Front Cover
 Title Page
 Table of Contents
 Editor's words of parting
 Model for the empowerment of a...
 The ecological dynamics of low-external-input...
 An integrated framework for solving...
 Identifying target groups for on-farm...

Group Title: Journal for farming systems research-extension.
Title: Journal of farming systems research-extension
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00071921/00014
 Material Information
Title: Journal of farming systems research-extension
Alternate Title: Journal for farming systems research-extension
Abbreviated Title: J. farming syst. res.-ext.
Physical Description: v. : ill. ; 23 cm.
Language: English
Creator: Association of Farming Systems Research-Extension
Publisher: Association of Farming Systems Research-Extension
Place of Publication: Tucson Ariz. USA
Publication Date: 1990-
Subject: Agricultural systems -- Periodicals -- Developing countries   ( lcsh )
Agricultural extension work -- Research -- Periodicals   ( lcsh )
Sustainable agriculture -- Periodicals -- Developing countries   ( lcsh )
Genre: periodical   ( marcgt )
Dates or Sequential Designation: Vol. 1, no. 1-
General Note: Title varies slightly.
General Note: Title from cover.
General Note: Latest issue consulted: Vol. 1, no. 2, published in 1990.
Funding: Electronic resources created as part of a prototype UF Institutional Repository and Faculty Papers project by the University of Florida.
 Record Information
Bibliographic ID: UF00071921
Volume ID: VID00014
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 22044949
lccn - sn 90001812
issn - 1051-6786

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Page i
        Page ii
    Table of Contents
        Page iii
        Page iv
    Editor's words of parting
        Page v
        Page vi
    Model for the empowerment of a local community through the analysis of exchange terms between policy criteria
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
    The ecological dynamics of low-external-input agriculture: A case study of hill farming in a developing country setting
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
    An integrated framework for solving problems in sustainable agriculture
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
    Identifying target groups for on-farm research: Characterizing farmers for soil fertility maintenance research in semi-arid areas of eastern Kenya
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
Full Text

Volume 7, Number 2
t 1997/2002


for Farming Systems


Table of Contents

v Editor's Words of Parting
I Model forte Empowerment ofa I oca I
Community Through the Analysis of 1-whange
Ternis Between Policy Criteria
Rct-nordo Rivera and Rulwn Dario I'4rado
23 The Ecological Dynamics of Lot%-ExIci nal- Input
Agriculture: A Case Study of Hill Fat ming in a
Developing Country Setting
Yanuma Ghtdc, Gintsh Shwakohmid Bi4mit
43 An Integrated Frainework for Solving Vtoblerns
in SuStainable AgriCUltUri,
/ohn S11111hers, I /It'// lVall ond chirelf, (' ;it '1111ml
59 Identifying Target Group for ()it [,it in
Rt-search Characterizing Farmers to-i Still
Fertility Maintenance Research hi Senil \iid
Areas of Eastern KenNa
II.Adc In't,1111111, /ohll 11. (biliti /fill/

for Farming Systems

Volume 7, Number 2, 1997/2002

Published by
the International Farming Systems Association

Journal for Farming Systems Research-Extension
ISSN: 1051-6786

John S. Caldwell
International Research Coordinator-Farming Systems
Japan International Research Center for Agricultural Sciences (JIRCAS)
1-2 Ohwashi, Tsukuba, Ibaragi 303-8686, Japan

The Journal for Farming Systems Research-Extension has been a publication of the
Association for Farming Systems Research-Extension (AFSRE), an international
society organized to promote the development and dissemination of methods and
results of participatory systems research and extension. The International Farming
Systems Association (IFSA) is the new organizational form of AFSRE. The objectives
of farming systems research and extension are the development and adoption,
through participation by farm households, rural communities and policy makers,
of technologies, practices and policies that promote the well being of agriculturally-
based people and places to meet the socio-economic and nutritional needs of farm
families and to foster the efficient and sustainable use of natural resources.

This will be the last print edition of the journal. The International Farming Systems
Association (IFSA) through its biannual meetings will foster the sharing of farming
systems research and extension by making papers available through the Web (http:
//conference.ifas.ufl.edu/), with hard copies on request.

IFSA Co-Presidents are:

Cornelia Butler Flora Constance L. Neely
Charles E Curtiss Distinguished Deputy Director
Professor of Agriculture SANREM CRSP
Director, North Central Regional Center 1422 Experiment Station Road
for Rural Development Watkinsville, GA 30677-1422
107 Curtiss Hall (706) 769-3792; (706) 769-1471 fax
Ames, IA 50011-1050 cneely@arches.uga.edu
(515) 294-8321; (515) 294-3180 fax

The IFSA Secretary/Treasurer is:

Dr. Virginia Cardenas
University of the Philippines Los Bailos
College, Laguna, PHILIPPINES
Fax (+63)(49) 536-3256 (+63)(49) 536-2489
VRC@mudspring.uplb.edu.ph vcard@laguna.net

Journalfor Farming Systems Research-Extension

Journal for Farming Systems Research-Extension
Volume 7, Number 2, 1997/2002

Table of Contents

v Editor's Words of Parting

1 Model for the Empowerment of a Local Community Through the
Analysis of Exchange Terms Between Policy Criteria
Bernardo Rivera and Ruben Dario Estrada

23 The Ecological Dynamics of Low-External-Input Agriculture: A Case
Study of Hill Farming in a Developing Country Setting
Yamuna Ghale, Ganesh Shivakoti and Bishnu Upreti

43 An Integrated Framework for Solving Problems in Sustainable
John Smithers, Ellen Wall and Clarence Swanton

59 Identifying Target Groups for On-Farm Research: Characterizing
Farmers for Soil Fertility Maintenance Research in Semi-Arid Areas of
Eastern Kenya
H. Ade Freeman, John M. Omiti and Patrick A. Audi

Vol. 7, No. 2,1997/2002

Editor's Words of Parting

Dear Readers,

This is the last regular issue of the Journal for Farming Systems Research/Extension
that the International Farming Systems Association (IFSA) will be publishing.
As you know, the Journal has had difficulties during the last number of years.
When the Journal was initiated over 10 years ago by the Association for Farming
Systems Research/Extension (AFSRE, the name used by the Association now
called IFSA until 1998), several donors provided a substantial portion of the cost.
In addition, the University of Arizona provided considerable support for much of
the editorial and business functions of maintaining the Journal. The first editor, Dr.
Tim Frankenberger, contributed enormous amounts of professional and personal
time to establish the Journal and its standards in the first part of the 1990s. In the
mid-1990s, Dr. George Axinn graciously donated personal time after retirement to
assume the editorship, and his.wife Dr. Nancy Axinn provided valuable support
on an entirely voluntary basis for several years. But such a voluntary arrangement
could not continue indefinitely. A number of people associated with the Journal
recognized that professional editorial support like what had been provided initially
by University of Arizona was needed, but no other institution appeared to offer this
kind of support. The Journal needed to pay its own way, either through income or
through grants.

From early on, the Association recognized that supporting the Journal through
income involved a structural and indeed philosophical dilemma. Even in 1992,
when the Association (as AFSRE) had over 400 subscribing members, the Journal
was not financially self-sustaining. The reason was due to the very purpose of
the Journal. The purpose of founding the Journal was to provide professionals
in farming systems and participatory research with an international forum for
publication and exchange of results and ideas that all could participate in. Since the
majority of professionals in farming systems and participatory research were, and
still are today, from developing countries, the subscription rate for the Journal was
set at an amount that professionals in developing countries could afford, $20 per
year for two to four issues per year. Professionals from developed countries initially
paid $65 per year, a figure that was subsequently raised several times, reaching $100
in the latter half of the 1990s. Printing was transferred from the United States to
the Philippines, to reduce costs. But even then, subscriptions by developed country
professionals only barely covered the direct costs of their issues. To make up the
difference through subsidization by developed country professionals would have
required subscription costs two to three times the cost of other professional journals,
and would have been self-defeating, causing subscription loss and consequent total
income decrease.

In the 1990s, there were major shifts in development philosophy that made it
difficult and finally impossible for the Journal to sustain the grant support it had
benefited from initially. Globally, economic neo-liberalism became the dominant
political and economic philosophy. There was a shift in nearly every country in the
balance between the public and private sector, leading to a reduction in the role of

Vol. 7, No. 2, 1997/2002

the state and increased devolution of development to the private sector. The goals of
development shifted from public support for assurance of food security, basic needs,
and rural well-being, to private enterprise-driven economic growth. Development
assistance for institution-building was given less priority, while loans and training
for agribusiness became more favored. In this environment, support for a journal
serving primarily public-sector professionals focused on helping develop public
goods for less economically-endowed farmers was not a high priority for donors.
The message to the Journal from potential donors was to become self-sustaining.

With only meager self-generated income, the end of donor support, no institutional
backing for professional editorial support, and reliance entirely on volunteer time, the
Journal became an untenable enterprise beyond the financial and human resources of
the Association. It is for this reason that IFSA finally came to a conclusion that it had
no choice but to terminate the Journal as a regular publication. Several alternatives
were explored, but each had problems. Electronic publishing would save costs, but
it would effectively remove the Journal from the hands of most developing country
professionals, who often do not have personal e-mail and internet access in their
own office, but have to rely on a central institutional site. The nature of the Journal
and its users made it inherently less attractive for commercial publishing and
income-generation through advertising, and there were also ethical questions about
which types of advertising would be congruent with the purposes and philosophy
of participatory farming systems. In an era of increased accountability, neither
institutions nor individual professionals could easily justify the kinds of financial
and time commitments required for the Journal, relative to its benefits that are long-
term and not easily quantifiable.

IFSA retains ownership of the name, Journal for Farming Systems Research-Extension.
Special issues based on its International Symposia held every two years will use
this name. On this limited basis, supported through revenue from the Symposia,
we hope the Journal may still appear on an occasional basis, and provide continuity
with its past.

On a personal note, it has not been easy to recognize and accept this conclusion.
All of us involved with the Journal during these last few years hoped that somehow
we could find a way to make this work, but the task was too great, the resources
too few, and time too little. To all who contributed time and moral support in spite
of these frustrations, I extend my deep appreciation. To all who felt disappointed
by our shortcomings and inability to maintain the Journal, I express my sincere
apologies, and ask for your understanding and tolerance. And finally, to all who
have supported the efforts by IFSA/AFSRE and various associations and networks
around the world, who have worked and continue to work to link formal science
and farmer knowledge, and to enable farmers and rural peoples to make that
linkage work for themselves, I extend my hopes for your perseverance and success.

John S. Caldwell

Journal for Farming Systems Research-Extension

Model for the Empowerment of a Local
Community Through the Analysis of
Exchange Terms Between Policy Criteria

Bernardo Rivera' and Ruben Dario Estrada2


The San Antonio sub watershed of the La Miel Watershed in Colombia has high
potential for soil degradation due to high precipitation, pronounced slopes,
construction of a hydroelectric complex, and lack of agricultural opportunities
for producers to improve their income. It was selected for research in order to
develop a pilot model to assess critical points and intervention strategies for
a conservation and development program. Basic information on land use and
socioeconomic conditions was determined by surveying 63 producers (25%
of the total in the subwater). The CROPWAT model was used to simulate
water use, the EPIC model to estimate soil losses, and the REPRESAS model
to estimate opportunity cost of sediments for the dam. Information generated
by the different models was integrated into a linear programming model
that optimizes producers' net income. It additionally incorporates soil loss,
water supply and job generation, as sustainability and equity criteria. The
main use of the model was to generate useful information for the producers
in the organization, as well as to negotiate with environmental politicians
and the enterprise constructing the dam. Four kinds of farms were identified
according to social, environmental and land use characteristics. Coffee was
the core of all the systems; according to the simulation model, it is the crop
that produces the highest soil loss (54 T/ha; the weighted mean in the whole
watershed is 16 T/ha). Results from the linear programming indicate that land
use is adjusted to the region's conditions. Assigning a value to the increment
in the water volume does not generate changes in the current land use.
Sedimentation levels are relatively low, and to reach substantial changes in
land use, a greater financial incentive than the one that the dam is willing to
provide is necessary. Reduction of up to 40% in coffee prices does not have an
important effect on the land use in the watershed, and poverty increases more
rapidly than erosion. The chances of capturing external resources are few. As a
result, the community formed a broad-based coalition to negotiate support for
education as a development strategy. It is also negotiating to participate in the
management of a new bioreserve in the Florencia Tropical Rain Forest.

Key words: models, hillsides, policy criteria, empowerment, hydroelectric

1 Bernardo Rivera. Eduquemos Foundation, Universidad de Caldas. Calle 65 No. 26-10.
Fax (68) 812975. Manizales, Colombia, brivera@cumanday.ucaldas.edu.co
2 Ruben Dario Estrada. CIAT-CONDESAN-COLCIENCIAS ECOFONDO Corporation,
A.A. 6713. Fax: (2) 4450273 Cali, Colombia, r.estrada@cgnet.com

Vol. 7, No. 2, 1997/2002


1. Introduction

The Florencia Tropical Rain Forest (Caldas), located on the east branch of the central
chain of the Andes mountains between 1,700 and 2,100 m, constitutes one of the last
forest strongholds in hillsides zones of the Andean region. In 1963, it had a forest
surface of 11,400 ha; in 1998 there were just 6,600 ha. The region is particularly rich
in diverse fauna and flora and is a key source of water for the region.

The source of the San Antonio River is in the Tropical Rain Forest of Florencia, an
important tributary of La Miel River watershed, where the energy complex Miel
I is currently under construction. Seventy eight percent of the energy used in the
country is hydraulic, and an increase in consumption of 1.1 to 1.5 megawatts per
person per year is expected for the next seven years.

The Eduquemos Foundation, with the technical support of Production Systems
Department- Universidad de Caldas and the Consorcio para el Desarrollo Sostenible
de la Ecorregi6n Andina (CONDESAN), with financial support of ECOFONDO
corporation and COLCIENCIAS, is implementing a proposal for a water use in
the communities of the Florencia Tropical Rain Forest. We began by generating
technical information to support the producers' and politicians' decision-making
processes, and then utilized that information for the training and organizing of the
community to manage the Tropical Rain Forest and the neighboring watersheds. The
project took place at the time that the Columbian government acquired land in the
Florencia Tropical Rain Forest as an institutional strategy to assure the conservation
of the Tropical Rain Forest and help reduce the rural poverty.

Conflicts between competitiveness and agroecological sustainability and diverse
ways in which stakeholders (agricultural producers and water and energy
consumers) are affected makes setting priorities on land use in a watershed difficult.
The research developed a pilot model to quantify the social benefits created by
alternative land management practices, and to see what benefits would actually
go to producers. We evaluated critical points of intervention for a conservation
and development program in the San Antonio river watershed. We undertook this
type of study because (a) the zone is considered to have high soil erosion potential,
caused by high precipitation and steep slopes, (b) one of the most efficient dams
in the world is under construction, where a cubic meter of sediment could have
a costly impact on farmer opportunities, and (c) agricultural producers have few
opportunities to improve their income through the agricultural productivity due
to adverse production factors such as high proportion of land in shade, lack of
improved varieties adapted to local conditions and inefficient use of fertilizers
caused by high rainfall. The degree to which peasant community organization
was strengthened and its decision-making process improved was used to measure
the impact of the research. We summarize the methodological process, the results
of potential scenario simulation of some potential scenarios, and advances in the
empowerment of the community.

Journal for Farming Systems Research-Extension


2. Methodology

2.1 Characterization of production systems
Secondary information was collected on location, weather, topography, hydrography,
and land use. Using this information, we designed a survey and applied to a stratified
sample of 63 producers, corresponding to 25% of properties in the watershed. The
farms were stratified based on ecosystem, altitude above sea level and location with
reference to the river. The survey identified current land use, total area, distribution
by crop, use and distribution of labor, use of external inputs, productivity, value of
crops and animals produced, and agricultural techniques, emphasizing topography
and crop and pasture distribution within the farm. The first reconnaissance of the area
identified landslides (mass movements) as a key factor impacting farmers, so a question
was included about the size of landslides in the farms. An agricultural technician, who
is also a producer, conducted all of the surveys to reduce errors in slope estimates.

Starting from a multivariate and principal component analysis, a conglomerate
analysis was done to typify farm groups with the highest internal homogeneity and
the highest external heterogeneity. This was accomplished following the methodology
proposed by the International Network on Research Methodologies for Production
Systems (Red Internacional de Metodologia de Investigaci6n de Sistemas de
Producci6n) (Escobar y Berdegu6, 1990; Rodriguez y Carvajal, 1996).

2.2 Determination of water use
To verify the secondary information regarding precipitation and its distribution, five
pluviometers were positioned in different places of the watershed. Water use and the
state of conservation of water sources was determined by a survey, which was applied
to 70% of the families. Evapotranspiration (measured using the Penman-Monteith
method) and water consumption of each crop in the different zones of the watershed
was estimated using the FAO CROPWAT model (Smith, 1993). Information on
precipitation, temperature, relative humidity, wind speed and radiation was obtained
from secondary sources. Time series information on precipitation is presented in Table
1. Water flow to the outlet point was estimated during a year through fortnightly
evaluations of speed (using the float method) and daily measures from the river area
(starting form the height level) to verify the amount of water originating in the San
Antonio River sub watershed flowing to La Miel watershed.

2.3 Determination of the effect of cultural practices on soil loss
The EPIC model (Environmental Policy Integrated Climate) was used to determine
soil loss in different zone and crop conditions. EPIC is based on a calibrated model
of USLE (Universal Soil Loss Equation) that determines soil loss as a function of
erosivity, erodibility, slope, cover and conservation practices (Wischmeier, W.H. and
Smith, D.D. 1978; Mitchell et al. nd.).

Five parcels corn, cassava, beans and fallow 12 x 2 m parcels were installed to
measure run off to verify the EPIC model's results (Arroyave et al., 1998). Farmers
were trained to manage the plots and collect information. Their participation helped
create consciousness of the effect of crop practices on the soil movement.

Vol. 7, No. 2, 1997/2002


For a year, we daily measured turbidity of the water flow in the San Antonio River
outlet, in order to verify total sediment. Total amount of solids in the water was
estimated from turbidity.

Table 1. Precipitation in San Antonio River watershed (mm)
Year 1 2 3 4 5 6 7 8 9 10 11 12 Total
1975 424 757 609 556 608 316 724 404 304 720 1119 819 7360
1976 659 459 556 760 578 197 5 140 547 792 454 406 5554
1977 292 428 594 185 475 467 272 350 404 728 592 515 5302
1978 314 349 645 1008 856 332 299 166 793 773 605 1156 7296
1979 163 622 711 797 358 417 338 671 854 864 600 630 7025
1980 456 498 375 353 633 265 166 200 702 738 1154 972 6510
1981 334 335 623 577 763 558 163 560 529 914 912 893 7161
1982 1039 873 241 370 712 136 41 79 323 431 670 713 5625
1983 386 322 480 835 528 270 250 200 338 742 483 447 5280
1984 567 834 548 511 647 384 586 514 904 1241 996 928 8660
1985 126 168 433 656 319 240 153 371 525 683 670 414 4759
1986 708 332 505 809 675 228 12 279 342 1018 655 805 6366
Mean 456 498 526 618 596 317 251 328 547 804 743 725 6408

2.4 Determination of the opportunity cost of sediments for the dam
The Represas model, developed by CIAT-CONDESAN (Estrada, R. 1998, Personal
Communication), was used. That model simulates the opportunity cost of
sediments, expressed as the present net value for ton of sediments based on the
benefits generated by the hydroelectric project, using the technical parameters that
the project's construction company published (HIDROMIEL, 1997).

2.5 Optimization model to analyze exchanges between different policy criteria
The information generated by the different models was integrated into a simple
linear programmatic model that optimizes net income (sales minus variable costs
in cash) of the producers in the watershed. We used the most commonly grown
crops, though they generally earn the lowest prices, in estimating the income from
agricultural production. Panela (raw sugar) production is countercyclical to coffee
production, due to labor availability. Coffee prices are mostly fixed by international
mechanisms rather than by local supply and demand. Livestock prices have local
stability. It was hardest to assign monetary values to forestry and fallow products
(firewood, timber, posts, and charcoal), due to lack of markets. Consequently, we

Journal for Farming Systems Research-Extension


used interviews to determine the opportunity costs of cooking with electricity or
petroleum-based fuels where electricity was not available.

Aside from adding the net income (competitiveness criteria) to the objective function,
substantial changes to land use that reduced soil loss and contributed more water
to the reservoir at two different times (maximum and minimum precipitation) were
added to the model as sustainability criteria. Additionally, the impact of changes in
the land use on employment was analyzed as an equity criterion.

The mental model for the construction of the mathematical model was related to
the competition for water between agricultural production and the reservoir. It was
also related to competition between current land use, its impact on the production
of sediment, and achieving a long and useful life for the dam. The starting point
to determine "shadow" prices was the agricultural production parameters in each
zone and the products' market value. The optimization exercise identified the point
where, given different values to water production (in two different seasons) and
sediment reduction, it would be feasible to initiate a substantial change in present
land use.

The model used the restrictions on land area, labor use and capital determined in
the characterization. Farm size and the area for basic foods (corn, bean and cassava)
were the only restriction over the model's decisions on land use. These foods are for
household consumption and do not contribute to the income. Yet, if they were not
grown, the household would incur substantial expense to purchase them. Limits
on labor were determined by its current availability, but the model had the option
of using family labor in farm activities or selling it for off-farm activities. Likewise,
restrictions on capital were determined by current use, but the model permitted
investment decisions by the producer, in which case it negatively affected the
income over the cost capital (10% in real terms).

