood
30
Pine Plantation
76
Old Field
176
Coppice Hardwood
283
Source: Monk, 1976; Sopper, 1975.


Table A-4. Average erosion rates for broad land use categories
in the United States.
Land Use / Sediment Loss
kg ha 1 year
Forest
85
Grassland
850
Abandoned Mines
8500
Cropland
17,000
Harvested Forest
42,500
Construction
170,000
Source: Pavoni, 1977; USEPA, 1973


289
Table A-5. Erosion rates for selected agricultural land uses in the
United States.f
Land Use
Soil Loss
Tons ha ^
All Cropland
10.53
Cultivated Cropland
11.44
Pastureland,
including native pastures
5.83
Rangeland
6.94
Forest Land
2.69
Grazed Forest
Land
9.41
fSheet and rill erosion.
Source: USDA, 1980.


290
Table A-6. Variation in sediment yield with land use in Malaysia.
Land Use in Watershed
Sediment Yield
kg ha ^ year-1
Rain Forest
4 to 34
Tea Plantations
673
Vegetable Farms
1009
Mining
495
Urban Areas
800
Source: Douglas, 1968; Morgan, 1979.


Table A-7. Erosion rates measured in the Caribbean and Latin America (soil losses in tons ha
yearl).
Land Use
Jamaicat
Puerto
Ricot
Colombia^
Brazil^
Clean-Tilled
Fallow
375 625
253 -
303
225 -
253

Field Crops
*
100 125
15.2 -
36.4
1.0 -
20.0
9.5 -
21.5
Field Crops,
Traditional
134.4


21.4
- 21.5
Field Crops,
on Contours



4.1 -
19.8
Field Crops,
Hillside Ditch and "Hills"
38.8



Field Crops,
Hillside Ditch and Contours
26.6
--


Field Crops,
Bench Terraces
17.5



Field Crops,
with Mulch

1.4


Field Crops, with Mix of Conservation
Practices
9.8 14.8



Pasture

4.4
3.0 -
7.1

Fertilized Pasture
5.0 12.5
1.5

1.2 -
2.7
Coffee, New


1.8 -
24

Coffee, New
with Terraces


0.2
--
Coffee, Old
without Terraces


0.6

Dense Forest
0.5 1.3
i
--
--
--
fsheng, 1973; Sheng and Michaelson, 1973.
jsmith and Abruna, 1955; Vicente-Chandler, 1976.
§Suarez de Castro and Rodriguez, 1955, 1962,
^Bertoni, 1966; Marques et al., 1961.
291


292
Table A-8. Runoff and soil loss in Senegal.t
Vegetation
Runoff
Soil Loss
%
^ -1 -1
tons ha yr
Natural Bush
0.9
0.2
Peanuts
22.8
6.9
Sorghum
34.1
8.4
Maize
30.9
10.3
fRainfall of 1100 to 1300 mm year~l.
Source: Charreau, 1972; Moutappa, 1973; Okigbo, 1977.


Table A-9. Runoff and soil loss under different land use in Uganda.t
Land Use
(As percent
Runoff
of precipitation)
Soil
(tons hal
Loss ^
season )
Natural Savanna
0. 5%
0.
1
Grazed Savanna
X10
4.
4
Grazed and Burned
Savanna
X20
2.
5
Cowpeas and Corn
X40
26 to
34
Bare Fallow
X80
81.
5
fA bulk density of
converted to tons
^Runoff for other
1.3 g cc
hal.
land uses
^ is assumed for all soil losses
is reported by increase over the
reported in
runoff from
2 -i ,
m ha and
the natural
savanna.
Source: Sperow and Keefer, 1975.
291


Table A-10. Soil losses in tea plantations compared to alternative land uses (in tons ha
year"l).f
Nation
Land Use
Forest Tea with Cover Crops Tea Vegetables Bare Fallow
Malaysia^
3.2 6.4 9.8
India§
o.4 to 2.1 3.1 to 14.4
Sri Lankafl
20.0 to 35.0 4.0
fSoil loss is converted to tons hal year_l using a soil density of 1.3 g cc-3 if data is reported
in m3.
|Data for Malaysia are reported in Lai (1977b).
§Data for India are reported in Lai (1977b) and Hasselo and Sikurajapathy (1965).
lData for Sri Lanka are reported in Lai (1977b) and Holland and Joachim (1933) .


Table A-ll. Variation of runoff and soil loss with land use in upland Mindanao
Vegetation
Runoff
(As percent of
Precipitation)
Soil Loss
During Cropping Period
(kg ha-1 day-1)
Soil
Over
Loss Increments
Primary Forest
Losses
Primary Forest
0.25
0.25
X
1.0
Softwood Trees
0.26
0.36
X
1.5
Imperata Grassland
3.00
0.50
X
2.0
New Abaca Plantation
0.35
0.59
X
2.5
10 Year Old Abaca Plantation 0.64
0.74
X
3.0
Logged Forest
1.73
-

New Corn Field
1.52
3.79
X
115.0
New Rice Field
1.08
1.81
X
70.0
2 Year Old Corn Field
1.78
15.06
X
60.0
12 Year Old Rice Field
11.64
145.14
X
600.0
Source: Kellman, 1969
295


APPENDIX B
SURVEYED CROSS SECTIONS OF RIVERS AND STREAMS


297
Fig. B-l. Mao River cross section.


-sJ
Fig. B-2. Amina River cross section.
298


Fig. B-3. Upper Bajamillo Stream cross section.
Fig. B-4. Prieto Stream cross section.
Fig. B-5. Greater Bajamillo Stream cross section.
299


HONDO STREAM
T
4
r
8
12
16
Fig. B-6.
Hondo Stream cross section.
22
21
20
19
0 2 4 6 8 10 12 14
Fig. B-7. Pananao Stream cross section.
300


APPENDIX C
STAGE -DISCHARGE CURVES FOR RIVERS, DERIVED BY
INDRHI


LEVEL READINGS ON HYDROMETRIC 6UA6E
DISCHARGE (Q) IN METERS
Fig. C-l. Stage discharge curves for rivers derived by INDRHI, Mao River


LEVEL READINGS ON HYDROMETRIC GUAGE
DISCHARGE ( Q) in m3 sec'
Fig. C-2. Stage discharge curves for rivers derived by INDRHI, Amina River.
303


APPENDIX D
MONTHLY RAINFALL FOR STATIONS 3 THROUGH 11


o
r :>
U
o
ci
r~
i j
ZJ1
n
o
/j
u
IFltflJ I¡M i NF NI I
19o 1 nniNnii I
I1E FIN
Fig
D-l
Monthly rainfall at the Santo Domingo climatological station, #3


f )
CO
Li
<\J
r -
u*

io
io
c.
CO
J-
r>
*-* n
- o
cr *
o_
r\.
l_J Cl
UJ fr*
"
0_
O
-f
J
19; :n nm hi ut i
-i9fn nn i ni ni i
ncnn
o
to
ro
i
mu
Fig. D-2. Monthly rainfall at the Jarabacoa climatological station, #4.
306


o
o
OJ
o
rj
r
zr
(O
o
(U
m
1960 P'.INI ill I
1961 Fill 1 NFfil l
MFllN
d
-
cr
Cl_
-i
L.
til
cr
L
ZT
o :
rj -
ro :
Fig. D-3. Monthly rainfall at the Mao climatological station, #5.
i
307