The sensitivity analysis of the model was completed by exploring different scenarios
through changes in the values assigned to greater availability of water as a result of
the dam and for sediment reduction and the reduction in the coffee price.

2.6 Community organization and training
Eight community participation committees (Comites Veredales de Participacion:
CVP) were organized, one for each watershed community, adapting the methodology
proposed by Ashby et al. (1996) to form local agricultural research committees
(Comites de Investigaci6n Agropecuaria Local: CIAL). The CVP facilitated the
diagnosis to increase participation and prioritize community training topics.

Two working groups were formed, with participation from market, HIDROMIEL,
the enterprise constructing the hydroelectric complex and the state, the Samana
Major's office, and CORPOCALDAS, the regional environmental authority to work
with civil society (the CVPs) to identify joint strategies to resolve the controversies
around the conservation proposals for the Florencia Tropical Rain Forest.

Vol. 7, No. 2, 1997/2002


3. Results

3.1 Typology of the production systems
The main components were determined through analysis of 56 of the 63 surveys.
The 16 variables with the highest discriminatory power between farms included
farm area, percent of landused in agricultural activities, percent in coffee, percent
in sugar cane, percent in pastures, altitude, farm weighted slope, producer age,
percent of male children on the farm, days worked per hectar, percent of days
worked purchased, days of labor sold, pig and poultry population, erosion index,
and technology index. The erosion index is a qualitative score by the producers on
the presence of erosion indicators. The technology index was calculated by assessing
the costs of fertilizers and pesticides used in agricultural activity and of mineral salts
and vaccines in livestock production. Four main components explained 57% of the
total variance. (Table 2).

Table 2. Proper values of the correlation matrix
Component Proper Value Difference Contribution Cumulate
1 3.23 0.72 0.20 0.20
2 2.51 0.55 0.16 0.36
3 1.96 0.54 0.12 0.48
4 1.42 0.15 0.09 0.57

The contribution made by each variable to each component is presented in Table
3. Structural variables such as, farm area, number of animal units, and percent of
the farm in coffee determined component 1. Percent of land in pasture and altitude
made up component 2. Off farm labor (number of days of labor sold) determined
component 3. Social variables (percent of male children on the farm) and slope of the
farm made up component 3.

Table 3. Contribution of vectors to the main components
Variable-Component 1 2 3 4
Farm Area 0.47 -0.03 0.06 -0.21
Altitude 0.19 -0.41 0.05 -0.12
% Children in the farm 0.06 0.00 -0.42 0.52
Animal Units 0.32 0.38 0.01 -0.01
Labor sold 0.12 -0.05 -0.54 -0.19
Weighted slope -0.05 -0.28 0.19 0.51
% Coffee -0.45 -0.07 0.16 -0.01
% Pasture 0.17 0.46 0.09 0.26

Journal for Farming Systems Research-Extension


Conglomerate analysis, using the information from the four main components,
resulted in six farm types (R2= 0.65). However, it was decided to regroup them
in four farm types clearly identified in the watershed: high zone (San Lucas and
San Antonio), middle high zone (La Vifia, La Cabafia and Encimadas), middle low
zone (Montesory) and low zone (Montecristo and Primavera), based on previous
knowledge and the community space that the combined types shared. Information
on the type of farm in each one of these zones and their corresponding technical
indices is shown in Table 4.

Table 4. Use of resources and mean income of type farms in 4 zones of San
Antonio River watershed
High Middle Middle Low
high low
Area (ha) 42.0 11.1 5.7 29.8
Coffee % 10.4 39.4 55.0 5.5
Pasture % 10.9 14.0 9.1 59.4
Sugar Cane % 0.3 4.5 10.3 6.3
Food Crops % 2.6 7.3 6.9 2.6
Scrub and Fallow % 75.8 34.8 18.7 26.2
Available labor 309 234 285 332
Coffee (labor) 262 262 304 170
Pasture (labor) 55 19 9 283
Sugar Cane (labor) 3 25 51 154
Food Crops (labor) 51 38 19 38
Scrub and fallow (labor) 127 15 4 31
Coffee ($x1000) 414 411 310 153
Pasture ($x1000) 0 6 0 71
Sugar Cane ($x1000) 0 0 3 0
Household crop ($x1000) 9 0 0 0
Scrub and fallow ($x1000) 0 0 0 0
Coffee ($x1000) 4.375 4.769 3.811 1.687
Milk and meat ($x1000) 434 148 57 2.944
Dry molasses ($x1000) 25 133 407 1.277
0 0 0 0
Food crops ($x1000) 2579 278 80 539
Scrub and fallow ($x1000) ____ __
1 USD= $1000

The high zone is located in the tropical rain forest and in its perimeter, at about
1,352 m. It has the largest average size farm (42 ha), and largest percentage of scrub
and fallow (76%); the owners are the oldest (55 years old) and they sell around 20%
of their available labor. Coffee is the most important income-generating crop, but
forest products (charcoal, firewood and timber) are also important income sources.

Vol. 7, No. 2, 1997/2002


In the watershed context; they are the farms that generate the greatest net incomes
(2.7 minimal wage).

The high middle zone is located on average at about 1,172 m, with farms with
the steep slope (205%). The farms average 11.1 ha, of which 4.4 ha are planted in
coffee. Farms obtain their cash income almost exclusively from coffee. The available
workforce is absorbed by the farm.

The lower middle zone (1,047m) has the highest density of families (102 out
of the 253 in the area). It has the smallest farms in the watershed (5.7 ha) and
the most intensive use of the workforce (71 day's labor/ha). Fifty-five percent
of the farm is planted to coffee, and this is in this zone of the highest coffee
productivity. These are the farms with the least investment, the lowest income
(1.6 minimum wages) and highest number of days worked off their own farm.

The low zone (857 m) is fundamentally different from the others because of livestock
production, the high investment this activity requires, and the diversification of
their income through sugarcane for panela production. These are larger farms (29.8
ha) and their owners are the youngest within the watershed (39 years old). These
farms have the highest number of cattle (21.2 animal units) and the highest carrying
capacity (1.2 animal units/ha).

Very intense land was expected in the region because of large family size and scarce
job alternatives outside of the agriculture and livestock sector. However, 49% of all
farms is forest and fallow, while in the high zone, almost 76% of the land in farms is
forest and fallow (thus explaining the larger farm size).

Farms have very steep slopes and under the traditional parameters of land quality
are not suitable for agriculture and livestock production. Plots slopes ranged from
75% to above 300%. The average slope in coffee plots is 164%, while for basic
household crops it is 158%.

Coffee constitutes the core of agriculture and livestock production systems, with
45% of the cultivated area. It uses 56% of the work days and generates 66% of farm
income. In the low zone, livestock generates more income than coffee (46% vs. 26%).
In all zones, the farmers feel that only the coffee crop merits the use of inputs and
improved technology. Only labor is invested in food crops and sugar cane. Purchase
of external inputs (vaccine, salts, drugs, etc.) is observed in the livestock activity, but
the cash investment in coffee is the highest.

Labor use is relatively intense: a total of 110,249 working days, which averages
28 days of household labor/per hectar or 436 days of household labor per farm.
Additionally, an average of 218 days labor/farm is hired, concentrated in the major
and minor coffee harvest. Contrary to what happens in regions where the annual
or seasonal crops are more important, farms in the San Antonio watershed use 60%
of crop labor for harvest. Likewise, they use 33% in weed control and only 7% in
preparation of the soil (slash and burn) and in sowing (without including labor used
for the establishment of pasture or for coffee planting, which are activities that are

Journal for Farming Systems Research-Extension


not implemented yearly). Due to steep slopes and constant precipitation, farms do
not use mechanization for planting. The traditional planting system is a minimal
one known as "a chuzo," which later controls weeds and covers the terrain.

Producers show their household priorities through the amounts they invest in high
school education for their children (294 USD/year), health (228 USD/year) and housing
improvement (203 USD/yea:). These are the fundamental components of their quality
of life. Producers report that they request credit for improved housing and utilities.

3.2. Water sources and uses
We rejected the initial hypothesis that the most rain was generated in the rain forest
and the high zone after rain gauges were installed in different sites of the watershed.
We found some rain gradient from the high to lower altitudes. The differences are
small compared to the total amount of rain. The precipitation information on
five different points of the watershed is presented in Table 5. The pattern of the
rain is similar throughout the watershed. Nevertheless, we identified a period of
maximum precipitation between September and April (689 mm/month) and minor
precipitation (457 mm/month) from May to August.

Table 5. Information on precipitation in different points in the watershed
(June 1996 to May 1997)
Middle Low
Middle Middle Middle Middle Low
Month High Zone HighZone 1 High Zone 2 High Zone 3 and Low

June 619 599 687 711 683
July 390 351 254 334 167
August 605 454 445 422 363
September 545 577 575 619 617
October 598 709 745 767 678
November 882 948 940 957 1040
December 760 897 713 726 625
January 330 540 331 475 320
February 868 766 868 972 920
March 570 484 472 509 445
April 821 845 858 834 655
May 581 352 477 466 405
Total 7569 7522 7365 7792 6918

The survey on water use revealed that almost all of the families' have their own
water sources, which are in relatively good conditions, and there are no important
aqueducts that empty into the San Antonio River. Therefore, the production of water
in the watershed may potentially impact the generation of energy in the hydroelectric
project Miel I. Aside from the use for domestic consumption, water is only used for

Vol. 7, No. 2, 1997/2002


animals and rainfed crop production. Table 6 shows water consumption for different
crops in two different seasons, as estimated by the CROPMAT Model. With the
available crop parameters, the model can predict similar water consumption values
among crops. In the case of the fallow, it was not possible to identify the coefficients
of corresponding crops. These coefficients were estimated in 3,000 and 1,000 m3/ha,
in the respective periods of maximum and minimum rainfall.

Table 6. Water consumption use (m3/ha) for diverse crops in two different
periods in the San Antonio River Watershed
Crops Maximum rainfall Minimum rainfall
period period
Coffee 7,260 2,470
Sugar cane 7,500 2,800
Pasture 7,000 3,000
Basic household crops 7,500 2,800

From a total production of 292 million cubic meters of water in the watershed, current
crops use 28.1 million for their growth and production. Therefore, 263.9 million cubic
meters per year would be contributed to the river, which is equivalent to 8.4 m3/
second, without considering human and livestock consumption. This information
was confirmed through daily estimates of flows at the end of the watershed. The
estimated value, taking into account height and speed measurements, was 7.2 m3
per second.

3.3. Soil loss assessment
Coffee is the crop that generates the most soil loss because producers in the area
do not use a shaded production system, have low crop density, and manually
weed every three months, which tramples the soil. Basic household crops (corn,
beans, yucca and banana) are the only ones that show low coverage during some
point of the cropping cycle, because the traditional slash and burn system is used
for its cultivation. However, these practices occur during the period of minimum
rainfall. Additionally, it is common that after one harvest the plot is left with fallow,
enormously reducing potential erosion. Table 7 shows that coffee is the crop that
contributes the most sediment to the river bed (49,434 T/year); sugar contributes
the least (984 T/year).

The 1997 estimates in the San Antonio River mouth indicate that total sediment loss
reached 48 T/ha. The monthly contribution of the sediment to the river is shown in
the Figure 1. It ranges from 1.5 to 5.8 T/ha in January and in March. Basic household
products in the zone are planted in March and August, when there is minimal soil
cover. There is higher rainfall in March than in August.

In addition to the estimated loss of soil due to the agriculture and livestock activity
(16 T/ha), landslides on 53% of the farms also contribute solids to the river bed.

Journalfor Farming Systems Research-Extension


Table 7. Estimated soil losses (TM/ha) in different zones by crop
Crops High Middle Middle Low Total Total
high low ha (No.) MT
Coffee 67 68 29 57 920 49,434
Pasture 8 3 2 5 776 3,888
Sugar cane 7 4 8 7 151 984
Basic crops 11 7 8 12 174 1,602
Fallow 4 1 1 1 1,951 6,346
Total 3,972 62,254

According to the survey, producers reported 3,248 linear meters of landslides.
Assuming an average width of 10m, depth of lm and density of 2 T/m3, the
contribution of sediments due to landslides is approximately 65,680 T, translating
to close to an average of 17 T/ha for all the watershed. The difference between
estimated losses of soil for agriculture and livestock activity and landslides (33 T/
ha) and the contribution of solids to the river (48 T/ha) is explained by the effect of
cattle trampling, road use by the communities, and the impact of the river over its
own banks. We did not estimate these factors in our research.

d 1.5
1.0 -

0.0 I I I I
1 2 3 4 5 6 7 8 9 10 11 12

Figure 1. Distribution of the total contribution per month of sediments to the
river (T/ha), based on the measurement of total solids in the river at the end of
the watershed

Vol. 7, No. 2, 1997/2002

-- 4.9


3.4 Opportunity cost for sedimentation of the dam
According to the results of the "Represas" Model, the savings that consumers
receive from sediment level reduction becomes important 83 years after the dam
construction. These benefits translate to the current net value $2.56 USD/T of
sedimention, considering a rate of sediments of 40 T/ha with increments of 10 T/ha
during years 10 and 20. The model was very sensitive to charges in the discount rate;
we used a rate of 5%, which is the normal rate for this type of investment.

3.5 Economic analysis of the agriculture and livestock activities
Table 8 presents a summary of the economic analysis of agriculture and livestock
activities for each of the four farm types. Cost of inputs and hired labor are
considered expenses (cash outflows). Sale of agriculture and livestock products and
labor are considered income. Sixty-five percent of the livestock is raised by equally
sharing the price received for the livestock between the owner of the land and the
owner of the animal; the amount received for the sale of meat was reduced to 65% of
the sale price and 50% of the sale price.

On average, farms in the watershed generate 2.04 minimum wages ($172,000
Columbian pesos per monthly minimum wage). We found a positive correlation
between the size of farms and number of minimum wages generated, except for the
low zone and marginal coffee land. In spite of their relatively larger farm size, the
division of livestock income reduces those producers' net incomes.

Table 8. Economic analysis (pesos/farm) of the agriculture and livestock activi-
ties implemented in distinct zones of the San Antonio River watershed
Zone Outflows Income Net incomes Income per
(Expenses, month *
High 2,433,266 8,073,060 5,639.794 2.73
Middle high 1,560,668 5,577,391 4,016,723 1.95
Middle lower 1,498,397 4,835,498 3,337,102 1.62
Low 3,828,282 7,675,501 3,847,218 1.86
Total Watershed 2,330,153 6,540,362 4,210,209 2.04
* Net incomes/month expressed in Minimum Legal Wage.

3.6 Analysis of the exchange among policy criteria
The restrictions of the linear programming model can be seen in Table 9. Each zone
has a maximum utilizable area of 3,972 ha. Working days are disaggregated into
household and hired labor (69,843 and 41,275 working days respectively), giving
the model the option of using the working days in farm activities or selling them for
off-farm activities.

In order to validate the model, the options were restricted to the current activities of
the farm to verify the soundness of the parameters. The results indicate an objective

Journal for Farming Systems Research-Extension


function (net income) of $3,655,000, $13,000 for family labor used and $8,000 for
hired labor. All available labor was used. The production of water for the river was
263.9 million cubic meters per year, after discounting use of rainfall by agriculture
and livestock activities. The estimated sediment load was of 60,121 MT per year.

Table 9. Restrictions used by the linear programming model
Zone Capital Area Labor
High 20,008 1,754 12,904
Middle high 38,133 967 20,385
Middle low 31,896 582 29,100
Low 11,409 669 7,453
TOTAL 101,446 3,972 69,842

The results of the linear programming indicate that current land use is appropriate
for the agroecological, technological and economic conditions of the region. The
increase in the objective function, after the optimization exercise, was only 8% (Table
10). The main land use changes proposed by the optimization model refer to the
reduction of pasture areas in the high, middle high and middle low zones; and an
increase of pasture areas in the low zone, replacing scrub and fallow. Coffee areas
increase in the middle high and middle low zones. Peasant logic of having pastures
to maintain animals is not incorporated into the model, which explains the difference
between the actual and optimal land uses. Animals allow farmers to have capital
assets, reduce risk, and eat meat. The optimal land use increases water production
slightly to 264.8 cubic meters, though sediment also increases to 63.950 MT.

The model's sensitivity to changes in the value of water, between $5 y $40/m3, was
minimal. Estimated water consumption for the CROPWAT model was very similar
among the different soil covers, explaining why the model shows little water use
responsiveness. An increase in water availability in the watershed did not lead to
land use changes, nor to increased water or sediment production.

Land use in the model changed more in response to a situation of no sedimentation
than to changes in water availability. Table 11 shows that the value of "no
sedimentation" increased the levels of sedimentation decrease. But an important
impact the amount of water in the reservoir is not achieved. The optimum response
point is found around $11,000/MT of sediments not produced. This is a value
that is too high for the dam at present, in that the dam is just under construction.
Consequently, paying $2,560/MT for the sediments not produced, a value that
the dam should be willing to pay for the benefits it represents will not result in
important changes will occur in land use in the watershed.

Vol. 7, No. 2, 1997/2002


Table 10. Comparison of current land use (ha) in the distinct zones of the wa-
tershed with the optimization model
Activity Zone Current Use Optimum Model
Coffee (ha) High 182 182
Middle high 381 465
Middle low 320 395
Low 37 37
Pasture (ha) High 191 0
Middle high 135 0
Middle low 53 0
Low 397 605
Sugar cane (ha) High 0 0
Middle high 44 0
Middle low 60 0
Low 42 10
Fallow (ha) High 1,330 1,526
Middle high 337 431
Middle low 109 146
Low 175 0
Objective function ($xl000xfarm) 3,653 3,992
Labor (No. days of labor) 111,051 111,118
Water provided (million m3) 263.9 264.8
Sediment (MT) 60,121 63,950

Table 11. Sensitivity analysis of the linear programming model response to
changes in the price of sediment over total production (MT) and contribution
of water to the watershed (millions of m3)
Value of non produced Contribution of Contribution of water
sediments (pesos/MT) sediments (MT) (millions of m3)
0 63,950 264.8
4,000 63,866 264.8
8,000 63,322 264.8
12,000 42,458 266.0
16,000 33,463 261.8
20,000 33,463 261.8
24,000 27,929 264.2

Figure 2 shows how the value of sediments not produced generates conflict with the
social dimension of the basin's production system. There may be a negative effect
on job generation; not only will the hiring of off-farm labor be reduced, but the
possibilities of present employment of family labor would also be affected.

Journalfor Farming Systems Research-Extension



2 101000 -

o 93000
63866 63322 42458 27929
Sediment (T)

Figure 2. Term of exchange between the reduction of the levels of sediment in the
watershed and the reduction of work days per year

Even a 40% reduction of coffee prices does not have an important effect over land
use in the watershed, as determined by the levels of sedimentation or the volume
increase in water supply. However, it does have a noticeable effect over employment
generation and income (Figure 3).

12 0

80 Workdays x 1000


40 Tons of Sediment x 1000

20 -

0 --
0 10 20 30 40 50

Reduction in Coffee Price %

Figure 3. Terms of exchange between the reduction of the percent decrease in
coffee price and the amount of labor (No. x 1000) and the quantity of generated
sediments (Tx1000)

Vol. 7, No. 2, 1997/2002


Developing better cultural practices for basic household crops and coffee activities
is a way of reducing sediments and the cost of not producing them. These practices
should have the potential to reduce the loss of soils and increase the productivity
of the crop. An investment of $60,000/ha/year for conservation practices on 751
hectares within the watershed could maintain producers' revenue and reduce
sedimentation, with $6,000/MT allocated for sediments that will not be produced.

3.7 Organization and empowerment of the community
Thevoluntary formation of eight CVP's was the cornerstone for sharing projectresults
and sensitizing the community to their own reality. The CVP's autonomously decided
to create the Associations of the Florencia Rain Forestand San Antonio River Watershed
communities (AVESELVA for its Spanish initials) as an autonomously legal entity.

Thanks to its convening power, ASVESELVA's negotiations facilitated the
development of the diagnosis and training process. Training priorities were
generated through consensus and included the following workshops: structure of
civil society's assets, endowment of resources and production potential of the San
Antonio River Watershed, the economic possibilities of conservation, negotiation
techniques, and strategies to resolve conflict.

The productive potential of the watershed, including job generation and future
options for the development of enterprises based on the tropical rain forest's
biodiversity, such as ecotourism, environmental services, sustainable agriculture,
extraction and bio-diversity (Vogel, 1996) were all themes addressed in the capacity-
building activities. The farmers discovered that a focus on biodiversity, better
positioned them to sell their property at a fair price to the Hydromiel enterprise. Their
position improved because it forced Hydromiel to revise the initial assessments and
redefine strategies. The organization process favored such positioning, despite the
enormous weaknesses of the initial phase, which was parallel instead of sequential.