Fig. D-4. Monthly rainfall at the Mata Grande climatological station, #6
lull
308


oS 0c OfiS G9S OE'ti Cln c£ h¡2 SI
l* W Ni Nkjlllid]jJad
Fig, D-5. Monthly rainfall at the Tavora climatological station, #7.
I
mu
309


Fig. D-6. Monthly rainfall at the Magua climatological station, #8.
OJ
o


o
Fig. D-7, Monthly rainfall at the Manabao climatological station, #9.
311


o
o
(O
o
r-
o
Z3'
iO
C3
J
ir
nmiu n.
FWINF HU
MEAN
I 'Jl.iU
Fig. D-8. Monthly rainfall at the Santiago Rodriguez climatological station, #10.
312


c:>
C'l
I
Fig. D-9.
Monthly rainfall
at the Partido climatological station,
#11.
313


APPENDIX E
DATA ON SEDIMENT CONCENTRATION, STAGE AND
DISCHARGE FOR MAO AND AMINA RIVERS


Legend for Tables E-l and E-2.
LEVLA
= hydrometric station reading, river stage, 7 A.M.
LEVLP
= hydrometric station reading, river stage, 5 P.M.
LEVLM
= hydrometric station reading, stage at flood peak.
LEVLF
= reading of river stage (distance below bridge) at sample time
SEDCA
= sediment concentration (g L "S of first sample.
SEDCB
= sediment concentration (g L of second sample.
SEDC
= average sediment concentration from both samples.
DLEVA
3 -1
= discharge rate (m sec ), derived from 7 A.M. stage.
DLEVP
3 -1
= discharge rate (m sec ), derived from 5 P.M. stage.
DLEVD
= average daily discharge rate (without flood peaks).
DLEVM
3 -1
= discharge rate (m sec ) at flood peak.
315


316
Table E-l. Mao River.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800101
0.42
0.42
-
-
-
-
-
9.6
9.6
9.6
-
800102
0.42
0.42
-
-
-
-
-
9.6
9.6
9.6
-
800103
0.61
0.46
-
-
-
-
-
17.8
11.0
14.4
-
800104
0.57
0.48
-
-
-
-
-
15.1
11.7
13.4
-
800105
0.44
0.42
-
-
-
-
-
10.3
9.6
9.9
-
800106
0.42
0.42
-
-
-
-
-
9.6
9.6
9.6
-
800107
0.65
0.55
-
-
-
-
-
20.3
14.3
17.3
-
800108
0.47
0.45
-
-
-
-
-
11.3
10.6
11.3
-
800109
0.43
0.42
-
-
-
-
-
9.9
9.6
9.8
-
800110
0.42
0.41
-
-
-
-
-
9.6
9.3
9.4
-
800111
0.41
0.40
-
-
-
-
-
9.3
8.9
9.1
-
800112
0.40
0.40
-
-
-
-
-
8.9
8.9
8.9
-
800113
0.40
0.39
-
-
-
-
-
8.9
8.6
8.8
-
800114
0.39
0.39
-
-
-
-
-
8.6
8.6
8.6
-
800115
0.39
0.39
-
-
-
-
-
8.6
8.6
8.6
-
800116
0.39
0.39
-
-
-
-
-
8.6
8.6
8.6
-
800117
0.38
0.37
-
-
-
-
-
8.3
7.9
8.1
-
800118
0.37
0.37
-
-
-
-
-
7.9
7.9
7.9
-
800119
0.37
0.37
-
-
-
-
-
7.9
7.9
7.9
-
800120
0.37
0.35
-
-
-
-
-
7.9
7.6
7.8
-
800121
0.36
0.36
-
-
-
-
-
7.6
7.6
7.6
-
800122
0.36
0.36
-
-
-
-
-
7.6
7.6
7.6
-
800123
0.35
0.35
-
-
-
-
-
7.3
7.3
7.3
-
800124
0.35
0.35
-
-
-
-
-
7.3
7.3
7.3
-
800125
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
-
800126
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
-
800127
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
-
800128
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
-
800129
0.34
0.33
-
-
-
-
-
7.0
6.7
6.8
-
800130
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800131
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800201
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800202
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800203
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800204
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800205
0.35
0.39
-
-
-
-
-
7.3
8.6
7.9
-
800206
0.34
0.33
-
-
-
-
-
7.0
6.7
6.8
-
800207
0.32
0.32
1.08
-
-
-
-
6.4
6.4
6.4
57
800208
0.87
0.55
-
-
-
-
-
36.9
14.3
25.6
-
800209
0.48
0.43
-
-
-
-
-
11.7
9.9
10.8
-
800210
0.42
0.40
-
-
-
-
-
9.6
8.9
9.3
-
800211
0.38
0.37
-
-
-
-
-
8.3
7.9
8.1
-


317
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800212
0.35
0.34
-
-
-
-
-
7.3
7.0
7.2
-
800213
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800214
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
800215
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
800216
0. 32
0.32
-
-
-
-
-
6.4
6.4
6.4
-
800217
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
-
800218
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
800219
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
800220
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
800221
0.51
0.64
-
-
-
-
-
12.8
19.6
16.2
-
800222
0.47
0.42
-
-
-
-
-
11.3
9.6
10.5
-
800223
0.39
0.37
-
-
-
-
-
8.6
7.9
8.3
800224
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
-
800225
0.34
0.33
-
-
-
-
-
7.0
6.7
6.8
-
800226
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800227
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
800228
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
800229
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
800301
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
800302
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-
800303
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-
800304
0.60
0.65
-
-
-
-
-
16.3
20.3
18.3
-
800305
0.47
0.42
-
-
-
-
-
11.3
9.6
10.5
800306
0.38
0.36
-
-
-
-
-
8.3
7.6
7.9
800307
0.82
0.59
-
-
-
-
-
32.7
15.9
24.3
800308
0.65
0.52
-
-
-
-
-
20.3
13.2
16.7
-
800309
0.44
0.43
-
-
-
-
-
10.3
9.9
10.1
-
800310
0.40
0.38
-
-
-
-
-
8.9
8.3
8.6
800311
0.37
0.36
-
-
-
-
-
7.9
7.6
7.8
800312
0.35
0.35
-
-
-
-
-
7.3
7.3
7.3
-
800313
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
-
800314
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800315
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
800316
0.32
0.30
-
-
-
-
-
6.4
5.8
6.1
-
800317
0.30
0.32
-
-
-
-
-
5.8
6.4
6.1
800318
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
800319
0.31
0. 31
-
-
-
-
-
6.1
6.1
6.1
800320
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
800321
0.30
0.30
-
145
0.021
0.021
0.021
5.8
5.8
5.8
800322
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-
800323
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-
800324
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-


318
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800325
0.29
0.29
-
-
-
-
-
5.5
5.5
5.5
-
800326
0.29
0.29
-
-
-
-
-
5.5
5.5
5.5
-
800327
0.29
0.29
-
145
0.013
0.013
0.013
5.5
5.5
5.5
-
800328
0.28
0.28
-
-
-
-
-
5.2
5.2
5.2
-
800329
0.28
0.28
-
-
-
-
-
5.2
5.2
5.2
-
800330
0.28
0.28
-
-
-
-
-
5.2
5.2
5.2
-
800331
0.28
0.28
-
-
-
-
-
5.2
5.2
5.2
-
800401
-
-
-
-
-
-
-
-
-
-
-
800402
-
-
-
-
-
-
-
-
-
-
-
800403
-
-
-
-
-
-
-
-
-
-
-
800404
-
-
-
-
-
-
-
-
-
-
-
800405
-
-
-
-
-
-
-
-
-
-
-
800406
-
-
-
-
-
-
-
-
-
-
-
800407
-
-
-
-
-
-
-
-
-
-
-
800408
-
-
-
-
-
-
-
-
-
-
-
800409
-
-
-
-
0.224
-
-
-
-
-
-
800410
-
-
-
-
-
-
-
-
-
-
-
800411
-
-
-
-
-
-
-
-
-
-
-
800412
-
-
-
-
-
-
-
-
-
-
-
800413
-
-
-
-
-
-
-
-
-
-
-
800414
-
-
-
-
-
-
-
-
-
-
-
800415
-
-
-
-
-
-
-
-
-
-
-
800416
-
-
-
-
-
-
-
-
-
-
-
800417
-
-
-
-
-
-
-
-
-
-
-
800418
-
-
-
-
-
-
-
-
-
-
-
800419
-
-
-
-
-
-
-
-
-
-
-
800420
-
-
-
-
-
-
-
-
-
-
-
800421
-
-
-
-
-
-
-
-
-
-
-
800422
-
-
-
-
-
-
-
-
-
-
-
800423
-
-
-
-
-
-
-
-
-
-
-
800424
-
-
-
-
-
-
-
-
-
-
-
800425
-
-
-
-
-
-
-
-
-
-
-
800426
-
-
-
-
-
-
-
-
-
-
-
800427
-
-
-
-
-
-
-
-
-
-
-
800429
-
-
-
-
-
-
-
-
-
-
-
800430
-
-
-
-
-
-
-
-
-
-
-
800501
0.75
0.80
1.98
-
-
-
-
27.2
31.1
29.1
151.2
800502
1.32
1.00
1.41
-
-
-
-
87.0
49.2
68.1
94.8
800503
1.17
1.04
1.63
-
-
-
-
67.9
53.3
60.6
115.7
800504
0.98
0.85
-
-
-
-
-
47.2
35.2
41.2
-
800505
0.77
0.73
1.40
-
-
-
-
28.7
25.7
27.2
93.9


319
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800506
0.87
0.81
1.56
-
-
-
-
36.9
31.9
34.4
108.9
800507
0.85
1.20
1.30
-
-
-
-
35.2
71.5
53.3
84.8
800508
0.96
0.84
1.10
-
-
-
-
45.2
34.3
39.8
59.8
800509
1.05
0.82
-
-
-
-
-
54.3
32.7
43.5
-
800510
0.80
0.78
1.05
-
-
-
-
31.1
29.5
30.3
54.3
800511
0.88
0.79
1.65
-
-
-
-
37.8
30.3
34.0
117.7
800512
1.06
0.87
-
-
-
-
-
55.4
36.9
46.2
-
800313
0.90
0.78
1.16
-
-
-
-
39.6
29.5
34.5
66.7
800514
0.93
0.84
1.08
-
-
-
-
42.3
34.3
38.3
57.6
800515
0.83
1.17
-
-
-
-
-
33.5
67.9
50.7
-
800516
0.95
0.92
1.63
-
-
-
-
44.2
41.4
42.8
115.7
800517
1.18
0.98
1.52
-
-
-
-
69.1
47.2
58.1
105.1
800518
1.03
0.92
2.12
-
-
-
-
52.2
41.4
46.8
166.1
800519
1.34
1..16
3.04
-
-
-
-
89.7
66.7
78.2
272.8
800520
1.30
1.21
-
-
-
-
84.3
72.7
78.5
-
800521
1.06
0.96
-
-
-
-
-
55.4
45.2
50.3
-
800522
1.03
0.95
-
-
-
-
-
52.2
44.2
48.2
-
800523
0.96
0.97
1.53
-
-
-
-
45.2
46.2
45.7
106.0
800524
1.07
0.94
2.14
-
-
-
-
56.5
43.3
49.4
168.3
800525
1.38
1.21
1.74
-
-
-
-
95.3
72.7
84.0
126.6
800526
1.12
1.07
1.59
-
-
-
-
62.0
56.5
59.3
111.8
800527
1.10
0.98
2.50
4
1.295
1.828
1.56
59.8
47.2
53.5
208.4
800528
1.32
1.56
2.75
-
0.168
-
-
87.0
122.6
104.8
237.6
800529
1.75
1.39
-
-
-
-
-
155.2
96.7
126.0
-
800530
1.20
1.15
2.30
-
-
-
-
71.5
65.5
68.5
185.8
800531
1.33
1.17
-
-
-
-
-
88.3
67.9
78.1
-
800601
1.44
1.35
-
81
0.257
0.217
0.24
104.0
91.1
97.5
-
800602
1.25
1.15
-
-
-
-
-
77.8
65.5
71.6
-
800603
1.05
0.99
-
-
-
-
-
54.3
48.1
51.2
-
800604
1.10
1.00
-
-
-
-
-
59.8
49.2
54.4
-
800605
1.22
1.10
-
-
-
-
-
74.0
59.8
66.9
-
800606
1.12
1.08
-
-
-
-
-
62.0
57.6
59.8
-
800607
0.96
0.93
-
-
-
-
-
45.2
42.3
43.8
-
800608
0.94
0.90
-
-
-
-
-
43.3
39.6
41.4
-
800609
0.92
0.88
-
-
-
-
-
41.4
37.8
39.6
-
800610
0.87
0.85
-
-
-
-
-
36.9
35.2
36.1
-
800611
1.17
1.65
-
-
-
-
-
67.9
137.5
102.7
-
800612
1.32
1.14
-
-
-
-
-
87.0
64.3
75.7
-
800613
1.30
1.32
-
-
-
-
-
84.3
87.0
85.6
-
800614
1.14
1.08
-
-
-
-
-
64.3
57.6
61.0
-
800615
1.00
0.96
-
-
-
-
-
49.2
45.2
47.2
-
800616
0.98
0.94
1.10
-
-
-
-
47.2
43.3
45.2
59.8