The optimization model indicates that opportunities to capture external resources
are scarce, even though generating changes in land use can reduce current indices
of sedimentation and can increase the supply of water to the hydroelectric project.
Nevertheless, the 1993 Law 99 (Law of the Environment) points out that the
enterprise that manages the project has the obligation to transfer 6% of the resources
obtained from the sale of energy to conservation and development. Three percent
is allocated to the Autonomous Regional Corporation (Corporaci6n Aut6noma
Regional), CORPOCALDAS, to invest in environmental improvement and basic
sanitation in the watersheds providing water to the reservoir; 1.5% is for the
municipality where the dam as constructed, and the remaining 1.5% is distributed
among the municipalities that represent the watersheds that provide water to the
reservoir (Art 45 of law 99, 1993 "transfers from the electricity sector"). The annual
sale of energy is estimated to be 70 million dollars, from which $4,200,000 USD are
to be reinvested by the Hydromiel. From the latter amount, $1,050,000 USD have to
be invested in the watersheds that supply water. The San Antonio River Watershed
possesses 5.2% of the total area of the providing watershed (3,972 of 77,000 ha).
However, according to results from our research, it provides the project with 8.5% of
the water (7.2 of 84.3 m3/second), meaning that the community now has a pressure

Journal for Farming Systems Research-Extension


mechanism for defining priorities and allocating resources. If the distribution of
resources is based on the area, the San Antonio river watershed should receive
$54,600 USD a year. However, if the distribution is based on the contribution of
water flows, the watershed should receive $89,250 USD a year-63% more than it
would based on its area.

The recognition that there are few options to capture resources in order to pave
the way for agricultural practices with less impact over loss of soil has caused the
community to identify education as a priority and a strategy for development.
Thanks to this community's foresight, there is a local level Environmental
Education Program at the University of Caldas. This program is made up of.
concentrated modules, in which 38 students from Florencia and neighboring
communities participated in the first year. At the same time, the National Learning
Service (SENA) is carrying out the Program of Development of Qualified Workers
in Natural Resources, where 20 young farmers are being trained.

Despite the fact that Law 99 establishes the acquisition of areas or ecosystems
of strategic interest for the conservation of natural resources, civil society must
participate in the definition of conservation and the process of acquisition and
administration of land. The purchase of property for conservation in the Florencia
Tropical Rain Forest was done without the participation of the local people. Local
farmers are working on a proposal regarding the structure of civil society's natural
reserves based on technical information and training as another step in the local
community's empowerment process. This autonomous strategy is to guarantee
conservation, restoration and sustainable development in the Florencia Tropical
Rain Forest. Panayotou (1996) puts the theory behind their action, thus natural
resources as state goods become no one's property and soon turn into "no resources"
for everyone. Their proposal is supported by Law 99 of 1996 that privileges the role
of civil society in planning and managing resources.

ASVESELVA's initiative and negotiation resulted in two venues to resolve
differences. Representatives from the Mayor's office in Samand, a local political
presence, CORPOCALDAS, the regional environmental authority, and HIDROMIEL,
the enterprise that is building the Miel I hydroelectric complex participated. They
recognized the community as an actor within the process of development and
conservation of natural resources and opened negotiation and decision-making
space with the participation of the community. Further, CORPOCALDAS has
handed over the surveillance and control of properties acquired to ASVESELVA
and has asked its active intervention in the formation of a management plan for the
reserve. This process is currently under negotiation.

4. Conclusions

Producers' current revenues are sufficiently high to generate a better standard
of living than they could earn in the city, considering that in areas with less
productivity, average income is higher than a minimum salary. This is without
considering the basic crops that the producer uses for household consumption.

Vol. 7, No. 2,1997/2002


Water used in agricultural activities and their production of sediments do not show
a negative environmental impact in the watershed. Without attempting to assign
a value to the impact on biodiversity, the fact that a very high proportion of the
watershed is in forest and fallow indicates that producers are maintaining some
level of biodiversity within the watershed.

The decision made by the regional institution in charge of natural resources to
acquire ownership of the Florencia Rain Forest is unjustifiable. It is unjust because
it maintains that its acquisition will contribute to the conservation of biodiversity
and to the solution to poverty problems by shifting the population that currently
occupies the watershed to urban areas. State property acquisition led to the
disappearance of more than half of the existing forest resources during the years
immediately following World War II, as well as a very low social profitability
(Panayotou, 1996). The fact that Florencia Rain Forest's resources are state property
does not mean that they have to be administered by the state. It is urgent to take
advantage of organized venues to resolve differences to identify participatory
mechanisms for planning, negotiating and administrating the acquired properties
in order to avoid replicating negative experiences. It is also necessary to advance
follow-up to property acquisition process and evaluate if the purchase of property
by the regional environmental authority translates to natural resource improvement
and poverty reduction for the future.

The changes in cultural practices suggest such a slight impact over water production
of available water that the dam scenario is not an attractive one. Access to more
water does not constitute an important point of negotiation for producers, especially
because they cannot control their contribution to the reservoir.

There seems to be a similar situation with regards to sediment production. Levels
of soil loss are relatively low. To achieve substantial changes in land use in the
watershed, a greater investment than is anticipated ($2, 560/MT) will be required.
Because the dam is under construction; therefore, the results suggest that this would
not be the best time for land use change interventions.

Although the region's conditions are ideal for high soil degradation, results show
an important rationality for current land management. Production systems have
adjusted to reduce losses by erosion, making them stable in the long run:

* Basic household crops are the only ones that present low ground coverage
during a portion of the cropping cycle. However, planted areas are very small.
* The rest of the crops used (coffee, sugar cane and pastures) have high ground
coverage, and the length of crops in the rotation is over 5 years.
* The proportion of fallow and native forest is relatively high (between 19% and
76% of total farm area, according to zone)
* A cropping system of minimum till and machete for weed control assure
permanent vegetation, which reduces erosion problems.
* Due to economic reasons and the current coffee crisis, pastures are preferred
by producers to increase the farmed area or increase productivity. Pastures
maintain a permanent cover and a root system that reduces erosion problems

Journal for Farming Systems Research-Extension


In general, the current situation seems very stable, since it almost seems impossible
to modify the present cultural practices. According to the model, a 40% reduction in
coffee prices would be insufficient to promote changes in land uses. If family labor is
available, coffee continues to be the best option for small producers in regions with
pronounced slope. Likewise, there are no production alternatives to pasture. Thus
changes in cropping techniques will lead to very little impact on erosion. Under
these conditions, an increase in poverty levels is more likely to come from reduction
in the price of coffee than from an increase in erosion.

Despite the previous arguments, the model points out that there are scenarios
of sediment reduction that can be an object of negotiation with Hydromiel.
One scenario that benefits both parties (maintaining income for producers
and less sediments for the dam) can be achieved by shared investments in the
development of producer cultural practices that have less impact on the land. This
means developing production technologies for basic household crops and coffee
production to reduce erosion and increase productivity to maintain revenues.

Landslides cause as much soil movement as agricultural activity, which suggests
that they deserve the same attention when determining impact over the dam's
performance. The research did not permit the identification of technological and
infrastructure (road building, for example) factors that determine the presence and
magnitude of these landslides.

This research points out the utility of these type of models for simulating scenarios
that are not feasibly in practice to understand terms of exchange between
conservation criteria, job generation and peasant productivity. However, as Rivera
and Estrada (1996) mention, models cannot respond to all questions. Other key
questions include: Which would be the best technology to use? Who should
direct technology transfer? What are the social and environmental benefits of
acquiring properties for their conservation? What factors determine the presence
and magnitude of landslides? These questions must be answered using other

The information generated by models constitutes a starting point for valuing the
assets that producers possess, as well as the internalizing of these values and the
way they could formulate the terms for an eventual negotiation with designers
of environmental policies and beneficiaries of the conservation process, such as
energy consumers and Hydromiel. As Panayotou (1996) mentions, civil society has
a critical role in the transition of the economy and society from its current inefficient
and unsustainable path to an efficient and sustainable one. An empowered local
community, like the one that has started to consolidate in the Florencia Tropical
Rain Forest, through training, formal organization, and information they generate
and share, gradually become part of processes of project planning and the opening
of spaces for negotiation. That local community organization can: (a) generate and
disseminate information about causes and consequences of environmental problems
and the inappropriate effect of some policies, (b) increase opportunities for people
to give their opinions and make choices, and (c) aggregate the will of social groups
with scarce political and economic power to exercise pressure upon the governments

Vol. 7, No. 2, 1997/2002


and for political reforms (Panayotou, 1996). Even though the greatest worries the
community had during the project were related to the properties acquisitioned by
the State, the best spaces for the organization to act upon in the immediate future are
related to planning and administration of the reserve, along with the distribution of
resources that are generated from the sale of energy.


The authors would like to thank Dr. Guillermo Carvajal from CORPOICA for his
contribution in the analysis of the main components, to Jose Fernando Arroyave,
Jeimar Alirio Tapasco, Diana Marcela Tangarife, Ana Milena Nieto, Jaime Humberto
Arias and Camilo Augusto Agudelo, students from the Systems Analysis Research
Group at Caldas University, as well as to Augusto Tangarife, local coordinator for
the Eduquemos Foundation, for their valuable contributions during the research.

Cornelia Flora, Edith Fernandez-Baca, Sylvia Gutierrez and Susan Fey collaborated
in the translation.


Arroyave, J.E; Tapasco, J. A.; Rivera, B.; Obando, EH. 1998. Viabilidad del uso del modelo de simulaci6n de
perdida de suelo EPIC en zonas de ladera. Manizales, Universidad de Caldas. (Trabajo de grado de
Ingenieria Agron6mica).

Ashby, J.; Gracia, T; Guerrero, M.; Quir6s, C.; Roa, J. I.; BeltrAn, J. A. 1996. Organizaci6n de agricultores
investigadores para su participaci6n en la investigaci6n agricola y en el desarrollo de tecnologias. pp
235-254. In: B. Rivera; R. Aubad (eds). El enfoque de sistemas de producci6n y la incorporaci6n de
criterios de political. Santa Fe de Bogota, Colombia, CORPOICA.

Escobar, G.; Berdegu6, J. (eds) 1990. Tipificaci6n de sistemas de producci6n agricola. RIMISP, Santiago de
Chile, Chile. 284p.

HIDROMIEL. 1997. Proyecto hidroel6ctrico Miel I. Descripci6n general del proyecto. Manizales. 12 p.

Mitchell, G.; Griggs, R.; Benson, V; Williams, J. no date. The EPIC model. Environmental Policy Integrated
Climate. Texas Agricultural Experiment Station, United States Department of Agriculture.

Panayotou, T. 1996. Ecologia-economia, medio ambiente y desarrollo. Pp. 11-22. In: B. Rivera; R. Aubad (eds).
El enfoque de sistemas de producci6n y la incorporaci6n de criterios de political. Santa Fe de Bogota,
Colombia, CORPOICA.

Rodriguez, P; Carvajal, G. 1996. Caracterizaci6n de sistemas de producci6n. Fasciculo 3 del M6dulo
Aplicaci6n del enfoque de sistemas de producci6n, de la series M6dulos para la capacitacio6n
metodol6gica de las UMATA. B. Rivera (ed). CORPOICA, Santa Fe de Bogota, Colombia. 16p.

Rivera, J.J.; Estrada, R. 1996. Cuantificaci6n ex-ante del intercambio entire equidad, productividad y
sostenibilidad para el disefo de alternatives tecnol6gicas. El caso del cultivo de arracacha en Colombia.
pp 101-112. In: B. Rivera; R. Aubad (eds). El enfoque de sistemas de producci6n y la incorporaci6n de
criterios de political. Santa Fe de Bogota, Colombia, CORPOICA.

Smith, M. 1993. CROPWAT, Programa de ordenador para planificar y manejar el riego. FAO, Roma, Italia.

Journalfor Farming Systems Research-Extension


Vogel, J.H. 1996. El uso exitoso de instruments econ6micos para fomentar el uso sustentable de la
biodiversidad: seis studios de caso de America Latina y el Caribe. Cumbre de las Americas sobre
desarrollo sustentable. Santa Cruz de la Sierra, Bolivia. Diciembre 6-8 de 1996. Traducci6n de Consuelo

Wischmeier, WH.; Smith, D.D. 1978. Predicting rainfall erosion losses fron croplan east of the rock
mountains. Guide for selection of practices Soil and water conservation. United States. Department
of Agriculture. Agricultural Handbook No. 282. 1965. 47p.

Vol. 7, No. 2, 1997/2002

The Ecological Dynamics of
Low-External-Input Agriculture:
A Case Study of Hill Farming
in a Developing Country Setting

Yamuna Ghale1, Ganesh Shivakoti2 and Bishnu Upreti


This paper on Low-External-Input Agriculture (LEIA) explores how
the indigenous hill farming systems in a developing country setting
of rural Nepal are functioning and which ecological properties of such
systems could contribute to the development of Low-External-Input and
Sustainable Agriculture (LEISA). Three cases were studied using direct
field observation, group discussion, and interviews, and the important
practices of LEIA were ranked into eight classes to analyse the ecological
sustainability of the hill farming systems. The study shows that hill farming
systems are complex, diverse, heavily reliant on local resources, and focus
on the renewability of production resources within the farm. All this can
greatly contribute to make the system ecologically sustainable. However,
demands for increased production put considerable pressure on farmers to
apply external inputs which, if in excess, may jeopardise the whole system.
A prerequisite for guidance in careful soil management is strengthening of
supportive research aimed at teaching farmers how to use external inputs.
This must agree with the principles of sustainable agriculture and take in to
account the site specific variations which are characteristic of Nepalese hill
farming systems. We defined the term Ecological Dynamics as "the pattern
of changes in structure and/or forces in processes which govern the
development of natural and artificial ecosystems". This definition assists
in the identification of trends in ecological sustainability. Understanding
indigenous farming systems reveals important ecological clues which
indeed are essential for the development of sustainable agriculture.

Key words: Low-External-Input Agriculture, Low-External-Input and
Sustainable Agriculture, Ecological Dynamics, Ecological Sustainability,
Hill farming system, Production Resources.

1 International Centre for Integrated Mountain Development (ICIMOD), Jawalakhel,
Kathmandu G.EO. Box 3226, Nepal.
2 School of Environment, Resources and Development, Asian Institute of Technology, PO.
Box 4, Klong Luang, Pathumthani 12120, Thailand.
3 Department of Social Sciences, Agrarian Law and Legal Anthropology Group,
Wageningen Agricultural University, Hollandswegl, 6706 KN, Wageningen, The

Vol. 7, No. 2,1997/2002


1. Introduction

Agriculture as a bio-physical and economic activity is a direct outcome of farmers'
decisions and actions, which are strongly influenced by their socio-economic as well
as bio-physical environment. In the case of hill farming, biodiversity of hill resources,
management of resources, and institutional imperatives for hill farming are part of
that environment. Farmer's traditional strategies in managing hill agriculture in
most of the developing countries are influenced by the rapid increase of population
and a resultant food demand straining both hill farming and farmers' conservation
strategies. This may lead to excessive exploitation of natural resources; alternatively
it could elicit improved coping strategies and efficiency in the use of scarce resources.
Therefore, the system must be carefully examined to find clues to the impact of
increased food demand on sustainable hill farming systems. Farmers' approaches to
natural resources, and their innovativeness to evolving indigenous technologies and
adjustment strategies in changing resource situations, constitute the developmental
dimensions of hill farming system. Promoting traditional farming systems, i.e.
Low-External-Input Agriculture (hereafter called LEIA) is essential for the design
of sustainable agriculture. LEIA, in this article, means the maximum utilization of
locally available resources combined with reduced dependency on external inputs
for agricultural production process. LEIA is strongly integrated with nature and
depends on natural process (ILEIA, 1995; 1996). However, LEIA is not necessarily
economically feasible. Erosive forms of locally available resources use in LEIA may
result to over exploitation of natural resources with adverse ecological and farming
systems effects and impacts on marginality and poverty. LEIA is common in hill
farming in most of the developing country settings including Nepal. Therefore, the
contribution of LEIA to Low-External-Input and Sustainable Agriculture (hereafter
called LEISA) is a pertinent issue. LEISA refers to a form of agriculture that makes
optimal use of locally available natural, social and human resources, but the use
of external inputs are not excluded but rather seen as complementary to the use
of local resources. LEISA is a mode of farming system which provides long-term
sustained yields through ecologically sound management. Therefore, LEISA are the
farming systems explicitly designed and managed with enhanced sustainability as
a principal objective. The distinguishing elements of LEISA from the LEIA are the
regeneration and optimum use of local resources complemented by external inputs
to develop a farming system economically feasible, economically sound, culturally
adapted and socially just.

For the transition from LEIA to LEISA, first it is necessary to analyse hill farming at
the systems level. While LEIA practices are identified from participatory research,
these practices could be used to develop LEISA. Complementarity between
these practices by optimal use of external inputs based on action research is a
precondition to successful transition (Basnyat, 1995). The role of farming system
research in this transition process is to find ways and means to make optimum use
of resources (regeneration as well as external) and minimize risks (Reijntjes and
Moolhuijzen, 1995). In this process, Sriskandarajah et al., (1991) propose viewing
farming systems as 'sustainable learning systems in constant co-evolution with
their environment'. So the focus of farming systems research and extension should
be on helping farmers to create learning systems from the bio-physical and socio-

Journalfor Farming Systems Research-Extension


cultural environment that surround them to successfully transit from LEIA to LEISA
(Reijntjes and Moolhuijzen, 1995).

Sustainability is a dynamic, complex, and contested concept. Sustainability and sus-
tainable agriculture have been defined and analysed in numerous ways. One way
of looking at sustainable agriculture is to minimize the use of external inputs by
regenerating internal resources more effectively or by a combination of both. This
view focuses more on ecologically sustainable and economically viable alternative
uses of local resources and knowledge. We base our research analysis on sustainability
related directly to concepts of agro-ecology, resources (Harrington, 1992), and
interlocking aspects of ecological, economic and social systems (SARE, 1997). In the
agro-ecological sense, sustainability is enhanced through ecosystem complexity with
its resilience and species diversity as consequences. The diversity over time and
space fosters the recycling of nutrients, increased efficiency in the use of moisture,
nutrients and sunlight, and the reduced role of plants and animals as weeds, pests
and diseases (Altieri, 1987; Oldemann, 1990). Sustainability in the sense of resources
focuses on the continuing availability of resources over time (Batie, 1989; Harrington,
1992). Sustainable agriculture needs to be viewed as consisting of social, economic and
ecological systems. So, the possible contribution of LEIA to LEISA is analysed on the
basis of LEISA principles and characteristics (Altieri, 1994).

Mixed plantations with arrangements made according to biological characteristics
and mutualistic interactions are positive ecological results of using biodiversity
which provides ecological services such as nutrient cycling, biological pest control,
and water conservation (Altieri, 1994). The higher biodiversity of plants, microbes,
and animals inherent to these systems support farm production. According to Ingels
and Campbell (1995), diversified LEIA farms would be ecologically resilient and
could buffer a farm in an ecological sense. Comprehensive systems level research
and extension of integrated pest management, rotational grazing, soil conservation
water conservation, covers crops, landscape diversity, nutrient management, agro-
forestry, and marketing contribute towards the promotion of LEISA (Lightfoot et
al., 1993).

In this paper we have defined the term "Ecological Dynamics" as 'the pattern of
changes in structure and/or forces in the processes which govern the development
of natural and artificial ecosystems'. The changes are both temporal and spatial.
The ecological dynamics in this study cover the practices of the complex farming
systems of LEIA which exhibit ecological properties such as biological diversity,
nutrient cycling capabilities, risk aversion/resilience development, soil and water
conservation features, pest suppressive potential, etc., sustained with the help
of ecological communities including humans, animals, pests, weeds, trees, soil
organisms, etc., within the farm.

2. Overview of the Study Sites and Methods

The study was conducted in a remote rural village in the mid-hills of eastern Nepal.
Three farms were selected at different positions along the elevational gradient

Vol. 7, No. 2, 1997/2002


within one Village Development Committee (VDC)'. The lowland farm (Farm 1)
"was located at an altitude of 700 meters from mean sea level whereas the midland
farm (farm 2) was located at 1100 meters and the upland farm (farm 3) was situated
at an altitude of 1900 meters. The slopes of all three farms were southeast facing.
Each of these three farms are analysed in subsequent sections. Interviews conducted
with thirty other farmers gave a general impression of the farming systems
which was substantiated through other research methods such as key informant
interviews, focus group interviews and semi-structured questionnaires (Mukherjee,
1993). The sites were assumed to reflect the indigenous farming system, under
minimal influence of external agro-technologies. Primary information was collected
from farmers through semi-structured interviews and direct observation by the
researcher. Important farm characteristics were ranked in eight purpose classes.
In-depth analysis of three farms was done to explore the ecological sustainability of
the farm. Analysis was supplemented by the information collected from thirty other
farmers within the study area. The research was exploratory in nature with primary
observation being processed by ranking into classes. Qualitative data collected
from different sources were analysed descriptively by using the LEISA concept.
The interpretation is mainly qualitative because of LEIA being rooted in farmers'
indigenous knowledge and decision-making process.

The major criteria for selection of the three farms are the representation of altitude
specific farming practices. Table 1 provides the details of case studies conducted
in the sites. The three farms represent three different points on the altitudinal
gradient, i.e., farm 1, 2 and 3 situated at different altitudes with a particular focus on
a specific land use pattern. For each case study, a detailed description is made based
on the response from farmers interviewed, while the general description is based
on information provided by other 30 interviewed farmers residing and practising
similar type of farming as those of case study farmers. On each farm, practices are
categorised in one out of eight purpose classes. There are four ranks which range
from very common practices to common and seldom to none practices.