320
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800617
0.96
0.93
1.71
-
-
-
-
45.2
42.3
43.8
123.6
800618
1.10
1.06
2.10
100
0.202
0.169
0.
18
59.8
55.4
57.6
164.0
800619
1.13
1.05
-
71
0.170
-
-
63.2
54.3
58.8
-
800620
0.98
0.93
1.75
-
-
-
-
47.2
42.3
44.7
127.6
800621
1.07
0.98
1.62
104
0.150
0.174
0.
16
56.5
47.2
51.8
114.7
800622
1.12
1.00
1.18
-
-
-
-
62.0
49.2
55.6
69.1
800623
0.94
0.94
1.35
104
0.280
0.426
0.
35
43.3
43.3
43.3
69.1
800624
0.99
0.94
-
-
-
-
-
48.1
43.3
45.7
-
800625
0.90
0.87
1.08
-
-
-
-
39.6
36.9
38.2
57.6
800626
0.95
0.88
-
-
-
-
-
44.2
37.8
41.0
-
800627
0.85
0.83
1.14
-
-
-
-
35.2
33.5
34.4
64.3
800628
0.86
0.83
-
-
-
-
36.0
33.5
34.8
-
800629
0.82
0.80
-
-
-
-
-
32.7
31.1
31.9
-
800630
0.78
0.77
1.50
-
-
-
-
29.5
28.7
29.1
103.2
800701
0.88
0.82
1.02
-
-
-
-
37.8
32.7
35.2
51.2
800702
0.88
0.79
-
-
-
-
-
37.8
30.3
34.0
-
800703
0.77
0.76
0.94
-
-
-
-
28.7
28.0
28.3
43.3
800704
0.79
0.76
-
-
-
-
-
30.3
28.0
29.1
-
800705
0.74
0.71
1.20
-
-
-
-
26.5
25.0
25.7
71.5
800706
0.80
0.76
1.00
-
-
-
-
31.1
28.0
29.5
49.2
800707
0.80
0.78
1.22
-
-
-
-
31.1
29.5
30.3
77.7
800708
0.85
0.80
-
-
-
-
-
35.2
31.1
33.1
-
800709
0.76
0.74
-
-
-
-
-
28.0
26.5
27.2
-
800710
0.75
0.73
1.04
-
-
-
-
27.2
25.7
26.5
53.3
800711
0.84
0.85
1.06
-
-
-
-
34.3
35.2
34.8
55.4
800712
0.77
0.73
-
-
-
-
-
28.7
25.7
27.2
-
800713
0.72
0.70
-
-
-
-
-
25.0
23.6
24.3
-
800714
0.69
0.68
-
-
-
-
-
22.9
22.2
22.6
-
800715
0.68
0.67
1.28
-
-
-
-
22.2
21.6
21.9
83.0
800716
0.79
0.73
1.13
-
-
-
-
30.3
25.7
28.0
63.2
800717
0.83
0.76
-
-
-
-
-
33.5
28.0
30.7
-
800718
0.76
0.73
-
-
-
-
-
28.0
25.7
26.8
-
800719
0.72
0.70
1.56
-
-
-
-
25.0
23.6
24.3
108.9
800720
1.04
0.88
1.32
-
-
-
-
53.3
37.8
45.5
86.6
800721
0.98
0.93
1.10
-
-
-
-
47.2
42.3
44.7
59.8
800722
0.88
0.83
1.34
-
-
-
-
37.8
33.5
35.7
88.4
800723
0.90
0.87
1.80
-
-
-
-
39.6
36.9
38.2
132.6
800724
0.95
0.86
-
-
-
-
-
44.2
36.0
40.1
-
800725
0.81
0.78
-
119
0.041
0.042
0.
05
31.9
29.5
30.7
-
800726
0.76
0.74
-
-
-
-
-
28.0
26.5
27.2
-
800727
0.72
0.72
-
-
-
-
-
25.0
25.0
25.0
-
800728
0.71
0.70
-
-
-
-
-
24.3
23.6
24.0
-


321
Table E-l~-continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800729
0.69
0.68
-
-
-
-
-
22.9
22.2
22.6
-
800730
0.66
0.65
-
-
-
-
-
20.9
20.3
20.6
-
800731
0.65
0.64
-
-
-
-
-
20.6
19.6
19.9
-
800801
0.63
0.62
1.14
-
-
-
-
19.0
18.4
18.7
-
800802
0.77
0.67
0.91
-
-
-
-
28.7
21.6
25.1
40.5
800803
0.75
0.66
-
-
-
-
-
27.2
10.9
24.1
-
800804
0.65
0.64
-
-
-
-
-
20.3
19.6
19.9
-
800805
0.63
0.65
1.12
-
-
-
-
19.0
20.3
10.6
62.0
800806
0.99
0.77
-
171
0.147
-
-
48.1
28.7
38.4
-
800807
0.70
0.67
-
-
-
-
-
23.6
21.6
22.6
-
800808
0.64
0.63
1.75
-
-
-
-
19.6
19.0
19.3
127.6
800809
0.89
0.77
1.02
-
-
-
-
38.7
28.7
33.7
51.2
800810
0.75
0.71
1.54
-
-
-
-
27.2
24.3
25.8
107.0
800811
0.95
0.80
-
-
-
-
-
44.2
31.1
37.7
-
800812
0.76
0.73
-
-
-
-
-
28.0
25.7
26.8
-
800813
0.69
0.67
-
-
-
-
-
22.9
21.6
22.2
-
800814
0.72
0.65
-
-
-
-
-
25.0
20.3
22.6
-
800815
0.65
0.64
1.00
-
-
-
-
20.3
19.6
19.9
49.2
800816
0.77
0.66
-
-
-
-
-
28.7
20.9
24.8
-
800817
0.65
0.64
-
-
-
-
-
20.3
19.6
19.9
-
800818
0.63
0.62
-
-
-
-
-
19.0
18.4
18.7
-
800819
0.60
0.58
0.66
-
-
-
-
16.3
15.5
15.9
20.9
800820
0.63
0.64
1.26
-
-
-
-
19.0
19.6
10.3
81.
8008212
0.73
0.64
0.96
-
-
-
-
25.7
19.6
22.7
45.2
800822
0.68
0.65
0.85
171
0.067
0.052
0.06
22.2
20.3
21.3
35.2
800823
0.67
0.64
0.82
-
-
-
-
21.6
19.6
20.6
32.7
800824
0.68
0.63
0.80
-
-
-
-
22.2
19.0
20.6
31.1
800825
0.67
0.64
0.86
-
-
-
-
21.6
19.6
20.6
36.0
800826
0.68
0.66
1.45
-
-
-
-
22.2
20.9
21.6
98.5
800827
0.90
0.77
1.18
-
-
-
-
39.6
28.7
34.1
69.1
800828
0.83
0.75
-
-
-
-
-
33.5
27.2
30.4
-
800829
0.70
0.68
1.00
-
-
-
-
23.6
22.2
22.9
49.2
800830
0.80
0.75
0.80
-
-
-
-
31.1
27.2
29.1
31.1
800831
0.77
0.73
0.94
-
-
-
-
28.7
25.7
27.2
43.3
800901
0.78
0.73
1.05
-
0.088
0.075
0.08
29.5
25.7
27.6
54.3
800902
0.78
0.73
-
-
-
-
-
29.5
25.7
27.6
-
800903
0.71
0.68
0.92
-
0.051
-
-
24.3
22.2
23.3
41.4
800904
0.75
0.71
-
-
0.027
-
-
27.2
24.3
25.8
-
800905
0.71
0.67
2.66
-
0.144
0.165
0.16
62.0
36.9
49.5
92.0
800907
0.96
0.87
2.16
-
-
-
-
45.2
36.9
41.1
170.4
800908
0.95
0.85
1.14
-
-
-
-
44.2
35.2
39.7
64.3


322
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800909
1.02
0.86
1.18
-
-
-
-
51.2
36.0
43.6
69.1
800910
0.81
0.80
-
-
-
-
-
31.9
31.1
31.5
-
800911
0.85
0.80
1.62
110
0.162
-
-
35.2
31.1
33.1
114.7
800912
0.92
0.85
-
-
-
-
-
41.4
35.2
38.3
-
800913
0.80
0.78
1.26
-
-
-
-
31.1
29.5
30.3
81.2
800914
0.87
0.82
1.15
93
0.584
-
-
36.9
32.7
34.8
65.6
800915
0.87
0.82
1.10
-
-
-
-
36.9
32.7
34.8
59.8
800916
0.84
0.78
1.12
-
-
-
-
34.3
29.5
31.9
62.0
800917
0.87
0.76
-
-
-
-
-
36.9
28.0
32.4
-
800918
0.73
0.71
-
-
-
-
-
25.7
24.3
25.0
-
800919
0.70
0.69
-
-
0.034
-
-
23.6
22.9
23.2
-
800920
0.67
0.67
0.89
-
-
-
-
21.6
21.6
21.6
38.7
800921
0.68
0.67
-
-
-
-
-
22.2
21.6
21.9
-
800922
0.66
0.65
1.05
-
-
-
-
20.9
20.3
20.6
54.3
800923
0.71
0.66
0.90
-
-
-
-
24.3
20.9
22.6
39.6
800924
0.70
0.66
-
-
-
-
-
23.6
20.9
22.3
-
800925
0.70
0.66
1.33
-
-
-
-
23.6
20.9
22.3
87.5
800926
0.80
0.73
0.83
90
0.012
0.861
0.44
31.1
25.7
28.4
33.5
800927
0.92
0.69
1.38
-
-
-
- .
41.4
22.9
21.2
92.0
800928
0.89
0.78
1.65
-
-
-
-
38.7
29.5
34'. 1
117.7
800929
0.89
0.80
1.80
-
-
-
-
38.7
31.1
34.9
132.6
800930
1.07
0.87
1.45
85
0.168
-
-
56.6
36.9
46.7
98.5
801001
1.16
0.92
1.45
-
-
-
-
66.7
41.4
54.0
98.5
801002
0.95
0.89
1.22
110
0.105
-
-
44.2
38.7
41.5
77.7
801003
0.89
0.85
-
-
-
-
-
38.7
35.2
36.9
-
801004
0.84
0.82
2.28
-
-
-
-
34.3
32.7
33.5
183.6
801005
1.04
0.92
2.14
-
-
-
-
53.3
41.4
47.3
168.6
801006
1.09
0.96
2.20
-
-
-
-
58.7
45.2
51.9
174.8
801007
0.90
0.86
1.65
-
-
-
-
39.6
36.0
37.8
117.7
801008
1.02
0.92
1.58
-
-
-
-
51.2
41.4
46.3
110.8
801009
0.96
0.90
1.87
-
-
-
-
45.2
39.6
42.4
139.8
801010
0.96
0.92
2.30
17
-
-
-
45.2
41.4
43.3
185.8
801011
1.27
1.00
1.44
-
-
-
-
80.3
49.2
64.7
97.6
801012
0.97
0.91
1.60
-
-
-
-
46.2
40.5
43.3
112.8
801013
0.99
0.89
0.97
-
-
-
-
48.1
-
43.4
46.2
801014
0.86
0.83
-
115
0.055
0.060
0.06
36.0
-
34.8
-
801015
0.80
0.78
-
119
0.038
-
-
31.1
-
30.3
-
801016
0.76
0.75
-
-
-
-
-
28.0
-
27.6
-
801017
0.74
0.73
1.31
-
-
-
-
26.5
-
26.1
85.6
801018
0.80
0.75
1.10
121
0.774
-
-
31.1
-
29.1
59.8
801019
0.85
0.83
1.65
-
0.392
-
-
35.2
-
34.4
117.7
801020
0.92
0.84
-
115
0.080
0.090
0.09
41.4
-
37.9
-


323
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
801021
0.83
0.80
1.25
-
-
-
-
33.5
-
32.3
80.3-
801022
0.94
0.87
1.95
-
-
-
-
43.3
-
45.1
148.1
801023
1.00
0.92
-
-
-
-
-
49.2
-
45.3
-
801024
0.85
0.82
-
-
-
-
-
35.2
-
33.9
-
801025
0.80
0.78
-
-
-
-
-
31.1
-
30.3
-
801026
0.76
0.75
1.10
-
-
-
-
28.0
-
27.6
59.8
801027
0.89
0.98
-
-
-
-
-
38.7
-
34.5
47.2 .
801028
0.81
0.78
-
Ill
0.210
-
-
31.9
-
30.7
-
801029
0.73
0.72
-
-
-
-
-
25.7
-
25.4
-
801030
0.71
0.70
-
-
-
-
-
24.3
-
24.0
-
801031
0.70
0.69
-
-
-
-
-
23.6
-
23.2
-
801101
0.69
0.68
-
-
-
-
-
22.9
-
22.6
-
801102
0.68
0.68
-
-
-
-
-
22.2
-
22.2
-
801013
0.67
0.66
-
-
-
-
-
21.6
-
21.2
-
801014
0.63
0.64
-
-
-
-
-
19.0
-
19.3
-
801105
0.63
0.63
1.22
89
0.134
0.275
0.20
19.0
-
19.0
78.5
801106
0.73
0.66
2.08
100
0.326
0.286
0.31
25.7
-
23.2
161.8
801017
1.10
0.86
-
-
-
-
-
59.8
-
47.9
-
801108
0.76
0.69
-
-
-
-
-
28.0
-
25.4
-
801109
0.68
0.66
-
-
-
-
-
22.2
-
21.6
-
801110
0.65
0.64
-
-
-
-
-
20.3
-
19.9
-
801111
0.64
0.63
-
-
-
- '
-
19.6
-
19.3
-
801112
0.63
0.62
-
-
-
-
-
19.0
-
18.7
-
801113
0.62
0.61
-
-
-
-
-
18.4
-
18.1
-
801114
0.61
0.60
-
-
-
-
-
17.8
-
17.0
-
801115
0.60
0.59
-
-
-
-
-
16.3
-
16.1
-
801116
0.58
0.58
-
-
-
-
-
15.5
-
15.5
-
801117
0.74
0.59
-
-
-
-
-
26.5
-
21.2
-
801118
0.57
0.56
-
-
-
-
-
15.1
14.7
14.9
-
801119
0.55
0.55
-
-
-
-
-
14.3
14.3
14.3
-
801120
0.55
0.54
-
-
-
-
-
14.3
13.9
14.1
-
801121
0.54
0.53
-
-
-
-
-
13.9
13.6
13.8
-
801122
0.53
0.52
-
-
-
-
-
13.6
13.2
13.4
-
801123
0.52
0.51
-
-
-
-
-
13.2
12.8
13.0
-
801124
0.51
0.56
-
-
-
-
-
12.8
14.7
13.8
-
801125
0.50
0.50
-
-
-
-
-
12.4
12.4
12.4
-
801126
0.49
0.49
-
-
-
-
-
12.1
12.1
12.1
-
801127
0.49
0.49
-
120
0.216
-
-
12.1
12.1
12.1
-
801128
0.49
0.49
-
-
-
-
-
12.1
12.1
12.1
-
801129
0.48
0.48
-
-
-
-
-
11.7
11.7
11.7
-
801130
0.48
0.48
-
-
-
-
-
11.7
11.7
11.7
-
801201
0.47
0.47
-
-
-
-
-
11.3
11.3
11.3
-