2.1 Characteristics of the Farms Studied
Many farming practices are common among the study farms; the differences that
exist are explained prior to our summarising the ranking of LEIA practices in the
hill farming system. These include practices of soil, water and fertility management,
including livestock and crop management.

The lowland terraces are developed in flat areas with inward slope for better
water stagnation to grow paddy and minimise soil loss. The upland terraces are
developed with outward slope to prevent the collapse of terraces by water flow and
to prevent the problem of water stagnation in upland crops. Effective methods of
soil stabilisation can be seen at the highland farm at Zalzale, ward number six of the
VDC. The farmer first develops relatively large terraces in sloping areas and plants
locally available perennial grasses. After the establishment of vegetation, the grasses
are removed from the topmost terrace and converted to annual crop land gradually,
leaving the risers and lower terraces in perennial grasses. This technique helps to
stabilise the soil by conserving moisture as well as preventing soil loss.

Journal for Farming Systems Research-Extension


The farmers in the lowlands and midlands use chemical fertilisers on wheat and
maize. Green manuring is common only in paddy nurseries due to the lack of
beneficial biomass and shortage of labour. In-situ green manuring crops are not
used because of lack of technical know-how and source seed. Irrigation is the
most important factor in water management at low and midland farms, where
sedimentation and inundation are common problems. Maintenance of terraces,
balance in terms of the period of water stagnation, and development of drainage
channels are common practices evolved to overcome these problems. These
problems are not found in highland farms due to existence of rain-fed farming
practices. Water harvesting is common in highland farm due to lack of perennial
sources of water.

Forest is one of the important components in the hill farming system as it directly
supplies fodder and litter and indirectly maintains ecological and environmental
balances, However, forest accessibility has emerged as a critical problem during
recent years. As forest accessibility decreased and population increased over the
period of time, marginal land was converted to annual crop land which led to a
shortage of fodder. As an initial step taken to solve this problem, the farmers have
started to grow oat (Avena sativa, Gramineae) seed, which has been perceived by
farmers as an excellent alternative to fodder, especially for milking stock. Thus,
farmers have started to produce oat seeds and feed for milking stock, which has
increased milk production, and these farmers have been satisfied with the result
achieved. The local breed of livestock is common on lowland and midland farms
because the local breeds are already adapted to the local environment and provide
stable production using local fodder and forage. The use of improved breeds in the
highlands is seen as an income generation strategy through high milk production.
In contrast to highland practices, grazing is very rare on low and midland farms due
to the scarcity of pasture land.

In the case of plant diseases and pests control, the use of local herbal extracts and
products and plant species of repellent character are more common on dry lands
of mid and highland farms. Crop rotation is an important strategy to control
weeds. One of the case study farmers has adopted an innovative approach of crop
rotation. Among the several crops, the farmer cultivates ginger as a cash crop, but
changes the site every year, citing the reason that ginger makes the soil pungent for
other crops-especially tubers and bulbs. This type of knowledge gained permits
farmers to use the soil very effectively which also helps in weed control. Although
the fallow period is comparatively longer on highland farms due to low cropping
intensity, fallowing is not a common practice of weed management because it is not
an economically feasible option for smallholders.

Cropping intensity is relatively high on low and midland farms compared to
highland farms due to the climatic condition of the locality. Use of chemicals in
storage is common on low and midland farms due to the ease of access to the
market. These chemicals are not, in general, easily available or affordable, although
farmers use them at minimal rates to protect the seeds for next year.

Vol. 7, No. 2, 1997/2002


3. Case Studies Findings and Discussion

Hill farming systems in the case studies area are heavily dependent upon the use of
local resources. Human and animal energy is regularly used and labour is pooled
by families in peak seasons. The farmers mostly use local varieties of crops. The
mobilisation of common property resources in relation to production processes is the
strong characteristic of the study area. From the study, it is observed that farming is
partially a collective business. Pretty (1995) explained that farmers worked together
on resource management, labour sharing, marketing, and other activities that would
be too costly or even impossible for the individual. The study shows that the farmers
are putting effort for efficient utilisation of indigenous knowledge and experiences
to manage their farms. Locally available resources like water, forest, and public
grazing lands have been utilised regularly by the local people. Working in a group
was seen as an important channel for the exchange of information and inputs. There
are several women farmer's groups which also serve as a platform for other off-farm
jobs and social activities. The male farmers act as a back-stopping committee for
them to make common property related decisions. The role of women farmers is
pertinent in the hill farming systems as they act as principal actors in the agricultural
production processes. Such local groups provide basis for discussion and collective
action, building consensus, and information exchange. The gender sensitisation
process is gradually started. Ostrom (1990), Bromley (1992) and Pretty (1995) support
this concept through their argument that when the common users come together, it
creates opportunities for collective action and mutual assistance and mobilising
resources on a self-sustaining basis. If the people feel more sense of obligation and
mutual rapport, it is an easy and effective way for resource conservation (Upreti,
1993). We discuss the findings of the case studies in the following section from the
perspective of ecological dynamics defined earlier.

The management practices of each purpose class in the hill farming systems and
their inter-relationships are discussed as follows:

3.1. Soil management
Hill farmers have their own strategies and approaches of soil management in
their farms. Case studies reveal that indigenous local soil management practices
like tillage, mulching, terracing, plantation, gully control, etc., are common in hill
farming systems. These practices rely mainly on local resources and indigenous
knowledge, which are temporally as well as spatially diverse in nature (Altieri, 1994;
Oldemann, 1996; AGRICOL/ILEIA, 1988). Nutrient recycling and soil conservation
are major strategies in soil management. Tillage practice is more common in low
land and midland farms as compared to the highland farm, whereas mulching is
seldom used in all these farms. Terracing is more common in the highland farm.
Gully control is more common in low and midland farms compared to the highland
farm mainly because of lesser severity of the problem in the highland farm.

As viewed by Ingels and Campbell (1995), the soil as a living medium should be
protected for long-term productivity and stability. Continuous vegetation cover
and root mass promote nutrient retention, reduce erosion, increase organic matter,
improve soil structure and texture, enhance fixation of nitrogen, prevent leaching,

Journal for Farming Systems Research-Extension


Table 1. Ranking of LEIA practices in hill farming.
Management practices FARM 1 FARM 2 FARM 3
Lowland Midland Highland
1. Soil management
Tillage ++ ++ +
Mulching -
Terracing + + ++
Plantation + +
Gully control + + -
2. Fertility Management
Use of fertilizer --
Use of FYM ++ ++ ++
Multiple cropping + ++ +
Cover crop -
*Legume + + +
In-situ manuring + + +
Green manure -
3. Water Management
Erosion control ++ ++ +
Sedimentation control + +
Inundation + +
Dams -- -- --
Water harvest + +
Irrigation ++ ++
4. Livestock management
Grazing ++
Fodder supply ++ ++ ++
Use of local breeds ++ ++ +
Feeding crop residues ++ ++ ++
5. Insect pests management
Use of chemicals -
Mechanical control + + -
Use o repellents -- -
Use of predators -- -- +
Multiple cropping + + ++
Crop rotation ++ ++
Trap cropping- -- -
Use of herbal extractions ++
6. Weed management
Use of chemicals -- --
Cultural practices ++ ++ ++
Fallowing- -- -
Crop rotation + + +
Burning of crop residue +

Table 1 continued on next page

Vol. 7, No. 2, 1997/2002


7. Crop management
* Inter cropping ++ ++ ++
* Cropping intensity ++ ++ +
* Crop rotation ++ ++ ++
8. Post harvest management
* Cleaning ++ ++ ++
* Drying ++ ++ ++
* Processing -- -- --
* Packing -
* Storage ++ ++ ++
Use of local resources
* Use of chemicals --
Note: There are four ranks, + + very common, + common, seldom, -- none.

promote the activity of decomposers, and keep the soil healthy. This provides a
strong base for sustainability (Jodha, 1993). These practices are common phenomena
adopted by the case study farmers and sample households.

3.2. Fertility management
Fertility management is a specific component of soil management in hill farming
systems. Study reveals that major activities in fertility management practised by
farmers are use of fertilizer and farm yard manure (FYM), multiple and cover
cropping, legume plantation, and in-situ and green manuring. Nutrient recycling
is the major strength of fertility management in Nepalese hill farmers' systems.
The farmers rely heavily on locally prepared manure, mainly FYM, and use green
manure in rice fields to a limited extent in all farms. Farmers from the highland area
do not use chemical fertilizer whereas farmers from the mid and lowland farms use
fertilisers and cover crops. Research from the Pakhribas Agricultural Centre (PAC),
at one of the hill farming system research sites in Nepal, also shows that paddy with
10 ton of asuro (Justicia adhatoda; Acanthaceae) leaves, a common practice used by
low-to-mid altitude farmers for generations, produces 40% higher yield than paddy
treated in the chemical fertilisers (Devkota, 1994).

Soil is protected and nutrients are managed by applying organic manure, growing
nitrogen fixing plants and deep rooted plants, enhancing crop rotation, fallowing,
and mulching to a lesser extent in all farms. This creates a favourable micro-climate.
The use of different crop residues and ash balances the soil life, increases soil fertility,
and satisfies the specific needs of the crops. Healthy soil is the main requirement for
ecologically sustainable farming, which produces healthy crops having optimum
vigour and less susceptibility to pests (Vaidya et al., 1995). Proper management of
soil, water, and nutrients can prevent the plant from having pest problems and make
efficient use of them. The farmers of India have also discovered that plants grown
without chemicals demand less quantity of water, and that trees need only 30%
irrigation water compared to the normal amount (Save and Sanghavi, 1992).

The living organisms in hill farming systems, like any subsistence farming system,
are inter-dependent and inter-linked. In the farms studied, nutrients are collected
from the forests, croplands, and/or rotational systems, or by including legumes

Journalfor Farming Systems Research-Extension


in the cropping pattern, by use of small amounts of chemical fertilisers together
with available amounts of FYM and green manure and other processes to enrich
the soil. Grazing in the pasture land enriches the soil with animal droppings. The
use of crop residues and by-products for animal feeding also helps in the recycling
of nutrients. Manure along with bedding material is retained in the field. During
the dry season, livestock graze in the cereal crop residues and also maintain the life
cycle of materials and wastes through effective recycling practices. Thus, the key to
the success of a farming system is crop-livestock integration, involving the recycling
of nutrients within the system as much as possible as evidenced by the field cases.
The LEISA system seeks to reduce the deficit between exports and imports by using
intrinsic recycling, natural replacement and minimising the artificial replacement
of lost material. The process of nutrient recycling of each farm can be seen in the
following figures 1, 2 and 3.

High incoming FYM N fixation by
runoff rainfall peas and soyabean

Occasional use of
Inward terrace mineral fertilisers

Broad bean and Asuro use for
soyabean as FIELD green manuring
in-situ manuring
in-s-itu manung .Occasional burning
r *- of straw and
/ fodder residues
Soft rock weathering
Decomposition *
Erosion/runoff OUTPUT,

Erosion/runoff I

Plant removal

Leaching _

- De-nitrification 4A

- Volatilisation a


SLOSS (fertility)

Fodder LIVESTOCK People

Night soil
Leaf litters Comoost/FYM (specific crops

like chyote
grown in the
ridges of night
soil pits)

Figure 1. Nutrient cycle in a lowland farm in the hill farming system in Nepal
Source: Adapted from Sherchan D.P and B. D. Gurung, 1995

Vol. 7, No. 2,1997/2002



High incoming FYM N fixation by
runoff rainfall peas and soyabean

Normal inward Occasional use of
terrace mineral fertilisers

Broad bean and Occasional use of green
soyabean as --FIELD asuro for manuring
in-situ manuring
in-situ manuringBurning of wasted straw


f f *iiiu~u

Leaching -(

- De-nitrification V

- Volatilisation


Plant removal

LOSS (fertility)

Fodder > LIVESTOCK People

V Night soil
Leaf litters --. Compost/FYM (specific crops
like chyote
grown in the
ridges of night
soil pits)

Figure 2. Nutrient cycle in a midland farm in the hill farming system in Nepal

In all of the three farms, fertility losses in general were reported as common
phenomena, but for some specific losses such as volatilisation, the farmers could not
identify as a potential cause of fertility reduction.

3.3. Water management
Farmers have a good idea of local level soil and water conservation processes. To
cope with water scarcity and unequal distribution of water, farmers follow different
strategies such as growing drought tolerant crops in low moisture areas, practising
water harvesting and moisture conservation technologies like collection ponds,
aquifer recharges, etc., developing terraces, and planting perennial shrubs and trees
in all three elevation farms. In areas with water availability, they generally practice
rice farming. Erosion control is the most common practice followed by farmers of

Journal for Farming Systems Research-Extension



High incoming
runoff relatively
high rainfall

Outward terraces \

Broad bean as
in-situ manuring


N fixation by
peas and bean

No use of
mineral fertilizers

Soft rock weathering



Leaching 4

- De-nitrification -

- Volatilisation


f I

V Burning of straw and
fodder waste

Decomposition (


Plant removal

LOSS (fertility)

Fodder LIVESTOCK People

S Night soil --
Leaflitters -- Compost/FYM (specific crops
like chyote
grown in the
ridges of night
soil pits)

Figure 3. Nutrient cycle in a highland farm in the hill farming system in

the study area. Water harvesting through dams is not practised at all for agricultural
purposes. However, some respondents explain that they collect water in ponds to
feed their animals. Water management is more common in low and midland farms
than in the high land farm.

3.4. Livestock management
Livestock management is an integral part of the hill farming system in the study
area. Hill farmers use local breeds of animals because they are adapted to the
local environment, can grow with minimum care, and are resistant against certain
diseases and pests. These local breeds and technologies make proper use of the
agro-ecological processes of predation, competition and parasitism to control pests,

Vol. 7, No. 2, 1997/2002


which is the best alternative against the use of chemicals. Grazing is more common
in the high land farm as compared to mid and lowland farms mainly because of
availability of grazing land in high altitude areas. Feeding crop residues like maize
straw, wheat straw, etc., are common in all farms. Farmers from the mid- and
lowland areas prefer local breeds because of high adaptability characteristics of local
breeds. In the highland area farmers are gradually moving to improved breeds due
to the availability of grasses and the profitability of marketing livestock products.
For livestock treatment, local measures and belief in the local faith healer is more
common than use of scientific veterinary practices. Farmers are well experienced in
identifying animals suffering from internal parasites and in control through local
medicines and other practices.

3.5. Insect pest management
Hill farmers have tremendous knowledge and experience in dealing with insects
and pests. For pest and disease control in plants, they apply local herbal extracts.
A mixture of different crops provides insurance against insect pest attack. Use of
local resistant varieties, botanical pesticides and repellents, and cultural practices
minimises pest interference. Use of ash, sawdust, bordering of main crops by
protective crops, etc., are very effective local measures practised by farmers to cope
with insects and pests. Crop rotation is one of the most common strategies used by
farmers to tackle the insect pest problem. However, farmers are not deliberately using
predators in their farms. Nevertheless, they explained that several insect predators
are naturally found in their farms, which are controlling many harmful insects.
Farmers of the study area explained that diversity decreases pests and diseases
epidemics and encourages and maintains the population of beneficial predators.
Altieri and Letourneau (1982) support the argument that in LEIA the population
of pests are regulated by natural assemblages of predators and parasites. Save and
Sanghavi (1992) reported that Indian farmers demonstrated that plants nourished
in a natural way are very healthy and develop strong resistance to disease so that
pesticides are not needed. The pests can be controlled biologically, and the plants
need care instead of cure. The weeds and soil-borne pests that survive in dry and
submerged condition have been controlled by flooded rice cultivation in the study
area. The crops rely on biological inter-dependencies that provide some biological
pest suppression. These LEIA practices can contribute greatly to developing LEISA
in hilly areas of developing countries.

3.6. Weed management
Indigenous weed management practices are a major component of the hill farming
systems in the study area. Farmers from this area were not using herbicides to control
weeds in their farms due to several factors like unavailability of herbicide, lack of
knowledge, and cost ineffectiveness. They explained that crop canopies effectively
suppress weed growth and minimise the need for weed control. Mulching provides
the benefit of moisture conservation and protection from weeds. Fallowing is not
common in low and midland farms due to the scarcity of land, whereas farmers from
the highland area do practice fallow to control weed and to graze animals. Burning of
crop residues is also practised in high land farms in contrast to low and midland farms.
Crop rotation is another strategy to control weeds in the study area. However, crop
rotation is done not only for weed control but also to match other farm requirements.

Journal for Farming Systems Research-Extension


3.7. Crop management
Intercropping, crop rotation, and crop intensification are the major crop management
practices of the study area. Generally farmers grow traditional varieties of crops.
They are saving seeds from previous years to continue farming. Pretty (1995)
and Chambers (1991) highlighted that when farmers grow and save traditional
seeds, they are also saving and reviving the principles, concepts and approaches
of sustainable agriculture which working close co-operation with nature, i.e., LEIA
(Carson, 1992).

Farming systems are both diverse and complex in the study area. Intercropping,
multiple cropping, crop-livestock integration techniques are the common practices
in the farms studied which enhance the biodiversity of the farm. Therefore, agro-
functional diversity can be achieved by combining plant and animal species that
have complementary characteristics. These are involved in positive, synergetic
interactions, so not only the stability, but also the productivity of LEIA systems can
be improved. The biodiversity of the farming systems is increased by manipulating
planting configurations or soil and water conditions, so it performs both production
and ecological functions. Biological diversity in the context of the study area helps to
achieve year-round food sufficiency and diverse diets, stable production, minimum
risk of failure, intensive production with limited resources, and reduced reliance
on scarce and expensive agricultural chemical inputs. It produces maximum return
at local levels of technology and develops production systems that rebuild the
regenerative capacities of their holdings.

Different practices of crop and land management can create temporal and spatial
diversity. From the study, it is found that the land immediately surrounding the
household is a most intensively cropped zone with vegetables, fruits, millets,
beans, ginger, etc. This is followed by a second, relatively larger but low manuring
area where short fallow is used for maize, paddy, wheat, and mustard. The
uncropped outer zone is generally used for grazing. The farmers follow cropping
practices for climatic suitability to the locality. The diversity and complexity of
the farming systems, including plants with different growth habits, canopies,
and root structures, make efficient use of space and nutrient resources, due to the
complementary biotopes of the crop and other species (Chand and Thapa, 1992).
These combine species diversity and structural diversity in time and space through
both vertical and horizontal organizations of crops. They exploit the full range of
micro-environments, which differ in soil, water temperature, altitude, slope, and
fertility within a field. Increasing the temporal or spatial diversity of plants in an
agro-ecosystem tends to prevent the build-up of crop-specific pests to economically
damaging levels. The correct spatial and temporal assemblage of crops, livestock,
trees, and soils enhances the interactions that support yields, which depend on
internal resources, recycling of nutrients, trophic relationships among plants, and
soil organisms that enhance biological pest control. These findings in temporal and
spatial diversity from the case study are in line with the argument put forward by
Altieri (1994) and Oldemann (1996).

Farmers have time-tested experience with coping strategies that have led to
mixed and diverse farming systems. Due to multiple cropping and crop-livestock

Vol. 7, No. 2, 1997/2002


integration, constant food production and soil coverage is ensured and the risk of
failure is minimised (Shivakoti et al., 1977). By multiple cropping and keeping high
levels of biomass in the soil under perennial grasses, crops, shrubs, trees, animals,
the soil organic matter and nutrients, they minimise the risks of drought, soil and
nutrient loss, pests and disease incidence.

3.8. Post harvest management
An indigenous approach to post harvest management is an unique strength of
the hill farming system. Proper cleaning and drying of storage grain and use of
local resources to save the stored grains are most common strategies of farmers.
However, special packing and application of chemicals to save stored grains are
not the priority of farmers. Hill farmers keep their food grain in locally made
storage bins. The accepted cosmo-vision is equally important for the storage grain
protection. The local residents have the strong belief that if the grains are dried
well in the sun, there will be no problem with mould. For this, the farmers follow
the lunar calendar believing that the grain should be dried in the sun on the day
of new moon light. This method is very effective in practice, and does not require
chemicals for the control of storage pests. Women farmers, who are the main actors
responsible for handling storage grain, explained that grains dried on the full moon
day have comparatively high storability and less incidence of weevils and other
storage grain pests. They also used several herbs to protect the stored seeds. Women
farmers explained that keeping grain in an area where air circulates well and
mixing different types of grains (for example, mixing maize and millet) gives high
storability, less damage by storage grain pests, and good quality. This practice is also
very common in other areas of eastern hills of Nepal (Upreti, 1995).

Earlier proposition by Ingels and Campbell (1995) that LEIA farmers maximize
reliance on natural, renewable and on-farm inputs as a sustainable approach, which
is least toxic and energy intensive, while maintaining the ecological basis of farming
systems, holds true for this study area as well. Therefore, the contribution of LEIA to
develop LEISA needs to be based on these elements, which are economically feasible
as well as less health hazardous.

4. Conclusion and Recommendations

The study shows that the farmers are managing the crops and lands in order to
get diverse production. They place primary importance on maximizing security
by producing subsistence crops and by growing some crops and livestock for
marketing as a source of reasonably stable income. For this reason, they grow
mostly local varieties of crops and local breeds of animals to get stable production,
and local crop varieties that have good storage capability. The straw-grain ratio is
another reason to discontinue modern varieties as their livestock depends on crop
residues. They diversify farming to reduce the risk of complete crop failure and
distribute labour, equipment, and other production inputs throughout the year
to reduce dependency on outside products. Diversified systems such as multiple
cropping and/or intercropping are common practices in the field to keep the
ground covered for most of the period of the year in order to reduce the risk of

Journal for Farming Systems Research-Extension


erosion, secure food availability and diverse diet, as well as maintain a broad genetic
pool. Hence, indigenous LEIA practises in terms of health care, food storage, pest
minimization, etc., as a result of farm diversification, facilitate sustainability of the
farming systems. Besides diversification as a means of risk aversion, we can make
the following conclusions from the discussion of the study.

Break-even point of LEISA
LEIA ^ ^ .-^-^ HEIA

Figure 4. Sustainability graphed against the use of external input

4.1. Low dependency on external inputs
Low dependency on external inputs is one of the basic tenets of LEISA. The farms
studied depend on external inputs to a limited extent. Farmers using supplemental
doses of chemical fertilizers do ensure that they achieve crop potential. Research
conducted by the Lumle Agricultural Center in Nepal found that a top dressing of
nitrogen fertilizer on maize, at the rate of 3 kg urea per ropani2 during the second
weeding and earthing-up, as a supplement to FYM, increased yield by 15% in mid
and high altitudes (Jayaswal and Gurung, 1992). This practice is common in the low
and midland areas, and farmers said that it is affordable with no adverse effects
on ecology and ensures stable production. Rotation with legumes also increases
reliance on biological fertility, reduces the need for external nitrogen, and helps
in pest management, which minimizes the demand of external pesticides and is
environmentally safe for the farming community. The farms in the study area are
barely influenced by outside technologies and inputs. This is illustrated in figure 4
below, where the farmers are not willing to limit themselves only to LEIA and are
inclined to increase use of external inputs but have not crossed the "break-even" point
of LEISA. Extensive use of external inputs and low level of internal regeneration
causes the change in the shape of the curve (fig. 4) at the break-even point. In other
words, input optimization causes the change in shape of curve in break-even point.
If more inputs are used without considering the law of diminishing return and
over-exploitation of resources, it will not be sustainable economically as well as
ecologically. However, the effect of increases in external inputs depends on the types
of external inputs, application methods, and geographical specificity. This kind of
trend in farming systems exhibits important elements of sustainability adapted
to their environment, reliance on local resources, and conservation of the natural
resources base, with a complementary use of external inputs. Thus, the limited
dependency on external inputs is an indicator of self-sustaining farming. A LEIA
approach in hill areas where few external inputs are available can be considered
as a basic step towards LEISA. The finding of this study is in line with the theory

Vol. 7, No. 2,1997/2002


developed by Wolfert et al. (in preparation), which seeks a balanced use of internal
and external inputs to make a farming system sustainable.

4.2. Local resource utilization:
The study reveals that the farmers are making proper use of local resources
on their farm. The farm production practices that exist in the study area such
as agroforestry, integrated pest management, integrated nutrient supply and
recycling, multiple cropping systems, integration of livestock, water and nutrient
harvesting and conservation, micro-climate management, and selection and
breeding of crops and livestock, based on maximum local resource use, form one
stable agro-ecological system. Additional inputs like integrated nutrient supply,
integrated pest management, appropriate selection of crop varieties and animal
breeds, optimal use of external inputs, high internal regeneration, and application
of indigenous knowledge and skills help to increase sustainability as they prevent
the overexploitation and degradation of local natural resources and also enhance
production process in hill farming system. Sustainability in the context of hill
farming system refers to the capacity to remain productive while sustaining the
resource base (Kessler and Moolhuijzen, 1994; Reijntjes et. al., 1992). Therefore,
the basic indicator of the sustainability of the hill farming system is the ability to
maintain and enhance its production performance without damaging long term
production potential. In other words, sustainability can be measured by: the
optimal use of local resources supplemented by external inputs, use of available
knowledge and experiences, maintenance of biological diversity, increased soil
water conservation capacity, use of risk aversion measures, pest management
and suppression, enhanced nutrient recycling, moblisation of common property
resources, etc. The resources mobilised by these interrelated components are the
potential features of LEIA that can make contributions in developing a LEISA
system. One of the basic challenges of sustainable agriculture is to make better use
of local resources (Pretty, 1995). This can be done by minimising external inputs and
regenerating internal resources more efficiently or by a combination of both. The
indigenous hill farming systems in Nepal and other developing countries clearly
exemplify this ability.

4.3. Local technology development and dissemination:
Nepalese farmers have been practising different indigenous systems through crop-
livestock-forest integration for centuries. Terracing, slicing terrace risers, flood
water harvesting, application of organic matter, in-situ manuring, and inclusion of
legumes in crop rotations are all built-in agronomic practices. These are the result
of continuous efforts over many generations that supply plant nutrients to the
field and enhance the soil life. Farmers differentiate between human managed soil
productivity and inherent soil fertility. Productivity is the indicator of soil fertility.
Indigenous practises are dependent upon local conditions and resource availability.
Mixed cropping is a very common crop management practice. These LEIA
technologies make use of the agro-ecological processes of predation, competition,
and parasitism to control pests effectively. Hence, local techniques of farm
management indeed are well adapted to their environment, rely on local resources,
and conserve natural resources sustainably. The development and dissemination of
local technology, the optimal use of natural resources, and the use of external inputs

Journalfor Farming Systems Research-Extension


in a restricted level to complement farm production, are the criteria of ecological

4.4. Adjustment to adverse conditions and resilience development
The more complex the ecological community, the more buffered it is against
disturbance. Stability can be thought of as the ability of an ecological community to
continue to function when stressed, because it is a living system and is comprised
of components that reproduce. Stability also embraces the idea of regeneration.
In ecological terms, the ability to regenerate following a stress-in other words,
resilience-is found on the LEIA farms studied here. Thus, the diversified LEIA
farms can cope with the situation of scarce resources and unforeseen farm problems
(drought, for example) and still be able to maintain subsistence production levels.

4.5. Recycling of resources
Recycling of farm-produced resources is the main ecological feature of LEISA. Use of
renewable products is a common characteristic found in the hill agriculture studied
here. The greater the recycling of resources, the more the system becomes more
sustainable (Axinn and Axinn, 1987). Interdependence between crops, livestock,
forests, and fodder is the key issue in hill farming systems, where the livestock,
forests and crops contribute to the synthesis of FYM that are the major sources of
plant nutrients (Shivakoti et al., 1997). The field, forest and pasture provide feed
and bedding materials to animals, and in return the fields and pasture land receive
nutrients from livestock, forming sustainable management of nutrients (Shrechan
and Gurung, 1995).

4.6. Development of local institutions
Emergence and re-strengthening the support of local institutions is a sign of
empowerment and a realisation of the importance of local common property
resource conservation and management. The emergence of milk collection centres,
dealers, users committees for irrigation, drinking water, and forests, women groups,
and an agricultural group are the collective binding force in the study area to
accelerate the development activities. This in turn has developed we-feeling among
the users, which is a successful indicator of sustainable farming. This social tie-up
is a strong force for common decision-making about resource use and sustainable
agricultural development. The rules and regulations formed by the local groups
help to control the over-use and/or misuse of resources.

In summary, it can be concluded that hill farmers in developing countries in general
and Nepalese hill farmers in particular are managing their farming systems by
careful management of soil, water, nutrients, and local resources as needed to make
farming sustainable. These systems are fine-tuned to the specific environmental
conditions along the altitudinal gradient. The basic principles behind the different
farming techniques in different altitudinal gradients are general but the specific
techniques are usually applicable only to a particular site. Therefore, understanding
indigenous farming systems reveals important ecological clues for the development
of alternative self-reliant, economically viable, ecologically sound, and socially just
sustainable agriculture and resource management systems.

Vol. 7, No. 2, 1997/2002


5. Recommendations

On the basis of this study which dealt with the ecological dynamics of LEIA and hill
farming system practices, we propose the following recommendations to be focused
on possible alternatives and measures for the development of LEISA:
(i) Crop dynamics should be emphasised so as to substitute improved biological
soil management, like nitrogen fixation, for increased chemical fertilising. A
participatory research on nutrient recycling and selection of appropriate green
manuring crops will be helpful for farmers to transit from LEIA to LEISA.
(ii) External input dynamics should be focused towards appropriate technologies
which can combine internal and external resources for conservation of
natural resources, enhancement of productivity, security, and reduction of
environmental damage. System oriented research will contribute to developing
an optimal level of external input use and to finding opportunities and
constraints for the development of LEISA.
(iii) Biodiversity dynamics should consider the profound knowledge of farmers
for agricultural innovation and development of LEISA. For that, research
on effective utilisation of indigenous knowledge and practices as well as
incorporation of suitable technologies to promote LEISA are the starting point
to be considered by agricultural research policies and programmes.
(iv) Nutrient dynamics need attention towards greater awareness of the potential
of nutrient recycling within the farm and efficient methods of manure
handling. Participatory technology development approach for effective
nutrient management and recycling involving farmers is an important step in
developing sustainable hill farming systems.
(v) Local institutional dynamics should pay attention to community based actions
through local groups and institutions for efficient use of production resources.
Action research on role of local institutions (which local institutions do what,
why and how) needs to be explored to mobilise them for successful transition
from LEIA to LEISA.
(vi) Governmental dynamics such as policies and strategies on research, extension,
marketing, and infrastructure should be emphasised to promote LEISA. To
improve the hill farming systems through LEISA approach, these needs to be
analysed in terms of their functioning and the opportunities and constraints of
institutional and policy process both at national and local levels.


The authors are thankful to Prof. Dr. R.A.A. Oldemann, two anonymous reviewers,
and Dr. John S. Caldwell for their comments and suggestions on the manuscript.


AGRECOL/ILEIA. 1988. Towards Sustainable Agriculture. In: Return to the good earth; damaging
effects of modern agriculture and the case for ecological farming. The Third World.
Altieri, M. A. 1987. Agro-ecology: scientific basis of alternative agriculture. Westview Press, Boulder, CO.

Journal for Farming Systems Research-Extension


Altieri, M. A. 1994. Biodiversity and pest management in agro-ecosystems. Haworth Press, Binghamton, New

Altieri, M. A., and D. K. Letourneau. 1982. Vegetation management and biological control in agro-ecosystems.
Crop Protection 1: 405-430.

Axinn, N.W, and G.H. Axinn. 1987. "The recycling ratio: a useful tool in farming systems
analysis" in How Systems Work, Proceedings of the Farming Systems Research Symposium,
University of Arkansas, Fayetteville.

Basnyat, B. B. 1995. Nepal's agriculture, sustainability, and intervention: Looking for new directions
(PhD thesis). Wageningen Agricultural University, The Netherlands.

Bromley, D.W 1992. Making the commons work theory, practice and policy. California Institute for Contemporary
Press, San Francisco.

Carson, B. 1992. An agroecological zonation approach to agricultural planning in Mountain
Environments. Pp. 307-328 in Sustainable Mountain Agriculture, Jodha, N. S., M. Banskota
and T. Partap (eds.). Vol. I. ICIMOD, Kathmandu.

Chambers R. 1991. To make the flip: strategies for working with undervalued resources agriculture. ILEIA
Newsletter, pp. 5-9.

Chand, S. P, and B. Thapa. 1992. Mountain agricultural technology development and diffusion:
the Pakhribas model, Nepal. Pp. 737-760 in Sustainable Mountain Agriculture, Jodha, N. S.,
M. Banskota and T. Partap (eds.). Vol. II. ICIMOD, Kathmandu.

Devkota R. 1994. Green manuring as an alternative to the chemical fertilizers (in Nepali) Sankalan, 48:7-9.

Harrington L. W 1992. Measuring sustainability: issues and alternatives. In: Let Farmers Judge: Experiences
in Assessing the Sustainability of Agriculture. ILEIA Newsletter, pp. 2-16.

ILEIA 1995. ILEIA Newsletter. 11, 4: 2.

ILEIA 1996. ILEIA Nerosletter. 12, 3: 2.

Ingels C. and D. Campbell. 1995. What is sustainable agriculture? UC Sustainable Agriculture Research &
Education Program. University of California, California.

Jayawal, J. P, and K. J. Gurung. 1992. A significant contribution of nitrogen top dressing in increasing the
production of maize. Prabidhi Sangalo 27: 3- 5.

Jodha, N. S. 1993. Sustainable and more productive mountain agriculture: problems and prospects. Paper
presented at International Symposium on Mountain Environment and Development,
ICIMOD, Kathmandu, Nepal.

Kessler, J. J. and M. Moolhuijzen. 1994. Low external input sustainable agriculture: expectations
and realities. Netherlands Journal of Agricultural Science, 42, 3: 181-194.

Lightfoot, C., E T. Dalsgaard, M. P Bimbao and E Fermin, 1993. Farmer participatory procedures for
managing and monitoring sustainable farming systems. Journal of Asian Farming Systems Association.
2 (2): 67-87.

Mukherjee, N. 1993. Participatory Rural Appraisal: Methodology and Applications. Concept Publishing
Company, New Delhi.

Oldemann, R. A. A. 1990. Forests: elements of silvology. Springer Verlag, New York.

Oldemann, R. A. A. 1996. Diagnosis of complex ecosystem. Info Base. Easy Software. Wageningen, The

Vol. 7, No. 2,1997/2002


Ostrom, E. 1990. Governing the commons: the evolution of institutions for collective action. Cambridge University
Press, New York.

Pretty, J. N. 1995. Regenerating agriculture: policies and practice for sustainability and self-reliance. Earthscan
Publications Ltd., London.

Reijntjes, C. and M. Moolhuijzen. 1995. Searching for new methods. ILEIA Newsletter 11, 2.

Reijntjes, C., H. Bertus and W B. Ann. 1992. Farming for the future: An introduction to low external input and
sustainable agriculture. The Macmillan Press Ltd., London.

SARE, 1997. Exploring sustainable agriculture: ways to enhance profits, protect the environment and improve the
quality of life.

Save, B. and A. Sanghavi. 1992. Mr. Save's way of natural farming, ILEIA Newsletter, 4:22-24.

Sherchan, D. B. and B. D. Gurung. 1995. An integrated nutrient management system for sustaining soil fertility
research in the hills. Pp. 50-62 in Challenges in mountain resources management in Nepal, processes,
trends and dynamics in middle mountain watersheds, Schreier H., E B. Shah, S. Brown (eds.).
Proceedings of a workshop held in Kathmandu. 10-12 April. IDRC, Ottawa, Canada and Singapore.

Shivakoti, G., J. Dixon and A. Shukla. 1997. Resource dynamics and farming system economics in Nepalese
hills-- sustainability analysis: a study of Purkot village development committee of Tanahu District, Nepal. A
report submitted to the Farm Management and Production Economics Service, Agricultural Services
Division, FAO, Rome.

Shivakoti, G., G.H. Axinn, N.H. Axinn and N.A. Khan. 1977. The role of livestock in the farming system of
Sharada Nagar Panchayat in Chitwan District, Nepah a review. Journal of the Institute. of Agriculture and
Animal Science 1(1) 112-124.

Sriskandarajah, N., R. J. Bawdwn, and R. G. Packham. 1991. Systems agriculture: a paradigm for
sustainability. AFSRE Newsletter 2 (3):1-5.

Upreti, B.R. 1995. Women's participation in development activities: a case study of Mechi Hill Development
Programme in Ilam District. Unpublished M. A. Thesis submitted to the Department of Sociology/
Anthropology, Tribhuvan University, Kathmandu, Nepal.

Upreti, B. R. 1993. The Effect of MHDP vegetable seed production programme on women farmers
of Shantidanda, MHDP and National Development Service. Kathmandu, Nepal

Vaidya, A., C. Turton, K. D. Joshi and J. K. Tuladhar. 1995. A systems analysis of soil fertility issues
in the hills of Nepal: Implications for future research. Pp. 63-80 in Challenges in mountain
resources management in Nepal, processes, trends and dynamics in middle mountain watersheds,
Schreier H., RB. Shah, S. Brown (eds.). Proceedings of a workshop held in Kathmandu, 10-
12 April, ICIMOD, IDRC, UBC.

Wolfert J., E. A. Goewie, F de Jonge and J. D. Van der Ploeg. Ecological production and production ecology as
contributions for realisation of practicalforms of sustainable farming (in preparation).

Wolfert J., E H. de Jonge and E. A. Goewie. A theory for ecological farming (in preparation).

End Notes

1 The Village Development Council (VDC) is the lowest administrative unit in Nepal. There are nearly
4000 VDCs in Nepal, with the total area for the whole country being 141,000 square kilometers.

2 A ropani is a unit of measurement of land in Nepal that is equal to 0.05 hectare.

Journalfor Farming Systems Research-Extension

An Integrated Framework for Solving
Problems in Sustainable Agriculture

John Smithers, Ellen Wall and Clarence Swanton'


Sustainable agriculture, understood to mean production systems that are
environmentally benign (or enhancing), economically viable, and socially
acceptable, continues to be a concern for farm operators, policy-makers,
and the general public. This is especially true in North America where the
agricultural industry is subject to a growing number of regulatory and
consumer demands in the midst of pressures from global competition.
This paper demonstrates how approaches from farming systems research
can be used successfully to address issues in sustainability assessment.
Farming systems research is well suited to sustainability issues because
both require the inclusion of multiple perspectives, systematic methods,
and holistic understanding. The framework presented is focused on, and
driven by, applied problems in sustainable agriculture at the farm level of
the farm and the local agricultural system. A research protocol is described
which follows a progression through conceptual, collaborative, analytical,
and evaluative domains representing four dimensions of an adaptive
problem solving approach. The application and utility of the approach is
discussed in terms of on-going Farming Systems Research at the University
of Guelph in Ontario, Canada.

1. Introduction

There is now broad consensus that workable definitions of sustainable agriculture
must include physical, biological, and socioeconomic elements (Acton and Gregorich,
1995; Benbrook, 1991; Schaller, 1990; Smit and Smithers, 1994). However, there is less
agreement on just how to merge these in practice. Sustainable agriculture generally,
and farming systems research specifically, need practical tools to assess and solve
problems related to the sustainability of existing and emerging agricultural systems
-especially in an integrative manner (Flora, 1992).

Despite the knowledge gained in traditional disciplinary research about physical,
biological, social, and economic processes in agriculture, there is an increasingly
well-recognized interest among agricultural researchers and planners to understand
how these processes are interrelated in different farming systems, or how these
systems affect the environmental and social health of rural communities. The former
issue is one which has been the focus of work by agricultural systems analysts (e.g.

1 Farming Systems Research, University of Guelph, Guelph, Ontario

Vol 7, No. 2, 1997/2002


Dent et al., 1995; Doyle, 1990)-but not in the explicit context of sustainability. The
latter has been attempted by a number of analysts (e.g. Stockle et al., 1994; Smit et al.,
1997), but not in a dynamic systems approach. Few would disagree that the multi-
faceted and largely philosophical nature of sustainable agriculture must be defined
and operationalized in ways that are conducive to "systematic and systemic"
description and analysis (Bawden, 1995).

Among those who are disadvantaged by the lack of progress on moving
sustainability from the conceptual to the practical are the very farm operators whose
family and community livelihoods hang in uneasy balance. North American farmers
find themselves subject to growing criticism regarding the environmental damage
from their practices (Beus and Dunlap, 1990; Kelly, 1996; Tisdall, 1992) despite the
fact that in some regions there have been both significant reductions in the use of
pesticide, chemical fertilizers, and fossil fuels coupled with increased adoption of
conservation-oriented soil management practices (Surgeoner, 1996).

It is incumbent upon sustainable agriculture researchers and advocates to work
toward implementing a clear conceptualization of the term and then contribute to
a process for evaluation which might, in some instances, include the development
of acceptable thresholds or management goals in pursuit of environmental quality
and food safety. In that way agricultural analysts, farm operators, and their critics
can evaluate farm operations and determine whether or not progress has been made
(Johnson, 1996).

The purpose of this paper is to present a problem-solving framework for researching
and alleviating problems in sustainable agriculture. The approach and concepts
described in the paper derive from previous scholarship in a variety of research
and environmental management paradigms, and from recent deliberation and
scholarly debate amongst members of an interdisciplinary farming systems research
program (FSR) at the University of Guelph.' The proposed framework establishes a
sequential, systematic, and strategically iterative approach that parallels the general
farming systems research process where objectives are established as an initial step
followed by information gathering and evaluation (Petheram, 1986). However, as
Flora (1992) has noted, the conventional 'steps' of FSR/E need to be modified to
reorient research toward sustainability assessment and intervention as opposed
to the historically dominant purpose of increasing productivity. The four main
components of the tradition-diagnosis, design, on-farm trials, and extension-
have been integrated into the FSR framework emphasizing the logical, sequential
nature of the research and the potential for change-both planned and incidental
-in the biophysical and socioeconomic dimensions of farming systems.

In keeping with Baker's (1993) assertion that farming systems research should be
more attuned to agricultural policy issues, the integrated FSR Project is interested in
working not only with community-based stakeholders and with commodity groups,
but also with policy-makers at various jurisdictional levels. In North America for
instance, the state has maintained a high profile in pursuing sustainability issues
although its effectiveness in doing so has been questioned (Hall, 1997). Thus, there
are opportunities and needs for policy-relevant contributions from agricultural

Journalfor Farming Systems Research-Extension


research of the type described here. There have also been several initiatives designed
to promote more acceptable practices where the impetus has come from producers
and associated farm groups themselves via partnerships with public institutions.2
Here also, there are needs for contributions in the area of participatory research
design and methods of evaluation.