324
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
801202
0.47
0.47
-
-
-
-
-
11.3
11.3
11.3
-
801203
0.46
0.46
-
-
-
-
-
11.0
11.0
11.0
-
801204
0.46
0.46
-
-
-
-
-
11.0
11.0
11.0
-
801205
0.45
0.45
-
-
-
-
-
10.6
10.6
10.6
-
801206
0.70
0.60
-
120
0.339
0.267
0.30
12.6
16.3
20.0
-
801207
0.81
0.66
-
-
-
-
-
31.9
20.96
26.4
-
801208
0.55
0.52
-
-
-
-
-
14.3
13.2
13.8
-
801209
0.50
0.48
-
-
-
-
-
12.4
11.7
12.1
-
801210
0.47
0.47
-
-
-
-
-
11.3
11.3
11.3
-
801211
0.46
0.45
-
-
-
-
-
11.0
10.6
10.8
-
801212
0.45
0.45
-
141
0.175
-
-
10.6
10.6
10.6
-
801213
0.45
0.45
-
-
-
-
-
10.6
10.6
10.6
-
801214
0.44
0.44
-
-
-
-
-
10.3
10.3
10.3
-
801215
0.48
0.47
-
-
-
-
-
11.7
11.3
11.5
-
801216
0.46
0.46
-
-
-
-
-
11.0
11.0
11.0
-
801217
0.45
0.44
-
-
-
-
-
10.6
10.3
10.5
-
801218
0.44
0.43
1.77
142
0.282
-
-
10.3
9.9
10.1
129.6
801219
1.02
0.70
-
-
-
-
-
51.2
12.6
37.4
-
801220
0.65
0.55
-
133
0.273
0.273
0.27
20.3
14.3
17.3
-
801221
0.50
0.49
-
-
-
-
-
12.4
12.1
12.3
-
801222
0.53
0.48
-
-
-
-
-
13.6
11.7
12.6
-
801223
0.47
0.46
-
-
-
-
-
11.3
11.0
11.2
-
801224
0.44
0.48
-
-
-
-
-
10.3
11.7
11.0
-
801225
0.44
0.43
-
-
-
-
-
10.3
9.9
10.1
-
801226
0.43
0.43
-
-
-
-
-
9.9
9.9
9.9
-
801227
0.43
0.43
-
-
-
-
-
9.9
9.9
9.9
-
801228
0.44
0.42
-
-
-
-
-
10.3
9.6
9.9
-
801229
0.42
0.42
-
142
0.148
0.222
0.19
9.6
9.6
9.6
-
801230
0.42
0.41
-
-
-
-
-
9.6
9.3
9.4
-
801231
0.42
0.41
-
-
-
-
-
9.6
9.3
9.4
-
810101
0.40
0.40
-
-
-
-
-
8.9
8.9
8.9
-
810102
0.40
0.40
-
-
-
-
-
8.9
8.9
8.9
-
810103
0.40
0.39
-
-
-
-
-
8.9
8.6
8.8
-
810104
0.39
0.39
-
-
-
-
-
8.6
8.6
8.6
-
810105
0.39
0.40
-
-
-
-
-
8.6
8.9
8.8
-
810106
1.35
0.90
-
100
-
-
-
91.1
39.6
65.3
-
810107
0.75
0.65
1.15
90
0.450
0.450
0.45
27.2
20.3
23.7
65.5
810108
0.85
0.72
-
114
0.286
0.224
0.26
35.2
25.0
30.1
-
810109
0.61
0.56
-
-
-
-
-
17.8
14.7
16.3
-
810110
0.54
0.55
0.90
134
0.292
-
-
13.9
14.3
14.1
39.6
810111
0.75
0.63
-
128
0.234
0.220
0.23
27.2
19.0
23.1
-
810112
0.68
0.60
-
-
-
-
-
22.2
16.3
19.3
-


325
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
810113
0.65
0.62
-
120
0.175
-
-
20.3
18.4
19.3
-
810114
0.60
0.58
-
-
-
-
-
16.3
15.5
15.9
-
810115
0.56
0.55
-
-
-
-
-
14.7
14.3
14.5
-
810116
0.53
0.52
-
-
-
-
-
13.6
13.2
13.4
-
810117
0.53
0.51
-
-
-
-
-
13.6
12.8
13.2
-
810118
0.86
0.68
-
-
0.114
-
-
36.0
22.2
29.1
-
810119
0.70
0.69
-
-
-
-
-
23.6
22.9
23.3
-
810120
0.63
0.60
-
131
0.603
-
-
19.0
16.3
17.7
-
810121
0.51
0.54
-
-
-
-
-
12.8
13.9
13.4
-
810122
0.53
0.51
-
-
-
-
-
13.6
12.8
13.2
-
810123
0.50
0.49
-
-
-
-
-
12.4
12.1
12.3
-
810124
0.48
0.48
-
-
-
-
-
11.7
11.7
11.7
-
810125
0.47
0.47
-
-
-
-
-
11.3
11.3
11.3
-
810126
0.47
0.46
-
-
-
-
-
11.3
11.0
11.2
-
810127
0.46
0.45
-
-
-
-
-
11.0
10.6
10.8
-
810128
0.44
0.44
-
-
-
-
-
10.3
10.3
10.3
-
810129
0.44
0.44
-
98
0.172
0.182
0.18
10.3
10.3
10.3
-
810130
0.46
0.43
-
-
-
-
-
11.0
9.9
10.5
-
810131
0.43
0.42
-
-
-
-
-
9.9
9.6
28.9
-
810201
0.58
0.93
0.86
113
0.470
0.470
0.47
15.5
42.3
28.9
36.0
810202
.77
0.66
-
-
-
-
-
28.7
20.9
24.8
-
810203
0.57
0.53
1.60
-
-
-
-
15.1
13.6
14.3
112.8
810204
0.85
0.66
1.04
-
-
-
-
35.2
20.9
28.1
157.6
810205
1.24
0.78
-
-
-
-
-
76.5
29.5
53.0
-
810206
0.68
0.63
-
-
-
-
-
22.2
19.0
20.0
-
810207
0.58
0.56
-
-
-
-
-
15.5
14.7
15.1
-
810208
0.54
0.53
-
-
-
-
-
13.9
13.6
13.8
-
810209
0.52
0.51
-
-
-
-
-
13.2
12.8
13.0
-
810210
0.51
0.49
-
-
-
-
-
12.8
12.1
12.4
-
810211
0.48
0.48
-
-
-
-
-
11.7
11.7
11.7
-
810212
0.47
0.47
-
-
-
-
-
11.3
11.3
11.3
-
810213
0.46
0.46
-
-
-
-
-
11.0
11.0
11.0
-
810214
0.45
0.45
-
-
-
-
-
10.6
10.6
10.6
-
810215
0.45
0.45
-
-
-
-
-
10.6
10.6
10.6
-
810216
0.47
0.45
-
-
-
-
-
11.3
10.6
11.0
-
810217
0.53
0.46
-
-
-
-
-
13.6
11.0
12.3
-
810218
0.44
0.44
-
-
-
-
-
10.3
10.3
10.3
-
810219
0.43
0.43
-
-
-
-
-
9.9
9.9
9.9
-
810220
0.42
0.50
0.80
-
-
-
-
9.6
12.4
11.0
31.1
810221
0.62
0.50
-
133
0.229
0.240
0.23
18.4
12.4
15.4
-
810222
0.46
0.45
-
-
-
-
-
11.0
10.6
10.8
-
810223
0.43
0.42
-
-
-
-
-
9.9
9.6
9.8
-


326
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
810224
0.42
0.41
-
-
-
-
-
9.6
9.3
9.4

810225
0.41
0.41
-
-
-
-
-
9.3
9.3
9.3

810226
0.40
0.40
-
-
-
-
-
8.9
8.9
8.9
810227
0.40
0.40
-
-
-
-
-
8.9
8.9
8.9
810228
0.40
0.40
-
-
-
-
-
8.9
8.9
8.9
-
810301
0.40
0.39
-
-
-
-
-
8.9
8.6
8.8
810302
0.39
0.39
-
-
-
-
-
8.6
8.6
8.6
-
810303
0.38
0.38
-
-
-
-
-
8.3
8.3
8.3
-
810304
0.38
0.38
-
-
-
-
-
8.3
8.3
8.3
810305
0.38
0.38
-
-
-
-
-
8.3
8.3
8.3
-
810306
0.38
0.37
-
-
-
-
-
8.3
7.9
8.1

810307
0.37
0.37
-
-
-
-
-
7.9
7.9
7.9
-
810308
0.36
0.36
-
-
-
-
-
7.6
7.6
7.6

810309
0.36
0.36
-
-
-
-
-
7.6
7.6
7.6
-
810310
0.36
0.35
-
-
-
-
-
7.6
7.3
7.5
-
810311
0.35
0.35
-
-
-
-
-
7.3
7.3
7.3
-
910312
0.35
0.35
-
-
-
-
-
7.3
7.3
7.3

810313
0.35
0.34
-
150
0.297
-
-
7.3
7.0
7.2
-
810314
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
810315
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0

810316
0.37
0.40
-
-
-
-
-
7.9
8'. 9
8.4

810317
0.35
0.33
-
-
-
-
-
7.3
6.7
7.0

810318
0.33
0.34
-
-
-
-
-
6.7
7.0
6.8
-
810319
0.33
0.32
-
65
-
-
-
6.7
6.4
6.5

810320
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4

810321
0.33
0.34
-
150
0.176
0.665
0.43
6.7
7.0
6.8
-
810322
0.34
0.33
-
-
-
-
-
7.0
6.7
6.8
810323
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4

810324
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
810325
0.33
0.34
-
-
-
-
-
6.7
7.0
6.8

810326
0.35
0.34
-
150
0.174
0.164
0.17
7.3
7.0
7.2
-
810327
0.45
0.57
-
128
-
-
-
10.6
15.1
12.9
-
810328
0.44
0.42
0.98
-
-
-
-
10.3
9.6
9.9
47.2
810329
0.85
0.65
-
125
0.10
-
-
35.2
20.3
27.7
-
810330
0.49
0.45
-
-
-
-
-
12.1
10.6
11.3
-
810331
0.41
0.39
-
-
-
-
-
9.3
8.6
8.9
-
810401
0.37
0.36
-
-
-
-
-
7.9.
7.6
7.8
-
810402
0.35
0.35
-
-
-
-
-
7.3
7.3
7.3
-
810403
0.34
0.34
-
-
-
-
-
7.0
7.0
7.0
-
810404
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
810405
0. 32
0.32
-
-
-
-
-
6.4
6.4
6.4
-
810406
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
-