Finally, we note that our focus is on developed industrialized agricultural systems.
Hence, some of the caveats associated with traditional farming systems research are
averted. As noted by Berdegue (1993), farming systems research has most often been
concerned with small-scale, resource-poor farmers in marginal areas of developing
countries. There, problems with credit availability, weak markets, powerless
farmer organizations, and the inability to take risks constrain their ability to adopt
alternative practices. In North America, where the challenges to farming are of a
different order, farming systems research can still bring its strengths and insights
to bear on a number of different issues, including identifying the limitations to the
adoption of sustainable agricultural practices. But an appreciation of the nature
of environmental, agricultural, and social systems, and of the multiple criteria of
sustainability, are fundamental.

2. Frameworks for Sustainability Assessment and Evaluation

Among the core objectives of the FSR Project is assessment of specified elements
of farming systems, prediction of the implications of change both within systems
and beyond their boundaries, and evaluation of these conditions and possible
responses in the context of the goal of sustainability. Such work draws naturally on
a range of research traditions and approaches. A variety of general and conceptual
frameworks have been designed to allow analysts to assess sustainability in other
types of managed and unmanaged ecosystems. A few are noted here for illustrative
purposes. Examples of conceptual frameworks for assessment include the Pressure-
State-Response Model emphasizing system processes (Rapport and Friend, 1991),
a Compliance-Diagnosis-Warning typology emphasizing system conformity to
standards (Cairns et al., 1992), and criteria of Vigour-Organization-Resilience,
invoking certain ecological properties of systems (Costanza et al., 1992). More
recently, Smit et al. (1997), and Gallopin (1996) have attempted to characterize and
assess sustainability in agriculture in the concepts, language, and methods of health.
While the usefulness of this approach and its methodological contributions are still
unresolved, it does highlight the important role of information (indicators) for
purposes of evaluation, and the goal of strategic intervention to address diagnosed
threats to sustainability or health.

Working from the perspectives of agroecosystem analysis, some analysts have
employed ecological network modeling approaches to assess system performance
(Dalsgaard et al., 1995). Tools such as these are appropriate, and necessary, to assess
in an iterative fashion the implications of different types and combinations of farm
management practices on certain attributes of the farm system. To the extent that
component models are linked via a macroscopic model, such approaches also have
the ability to account for changes throughout a whole farming system.

Vol. 7, No. 2, 1997/2002


Elsewhere, Stockle et al. (1994) proposed a framework that accommodates the use of a
weighted index of farming system sustainability based on combined adjusted scores
pertaining to selected system attributes and perceived constraints to sustainability.
While the method does not account for linkages and interdependencies within
the system, it does provide a summative evaluation of the relative sustainability
of a farming system. Dumanski et al. (1992) employ the concept of "framework
as pathway" leading to a systematic approach for assessing the prospects of
sustainable land management systems. This framework is underpinned by certain
specified preconditions for sustainability and moves systematically through system
description, problem/constraint identification, and the application of diagnostic
criteria and performance indicators. This results in an assessment of the likelihood
of sustainability. A strength of this method is its explicit identification of a set of
criteria upon which the remainder of the framework is built and against which the
system is judged.

The present framework is distinctive for at least two reasons. One is the central
role played by the clear identification of a problem or set of problems located in a
specific region and/or for a particular farming community. As a result of this, the
conceptual (and arguably more abstract) aspects of sustainability form the context
rather than the focus of research efforts.

A second distinction is the overt use of the classic farming systems approach including
the designation of different "targeting" domains (Wotowiec and Hildebrand, 1988;
Moore, 1995). For the FSR framework, they range from the conceptual, where
elements of sustainability and farming systems are described and form the theoretical
rationale for evaluation, to the consultative where input from the farm community
is sought to determine problematic issues in sustainable agriculture. Based on the
specification of a particular issue, the analytic context follows ard involves the
choice of indicators, their measurement, and computation in appropriate models.
Subsequent to analysis is the evaluation domain where results are generated that
address the initial sustainability issue/problem and supply useful information for
farm decisions and policy-making. Clearly, each one of these domains requires
further articulation in its own right. For example, work in the analytic domain
involves the development and application of diagnostic indicators and formal
analytic structures for their manipulation and assessment. One potentially useful
strategy involves the development and testing of certain indicators for agreed upon
assessment endpoints such as quality of life, soil and/or water quality, economic
viability, etc., for which a wide variety of applications is anticipated. This preliminary
identification of indicators, and of appropriate integrating models is currently either
underway or in development (to the extent that it can be done generically) and will
be reported elsewhere. For this paper however, the focus is on the research approach
and on the selection and sequencing of components for effective consultation,
analysis, communication, and adjustment.

3. A Framework for Problem-Solving in Sustainable Agriculture

Details of a problem solving framework have been developed as part of the evolu-
tion of the broader interdisciplinary research initiative.3 Practical experience in

Journal for Farming Systems Research-Extension


a multi-disciplinary group, constrained by discipline-related perspectives and
institutional structures, created the need for a more systematic and systemic
approach to sustainable agriculture issues. The proposed framework (Figure 1)
uses a standard farming systems model that can be adapted to a particular case.
Employing a systems approach to farming encourages the kind of thinking that
is needed to understand sustainability problems, namely interpretations that are
holistic, hierarchical, and dynamic (Waltner-Toews, 1996).


Identify issue and
establish context.

problems, goals
and objectives.

Specify farm
Identify end points
and indicators.

Assess conditions

offer advice.


Conservation Financial Crop/Livestock

I.. Farming
'sity System -

Adaptability Natural Human/Social I

Problem Identification I
I goals hypotheses
objectives researchdesign

Aicators Subsystems and Parameters
! e Finance (income...)
sC Nat. Resource (soil...)
essto swvices Human (decisionmaking...)
Products (enterprise...)

Measure; Monitor; Model







Figure 1.A Framework for Problem Solving in Sustainable Farming Systems Research
and Extension

4. Conceptual Domain

In the first stage of the FSR project and research framework, basic definitions
and assumptions are sorted out and agreed upon to provide the rationale and
structure for understanding particular problems. The logic and effectiveness of the
assessment and problem-solving exercise depend on establishing the researchers'

Vol. 7, No. 2,1997/2002


ability to link specific details and findings into a broader set of concepts and ideas
that provide the means for communication with stakeholders and other interest
groups. The three elements highlighted at this stage concern the definition of
criteria for sustainability, a generalized understanding of farming systems, and
the identification and cose specification of relevant and researchable issues or

Defining Sustainable Agriculture
According to Harrington (1992) and others, sufficiently comprehensive definitions
of sustainable agriculture will include ecological, social, economic, and ethical dimensions.
Another approach to defining sustainable agriculture is to delineate a number of key
elements that capture the spirit of sustainability (or health) and in turn can be treated as both
sustainability criteria and/or goals. Based on reviews of ecosystem health or integrity (Kay and
Schneider, 1992; Rapport, 1995; Woodley et al., 1993), agroecosystem sustainability (Altieri
et al., 1983; Conway, 1987; Dalsgaard et al., 1995; Gallopin, 1995; Marten, 1988) and other
"sustainability literature (Bryden, 1994; WCED, 1987), four such system-level attributes
are suggested as forming the basic elements of sustainability: Conservation, Adaptability,
Capacity, and Productivity. While it would be possible to select other such properties (and in
different cultural and technological contexts others might be appropriate), this set captures a
useful set of integrating (between subsystems) themes which transcend subsystems and span
a variety of recently expressed threats and issues in sustainable agriculture.

The notion of Conservation accommodates issues of cycling, subsidy, and efficiency
in the use of resources which may be brought to bear in analysis of existing and
emerging farming systems. Concern for Adaptability acknowledges the potential
importance of system stability and diversity in understanding the impact of selected
stresses and the likelihood and effect of selected responses. The concept of Capacity
includes references to the system's various resources which define the opportunities
for coping and response, its resilience, and equity. Finally, Productivity recognizes
basic concerns for system outputs in terms of the satisfaction of expectations and
demands, both personal and social, and for future viability and the likelihood of a
continuance in farming under specified conditions.

As noted above, such properties are hardly unrelated-and debate over their
meaning and potential relevance invariably has them converging with each other,
and with the overriding notion of sustainability. Like definitions of sustainable
agriculture, these components are intentionally not expressed in precise (system/
subsystem specific) terms because they reflect values and goals that have general
applicability in society-the cultural and institutional context for that agricultural
system. When the terms are broad, they are also ideally suited to refer to both
the socioeconomic and non-human or biophysical elements of farming systems,
thereby enhancing integrated farming system analysis. For instance, productivity,
as a criterion for sustainability, can be applied not only to the services available
from natural ecosystems in the farming system but also to both the livestock and/or
cropping systems. The concept of productivity is also applicable to the human/
social resource dimension of farming systems in terms of how well individuals and
social groups involved in farming can fulfil their roles.

Journal for Farming Systems Research-Extension


Farming System Components and Interactions
Also significant at the conceptual level of the assessment framework is the
definition of "farming systems." For the most part, farming systems can be defined
in terms of an overall approach to farming, reflecting goals, abilities, resources,
and circumstances pertinent to the farm operation-all of which result in decisions
concerning the composition and operation of these human-made systems (e.g. Ikerd,
1993). While farming systems are largely defined by their main inputs and products
(cultivation, crop selection, pest management, land allocation, etc.), the focal point
for that system is often viewed as the farm operator as decision-maker.

A key feature of the internal components of any farming system is their diverse
nature, a quality that may be hidden in the versatility of farm operators but is
patently obvious when outside experts are called in to aid in analysis. Thus, any
research team investigating sustainable agriculture issues in particular farming
systems will have to include proficiency in the theory and methods of a number
of disciplines. For instance, the present FSR group at the University of Guelph has
representation from animal and crop science, agricultural economics, engineering,
ecology, geography, land resources science, rural extension, and sociology to foster
a comprehensive and inclusive approach to both the conceptualization of systems
and problems, and to their analysis.

A principal concern in sustainability assessment is to ascertain the basis for deci-
sions through an understanding of both the elements of the farm system (about
which decisions are made) and the social/cultural context that motivates and
directs farm-level decision-making. Further, in keeping with a focus on integration
and feedback within systems, there is (or should be) interest in the recursive nature
of these relationships, recognizing that management decisions have consequences at
both the farm scale and beyond. Farming systems are not simply a sum of discrete
components, they are complex and highly interactive with significant feedbacks.

Merging concerns about sustainability with farming system components has the
potential to generate a wide-ranging set of issues, many of which are inter-related.
Among the most abiding of agricultural sustainability-related issues in Ontario
and throughout North America is the desire to ensure the conservation of natural
features and resources while enhancing and improving the viability of farming
systems in the same region. Although these can be viewed as mutually exclusive, in
some areas progress is being made in achieving a satisfactory balance (Ontario Farm
Environmental Coalition, 1994).

Problem Identification
The identification of a particular problem or threat to sustainability focuses efforts
on dealing with the practicalities and key attributes of a specific difficulty. The
importance of such an exercise is well recognized in applied assessment research
where issues of social/political values and multiple interests apply. For example,
a defining feature of most environmental impact assessment frameworks is the
process of 'scoping,' where valued environmental components and associated issues
are defined and carried forward into the assessment process. The same logic holds
here. In North America, particular sustainable agriculture concerns are being raised

Vol. 7, No. 2,1997/2002


by a number of groups and individuals including the state, commodity groups,
farm operators and their families, representative farming organizations, consumer/
community groups, and members of the public. In most cases, sustainability issues
arise because there has been direct experience with a significant problem (e.g., poor
returns from crop and/or livestock production) or there is the potential for one to
develop (e.g., the loss of a valuable wetland).4

The range of sustainability-related issues affecting agriculture is well documented
and includes a host of social, economic and environmental factors and processes.
Among the more commonly voiced concerns in the Southern Ontario agricultural
system is the potentially negative consequences for human and environmental
health from excess nitrogen and phosphorous as well as other pollutants, in ground
water (Goss and Barry, 1995). FSR at the University of Guelph has been directly
involved in several initiatives aimed at addressing these problems. An example
from recent consultations with resource management agencies and agricultural
sector partners serves to illustrate both the opportunity and the protocol for issue-
driven research collaboration.

Current concern for water quality in a local, largely agricultural, subwatershed has
provided an impetus for discussions on a potential collaboration between FSR, a
local resource management agency, the local farming community, and a downstream
urban municipality to improve water quality in the region. At this point, hypotheses
and research designs can be developed in collaboration with interested parties and
in the context of their goals and objectives. Waltner-Toews and Wall (1997) note that
for agricultural sustainability (or health) issues, it is useful to recognize that two
kinds of goals are at work: primary and operative. The former refer to those ideal
aspirations that act as overall long-term objectives while operative goals are sub-
ordinate and considered the means to achieving primary goals. With respect to the
FSR example, the primary goal is to improve water quality in the watershed while
the operative goals are more specific and include (among other things): correcting
manure management practices; modifying fertilizer handling; and protecting
riparian areas.

Distinguishing goals in this manner acknowledges issues of scale and hierarchy, a
distinguishing feature of farming systems research. Primary goals are clearly at a
broader scale than the operative ones; the former encompass the latter. Specifying
goals is also useful for accommodating the potential tension between "science and
farmer participation," recently articulated by Caldwell and Christian (1996). In the
southern Ontario example, the primary aim of improving water quality is one to
which all interests can readily agree. When the focus moves down the scale to more
specific operative goals and/or objectives, a logical division arises with certain farm
operators (e.g., livestock producers) and certain farming system experts (e.g., land
resource scientists) taking on specific projects motivated by particular objectives.
Although conflicts might arise at this level between the interests of researchers and
farm operators (in that both may resent and resist changes suggested by the other)
it is possible for such differences to be resolved in light of their common attachment
to the primary goal.

Journal for Farming Systems Research-Extension


5. Analytical Domain

When the problemss' has been clearly identified and communicated, and the goals,
hypotheses and research design have been determined, the next stage of the research,
and of the framework, involves more formal integrated analysis. Here the concepts,
theories and methods of the natural, agricultural, and social sciences are brought to
bear on how well defined issues and tasks, and on a conceptual understanding of
the farming system. It is at this stage that effective interdisciplinarity is most critical
and its challenges most visible. The key challenge is to determine those components
of the farm system that are salient for the task at hand, and to select and apply
paradigms and methods that are appropriate for that (usually multidimensional)
problem. The former effectively represents a scientific 'scoping' of the research task.
The latter demands an appreciation of the potential contributions and limitations of
methods in the natural and social sciences. The desired end is an integrated research
strategy in which the environmental, economic, and social dimensions of sustainable
agriculture problems are considered together. The specific mix of approaches
(environmental systems modeling for purposes of prediction, behavioral research
to explain farm decision-making, statistical analysis and validation of proposed
indicators, etc.) and the extent to which they are fully (and physically) linked or
operate in a parallel and complementary fashion will vary in each application.

Specifying the Farming System
Farming systems are comprised of sub-systems at smaller scales of aggregation,
and exist as part of larger systems in a hierarchical fashion (Allen and Starr, 1982;
Izacs and Swift, 1994; Weber, 1996). According to Bird et al. (1984), the first step in
a systems approach is to identify the boundaries (and the elements) of the system
under consideration, which leads to defining those components of the farming
system which are internal to this realm and those which lie outside it. The former
include crop/livestock systems, farm economics, human interactions, and on-farm
ecosystems which can be examined in terms of how each one influences and is
influenced by the others. Here data are needed to identify and assess important
flows, relationships and connections. By contrast, systems pictured as external to
the farming system refer to such elements as markets, policy, climate, and natural
resources, all of which are of particular interest for sustainability issues.

The designation of a particular problem will implicitly highlight one of the
subsystems within the farming system. The southern Ontario example cited above
would begin with a focus on crop and/or livestock subsystems. However, the
selection of key initial subsystems does not suggest that the research is artificially
limited to these aspects of the farming system alone. Quite the opposite. Other
subsystems of the farm and the local agricultural system are dealt with in time and
in terms of their relationship to aspects of the initially relevant subsystems and the
broader sustainability issue. Thus, if and when barriers to adoption of better manure
management are studied, conditions in both the financial and human/social aspects
of the farming system (as well as a number of external factors) would become

Vol. 7, No. 2, 1997/2002


Choosing Measurement Endpoints and Indicators
Essential to the sustainability analysis is the selection of appropriate indicators
which will depend on a number of factors including the designated problem, scale,
type of farming system (e.g., livestock, cash crop, horticulture, etc.) and specific
component of farming system implicated in the sustainability issue. Thus it may be
that attention is directed at all the farming systems in a region where the problem
identified concerns soil erosion; it may be all livestock operations in the case of
concern over nitrate loading in a damaged watershed; or it may be one highly
diversified industrial farming system that is perceived as problematic because it is
threatening the sustainability of neighboring smaller farming systems.

Also important for indicator selection are the significant endpoints chosen; they
will have a direct connection to goals and objectives that are driving the research
(Cairns et al., 1992). In the case of the southern Ontario watershed, for instance, a
target parts per million measure of nitrogen and phosphorous will be decided on
as a desirable endpoint for the watershed. Various changes in farming practices can
then be monitored with respect to how they affect the N and P targets.

Regardless of which endpoints are chosen, all significant variables need to be
operationalized into a number of measurable indicators. Because there has been
significant interest in the development and testing of agri-environmental indicators
of sustainability and health (e.g., Agriculture Canada, 1993; Smit et al., 1997;
Benbrook and Mallinckrodt, 1994; Eswaran et al., 1993), there is a wide array of
choice for sustainability analysts.5 For instance, soil quality can be reflected with
specific values for nutrient balances, waterholding capacity, or percentage of organic
matter. Likewise, quality of life indicators include both "objective" measures such as
net income and "subjective" assessments related to various perceptions regarding
the availability of services, conditions in farming and,the rural community, and
future outlook (Filson, 1997).

Empirical analysis has also documented ranges for a number of measures that will
affect soil and water quality, agricultural economic viability, and human physical
and social health. What science cannot do, however, is dictate the acceptability of
these impacts. Such decisions are related to the values belonging to those who enlist
the help of farming systems research for dealing with specific sustainability issues.

Measuring, Monitoring, Modeling
A fifth step in the problem solving framework is directed at assessing conditions
and determining outcomes based on the nature of data gathered and approach
selected for its manipulation and analysis. Opportunities for explanation and
prediction arise from the use of rigorous analytical methods based on inferential
statistical analyses, simulations, and optimization techniques. Of special interest are
mathematical modeling techniques for the analysis of agricultural systems. At this
analytical stage individual researchers, working within accepted methodologies
appropriate to their disciplinary interests and roles, generate findings specific to
selected issues (e.g. developing and testing indicators of quality of life), but also
contribute information to a more integrated venture. Current work is focused on
the development of functional optimization and simulation frameworks where

Journalfor Farming Systems Research-Extension


trade-offs and interactions among the components can be assessed systematically
and iteratively (Yiridoe and Weersink, 1997). The use of such models for addressing
sustainable agriculture issues overcomes some of the limitations that are associated
with farming systems research and extension work. As it has been practiced for
most of its history, farming systems analysis has precluded opportunities for long
(or even medium) term experimentation as well as for transferring findings to
conditions in other locations (Harrington, 1992). With the adoption of modeling
techniques, it is possible to extend both the temporal and a spatial component in the
analysis thereby increasing the potential information available for evaluation.

6. Evaluation and Decision-making

Integrated farming systems research represents an example of what Waltner-Toews
et al. (1997) have termed 'mandated science.' The clear intent is to offer information
and advice as a basis for improved, or at least informed, decision-making at the farm
scale and beyond. Hence, from the analytic exercise described above, there is need
for the broader question "so what for sustainability?". What are the implications
of findings from an integrated analysis for the problem initially described (recall
the water quality example earlier) and, just as importantly, for the sustainability of
regional farming systems?

So while Figure 1 suggests a terminal point in the proposed problem-solving
framework devoted to discussion of alternative actions and solutions, it is in many
ways yet another beginning. Recognition of the interconnectedness of systems begs
further questions. Solution options for one issue must be considered relative to
their implications elsewhere in the system. In effect, there are not only responses
to issues, but also potential issues from responses. Systems-oriented research must
accommodate not only first order outcomes associated with certain courses of
actions, but also the recursive nature of systems. Bounding of the research scope can
and should occur in consultation with sector partners, but it should be underpinned
by a commitment to understand the systemic aspects of farm-level sustainability
and associated actions to that end. Adequate assessment of the prospects for
sustainability in farming systems requires a system-level perspective. While
the approach outlined here is driven by locally (and perhaps narrowly) defined
concerns which often focus on selected issues and subsystems, concern for the
sustainability of whole farm systems, and of farming regions demands appreciation
of interactions within and between those systems, and of the implications of change
-socially, economically and environmentally.

7. Conclusions

This review of the sustainability problem-solving framework being used by the FSR
team at the University of Guelph has addressed important issues in both farming
systems analysis and sustainable agriculture. The framework, which is grounded
on the premise that North American farm operators will only be able to adopt
more sustainable practices when they have clear standards to gauge their success

Vol. 7, No. 2, 1997/2002


or failure, has been presented as an integrated and practical method focused on
solving problems in sustainable agriculture.