327
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
810407
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
810408
0.32
0.34
-
-
-
-
-
6.4
7.0
6.7
-
810409
0.41
0.46
-
140
0.175
2.622
1.40
9.3
11.0
10.1
-
810410
0.37
0.34
-
-
-
-
-
7.9
7.0
7.5
-
810411
0.35
0.34
-
-
-
-
-
7.3
7.0
7.2
-
010412
0.33
0.33
-
-
-
-
-
6.7
6.7
6.7
-
810413
0.32
0.32
-
-
-
-
-
6.4
6.4
6.4
-
810414
0.32
0.31
-
-
-
-
-
6.4
6.1
6.2
810415
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
810416
0.31
0.31
-
-
-
-
-
6.1
6.1
6.1
-
810417
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
810418
0.30
0.30
-
-
-
-
-
5.8
5.8
5.8
-
810419
0.29
0.29
-
-
-
-
-
5.5
5.5
5.5*
-
810420
0.29
0.29
-
-
-
-
-
5.5
5.5
5.5
-
810421
0.29
0.29
-
-
-
-
-
5.5
5.5
5.5
-
810422
0.31
0.29
-
-
-
-
-
6.1
5.5
5.8
-
810423
0.35
0.30
-
150
0.259
0.246
0.25
7.3
5.8
6.6
810424
0.30
0.29
-
-
-
-
-
5.8
5.5
5.7
-
810425
0.33
0.32
-
-
-
-
-
6.7
6.4
6.5
-
810426
0.29
0.29
-
-
-
-
-
5.5
5.5
5.5
-
810427
0.29
0.34
1.64
-
-
-
-
5.5
7.0
6.3
116.7
810428
1.18
0.71
1.34
0
1.385
1.385
1.38
69.1
24.3
46.7
88.4
810429
1.14
0.74
1.40
-
-
-
-
64.3
26.5
45.4
93.9
810430
0.83
0.64
1.04
-
-
-
-
33.5
19.6
26.2
53.3
810501
0.86
0.70
2.12
-
-
-
-
36.0
23.6
29.8
166.1
810502
1.15
0.85
1.95
41
1.811
-
-
65.5
35.2
50.4
148.1
810503
1.68
1.17
3.95
93
0.383
0.386
0.38
142.7
67.92105.3
391.1
810504
1.83
1.37
1.80
-
-
-
-
170.1
93.9
132.0
132.6
810505
1.27
1.21
1.85
0
3.229
0.700
1.96
80.3
72.7
76.5
137.6
810506
1.34
1.22
1.24
-
-
-
-
89.7
74.0
81.8
79.4
810507
1.18
1.03
-
-0
0.113
-
-
69.1
52.2
60.7
-
810508
0.96
0.94
1.14
95
1.886
-
-
45.1
43.3
44.2
64.3
810509
1.03
1.45
2.80
0
2.572
1.315
1.94
52.2
105.5
78.9
243.6
810510
1.92
1.39
2.08
-
-
-
-
187.8
96.7
142.2
161.8
810511
1.55
1.35
2.48
0
1.537
2.888
2.13
121.0
91.1
106.0
206.1
810512
1.67
1.42
2.24
0
0.275
0.228
0.25
141.0
101.0
121.0
179.2
810513
1.53
1.72
2.06
0
0.269
1.144
0.71
117.8
149.8
133.8
159.7
810514
1.55
1.40
-
-
-
-
-
121.0
98.1
109.5
-
810515
1.47
2.35
-
200
0.198
0.494
0.35
108.5
260.1
184.3
-
810516
1.51
1.39
2.74
-
-
-
-
114.6
96.7
105.7
.236.4
810517
1.90
1.51
1.70
0
1.067
-
-
183.8
114.6
149.2
122.6
810518
1.37
1.35
2.70
-
0.051
0.982
0.52
93.9
91.1
92.5
231.7


328
Table E-l--continued.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
810519
1.51
1.38
-
-
-
-
-
114.6
95.3
105.0
-
810520
1.26
1.20
-
-
-
-
-
79.0
71.5
75.3
-
810521
1.16
1.11
-
-
-
-
-
66.7
60.9
63.8
-
810522
1.09
1.02
1.40
-
-
-
-
58.7
51.2
54.9
93.9
810523
1.10
1.17
-
-
-
-
-
59.8
67.9
63.8
-
810524
1.00
0.98
1.82
-
-
-
-
49.2
47.2
48.2
134.7
810525
1.47
1.34
1.74
-
0.336
0.
252
0
.29
108.4
89.7
99.1
126.6
810526
1.31
2.26
2.60
64
2.952
2.
807
2
.88
85.6
262.4
174.0
220.0
810527
1.33
1.18
1.53
-
-
-
-
88.3
69.1
78.7
106.0
810528
1.23
1.64
-
61
1.099
1.
239
1
.17
75.2
135.8
105.5
-
810529
1.35
1.23
-
-
-
-
-
91.1
75.2
83.1
-
810530
1.13
1.08
1.48
-
-
-
-
63.2
57.6
60.4
101.3
810531
1.12
1.07
1.70
-
-
-
-
62.0
56.5
59.3
122.6
810601
1.16
1.08
1.78
99
0.454
0.
363
0
.41
66.7
57.6
62.1
130.6
810602
1.19
1.40
1.64
90
0.169
0.
096
0
.13
70.3
98.1
84.2
116.7
810603
1.16
1.10
1.28
-
-
-
-
66.7
59.8
63.2
83.0
810604
1.12
1.08
-
-
-
-
-
62.0
57.6
59.8
-
810605
1.05
0.98
1.44
-
-
-
-
54.3
47.2
50.7
97.6
810606
1.12
1.35
1.24
-
-
-
-
62.0
91.1
76.6
79.4
810607
1.06
2.30
-
-
-
-
-
55.4
272.1
163.7
-
. 810608
1.16
1.10
1.37
-
-
-
-
66.7
59.8
63.2
91.1
810609
1.17
1.00
-
-
-
-
-
67.9
49.2
58.4
-
810610
0.95
0.96
-
90
0.363
-
-
44.2
45.2
44.7
-
810611
0.90
0.87
-
-
-
-
-
39.6
36.9
38.2
-
810612
0.85
0.84
-
-
-
-
-
35.2
34.3
34.8
-
810613
0.83
0.81
1.36
-
-
-
-
33.5
31.9
32.7
90.2
810614
0.81
0.81
1.50
-
-
-
-
35.2
31.9
33.5
103.2
810615
0.92
0.84
-
-
-
-
-
41.4
34.3
37.9
-
810616
0.85
0.81
1.44
-
-
-
-
35.2
31.9
33.5
97.6
810617
0.96
0.84
1.40
-
-
-
-
45.2
34.3
39.8
93.9
810618
1.00
0.84
-
-
-
-
-
49.2
34.3
41.8
-
810619
0.84
0.80
1.24
-
-
-
-
34.3
31.1
32.7
79.4
810620
0.84
0.81
1.10
-
-
-
-
34.3
31.9
33.1
59.8
810621
0.83
0.80
-
-
-
-
-
33.5
31.1
32.3
-
810622
0.83
0.82
-
-
-
-
-
35.5
32.7
33.1
-
810623
0.75
0.74
1.42
-
-
-
-
27.2
26.5
26.8
95.7
810624
1.04
0.85
1.10
-
-
-
-
53.3
35.2
44.2
59.8
810625
0.85
0.81
1.19
-
-
-
-
35.2
31.9
33.5
70.3
810626
0.89
0.84
1.38
-
-
-
-
38.7
34.3
36.5
92.0
810627
0.90
0.85
1.09
-
-
-
-
39.6
35.2
37.4
58.7
810628
0.86
0.83
1.66
-
-
-
-
36.0
33.5
34.8
118.6
810629
0.98
1.80
1.98
-
-
-
-
47.2
164.5
105.8
151.2
810630
1.15
1.23
1.57
-
-
-
-
65.5
75.2
70.4
109.9


329
Table E-2--Amina River.
D
L
L
L
L
S
S
S
D
D
D
D
A
E
E
E
E
E
E
E
L
L
L
L
T
V
V
V
V
D
D
D
E
E
E
E
E
L
L
L
L
C
C
C
V
V
V
V
A
P
M
F
A
B
A
P
D
M
800101
0.57
0.57
-
-
-
-
-
3.2
3.2
3.2
-
800102
0.56
0.56
-
-
-
-
-
2.9
2.9
2.9
-
800103
0.59
0.58
-
-
-
-
-
3.8
3.5
3.7
-
800104
0.72
0.71
-
-
-
-
-
8.0
7.6
7.8
-
800105
0.60
0.58
-
-
-
-
-
4.1
3.5
3.8
-
800106
0.59
0.59
-
-
-
-
-
3.8
3.8
3.8
-
800107
0.90
0.84
-
-
-
-
-
16.7
13.3
15.0