The framework is integrated in several ways. First and in the most general terms,
it draws together traditional farming systems methods and sustainable agriculture
issues. Once a specific problem has been articulated, the system functioning can be
analyzed and results implemented in improved techniques. A second integration
feature involves the multi-disciplinary approach that the FSR framework embodies.
Although selected analyses may be confined to disciplinary fields, they become
synthesized when solutions are generated in a latter phase and in the context of a
primary goal. This process requires integration across disciplines and between the
realms of public interest and scientific expertise. In the FSR framework neither can
operate effectively without the other. Equally possible, and highly desirable, is the
development of more formally integrated models which capture the linkages and
interactions between environmental, economic and social dimensions of farming
systems and facilitate whole system level prediction. The development of such tools
is a needed component and a valuable byproduct of systems evaluation research.

Sustainability has not been, is not, and never will be afait accompli. As values, goals,
and objectives evolve within the sociocultural context that informs agricultural
practices, so too will the identification of problems and the solutions associated
with them. The research framework described in this paper offers a structure that
accommodates and incorporates different needs, perspectives and uses, and has
been presented here as a viable and practical approach in applied farming systems
research in support of sustainable agricultural systems.


The authors appreciate the valuable input from other members of the Farming
Systems Research Team at the University of Guelph including Drs. M. Miller, M.
Goss, J. Ogilvie, J. Gibson, T. Hunt, A. Weersink, and G. Filson, as well as the helpful
advice of Dr. P Hildebrand, University of Florida. We also gratefully acknowledge
the Ontario Ministry of Agriculture, Food, and Rural Affairs and the Ontario
Agricultural College for their financial support of the Integrated Farming Systems
Research Project.

1. The University of Guelph has initiated a project in Integrated Farming
Systems Research and Extension (FSR/E). The broad research goal of this
inter-disciplinary project is to develop and test an approach for assessing the
sustainability of Ontario's farming systems. An initial objective of the project
is the development of an analytic, indicator-based, framework for assessing
the sustainability of current farming systems and for exploring the impacts
of changes and/or stresses on those systems. The desired outcome is an
improved understanding of the prospects for, and barriers to, sustainability in
various farming systems and the delivery of assessment tools which will be of
use to researchers, extension personnel, farmer operators, and policy-makers

Journal for Farming Systems Research-Extension


in improving Ontario's farming systems. For more information on FSR at
University of Guelph, refer to our web page: http://tdg.uoguelph.ca/www/

2. One of the most significant is a joint government and industry effort called the
Ontario Environmental Farm Plan (Ontario Farm Environmental Coalition,
1993) which has been reviewed and adopted by approximately 17% of the
provincial farming community. By the summer of 1997, roughly 11,000 farmers
in Ontario have gone through the farm plan and modified their operations on
the basis of 23 criteria pertaining to measures for everything from improving
soil and water quality to enhancing natural environment of the farm property.
Over half of these farmers have taken the further step of submitting their
changes to a peer review process and thereby receiving official documentation
regarding the environmental aspects of their operations.

3. The FSR Project appreciates the contribution Peter Hildebrand made in this
regard during his visit with the group.

4. Another approach to specifying particular sustainability problems is to
rephrase them in terms of constraints to sustainability (Stockle et al., 1994).
Thus, in the example cited earlier, the concern with poor returns from crop
or livestock production can be considered in light of the conditions that
are leading to poor productivity. Adopting this perspective emphasizes the
need to address underlying or root causes of problems and encourages those
involved not to accept short term or superficial solutions.

5. Earlier indicator efforts could be described as relatively conceptual involving a
search for particular indicators in light of general sustainability problems. The
integrated framework presented here somewhat reverses the emphasis and
suggests instead that the more productive search is for particular sustainability
problems that can be addressed with generally accepted indicators.


Acton, D. and L. Gregorich (eds) 1995. The Health of our Soils: Toward Sustainable Agriculture in Canada. Centre
for Land and Biological Resources Research, Agriculture and Agri-Food Canada, Ottawa.
Agriculture Canada. 1993. Developing Environmental Indicators for Agriculture. Agriculture Canada
Environmental Indicator Working Group Discussion Paper, Ottawa, June 1993.
Allen, T. and T. Starr. 1982. Hierarchy Perspectives for Ecological Complexity, University of Chicago Press,

Altieri, M.A., K. Letourneau and J. Davis. 1983. Developing Sustainable Agroecosystems. BioScience
33(1): 45-49.

Baker, D. 1993. Inability of Farming System Research to Deal With Agricultural Policy. Journal of Farming
Systems Research/Extension. 4 (1): 67-86.
Bawden, R. 1995. On the Systems Dimension in FSR. Journal of Farming Systems Research/Extension. 5 (2):

Vol 7, No. 2,1997/2002


Benbrook, C. 1991. Introduction. Pp. 1-12 in: National Research Council, Sustainable Agriculture Research and
Education in the Field. National Academy Press, Washington, D.C.

Benbrook, C. and E Mallinckrodt. 1994. Indicators of Sustainability in the Food and Fiber Sector. A paper
prepared for the SARD FORUM.

Berdegud, J. 1993. Challenges of Farming Systems Research and Extension. Journal of Farming Systems
Research/Extension. 4(1): 1-10.

Beus, C. and R. Dunlap. 1990. Conventional versus Alternative Agriculture: The Paradigmatic Roots of the
Debate. Rural Sociology. 55(4): 590-616.

Bird, E., T. Edens, E Drummond, and E. Groden. 1984. Design of Pest Management Systems for Sustainable
Agriculture. In C.A. Francis et al. (eds.) Sustainable Agriculture in Temperate Zones. Wiley, New York.

Bryden, J. (ed.) 1994. Towards Sustainable Rural Communities. University of Guelph, Guelph, Ontario.

Caldwell, J. and A. Christian. 1996. Reductionism, Systems Approaches, and Farmer Participation: Conflicts
and Contributions in the North American Land Grant System. Journal for Farming Systems Research-
Extension. 6(2):33-44.

Cairns, J. P McCormick, and B. Niederlehner. 1992. A Proposed Framework for Developing Indicators of
Ecosystem Health. Hydrobiologia. 263:1-44.

Conway, G.R. 1987. The Properties of Egroecosystems. Agricultural Systems 24: 95-117.

Costanza, R., B. Norton, and B. Haskell (eds) 1992. Ecosystem Health. Island Press, Washington, D.C.

Dalsgaard, J., C. Lightfoot, and V Christensen 1995 Towards Quantification of Ecological Sustainability in
Farming Systems Analysis. Ecological Engineering, 4:181-189.

Dent, B., G. Edwards-Jones, and M.J. McGregor. 1995. Simulation of Ecological, Social and Economic Factors
in Agricultural Systems. Agricultural Systems, 49:337-351.

Doyle, C. J. 1990. Application of Systems Theory to Farm Planning and Control: Modeling Resource
Allocation. Pp.89-112 in: Jones, J.G.W and Street, PR. (eds.), Systems Theory Applied to Agriculture and
the Food Chain. Elsevier, London.

Dumanski, J. H. Eswaran, and M. Latham. 1992. A proposal for an international framework for evaluating
sustainable land management. Pp. 25-45 in Dumanski, J. et al. (eds.) Evaluation for sustainable land
management in the developing world. IBSRAM, Bangkok.

Eswaran, H., E. Pushparajah, and C. Ofori. 1993. Indicators and Their Utilization in a Framework for
Evaluation of Sustainable Land Management. Paper from SARD FORUM.

Filson, G. 1997. A Farming Systems Approach to Testing Indicators of Dairy Farmers' Quality of Life and
Environmental Attitudes and Behaviour. FSR Research Project at the University of Guelph, Guelph,

Flora, C. 1992. Building Sustainable Agriculture: A New Application of Farming Systems Research and
Extension. Journal of Sustainable Agriculture 2(3):37-50.

Gallopin, G. 1995. The Potential of Agroecosystem Health as a Guiding Concept for Agricultural Research.
Ecosystem Health 1(3): 129-140.

Goss, M.J. and D.A. Barry, 1995. Ground Water Quality: Responsible Agriculture and Public Perceptions.
Journal ofAgricultural and Environmental Ethics. 8:52-64.

Hall, A. 1997. Sustainable Agriculture and Neoliberalism. Paper presented at the Annual Meeting for Rural
Sociological Society, August, Toronto, Ontario.

Journal for Farming Systems Research-Extension


Harrington, M. 1992. Measuring Sustainability. Journal of Farming Systems Research/Extension. 3(1): 1-20.

Ikerd, J.E. 1993. The Need for a Systems Approach to Sustainable Agriculture. Agriculture, Ecosystems and
Environment. 46:147-160.

Izacs, A-M.N. and M.J. Swift, 1994. On Agricultural Sustainability and its Measurement in Small-scale
Farming in Sub-Saharan Africa. Ecological Economics, 11:105-125.

Johnson, P 1996. Keynote Address. Great Lakes Agriculture Summit, Kellogg Centre for Continuing
Education, Michigan State University, April 23-24.

Kay, J. and E. Schneider. 1992. Thermodynamics and Measures of Ecological Integrity. In: Ecological indicators:
problems and approaches. S.A. Lewin, M.A.Harwell, J.R. Kelly, and K.D. Kimball (eds). Springer-Verlag,
New York.

Kelly, K. 1996. Keynote Address. Great Lakes Agriculture Summit, Kellogg Centre for Continuing Education,
Michigan State University, April 23-24.

Marten, G.G. 1988. Productivity, Stability, Sustainability, Equitability and Autonomy as Properties for
Agroecosystem Assessment. Agricultural Systems. 26:291-316.

Moore, K. 1995 The Conceptual Basis for Targeting Farming systems: Domain, Zones, and Typologies.
Journal of Farming systems Research/Extension, 5(2): 19-38.

Ontario Farm Environmental Coalition (OFEC). 1994. Ontario Environmental Farm Plan.

Petheram, R. J. 1986. Farming Systems Research at BPT-Some Progress and Constraints. Pp. 32-46 in: J.
Levine and M. Sabrani (eds.), Proceedings of the Workshop on Farming Systems Research and Development.
Central Research Institute for Animal Sciences Ciawi-Bogor, March 1985.

Rapport, D. 1995. Ecosystem Services and Management Options as Blanket Indicators of Ecosystem Health.
Journal ofAquatic Ecosystem Health. 4:97-105.

Rapport, D. and A. Friend 1979. Towards a Comprehensive Framework for Environmental Statistics: A
Stress-response Approach. Cat.No.11-510, Statistics Canada, Ottawa.

Schaller, N. 1990. Mainstreaming Low-input Agriculture. Journal of Soil and Water Conservation 45:9-23.

Smit, B. and J. Smithers 1994. Sustainable Agriculture: Interpretations, Analyses and Prospects. Canadian
Journal of Regional Science.

Smit, B., D. Waltner-Toews, D. Rapport, E. Wall, G. Wichert, and E. Gwyn. 1997. Agroecosystem Health:
Analysis and Assessment. University of Guelph, Guelph, Ontario.

Stockle, C., R. Papendick, K. Saxton, G. Campbell, and E van Evert. 1994. A Framework for Evaluating the
Sustainability of Agricultural Production Systems. Journal of Sustainable Agriculture.

Surgeoner, G. 1996 Keynote Presentation, Great Lakes Agriculture Summit, Kellogg Centre for Continuing
Education, Michigan State University, April 23-24.

Tisdall, P, 1992. Approaches to Sustainable Agriculture: Seven Case Studies. Discussion Paper from the Science
Council of Canada, Ministry of Supply and Services, Ottawa.

Waltner-Toews, D. 1996. Ecosystem Health-a Framework for Implementing Sustainability in Agriculture.
Bioscience 46(9): 686-689.

Waltner-Toews, D. and E. Wall. 1997. Emerging Perplexity: In Search of Post-Normal Questions for
Agroecosystem Health. Social Science and Medicine (forthcoming).

WCED 1987. Our Common Future. Oxford University Press, New York.

Vol. 7, No. 2, 1997/2002


Weber, G. 1996. Heterogeneity and Complexity in Farming systems: Towards an Evolutionary Perspective.
Journal for Farming Systems Research-Extension. 6(2):15-32.

Woodley, S. J. Kay, and G. Francis (eds). 1993. Ecological Integrity and the Management of Ecosystems. St. Lucie
Press, Delray Beach, Florida.

Wotowiec, P and P Hildebrand. 1988. Research, .Recommendation and Diffusion Domains: A Farming
Systems Approach to Targeting. In S.V Poats, M. Schmink and A. Spring (eds.) Gender Issues in Farming
Systems Research and Extension. Westview Press. Boulder

Yiridoe, E. And A. Weersink, A. 1997. A Review and Evaluation of Agroecosystem Health Analysis: The Role
of Economics. Agricultural Systems. 55(4):601-626.

Journal for Farming Systems Research-Extension

Identifying Target Groups for On-farm Research:
Characterizing Farmers for Soil Fertility Maintenance
Research in Semi-arid Areas of Eastern Kenya

H. Ade Freeman', John M. Omiti' and Patrick A. Audi2


Classification of farmers into meaningful target groups is a key step in
diagnosing farmer priority problems as well as identifying opportunities
for conducting research with farmers. This study uses logit and cluster
analysis to identify target groups of farmers for on-farm research based
on variables representing their capacity to invest in alternative soil
fertility maintenance technologies in a semi-arid area of eastern Kenya.
The characteristics of farmers in each cluster suggested a range of soil
fertility maintenance options for experimentation with farmers. The
study demonstrates the practical utility of using an objective classification
to assist decision-making in setting priorities for on-farm research and
identifying opportunities that exploit the diversity of farmers' situation
and farming systems.

1. Introduction

Depletion of soil fertility on smallholder farms is a widespread and critical
constraint to food production and farm livelihood in many areas of Sub-Saharan
Africa (Sanchez et al., 1996; Kumwenda et al., 1996; IFPRI, 1995). Estimates of soil
nutrient balances are negative for key nutrients in several locations in sub-Saharan
Africa (Smaling et al., 1997). This problem is most intense in semi-arid areas of East
Africa where soils have low inherent fertility and other physical and soil chemical
characteristics limit crop growth. Farmers in semi-arid areas of eastern Kenya
cited declining soil fertility, water availability, and pests and diseases as the most
important constraints to increased food production (KARI 1996a, 1996b). Studies
in this area indicate that many farmers have adopted improved maize varieties but
crop yields are below their productive potential (Muhammad and Parton, 1992;
Kimotho et al., n.d.). For example, even in seasons when water availability is not a
limiting constraint to crop growth, farmers typically harvest less than 20% of maize
yields that could be obtained when the appropriate type and quantity of inorganic
fertilizer is applied (KARI, 1995). This suggests that adoption of improved crop
varieties will not contribute much to improving crop productivity in semi-arid areas
of eastern Kenya unless the soil fertility constraint is resolved.

1 ICRISAT, R O. Box 39063, Nairobi, Kenya
2 NDFRC-Katumani, P O. Box 340, Machakos, Kenya

Vol. 7, No. 2,1997/2002


In areas where livestock is an important component of the farming system, farmers
traditionally use animal manure to maintain soil fertility. However, the quantity
of animal manure available on-farm is declining due to reductions in livestock
holdings attributable to periodic droughts, degrading pastures and, population
pressure on land (Tiffen et al., 1994). The limited quantity of manure available is
also manifested in thin manure markets that impede widespread trade in the input
(Omiti et al., 1998). In addition to the limited quantities available for crop growth,
extant manure storage and management practices lead to substantial nutrient losses
that reduce the quality of manure (Probert et al., 1995).

Inorganic fertilizers have been an important component of soil fertility maintenance
strategies in both developed and developing countries where there has been
substantial crop productivity increases (Sanders and Vitale, 1996; Sanchez et al.,
1996). However, its use by smallholder farmers in semi-arid areas has remained
low because inadequate and irregular rainfall leads to production risks and highly
variable returns (Probert et al., 1994). Such reasoning implies that inorganic fertilizer
cannot be an important component of smallholder farmers' fertility management
strategies in semi-arid areas. Nonetheless, recent research findings from semi-
arid areas of eastern Kenya provide evidence of increasing farmer adoption of
inorganic fertilizer particularly after fertilizer markets were liberalized in the early
1990s (Murithi and Shiluli, 1993; Freeman and Kaguongo, 1998). Fertilizer market
liberalization has resulted in a sharp growth in the number of private fertilizer
retail traders that is associated with a trend toward greater availability, smaller
more affordable fertilizer packages, and greater use of fertilizer. The willingness of
smallholder farmers to invest in inorganic fertilizer in semi-arid areas, even though
in small quantities, suggests that the use of appropriate types of inorganic fertilizer
can be a component of a sustainable soil fertility maintenance strategy for many

A practical approach to develop soil fertility technology options which farmers are
more likely to use is to augment their existing soil fertility maintenance practices
with small quantities of inorganic fertilizer (CARMASAK, 1996). The integrated
use of organic and inorganic sources of nutrients, hereafter referred to as fertilizer-
augmented soil enrichment, FASE, aims to maximize the effectiveness of current
fertility maintenance strategies with the objective of improving the productivity and
sustainability of the system (McCown and Keating, 1992).

As with most other technological innovations, FASE is not an optimal soil
fertility maintenance strategy for every farmer because they differ in their
socio-economic and natural circumstances. An important step in the process to
assess the contribution of FASE technologies to maintenance of soil fertility is
the characterization of farmers into homogeneous or target groups for on-farm
research. The concept of homogeneous farmer groups occupies a major part
of applied research on technology generation for smallholder farming systems
in developing countries. In earlier Farming Systems Research and Extension
literature, the recommendation domain was used to classify farmers into relatively
homogeneous groups with similar practices and circumstances and for whom it is
likely the same recommendations or technologies will be appropriate (CIMMYT,

Journal for Farming Systems Research-Extension


1988). Though similar in many respects, the concept of target groups as used here is
quite different from earlier use of recommendation domains which was primarily an
approach for targeting location specific technologies (Hilderbrand, 1986). The target
groups defined for soil fertility maintenance research in this study recognize the
diversity in farmers' priority production problems and potential solutions to these
problems. Farmers are classified into relatively homogeneous groups derived from
a matrix of current fertility management practices and farming system variables.
The resulting cells within the matrix becomes the basis for experimenting with a
"menu" or "basket" of technology options that farmers can use to improve soil
fertility and crop productivity (Waddington et al., 1998). This framework provides
a decision-based approach to identifying target groups for on-farm research. It is
more useful than the recommendation domain approach because it expands the
opportunity set farmers face, allowing them to experiment with a range of options
that are consistent with their socio-economic and bio-physical circumstances as well
as their risk preferences and investment priorities.

This study describes a procedure for identifying target groups for on-farm research
on soil fertility maintenance strategies in which inorganic fertilizer is an important
component in a clearly identified agro-ecological zone in eastern Kenya. This
implies that decisions relating to the choice of agro-ecological zones have been made
and the focus is on socio-economic variables that are likely to influence household
investment decisions in a fertility maintenance strategy involving inorganic fertilizer.
Investment in inorganic fertilizer competes for limited household resources such
as cash, labor, and time. In making investment decisions, households therefore
compare the opportunity cost of resources invested in inorganic fertilizer with those
in alternative investments such as in the livestock sector and in off-farm activities.
The opportunity cost of investing in fertilizer is determined, among other things, by
household specific incentives and its capacity to make the investment. These two
factors, though related, might be influenced by different variables. For example,
household specific incentives include the actual and perceived level of returns to
fertilizer, relative returns and risks to alternative investments in the farm and non-
farm sector, and socio-economic characteristics. In contrast, the capacity to invest is
a function of the household asset portfolio and the cash flow resulting from them.

2. Data

The study was conducted in semi-arid eastern Kenya. This area has a bi-modal
rainfall distribution with average annual rainfall of between 600 and 800 mm. The
production system is characterized as a mixed crop-livestock system. The major
food crops are maize, beans, pigeonpeas, and cowpeas grown mostly as a cereal/
legume intercrop. Cultivation of cotton, sunflower, fruit trees, and sale of livestock
and livestock products provide the main sources of cash income. The major sources
of soil nutrients are from animal manure and inorganic fertilizer, with many farmers
using combinations of manure and inorganic fertilizer (as will be seen in Table 4 to
follow in the results). In these areas manure and inorganic fertilizer are often treated
as substitute rather than complementary inputs.

Vol. 7, No. 2, 1997/2002


Participatory rural appraisals conducted in this area provided background information
on farmers' soil and water management practices as well as their perception of
production constraints and opportunities for solution of the identified constraints
(KARI, 1996a; KARI, 1996b). The participatory appraisals were supplemented by a
survey of 100 randomly selected households in the same area. Data on household and
farm characteristics, resource endowments, crop production, input use, sources and
disposal of income, current and potential soil and water management practices and
crop production constraints were collected in 5 villages in Kionyweni sub-location of
Machakos district by structured questionnaires in September 1996.

3. Analytical framework for characterizing farmers

Although the importance of classifying farmers into recommendation domains is
well documented in the applied research literature (Franzel, 1981; Harrington and
Tripp, 1984; Williams, 1994) procedures for delineating recommendation domains are
less developed. In many cases recommendation domains have been delineated using
variations in household resource use, production or marketing goals, and cropping
systems and/or agro-ecological conditions. While such approaches are interesting,
their practical utility in applied research is limited by the fact that the classification
scheme selected might not relate to farmers' priority production problems.

A fundamental characteristic of smallholder agriculture in the study area is the
complex behavioral patterns of farm households. For the most part, these systems
are characterized by diversity in farmers' socio-economic and natural circumstances
and varying degrees of engagement in product and input markets with multiple
constraints. These complexities imply that methods used in classifying farmers
for on-farm research should be robust enough to gain a better understanding
of the likely fertility management decisions farmers are making. The paper uses
multivariate statistical techniques to characterize farmers into target groups for
on-farm research on maintenance of soil fertility with the objective of improving
the productivity of the system. This approach to classifying farmers for applied
research is more useful than simple descriptive or bi-variate approaches because it
provides an objective method for handling multiple relationships that characterize
smallholder farming systems. Moreover, the availability of computer technology
and appropriate software programs has made it possible to easily analyze survey
data that otherwise would seem complex or intractable.

A two-stage approach is used to classify farmers. In the first stage a discrete choice
statistical model, logit analysis is used to reduce the initial variables on farmers'
circumstances into a smaller number of variables that predicts the probability
that fertilizer will be a component of farmers' fertility maintenance strategy. This
procedure provides a reduced set of variables that are used as delineating variables
to characterize farmers according to their potential to invest in maintenance of soil

Cluster analysis is used in the second stage to identify homogeneous or target
groups of farmers for on-farm research. These groups reflect variations across farms

Journal for Farming Systems Research-Extension


and farmers' access to productive resources that are likely to influence their ability
to invest in maintenance of soil fertility. The subsequent classification is such that
each farmer is very similar to other farmers in the cluster with respect to their ability
to invest in a soil fertility maintenance strategy. Thus farmers are homogeneous
within clusters but heterogeneous between clusters (Hair et al., 1992).

The dependent variable in the logit regression is a binary variable that takes a value
of 1 if a farmer incorporated fertilizer in current soil fertility maintenance and 0
otherwise. The explanatory variables shown in Table 1 comprised both continuous
and binary variables representing household and farm characteristics, resource
endowment, production, marketing, and income variables. K-means cluster
algorithm was used to group farmers with similar characteristics into clusters (SPSS,
1977). The data were standardized by transforming each variable into its z-score
with 0 mean and standard deviation of 1 (SPSS, 1977). This procedure eliminated
any spurious effects that might result from unequal variances of variables included
in the analysis (Hair et al., 1992).

Table 1: Description of variables in the Logit regression
Variable Type Description
AGE Continuous Age of household head in years
GENDER Binary Sex of farmer: 1 if male, 2 otherwise
CA RATIO Continuous Ratio of total children to total adults
ED1 Binary 1 if farmer has formal education, 0 otherwise

AG. TRAIN Binary Farmers participation in agricultural training

FARMSIZE Continuous Farm size in acres
NO F LAB Continuous Number of family members working on the farm
MORGAN Binary Member of organization
C_CROP Continuous 1 if farmer grows cash crop, 0 otherwise
GRAINSAL Binary 1 if regular produce of grain sale, 0 otherwise
MILKSALE Binary 1 if farmer sells milk regularly, 0 otherwise
LIV.SALE Binary 1 if farmer sells livestock, 0 otherwise
OFFEEMP Binary 1 if farmer is engaged in off-farm activities,
0 otherwise
OFFFEOTH Binary 1 if members of family have off-farm development, 0
V1 Binary 1 if wandue or kiaani 0 if utani, kangethe or windala
CATTLE 1 Continuous Number of cows and bulls

In cluster analysis, it is possible to form as many numbers of clusters as possible
depending on the rationalization of the classification scheme. Hence, it is important
to determine the number of clusters that should be formed. Although there are
several guidelines for selecting the number of clusters there is no objective procedure
for doing so (Hair et al., 1992). In this analysis, intuitive conceptual relationships

Vol. 7, No. 2, 1997/2002


based on past research on soil fertility management in the area suggested three
target groups of farmers for on-farm research on FASE strategies (McCown, 1996).
It is hypothesized that there exists a minority of farmers who practice productive
and sustainable agriculture because they have the information and resources to
invest in soil fertility maintenance. A second category of farmers cannot invest in
soil fertility maintenance irrespective of availability of information on technological
possibilities because they do not have the resources to re-allocate to investments
in soil productivity. A third group comprises farmers who have access to financial
resources that can be re-allocated to investments in soil fertility maintenance if the
marginal returns to those investments provide sufficient incentives given their level
of risk aversion and investment portfolio.

4. Empirical results

Table 2 shows results from the logit regression. Hypothesis tests are based on the
Wald statistic which has a chi-squared distribution. The estimated model fitted the
data fairly well, correctly predicting 68% of farmers' current fertility maintenance
practices. Six variables were significant at the 20% level of significance or better.
Farmers who had access to cash resources through regular milk sale, cultivated a
cash crop, had off-farm remittance from family members, and had smaller family
labor available for farm work were more likely to use fertilizer as a component of
their soil fertility maintenance strategy. Farmers who sold food grain were less likely
to include fertilizer in their fertility maintenance strategy. Income from crop and milk
sales as well as remittances from other family members were important determinants
of the likelihood of fertilizer use because they provided financial resources which
facilitated investments in soil fertility maintenance. Family labor size was negatively
related to the likelihood of fertilizer use because farmers with large family labor
supply would probably rely on the more labor-intensive application of manure.
Significant explanatory variables which positively influenced the probability that
fertilizer will be a component of farmers' soil fertility management strategies
were selected as delineating variables for the cluster analysis. Livestock sale was
significant at only the 30% level of significance, but it was included as a delineating
variable because it could provide cash that is likely to influence fertilizer use.

Table 3 reports the numbers of farmers in each cluster and the mean and standard
deviation of standardized values of the delineating variables. Farmers were not.
equally distributed across clusters although the distribution was fairly close. In
cluster 1, all the income and marketing variables; represented by cash crop, regular
milk sales, off farm income from other family members, and livestock sales, were
below the overall mean for all farmers. Cluster 2 had the largest standard deviation
above the sample mean for cash cropping activities and regular milk sales and the
second largest standard deviation above the sample mean for livestock sale. Cluster
3 had the largest standard deviation above the sample mean for livestock sales and
off farm employment of other family members. The mean values for cash crop and
regular milk sales were also below the sample mean. Clusters 1 and 3 had the same
standard deviation below the sample mean for cash cropping activity but cluster 3
had a larger standard deviation for regularity of milk sales.

Journal for Farming Systems Research-Extension


Table 2: Results of Logit regression
Variable Coefficient estimate Significance a
AGE -.0187 -.3964
GENDER -.2902 -.6510
CA RATIO -.0621 -.6791
ED1 .2987 .6933
AG. TRAIN -.1781 -.7836
FARMSIZE -.0099 -.7967
NO F LAB -.2576 -.1619*
MORGAN. .2736 .5819
C CROP .9467 .1701*
GRAINSAL -.9486 -.1149*
MILKSALE 1.2367 .1084*
LIV.SALE .7087 .2785
OFFEEMP -.4992 .3643
OFFFEOTH 1.7386 .0156*
V1 .9014 .0987
CAITLE1 .1274 .5477
Constant -1.2875 -.4546
a Significance level for Wald Statistics
* Significant at 20% level.

Table 3. Final Cluster Centers*


1 2 3
NO F LAB -.67943 .04608 .73891
C CROP -.57446 1.06685 -.57446
MILKSALE -.46617 .79175 -.37984
LIV.SALE -.80505 .26000 .63590
OFFFEOTH -.11516 -.04537 .18728
NO. OF CASES 35.000 35.000 30.000
Results are standardized scores

The key variables which determined farmers ability to invest in a FASE strategy
are income and marketing variables since they provided the financial resources
which could be used for investments in inputs for maintaining soil fertility. The
results from the cluster analysis suggested that farmers in cluster 2 had the highest
potential for investing in a FASE strategy because they had more access to financial

VoL 7, No. 2,1997/2002


resources from cash cropping activities and milk sales. Farmers in cluster 3 had
potential for investments in FASE strategies because they obtained income from
livestock sales and remittances from family members with off-farm employment.
Given sufficient incentives these farmers were likely to have sufficient flexibility to
invest in a FASE strategy if they perceived the incremental return to be high enough.
Farmers in cluster 1 had limited income sources and, therefore, did not have the
financial resources or had limited potential to invest in a FASE strategy involving
cash investments in inorganic fertilizer.

Table 4 shows descriptive statistics for the three clusters. Cluster 1 had 35 farmers
with a relatively small family size compared to the other two clusters. The household

Table 4. Mean and standard deviation of selected characteristics of the
household clusters
VARIABLE Cluster 1 Cluster 2 Cluster 3 Whole Sample
Sample Size n=35 n=35 n=30 n=100
Age (years) 45.83 51.97 50.43 49.36
(17.89) (15.67) (14.54) (16.23)
Education of head of 1.34 1.09 1.07 1.17
household (1.06) (1.24) (1.14) (1.15)
(average years in school)
Household size (no.) 4.80 6.37 8.03 6.32
(2.22) (2.44) (3.13) (2.89)
Farm size (acres) 5.41 8.68 6.51 6.89
(4.40) (12.23) (4.39) (8.10)
Cultivated land (acres) 2.92 4.66 3.73 3.77
(1.86) (4.08) (2.60) (3.07)
Pasture (acres) 2.48 4.01 2.74 3.09
(3.60) (9.25) (2.77) (6.04)
Cattle (Nos.) 0.77 200 0.93 1.25
(1.06) (1.72) (1.20) (1.46)
Sheep (Nos.) 0.31 2.20 0.47 1.02
(1.08 (6.81) (0.94) (4.16)
Goats (N6s.) 2.26 7.29 4.77 4.77
(3.79) (13.38) (4.14) (8.71)
Oxen (Nos.) 0.63 1.23 1.43 1.08
(0.97) (1.22) (1.25) (1.19)
Family members in off-farm 0.89 1.14 1.03 1.02
employment (0.93) (1.03) (1.19) (1.04)
Ratio of adults to children 1.78 1.45 1.89 1.70
(2.06) (1.39) (1.40) (1.66)
Standard deviations in parenthesis
Source: Survey Results

Journalfor Farming Systems Research-Extension


heads were relatively young and had, on average, the smallest cultivated and
pasture areas. Unlike the other clusters, many members of this cluster did not own
any livestock. These households applied the smallest quantity of manure on their
farms probably because they owned few animals. Consistent with the cluster result,
these households have the least access to financial resources that could be used
for making investments. They hardly produced a marketable surplus of grain or
cash crops and therefore were not actively engaged as sellers in the rural economy.
However, a number of them were employed in off-farm activities but it was also
likely that the income earned from these activities mostly supported household
subsistence needs. The relatively small number of farmers using inorganic fertilizer
in this cluster suggests that many of these households are below the cash income
threshold necessary for farmers to make investments in fertility maintenance
strategies involving use of inorganic fertilizer.

Cluster 3 had 30 farmers and reported the largest family size. These farmers had
modest farm sizes, owned more livestock than cluster 1 household, and therefore
had a higher proportion of farmers applying manure compared to cluster 1. About
one third of the members of this cluster regularly sold grain. A considerable number
were also engaged in off-farm income generating activities. A higher proportion of
farmers in cluster 3 applied-fertilizer on their farms compared to cluster 1 farmers
(Table 5). Given the right level and stability of incentives and/or the information
to correctly assess the potential benefits from their investments, these farmers are
likely to invest in inorganic fertilizer because they have resources that could be re-
allocated to making capital investments in soil productivity.

Table 5: Selected qualitative features of the household clusters
Variable Cluster 1 Cluster 2 Cluster 3
Sample size n=35 n=35 n=30
Female (%) 26 24 23
Male (%) 74 76 77
Has agricultural training (%) 20 63 13
Taken agricultural loan (%) 0 6 0
Has at least primary education (%) 69 51 53
Grows cash crops (%) 0 71 0
Sells grain regularly (%) 6 37 30
Nutrient source
Inorganic fertilizer (%) 16 43 41
Manure (%) 27 33 40
Involved in off-farm income activity 40 31 30
Source: Survey data

Cluster 2 had 35 farmers who, on average, had the largest farm size and moderate
family size. They had the largest livestock herds, especially small ruminants, and

Vol. 7, No. 2, 1997/2002


applied the largest quantity of manure on their farms compared to farmers in cluster
1 and 3, presumably because they had more manure available. This finding confirms
other results that show a strong positive relationship between the size of livestock
holding and the level of manure application in this area (Omiti et al., 1998). Most of.
the farmers in this cluster cultivated a cash crop and regularly sold grain. They also
had more diverse sources of income and off-farm employment. It is therefore not
surprising that this custer had the largest proportion of farmers using inorganic
fertilizers, although in small quantities.

5. Discussion and conclusions

This paper used logit and cluster analysis to classify farmers into target groups for
on-farm research on soil fertility maintenance based on their capacity to invest in
alternative technologies. These clusters provide a framework for conducting a range
of practical experimentation on soil fertility maintenance options with different
groups of farmers. For example, many farmers in cluster 2 and 3 have financial
resources that can be used for investment in inorganic fertilizer although at varying
levels of intensity. Given the right type of information they can correctly perceive
the benefits from using inorganic fertilizer. These farmers also own livestock that
provide manure which can be used for maintaining soil fertility. Farmers in these
dusters would therefore be interested in testing options that maximize the returns
to the investments they are already making in maintaining soil fertility. Some
farmers in these clusters would be interested in experimentation to understand
the effects of continuous fertilizer application on particular fields, options that
improve fertilizer use efficiency through correct timing, and methods of application
or questions relating to the best options for targeting limited quantities of inorganic
fertilizer. In addition, some of these farmers could be interested in options involving
balanced use of nutrients or those offering improved nutrient use efficiency through
increased exploitation of the complementarities between inorganic fertilizer and
manure. On the other hand, the characteristics of many farmers in cluster 1 imply
that they cannot realistically benefit from strategies involving investments in
inorganic fertilizer because they do not have the financial resources to invest in
purchased inputs. But many of these farmers have access to and are using small
quantities of low quality manure. Experimentation on options that would help these
farmers make better use of the limited quantities of manure available should be
appealing to these groups of farmers especially if the interventions have relatively
low labor and cash requirements. Such interventions would include those that
improve manure quality and manure use efficiency through better targeting and
improved distribution to cropland.

This study shows the usefulness of classifying farmers into meaningful target groups
for on-farm research based on their priority research problems and the opportunities
for testing alternative options to resolve these problems. Such an approach assists
decision-making by helping to set priorities for on-farm research and identifying
key research opportunities that exploit the diversity of farmers' situation and
farming systems. Further research and testing with farmers will help establish the
soil fertility benefits of various technological options on-farm, their profitability, and

Journal for Farming Systems Research-Extension


farmer acceptance. This type of framework enhances the probability of adoption of
soil fertility maintenance technologies because it improves the capacity of research
to develop practical recommendations that smallholder farmers are more likely to


This study is the outcome of collaborative research between the Kenya Agricultural
Research Institute and the International Crops Research Institute for the Semi-
Arid Tropics. We are grateful to the Rockefeller Foundation and International
Crops Research Institute for the Semi-Arid Tropics for funding the project. The
authors thank David Rohrbach, Subhash Chandra, Richard Coe, Steve Franzel,
two anonymous journal reviewers, and the journal editor for helpful comments on
earlier drafts of this paper. The views expressed in this paper are those of the authors
and should not be attributed to their respective institutions.


CARMASAK. 1996. Learning with farmers how to better manage crops and croplands in semi-arid
eastern Kenya using field experiments, simulation models and group discussions. In Collaboration
on Agricultural/Resource Modeling and Applications in Semi-Arid Kenya (CARMASAK) Project, Kenya
Agricultural Research Institute (KARI), Kenya.

CIMMYT. 1988. From agronomic data to farmer recommendations: an economics training manual, CIMMYT,

Franzel, S. 1981. Identifying farmer target groups in an area: methodology and procedures. In Farming
Systems Newsletter, No.4, CIMMYT East African Economics Program.

Freeman, H.A., and W Kaguongo. 1998. Fertilizer market reforms and farmers changing soil fertility
management practices: evidence from semi-arid Kenya. Paper presented at the 15th international
symposium of the Association of Farming Systems Research-Extension, November 29 December 4,
1998, Pretoria, South Africa.

Hair, J.E, R. E. Anderson, R.L. Tatham, W C. Black. 1992. Multivariate data analysis with readings, Maxwell
Macmillan Canada, Inc., Ontario.

Harrington, L.W, and R. Tripp. 1984. Recommendation domains: a framework for on-farm research. In
Economics Program Working Paper 02-84, CIMMYT.

Hilderbrand, EE. 1986. The concept of homogeneous systems and its usefulness. In Perspectives of farming
systems research and extension, P Hilderbrand (ed.). Lynne Rienner Publishers Inc. Boulder, Colorado.

IFPRI. 1995. A 2020 vision for food agriculture and the environments: the vision, challenge, and recommended action,
IFPRI, Washington, D.C.

KARL 1995. National Dryland Farming Research Center-Katumani, Regional Research Program, KARI,

KARl. 1996a. A participatory study of farmers constraints, opportunities and research needs in the hilly masses of
Eastern Kenya, KARI, Kenya.

KARL 1996b. Participatory rural appraisal: a case study of Kasikeu sub-location, Makueni District, KARI, Kenya.

Vol. 7, No. 2, 1997/2002


Kimotho, L.M., EK. Asambu, R.E. Hudgens, and J.K. Ransom. (n.d.) Crop management factors causing
maize grain yield reduction in eastern Kenya. Paper presented at 15th Biennial Weed Science Society
of East Africa, Morogoro, Tanzania.

Kumwenda J.D.T., S. R. Waddington, S. S. Snapp, R.B. Jones, and M.J. Blackie. 1996. Soil fertility management
research for the maize cropping smallholders of Southern Africa: a review. In Natural Resources Group
Paper 96-02, CIMMYT.

McCown R.B. and B.A. Keating. 1992. Looking forward: finding a path for sustainable farm development.
In A Search for Strategies for Sustainable Dryland Cropping in Semi-Arid Eastern Kenya, Probert, M.E. (ed.).
ACIAR Proceedings no. 41.

McCown, R. 1996. A discussion paper as input to the on-farm research strategic and operational plan.

Muhammad, L.W, and K.A. Parton. 1992. Smallholder farming systems in semi-arid eastern Kenya basic
issues relating to the modeling of adoption. In A search for strategies for sustainable dryland cropping in
semi-arid Eastern Kenya, Probert, M.E. (ed.). ACIAR Proceedings no. 41.

Murithi, EM. and M.C. Shiluli. 1993. Effects of the liberalization of fertilizer markets on the distribution and
use of fertilizer on food crop production: A study on Embu and Meru districts of Kenya. In Cereal grain
policy analysis in the National Agricultural Research Systems of Eastern and Southern Africa, W Mwangi et
al. (eds.). Addis Ababa: CIMMYT SADC/ICRISAT.

Omiti, J.M., H.A. Freeman, and C. Bett. 1998. Soil fertility management practices in semi-arid agriculture: results
of a baseline survey in Machakos District of Eastern Kenya, International Crops Research Institute for the
Semi-Arid Tropics Research Bulletin.

Probert, M.E., J.R. Okalebo, and R.K. Jones. 1995. The use of manure on smallholders' farms in semi-arid
eastern Kenya. Experimental Agriculture, Vol. 31, pp. 371-381.

Probert, M.E., B.A. Keating, M.N. Siambi, and J.R. Okalebo. 1994. Management of soil fertility in climatically
risky environments. In Soil fertility and climatic constraints in dryland agriculture, E.T. Caswell and J.
Simpson (ed.). Proceedings of an ACAIR/ACCAR workshop held at Harare, Zimbabwe, 30 August-1
September, 1993.

Sanchez, PA, A., M. Izac, I. Valencia and C. Pieri. 1996. Soil fertility replenishment in Africa: a concept note.
In Achieving greater impact from research investments in Africa, Breth, S.A. (ed.).

Sanders J.H., and J.D. Vitale. 1990. Technology development for traditional cereals in the Sahelian Countries,
Invited paper presented at the African Farming Systems Symposium, Burkina Faso.

Smaling, E.M.A., S.M. Nandwa and B. Hansen. 1997. Soil fertility in Africa is at stake. In Replenishing soil
fertility in Africa, Buresh, R.J. et al. (eds.), SSSA Special Publication Number 51.

SPSS. 1977. SPSS Base 7.5 application guide, SPSS Inc., USA.

Tiffen, M., M. Mortimore, and E Gichuki. 1994. More people less erosion, ACTS Press, Nairobi, Kenya.

Waddington, S.R., R. Gilbert, and K.E. Giller. 1997. Best bet technologies for increasing nutrient supply for
maize on smallholder farms. In Soil fertility research for maize-based systems in Malazi and Zimbabnoe,
Waddington, S. et al. (eds.). Proceedings of the Soil Fertility Net Results and Planning Workshop, July
7-11, 1997, Harare, Zimbabwe.

Williams, T.O. 1994. Identifying target groups for livestock improvement research: the classification of
sedentary livestock producers in western Niger. Agricultural Systems, 46, pp. 227-237.

Journalfor Farming Systems Research-Extension

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs