Citation
The Organic matter budget and energy flow of a tropical lowland aquatic ecosystem

Material Information

Title:
The Organic matter budget and energy flow of a tropical lowland aquatic ecosystem
Creator:
Brinson, Mark M. ( Dissertant )
Lugo, Ariel E. ( Thesis advisor )
Anthony, David S. ( Reviewer )
Griffin, Dana G. ( Reviewer )
Nordlie, Frank G. ( Reviewer )
Popenoe, Hugh I. ( Reviewer )
Shanor, Leland ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1973
Language:
English
Physical Description:
251 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Bodies of water ( jstor )
Dry seasons ( jstor )
Lakes ( jstor )
Oxygen ( jstor )
Primary productivity ( jstor )
Rain ( jstor )
Respiration ( jstor )
Rivers ( jstor )
Surface runoff ( jstor )
Watersheds ( jstor )
Aquatic ecology -- Guatemala ( lcsh )
Botany thesis Ph. D
Dissertations, Academic -- Botany -- UF
Water-supply -- Guatemala ( lcsh )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
This study examines the influence of regional coupling mechanisms on the organic matter budget of a lowland tropical lake and documents the principal energy flows that contribute toward making the watershed a cohesive ecological unit. the downhill flow of organic matter (OM) and water from the terrestrial (6,860 km ) to the aquatic (717 km 1 ecosystem was quantified and evaluated as to its effect on physical conditions and metabolic activity of Lake Izabal, Guatemala. The lake type is warm polymictic due to its shallow mean depth (11.6 m) . Its water mass has a short residence tin;e (6.6 months), and the annual gross 2 productivity is high (1,592 g OM/m ). The hydrologic regime exerted control on OM flow into the lake. Mean runoff for the watershed was calculated as 6S percent of mean annual rainfall (2,992 mm), most of which occurred during the 9-month wet season (greater than 100-mm monthly rainfall). Organic matter runoff was characterized by an initial "flushing" of the watershed at the beginning of the wet season, when the particulate fraction (greater than O.SO microns 1 diameter) constituted a large portion of the total organic flow. During the remainder of the year, OM runoff was nearly all contained in the dissolved fraction (less than 0.80 microns diameter). Approximately 50 percent of the total OM runoff occurred during the 3 wettest months of the year. During the dry season when OM inputs from the watershed were low, the lake experienced a net gain in OM when gross primary productivity exceeded community respiration. During the wet season, there was a net loss of OM in spite of increased inputs of allochthonous organic detritus. This loss increased as the wet season progressed, due to a combination of decreased rates of gross primary productivity and increased rates of community respiration. The periodicity of OM accrual and loss provides a mechanism, apparently controlled by hydrologic patterns, by which steadystate conditions can be achieved on an annual basis. Daily rates of gross primary productivity ranged from 1.15-7.31 g 2 2 /m day and planktonic respiration rates from 0.50-8.38 g /m day. The average daily values for the organic matter budget were calculated for the 8-month sampling period and were represented by five principal 2 flows. The two OM sources were gross primary productivity (3.730 g/m 2 day) and OM imports (0.632 g/m day). The three OM losses were by OM exports (0.452 g/m day), planktonic respiration (3.875 g/m" day), and 2 respiration of the bottom muds (0.36 g/m" day). The mean residence time 2 for the average OM content of the lake (71.08 g/m ) was 16.3 days. Seasonal periodicity was expressed in the net plankton by a bimodal pulse of abundance. Peaks in plankton density occurred at the end of the dry season (April-May) and followed the initial period of heavy rainfall (August-September). Causal factors for this response remain undetermined. The connection of the lake to the marine environment, via a 42-kir. long waterway, allows additional mechanisms for ecosystem coupling. Evidence was collected to demonstrate the control of the Na:Cl ratio of lake water by dry-season penetration of brackish water into the lake. The waterway also provides marine vertebrates and invertebrates access to a fresh-water environment. Periodicity of OM metabolism in the lake, high productivity of surrounding lagoons and coves, and brackish water penetration from the coastal marine ecosystem are discussed as factors influencing consumer activity and seasonal migration. The lake's fisheries, dependent on the marine contingent of fishes, may best be managed by utilizing the understanding of regional coupling mechanisms to prevent fisheries deterioration and to ensure continued yields.
Thesis:
Thesis (Ph. D.)--University of Florida, 1973.
Bibliography:
Includes bibliographical references (leaves 244-250).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Mark McClellan Brinson.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
022293554 ( AlephBibNum )
37876216 ( OCLC )
ACZ2808 ( NOTIS )

Downloads

This item has the following downloads:


Full Text




















THE ORGANIC MATTER BUDGET AND ENERGY FLOW
OF A TROPICAL LOWLAND AQUATIC ECOS 'STEM








By





MARK McCLELLAN BRINSON
1


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF F1.ORID~A IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR Fi;E DEGREE OF DOCTOR OF PHILOSOPHY





UNIVERSITY OF FLORIDA
1973























Copyright by


Mark McClellan Brinson

1973



















Digitized by the Internet Archive
in 20091withp~iuding from
University of Florida, George A. Smathers Libraries


http://www.archive.org/details/organicmatterbud00brin








ACKNOWLEDGMENTS


The author wishes to express his appreciation to Dr. Ariel Lugo

for ecological insight and logistical support which helped to insure the

success and feasibility of this investigation. Sincere appreciation goes

to Dr. Frank Nordlie for helpful discussions on Lake izabal, and to the

other members of the supervisory committee, Dr. David Anthony, Dr. Dana

Griffin, III, and Dr. Leland Shanor, for critically reviewing the manu-

script.

Financial support that was provided by the Foreign Area Fellowship

Program during the field study and write-up period is gratefully acknow-

ledged. The Center for Tropical Agriculture provided assistance during

a reconnaissance trip prior tc the field research. Dr. Hugh Popenoe's

continued i-;ntrest in the Izabil Watershed provided the frinme.'ork for

research on a regional level.

Successful affiliation was established with employees of the Depart-

ment of Wildlife, Ministry of Agriculture, Guatemala. This was possible

through the efforts of its Chief, Jose Ovidio de Le6n, who was also res-

ponsible for processing equipment through customs. Adolfo Orosco Pardo

assisted in counting much of the plankton and Renato Z6niga helped with

a phase of the field operations.

Thanks goes to the many people of El Estor who helped to provide day-

to-day needs, and to the EXMIBAL mining company for supplying climatolog-

ical data during the period of study. I am particularly grateful for my

wife's assistance and encouragement during our memorable experiences in the

Guatemalan lowlands.













TABLE OF CONTENTS


Page


ACKNOWLEDGMENTS................................................. iv

LIST OF TABLES .................................................. x

LIST OF FIGURES ................................................. xii

ABSTRACT...... ................................................. xvi

INTRODUCTION..................................................... 1

Terrestrial Ecosystem Exports............................. 2

Inflows to Aquatic Ecosystems ............................. 5

REGIONAL SETTING ................................................ 9

Climate................................................... 9

Lakeside Temperature, Rainfall and Solar Radiation.... 9

Precipitation for the Remainder of the Watershed...... 16

Geology and Soils......................................... 19

The Lake .................................................. 21

The Fisheries and People.................................. 24

Immigration and Labor Alternatives.................... 25

Fishing Regulations and Yields........................ 26

Fish Behavior and Fisheries Management................ 28

HYDROLOGY AND WATER CHARACTERISTICS............................. 32

Methods.................................................... 33

General Methodology--Field Stations, Logistics,
and Schedule....................................... 33













TABLE OF CONTENTS continued




Hydrologic Measurements ..............................

Analyses of Water Characteristics.....................

Results and Discussion ...................................

River Discharge and Runoff...........................

Water Budget........................................

Water Characteristics................................

Thermal properties and circulation patterns
of the lake....................................

The seasonal pattern of water characteristics....

Mineral analyses and the Na:Cl ratio.............


Brackish water movement into Lake Izabal.

Summary Statement ................................

NET PLANKTON AND BENTHIC COMMUNITIES..................

Methods ..........................................

Results and Discussion.............................

Phytoplankton and Zooplankton Communities....

Diatoms..................................

Green algae..............................

Blue-gredialgae .........................

Other algae..............................

Phycomycetes............................


Page

36

38

41

41

52

60


60

66












TABLE OF CONTENTS continued


Page

Copepods ......................................... 117

Cladocera......................................... 117

Rotifers................................*....... 122

Possible Controlling Factors......................... 122

Benthic Community .................................... 130

Summary Statement..........................*.... ......... 133

METABOLISM AND ORGANIC MATTER .................................. 138

Methods.............................**....** ........... .. 139

Chemical Oxygen Demand................................ 139

Fraction definitions and conversion criteria..... 139

Collection and treatment of samples .............. 140

Dissolved Oxygen Concentration Determinations........ 143

Biological oxygen demand......................... 144

Respiration of bottom muds and their organic
content ........................................ 145

Light and dark bottle method..................... 147

Results and Discussion................................... 148

Concentrations of Chemical Oxygen Demand.............. 149

Swamp waters ..................................... 149

Rio Polochic distributaries....................... 152

Small rivers............................ ......... 152














TABLE OF CONTENTS continued


Page

Lake stations and outlet......................... 153

Monthly Flows of COD................................. 154

Respiration Rates (BOD) ............................... 160

Respiration and Organic Content of the Bottom Muds... 168

Measurements of Primary Productivity................. 175

Seasonal rates of gross primary productivity
and respiration ................................ 178

Efficiency of gross primary productivity......... 185

Light penetration and ccm:ponation depth ......... 19

Balance of the Organic Matter Budget................. 193

Summary Statement........................................ 202

CONCLUSIONS.................................................... 204

Seasonal Pulse of Organic Detritus Movement.............. 208

Movement of Organic Matter to Site of Consumption.... 208

Mechanisms for Steady-State Balance.................. 210

Oscillations in Food Concentration for Consumers.......... 212

Consequences of the Connection to a Marine Ecosystem..... 213

Ecosystem Management ..................................... 213

Fishery Management ................................... 214

Modern Agricultural and Industrial Man............... 215

APPENDICES..................................................... 220


viii













TABLE OF CONTENTS continued


Page

LITERATURE CITED.............................................. 244

BIOGRAPHICAL SKETCH............................................. 251













LIST OF TABLES


Table Page

1 Sampling dates for chemical oxygen demand (COD), bio-
logical oxygen demand (BOD), and primary productivity
experiments (Prod).......................................... 34

2 Monthly discharges (m3 x 106) of rivers and watersheds
draining into Lake Izabal ................................... 50

3 Summary of hydrologic data for the major watersheds
draining into Lake Izabal................................... 51

4 Monthly free water surface evaporation measurements
for selected warm climates and warm seasons................. 53

5 Summary of water budget for Lake Izabal (November
1971-October 1972).......................................... 55

6 Ranges of temperature, pH, total alkalinity (T.A.),
specific conductance, and dissolved oxygen satura-
tion for all stations sampled at Lake Izabal................ 67

7 Sodium and chloride concentrations (mg/liter) and
Na:C1 ratios for lake stations and distributaries
of the Rio Polochic .......................................... 79

8 Species list of net plankton frequently collected from
Lake Izabal................................................. 105

9 Bottom fauna of Lake Izabal expressed in numbers/m2......... 132

10 Characteristic oxygen equivalents (O.E.) in approxi-
mate order of their limnological importance................. 141

11 Monthly COD inflows (g COD x 106) for individual rivers
and watersheds, total monthly inflows to Lake Izabal,
and monthly and total outflow at San Felipe.................. 155

12 Relative contribution (percent) of organic matter run-
off into Lake Izabal from the Rio Polochic Valley and
from the minor watersheds................................... 159

13 Hourly and daily respiration rates of Lake Izabal
bottom muds ................................................. 173













LIST OF TABLES continued


Table Page

14 Metabolism calculations of light and dark bottle experi-
ments for Stations A, B, and C during 1972................... 181

15 Total daily incoming radiation averaged on a weekly and
monthly basis from October 1971 through October 1972........ 186

16 Daily efficiencies of energy conversion from visible
solar energy to energy fixed by gross primary pro-
ductivity (Pg) .............................................. 188

17 Calculation of light intensity at compensation depths
as percent of surface intensity ............................. 194

18 Summation of organic matter (OM) flows for Lake Izabal
from March to October 1972................................... 198

19 Daily and annual rates of gross primary productivity (Pg)
for some tropical lakes and comparative ranges for
temperate lakes............................................. 201













LIST OF FIGURES


Figure Page

1 The Izabal Watershed ........................................ 11

2 Monthly averages of the maximum, minimum, and calculated
mean temperatures at Las Dantas for the period of study
(October 1971-1972)......................................... 15

3 Temperature-rainfall climate diagram for Las Dantas and
Mariscos during the year of study........................... 18

4 Bathymetric map of Lake Izabal illustrating sampling
stations .................................................... 23

5 Calibration curves for the small and large probes used
for specific conductance measurements between 100 and
30,000 imho/cm at 25 C...................................... 40

6 Monthly estimates of all inputs to the lake (runoff and
direct rainfall) and the monthly averages of lake
water levels................................................ 44

7 Linear relationship between monthly rainfall and monthly
discharge of rivers emptying into Lake Izabal............... 46

S The seasonal march of monthly discharge rates for some
major rivers emptying into Lake Izabal...................... 48

9 Relationship between velocities at the San Felipe outlet
from direct field measurements and those calculated by
balancing the water budget................................... 57

10 Summary diagram of water storage and annual flows........... 59

11 Vertical temperature profiles of Lake Stations A, A-B,
and B recorded at approximately monthly intervals........... 63

12 Vertical temperature profiles of Lake Stations B-C and C
recorded at approximately monthly intervals................. 65

13 Seasonal changes in dissolved oxygen concentration,
temperature, total alkalinity (T.A.), conductivity,
and pi! for lake stations .................................... 69












LIST OF FIGURES continued


Figure Page

14 Seasonal changes in dissolved oxygen concentration,
temperature, total alkalinity (T.A.), conductivity,
and pH for swamp waters ...................................... 72

15 Seasonal changes in dissolved oxygen concentration,
temperature, total alkalinity (T.A.), conductivity and
pH for distributaries of the Rio Polochic (Comercio,
Cobin, and Bujajal).......................................... 74

16 Seasonal changes in dissolved oxygen concentration,
temperature, total alkalinity (T.A.), conductivity, and
pH for small rivers (Sauce, San Marcos, and Manacas Creek)... 76

17 Map of the Rio Dulce-El Golfete system....................... 82

18 Conductivity profiles (March 22-23, 1972) along the Rio
Pulse fro: San: Felipe (Station 1) to Amatique Pay
(Station 11) ................................................. 84

19 Temperature profiles at two stations illustrating a
slight temperature increase associated with the halocline.... 86

20 Observed diurnal change in the conductivity profile of
the waters at the lower reaches of El Golfete (Station 8).... 88

21 Conductivity of the ground water in a swamp forest at
Cuatro Cayos ................................................. 91

22 Conductivity profiles along the Rio Dulce from San Felipe
(Station 1) to the lower reaches of El Golfete
(Station 8).................................................. 93

23 Conductivity profiles taken at several stations in Lake
Izabal and the upper reaches of Rio Dulce (Stations
1 and 2) ..................................................... 95

24 Bottom profile of the Lake Izabal-Rio Dulce system,
showing locations of sampling stations...................... 98

25 Seasonal changes in abundance of pennate diatoms and
Melosira granulata........................................... 107


xiii













LIST OF FIGURES continued


Figure Page

26 Seasonal changes in abundance of Staurastrum leptocladum
and S. pingue................................... .. .......... 110

27 Seasonal changes in abundance of Pediastrum simplex
and Staurastrum tohopekaligense.............................. 112

28 Seasonal changes in abundance of phycomycetes and
Anacystis cyanea ............................................. 115

29 Seasonal changes in abundance of copepods.................... 119

30 Seasonal changes in abundance of cladocera................... 121

31 Seasonal changes in abundance of colonial rotifers........... 124

32 Seasonal changes in abundance of solitary rotifers .......... 126

33 Total net-plankton abundance (units or organisms per
liter) represented by phytoplankton, zooplankton,
and phycomycetes............................................. 136

34 Concentrations of particulate and dissolved COD (mg/liter)
during the sampling period for (a) swamp waters, (b) Rio
Polochic distributaries, (c) small rivers, and (d) lake
stations (A, B, C, and San Felipe) ........................... 151

35 Rates of organic matter inflows and outflows of Lake
Izabal for the lake as a whole (g COD x 109/month) and
for an average m2 of surface area (g COD/m2 day).............. 157

36 Respiration rates (1-day and 5-day BOD) for (a) swamp
waters, (b) Rio Polochic distributaries, (c) small
rivers, and (d) lake stations ................................ 162

37 Dissolved oxygen concentrations of water samples during
7-day incubation periods..................................... 165

38 Relationship between oxygen consumption rates (BOD) and
total COD concentrations for the four water types
characterized................................................ 167












LIST OF FIGURES continued


Figure Page

39 Organic composition of Ekman samples from Lake Izabal
bottom muds ..................................... ............. 170

40 Respiration rates of mud samples from Lake Izabal............. 172

41 Example of curves generated from light and dark bottle
experiments from which metabolism is determined
planimetrically .............................................. 177

42 Method for calculating the number of hours of effective
light per day................................................ 180

43 Gross primary productivity (Pg), 24-hour respiration
(R24) and Pg/R24 ratios at Station A, B, and C............... 184

44 Scchi disk transparency measurements recorded at
Stations A, B, and C......................................... 191

45 Simplified model of the principal organic matter flows
and storage in Lake Izabal as averaged over the period
of study ..................................................... 196

46 Summary diagram of energy and matter flows and storage
that characterize the Izabal Watershed ....................... 206

47 Summary diagram of energy and matter flows and storage
of the 1zabal Watershed which includes some of the
possible influences of development by modern agricul-
tural and industrial man ..................................... 217







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


THE ORGANIC IIMATTER BUDGET AND ENERGY FLOW
OF A TROPICAL LOWLAND AQUATIC ECOSYSTEM


By

Mark McClellan Brinson

December, 1973


Chairman: Ariel E. Lugo
Major Department: Botany


This study examines the influence of regional coupling mechanisms

on the organic matter budget of a lowland tropical lake and documents

the principal energy flows that contribute toward making the watershed

a cohesive ecological huit. The downhill flow of organic matter (OM)
2
and water from the terrestrial (6,860 km ) to the aquatic (717 km)

ecosystem was quantified and evaluated as to its effect on physical

conditions and metabolic activity of Lake Izabal, Guatemala. The lake

type is warm polymictic due to its shallow mean depth (11.6 m). its

water mass has a short residence time (6.6 months), and the annual gross

productivity is high (1,592 g OM/m2).

The hydrologic regime exerted control on OM flow into the lake.

Mean runoff for the watershed was calculated as 65 percent of mean annual

rainfall (2,992 mm), most of which occurred during the 9-month wet season

(greater than 100-mnm monthly rainfall). Organic matter runoff was char-

acterized by an initial "flushing" of the watershed at the beginning of

the wet season, when the particulate fraction (greater than 0.S0 microns







diameter) constituted a large portion of the total organic flow. During

the remainder of the year, OM runoff was nearly all contained in the dis-

solved fraction (less than 0.80 microns diameter). Approximately 50 per-

cent of the total OM runoff occurred during the 3 wettest months of

the year.

During the dry season when OM inputs from the watershed were low,

the lake experienced a net gain in OM when gross primary productivity

exceeded community respiration. During the wet season, there was a net

loss of OM in spite of increased inputs of allochthonous organic detritus.

This loss increased as the wet season progressed, due to a combination of

decreased rates of gross primary productivity and increased rates of com-

munity respiration. The periodicity of OM accrual and loss provides a

mechanism, apparently controlled by hydrologic patterns, by which steady-

state conditions can be achieved on an annual basis.

Daily rates of gross primary productivity ranged from 1.15-7.31 g

0 /m2 day and planktonic respiration rates from 0.50-8.38 g 0 /m day.

The average daily values for the organic matter budget were calculated

for the 8-month sampling period and were represented by five principal

flows. The two OM sources were gross primary productivity (3.730 g/m2

day) and OM imports (0.632 g/m2 day). The three OM losses were by 0M

exports (0.452 g/m2 day), planktonic respiration (3.875 g/m day), and

respiration of the bottom muds (0.36 g/m2 day). The mean residence time

for the average OM content of the lake (71.08 g/m ) was 16.3 days.

Seasonal periodicity was expressed in the net plankton by a bimodal

pulse of abundance. Peaks in plankton density occurred at the end of the

dry season (April-May) and followed the initial period of heavy rainfall

(August-September). Causal factors for this response remain undetermined.

xvii







The connection of the lake to the marine environment, via a 42-km

long waterway, allows additional mechanisms for ecosystem coupling.

Evidence was collected to demonstrate the control of the Na:Cl ratio of

lake water by dry-season penetration of brackish water into the lake.

The waterway also provides marine vertebrates and invertebrates access

to a fresh-water environment. Periodicity of OM metabolism in the lake,

high productivity of surrounding lagoons and coves, and brackish water

penetration from the coastal marine ecosystem are discussed as factors

influencing consumer activity and seasonal migration. The lake's fish-

eries, dependent on the marine contingent of fishes, may best be managed

by utilizing the understanding ef regional coupling mechanisms to pre-

vent fisheries deterioration and to ensure continued yields.


xv ii











INTRODUCTION


Many years ago Forbes (1887) discussed lakes as microcosms, and

in so doing emphasized their isolation from the terrestrial ecosystem.

Much of the limnological work since that time has taken this myopic

view of lakes with little regard to the activities beyond their boundaries.

It is these extra-lacustrine activities that give lakes their character-

istics, and they can be dealt with effectively only by expanding the

ecosystem boundaries to include the whole of the watershed. The co-

hesive nature of the watershed, and its well-defined boundaries, are

characteristics that make it a conceptually attractive unit for ecolo-

gical study.

Powered by the proper energy sources, water is the common denominator

that couples this ecological unit by virtue of its geological constraints

(downhill flow) and its biological indispensibility. The characteristics

of this water, its flux from the terrestrial to the aquatic subsystem,

and the effect of this flux on the lacustrine ecosystem are all impor-

tant components of the present study.

The study was conducted in the Izabal Watershed on the Caribbean

slope of Guatemala. Because of the high rainfall and seasonal climate

of the region and the suspected short residence time of the lake's

waters, I hypothesized that upstream activities in the terrestrial sub-

system would be reflected by short-term (<1 year) responses in the down-

stream subsystem of the lake. The main focus of the study was to examine









the metabolic activity of the lake in response to organic matter inflow

from the watershed.

Tailing (1969), in reference to the poorly understood seasonality

of shallow tropical lakes with no long-term thermal stratification,

pointed out that "any periodicity [discovered] acquires a new interest."

In the Izabal Watershed, this interest may necessarily extend beyond

the boundaries of the lake to determine the extent of ecosystem coupling

and matter exchange between subsystems. The connection of the lake to

a marine environment adds to the difficulty of evaluating the watershed

as an isolated unit or closed system. This adds new dimensions for

possible mechanisms for ecosystem coupling.

Depending on the degree to which these coupling mechanisms create

interdependency on a regional scale, schemes for ecosystem management

and land use planning should demonstrate awareness of both the potential

benefits and inherent dangers of manipulation of isolated subsystems.

Therefore, as Odum (1971) emphasized, "it is the whole drainage basin,

not just the body of water, that must be considered as the minimum eco-

system unit when it comes to man's interests." Thus, there is an urgent

need for understanding these interactions on a regional level.


Terrestrial Ecosystem Exports


Naturally forested areas are extremely effective in recycling

materials, thereby preventing losses to downhill processes (Bormann et al.

1969). However, organic matter fixed by terrestrial photosynthesis under-

goes some leakage that eventually appears downstream. The degree to

which this leakage occurs depends on factors which characterize the









ecosystem. Some of the factors that significantly affect organic matter

runoff are: (1) seasonal phenomena, (2) runoff intensity, (3) ecosystem

perturbation, and (4) topography and other geological characteristics

of the watershed.

In temperate regions, the organic matter inputs to streams located

in forested areas are seasonal, being highest in the autumn during the

seasonal leaf drop (Kaushik and Hynes 1968). In tropical latitudes where

rainfall is the seasonal control, a deciduous forest presumably could

experience a similar pulse during the dry season. However, since water

is the medium that transfers organic matter downhill, one might expect

the downhill transfer of organic matter to be seasonally coincident with

high rates of runoff.

Although there are no data from tropical regions to support the

assumption that water runoff and organic matter runoff are positively

correlated, there is some evidence for this in the temperate zone. In

a Piedmont stream in the southeastern USA,Nelson and Scott (1962) found

that,although the dissolved and colloidal organic matter fraction of

river water increased with discharge rate, the particulate fraction in-

creased at a much more rapid rate. At low to moderate flows the dis-

solved and colloidal organic matter concentrations were two to ten times

higher than the particulate, while,during high flow rates, the particulate

organic fraction increased to double the concentration of the dissolved.

Causal factors which increased the particulate portion so dramatically

during high discharge were: (1) greater surface runoff associated with

heavy rainfalls, and (2) the flushing effect of high water. Therefore,

with increased rates of river discharge, organic matter flows increase at








a proportionately greater rate than water flows. There is no reason to

expect that these factors would operate differently in tropical latitudes.

Therefore, measurements of the organic matter runoff from the Izabal

Watershed to Lake Izabal would necessarily include a range of runoff

intensities in order to characterize size fractions of organic runoff

and to achieve a good estimate of absolute quantities.

Perturbation of terrestrial ecosystems by deforestation decreases

the ability of watersheds to prevent downstream losses by destroying the

mechanisms and adaptations for the recycling of matter. Regions of the

Izabal Watershed have received some alteration from deforestation and

agriculture. Again, specific data for estimating the magnitude of organic

runoff change by deforestation must be drawn from temperate-zone ecosystems.

The northern hardwoods of the Hubbard Brook Experimental Forest,

New Hampshire, provide the model on which to judge effects of perturba-

tion. There, drainage streams exported 5.3 grams of organic matter per
2
m of watershed area annually (Bormann et al. 1969). After clear-cutting

of the forest, organic matter losses doubled during the first two years.

These ecosystem exports represent inputs for downstream ecosystems

(Likens et al. 1970). If the downstream ecosystem were a lake, then

depending on its size, these can represent a significant source of organic

matter and illustrate the importance of one-way coupling between a ter-

restrial and aquatic ecosystem. The absolute values for tropical regions

may be different, but nevertheless deforestation could be expected to

result in increases in organic matter runoff. The only assumption neces-

sary for arriving at this conclusion is that the mechanisms for recycling

matter possessed by ecosystems of both latitudes would be lost or severely

damaged upon destruction of the forests.








Finally, the topography and other geological characteristics of

watersheds are factors that should be considered in the regulation of

organic matter runoff. Runoff waters from the steep mountain environment

in the Amazon Basin have higher concentrations of dissolved and sus-

pended organic matter than those of the lower Amazon (Gibbs 1967).

However, in large watersheds there may be a great deal of spatial varia-

tion in topography and geology as well as the previously discussed vari-

ables of runoff intensity and ecosystem perturbation. The larger the

watershed, the more these spatial variations tend to become integrated

by the confluence and mixing of tributaries by the time downstream

measurements are taken. Also the metabolic activities of riverine eco-

systems could be expected to modify both the quantity and quality of

organic matter after it is received upstream. In the Izabal Watershed,

the majority of the drainage areas become confined to one major river

before the waters discharge into the lake. Thus, any peculiarities in

organic matter sources will tend to cancel one another by the time the

water discharges into the lake.


Inflows to Aquatic Ecosystems


Now that the characteristics of organic matter runoff have been dis-

cussed and loosely established, it is essential to determine the quali-

tative and/or quantitative influence that this organic source may have

on downstream aquatic ecosystems. Showing that the energy fixed in the

form of organic matter in the terrestrial ecosystem can be utilized in

the aquatic ecosystem would establish that an energetic coupling or flow

occurs on a regional level.









There is good evidence that aquatic ecosystems adapt to organic

detritus inputs by utilizing them as a source of energy. This has been

demonstrated for estuaries (Teal 1962; Darnell 1967; Heald 1971; Cooper

and Copeland 1973) and in flowing waters (Odum 1956; Nelson and Scott

1962). That the Lake Izabal ecosystem should be an exception to this

concept would seem anomalous. Some workers have had the insight to com-

pare the relative contribution of organic matter import to the total

metabolism of coastal embayments. Of the total organic increment,

allochthonous inputs accounted for about one-half the total for the

Strait of Georgia (Seki et al. 1968), 7-21% in Moriches Bay on Long

Island (Barlow et al. 1963), and about 97% in the turbid and polluted

Chao Phya River estuary in Thailand (Pescod 1969). In addition to its

importance to the total metabolism of some areas, organic detritus has

been shown specifically to be a principal food item for many estuarine

vertebrates and invertebrates (Darnell 1961, 1967; Odum and de la Cruz

1967; W. Odum 1971). The basis for the nutritive and presumably the

energetic value of particulate detritus is the high quality of organic

matter (e.g. proteins) associated with the micro-organisms that colonize

the particles. Wherever ecosystems exist that have organic detritus

inputs, there is good evidence that the organisms are adapted to utilizing

the organic component as a source of energy.

Lakes, however, have apparently received the least attention of all

aquatic ecosystems that may derive some of their energy from organic

matter transported from outside their boundaries.

Again, specific examples that demonstrate this are drawn from studies

in temperate latitudes. Two independent approaches have been used to









measure allochthonous organic matter sources for lakes. One is to

directly monitor the organic matter runoff from a watershed--the ap-

proach used in the present study. This was done by McConnell (1963) in

small impoundments of the semi-arid southwestern USA. There, relatively

unleached oak litter entered a lake by surface runoff at an annual rate
2
of about 750 g dry weight per m of lake surface. This allochthonous

source of organic matter supplied the lake with approximately one-third

of the total organic matter increment; the remainder was supplied by

primary production.

The other approach is by indirect measurements and is feasible

only in lakes that thermally stratify. In eutrophic Lake Mikalajki,

Poland, Lawacz (1969) trapped seston as it sank from the trophogenic

zone to below the epilimnion. The total organic matter production by

this method was about half again as great as that of the plankton over

a year as measured by the oxygen method. Lawacz attributed the difference

in production to unmeasured dissolved organic substances that are pro-

duced in the littoral zone. These substances presumably move into the

pelagial zone and sink after transformation into particulate form.

Most attempts to quantify the organic sources of tropical fresh-

waters have been made by measuring the primary productivity of relatively

large lakes. The size of these lakes would tend to diminish the impor-

tance of an allochthonous organic source, and thus the "lake as a micro-

cosm" view is justified, at least for organic matter production.

As pointed out by Ruttner (1963), the quantity and composition of

allochthonous organic matter depends on the ratio of the surface area of

the lake to that of the watershed. In general, allochthonous organic









sources will be of greater importance to small lakes with large water-

sheds than to large lakes with small watersheds. Already discussed are

the important influencing factors such as climate, the morphological and

geological character of the watershed, and plant cover of the watershed.

Considering these factors, the humid tropics are a likely region

where lakes may receive flows of significant magnitude from outside their

boundaries. These flows can have a marked influence on the limnological

characteristics of these lakes. Lakes in which the residence time of the

water mass is relatively short are likely candidates for such tightly

coupled influences. In the context of ecosystem management, it follows

that alteration or perturbation of watersheds could have a profound

effect on the activities within such lakes. Similarly, these lakes may

be our most sensitive indicators of changes in upstream activity.

Thus, the objectives of this study are to (1) determine the

quantity and seasonal distribution of organic matter transfer from the

terrestrial to the lacustrine ecosystem, (2) quantitatively compare the

allochthonous organic matter source of the lake with the organic matter

derived by in situ primary production, (3) evaluate how this one-way

regional coupling mechanism influences metabolic activities in the lake,

and (4) document the principal energy flows that contribute toward

making the watershed a cohesive ecological unit.












REGIONAL SETTING


The Izabal Watershed extends 205 km eastward from the interior of

Guatemala to the Caribbean coast. The orientation of this 50-km wide

strip of terrain is east-west, located at 88041' W to 90034' W and

15003' N to 15052' N. The drainage pattern is from the highlands in

the western sector, downward to Lake Izabal only a few meters above

sea level, and then through the Rio Dulce, finally reaching Amatique

Bay in the Gulf of Honduras on the Caribbean. However, for the purpose

of this study, the watershed will not include the area downstream from

Lake Izabal, but only the region (6,860 km2) that drains into the lake

(Figure 1).

Two mountain ranges delimit the Izabal hydrologic unit; the Sierra

de las Minas and the Montanas del Mico lay end-to-end and parallel the

southern boundary, while the Sierra de Santa Cruz and Sierra de Chama

create the divide to the north. The major rivers are the Rio Polochic

and the Rio Cahab6n, which converge into a massive river that is actively

building a delta across the end of the lake. Other rivers and small

streams, about 40 in all, drain less extensive areas and often are inter-

mittently dry during periods of low rainfall.


Climate


Lakeside Temperature, Rainfall and Solar Radiation

The seasons associated with rainfall and temperature changes near Lake

Izabal were described by Snedaker (1970). The wet season begins abruptly

9



















Figure 1.- The Izabal Watershed. Rainfall stations and geographical features of regional
importance are labeled.






















o R.O.nll




So~dy A,.e, loPdoay

--- S-11I WogIr.Ihd Bo-nd-.I.s


R.Infoll i.



0 10 20 30 40 50KM.


Amofilue Bay


___


----------i


S.err, a* C*om6









sometime in May and tapers off toward the end of the year; the dry sea-

son is from December through April. The daily temperature range is

greatest during the drier months, reflecting the ameliorating effect of

rainfall. The average monthly temperature range is 4.4 C over the year

and the annual mean is 25.2 C. The lowest temperatures generally occur

in December and January and are associated with cold air masses moving

in from the north during the winter.

Finca Murcielagos, located on the north-central shore of the lake,

received an annual average of 2,004 mm precipitation over a 6-year per-

iod (Snedaker 1970). Rainfall increases to the east and the west of

this location. No month is without rainfall, while July is the wettest

and February the driest month. August has less precipitation than

either of the adjacent months which coincides with the caniciilas (dog

days), a term that describes a week to 10-day period of dry and cloud-

less days. Snedaker reports that during his 6-year study, 45.5% of the

days experienced at least 1 mm of precipitation, and more precipitation

tends to fall at night (ca. 71% of the total).

Lake Izabal has a pronounced local effect on the climate, especially

the solar radiation. Mornings are generally cloudless in the lowlands,

but in the lake area, clouds begin to accumulate over the surrounding

terrestrial region from about 10 a.m. to noon. Differential solar heat-

ing over the land causes uplifting convective air currents and condensa-

tion of the water vapor while the sky above the relatively cooler lake

remains clear. About mid to late afternoon these clouds move horizontally

across the lake. Snedaker (1970) recorded hourly values of net incident









radiation during 1964 and calculated the total daily net radiation

received for May and June at 408.1 and 494.1 langleys/day, respectively.

The climate near the lake during the year of study (October 1971-

1972) was characterized by examining records maintained by the EXMIBAL

mining company on the north and south shores of the lake. The records

included daily rainfall and maximum and minimum temperatures from the

north shore at Las Dantas, and daily rainfall from the south shore at

Mariscos, 32 km to the east-southeast. Average temperatures were cal-

culated from the maximum and minimum temperatures by interpolation using

the same relationship that Snedaker (1970) observed from his hourly

readings at Finca Murci6lagos. Figure 2 shows that the two warmest

months were April and May during the dry season. The decrease in tem-

perature after May can be attributed to amelioration by rainfall during

the wet season. Temperatures increased until September and thereafter

decreased until February. The low November temperature, which interrupts

an otherwise continual decrease from September through February, can be

attributed to an unseasonably cool,week-long persistence of stormy

weather caused by a hurricane on the Atlantic coast.

Mariscos received 3,236 mm rainfall during the year of study, which

is considerably more than the 2,210 recorded at Las Dantas (Appendix,

Table A). The month of heaviest precipitation for both stations was

August (630 mm at Mariscos and 426 mm at Las Dantas). The rainfall at

Las Dantas during the year of study is in close agreement with a five-

year average (1963-1967) for the same station and slightly higher than

the 2,004 nmm average from 1961-1967 at Finca Murci6lagos (Snedaker 1970).

The climate diagram, which uses the conventions of Walter and Leith

(1960-67), illustrates the monthly march of precipitation and average















Figure 2.-


Monthly averages of the maximum, minimum, and calculated
mean temperatures at Las Dantas for the period of study
(October 1971-1972).




















































J F M A M J J A SO
MONTH


O 25

F-
L



I-^-
cii
L


E
H-









ambient temperature (Figure 3). Rainfall is the average of the records

at Mariscos and Las Dantas during the year of study and temperatures are

those of Snedaker (1970). The shaded area above 100 mm represents months

when rainfall was in excess of evapotranspiration, and is the period when

the most surface runoff can be expected. The stippled area represents

the period when evaporation is greater than rainfall, implying a water

deficit.


Precipitation for the Remainder of the Watershed

The low topography near the eastern part of the lake allows trade

winds to pass unobstructed into the basin. Orographic rains occur as

the moisture-laden air masses sweep inland from the Caribbean and upward

over the Polochic Valley. There is considerable spatial variation in

rainfall as illustrated by the isopleths in Figure 1. The data on

which these isopleths are based appear in the Appendix, Table B and also

include information from the Instituto Geografico Nacional (1966).

Heaviest rainfall, averaging close to 4,000 mm, is concentrated in the

north-central region of the Polochic Valley. Radiating from that area,

the average decreases to values below 2,000 mm. San Juan received 6,128 mm

in 1969, the highest amount recorded for any station in a single year.

To arrive at an estimate of the average rainfall for the Izabal

Watershed exclusive of the lake, the areas between the estimated rain-

fall isopleths (Figure 1) were determined planimetrically. For the year

of study, average rainfall for the watershed as a whole was calculated at

2,992 mm.















Figure 3.- Temperature-rainfall climate diagram for Las Dantas and
Mariscos during the year of study. Average monthly
temperatures were calculated from a 6-year record at
Finca lurcielagos.























30 7 / 60


oC20- 0
1 O- -20

0- -0
Sf mamj j a s o nd

Month









Geology and Soils


The Izabal Watershed lies in the physiographic province known as

the Central American Mountain System, just to the north of the volcan-

ically active Pacific Cordillera (Walper 1960). The orientation of the

rivers in the watershed is controlled by the east-west faulting and

folding which has produced a series of anticlinal mountains. This major

zone of faulting,in which the Polochic Valley lies, is postulated as

being tectonically related with the fault zone of the Cayman trench

(Bartlett trough) in the northern Caribbean (Walper 1960). The lake

occurs in a block fault basin (Dengo and Bohnenberger 1969).

The Rio Polochic originates some 2,100 m above sea level and passes

a distance of 100 km before reaching the lake. The headwaters lie be-

tween the Sierra de Pansal to the south and the Sierra de Xucaneb and

Sierra Tzalamila to the north (Popenoe 1960). To the north of these

two latter confluent ranges, the Rio Cahab6n begins its eastwardly flow,

finally connecting with the Rio Polochic 53 km downstream. These waters

become distributed in the multilobate Polochic Delta before they dis-

charge into the lake. As levees of the distributaries protrude into the

lake and deposit alluvium, shallow coves and lagoons become isolated in

the delta region, providing interesting ecosystems for the study of

seasonal succession and metabolism of the plankton (Brinson 1973).

The structure and stratigraphy of the watershed is complex and in-

completely understood (Walper 1960). West of the lake, and at higher

elevations, prominent cliffs of sedimentary limestone and interbedded

dolomite constitute massive beds of Permian age (Roberts and Irving 1957).

Nearby at Cahab6n, beds of terrestrial conglomerate and sandstone predominate









and are believed to be of Tertiary age. Igneous rocks occur in lenti-

cular arrangement along a fault which parallels the north shore of the

lake. This serpentine area is believed to be of late Paleozoic and

early Mezozoic age. The mountains that parallel the southern area of

the watershed are metamorphic pre-Cambrian rocks consisting of undif-

ferentiated schist, gneiss, phyllite, quartzite, and marble.

The soils of the Izabal Watershed, as all the soils of Guatemala,

have been classified and mapped by Simmons et al. 1959. Soils around

the north and east perimeter of the lake were examined by Tergas (1965)

in relation to the primary production of natural vegetation. Some of

these soils have high ratios of calcium to magnesium (1:1.2 to 1:6) as

a result of their derivation from serpentine rock. Popenoe (1960) in

his remarkable study of the response of soils of the Polochic Valley to

shifting cultivation (slash-and-burn), described many of the soil pro-

perties. He stated that "Erosion is very slight on the steep lands of

the Polochic Valley, probably due to excellent soil physical conditions"

imparted by the relatively low bulk densities of the topsoils. Consider-

ing the diversity of parent material and the extremes in climatologic

regimes of the watershed, a more complete description of the soils would

be inappropriate and would necessarily include inaccurate generalizations.

The lithology of the Izabal Watershed is a heterogeneous one,

ranging from old metamorphic to relatively young sedimentary rocks. With-

in this mosaic, one would expect to find significant spatial variation in

weathering and differences in the ionic composition of the headwaters of

streams. However these local irregularities can be expected to cancel

each other as tributaries converge, so that the downstream parts of the

rivers will tend to resemble one another.








The Lake


Lake Izabal is located between 15024' N to 15038' N and 88058' W to

89025' W and lies only a few meters above sea level. The central basin

of the lake is a broad, nearly flat plain reaching a maximum depth of

about 16 m near the center. Thus most of the volume of the lake is be-

low sea level. Most of the shallow areas are the coves and lagoons at

the west end of the lake, bordering the delta of the Rio Polochic. Accord-

ing to the bathymetric map (Figure 4) adapted from Brooks (1969) only

9.7% of the area of the 717 km2 lake is less than 4.6 m (15 feet) deep;

50% of this occurs in the shallow areas near the delta, and the remain-

der around the perimeter of the lake. The volume of the lake is 8,300 x

10 m 3 giving it a mean depth (volume/area) of 11.6 m.

Nearly 50 streams and rivers flow into the lake, and while the

greatest number flow into the eastern, northern, and southern edges, the

greatest volume is received through the Polochic Delta to the west. The

outlet to the lake is at San Felipe, where the lake water enters to Rio

Dulce on its flow to Amatique Bay in the Gulf of Honduras on the Caribbean

Sea. About midway between San Felipe and the coastal port of Livingston

(a distance of 42 km) the Rio Dulce broadens into a large shallow area

(4.5 m depth) known as El Golfete.

Tsukada and Deevey (1967) suggested, on the basis of sediment cores

that ended in sand and gravel, that lacustrine conditions were established,

or reestablished, relatively late during the time of ecstatic rise of sea

level. Brooks (1969) speculated that the Izabal Basin originated as long

ago as the Miocene, and pointed out, on the basis of his inability to




















Figure 4.- Bathymetric map of Lake Izabal illustrating sampling stations. A, A-B, B, B-C,
and C are the locations of sampling stations on the lake. Adapted from map by
Brooks (1969).
























SB-C J

A-B S -- 1







Manscos
*- ^ LLake Izabal
Guatemala





Scole
0 2 4 km
Depth in Feet









find salt traces in the interstitual waters of sediment cores, that

marine conditions have been absent in the recent past.

The shallow littoral zone of the lake is narrow and subject to the

abrasive action of waves. In some of the protected areas, especially

along the north and east shores, the forest grows to the water's edge.

The shallow bottom consists of large pebbles where the submerged aquatic

macrophyte, Vallisneria, grows in sparce densities. Sandy beaches are

more common along the south shore where there is greater exposure to

wave action created by prevailing northeasterly winds. Occasionally

isolated individuals or aggregates of water lettuce (Pistia stratiotes)

washed in from the shallow lagoons and black-water creeks of the delta

region can be seen floating on the surface of the lake.

A limnological survey of the lake was made by Nordlie (1970) in

August 1969 and March 1970. His most surprising discovery was the rela-

tively high densities of a member of the Tanaidacae, a bottom-dwelling

crustacean with marine affinities. Nordlie found the planktonic community

to have only moderate primary production. On his March visit he observed

extremely low densities of net phytoplankton, whereas in August, they

were present in "bloom" densities. Since most of the present study con-

tains detailed seasonal descriptions of the same phenomena that Nordlie

observed on his visits, his findings will be discussed in more detail in

the chapters that follow.


The Fisheries and People


The Izabal Basin has boen,until recently, a region of only modest

human activity. The lowlands near the lake were believed to have been









settled sparsely during Maya times (Voorhies 1969). Major ceremonial

centers were completely absent from this area although there were cen-

ters immediately to the south (Copin) and to the north scattered through-

out the Pet6n. During the current century, the population did not in-

crease substantially until malaria control was available and a road was

completed from Cobdn to El Estor in 1948. This route replaced, at least

during the dry season, the older route up the Rio Polochic to Panz6s

where the Verapaz Railway connected with Pancajche. A road terminating

at Pancajche completed the journey to the central highlands of Guatemala.

Prior to this, 19th-century ships sailed in from the Caribbean to the colonial

town of Izabal on the south shore of the lake. From this location, mule

transportation provided an overland access to the central highlands. In

the last decade, a spur road was completed, connecting the village of

Mariscos on the south shore with the Atlantic Highway that couples Guate-

mala City with Puerto Barrios and Puerto Santo Tomas de Castilla on the

Caribbean. Now the people living around the perimeter of the lake who

have access to dugout canoes with outboard motors can reach the road

terminating at Mariscos. From there the bus trip to Guatemala's capital

city lasts only 4-5 hours. Likewise, the access provided by the lake to

its perimeter opened the slopes of the Izabal Basin to agriculture.


Immigration and Labor Alternatives

The influx of Kekchi Indians from the upper Polochic Valley, and of

people from other parts of Guatemala, was initiated by the discovery) of

a rich nickel deposit near the northwestern region of the lake. All the

exploratory work for mining the ore has been completed, but the construction

of the processing plant for the extraction of the miiner;al is still peddling









financial support and legal agreements. The already massive capital

outlay, made possible through subsidization by North American firms,

has dwarfed other business interests in the valley. In the past 0O

years the town of El Estor has grown from a sleepy Indian village to a

bustling community.

The labor force required for the initial clearing and exploration

of the mining area is now largely unemployed. Although some families

have been forced to leave, many remain with the hope that they will again

work for the attractive wages paid by the mining company. Some of these

desperate people have shifted their means of living to agriculture and

fishing. Carter (1969) in an anthropological monograph on the Kekchi

cultivators described the problems and successes of these Indians in

their efforts to apply highland methods of shifting cultivation to the

lowland areas near the lake.

The fishing, which became more effective with the introduction of

nylon gill nets in the early 1960's, has since become so unprofitable

that it now offers employment and income for only a few dozen individuals

around the lake. Holloway (1948) recognized long ago the potential food

source of the predominately marine fishes that inhabit the lake, but

made no prediction as to what the carrying capacity of this resource

might be.


Fishing Regulations and Yields

Current Guatemalan fishing regulations, passed into law in 1936,

apply to all inland waters regardless of size or location. One of the

restrictions of the law is that gill nets can be no greater than 36 meters

in length nor have a: mesh si-e of less than 7 inches. However, gill nets









as long as one kilometer with 2 1/2-inch mesh are frequently used during

the fishing season. Initially, excellent yields from the relatively

virgin fisheries were the incentive for people with capital to invest in

this equipment. However, the subsequent decline in yields possibly may

have been the result of exceeding the carrying capacity of the fisheries.

The fishermen react to the decreased yield per unit effort in

several ways. Some cease fishing altogether, while others send their equip-

ment and hired labor northward to fish in the Rio de la Pasi6n in the

Peten when the'demand preceding Easter forces prices upward. Most owners

of gill nets have several thousands of dollars invested in equipment, but

fishing is usually subsidiary to their main business interests.

The general disregard for the irrelevant and unrealistic laws

favors individuals capable of making large capital investments and pe-

nalizes to exclusion those individuals fishing at a subsistence level.

Dickinson (in press) has discussed in detail the sociological implica-

tions of the fisheries as well as thoroughly documenting the geographi-

cal features of the region.

Since there is no governmental agency to keep records of inland

fish catches, past yields from the lake cannot be estimated. Carr (1971)

estimated current yields during a month-long study by observing the

weight of catch per unit length of net. He calculated a daily mean fresh-

weight of 11.6 pounds per 100 yards (5.75 kg/100 m) of net, and based

on other assumptions and measurements, estimated the lakeside yield to be

2,808 kg dry weight per week during the 1970 season. Some fishermen



Most of the catch is salted and dried, accounting for 59% loss of fresh
weight. They arc marketed beyond the I;t abal BRasin, iad only a small
percentage of the total catch remains for local consumption.









fish year round while others may be active only when fish prices are

high (2 or 3 months of the year). If an average fisherman worked 5

months of the year, then the lake would yield 56,160 kg dry weight

annually (0.071 g/m" yr).


Fish Behavior and Fisheries Management

No data exist on the standing crop of fish, on growth increments

of the population or on recruitment from immigration. Since much of

the catch consists of euryhaline marine species, immigration is probably

an important factor to consider in fisheries management. If the fisher-

ies resource of the lake is being overexploited, as it presumably is

(Carr 1971), then the immigration route through the Rio Dulce--El Golfete

region is a logical area for control measures. Disgruntled fishermen

that fish ony n on the lake are aware of this migration route, and of the

gill nets set across these routes by their counterparts in the Rio Dulce

region.

Fishermen as well as large flocks of cormorants, terns, gulls, and

scattered pelicans find another popular fishing region in the shallow

coves and black-water lagoons of the Polochic Delta. There the fishing

is seasonal for both man and birds. During the dry season, these waters

are stagnant and perpetually in bloom with high densities of phytoplankton

(Brinson 1973). The increased consumer activity, in response to the

highly concentrated food source, coincides with the season when fishing

is legal. As law dictates, the fishing ceases in June when heavy rains

mark the beginning of the wet season. Ironically, these events coincide

with the migration of fish to the open water of the lake where they be-

come more dispersed and harder to catch. Fish-eating birds also disperse,









and only a few pelicans remain. The fishing laws thus present a paradox

by allowing fish to be caught when they are easiest to catch, and by

timing the open season with the high prices preceding Easter which

provides incentive for economic gain.

The black anaerobic waters that flush the delta are not completely

devoid of fish. Gobies (Gobionellus sp.) which are particularly poor

swimmers, can be seen gulping air under the sudd vegetation bordering

the lagoons. Vultures and egrets prey upon these fish, but the main

predators are schools of large tarpon (Megalops atlantica) that travel

the backwaters of the delta. By being facultative air breathers, the

tarpon are well adapted to the anaerobic environment.

Marine fish, such as the tarpon, make up the majority of the fish

caught in the lake. In order of decreasing yield they include Chloroscom-

brus sp. (zapatero or leather-jacket jack), several species of catfish

including Bagre sp. and Arius spp. (vaca and chunte), and the prized

Centropomus undecimalis (robalo or snook). The presence of schools of

sardines, anchovies (Anchoviella), and other small herbivorous fish

apparently provide much of the food for the larger carnivores. The impor-

tant fresh-water fishes include Cichlasoma gutulatum mojarraa) and

Brycon g'uatemalensis (machaca).

Some other predominately marine animals, although not directly

important to the fishing economy, are conspicuous components of the

fauna. Blue crabs (Callinectes sp.) are occasionally caught in nets, and

the barnacle, Balanus improvisus, attached to pilings in the eastern end

of the lake are testimony of the seasonally brackish water in that area.

Porpoises (Tursiops truncatus) apparently do not enter the lake, but do









follow the front of high-salinity water as far as the upper Rio Dulce.

Man has noticably reduced the abundance of some of the large aquatic

vertebrates which probably has altered aquatic food chains. Both shark

(Carcharhinus leucas) and sawfish (Pristis perotteti and P. pectinatus)

are reported to inhabit the lake (Thorson et al. 1966) although their

presence has gone unnoticed for the past 8 to 10 years. Fishermen assert

that nets frighten sharks, and that they have been absent from the lake

since gill nets became prevalent. Crocodiles (Crocodylus acutus), once

conspicuous carnivores of the aquatic community, have suffered consider-

able reduction of their populations as a result of hunting pressure.

Hunting may have also been responsible for reducing the manatee, a large

herbivore, to its present day low population density. Two species of

turtles, Dermatemys mawi and Pseudemys scripta ornata, are occasionally

caught by baited hook or by gill nets. However, they seem to be rela-

tively abundant and could serve as a potential source of food for the

local human population (Carr 1971). Based on the consequences of exploita-

tion of other components of the aquatic community, the "potential food

source" offered by turtles would be short-lived, at best. Undoubtedly,

the boney fishes will continue to be the principal aquatic resource for

human exploitation.

Considering the migration routes and seasonal activity of the fishes,

it is doubtful if enforcement of the law during the veda, or prohibition

period, would significantly relax fishing below its present intensity.

Complete enforcement of the law, which would limit the length of gill nets

to 36 meters, would paralyze the fisheries completely and be socially

dysfunctional (Dickinson, in press). There is an urgent need for fisheries





31



management on Lake Izabal. Any management, however, must demonstrate

an understanding of the role of the fishes in the ecosystem as well as

the needs of the commercial and subsistance-level fishermen.

Undoubtedly some sacrifices in the habits of the fishermen would

be necessary before the fisheries could reach a steady-state level of

maximum sustained yield. Proper management of the fisheries may neces-

sarily extend beyond the boundaries of the aquatic ecosystem if regional

coupling mechanisms are operative as hypothesized in the Introduction.













HYDROLOGY AND WATER CHARACTERISTICS


The principal objective of monitoring the hydrological regime of

the Izabal Watershed was to enable the calculation of rates of inflow and

outflow of organic matter to and from the lake. Subsidiary to this pur-

pose was the need to characterize the hydrologic properties of a lake

that historically has received almost no limnological attention until

recently (Brooks 1969; Nordlie 1970). Even these recent studies lack

the perspective of a long-term study necessary for a fuller understand-

ing of the ecological implications of a seasonal hydrological regime.

For example, Brooks (1969) labeled the likelihood of brackish water

penetration into tne lake as a "common misconception" although its oc-

currence is common knowledge among the non-scientific local inhabitants.

Other "anomalies" of the hydrology and water characteristics might be

of ecological significance, and their occurrence, if undetected, would

unknowingly detract from a more complete understanding of the Izabal

Watershed ecosystem.

Information on runoff characteristics of watersheds in the humid

tropics has been approached mainly by calculating the excess of pre-

cipitation over evapotranspiration from empirical formulae applied to

rainfall and ambient temperature. Direct measurements of runoff are few,

and the information presented in the following chapter should be valu-

able for comparison with studies that do exist.









Methods


General Methodology--Field Stations, Logistics, and Schedule

The sampling stations are indicated in Figure 4 and encompass nine

river-mouth stations, three main lake stations, and the outflow station

at San Felipe. To achieve the objective of the study of estimating the

lake's organic matter budget as modified by seasonal changes, samples

were collected during both the dry and wet seasons. The dry-season data

were collected during the months of March, April, and May. The clear,

cloudless days during the warmer dry season helped to distinguish it

from the wet season. January, February, and June through October were

considered wet-season months because the average local precipitation was

greater than 100 mm per month.

The river-m south stations were sampled to determine the qu'aity and

quantity of water entering the lake and the outflow station was sampled

to determine the quality and quantity of exports. Lake stations A, B,

and C were sampled for chemical oxygen demand (COD) and biological oxygen

demand (BOD) as well as for light and dark bottle primary production

determinations (Table 1). Concurrently with COD collections, the follow-

ing data were recorded from surface and bottom samples: dissolved oxy-

gen concentration, pH, total alkalinity, and temperature.

Net plankton was collected from Stations A, B, and C at approximately

three-to four-week intervals. On these collection trips temperature pro-

files and Secchi disk transparencies were determined at the three sta-

tions as well as at stations midway between them. These intermediate

stations are referred to in the text as A-B and B-C.

The large size of Lake liablil and the lonI d istnc .s between s.ampi il.

stations required the use of a relatively fast and reliable mode of












Table 1.- Sampling dates for chemical oxygen demand (COD), biological
oxygen demand RBOD); ;Ind primary productivity experiments
(Prod). Numbers represent day of month



WET SEASON DRY SEASON
January February March April
Location COD BOD Prod COD BOD Prod COD BOD Prod COD BOD Prod


Station A 3 11,29a 2a 28 4a 21 10 24

Station B 3 28 21

Station C 28 21

San Felipe 3 8 28 21

Rio Oscuro 25 10 11 1

Amatillo 25 10 11 10

El Padre Cr. 17 10 11

Rio Polochic
Comercio 25 10 11 1
Coban 17 10 14 11
Bujajal 17 10 14 11 1

Rio San Marcos 28 21

Rio Manacas Cr. 28 21

Rio Sauce 28 21 10

Rio Tunico -

Largartos

SPrimary productivity experiments were not duplicated.
Date during month of September.











Table 1.- extended


May
COD BOD Prod


10,19 28

19 3

19 5,29















10

19

19


June
COD BOD Prod


20

5,19

18


WET SEASON
July
COD BOD Prod COD


August October
BOD Prod COD BOD Prod


5,23

5

5

5

1,23

1





1


1
1,23
1

5

5

1


27,27

26b

25b


- 10 -









transportation. For this a 25-hp outboard motor was brought from the USA

and a 24-foot dugout chance (25 hundredweight capacity) was purchased at

the lake. Routine sample collection required approximately seven hours

for stations at the western end of the lake in the Polochic delta. For

the remaining stations to the east of El Estor, the samples were col-

lected on another day as ten hours were required due to the longer dis-

tances involved.

Ice could be purchased for storing samples during collection trips

except during July and August when the ice plant closed for a minor re-

pair. Several weeks of research time were lost due to breakdowns of the

outboard motor and the car, which was used to transport equipment to the

dock. At least monthly trips to Guatemala City were necessary to pur-

chase repair parts and laboratory supplies.

A room of a concrete-floored house approximately ]/2 km from the

dock site was used for the laboratory. Electricity was available three

or four hours a day after dusk from a privately owned diesel generator

whose output ranged between 85-105 volts A.C. The low and variable

voltage was insufficient to operate the spectrophotometer in spite of

the constant voltage supply transformer. For this reason, chlorophyll

determinations were not made.


Hydrologic Measurements

Discharge rates of several of the major rivers entering the lake

were measured at monthly (occasionally twice-monthly) intervals from

November 1971-October 1972. The float-and-dye method (Welch 1948) was

used to determine the velocity by sprinkling fluorescein dye powder on








the water surface and recording the time elapsed in traveling between

two floats 50 m apart.

When the velocity was slow (less than 0.3 m/sec) a 25-m distance

was used. This procedure was repeated three times and the velocity was

estimated to be the mean of the three readings. Measurements were deter-

mined far enough upstream from the river's mouth to avoid backforce

from the river entering the lake. The cross-sectional areas of the

rivers were measured by determining the width and five depths (two near

the banks, one in the middle, and two halfway between these). A nail was

driven into a tree trunk on the river bank at each station to provide a

permanent point of reference for measuring changes of level in the river.

Discharge rates (cross-sectional area multiplied by velocity) for all

rivers, with the exception of the San Felipe outflow, were multiplied by

a factor of 0.8 in order to correct for frictional resistance of the

stream bed and banks (Welch 1948). At the San Felipe outlet where fric-

tional resistance is small because of the large cross-sectional area, a

factor of 0.9 was used.

Lake levels were recorded at frequent but variable intervals through-

out the study year on the municipal dock at El Estor. Two other stations

served as level markers on the lake: one on the western end at the mouth

of Rio El Padre Creek, and the other at the eastern outlet at San Felipe.

The EIXIBAL mining company also recorded lake levels at the plant site

2 km west of Las Dantas.

Temperature profiles were recorded with a YSI Model 51A oxygen meter

and a YSI 5419 oxygen/temperature pressure-compensated probe with a 50-

foot lead. Calibration was performed in the field for each profile, using

a mercury thermometer as the standard.









Analyses of Water Characteristics

Samples collected for analyses of pH, total alkalinity, and speci-

fic conductance were usually the same as those used for COD analyses.

Methods of collection are described in the section "Metabolism and

Organic Matter." Samples for 02 determinations were collected with a

2- or 3-liter Van Dorn bottle, transferred to 300-ml BOD bottles, and

fixed in the field. Temperature was recorded with a mercury thermom-

eter while the water was in the Van Dorn bottle. All other procedures

and determinations were made in the laboratory, usually within seven to ten

hours from the time the first sample was collected. Total alkalinity

and pH were determined first using a Beckman Model N2 pH meter and Beck-

man glass electrodes. Calibration was performed with factory-prepared

buffer solutions (pH 7.0 and pH 4.5). Total alkalinity (carbonate

plus bicarbonate) was determined by titration of duplicate 100-ml aliquots

to pH 4.5 with a 0.02N HC1 solution. The pH meter scale could be read

to an accuracy of 0.5 pH unit.

Specific conductance was determined with a Beckman Model RB3 Solu-

Bridge and readings were adjusted to 25 C. Two probes were necessary

for the range of conductivities encountered. A small, more sensitive

probe was used in the range of 50-800 pmho/cm and a large probe, one-

tenth as sensitive, was used in the range of 700-40,000 Imho/cm. Fig-

ure 5 illustrates the calibration curves determined with dilutions of a

standard KC1 solution of known specific conductance (Golterman 1969).

Readings of lake water, all within the sensitivity range of the small

probe, were corrected by adding 20 umho/cm, based on the calibration

curve. Higher conductivity readings, accomplished with the large probe,



















Figure 5.- Calibration curves for the small and large probes used for specific conductance
measurements between 100 and 30,000 pmho/cm at 25 C. Small probe readings
(magnified on insert) were corrected by adding 20 pmho/cm to the meter readings
between 50 and 700 pmho/cm. Large probe readings were left uncorrected.





40











(wL/oqwrlj) O1 pJepueIls




0 00
8 o


0O
~~-0



--


o 0
,- E ,. 0

I "'

O 10



GO O ,
-0
0 O"
) 0)











(w:/oNwH) 1:M pJepuelS
SrO





\ N



(w/o/oLurI) I3Y1 pjepue-4s








required less precision since detecting differences in the salinity

gradients was more important than acquiring absolute values. Neverthe-

less, the calibration curve is in close agreement with the standard KC1

solutions (Figure 5).

Samples for mineral analysis were collected April 14 and October 23,

1972. The samples were prepared by filtering the water through membrane

filters (47 mm-diameter, 0.80-p nominal pore size) to remove the seston.

The samples were transferred to 0.5-liter bottles and 1% formalin was

added as a preservative. The April samples were stored in polyethylene

bottles and the October samples in amber glass bottles. Both groups

were flown to Gainesville and mineral analyses were performed during

March 1973. An atomic absorption spectrophotometer (Perkin-Elmer 303)

was used for determination of Ca, Mn, Mg, Si, Fe, Ni, Zn, and Al.

Potassium and Na were analyzed with a flame emission spectrophotometer

(Beckman DU), phosphorus by colorimetric determination with molybdate

(Golterman 1969), NO, and C1 with specific ion electrodes in conjunction

with an Orion Potentiometer Model 801.


Results and Discussion


River Discharge and Runoff

Discharge rates from six rivers were used to estimate runoff from

the Izabal Watershed. All values were calculated from velocities and

cross-sectional area measurements of the rivers (Appendix, Table C)

except for the values recorded for the Rio Polochic distributaries and

the Rio San Marcos in August. These low August values are noteworthy

since they occurred during the month of heaviest rainfall. During this




'S


period the level of the lake had risen to as high as 1.15 m (August 21)

above the lowest level recorded during the preceding dry season (Figure

6) and resulted in flooding of the delta areas. During flooding the

rivers were not confined to channel flow, but moved as a sheet across

the deltas. If August discharges had been calculated from velocity mea-

surements in the river channels, gross underestimates of runoff would

have resulted. To estimate the August runoff, linear regression formulae

were calculated from the other measurements using rainfall at Las Dantas

as the independent variable and monthly discharge as the dependent vari-

able. In this way August discharges could be calculated by extrapolation

from rainfall, assuming a linear relationship between discharge and rain-

fall. Values for other months were similarly calculated where data were

missing (e.g. February). These extrapolated values appear as open cir-

cles in Figure 7 and were used for the August values in Figure 8.

Discharge rates at the mouths of the Rio San Marcos and Rio Sauce

were low compared to the other rivers measured (Figure S). The similar

patterns of discharge for the Rio Polochic distributaries (Comercio,

Coban, and Bujajal) can be attributed to their common origin. Whereas

the Polochic distributaries demonstrated a sharp increase in discharge

during June, an increase of similar magnitude for Rio Oscuro did not

occur until July. This can be attributed to differences in local rain-

fall regimes. During the year of study the wet season in the arel of the

lake began in June,which accounted for the July increase in Rio Oscuro's

discharge (Figure S). The wet season in some regions of the Polochic

Valley began in May,which may have accounted for the June increases in

the discharges of the Polochic distributaries.














Figure 6.-


Monthly estimates of all inputs to the lake (runoff and
direct rainfall) and the monthly averages of lake water
levels. Lake levels are the number of meters above the
lowest recorded week which occurred during the second
week of March 1972.

































Lake
Lake


0
bo
Level (nm)















Figure 7.-


Linear relationship between monthly rainfall and monthly
discharge of rivers emptying into Lake Izabal. The open
circles represent extrapolated values for February and
August for which discharge data were missing or required
correction.

































200 400 600 800 0
Discharge (m3 x 106/month)














Figure 8.- The seasonal march of monthly discharge rates for some
major rivers emptying into Lake Izabal.


















Oscuro


S dry
season


Comerc io


J F M A M J J
MONTH


A S O N D


0 1



C,-)

E



(T
U




ifl


200









Table 2 summarizes the discharge rates of all rivers and water-

sheds draining into Lake Izabal. Included is runoff from the water-

sheds to the north and south of the lake which was estimated from the

runoff of the Rio Sauce and Rio San Marcos watersheds. The annual dis-

charge from rivers emptying into the extreme western end of the lake

was greater than ten times the contribution of the remaining watersheds.

The total runoff volume of all watershed areas into Lake Izabal was es-

timated to be 13,290.9 m x 106 during the year of study.

Instead of expressing runoff as a volume, it can be compared directly

to rainfall by conversion to equivalent units. This was accomplished by

dividing the known runoff volume (m 3) by the area of the watershed (m 2)

(Table 3). The annual runoff value of 8,253 mm calculated for the Rio

Oscuro watershed was much toohigh since it is unlikely that rainfall

ever reached this value in any part of the watershed. The reason for

this overestimate was that at flood stage, the Rio Polochic apparently

overflowed into the southern branch of the Rio Oscuro headwaters

(Riachuelo Suncal). This resulted in high discharge rates for the Rio

Oscuro due partly to water originating from the Polochic watershed. By

combining the Rio Polochic and the Rio Oscuro areas (5,480 km ) and their

annual runoff volumes (12,112 m3 x 10), the combined runoff would be

2,210 mm, a more realistic value.

Runoff for all watersheds averaged 1,937 mm. Since the average

rainfall for the Izabal Watershed was 2,992 mm, the portion of the

precipitation lost as runoff was 65%. This is in close agreement with

the watershed of Lake Lanao, Phillipines (Frey 1969) which loses 67% of

the 2,873 mm rainfall it receives. Other tropical watersheds receiving











o'i -o o o'i o


t ro t )O o








N t LO 0
'-4-4.




V b,- Ll


- G t- 00 C'
~~bcCNJ


301 N





ts 00 L -




r moo l

c"J -4 1-4


CN co 0

cc r-
1-I Ln


oG ri-Ln CIA Lfn

01 C -O 4 -
-o4 1-1
-cObb


n C \o r \
1-1 CIAi --1 -1


tf
,-







c-,








0
























U u
0.)


C)
0rt






























CD u
ocA
.HE














0
o d
*i l
kai












'-' (


m T

4-) 1- C)'






4-) C 00




0

;-4
I)









0)
co




Ui)
U /-'

3 \
CT -
LD


iLfll
oLDo0
OTwou


r-lCD CO
00 o


LI)


L V)
OVO t


co

r 1


\0NC0C)tn






N O o -1 in

t i Ln co t -T


i)n 00

C r
--4 -
r~ O'


Co r~'
ON

oco
CT] r~


co ~ om
cO- 01 co 00

N 11r i
. . ..~N
CU O~


) Cq) i- n-1
McticNC


SI o o
\D ft L


04J > C- C C4J
Ozoo <'20m<'EO


00 LO \


- o 00 -
r
c o3


I 01
4-4

E-


-d
0







; C











-I
ct

*rl





































-1U
0





C


































-4
c



(O


























o




O 3














0C







6o
,c(
i-i
aO3
a^










*i-
1/1

(1) *T
?-


n N C


\oD CI C'


cl
~Jc'Jcc'
CO L\h I


(t
* * CD

OO Ct0


00
Sco





I
cc

,i


r- r~
. Co
r


in cr-O '

Lo i^


'T Lfl


coN
CO0 1
CIA


o
0









o


O
U




*0'--
rt






O
C-
U


0




O
u


Cl N-
NOT
b0


co t- a o

cq C) C:) 'T
l-I I- I r-f 1-
















Table 3.- Summary of hydrologic data for the major watersheds
draining into Lake Izabal


Watershed

Rio Polochic

Rio Oscuro

Rio Sauce

Rio San Marcos

North Watershed

South Watershed


All Watersheds


Area
(km2)

5,247

233

300

170

474

438


Volume of
Annual Runoff
(nm x 106)

10,189

1,923

257

144

406

372


6,862 13,291


Annual Runoff
(mm)

1,942

8,253

857

847

857

847


1,937









less precipitation lose between 40-50% of the rainfall as runoff (Golley

et al. 1971).

Snedaker (1970) calculated runoff for Finca Murci6lagos by a sim-

ple method devised by Holdridge (1967) which requires knowledge of only

mean annual biotemperature and annual rainfall. Using this method,

Snedaker's runoff estimate was 902 mm or 45% of the rainfall. This is

in close agreement with my value for the Rio Sauce watershed (Table 3)

which is located on the north shore of the lake near Murci6lagos. Dur-

ing the period of study, runoff was 857 mm or 39% of the rainfall (Las

Dantas records).


Water Budget

The water budget of the lake was calculated at monthly intervals

from inputs by river runoff (already discussed) and by direct precipi-

tation, in addition to the outputs by evaporation and losses though the

San Felipe outlet. The monthly contribution by direct precipitation was

calculated from the average rainfall at Las Dantas and Mariscos.

Since evaporation from the free water surface of the lake was not

measured, a literature search was made to arrive at a reasonable esti-

mate. Free surface evaporation estimates from tropical latitudes appear

to be few and the data presented in Table 4 include values for more

northern latitudes. However, summer climate regimes, especially in

Florida, approximate the year-round climate of the Lake Izabal region.

The value of 161 mm/month was chosen as the representative evaporation

by averaging the estimates from the humid areas of Lake lelene, Anderson

Cue, Lake Michie, Lake Chad, and the Caribbean Lowlands (Table 4). This















Table 4.- Monthly free water surface evaporation measurements for
selected warm climates and warm seasons


Region or Lake


Monthly
Evaporation
(mm)


Lake Helene, Floridaa

Anderson Cue Lake, Floridaa

Lake Elsinore, Calif.a

Lake Tiberias, Israela

Polish Lakesb

Lake Chad, Africac

Caribbean Lowlands

Lake Michie, N.C.a

Lake Colorado City, Texasa

East Africa (Nile Region)


128

155

231

193

140-180

188

174-254

118

221-251

90-120


1962 Pride et al. 1966

1966-68 Brezonik et al. 1969

Szeicz & Endr6di 1969

1949 Reiser 1969

several Debski 1966

Grove 1972

Ray 1931

1962-64 Turner 1966

1955 Harbeck et al. 1959

Talling 1966


asked on an average of 5 warmest months (y-September
Based on an average of 5 warmest months (May-AuguSeptember).
nuBased on an average of 4 warmest months (12.ay-August).
Annual total divided by 12.


Year


Source









is equivalent to 115 x 106 m /month for a surface the size of Lake Izabal

and will be considered constant throughout the year.

Changes in lake level (Appendix, Table D, see also Figure 6) re-

presented integrated results of both inputs and outputs. The volume of

the lake was calculated from the bathymetric map prepared by Brooks
6 3
(1969) and was estimated to be 8,300 x 10 m .Changes in volume were

calculated by multiplying changes between mean monthly lake levels by

the surface area of the lake.

Table 5 is a summary of the monthly contributions and losses for

Lake Izabal. Adding the inputs from runoff and direct precipitation,

subtracting the evaporation, and subtracting the positive or negative

change in volume resulted in the final value which estimates the monthly

loss through the San Felipe outlet. Direct measurement of discharge

from the San Felipe outlet was inadequate due to extreme variations in

surface velocity. On one day I even observed that the flow had reversed,

and further inquiry led to the conclusion that this was a common, if

not daily phenomenon. Apparently prevailing northeasterly winds are

capable of shifting the leeward level of El Golfete above the level of

the lake, thus generating the variation in discharge or reversal of

flow at San Felipe. In spite of this difficulty, there was sufficient

agreement between the velocities measured in the field and those cal-

culated (Figure 9) to regard the latter values as good estimates of

discharge at San Felipe. The direct field measurements were multiplied

by a factor of 0.9 to correct for frictional resistance of the banks and

bottom (Welch 1948).

The annual water budget for the lake and the watershed is summarized

in Figure 10. The average residence time of the water in the lake was





Table 5.- Summary of water budget for Lake Izabal (November 1971-October 1972)


Direct
Rainfallb


246.7
74.2
106.5
74.2
67.8
49.1
20.8
212.6
305.8
378.6
194.5
218.7

1,949.5


m3 x 10 /month

Evaporationc


-115
-115
-115
-115
-115
-115
-115
-115
-115
-115
-115
-115

-1,380


Change Lake
Volume


+122.6
-171.4
+11.5
-88.9
+23.7
+15.1
-38.7
+67.4
+302.6
+253.8
-249.5
-151.3

+96.9


Loss at Outlet
(San Feline1e


(SnFli,)


1,324.4
877.4
789.5
520.7
322.3
236.8
311.5
1,062.4
1,648.2
3,094.2
1,804.9
1,771.2

13,763.5


a
aAll watershed areas.
bCalculated from the average rainfall of two lakeside stations, Mariscos and Las Dantas.

SEstimated from pan evaporation from other localities (Table 4).
d
Calculated from level changes of lake (Table D, Appendix).
e The sum of co s a, b, and c, and the opposite sign of values in column d.
The sum of columns a, b, and c, and the opposite sign of values in column d.


Month


Runoffa


Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct


1,315.3
746.8
809.5
472.6
393.2
317.8
367.0
1,032.2
1,760.0
3,084.4
1,475.9
1,516.2

13,290.9


Total















Figure 9.-


Relationship between velocities at the San Felipe outlet
from direct field measurements and those calculated by
balancing the water budget. Perfect agreement between
the two estimates would fall on the dashed line. Direct
field measurements were multiplied by a factor of 0.9 to
correct for frictional resistance of the banks and bottom.





















__
U8-
U /

E ,


0
_.6- / 0

0 /
O /
:>A- /
> "

0 / 0

/
M / o

U /

0 .2 ,4 .6 .8
Measured Velocity (m/sec)





























V)
U-,
U-,
V)





0

U-,











'-4



-4
Olo
U-,
















0
4J





4-4







0
4-,






5




























U-.
C)



4-,


0


o





Cl .




C)













20,531 Ra
Rainfall Rain
SRainfall


1,950
Rainfall


7,240
Evapotranspiration


13,291
Runoff


8,300 1
Lake Evap


Outflow








63
calculated by dividing its volume (8,300 x 10 m ) by the 12-month loss

by direct evaporation from the surface and the outflow at san Felipe

15,143.5 x 10 m ). This estimated the residence time of the water to

be 6.6 months or 0.55 )ear.


Water Characteristics

Lake Izabal, like other lakes that lack sufficient depth to develop

a hypolimnion, would be classified as a third-class lake (Hutchinson

1957) and could fall, functionally at least, into the class of polymictic

which is usually reserved for lakes of high mountain regions in equatorial

latitudes. Because of Lake Izabal's large area, shallow depth, and high

influx of relatively colder waters from rivers, some anomalous and in-

teresting temperature patterns develop.

The Tzabal Watershed lies in an area of diverse geological forma-

tions, and as a result, receives a variety of water types. Partly be-

cause of the large drainage area of a single influent river, the Rio

Polochic, much of the water from the various rock types has already

mixed before discharging into the lake. Regardless, the large deltaic

swamp at the western end of the lake has distinct modifying effects on

some of these waters. The proximity of the lake to the sea coast and

its connection with a marine environment cannot be overlooked as a poten-

tial factor that could influence the ionic composition of this fresh-

water lake.


Thermal properties and circulation patterns in the lake

Temperature profiles were recorded at approximately monthly intervals

at five equidistantly spaced stations along the length of the lake (Figure









4). Figures 11 and 12 illustrate the temperature profiles for each of

the five stations (A, A-B, B, B-C, and C). For most stations and dates

of sampling, the profiles were isothermal for the first 10-12 m, with

the exception of the upper 2-3 m which were heated directly by solar

radiation. This surface stratification was only temporary and dis-

appeared rapidly either at sundown or with windy conditions during the

day.

More persistent thermal stratification often occurred in the

bottom 1-2 m. This was most pronounced at Station A after June when

the colder waters from the nearby Rio Polochic created a density current

resulting in a cooler layer along the bottom. At Station A-B this sea-

sonal change was more noticable, as the profile is isothermal until

June, the beginning of the wet season. Stations B and B-C showed the

same characteristics but the thermoclines were not as sharp as the more

westerly stations.

Station C was located 30 km east of the Rio Polochic and little

influence from the density layers was expected. The stratified layer

found on April 8 can only be explained if the 2 C increase in tempera-

ture of the upper column (since March 16) failed to circulate with the

bottom meter. More interesting was the noticeable increase in bottom

water temperature on June 13, due to the arrival of a warmer but denser

water mass of high specific conductance (465 mho/cm at 13 m and only

248 umho/cm at the surface). This warmer water originated from the

brackish conditions that developed at the San Felipe outlet during the

proceeding dry season (p. 89).

A dry-season increase in temperatures was recorded for all sta-

tions from a low in February to a maximum in June. For most of the














Figure 1].- Vertical temperature profiles of Lake Stations A, A-B,
and B recorded at approximately monthly intervals.






Depth (m)
Sr O OO ro
-I I 0 I ii N) 0






















~ 8
j-^-1^y


"o
K0





InDO
Zi R


Depth (m)
ar 0 ( ) .t ro
r i 1 ; I I I i r i r ,


\ 1



0











10
0,




iSCDO
~------^,-^-^j




-----------7|



-.___.___T
-\ -^
\ ''os
-i* * i^ -
\ -
0-.-. M

^ --------- --- ^

-v -g1'
\- 'o


Depth (m)
o 0 0) Z, M) o











-t -




]2: Y










'Il







CD0


?--- r




































C)
a

0







cT5

0
a,








41)
cT5
4-3


C)


-4



0)
>
















t4-4 4-)

A -4H



'--4




Cd
7d,


















*d 0

-4


























1110
3Jan
0 25 27
2 7
27

4

6-

8

10-

12

14


STATION
'345 1140 1024 (8503
25Feb 16ar OApr 23May
26 28 26 28 28 30 30 31
S I r1 '











i i
: ; T '


1429 1218
25 Feb 16 Mar
26 28 26 2


B-C
0905
13Jun
30 31
r"m


STATION C
1059 0925 0945
8Apr 23May 13Jun
27 23 0 31 30 32












6'


1007
19Aug
27 29


r


0900
17SD
29 30


1050 0930
19Aug 17Sep
29 30 29 31


Iyrr


1015
260ct
28 30
(-- -


1030
80ct
28 30









I









water mass, the minimum was approximately 25.5 C and the maximum 30.4 C,

a difference of 4.9 C.


The seasonal pattern of water characteristics

The water samples were arbitrarily divided into four groups--

lake stations (A, B, C and San Felipe), swamp waters (Oscuro, Amatillo,

El Padre), Polochic distributaries (Comercio, CobAn, Bujajal), and

small rivers (Sauce, San Marcos, Manacas Creek, TAnico). These groups

are based not on water characteristics per se, but on the origin and

locality of the water. The same divisions will be recognized in the

treatment of the organic matter data. The results of measurements of

pH, total alkalinity, specific conductance, dissolved oxygen concentra-

tion, and temperature are presented graphically by station to illustrate

the magnitude of sepsnnil changes (Figure 13-36). Table 6 summarizes

the ranges of the extremes measured as well as the approximate ranges

for more representative values.

Compared with the other groups, the lake stations (Figure 13) were

least variable seasonally for all parameters except specific conductance.

This exception was due to the upstream movement of brackish water from

the Rio Dulce, through the San Felipe outlet, and into the lake. Total

alkalinity showed little seasonal change. Some stratification was noted

at Station A due to the density current created by the Rio Polochic

waters of slightly higher total alkalinity. A decrease occurred from

above pH 8 in January to below pH 7 in March, and then showed a trend

toward increase after July. Temperatures increased during the dry season

and began to decrease after June. Dissolved oxygen concentrations at

the surface probably varied daily as much as they did seasonally.














Table 6.- Ranges of temperature, pH, total alkalinity (T.A.), specific
conductance, and dissolved oxygen saturation for all stations
sampled at Lake Izabal


Lake Stations
(A, B, C, &
San Felipe)





Swamp Waters
(Oscuro, Amatillo,
El Padre Cr.)





Rio Polochic
(Comercio, Cobin,
Bujajal)





Small Rivers
(Sauce, San Marcos
Manacas Cr., Tinico)


Temp. (C)
pH
T.A. (meq/liter)
Cond. (pmho/cm)
02 (% Sat.)



Temp. (C)
pH
T.A. (meq/liter)
Cond. (iimho/cm)
02 (% Sat.)



Temp. (C)
pH
T.A. (meq/liter)
Cond. (Qmho/cm)
02 (% Sat.)



Temp. (C)
pH
T.A. (meq/liter)
Cond. (pmho/cm)
02 (% Sat.)


Most Values


26.0-30.0
6.00-7.00
1.70-1.80
175-200
80-105



25.0-31.0
5.50-7.00
1.00-2.10
100-200
0-100



23.5-30.5
5.50-6.75
1.50-2.00
170-225
75-95



25.0-29.0
6.00-7.00
1.00-2.75
100-250
90-100


Extremes
Measured


24.1-31.4
5.60-8.25
1.60-2.00
150-465a
8-107



23.8-31.3
5.20-7.30
0.82-2.37
65-230
0-123



23.0-30.9
5.20-8.00
1.59-2.19
157-260
47-101



24.2-32.8
5.75-8.20
0.80-2.82
79-268
84-105


a The highest conductivity was recorded at
due to a localized brackish water mass.


San Felipe (5,000 mmho/cm)


II~ _~ _



















Figure 13.-


Seasonal changes in dissolved oxygen concentration, temperature, total alkalinity
(T.A.), conductivity, and pH for lake stations. Solid circles represent surface
readings and clear circles the bottom readings at 11.5 m (Station A), 15 m
(Station B), 12 m (Station C), and 10 m (San Felipe).


























Sf mam j a s o nd j


SAN FELIPE


5,000 A 355
,,v,


mamjj ason


MONTH


I,









Stratification of oxygen is highly variable depending on the frequency

and the depth to which vertical circulation occurred.

The swamp waters (Figure 14) became diluted at the beginning of

the wet season as reflected by the decreases below dry-season values in

conductivity and total alkalinity after June. The pH was variable, but

tended to decrease during the dry season. Both temperature and dissolved

oxygen concentration decreased dramatically at the beginning of the wet

season. The flushing of these rivers by the colder and poorly oxygenated

waters originating in the surrounding swamp forest also destroyed the

stratification that had been established during the dry season.

The water of the Rio Polochic distributaries (Figure 15) was

characterized by dry-season decreases in pH, increases in conductivity

and temperature, little change in total alkalinity, and absence of low

dissolved oxygen concentrations.

All the small rivers (Figure 16), except for the Rio San Marcos,

ceased flowing during the dry season. Thus changes at the beginning of

the wet season are less marked in the San Marcos water than in the other

small rivers. The seasonal trends were similar to those of the Rio

Polochic distributaries described above.

The implications of these seasonal changes in water quality for

metabolic activity of the lake and plankton abundance will be discussed

in the following sections.


Mineral analyses and the Na:Cl ratio

Samples of lake and river water that were analyzed for dissolved

minerals were collected during the dry season (April 14, 1972) and the

wet season (October 23, 1972). The results (Appendix, Table E) show



















Figure 14.- Seasonal changes in dissolved oxygen concentration, temperature, total alkalinity
(T.A.), conductivity, and pH for swamp waters. Solid circles represent surface
readings and clear circles the bottom readings at 5 m (Rio Oscuro), 3 m
(Amatillo), and 2 m (El Padre Creek).















































MONTH



















Figure 15.- Seasonal changes in dissolved oxygen concentration, temperature, total alkalinity
(T.A.), conductivity and pH for distributaries of the Rio Polochic (Comercio,
Coban and Bujajal).






COMERCIO


j f ma m j j a s o n


mamj j ason
MONTH


COBAN


BUJAJAL



















Figure 16.- Seasonal changes in dissolved oxygen concentration, temperature, total alkalinity
(T.A.), conductivity, and pH for small rivers (Sauce, San Marcos, and Manacas Creek).









RIO MANACAS CR


j fmam j as ond j f mamj J as on


'MONTH


RIO; SAUCE


RIO SAN MARCOS









that Mn, Fe, Zn, and Ni were detected only occasionally, and always at

the lower range of the sensitivity of the tests. Attributing signifi-

cance to these results would be unwarranted.

The presence of chloride apparently interfered with the measure-

ment of nitrate for the brackish waters of the Rio Dulce Salt Spring

and Amatique Bay (October 23) yielding suspiciously high values of 2.7

and 8.5 mg/liter, respectively. Except for the Rio Agua Caliente and

Rio Sauce during the wet season, nitrate was near the limit of detection

(0.62 mg/liter) of the specific ion electrode. Dry-season concentra-

tions of nitrate were greater than 1 mg/liter in some of the swamp

waters (Oscuro Bay, Amatillo, El Padre Creek, and Ensenada El Padre)

as well as the Comercio and Coban distributaries of the Rio Polochic.

Discolved phosphate concentrations ranged between 0.04-0.17: mg/liter

and in this range of sensitivity the significance of the results is

subject to question.

The high concentrations of Ca, Mg, and Si in the Rio Agua Caliente

can be attributed to the hot spring at its origin. However, the low

discharge of this river would have resulted in less contribution to the

lake of these ions than less concentrated rivers with higher discharges.

The Rio Sauce drained a limestone area (probably dolomite) and differed

from the lake water by its higher Mg and Si concentration and lower Ca.

Rio Manacas Creek during the dry season had a Mg:Ca ratio greater than

one. The Rio San Marcos was notably more dilute than the lake water

as suggested by the consistently lower specific conductance of the river

throughout the year (Figure J6). Silica concentrations for the October

23 samples may be high due to storage for three months in glass bottles.









The analyses for sodium and chloride in the lake and Rio Polochic

waters, however, have some interesting implications for understanding

circulatory patterns of the lake water. These data and the ratios of

sodium to chloride are presented in Table 7. On April 14 both Na and

C1 were higher at San Felipe and Station C at 12.5-m depth than in the

rest of the lake. The appearance of slightly brackish water was noted

also with conductivity determinations at San Felipe in April but not

until June at Station C (Figure 13). Thus the mineral analyses of Na

and C1 provided a more sensitive method than conductivity measurements

for detection of brackish water as it entered the lake from the Rio

Dulce.

The average Na:C1 ratio for the April 14 lake station samples

excludingg San Felipe :ind Station C, 12.5) was 0.52. This value is

slightly lower than the ratio 0.56 for sea water (Remane and Schlieper

1971). On October 23 the high ratios of 1.33 at Station A (11.5 m)

and 1.69 at Station B (15 m) distinguished them from the ratios of the

remaining lake stations. The origin of these high Na:Cl ratios is

apparent by comparing them with the October 23 average of the Polochic

samples which was 1.76. This provides a conclusive check for the exis-

tance of the density current implied by the 02, alkalinity and specific

conductivity stratification at Station A (Figure 13) after the initia-

tion of the wet season.

If the Rio Polochic had been the only source of water for the lake,

then the lake water would have been more dilute than it was. Even on

April 14, when Na and C1 concentrations were higher in the Rio Polochic,

they were not as high as the more dilute lake water on October 23














Table 7.- Sodium and chloride concentrations (mg/liter) and Na:C1 ratios
for lake stations and distributaries of the Rio Polochic.
Averages are calculated for selected stations



14 April 1972 23 October 1972
mg/liter Na:C1 mg/liter Na:C1
Na C1 ratio Na C1 ratio
Lake Stations

Station A Surf 4.5 7.8 .58 3.3 4.1 .80
Station A 11.5 m 4.3 8.2 .52 3.2 2.4 (1.33)
Station B Surf 4.3 8.8 .49 3.5 6.0 .58
Station B 15 m 4.3 8.8 .49 2.7 1.6 (1.69)
Station C Surf 4.3 8.2 .52 3.6 5.5 .65
Station C 12 m 8.2 10.5 3.6 6.4 .56
San Felipe -
Surf & 10 m 7.7 15.6 3.7 4.6 .80
Average .52 .68a

Rio Polochic
Comercio 3.1 4.9 .63 1.9 1.0 1.90
Coban 3.1 4.5 .69 2.2 1.0 2.20
Bujajal 3.1 1.9b 2.0 1.7 1.18
Average .66 1.76


SNumbers in parentheses not calculated in average.
Accuracy of determination questionable.









(excluding Station A-11.5 m, and Station B-15 m). -By examining Table E

of the Appendix, there is no evidence for possible sources of high Na

or C1 from the rivers sampled, except Rio Agua Caliente. However, its

wet-season discharge was low and the river did not flow at all through-

out most of the dry season. Further confirmation is available from

August 1969, when Brooks (1969) detected Na and C1 concentrations in

the Rio Polochic to be 3.2 and 2.2 mg/liter respectively. Higher con-

centrations (4.8 mg Na/liter and 7.5 mg Cl/liter) were reported by

Brooks for lake water samples.


Brackish water movement into Lake Izabal

To determine the seasonality and extent of movement of the saline-

fresh water interface between Lake Izabal and Amatique Bay, several

sampling trips were made into the Rio Dulce-Fl Golfete region (Figure

17). Two of these trips traversed the area between the San Felipe out-

let and the coastal port of Livingston; one was made during the dry

season (March 22-23, 1972) and the other during the wet season (October

26, 1972). A third transect (May 13) included only Stations 1-8. The

sampling stations are numbered from 1 to 11 in Figure 17.

On the first transect (March 22-23) a salt-water wedge was observed

extending from Station 11 at Livingston to El Golfete between Stations 5

and 6 (Figure 18). The deep high-salinity water was nearly isothermal

(29.4-29.6 C) and slightly warmer than the surface waters (Figure 19).

A strong (ca. 0.6 m/sec) outgoing current in the upper 1-2 m of the Rio

Dulce below El Golfete marked the interface with the relatively motion-

less deep layers. An increase in conductivity at Station 8 over a 17-

hour period was attributed to downstream tidal forces (Figure 20). This

































0

0
cd





0
r-I,




4-)
C 0






4-)
cd~

r-
cd-








Ln~ Ln



*H j 4 -






.4-)0
a)

0 C V )
I4J








tR 0 0
0c,~c


















4- -4
o
Pi-L 0 0







0 c
--4c
0~c 0








--4k
0ao









BAHIA DE

SAIW4TIQUE


LAGUNA


p- 2 Kilostsrm
I I, M 0ile$


LAGO DE
IZABAL


LAKE
STATIONS




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EMXOE1GF0_YS03S3 INGEST_TIME 2011-07-15T20:16:57Z PACKAGE UF00098368_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

THE ORGANIC R-MTER BUDGET AIvID ENERGY FLOW OF A TROPICAL LOWLAND AQUATIC ECOSYSTEM By MARK McCLELLAN BRINSON A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FL01?IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1973

PAGE 2

Copyright by Mark McClellan Brinson 1973

PAGE 3

Digitized by the Internet Archive in 2009iwitMbmding from University of Florida, George A. Smathers Libraries http://www.archive.org/details/organicmatterbudOObrin

PAGE 4

ACICNOWLEDGMENTS The author wishes to express his appreciation to Dr. Ariel Lugo for ecological insight and logistical support which helped to insure the success and feasibility of this investigation. Sincere appreciation goes to Dr. Frank Nordlie for helpful discussions on Lake Izabal, and to the otlier members of the supervisory committee, Dr. David Anthony, Dr. Dana Griffin, 111, and Dr. Leland Shancr, for critically reviewing the manuscript. Financial support that was provided by the Foreign Area Fellowship Program during the field study and write-up period is gratefully acknowledged. The Center for Tropical Agriculture provided assistance during a reconnaissance trip prior to the field research. Dr. Hugh Popenoe's continned i-^t'^rest in '•"'"'e Izabal Witershed provided the fr?-me'"ork for research on a regional level. Successful affiliation was established with employees of the Department of Wildlife, Ministry of Agriculture, Guatemala. This was possible through the efforts of its Chief, Jose Ovidio de Leon, who was also responsible for processing equipment through customs. Adolfo Orosco Pardo assisted in counting much of the plankton and Renato Zuniga helped with a phase of the field operations. Thanks goes to the many people of El Estor who helped to provide dayto-day needs, and to the EXMIRAL mining company for supplying climatological data during the period of study. I am particularly grateful for my wife's assistance antl encouragement during our memorable experiences in the Guatemalan lowlands. IV

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGMENTS ^^ LIST OF TABLES ^ LIST OF FIGURES ^^^ ABSTRACT '. • xvi INTRODUCTION ^ Terrestrial Ecosystem Exports 2 InfloKS to Aquatic Ecosystems 5 REGIONAL SETTING ^ Climate • ^ Lakeside Temperature, Rainfall and Solar Radiation 9 Precipitation for the Remainder of the Watershed 16 Geology and Soils ^^ The Lake ^^ The Fisheries and People 24 Immigration and Labor Alternatives 25 Fishing Regulations and Yields 26 Fish Behavior and Fisheries Management 28 HYDROLOGY AND WATER CHARACTERISTICS 32 Methods ^^ General Methodology--Field Stations, Logistics, and Schedule •^'"'

PAGE 6

TABLE OF CONTENTS continued Page Hydrologic Measurements 36 Analyses of Water Characteristics 38 Results and Discussion 41 River Discharge and Runoff 41 Water Budget . 52 Water Characteristics 60 Thermal properties and circulation patterns of the lake 60 The seasonal pattern of water characteristics.... 66 Mineral analyses and tlie Na:Cl ratio 70 Brackish water movement into Lake Izabal 80 Summary Statement 99 NET PLANKTON AND BENTHIC COMMUNITIES 102 Methods 103 Results and Discussion 104 Phytoplankton and Zooplankton Communities 104 Diatoms 104 Green algae 108 Blue-green algae 115 Otlicr algae 116 Phycomycetes 116 VI

PAGE 7

TABLE OF CONTENTS continued Page Copepods '^' Cladocera ^^"^ Rotifers ^^^ Possible Controlling Factors 122 Benthic Community 1-^^ Summary Statement ^^-^ METABOLISM AND ORGANIC MATTER i38 Methods ^ "'^ Chemical Oxygen Demand ^^^ Fraction definitions and conversion criteria 139 Collection and treatment of samples 140 Dissolved Oxygen Concentration Determinations 143 Biological oxygen demand 144 Respiration of bottom muds and their organic content '•^^ Light and dark bottle method 147 Results and Discussion 148 Concentrations of Chemical Oxygen Demand 149 Swamp waters 149 Rio Polochic distributaries 152 Smal 1 rivers 1 ^^ o VIX

PAGE 8

TABLE OF CONTENTS continued Page Lake stations and outlet 153 Monthly Flows of COD 154 Respiration Rates (BOD) , 160 Respiration and Organic Content of the Bottom Muds... 168 Measurements of Primary Productivity 175 Seasonal rates of gross primary productivity and respiration 1 78 Efficiency of gross primary productivity 185 Light penetration and cc".pen:;aticn depth 1S9 Balance of the Organic Matter Budget 195 Summary Statement 202 CONCLUSIONS 204 Seasonal Pulse of Organic Detritus Movement 20S Movement of Organic Matter to Site of Consumption.... 208 Mechanisms for Steady-State Balance 210 Oscillations in Food Concentration for Consumers 212 Consequences of the Connection to a Marine Ecosystem 213 Ecosystem Management 213 Fishery Management 214 Modern Agricultural and Industrial Man 215 APPENDICES 220 VI 11

PAGE 9

TABLE OF CONTENTS continued Page LITERATURE CITED 244 BI0GR.-\P!1ICAL SKETCH 251 IX

PAGE 10

LIST OF TABLES Table Page 1 Sampling dates for chemical oxygen demand (COD) , biological oxygen demand (BOD), and primary productivity experiments (Prod) 34 2 Monthly discharges (m^ x 10 ) of rivers and watersheds draining into Lake I zabal 50 3 Summary of hydrologic data for the major watersheds draining into Lake I zabal 51 4 Monthly free water surface evaporation measurements for selected warm climates and warm seasons 53 5 Summary of water budget for Lake 1 zabal (November 1971-October 1972) 55 6 Ranges of temperature, pH, total alkalinity (T.A.)j specific conductance, and dissolved oxygen saturation for all stations sampled at Lake Izabal 67 7 Sodium and chloride concentrations (mg/ liter) and Na:Cl ratios for lake stations and distributaries of the Rio Polochic 79 8 Species list of net plankton frequently collected from Lake Izabal 105 9 Bottom fauna of Lake Izabal expressed in numbers/m 132 10 Characteristic oxygen equivalents (O.E.) in approximate order of their limnological importance 141 11 Monthly COD inflows (g COD x 10^) for individual rivers and watersheds, total monthly inflows to Lake Izabal, and monthly and total outflow at San Felipe 155 12 Relative contribution (percent) of organic matter runoff into Lake Izabal from the Rio Polochic Valley and from the minor watersheds 159 13 Hourly and daily respiration rates of Lake Izabal bottom hukIs 1"73

PAGE 11

LIST OF TABLES continued Table P^ge 14 Metabolism calculations of light and dark bottle experiments for Stations A, B, and C during 1972 181 15 Total daily incoming radiation averaged on a weekly and monthly basis from October 1971 through October 1972 186 16 Daily efficiencies of energy conversion from visible solar energy to energy fixed by gross primary productivity (Pg) ISS 17 Calculation of light intensity at compensation depths as percent of surface intensity 1S4 18 Sunmiation of organic matter (CM) flows for Lake Izabal from March to October 1972 • 198 19 Daily and annual rates of gross primary productivity fPg) for some tropical lakes and comparative ranges for temperate lakes 201 XI

PAGE 12

LIST OF FIGURES Figure Page 1 The Izabal Watershed 11 2 Monthly averages of the maximum, minimum, and calculated mean temperatures at Las Dantas for the period of study (October 1971-1972) 15 3 Temperature-rainfall climate diagram for Las Dantas and Mariscos during the year of study 18 4 Bathymetric map of Lake Izabal illustrating sampling stations 25 5 Calibration curves for the small and large probes used for specific conductance measurements between 100 and 30,000 ymho/cm at 25 C , 40 6 Monthly estimates of all inputs to the lake (runoff and direct rainfall) and the monthly averages of lake water levels 44 7 Linear relationship between monthly rainfall and monthly discharge of rivers emptying into Lake Izabal 46 8 The seasonal march of monthly discharge rates for some maj or rivers em.ptying into Lake 1 zabal 48 9 Relationship between velocities at the San Felipe outlet from direct field measurements and those calculated by balancing the water budget 57 10 Summary diagram of water storages and annual flows 59 11 Vertical temperature profiles of Lake Stations A, A-B, and B recorded at approximately monthly intervals 63 12 Vertical temperature profiles of Lake Stations B-C and C recorded at approximately monthly intervals 65 13 Seasonal changes in dissolved oxygen concentration, temperature, total alkalinity (T.A.), conductivity, and pH for lake stations 69 XI 1

PAGE 13

LIST OF FIGURES continued Figure ^^^e 14 Seasonal changes in dissolved oxygen concentration, temperature, total alkalinity (T.A.), conductivity, and pH for swamp waters 72 15 Seasonal changes in dissolved oxygen concentration, temperature, total alkalinity [T.A.), conductivity and pH for distributaries of the Rio Polochic (Coinercio, Coban, and Bujajal) 74 16 Seasonal changes in dissolved oxygen concentration, temperature, total alkalinity (T.A.), conductivity, and pH for small rivers (Sauce, San Marcos, and Manacas Creek)... 76 17 Map of the Rio Dulce-El Golfete system. 82 18 Conductivity profiles (Marrh 22-23, 1972) alnnp the Rio Dulce frc:u San Feliiie (Station 1) to Ajp.?tique ^-'^y (Station 11) 84 19 Temperature profiles at two stations illustrating a slight temperature increase associated with the halocline. . . . 86 20 Observed diurnal change in the conductivity profile of the waters at the lower reaches of El Golfete (Station 8) 88 21 Conductivity of the ground water in a swamp forest at Cuatro Cayos 91 22 Conductivity profiles along the Rio Dulce from San Felipe (Station 1) to the lower reaches of El Golfete (Station 8) ' 93 23 Conductivity profiles taken at several stations in Lake Izabal and the upper reaches of Rio Dulce (Stations 1 and 2) 95 24 Bottom profile of the Lake Izabal-Rio Dulce system, showing locations of sampling stations 98 25 Seasonal changes in abundance of pennate diatoms and Melosirg gran ulata 107 Xlll

PAGE 14

LIST OF FIGURES continued Figure : Page 26 Seasonal changes in abundance of St aura strum leptocladum and S^. p ingue 110 27 Seasonal changes in abundance of Pediastrum simplex and Staurastrum tohopekaligense 112 28 Seasonal changes in abundance of phycomycetes and Anacystis cyanea 115 29 Seasonal changes in abundance of copepods lli^ 30 Seasonal changes in abundance of cladocera 121 31 Seasonal changes in abundance of colonial rotifers 124 32 Seasonal changes ir\ abundance of solitary rotifers 126 33 Total net-plankton abundance (units or organisms per liter) represented by phytoplankton, zooplankton, and phycomycetes 136 34 Concentrations of particulate and dissolved COD (mg/liter) during the sampling period for (a) swamp waters, (b) Rio Polochic distributaries, (c) small rivers, and (d) lake stations [A, B, C, and San Felipe] 151 35 Rates of organic matter inflows and outflows of Lake Izabal for the lake as a whole (g COD x 109/month) and for an average mof surface area (g C0D/m2 day) 157 36 Respiration rates (1-day and 5-day BOD) for (a) swamp waters, (b) Rio Polochic distributaries, (c) small rivers, and (d) lake stations 162 37 Dissolved oxygen concentrations of water samples during 7-day incubation periods 165 38 Relationship between oxygen consumption rates (BOD) and total COD concentrations for the four water types characterized 1"' -7 xiv

PAGE 15

LIST OF FIGURES continued Figure Page 39 Organic composition of Ekman samples from Lake Izabal bottom muds 170 40 Respiration rates of mud sam.ples from Lake Izabal 172 41 Example of curves generated from light and dark bottle experiments from uiiich metabolism is determined planimetrically 177 42 Method for calculating the number of hours of effective light per day 180 43 Gross primary productivity (Pg), 24-hour respiration (R24) and Pg/R24 ratios at Station A, B, and C 184 44 Sccchi dislc transparency measurements recorded at StdtioiiS rv, L) , anci C 191 45 Simplified model of the principal organic matter flows and storages in Lake Izabal as averaged over the period of study 196 46 Summary diagram of energy and matter flows and storages that characterize the Izabal Watershed 206 47 Summary diagram of energy and matter flows and storages of the Izabal Watershed which includes some of the possible influences of development by modern agricultural and industrial man 217 XV

PAGE 16

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE ORGANIC MA^TTER BUDGET AND ENERGY FLOW OF A TROPICAL LOWLAND AQUATIC ECOSYSTEM By Mark McClellan Brinson Decemberj 1973 Chairman: Ariel E. Lugo Major Department: Botany This study examines the influence of regional coupling mechanisms on the organic matter budget of a lovland ti-opical lake and documents the principal energy flows that contribute toward making the waterslied h coiiesivo ecologicaj. uiiit. TiiC uowrinili j-xOW ci organic mai^tei i.Oi>i^ 2 2 and water from the terrestrial (6,860 km ) to the aquatic (717 km 1 ecosystem was quantified and evaluated as to its effect on physical conditions and metabolic activity of Lake I::abal, Guatemala. The lake type is warm pol>Tnictic due to its shallow mean depth (11.6 m) . Its water mass has a short residence tin;e (6.6 months), and the annual gross 2 productivity is high (1,592 g OM/m ). The hydrologic regime exerted control on OM flow into the lake. Mean runoff for the watershed was calculated as 6S percent of mean annual rainfall (2,992 mm), most of which occurred during the 9-month wet i-eason (greater than 100-mm monthly rainfall). Organic matter runoff was characterized by an initial "flushing" of the watershed at tlie beginning of the wet season, when the particulate fraction (greater than O.SO microns XV 1

PAGE 17

diameter) constituted a large portion of the total organic flow. During the remainder of the year, OM runoff was nearly all contained in the dissolved fraction (less than 0.80 microns diameter). Approximately 50 percent of the total OM runoff occurred during the 3 wettest months of the year. During the dry season when OM inputs from the watershed were low, the lake experienced a net gain in OM when gross primary productivity exceeded comjr.unity respiration. During the wet season, there was a net loss of OM in spite of increased inputs of allochthonous organic detritus. This loss increased as the wet season progressed, due to a combination of decreased rates of gross primary productivity and increased rates of community respiration. The periodicity of OM accrual and loss provides a mechanism, apparently controlled by hydrologic patterns, by v/hich steadystate conditions can be achieved on an annual basis. Daily rates of gross primary productivity ranged from 1.15-7.31 g 2 2 /m day and plaiiktonic respiration rates from 0.50-8.38 g /m day. The average daily values for the organic matter budget were calculated for the 8-month sampling period and were represented by five principal 2 flows. The two OM sources were gross primary productivity (3.730 g/m 2 day) and OM imports (0.632 g/m day). The three OM losses were by OM exports (0.452 g/m day), planktonic respiration (3.875 g/m" day), and 2 respiration of the bottom muds (0.36 g/m" day). The mean residence time 2 for the average OM content of the lake (71.08 g/m ) was 16.3 days. Seasonal periodicity was expressed in the net plankton by a bimodal pulse of abundance. Peaks in plankton density occurred at the end of the dry season (April-May) and followed the initial period of heavy rainfall (August-September). Causal factors for this response remain undetermined. XV ii

PAGE 18

The connection of the lake to the marine environment, via a 42-kir. long waterway, allows additional mechanisms for ecosystem coupling. Evidence was collected to demonstrate the control of the Na:Cl ratio of lake water by dry-season penetration of brackish water into the lake. The waterway also provides marine vertebrates and invertebrates access to a fresh-water environment. Periodicity of OM metabolism in the lake, high productivity of surrounding lagoons and coves, and brackish water penetration from the coastal marine ecosystem are discussed as factors influencing consumer activity and seasonal migration. The lake's fisheries, dependent on the marine contingent of fishes, may best be managed by utilizing the understanding of regional coupling mechanisms to prevent fisheries deterioration and to ensure continued yields. XVI 11

PAGE 19

INTRODUCTION Many years ago Torbes (1887) discussed lakes as microcosms, and in so doing emphasized their isolation from the terrestrial ecosystem. Much of the limnological u-ork since that time has taken this myopic view of lakes with little regard to the activities beyond their boundaries. It is these extra-lacustrine activities that give lakes their characteristics, and they can be dealt with effectively only by expanding tJie ecosystem boundaries to include the whole of the watershed. The cohesive nature of the watershed, and its well-defined boundaries, are characteristics that make it a conceptually attractive unit for ecological study. Powered by the proper energy sources, water is the common denominator that couples this ecological unit by virtue of its geological constraints (downhill flow) and its biological indispensibility. The characteristics of this water, its flux from the terrestrial to the aquatic subsystem, and the effect of this flux on the lacustrine ecosystem are all important com.ponents of the present study. The study was conducted in the Izabal Watershed on the Caribbean slope of Guatemala. Because of the high rainfall and seasonal climate of the region and the suspected short residence time of the lake's waters, I hypothesized that upstream activities in the terrestrial subsystem would be reflected by short-term (<1 year) responses in the downstream subsystem of the lake. The main focus of the study was to exiimine

PAGE 20

the metabolic activity of the lake in response to organic matter inflow from the watershed. Tailing (1969), in reference to the poorly understood seasonality of shallow tropical lakes with no long-term thermal stratification, pointed out that "any periodicity [discovered] acquires a new interest." In the Izabal Watershed, this interest may necessarily extend beyond the boundaries of the lake to determine the extent of ecosystem coupling and matter exchange between subsystems. The connection of the lake to a marine environment adds to the difficulty of evaluating the watershed as an isolated unit or closed system. This adds new dimensions for possible mechanisms for ecosystem coupling. Depending on the degree to which these coupling mechanisms create interdependency on a regional scale, schemes for ecosystem management and land use planning should demonstrate awareness of both the potential benefits and inherent dangers of manipulation of isolated subsystems. Therefore, as Odum (1971) emphasized, "it is the whole drainage basin, not just the body of water, that must be considered as the minimum ecosystem unit when it comes to man's interests." Thus, there is an urgent need for understanding these interactions on a regional level. Terrestrial Ecosystem Exports Naturally forested areas are extremely effective in recycling materials, thereby preventing losses to do\\Tiliill processes (Bormann et al. 1969). However, organic matter fixed by terrestrial photosynthesis undergoes some leakage that eventually appears downstreiim. The degree to which this leakage occurs depends on factors which characterize the

PAGE 21

ecosystem. Some of the factors that significantlyaffect organic matter runoff are: (1) seasonal phenomena, (2) runoff intensity, (3) ecosystem perturbation, and (4) topography and other geological characteristics of the watershed. In temperate regions, the organic matter inputs to streams located in forested areas are seasonal, being highest in the autum.n during the seasonal leaf drop (Kaushik and Hynes 1968). In tropical latitudes where rainfall is the seasonal control, a deciduous forest presumably could experience a similar pulse during the dry season. However, since water is the medium that transfers organic matter downhill, one might expect the downhill transfer of organic matter to be seasonally coincident witli high rates of runoff. Although there are no data from tropical regions to support the assumption that water runoff and organic matter runoff are positively correlated, there is some evidence for this in the temperate zone. In a Piedmont stream in the southeastern USA, Nelson and Scott (1962) found that, although the dissolved and colloidal organic matter fraction of river water increased with discharge rate, the particulate fraction increased at a much more rapid rate. At low to moderate flows the dissolved and colloidal organic matter concentrations were two to ten times higher than the particulate, while, during high flow rates, the particulate organic fraction increased to double the concentration of the dissolved. Causal factors which increased the particulate portion so dramatically during high discharge were: (1) greater surface runoff associated with heavy rainfalls, and (2) the flushing effect of high water. Therefore, with increased rates of river discharge, organic matter flows increase at

PAGE 22

a proportionately greater rate than water flows. There is no reason to expect that these factors would operate differently in tropical latitudes. Therefore, measurements of the organic matter runoff from the Izabal Watershed to Lake Izabal would necessarily include a range of runoff intensities in order to characterize size fractions of organic runoff and to achieve a good estimate of absolute quantities. Perturbation of terrestrial ecosystems by deforestation decreases the ability of watersheds to prevent dovmstream losses by destroying the mechanisms and adaptations for the recycling of matter. Regions of the Izabal Watershed have received some alteration from deforestation and agriculture. Again, specific data for estimating the magnitude of organic runoff change by deforestation must be drawn from temperate-zone ecosystems. The northern hardwoods of the Hubbard Brook Experimental Forest, New Hampshire, provide the model on which to judge effects of perturbation. There, drainage streams exported 5.3 grams of organic matter per m" of watershed area annually (Bormann et al. 1969). After clear-cutting of the forest, organic matter losses doubled during the first two years. These ecosystem exports represent inputs for do\mstream ecosystems (Likens et al . 1970). If the downstream ecosystem were a lake, then depending on its size, these can represent a significant source of organic matter and illustrate the importance of one-way coupling between a terrestrial and aquatic ecosystem. The absolute values for tropical regions may be different, but nevertheless deforestation could be expected to result in increases in organic matter runoff. The only assumption necessary for arriving at this conclusion is that the mechanisms for recycling matter possessed by ecosystems of both latitudes would be lost or severely damaged upon destruction of the forests.

PAGE 23

Finally, the topography and otlier geological characteristics of watersheds are factors that should be considered in the regulation of organic matter runoff. Runoff waters from the steep mountain environm.ent in the Amazon Basin have higher concentrations of dissolved and suspended organic matter than those of the lower i'^jTiazon (Gibbs 1967). However, in large watersheds there may be a great deal of spatial variation in topography and geology as well as the previously discussed variables of runoff intensity and ecosystem perturbation. The larger the watershed, the more these spatial variations tend to become integrated by the confluence and mixing of tributaries by the time downstream measurements are taken. Also the metabolic activities of riverine ecosystems could be expected to modify both the quantity and quality of organic matter after it is received upstream. In the Izabal Watershed, the majority of the drainage areas become confined to one major river before the waters discharge into the lake. Thus, any peculiarities in organic matter sources will tend to cancel one another by the time the water discharges into the lake. Inflows to Aquatic Ecosystems Now that the characteristics of organic matter runoff have been discussed and loosely established, it is essential to determine the qualitative and/or quantitative influence that this organic source may have on downstream aquatic ecosystems. Showing that the energy fixed in the form of organic matter in the terrestrial ecosystem can be utilized in the aquatic ecosystem would establish that an energetic coupling or flow occurs on a regional level.

PAGE 24

There is good evidence that aquatic ecosystems adapt to organic detritus inputs by utilizing therr. as a source of energy. This has been demonstrated for estuaries (Teal 1962; Darnell 1967; Heald 19 71; Cooper and Copeland 1972) and in flov/ing waters (Odum 1956; Nelson and Scott 1962). That the Lake Izabal ecosystem should be an exception to this concept would seem anomalous. Some workers have had the insight to compare the relative contribution of organic matter import to the total metabolism of coastal embayments. Of the total organic increment, allochthonous inputs accounted for about one-half the total for the Strait of Georgia (Seki et al. 1968), 7-21% in Moriches Bay on Long Island (Barlow et al. 1963), and about 97% in the turbid and polluted Chao Phya River estuary in Thailand (Pescod 1969) . In addition to its importance to the total metabolism of some areas, organic detritus lias been shown specifically to be a principal food item for many estuarine vertebrates and invertebrates (Darnell 1961, 1967; Odum and de la Cruz 1967; W. Odum 1971) . The basis for the nutritive and presumably the energetic value of particulate detritus is the high quality of organic matter (e.g. proteins) associated with the micro-organisms that colonize the particles. Wherever ecosystems exist that have organic detritus inputs, there is good evidence that the organisms are adapted to utilizing the organic component as a source of energy. Lakes, however, have apparently received the least attention of all aquatic ecosystems that may derive some of their energy from organic matter transported from outside their boundaries. Again, specific examples that demonstrate this are drawn from studies in temperate latitudes. Two independent approaches have been used to

PAGE 25

measure allochthonous organic matter sources for lakes. One is to directly monitor the organic matter runoff from a watershed--the approach used in the present study. This was done by McConnell (1965) in small impoundments of the semi-arid southwestern USA. There, relatively unleached oak litter entered a lake by surface runoff at an annual rate of about 750 g dry weight per m" of lake surface. This allochthonous source of organic matter supplied the lake with approximately one-third of the total organic matter increment; the remainder was supplied by primary production. The other approach is by indirect measurements and is feasible only in lakes that thermally stratify. In eutrophic Lake Mikalajki, Poland, Lawacz (1969) trapped seston as it sank from the trophogenic zone to below the epilim.nion. The total organic matter production by this method was about half again as great as that of the plankton over a year as measured by the oxygen method. Lawacz attributed the difference in production to unmeasured dissolved organic substances that are produced in the littoral zone. These substances presumably move into the pelagial zone and sink after transformation into particulate form. Most attempts to quantify the organic sources of tropical freshwaters have been made by measuring the primary productivity of relatively large lakes. The size of these lakes would tend to diminish the importance of an allochthonous organic source, and thus the "lake as a microcosm" view is justified, at least for organic matter production. As pointed out by Ruttner (1965), the quantity and composition of allochthonous organic matter depends on the ratio of the surface area of the lake to that of the watershed. In general, allochthonous organic

PAGE 26

sources will be of greater importance to small lakes with large watersheds than to large lakes with small watersheds. Already discussed are the important influencing factors such as climate, the morphological and geological character of the watershed, and plant cover of the watershed. Considering these factors, the humid tropics are a likely region where lakes may receive flows of significant magnitude from outside their boundaries. These flows can have a marked influence on the limnological characteristics of these lakes. Lakes in which the residence time of the water mass is relatively short are likely candidates for such tightly coupled influences. In the context of ecosystem management, it follows that alteration or perturbation of watersheds could have a profound effect on the activities within such lakes. Similarly, these lakes may be our most sensitive indicators of changes in upstream activity. Thus, the objectives of this study are to (1) determ.ine the quantity and seasonal distribution of organic matter transfer from the terrestrial to the lacustrine ecosystem, (2) quantitatively compare the allochthonous organic matter source of the lake with the organic matter derived by in situ primary production, (3) evaluate how this one-way regional coupling mechanism influences metabolic activities in the lake, and (4) document the principal energy flows that contribute toward making the watershed a cohesive ecological unit.

PAGE 27

REGIONAL SETTING The Izabal Watershed extends 205 km eastward from the interior of Guatemala to the Caribbean coast. The orientation of this 50-km wide strip of terrain is east-west, located at 8S°41' W to 90°54' W and 15°03' N to 15°52' N. The drainage pattern is from the highlands in the western sector, downward to Lake Izabal only a few meters above sea level, and then through the Rio Dulce, finally reaching Amatique Bay in the Gulf of Honduras on the Caribbean. However, for the purpose of this study, the watershed will not include the area downstream from 2 Lake Izabal, but only the region (6,860 km ) that drains into the lake Cl-igure 1) . Two mountain ranges delimit the Izabal hydrologic unit; the Sierra de las Minas and the Montanas del Mico lay end-to-end and parallel the southern boundary, while the Sierra de Santa Cruz and Sierra de Chama create the divide to the north. The major rivers are the Rio Polochic and the Rio Cahabon, which converge into a massive river that is actively building a delta across the end of the lake. Other rivers and small streams, about 40 in all, drain less extensive areas and often are intermittently dry during periods of low rainfall. Climate L akeside Temperature, Rainfall and Solar Radiation The seasons associated with rainfall and temperature changes near Lake Izabal were described by Snedaker (.1!^70J. The wet season begins abruptly

PAGE 28

ctf

PAGE 29

11

PAGE 30

12 sometime in May and tapers off toward the end of the year; the dry season is from December through April. The daily temperature range is greatest during the drier months, reflecting the ameliorating effect of rainfall. The average m.onthly temperature range is 4.4 C over the year and the annual mean is 25.2 C. The lowest temperatures generally occur in December and January and are associated with cold air masses moving in from the north during the winter. Finca Murcielagos, located on the north-central shore of the lake, received an annual average of 2,004 mm precipitation over a 6-year period (Snedaker 1970). Rainfall increases to the east and the west of this location. No month is without rainfall, while July is the wettest and February the driest month. August has less precipitation than either of the adjacent months which coincides with the caniculas fdog days), a term that describes a week to 10-day period of dry and cloudless days. Snedaker reports that during his 6-year study, 45.5-0 of the days experienced at least 1 mm of precipitation, and more precipitation tends to fall at night (ca. 71% of the total). Lake Izabal has a pronounced local effect on the climate, especially the solar radiation. Mornings are generally cloudless in the lowlands, but in the lake area, clouds begin to accumulate over the surrounding terrestrial region from about 10 a.m. to noon. Differential solar heating over the land causes uplifting convective air currents and condensation of the water vapor while the sky above the relatively cooler lake remains clear. About mid to late afternoon these clouds move horizonally across the lake. Snedaker (1970) recorded hourly values of net incident

PAGE 31

13 radiation daring I9bi and calculated the total daily net radiation received for May and June at 408.1 and 494.1 langleys/day, respectively. The climate near the lake during the year of study (October 19711972) was characterized by examining records maintained by the EXMIBAL mining company on the north and south shores of the lake. The records included daily rainfall and maximum and minimum temperatures from the north shore at Las Dantas, and daily rainfall from the south shore at Mariscos, 32 kiii to the east-southeast. Average temperatures v%'ere calculated from the maximum and minimum temperatures by interpolation using the same relationship that Snedaker (1970) observed from his hourly readings at Finca Murcielagos. Figure 2 shows that the two warmest months were April and May during the dry season. Ihe decrease in temperature after May can be attributed to amelioration by rainfall during the wet season. Temperatures increased until September and thereafter decreased until February. The low November temperature, which interrupts an otherwise continual decrease from September through February, can be attributed to an unseasonably cool, weeklong persistence of stormy weather caused by a hurricane on the Atlantic coast. '" ' Mariscos received 3,236 mm rainfall during the year of study, which is considerably more than the 2,210 recorded at Las Dantas (Appendix, Table A). The month of heaviest precipitation for both stations was August (630 mm at i\kiriscos and 426 mm at Las Dantas). The rainfall at Las Dantas during the year of study is in close agreement with a fiveyear average (1963-1967) for the same station and slightly higher than the 2,004 nm average from 1961-1967 at Finca Murcielagos (Snedaker 1970). The climate diagram, which uses the conventions of Walter and Lcith (1960-67), illustrates the monthly marcli of precipitation and average

PAGE 32

Figui'e 2.Monthly averages of the maximum, minimum, and calculated mean temperatures at Las Dantas for the period of study (October 1971-1972).

PAGE 33

15 Las Dantas season 20 J F M A M J J A MONTH S O N D

PAGE 34

16 ambient temperature (Figure 3). Rainfall is the average of the records at Mariscos and Las Dantas during the year of study and temperatures are those of Snedaker (1970). The shaded area above 100 mm represents months when rainfall was in excess of evapotranspiration, and is the period when the most surface runoff can be expected. The stippled area represents the period when evaporation is greater than rainfall, implying a water deficit. Precipitation for the Remainder of the Watershed The low topography near the eastern part of the lake allows trade winds to pass unobstructed into the basin. Orographic rains occur as the moisture-laden air masses sweep inland from the Caribbean and upward over the Polochic Valley. There is considerable spatial variation in rainfall as illustrated by the isopleths in Figure 1. The data on which these isopleths are based appear in the Appendix, Table B and also include information from the Instituto Geografico Nacional (1966). Heaviest rainfall, averaging close to 4,000 nmi, is concentrated in the north-central region of the Polochic Valley. Radiating from that area, the average decreases to values below 2,000 mm. San Juan received 6,128 mm in 1969, the highest amount recorded for any station in a single year. To arrive at an estimate of the average rainfall for the Izabal Watershed exclusive of the lake, the areas between the estimated rainfall isopletlis (Figure 1) were determined planimetrically. For the year of study, average rainfall for the watershed as a whole was calculated at 2,992 mm.

PAGE 35

Figure 3 Temperature-rainfall climate diagram for Las Dantas ana Mariscos during the year of study. Average monthly temperatures were calculated from a 6-year record at Finca Murcielagos.

PAGE 36

18 3020^ 'C 100-600 -400 -200 mm 20 -0 T » 1 1 1 1 1 rJ fmamj j asond J Month

PAGE 37

19 Ge ology and Soils The Izabal Watershed lies in the physiographic province known as the Central American Mountain System, just to the north of the volcanically active Pacific Cordillera (Walper 1960). The orientation of the rivers in the watershed is controlled by the east-west faulting and folding which has produced a series of anticlinal mountains. This major zone of faulting, in which the Polochic Valley lies, is postulated as being tectonically related with the fault zone of the Cayman trencli (Bartlett trough) in the northern Caribbean (Walper 1960). The lake occurs in a block fault basin (Dengo and Bohnenberger 1969) . The Rio Folochic originates some 2,100 m above sea level and passes a distance of 100 km before reaching the lake. The headwaters lie between the Sierra de Pansal to the south and the Sierra de Xucaneb and Sierra Tzalamila to the north (Popenoe 1960) . To the north of these two latter confluent ranges, the Rio Cahabon begins its eastwardly flow, finally connecting with the Rio Polochic 53 km downstream. These waters become distributed in the multilobate Polochic Delta before they discharge into the lake. As levees of the distributaries protrude into the lake and deposit alluvium, shallow coves and lagoons become isolated in the delta region, providing interesting ecosystems for the study of seasonal succession and metabolism of the plankton (Brinson 1973). The structure and stratigraphy of the watershed is complex and incompletely understood (Walper 1960) . West of the lake, and at higher elevations, prominent cliffs of sedimentary limestone and interbedded dolomite constitute massive beds of Permian age (Ftoberts and Irving 1957). Nearby at Cahabon, beds of terrestrial conglomerate and sandstone predominate

PAGE 38

20 and are believed to be of TeiLiary age. Igneous rocks occur in lenticular arrangement along a fault which parallels the north shore of the lake. This serpentine area is believed to be of late Paleozoic and early Mezozoic age. The mountains that parallel the southern area of the watershed are metamorphic pre-Cambrian rocks consisting of undifferentiated schist, gneiss, phyllite, quartzite, and marble. The soils of the Izabal Watershed, as all the soils of Guatemala, have been classified and mapped by Simmons et al. 1959. Soils around the north and east perimeter of the lake were examined by Tergas (1965) in relation to the primary production of natural vegetation. Some of these soils have high ratios of calcium to magnesium (1:1.2 to 1:6) as a result of their derivation from serpentine rock. Popenoe (1960) nn his remarkable study of the response of soils of the Polochic Valley to shifting culti\'ation (slash-and-burn) , described many of the soil properties. He stated that "Erosion is very slight on the steep lands of the Polochic Valley, probably due to excellent soil physical conditions" imparted by the relatively low bulk densities of the topsoils. Considering the diversity of parent material and the extremes in climatologic regimes of the watershed, a more complete description of the soils would be inappropriate and would necessarily include inaccurate generalizations. The lithology of the Izabal Watershed is a heterogeneous one, ranging from old metamorphic to relatively young sedimentary rocks. Within this mosaic, one would expect to find significant spatial variation in weathering and differences in tlic ionic comjiosition of the Iicadwaters of streams. However these local irregularities can be expected to cancel each otlier as tributaries converge, so tliat tlie doMistream parts of tlie rivers will tend to resemble one anotlier.

PAGE 39

21 The Lake Lake Izabal is located between 15°24' N to 15°38' N and 88°58' W to 89°25' W and lies only a few meters above sea level. The central basin of the lake is a broad, nearly flat plain reaching a maximum depth of about 16 m near the center. Thus most of the volume of the lake is below sea level. Most of the shallow areas are the coves and lagoons at the west end of the lake, bordering the delta of the Rio Polochic. According to the bath>'Tiietric map (Figure 4) adapted from Brooks (1969) only 9.7% of the area of the 717 km" lake is less than 4.6 m (15 feet) deep; 50% of this occurs in the shallow areas near the delta, and the remainder around the perimeter of the lake. The volume of the lake is 8,300 x 10 m"^, giving it a mean depth (volume/area) of 11.6 m. Nearly SO streams and rivers flow into the lake, and while the greatest number flow into the eastern, northern, and southern edges, the greatest volume is received through the Polochic Delta to the west. The outlet to the lake is at San Felipe, where the lake water enters to Rio Dulce on its flow to Amatique Bay in the Gulf of Honduras on the Caribbean Sea. About midway between San Felipe and the coastal port of Livingston (a distance of 42 km) the Rio Dulce broadens into a large shallow area (4.5 m deptli) known as El Go 1 fete. Tsukada and Deevey (1967) suggested, on the basis of sediment cores that ended in sand and gravel, tliat lacustrine conditions were established, or reestablished, relatively late during the time of custatic rise of sea level. Brooks (1969) speculated that the Izabal Basin originated as long ago as the Miocene, and pointed out, on the basis of his inability to

PAGE 40

U X

PAGE 41

23

PAGE 42

24 find salt traces in t!ie interstitual waters of sediment cores, that marine conditions have been absent in the recent past. The shallow littoral zone of the lake is narrow and subject to the abrasive action of waves. In some of the protected areas, especially along the north and east shores, the forest grows to the water's edge. The shallow bottom consists of large pebbles where the submerged aquatic macrophyte, Vallisneria , grows in sparce densities. Sandy beaches are more common along the south shore where there is greater exposure to wave action created by prevailing northeasterly winds. Occasionally isolated individuals or aggregates of water lettuce ( Pistia stratiotes) washed in from the shallow lagoons and black-water creeks of the delta region can be seen floating on the surface of the lake. A limnological survey of the lake was made by Nordlie (1970) in August 1969 and March 1970. His most surprising disco'/ery v;as the relatively high densities of a member of the Tanaidacae, a bottom-dwelling crustacean with marine affinities. Nordlie found the plaiiktonic community to have only moderate primary production. On his March visit he observed extremely low densities of net phytoplankton, whereas in August, they were present in "bloom" densities. Since most of the present study contains detailed seasonal descriptions of the same phenomena that Nordlie observed on his visits, his findings will be discussed in more detail in the chapters that follow. The Fisheries and People The Izabal Basin has been, until recently, a region of only modest human activity. The lowlands near the lake were believed to have been

PAGE 43

settled sparsely durina Maya tinies (Voorhies 1969). Major ceremonial centers were completely absent from this area although there were centers immediately to tlie south (Cop^n) and to the north scattered throughout the Peten. During the current century, the population did not increase substantially until malaria control was available and a road was completed from Coban to El Estor in 1948. This route replaced, at least during the dry season, the older loute up the Rio Polochic to Panzos where the Verapaz Railway connected with Pancajche. A road terminating at Pancajche completed the journey to the central highlands of Guatemala. Prior to this, 19th-century ships sailed in from the Caribbean to the colonial town of Izabal on the south shore of the lake. From this location, mule transportation provided an overland access to the central highlands. In the last decade, a spur road was completed, connecting the village of Mariscos on the south shore with the Atlantic Highway that couples Guatemala City with Puerto Barrios and Puerto Santo Tomas de Castilla on the Caribbean. Now the people living around the perimeter of the lake who have access to dugout canoes with outboard motors can reach the road terminating at Mariscos. From there the bus trip to Guatemala's capital city lasts only 4-5 hours. Likewise, the access provided by the lake to its perimeter opened the slopes of the Izabal Basin to agriculture. Immigration and Labor Alternatives The influx of Kekchi Indians from the upper Polochic Valley, and of people from other parts of Guatemala, was initiated by the discovery of a rich nickel deposit near the nortliwestern region of the lake. All the exploratory work for mining the ore has been completed, but the construction of tlie jirocessing plruit for the extraction of tlic mineral is still pending

PAGE 44

26 financial support and legal agreements. The already massive capital outlay, made possible through subsidization liy N'orth Ajnerican firms, has dwarfed other business interests in the valley. In tlie past 10 years the town of El Estor has grown from a sleepy Indian village to a bustling community. The labor force required for the initial clearing and exploration of the mining area is now largely unemployed. Altliough some families have been forced to leave, many remain with the hope that they wilJ again work for the attractive wages paid by the mining company. Some of these desperate people have shifted their means of living to agriculture and fishing. Carter (1969) in an anthropological monograph on the Kekchi cultivators described the problems and successes of these Indians in their efforts to applv highland methods of shifting cultivation to the lowland areas near the lake. The fishing, which became more effective with the introduction of nylon gill nets in the early 1960's, has since become so unprofitable that it now offers emplo>'ment and income for only a few dozen individuals around the lake. Holloway (1948) recognized long ago the potential food source of the predominately marine fishes that inhabit the lake, but made no prediction as to what the carrying capacity of this resource might be. Fishing Regulations and Yields Current Guatemalan fishing regulations, passed into law in 1936, apply to all inland waters regardless of size or location. One of the restrictions of the law is that gill nets can be no greater than 36 meters in IcH'.'tli nor Iiavc a mesh size ol" less than 7 iiiclics. However, gill nets

PAGE 45

27 as long as one kilometer with 2 1/2-inch mesh are frequently used during the fishing season. Initially, excellent yields from the relatively virgin fisheries were the incentive for people with capital to invest in this equipment. However, the subsequent decline in yields possibly may have been the result of exceeding the carrying capacity of the fisheries. The fishermen react to the decreased yield per unit effort in several ways. Some cease fishing altogether, while others send their equipment and hired labor northward to fish in the Rio de la Pasion in the Peten when the demand preceding Easter forces prices upward. Most owners of gill nets have several thousands of dollars invested in equipment, but fishing is usually subsidiary to their main business interests. The general disregard for the irrelevant and unrealistic laws favors individuals capable of making large capital investments and penalizes to exclusion those individuals fishing at a subsistance level. Dickinson (in press) has discussed in detail the sociological implications of the fisheries as well as thoroughly documenting the geographical features of the region. Since there is no governmental agency to keep records of inland fish catches, past yields from the lake cannot be estimated. Carr (1971) estimated current yields during a monthlong study by observing the weight of catch per unit length of net. He calculated a daily mean freshweight of 11.6 pounds per 100 yards (5.75 kg/100 m) of net, and based on other assumptions and measurements, estimated the lakewide yield to be 9 ,808 kg dry weight per week during the 1970 season. Some fishermen Most of the catch is salted and dried, accounting for 59". loss of fresh Kcigiit. The>are marketed bc\-onJ the Izabal Ba;;in, and only a small percentage of the total catch remains for local consumption.

PAGE 46

28 fish year round while others may be active only when fish prices are high (2 or 3 months of the year) . If an average fisherman worked 5 months of the year, then the lake would yield 56,160 kg dry weight annually (0.071 g/m yr) . Fish Behavior and Fisheries Management No data exist on the standing crop of fish, on growth increments of the population or on recruitment from immigration. Since much of the catch consists of euryhaline marine species, immigration is probably an important factor to consider in fisheries management. If the fisheries resource of the lake is being overexploited, as it presumably is (Carr 1971), then the immigration route through the Rio Dulce--El Golfete region is a logical area for control measures. Disgruntled fishermen that fish only on the lake are aware of this migration route, and of the gill nets set across these routes by their counterparts in the Rio Dulce region. Fishermen as well as large flocks of cormorants, terns, gulls, and scattered pelicans find another popular fishing region in the shallow coves and black-water lagoons of the Polochic Delta. There the fishing is seasonal for both man and birds. During the dry season, these waters are stagnant and perpetually in bloom with high densities of phytoplankton (Brinson 1973). The increased consumer activity, in response to tlie highly concentrated food source, coincides with the season when fishing is legal. As law dictates, the fishing ceases in June when heavy rains mark the beginning of the wet season. Ironically, tliesc events coincide with the migration of fish to the open water of the lake where they become move dispersed and harder to catch. Fisli-cating birds also disperse.

PAGE 47

29 and only a few pelicans remain. The fishing laws thus present a paradox by allowing fish to be caught when they are easiest to catch, and by timing the open season with the high prices preceding Easter which provides incentive for economic gain. The black anaerobic waters that flush the delta are not completely devoid of fish. Gobies ( Gobionellus sp.) which are particularly poor swimmers, can be seen gulping air under the sudd vegetation bordering the lagoons. Vultures and egrets prey upon these fish, but the main predators are schools of large tarpon ( Megalops atlantica ) that travel the backwaters of the delta. By being facultative air breathers, the tarpon are well adapted to the anaerobic environment. Marine fish, such as the tarpon, make up the majority of the fish caught in the lake. In order of decreasing yield they include Chlo roscombrus sp. ( zapatero or leather-jacket jack), several species of catfish including Bagre sp. and Arius spp. ( vaca and chunte ) , and the prized Centropomus undecimalis ( robalo or snook) . The presence of schools of sardines, anchovies ( Anchoviella) , and other small herbivorous fish apparently provide much of the food for the larger carnivores. The important fresh-water fishes include Cichlasoma gutulatum ( mojarra) and Brycon guatcmalensis ( machaca ) . Some other predominately marine animals, although not directly important to the fishing economy, are conspicuous components of the fauna. Blue crabs ( Callinectes sp.) are occasionally caught in nets, and the barnacle, Balanus improvisus , attached to pilings in the eastern end of the lake arc testimony of the seasonally brackish water in that area. Porpoises ( Tursiops truncatu s) apparently do not enter the lake, but do

PAGE 48

30 follow the front of high-salinity water as far as the upper Rio Dulce. Man has noticably reduced the abundance of some of the large aquatic vertebrates which probably has altered aquatic food chains. Both shark ( Carcharhinus leucas ) and sawfish ( Pristis perotteti and P_. pectinatus ) are reported to inhabit the lake (Thorson et al. 1966) although their presence has gone unnoticed for the past 8 to 10 years. Fishermen assert that nets frighten sharks , and that they have been absent from the lake since gill nets became prevalent. Crocodiles ( Crocodylus acutus) , once conspicuous carnivores of the aquatic community, have suffered considerable reduction of their populations as a result of hunting pressure. Hunting may have also been responsible for reducing the manatee, a large herbivore, to its present day low population density. Two species of turtles, Dermatemys mawi and Pseu demys scripta o rnat a, are occasionally caught by baited hook or by gill nets. However, they seem to be relatively abundant and could serve as a potential source of food for the local human population (Carr 1971) . Based on the consequences of exploitation of other components of the aquatic community, the "potential food source" offered by turtles would be short-lived, at best. Undoubtedly, the boney fishes will continue to be the principal aquatic resource for human exploitation. Considering the migration routes and seasonal activity of the fishes, it is doubtful if enforcement of the law during tlie veda , or prohibition period, would significantly relax fishing below its present intensity. Complete enforcement of the law, which would limit the length of gill nets to 36 meters, would paralyze the fisheries completely and be socially disfunctional (Dickinson, in press). There is an urgent need for fisheries

PAGE 49

31 management on Lake Izabal. Any management., however, must demonstrate an understanding of the role of the fishes in the ecosystem as well as the needs of the commercial and subsistancelevel fishermen. Undoubtedly some sacrifices in the habits of the fishermen would be necessary before the fisheries could reach a steady-state level of maximum sustained yield. Proper management of the fisheries may necessarily extend beyond the boundaries of the aquatic ecosystem if regional coupling mechanisms are operative as hypothesized m the Introduction.

PAGE 50

HYDROLOGY AND WATER CHARACTERISTICS The principal objective of monitoring the hydrological regime of the Izabal Watershed was to enable the calculation of rates of inflow and outflow of organic matter to and from the lake. Subsidiary to this purpose was the need to characterize the hydrologic properties of a lake that historically has received almost no limnological attention until recently (Brooks 1969; Nordlie 1970) . Even these recent studies lack the perspective of a long-term study necessary for a fu]ler understanding of the ecological implications of a seasonal hydrological regime. For example. Brooks (1969) labeled the likelihood of brackish water penetration into tne lake as a ''common misconception" although its occurrence is common knowledge among the non-scientific local inliabitants. Other "anomalies" of the hydrology and water characteristics might be of ecological significance, and their occurrence, if undetected, would unknowingly detract from a more complete understanding of the Izabal Watershed ecosystem. Information on runoff characteristics of watersheds in the humid tropics has been approached mainly by calculating the excess of precipitation over evapotranspiration from empirical formulae applied to rainfall and ajnbient temperature. Direct measurements of runoff are few, and tl\e information presented in the following chapter should be valuable for comoarison with studies that do exist. 32

PAGE 51

33 Metliods General Methodology--Field Stations, Logistics, and Schedule The sampling stations are indicated in Figure 4 and encompass nine river-mouth stations, three main lake stations, and the outflow station at San Felipe. To achieve the objective of the study of estimating the lake's organic matter budget as modified by seasonal changes, samples were collected during both the dry and wet seasons. The dry-season data were collected during the months of March, April, and May. The clear, cloudless days during the warmer dry season helped to distinguish it from the wet season. January, February, and June through October were considered wet-season months because the average local precipitation was greater than 100 mm per month. The rivermouth stations v/ere sampled to detei'mine the quality and quantity of water entering the lake and the outflow station was sampled to determine the quality and quantity of exports. Lake stations A, B, and C were sampled for chemical oxygen demand (COD) and biological oxygen demand (BOD) as well as for light and dark bottle primary production determinations (Table 1). Concurrently with COD collections, the following data were recorded from surface and bottom samples: dissolved oxygen concentration, pH, total alkalinity, and temperature. Net plankton was collected from Stations A, B, and C at approximately three-to four-week intervals. On tlicsc collection trips temperature profiles and Secchi disk transparencies were determined at tlic three stations as well as at stations midway between them, lliese intermediate stations are referred to in the text as A-B and B-C. The large size of Lake 1-abai and llic ]t)iig distances between sainpiiiig stations required the use of a relatively fast and reliable mode of

PAGE 52

34 Table 1.Sampling dates for chemical oxygen demand (COD), biologscal oxygen demand (ROD), and primary productivity experiments (Prod) . Numbers represent day of month

PAGE 53

35 Table 1.extended

PAGE 54

36 transportation. For this a 25-hp outboard motor was brought from the USA and a 24-foot dugout cance (25 hundredweight capacity) vv-as purchased at the lake. Routine sample collection required approximately seven hours for stations at the western end of the lake in the Folochlc delta. For the remaining stations to the east of El Estor, the samples were collected on another day as ten hours were required due to the longer distances involved. Ice could be purchased for storing samples during collection trips except during July and August when the ice plant closed for a minor repair. Several weeks of research time were lost due to breakdowns of tlie outboard motor and the car, which was used to transport equipment to the dock. At least monthly trips to Guatemala City were necessary to purchase repair parts and laboratory supplies. A room of a concrete-floored house approximately ]/2 km from the dock site was used for the laboratory. Electricity was available three or four hours a day after dusk from a privately owned diesel generator whose output ranged between 85-105 volts A.C. The low and variable voltage was insufficient to operate the spectrophotometer in spite of the constant voltage supply transformer. For this reason, chlorophyll determinations were not made. Hydrologic Measurements Discharge rates of several of the major rivers entering the lake were measured at montlily (occasionally twice-monthly) intervals from November 1971-Octobcr 1972. The f loat-and-dye method (Welch 1948) was used to determine the velocity by sprinkling fluorescein dye powder on

PAGE 55

37 tlie water surface and recording the time elapsed in traveling between two floats 50 m apart. When the velocity was slow (less than 0.3 m/sec) a 25-m distance was used. This procedure was repeated three times and the velocity was estimated to be the mean of the three readings. Measurements were determined far enougli upstream from the river's mouth to avoid backforce from the river entering the lake. The cross-sectional areas of the rivers were measured by determining the width and five depths (two near the banks, one in the middle, and two halfway between these) . A nail was driven into a tree trunk on the river bank at each station to provide a permanent point of reference for measuring changes of level in the river. Discliarge rates (cross-sectional area multiplied by velocity) for all rivers, with the exception of the San Felipe outflow, were multiplied by a factor of 0.8 in order to correct for frictional resistance of the stream bed and banks (Welch 1948). At the San Felipe outlet where frictional resistance is small because of the large cross-sectional area, a factor of 0.9 was used. Lake levels were recorded at frequent but variable intervals throughout the study year on the municipal dock at El Estor. Two other stations served as level markers on the lake: one on the western end at the mouth of Rio El Padre Creek, and the other at the eastern outlet at San Felipe. The EXWIBAL mining company also recorded lake levels at the plant site 2 km west of Las Dantas. Temperature profiles were recorded with a YSI Model 51A oxygen meter and a YSI 5419 oxygen/temperature pressure-compensated probe with a 50foot lead. Calibration was performed in the field for each profile, using a mercury thermometer as the standard.

PAGE 56

38 Analyses of Water Characteristics Samples collected for analyses of pH, total alkalinity, and specific conductance were usually the same as those used for COD analyses. Methods of collection are described in the section "Metabolism and Organic Matter." Samples for determinations were collected with a 2or 5-liter Van Dorn bottle, transferred to 300-ml BOD bottles, and fixed in the field. Temperature was recorded with a mercury thermometer while the water was in the Van Dorn bottle. All other procedures and determinations were made in the laboratory, usually within seven to ten hours from the time the first sample was collected. Total alkalinity and pH were determined first using a Beckman Model N2 pH meter and Beckman glass electrodes. Calibration was performed with factory-prepared buffer solutions (pH 7.0 and pH 4.5). Total alkalinity (carbonate plus bicarbonatej was determined by titration of duplicate 100-ml aliquots to pH 4.5 with a 0.02N HCl solution. The pH meter scale could be read to an accuracy of 0.5 pH unit. Specific conductance was determined with a Beckman Model RB3 SoluBridge and readings were adjusted to 25 C. Two probes were necessary for the range of conductivities encountered. A small, more sensitive probe was used in the range of 50-800 ymho/cm and a large probe, onetenth as sensitive, was used in the range of 700-40,000 ymho/cm. Figure 5 illustrates the calibration curves determined with dilutions of a standard KCl solution of known specific conductance (Golterman 1969). Readings of lake water, all within the sensitivity range of the small probe, were corrected by adding 20 pmho/cm, based on the calibration curve. Higher conductivity readings, accomplished with the large probe.

PAGE 57

o

PAGE 58

40 (UJD/omjjr1)o>i pjepue^s o o o CO (uuD/ogiuri) \o>\ pjepuE-^s

PAGE 59

41 required less precision sJiice detecting differences in the salinity gradients was more important than acquiring absolute values. Nevertheless, the calibration curve is in close agreement with the standard KCl solutions (Figure 5) . Samples for mineral analysis were collected April 14 and October 23, 1972. The samples were prepared by filtering the water through membrane filters (47 mm-diameter, (J.bO-y nominal pore size) to remove the seston. The samples were transferred to 0.5-liter bottles and 1% formalin was added as a preservative. The April samples were stored in polyethylene bottles and the October samples in amber glass bottles. Both groups were flown to Gainesville and mineral analyses were performed during March 1973. An atomic absorption spectrophotometer (Perkin-Elmer 303) was used for determination of Ca, Mn, Mg, Si, Fe, Ni, Zn, and Al . Potassium and Na were analyzed with a tlame emission spectrophotometer (Beckman DU) , phosphorus by colorimetric determination with molybdate (Golterman 1969), NO and CI with specific ion electrodes in conjunction with an Orion Potentiometer Model 801.' Results and Discussion River Discharge and Runoff Discharge rates from six rivers were used to estimate runoff from the Izabal Watershed. All values were calculated from velocities and cross-sectional area measurements of the rivers (Appendix, Table C) except for the values recorded for the Rio Polochic distributaries and the Rio San Marcos in August. These low August values are noteworthy since they occurred during the month of heaviest rainfall. During this

PAGE 60

42 period the level of the lake had risen to as high as 1.15 m (August 21) above the lowest level recorded during the preceding dry season (Figure 6) and resulted in flooding of the delta areas. During flooding the rivers were not confined to channel flow, but moved as a sheet across the deltas. If August discharges had been calculated from velocity measurements in the river channels, gross underestimates of runoff would have resulted. To estimate the August runoff, linear regression formulae were calculated from the other measurements using rainfall at Las Dantas as the independent variable and monthly discharge as the dependent variable. In this way August discharges could be calculated by extrapolation from rainfall, assuming a linear relationship between discharge and rainfall. Values for other months were similarly calculated where data were missing (e.g. February). These extrapolated values appear as open circles in Figure 7 and were used for the August values in Figure 8. Discharge rates at the moutlis of the Rio San Marcos and Rio Sauce were low compared to the other rivers measured (Figure 8) . The similar patterns of discharge for the Rio Polochic distributaries (Comercio, Coban, and Bujajal) can be attributed to their common origin. Whereas the Polochic distributaries demonstrated a sharp increase in discharge during June, an increase of similar magnitude for Rio Oscuro did not occur until July. This can be attributed to differences in local rainfall regimes. During the year of study the wet season in the area of tlie lake began in June, which accounted for the July increase in Rio Oscuro "s discharge (Figure 8) . The wet season in some regions of the Polochic Valley began in May,wluch may have accounted for the June increases in the discharges of the Polochic distributaries.

PAGE 61

Figure 6.Monthly estimates of all inputs to the lake (runoff and direct rainfall) and the monthly averages of lake water levels. Lake levels are the number of meters above the lowest recorded week which occurred during the second week of March 1972.

PAGE 62

44 M J J A MONTH

PAGE 63

Figure 7.Linear relationship between monthly rainfall and monthly discharge of rivers emptying into Lake Izabal. The open circles represent extrapolated values for February and August for which discharge data were missing or required correction.

PAGE 64

46 600 ConnercioJ 400 E E 'ct a: 200 y:>67.99* 2.06 X Coban y =-75.23 *0.74x y =-31.17 1.12 X 200 400 600 800 200 400 Discharge (m3 x lOS/month) 600 Oscuro y:--103.08*0.75x -1 1 I L. y =146.85* 3.92 X San Marcos •y:6.09* 1.81 X O 200 400 50 100 20 40 Discharge im^xlO^/month)

PAGE 65

Figure 8.The seasonal march of monthly discharge rates for some major rivers emptying into Lake Izabal.

PAGE 66

48 JFMAMJJASOND MONTH

PAGE 67

49 Table 2 summarizes the discharge rates of all rivers and watersheds draining into Lake Izabal. Included is runoff from the watersheds to the north and south of the lake which was estimated from the runoff of the Rio Sauce and Rio San Marcos watersheds. The annual discharge from rivers emptying into the extreme western end of tlie lake was greater than ten times the contribution of the remaining watersheds. The total runoff volimie of all watershed areas into Lake Izabal was estimated to be 13,290.9 m"^ x 10 during the year of study. Instead of expressing runoff as a voKune, it can be compared directly to rainfall by conversion to equivalent units. This was accomplished by dividing the known runoff volume (m ) by the area of the watershed (m ) (Table 3). The annual runoff value of 8,253 mm calculated for the Rio Oscuro watershed v;as much too high since it is unlikely that rainfall ever reached this value in any part of the watershed. The reason for this overestimate was that at flood stage, the Rio Polochic apparently overflowed into the southern branch of the Rio Oscuro headwaters (Riachuelo Suncal) . This resulted in high discharge rates for the Rio Oscuro due partly to water originating from the Polochic watershed. By 2 combining the Rio Polochic and the Rio Oscuro areas (5,480 km ) and their annual runoff volumes (12,112 m x 10 ), the combined runoff would be 2,210 mm, a more realistic value. Runoff for all watersheds averaged 1,957 mm. Since the average rainfall for the Izabal U'atershed was 2,992 mm, the portion of the precipitation lost as runoff was 65'o. This is in close agreement with the watershed of Lake Lanao, Phillipines (Frey 1969) which loses 67% of the 2,873 mm rainfall it receives. Other tropical waterslieds receiving

PAGE 68

50 •t-> o o

PAGE 69

51 Table 3.Sununary of hydrologic data for the major watersheds draining into Lake Izabal

PAGE 70

52 less precipitation lose between 40-50-6 of the rainfall as runoff (Golley et al. 1971). Snedaker (1970) calculated runoff for Finca Murcielagos by a simple method devised by Holdridge (1967) which requires knowledge of only mean annual biotemperature and annual rainfall. Using this method, Snedaker 's runoff estimate was 902 mm or 45% of the rainfall. This is in close agreement with my value for the Rio Sauce watershed (Table 3) which is located on the north shore of the lake near Murcielagos. During the period of study, runoff was 857 mm or 39% of the rainfall (Las Dantas records) . W ater Budget The water budget of the lake was calculated at monthly intervals from inputs by river runoff (already discussed] and by direct precipitation, in addition to the outputs by evaporation and losses though the San Felipe outlet. The monthly contribution by direct precipitation was calculated from the average rainfall at Las Dantas and Mariscos. Since evaporation from the free water surface of the lake was not measured, a literature search was made to arrive at a reasonable estimate. Free surface evaporation estimates from tropical latitudes appear to be few and the data presented in Table 4 include values for more northern latitudes. However, summer climate regimes, especially in Florida, approximate the year-round climate of the Lake Izabal region. The value of 161 mm/month was chosen as the representative evaporation by averaging the estimates from the humid areas of Lake llelenc, Anderson Cue, Lake Michie, Lake Chad, and the Caribbean Lowlands (Table 4). This

PAGE 71

53 Table 4.Monthly free water surface evaporation measurements for selected warm climates and warm seasons Region or Lake Monthly Evaporation (mm) Year Source Lake Helene, Florida Anderson Cue Lake, Florida' Lake Elsinore, Calif. Lake Tiberias, Israel Polish Lakes Lake Chad, Africa^ Caribbean Lowlands Lake Michie, N.C,^ Lake Colorado City, Texas^ East Africa (Nile Region) 128 1962 Pride et al. 1966 155 1966-68 Brezonik et al. 1969 231 Szeicz ^ Endrodi 1969 193 1949 Reiser 1969 140-180 several Debski 1966 188 Grove 1972 174-254 Ray 1931 118 1962-64 Turner 1966 221-251 1955 Harbeck et al. 1959 90-120 Tailing 1966 Based on an average of 5 warmest months (MaySeptember) Based on an average of 4 warmest months (May-August) . Annual total divided by 12.

PAGE 72

54 ft T is equivalent to 115 x 10 m /month for a surface the size of Lake Izabal and will be considered constant throughout the year. Changes in lake level (Appendix, Table D, see also Figure 6) represented integrated results of both inputs and outputs. The volume of the lake was calculated from the bathymetric map prepared by Brooks (1969) and was estimated to be 8,300 x 10 m . Changes in volume were calculated by multiplying changes between mean monthly lake levels by the surface area of the lake. Table 5 is a summary of the monthly contributions and losses for Lake Izabal. Adding the inputs from runoff and direct precipitition, subtracting the evaporation, and subtracting the positive or negative change in volume resulted in the final value which estimates the monthly loss t'nrough the San Felipe outlet. Direct measurement of discharge from the San Felipe outlet was inadequate due to extreme variations in surface velocity. On one day I even observed that the flow had reversed, and further inquiry led to the conclusion that this was a common, if not daily phenomenon. Apparently prevailing northeasterly winds are capable of shifting the leeward level of El Golfete above the level of the lake, thus generating the variation in discharge or reversal of flow at San Felipe. In spite of this difficulty, there was sufficient agreement between the velocities measured in the field and those calculated (Figure 9) to regard the latter values as good estimates of discliarge at San Felipe. The direct field measurements were multiplied by a factor of 0.9 to correct for frictional resistance of the banks and bottom (Welch 194 8) . The ainiual water budget for the lake and the watcrslicd is summarized in Figure 10. The average residence time of the water in the lake was

PAGE 73

55 CN u 6) O +J o o I 1—1 r^ U 0) i > o 2: C o VD to +-I

PAGE 74

Figure 9.Relationship between velocities at the San Felipe outlet from direct field measurements and those calculated by balancing the water budget. Perfect agreement between the two estimates would fall on the dashed line. Direct field measurements were multiplied by a factor of 0.9 to correct for frictional resistance of the banks and bottom.

PAGE 75

u-8

PAGE 76

o in 1/5 u w •p H c 3 o r-l 1 o
PAGE 77

59
PAGE 78

60 calculated by dividing its volume (8,300 x 10 m ) by the 12-month loss by direct evaporation from the surface and the outflow at san Felipe 15,143.5 X 10 m ). This estimated the residence time of the water to be 6.6 months or 0.55 year. Water Characteristics Lake Izabal, like other lakes that lack sufficient depth to develop a hypolimnion, would be classified as a third-class lake (Hutchinson 1957) and could fall, functionally at least, into the class of pol}TTiictic which is usually reserved for lakes of high mountain regions in equatorial latitudes. Because of Lake Izabal 's large area, shallow depth, and liigh influx of relatively colder waters from rivers, some anomalous and interesting temperature patterns develop. The Izabal Vv'atershed lies in an area of diverse geological formations, and as a result, receives a variety of water types. Partly because of the large drainage area of a single influent river, the Rio Polochic, much of the water from the various rock types has already mixed before discharging into the lake. Regardless, the large deltaic swamp at the western end of the lake has distinct modifying effects on some of these waters. The proximity of the lake to the sea coast and its connection with a marine environment cannot be overlooked as a potential factor that could influence the ionic composition of this freshwater lake. Thermal ]n-operties and circulation patterns in tlie lake Temperature profiles were recorded at approximately monthly intervals at five equidistantly spaced stations along the length of the lake (Figure

PAGE 79

61 4). Figures 11 and 12 illustrate the temperature profiles for each of the five stations (A, A-B, B, B-C, and C) . For most stations and dates of sampling, the profiles were isothermal for the first 10-12 m, with the exception of the upper 2-5 m vvhich were heated directly by solar radiation. This surface stratification was only temporary and disappeared rapidly either at sundown or with windy conditions during the day. More persistant thermal stratification often occurred in the bottom 1-2 m. This was most pronounced at Station A after June when the colder waters from the nearby Rio Polochic created a density current resulting in a cooler layer along the bottom. At Station A-B this seasonal change was more noticable, as the profile is isothermal until June, the beginning of the wet season. Stations B and B-C showed the same characteristics but the thermoclines were not as sharp as the more westerly stations. Station C was located 30 km east of the Rio Polochic and little influence from the density layers was expected. The stratified layer found on April 8 can only be explained if the 2 C increase in temperature of the upper column (since March 16) failed to circulate with the bottom meter. More interesting was the noticeable increase in bottom water temperature on June 15, due to the arrival of a warmer but denser water mass of high specific conductance (465 ymho/cm at 15 ra and only 248 ymho/cm at tlie surface). This warmer water originated from the brackish conditions that developed at the San Felipe outlet during the proceeding dry season (p. 89). A dr>--season increase in temperatures was recorded for all stations from a low in February to a maximum in June. For most of the

PAGE 80

Figure 11.Vertical temperature profiles of Lake Stations A, A-B, and B recorded at approximately monthly intervals.

PAGE 81

63 Or 2 4 Q 101214o2 ^6 Q 8 Q 10112 14 STATION A O90O 0746 0620 0630 0656 0700 -SMjr BADT aiApr i3May UJun yjul ^ y ?7 ge ?9 g9 30 ?9 y> ? e 30 27 STATION A-B 072O 07J0 0600 0847 0755 0830 23May 13Jun 9Jul 19 Aug 17Sep SOrt 30 31 29 31 28 30 29 3.3 28 30 28 30 ~1 0820 26 on 26 28 STATION B 0952 1210 1015 0945 0600 COX 3Jan 25Feb 16 Mar 8Apr 23May i3Jun ^t l^ 25__27^ 2926 28 28S30313031 0930 19 Aug 28 30 7

PAGE 82

+J

PAGE 83

65 •0m « — •— • — o— -•~~9 » • • O • •— 9— •— • » — • •— o_ IPS -»— • — e — •— • — • — • — • « — « — ^ -• — • — • — ^ » — • — • — « — • — • — e — e — •^"^ .„ » — « — • — • — •— » — • — • — • — •I I I I I i I I I I 1 1 1 1 1 O C\j q^ (D 00 O C\J 1 1__1 1 I I J I I L. J 1 I i I C\J CT) C» O CNJ

PAGE 84

66 vvater mass, the mininiiijT. was approximately 25.5 C and the maximum 30.4 C, a difference of 4.9 C. The seasonal pattern of water characteristics The water samples were arbitrarily divided into four groups-lake stations (A, B, C and San Felipe), swamp waters (Oscuro, Amatillo, El Padre), Polochic distributaries (Comercio, Coban, Bujajal), and small rivers (Sauce, San Marcos, Manacas Creek, Tunico) . These groups are based not on water characteristics per se, but on the origin and locality of the water. The sarne divisions will be recognized in the treatment of the organic matter data. The results of measurements of pH, total alkalinity, specific conductance, dissolved oxygen concentration, and temperature are presented graphically by station to illustrate the magniti'de of sers^rni rhTPges (Figure 13-16). T-ihle 6 summarizes the ranges of the extremes measured as well as the approximate ranges for m.ore representative values. Compared with the other groups, the lake stations (Figure 13) were least variable seasonally for all parameters except specific conductance. This exception was due to the upstream movement of brackish water from the Pdo Dulce, through the San Felipe outlet, and into the lake. Total alkalinity showed little seasonal change. Some stratification was noted at Station A due to the density current created by the Rio Polochic waters of slightly higher total alkalinity. A decrease occurred from above pH 8 in January to below pH 7 in March, and then sliowcd a trend toward increase after July. Temperatures increased during the dry season and began to decrease after June. Dissolved oxygen concentrations at rlie surface probably varied daily as much as tliey did seasonally.

PAGE 85

67 Table 6,Ranges of temperature, pH, total alkalinity (T.A.)j specific conductance, and dissolved oxygen saturation for all stations sampled at Lake Izabal Most Values Extremes Measured Lake Stations (A, B, C, & San Felipe) Temp. (C) pH T.A. (meq/ liter) Cond. (ymho/cm) O2 (% Sat.) 26.0-30.0 6.00-7.00 1.70-1.80 175-200 80-105 24.1-31.4 5.60-8.25 1.60-2.00 150-465^ 8-107 Swamp V/aters (Oscuro, Amatillo, El Padre Cr.) Temp. (C) pK T.A. (meq/liter) Cond. (limho/cm) 25.0-31.0 5.50-7.00 1.00-2.10 100-200 o-]no 23.8-31.3 5.20-7.30 0.82-2.37 65-230 0-113 Rio Polochic (Comercio, Coban, Bu j a j a 1 ) Temp, (C) pH T.A. (meq/liter) Cond. (ymlio/cm) O2 (% Sat.) 23.5-30.5 5.50-6.75 1.50-2.00 170-225 75-95 23.0-30.9 5.20-8.00 1.59-2.19 157-260 47-101 Small Rivers (Sauce, San Marcos Manacas Cr., Tunico) Temp. (C) pH T.A. (meq/liter) Cond. (ymho/cm) O2 ("6 Sat.) 25.0-29.0 6.00-7.00 1.00-2.75 100-250 90-100 24.2-32.8 5.75-8.20 0.80-2.82 79-268 84-105 The highest conductivity was recorded at San Felipe (5,000 ymho/cm) due to a localized brackish water m,ass.

PAGE 86

X

PAGE 87

69 CO (D ^ C\J NaOAXO oo CO CO OJ CO ^ CM ^ O O) dHTil V 1 'aNOD S03 o Hd

PAGE 88

70 Stratification of oxygen is highly variable depending on the frequency and the depth to wliich vertical circulation occurred. The swamp waters (Figure 14} became diluted at the beginning of the wet season as reflected by the decreases below dry-season values in conductivity and total alkalinity after June. The pH was variable, but tended to decrease during the dry season. Both temperature and dissolved oxygen concentration decreased dramatically at the beginning of the wet season. The flushing of these rivers by the colder and poorly oxygenated waters originating in the surrounding swamp forest also destroyed the stratification that had been established during the dry season. The water of the Rio Polochic distributaries (Figure 15) was characterized by dry-season decreases in pH, increases in conductivity and tem-neraturCj little change in total alkalinityj and absence of low dissolved oxygen concentrations. All the small rivers (Figure 16), except for the Rio San Marcos, ceased flowing during the dry season. Thus changes at the beginning of the wet season are less marked in the San Marcos water than in the otiier small rivers. The seasonal trends were similar to those of the Rio Polochic distributaries described above. The implications of these seasonal changes in water quality for metabolic activity of the lake and plankton abundance will be discussed in the following sections. Mineral analyses a nd the Na:Cl ratio Samples of lake and river water that were analyzed for dissolved minerals were collected during the dry season (April 14, 1972) and the wet season (October 25, 1972). The results (Appendix, Table E) show

PAGE 89

X

PAGE 90

72 CO CD (J5)ii/6uu) N39AXO O O CO O ^, n c\) cvj c\J 0)dN3i CM C\J C\i ^ Vl AilAliDnaNOD r
PAGE 91

c u U (D B o •H !h O 3 -• 4-> U d) O o a, (D o H +-> P c o o o <4-l o (D •H Oj (-> -1 C +-> a> tn bO-H X! O ^1 o O (/I to Ph C 03 +-> >
PAGE 92

74 N39AXO 0)dN3i 'Vl QNOD Hd

PAGE 93

^,

PAGE 94

76 I/) o (J DC < < in O U iTi < < o a: UJ u < O q: _i I I J — I — i-I I I l__l L. \ -\ J I I L. 2 o c o (/) E c "D C o to rd E E T3 C o E CO E CO CD O

PAGE 95

77 that Mil, Fe, Zii, and Ni were detected only occasionally, and always at the lov;er range of the sensitivity of the tests. Attributing significance to these results would be unwarranted. The presence of chloride apparently interfered with the measurement of nitrate for the brackish waters of the Rio Dulce Salt Spring and Amatique Bay (October 25) yielding suspiciously high values of 2.7 and S.5 mg/liter, respectively. Except for the Rio Agua Caldente and Rio Sauce during the wet season, nitrate was near the limit of detection (0.62 mg/liter) of the specific ion electrode. Dry-season concentrations of nitrate were greater than 1 mg/liter in some of the swamp waters (Oscuro Bay, Amatillo, El Padre Creek, and Ensenada El Padre) as well as the Comercio and Coban distributaries of the Rio Polochic. Dis?o1\'ed phosphate concentrations ranged between 0.04-0. 17E mg/liter and in this range of sensitivity the significance of the results is subject to question. The high concentrations of Ca, Mg, and Si in the Rio Agua Caliente can be attributed to the hot spring at its origin. However, the loiv discliarge of this river wouJd have resulted in less contribution to the lake of these ions than less concentrated rivers with higher discharges. The Rio Sauce drained a limestone area (probably dolomite) and differed from Lhe lake water by its higher Mg and Si concentration and lower Ca. Rio Manacas Creek during the dry season had a Mg:Ca ratio greater than one. The Rio San Marcos was notably more dilute t!ian the lake water as suggested by the consistantly lower specific conductance of the river throughout the year (Figure J6). Silica concentrations for the October 23 samples may be high due to storage for three months in glass bottles.

PAGE 96

78 The analyses for sodium and chloride in the lake and Rio Polochic waters, however, have some interesting implications for understanding circulatory patterns of the lake water. These data and the ratios of sodium to chloride are presented in Table 7. On April 14 both Na and CI were higher at San Felipe and Station C at 12.5-m depth than in the rest of the lake. The appearance of slightly brackish water was noted also with conductivity determinations at San Felipe in April but not until June at Station C (Figure 13) . Thus the mineral analysts of Na and CI provided a more sensitive method than conductivity measurements for detection of brackish water as it entered the lake from the Rio Dulce. The average Na:Cl ratio for the April 14 lake station samples (excluding San Felipe and Station C, 12.5) was 0.5?.. This value is slightly lower than the ratio 0.56 for sea water (Remane and Schlieper 1971). On October 23 the high ratios of 1.33 at Station A (11.5 m) and 1.69 at Station B (15 m) distinguished them from the ratios of the remaining lake stations. The origin of these high Na:Cl ratios is apparent by comparing them with the October 23 average of the Polochic samples which was 1.76. This provides a conclusive check for the existance of the density current implied by the 0_, alkalinity and specific conductivity stratification at Station A (Figure 13) after the initiation of the wet season. If the Rio Polochic had been the only source of water for the lake, then the lake water would have been more dilute than it was. Even on April 14, when Na and CI concentrations were higher in the Rio Polochic, they were not as high as the more dilute lake water on October 23

PAGE 97

79 Table 7.Sodium and chloride concentrations (mg/liter) and Na:Cl ratios for lake stations and distributaries of the Rio Polochic. Averages are calculated for selected stations 14 April 1972 25 October 1972 mg/liter Na:Cl Na CI ratio Lake Stations Station A -

PAGE 98

80 (excluding Station A-li.S m, and Station B-IS m) . "By examining Table E of the Appendix, there is no evidence for possible sources of high Na or CI from the rivers sampled, except Rio Agua Caliente. However, its wet-season discharge was low and the river did not flow at all throughout most of the dry season. Further confirmation is available from August 1969, when Brooks (1969) detected Na and CI concentrations in the Rio Polochic to be 3.2 and 2.2 mg/liter respectively. Higher concentrations (4.8 mg Na/ liter and 7.5 mg CI/ liter) were reported by Brooks for lake water samples. Brackisli water movement into Lake Izabal To determine the seasonality and extent of movement of the salinefresh water interface between Lake Izabal and Amatique Bay, several sampling trips were made into the Rio Dulce-Hl Golfete region (Figure 17). Two of these trips traversed the area between the San Felipe outlet and the coastal port of Livingston; one was made during the dry season (March 22-23, 1972) and the other during the wet season (October 26, 1972). A tliird transect (May 13) included only Stations 1-8. The sampling stations are numbered from 1 to 11 in Figure 17. On the first transect (March 22-23) a salt-water wedge was observed extending from Station 11 at Livingston to El Golfete between Stations 5 and 6 (Figure 18). The deep high-salinity water was nearly isothermal (29.4-29.6 C) and slightly warmer than the surface waters (Figure 19). A strong (ca. 0.6 m/sec) outgoing current in the upper 1-2 m of the Rio Dulce below Fl Golfete marked the interface with the relatively motionless deep layers. An increase in conductivity at Station 8 over a 17hour period was attributed to downstream tidol forces (Figure 20). This

PAGE 99

(D

PAGE 100

82

PAGE 101

U-, C CO 6 o >o o 1—1 3 Q O • H «; (1) ,c +j C o CM '

PAGE 102

84 UJ 'Hid3a

PAGE 103

Figure 19.Temperature profiles at two stations illustrating a slight temperature increase associated with the haiocline. On Station S, the halocline occurs at 5 m, on Station 9, at 5.5 m.

PAGE 104

86 TEMP °C 29 30 31 32 C 23. 4 6LU Q 910II,21314 .J MAR. 23. 1972 STA. 9 \ o 1 o ^3> o I 1 o I O 1 o liiiiiliipill«ili iW

PAGE 105

u

PAGE 106

88 o o o o" o CM

PAGE 107

89 resulted in an upward displacement of the halocline by approximately 2 Til, demonstrating the mechanism for upstream movement of the highsalinity water mass. On March 22 the specific conductance of the ground water (ca. 30-cm depth) of the swamp forest at Cuatro Cayos in the southwestern end of El Golfete was measured. The conductivity increased from the edge of the island to 5-m distance from the shore, and then decreased gradually to a distance of 55 m where the ground-water salinity was still higher than the surrounding El Golfete water (Figure 21) . The presence of this higher salinity water indicated that salt-water intrusion had occurred during dry seasons prior to 1972 and that the swamp forest, where som.e red mangroves were present, maintained the brackish condition throu.'^bont the wet season. On May 13 brackish water was detected as far upstream as Station 1 at San Felipe (Figure 22) on a transect which was conducted downstream to the upper end of the lower Rio Dulce at Station 8. The main mass of the salt-water wedge, determined at ca, 14,000 ymho/cm on March 22 at Station 7, had moved 7.6 km across the Golfete to Station 4 in 22 days, or an average of 345 m/day. The last dry-season measurements on June 13 included lake stations whose conductivities were measurably higher than previously recorded. At Station C the reading of 465 pmho/cm at the bottom (Figure 13) and other measurements above 200ijmho/cm were detected between San Felipe and Station C (Figure 23). 1 failed to detect a continuously declining gradient of conductivity between San Felipe and Station C on June 13. The presence of higli conductivity water at San Felipe (Station 1, 10 m)

PAGE 108

U)

PAGE 109

91 0001 iuo/o^mn 'AiiAiionaNOO

PAGE 110

o

PAGE 111

93 ^ 'HldBQ

PAGE 112

Figure 23.Conductivity profiles taken at several stations in Lake Izabal and the upper reaches of Rio Dulce (Stations 1 and 2) . Notice that a detectable increase in conductivity occurs as far as 12.2 km inside the lake at a time corresponding to the end of the dry season.

PAGE 113

!• 23 4 5 ICL 8910 II 12 ' h ^ 14154o — »o— o-oo 20 Km 7.5 Km June 13, 1972 Volue In Km represents distance of lake stations from Station 1. \ 12.2 Km o V STA. I o 100 1000 CONDUCTIVITY, ^Mho/cm

PAGE 114

96 on June 9 and its absence on June 13 (only 365 iJmho/cm) emphasized the haline discontinuity in this region. This can be attributed to the movement of discrete water masses and their eventual mixing with lake water which diluted them,, thus dispersing the denser high-salinity layers. The movement of brackish water into the lake was probably not a continual flow, but an occasional spill-over depending on wind-generated currents and tidal forces. Gravitational force also may have caused denser masses of high-conductivity water to sink to deeper regions of the bottom profile. The dry-season current at San Felipe shifts with the prevailing winds, lending support to the belief that the level of El Golfete and the lake were approximately the same. El Golfete (4.5-m depth) served as only a temporary barrier for the upstream movement of the salt-water wedge. San Felipe Bay constituted another barrier (10 m) between the deeper region of the San Felipe narrows (15 m) and the deeper lake (Figure 24) . On a given year, the amount of brackish water that crosses these barriers probably depends on the intensity and length of the dry season. Two local inhabitants informed me that on some years they have noted brackish surface water (by taste) as far into the lake as Finca Jocolo on the north shore (ca. 7 km from San Felipe) and Finca Icacal (ca. 8 km from San Felipe). Popenoe (personal communication) has detected a brackish taste in surface water as far westward as Zapotillo near El Estor. An even more widespread occurrence of brackish water is implicit in these observations based on the stratification patterns that developed during the year of study. This would suggest that during some years the lake receives much larger quantities of brackish water than was observed during the 1972 dry season. This may

PAGE 115

M^

PAGE 116

98 uoisSuTAT'] :;b jcg puss SDina "H a3M0T O^ISJTOD IH -adTiaj UBS 3 uoTaB:^s a uoiaBr^S V uot:;b:is sJtpBd IH BpBuasugo z. c o • H 4-> 0) +•> CO (m) qtldoQ

PAGE 117

99 explain ^
PAGE 118

100 during 1972. Over 90% of the runoff volume entered the lake through the Rio Polochic Delta at the western end of the lake, and was largely responsible for the relatively short residence time of 6.6 months for the water mass of the lake. These colder river waters created profilebound density currents that spread over the broad basin of the lake, causing temporary thermal stratification of the lower 1-2 m. The inflow of colder waters and the decreased insolation later in the wet season, resulted in a decrease of the temperature of the water column to ca. 5 C below the dry-season temperature maximum. Stratification was usually absent in the lower layers during the dry season, but present in the surface layers on calm days due to localized heating of the upper 2-3 m. Thus the lake could be classified as warm pol)nnictic. Whereas wet-season influences originated at the western end of the lake, dry-season influences were noticed at the opposite end. At the lake's outlet to the sea, where dry-season outflow was low and sometimes absent, high-conductivity water worked its way up from the lower Rio Dulce and over the barriers imposed by the shallow El Golfete and San Felipe Bay. This penetration of high-conductivity water was low during the year of study, apparently due to the brief dry season. However, during other years, the penetration may be much greater, and there is evidence that the concentration of Na and CI ions in the lake water, although more dilute tlian in most hard waters, is regulated by the seasonal influx of water of marine origin. Other than tliis, the ionic composition of the water showed no interesting anomalies, except for the high Mg:Ca ratios of some of the smaller rivers that apparently drain serpentine and dolomite form.ations.

PAGE 119

101 River discharge was greatest in August, the wetLest month, just three months after the beginning of the wet season. Runoff waters were most dilute during July and August. Although conductivity, alkalinity and temperature of the lake waters roughly paralleled the seasonal changes of the rivers, the magnitude of change was much smaller, owing to the buffering effect of the large water mass of the lake.

PAGE 120

NET PLANKTON AND BENTHIC COMMUNITIES An understanding of the seasonal changes in plankton abundance would require not only a constant monitoring of many environmental parameters, but also a knowledge of the requirements of each species and their influence on one another. Even under the controlled conditions of chemostats, this problem is formidable. The objective of following the seasonal changes in net-plankton abundance was not to acquire precise information on the population dynamics of individual species, but to gain insight on the gross seasonal periodicity of the biotic component of the lacustrine ecosystem. Only part of ilie plankton coiiuuunixy is sampled when collected witli a net, and some important plankters are probably unrepresented in the samples, while others, that may vary in size, could be underestimated due to selection for larger individuals. Thus, the net plankton does not represent a natural assemblage of organisms, but rather an artificially selected community whose lower limit in size is determined by the mesh aperture of the plankton net. By using vertical tows, the collection method integrates the numbers of organisms in the water column by sampling most of its length, but ignores vertical stratification wliich probably would be of less interest in a shallow lake than a deep one. 102

PAGE 121

103 Methods Phytoplankton and zooplankton were sampled 15 times from January 6 to October 26, 1972, at threeto four-week intervals at Stations A, B, and C (Figure 4). Three vertical tows were made at each station with a No. 20 plankton net (25-cm diameter) and the tows approximated the depth of the stations. At Stations A, B, and C, the tow lengths were 12.5 m, 14.65 m, and 11.80 m, respectively. The three tows were combined at each station and anestlietized with absolute ethyl alcohol at about 5% by volume. Upon return to the laboratory, the samples were diluted so they could be counted and were fixed with a buffered formalin solution (40% formaldehyde) at 2% concentration by volume. The volumes of the plankton tovs's were calculated as the length of the tows multiplied by the area of the net (0.049 m ). Counts were made of three one-ml subsamples of the thoroughly mixed samples. Most of the larger plankton were counted with a dissecting microscope at 45x magnification and a binocular compound microscope was used for counting rotifers (lOOx magnification) . Identifications were made with the aid of Ward and Whipple's Fresh-Water Biology (Edmondson 1959), Fresh-Water Algae of the United States (Smith 1950), and the Fresh-Water Invertebrates of the United States (Pennak 1953) . Pennate diatoms were counted as one unit per cell, and no attempt was made to distinguish species. Each Melosira granulata filament was counted as one unit as were Pediastrum simplex colonies, Anacystis c yanea gelatinous masses, and Eudorina sp. colonies. Calanoid and cylopoid copepods were counted as two groups; each individual was counted as one unit. Adults and copcpodids were not distinguished from each other.

PAGE 122

104 but the nauplii were counted seperately. Cladoccra were identified and counted to species except for the Bosminids; no distinction was made between Eubosmina tubiscen and Bosmina longirostris . Cladoceran eggs were counted as one group. Rotifers were identified and counted to genus or species. Individuals within the colonial forms were counted as discrete units. Collections of bottom fauna were made between October 10-16, 1972 2 with a 9 X 9 in (522.6 cm ) Ekman sampler. The contents of two grabs at each station were sieved through a No. 40 mesh screen to retain the organisms. Results and Discussion Phytoplan kton and Z o opj. ankt on Communi ties The algae were represented by several diatoms of the pennate tynpe and one centric, 3 blue-green algae, 3 desmids, 3 other species of green alage, and 1 dinoflagellate. Nineteen species of zooplankton were counted which included 2 calanoid and 2 cyclopoid copepods, 5 cladocerans, and 9 rotifers. Table 8 is a partial list of species. Diatoms Diatom populations, represented by the dominant S\Tiedra ulna and other less numerous unidentified species, were generally abundant throughout the sampling period (Figure 25) . A noteworthy feature of tlie seasonal change was that increases and decreases followed the same general trend for all stations. The maximum density was on July 29 when 421.5 organisms/ liter were present at Station C. The early high density at Station A (April 50) coincided with the presence of extremely high numbers of Mclosira granulata at that station (Figure 25).

PAGE 123

105 Table 8.Species list of net plankton frequently collected from Lake Izabal Algae Diatoms Melosira granulata (Ehrenberg) Ralfs "S>Tiedra ulna (Nitzsch) Ehrenberg Blue-green Algae Anacystis cyanea (Kiitzing) Drouet 5 Daily Anabaena f los-aquae (Lyngby) Brebisson L>nigbya sp. Green Algae Staurastrum pingue Telling • S. le ptocladum Nordst. var. denticu latum S. tohopekaligense Wolle Cosmarium sp. Pediastrum simplex var. duodenarium (Bailey) Rabenhorst Eudorina sp. Coelastrum sp. Pyrrophyta Ceratium hirun dinella [O-F. Mueller) Schrank Cladocera Bosmina longirostris (O.F. Mueller) Eubosmina tubicen (Brehm) Moina micrura Kurz Ceriodaphnia lacustris Birge Diaphanosoma brach)airum Copepoda Diaptomus dorsalis Marsh Pscudodiaptomus culebrensi s Marsh Mesocyclops edax (E.A. Forbes) Mesocyclops ( Thermocyclops ) inversus Kiefer Rotif era Brachionus falcatus Zacharias B. liavanaensis Rousselet Yeratella cochlcaris (Gosse) Filinia pej Icri Hutchinson Conochilus unicornis Rousselet Con ochiloide s dossuarius (Hudson) Sinantherina sp. Hcxarthra sp. Platyias sp.

PAGE 124

Figure 2S.Seasonal changes in abundance of pennate diatoms and Melosira granulata.

PAGE 125

107 C\J o X I/) c Melosira granuiata C Q _»__»_i_, .J — i — -U i— ^4=A--e L-^ .B 2 1 O £±— ,* L-e Jt^-'^-T--^ l_t ai m—i • ^ _A kix season -* — »F M A M J J A MONTH S O

PAGE 126

108 M. granulata filaments were present in high densities only at Station A but occasionally appeared in smaller numbers at Stations B and C. The origin of the M elosira at Station A was from horizontal displacement rather than in situ growth. It is a common dry-season occurrence for a steady breeze to blow every morning from the Polochic Valley, over the delta, and across the lake. On one morning I observed this phenomenon from a hill at Las Dantas overlooking the lake. A large plume of water whose color differed from lake water, moved out from Ensenada Los Lagartos and into the offshore area of the lake. M. granulata is of widespread occurrence in the summer plankton of eutropiiic temperate lakes. It characteristically occurs in lakes with blue-green algae blooms, particularly Anacystis cyanea , but their maxima do not necp'^'sa'^i ]y coincide (Hutchinson ]9<^7). Green alg ae St aura strum spp. were found only in low concentrations at the beginning of the sampling period (Figure 26 and 27) . Thereafter a bloom of Staurastrum pingue occurred at Station A on April 30 when 2,548 organisms/ liter were present. This was followed on the next sampling date by a less abrupt increase at Station B. Part of this abundance at Station B could have been a result of horizontal redistribution from Station A rather than in situ growth. If that were so, then the horizontal currents did not carry plankton to Station C as there was no evidence for an increase there. S^. pingue generally occurs in hard, productive waters and is often associated with blue-green algae and diatoms (Hutchinson 1967).

PAGE 127

Figure 26.Seasonal changes in abundance of Staurastrum leptoc ladum and S^. pingue .

PAGE 128

110 1 1 _C Staurastrurn ieptocladum _B 0) CM. b X C 3 4 A *-r-» — *~1 — > M
PAGE 129

Figure 27.Seasonal changes in abundance of Pediastrum simplex and Staurastrum tohopekaligense.

PAGE 130

112

PAGE 131

113 S^. 1 eptoc 1 ad'jgp. var . denticulatum was the least abundant species of desmid and no large increases occurred until September when a maximum of 124.3 organisms/liter was reached at Station A (Figure 26). S. tohopekaligense pulsed twice during the study period, simultaneously reaching high densities at all three stations (Figure 27). The April 30 high (466 organisms/liter at Station B) occurred when S^. pingue bloomed at Station A and the increase before September 19 coincided with the increases of S^. pingue (particularly Stations A and B) . Pediastrum s imp 1 ex var. duodenarium showed an overall increase from the beginning of the sampling period until May (Figure 27). At Station A the density began to decrease prior to the decrease at Stations B and C, a sim.ilar pattern to that of S_. pingue . The subsequent decline was followed by a higher peak in late July and August, reaching a maximum of 262 organisms/liter on July 29 at Station C. The most striking feature of the seasonal changes in the green algae was the prevalence of two periods of increase for the majority of species and stations, the first during April-May and the second during August-September. Blue-green algae Anacystis cyane a was the only species of blue-green algae present at all times of the year and in quantities worthy of reporting. Modest densities were present in January, but subsequent lower densities persisted until July (Figure 28). The extremely high densities encountered on July 29 at all tliree stations (543-899 organisms/liter) was such an abrupt increase that its occurrence could not have been predicted on the basis of increases prior to that date (except porliaps Station B). At tlie

PAGE 132

Figure 28.Seasonal changes in abundance of phycomycetes and Anacystis cyanea.

PAGE 133

115 (D O X in +-» E 3 J F M A M J MONTH A 5 O

PAGE 134

116 time of this pulse, green algae and diatcns were present in low densities relative to other sampling periods. Other algae Ceratium hirundinel la was present in low numbers (up to 19 organisms/ liter) during some periods and completely absent from the plankton at other times. Some forms of C_. hirundinella are eurytopic, but the species is most characteristic of eutrophic warm waters with a slightly alkaline pH. It is present in the epilimnion of many temperate lakes in the summer, but also was found in the small lakes of Java (Ruttner 1952, in Hutchinson 1967). Eudorina sp. was never more abundant than 1 colony/liter and was often absent, while Coelastrum sp. was nearly always present and reached a maximum density of 17 colonies/liter. Phycomycetes One of the most interesting observations was the occurrence of an organism believed to be an aquatic phycomycete. The free-living cottony masses did not appear to be attaclied to particles or otlier organisms. It was abundant in nearly all of the plankton samples until its disappearance at the end of the sampling period in September and October (Figure 28). The pulses of abundance did not coincide witli pulses of the dominant phytoplankton. The first pulse, evident at Stations B and C, occurred in February and March and preceded tlie dryseason blooms of phytoplankton. Likewise, the maximum abundance for the year was observed at Stations A and B whicli preceded wet-season pliytoplankton blooms. Failure to find cases wlicre tills or similar phycomycetes liave been reported previously as a component of lake plankton makes this observation an unusual curiosity.

PAGE 135

117 Copepods Cyclopoid copepods were represented by T herniocyclo ps inversus and Mcso cyclops edax , the latter being more abundant. Adults and copepodids were grouped together and were more abundant during the first half of th.e sampling period thnn the second (Figure 29). No abrupt changes v.'ore noted, and the lake appeared to be relatively homogeneous at all sampling stations for any particular sampling date. Of the calanoid copepods, Diaptomus dorsal is was far more abundant than Pseudodi optomus cule brensi s . Calanoids were generally less abundant than cyciopoids during tlie beginning of the sampling period, while tlie opposite vius true for July-September (Figure 29). Tliere seemed to be more similarities in cl\0-nges between Stations A and B than for Station C and the ctlier stations. However, the nauplii (of both calanoids and cyciopoids) had a very homogeneous distribution th.roughout the lake for any particular collection. Relative to collections preceding and after July 9, densities on that date were markedly low. There was some evidence of decreased egg numbers between May and June before the decrease in nauplii in July, but no comparable explanation exists for the sharp increase in nauplii at the end of July. Cladocera D iaphanosoma brachyuru m increased throughout the dry-season months (March-May) at all stations (Figure 30). The decrease that followed was most abrupt at Station A, less at Station B, and least at Station C. Moina m icrura and Bosmina longirostris showed no evidence of seasonal synclirony in abundance at the sampling stations. Little can be said about

PAGE 136

Figure 29.Seasonal changes in abundance of copepods,

PAGE 137

119 J F M A M J J A S O MONTH

PAGE 138

Figure 30.Seasonal changes in abundance of cladocera.

PAGE 139

121 Ceriodaphnia J FMAMJ JASO MONTH

PAGE 140

122 Ceriodaphiiia lacustris except that it was present in moderate numbers throughout most of the sojupling period. Rotifers No distinct seasonal trends in abundance could be detected for rotifers. IVhere rapid increases did occur, they were more conmion in colonial rotifers at Station A (Figure 31). For example Conoc hilus unicornis increased to a maximum of 79 organisms/ liter at Station A during August, Conochiloides dossuarius increased rapidly to 24 organisms/liter at the same station in April, while the greatest increase in Sinantherina was between September and October at Station A (to 59 organisms/liter). The non-colonial rotifers showed no distinct seasonal trends (Figure 32) . This was likely due to the long interval between collection dates relative to the generation times of most rotifer populations. Possible Controlling Factors The most apparent feature of seasonal changes in phytoplanlr.ton populations was the late dry-season pulse in April and May and a later wet-season pulse in August and September. Only the dry-season pulse in Melosira granulata at Station A can be attributed to horizontal movement rather than in situ growth. The synchrony between stations and between species of algae for the two pulses indicate that the factors controlling growth were present at the same time throughout tlie lake and had similar effects on most of the species. Environmental factors commonly accepted as regulators of phytoplankton growth are solar radiation, temperature, and nutrients. Of these, only solar radiation and temperature were

PAGE 141

Figure 31.Seasonal changes in abundance of colonial rotifers,

PAGE 142

124 20 10 Q Sinantherina 10 Q Conochiloldes J P'MAMJ JASO MONTH

PAGE 143

Figure 32.Seasonal changes in abundance of solitary rotifers.

PAGE 144

126 Brachionus 5 5 5 10 5 10 i-> ^ 10 •t^ ^ 10 10 10 10 10 *^ -!-^ r^ t -B I P ' » I [I) ll ..I » I t I ft .. 1.1 ( 1 9Filinia Keratella JFMAMJJA50 MONTH

PAGE 145

127 monitored at sufficient intervals to enable detecting seasonal trends; saJTiples for nutrient analysis were collected only twice ("April 14 and October 23). Grazing by zooplankton may be an additional consideration, but it is unlikely that the zooplankton would have effectively grazed some of the larger phytoplankton such as Pcdia strum and Ana cystis . Diatoms, however, may have been subject to some control by grazing. 7\nderson (195S) noted a phytoplankton minimum when nutrient deficiency was unlikely in Lake Lenore, Washington, but Moina hutchinsoni were moderately abundant. Thus, it is intere?ting to note that M. mi crura experienced an annual maxim.um at Station A when diatoms had decreased (Figure 30) . For the lake stations, and particularly Stations A and B, the surface and bottom temperatures were highest during the dry season and decreased in July at the beginning of the wet season (Figure 13) . This was a result of an increase in discharge to the lake of river water of lower temperature, noticeable in the deep samples from Stations A and B. The lower insolation in July also could have contributed to the decrease in water temperature. It is possible that the decrease in solar radiation could liave been partially responsible for the lower numbers of phytoplankton between the April-May and August-September pulses. Secchi disk transparencies in July for the three stations were among the lowest observed during the year (Figure 44). In the absence of large densities of plankton, the decreased transparency might be attributed to higher turbidity induced by wetseason inputs of silt and detritusladen waters from the watershed. Floating debris, ranging from leaves to large logs, were especially abundant in the area of Station A, and as a result, navigation was often hazardous.

PAGE 146

128 In the search for an explanation of the decrease during July for many of the planktonic species, the possibility of inhibitory substances cannot be overruled. This v/as the month of initial flushing of the colored waters of the lagoons (Amatillo, Lagartos) and rivers (El Padre Creekj Rio Oscuro) of the Polochic Delta. The water exported from this area to the lake differed sharply from lake water in its organic matter concentration (Figure 34) , dissolved oxygen, pH, total alkalinity, and conductivity (Figure 14). Although a complete list of potentially inhibitory substances from these anaerobic or near-anaerobic waters is not in order here, a few would include hydrogen sulfide, tannins, and heavy metals (Cu, Zn, Al) either in ionic form or present as complexes with organic matter. In spite of the possibility of inhibitory srbstanres, some of the largest pulses in phytoplankton occurred soon after the July 9 low. It is interesting that some substances, inhibitory to certain species, are apparently stimulatory to others. For example, Lefevre et al. (1952, in Hutchinson 1967) collected pond water at different times of the year to test its effect on plankton cultures. Water collected in October, when there was a great deal of decomposition of higher plants, was stimulatory to two species of Pediastrum (£. boryanum and P^. clathraturm var. punctulatum) but was inhibitory to two species of Cosmarium . Similarly, substances which may be inhibitory at high concentrations may be stimulatory at low concentrations. If the low numbers of phytoplankton sam.plcd at a time coinciding witli the initial flushing of the watershed were evidence of inhibition, then subsequent dilution of the lake by wet-season rains could have rendered such substances innocuous or even stimulatory.

PAGE 147

129 Data on critical nutrients arc both scanty and unreliable, but it is unlikely that these would have been limiting. For example, silica ranged from 10-14 mg/liter for both sampling periods (Appendix, Table F.) well above 0-1 mg/liter, a range for which there is evidence suggesting a limitation of diatom growth in nature (Lund 19S0) . Because no samples were analyzed for nutrients during the July 9 low in phytoplankton density, no conclusions can be drawn from control by nutrients. The dramatic increase of Anacystis cyanea on July 29 is noteworthy because it marked the increase of the other phytoplankton populations. Blue-green algae are typically found in waters with high dissolved organic matter and some of the highest concentrations of organic matter in the lake were detected in July (Figure 34). However, no specific organic substance has been found to facilitate their growth (Hutchinson 1967). Nevertheless, the sudden appearance of high densities of A. cyane a on July 29 marked the beginning of conditions apparently favorable to the growth of other phytoplankton. A. cyanea persisted until the end of the sampling period. A. cyanea is one of the greatest problem algae in eutrophic lakes of northern latitudes, particularly when it forms massive summer water blooms. In India, it develops permanent blooms in artificial temple tanks. Maximum numbers occur in July and decrease to a minimum a month later in August. No changes in temperature or phosphate concentration is observed during the bloom reduction (Hutchinson 1967). Grazing by zooplankton can be discounted as a factor contributing to the July 9 low in phytoplankton not only because it is likely that most of the algae were too large to serve as a direct food source, but also because many species of the zooplankton were present in low numbers

PAGE 148

130 on that date. Nauplii underwent a sharp decrease as well. An exception was the pulse in Moina micrvira at Station A which reached a maximum of 21 organisms/liter. In spite of the homogeneous distribution and the distinct bimodal pulse demonstrated by the net plankton during 1972, it would be tenuous, at best, to expect this to be a regular annual occurrence. However, Nordlie (1970) also has evidence of seasonality in the plankton abundance of Lake Izabal. In August 1969 the phytoplankton diversity was approximately the same along the east-west axis of the lake. Presumably it was also relatively abundant. Heaviest zooplankton populations were at the west end and their abundance decreased in the eastern region. This may have been a situation similar to my August 1972 observations. During March 1970, Nordlie's samples showed the greatest phytoplankton diversity at the east end of the lake, while almost none were present from the middle and west end. Zooplankters were similarly distributed as in 1969 but more dense. Although 1 found phytoplankton to be much less abundant in March 1972 than August 1972, it was much more abundant than implied by Nordlie for his March 1970 samples. It is possible that Nordlie's sampling in March preceded a dry-season pulse in abundance, sucli as that observed in April and May 1972, Benthic Community The most interesting curiosity of the bottom fauna was the occurrence of Tanaidacea, first reported from the lake by Nordlie (1970). Several species of this family are known from fresh waters, but it is likely that all are dependent on slightly more saline water than is usual far from the coast (Hutchinson 1967). The highest density recorded

PAGE 149

131 was 296 organisms/m^ at San Felipe Bay (Table 9) where the penetration of brackish water from the coast is at least a seasonal event. It is doubtful, however, that the specimens collected from the outfalls of the Rio Polochic distributaries (Comercio and Coban) ever come in contact with water of conductivity greater than ca. 200 ymlio/cm. More sampling, both on a spatial and seasonal basis, would be necessary to adequately establish distributional and temporal patterns. Nordlie 2 (1970) calculated abundances of Tanaidacea as high as 1,378/m in previous collections from the lake. Members of the family Chaoborinae, presumably Chaoborus , were present in the bottom deposits both in larval and pupal stages. Den2 sities were as high as 77 organisms/m and they were present at all localities sampled except Station 15 (Figure 39) where none were found. They also were collected occasionally in net plankton tows. The Tendipedinae were found at all stations and ranged between 10-77 organisms/m Adults were seen frequently and in massive numbers above the lake. Oligochaeta were present at five of the twelve stations with densities 2 ranging from 19-153 organisms/m . The gastropods, most of which were not alive when collected, were present at six of the twelve stations. The average density of 29 organisms/m" for the Chaoborinae was 2 below that found by Deevey (1957) in Lake Amatitlan (40/m ), and much 2 lower than his mean for Lake Giiija (l,27S/m ). Deevey found no Tendi2 pidac at Lake Giiija but reported a mean of 673 organisms/m for Lake Amatitlan, an order of magnitude greater than my samples at Lake Izabal (Table 9). The density of bottom fauna was strikingly low compared to lakes in temperate latitudes (Deevey 1941), but it must be remembered that

PAGE 150

132 \D I O ^1 (U o +J o o •xi (D P O o o (U CTi fH to in -H • w w
PAGE 151

IJJ these numbers represent standing crop and not rates of biomass increment. Although the higher year-round temperatures of tropical lakes may result in growth rates of benthic organisms as much as tX'io or three times those of temperate lakes, it is doubtful if this factor would make their production comparable. A plausible explanation for the difference is that the more intense metabolism of the free water would tend to be more complete in tropical lakes, so that less surplus energy reaches the sediments and their fauna (Deevey 1957) . This may be true for monomictic lakes (studied by Deevey) which rem.ain stratified throughout most of the year, but for lakes such as Lake Izabal that mix frequently to the bottom, it is unlikely that the benthic organisms would be restricted by energy availability. A more reasonable explanation may be predation by bottomfeeding fishes. Such fishes would have access to benthic organisms in all areas of the lake since they are not restricted by an anaerobic zone in Lake Izabal as they are seasonally, at least, in monomictic lakes. Bottom-feeding fishes are well represented in the ]ake and include marine catfish and several species of cichlids. The presence of substantial amounts of organic matter in the surface muds will be discussed as a potential source of food for detritus consumers in the section, "Metabolism and Organic Matter." Svurmiary Statement The net phytoplankton assemblage of Lake Izabal was characteristic of that found in many other shallow productive lakes with moderately hard waters. Most of these species have a wide latitudinal distribution. The predominant taxonomic groups included diatoms, myxopliycetes, desmids, and

PAGE 152

134 members of the chlorococcales. The zooplankton composition had a slightly more tropical flavor, especially with regard to the copepods. For example, species of Mesocyclops were present, a genus of predominantly tropical distribution, and Diaptomus dorsalis , also present, is common in countries bordering the Caribbean. Cladocera were less abundant than copepods, and while the rotifers were more diverse than other zooplankton groups, their abundance was much lower. The only surprising feature of the benthic community was the relatively abundant occurrence of a member of the Tanaidacea which is a group characteristic of marine and brackish waters. The most apparent feature of seasonal change was the late dry-season pulse of phytoplankton in April and May followed by a wetseason pulse in August and September (Figure 53). These pulses did not seem to coincide with rates of gross primary productivity except in late September and October when both phytoplankton abundance and primary productivity rates decreased (see following section. Metabolism and Organic Matter) . Apparently differences in the standing crop of net phytoplankton over periods of 3 or 4 weeks were an insensitive indicator of daily rates. This could be attributed to the high variation in rates of turnover possible in a system with low storage capacity. In most cases the pulses could be attributed to in situ growth rather than to distribution by horizontal currents. The interluding low abundance between the two pulses coincided with the inception of the wet season when runoff increased dramatically, discharging large quantities of silt and organic debris into the lake. The resulting decrease in transparency of the lake water may have been partly responsible for the low phytoplankton abundance during

PAGE 153

>%

PAGE 154

136 o

PAGE 155

137 this period. Other controlling factors thnt cannot be overruled are the influx of inhibitory or toxic substances, moderate decreases in temperature, or changes in nutrient availability. The low abundance In June and July was interrupted by a bloom of Anacystis cyanea in late July whicli shortly preceded an increase in most other species of phytoplankton. Zooplankton populations on the whole were comparatively more stable than phytoplankton populations.

PAGE 156

METABOLISM AND ORGMIC MATTER This chapter, which represents the core of the present study, attempts to estimate the magnitude of the principal energy flows responsible for the metabolism of the lake ecosystem. One of these energy flows, that from allochthonous detritus input, received special attention. In lakes that approximate the surface area of Lake Izabal, allochthonous energy sources are generally regarded as insignificant contributions to total metabolism. This is undoubtedly true for lakes whose replacement time for the water mass is in the tens or hundreds of years. However, by virtue of the relatively small volume of Lake Izabal in relationship to its surface area, as well as it location within a watershed receiving high rainfall, the annual replacement of a large portion of the water mass by runoff could conceivably produce an energy surplus or deficit. This would ultimately depend on the organic matter concentrations of the inflowing and outflowing waters. Other important metabolic compartments in aquatic ecosystems are the plankton and the bottom muds. In the gradient from deep to shallow lakes, the proportion of total metabolism attributed to the bottom muds becomes greater. Thus some attention was given to the energy flow of the benthos, although the metabolism of the planktonic compartment was expected to be of greater magnitude. 138

PAGE 157

139 Methods Chemical Oxygen Demand The concentration of organic compounds in the lake and river waters was determined by oxidative digestion with potassium dichromate and sulfuric acid, commonly known as "wet oxidation." Several procedures were reviewed and a number of modifications were made before routine analysis was established. Silver sulfate (Ag SO ) used as a catalyst for the oxidation is particularly effective for straight-chain alcohols and acids (Golterman 1969} but certain phenolic compounds are generally resistant to oxidation, even in the presence of a catalyst. In spite of this and other disadvantages, the method was generally useful in the absence of more sophisticated instrumentation. Fraction definitions and conversion criteria The limnological application of organic analysis by quantitative dichromate oxidation has been reviewed and tested by Maciolek (1962). To maximize the information that could be extracted from a single test, the samples were separated to distinguish three size fractions. Total_ COD represents all the organic matter sizes in the sample, dissolved and particulate. Dissolved COD is the fraction remaining after vacuum filtration through a O.SO-y pore membrane filter and theoretically should contain particles no greater in diameter than the filter's pore size as well as soluble organic matter. Particulate COD refers to organic matter greater than O.SO \} in diameter, calculated by subtracting the total COU from the dissolved COD concentration. Net Particulate COD is the fraction collected with a No. 20 plankton net wliose mesh aperture is 76 y (Welch 1948).

PAGE 158

140 Theoretically tliis sliuulu represent particles whose diameter is greater than the aperture size (approximately 100 times the size of the particulate COD) and corresponds to the organic matter of the net plankton. Maciolek (1962) reported COD as Oxygen Consumed (O.C.)He pointed out that a theoretical Oxygen Equivalent (O.E.) must be assumed in order to convert COD (or O.C.) to organic matter. The O.E. is constant for an individual pure compound (e.g., the O.E. of hexose is 1.06). The mg Oxygen Consumed divided by the assumed O.E. for representative organic compounds yields the weight of the organic compound, mg O.C. (or COD) „ • m ^^ —2 ;^— = — — ^^ = mg Organic Matter. mg O.E. " Several approaches can be used to determine the O.E., including calculation from elemental composition and from proximate composition determined from proteins, lipids, and carbohydrates. Table 10 lir,ts characteristic O.E. 's that were reported by Maciolek (1962) in approximate order of their limnological importance. In the section-Balance of the Organic Matter Budget -where the energy budget of the lake is calculated, a value of 1.44 O.E. was selected to convert COD values to organic matter. The relationship between combustion calorimetry and oxygen consumed is close enough to permit an accurate caloric estimate by quantitative oxidation (Maciolek 1962). A value of 3.4 gcal per mg of O.C. is suggested, which would be equivalent to approximately 4. 86 gcal per mg organic matter. Collection and treatment of samples Water samples were collected in the field in clean 500-ml narrowneck amber glass bottles. When deep samples were collected a 2or 3liter Van Dorn bottle was used. Surface waters were generally taken by

PAGE 159

141 Table 10.Characteristic oxygeji equivalents [O.E.j in approximate order of their limnological iipportance^

PAGE 160

142 immersing the collection bottle beneath the surface" (10-25 cm). The samples were stored on ice during transport to the laboratory. Generally there was a maximum lapse of 7-8 hours between the first field collection and the laboratory treatment of the samples. The procedure for chemical oxygen demand (American Public Health Association 1965) was modified slightly. Mercuric sulfate (Hg SO ) was not added to eliminate interference by chloride because of the low concentration of this ion (<12 ppm) . One hundred-ml samples were ovendried at 95 C in 200 or 250-ml Erlenmeyer flasks requiring approximately 18 hours for complete evaporation of the water. Ten ml of 0.05N^K^Cr^O were added to the flask followed by 25 ml of concentrated H SO in which 5 g/ liter of Ag SO was dissolved as a catalyst. The sainples were digested for three hours in boiling rain water in unsealed pressure cookers to facilitate even heat distribution of the samples. The mouths of the flasks were covered with aluminum foil and the necks of the flasks provided a refluxing surface which may have minimized the loss of volatile organics. After digestion, 100 ml of distilled water were added and the flasks were cooled to room temperature in a water bath. Titrations of blanks and samples were made with a 0.025N (approx.) Fe(NH^)2(S0^) 2611 solution and three drops of ferroin solution (G. Frederick Smith Co.) were used as an indicator. To determine the dissolved fraction of the total organic matter, 100 ml of sample were vacuum filtered through a O.SO-ypore Metricel GA-4 membrane filter (Gelman Instrument Co.) The fiJters were first rinsed with distilled water to rid them of possible organic contamination and were then purged before filtration of the sample with approximately 200 ml of

PAGE 161

143 of the sample to clear the distilled v.'ater. The dissolved organic matter samples were treated as described above and the particulate organic matter was determined by subtraction. A No. 20 plankton net (25-cm diameter) was used for filtering water and concentrating samples for the determination of large particulate organic matter. River samples were collected by holding the net below the surface in the current (of knoun velocity) for a known time period (usually less than one minute). When velocities were too great for holding the net in position, a bucket was used to remove samples just below the surface and 200 liters were poured through the plankton net. Samples were stored on ice until they were transported to the laboratory. They were always analyzed the same day as collected. The procedure for digestion and analysis was modified from the method of Golterman (1969) . Samples were concentrated to 50 ml or, if the total sample was too high in organic matter for this procedure, an aliquot (1/2 or 1/4 of the total sample) was used instead. Since the sam.ples were high in organic matter concentration, they were not dried and reagents that were added for digestion were the same as described above except that l.OOON^K Cr was the oxidant. The aluminum foil-capped flasks were digested for three hours. After dilution vv'ith SO ml of distilled water, the samples were coiled and titrated with 0.25N (approx.) Fe(NH ) (SO ) '6 HO in the presence of 3 drops of ferroin indicator 4 4i 4 2 2 soluti on. Dissolved Oxygen Concentration Determinations The precision of methods for the determination of dissolved oxygen in natural waters makes possible the measurement of the metabolic activities

PAGE 162

144 of cojisumers and producers over relatively short periods of time. The method used for dissolved oxygen in this study is a modification of the Winkler method (Golterman 1969] . The modification employs a much higher concentration of KI than normally used, thus reducing errors due to volatilization of I_ and interference by organic matter. Hydroxides dissolve more readily and the starch endpoint is sharper. Nitrite (NO ) interference was eliminated by use of sodium azide. Approximately 0.010N_ sodium thiosulfate was used to titrate duplicate 100-ml aliquots from each bottle resulting in a precision of about 0.03 mg 0^/liter. Ground glass-stoppered BOD bottles of 300-ml capacity were used for nearly all water collections for oxygen determination. The MnSO reagent and alkaline iodine-azide solution were added immediately after sampling and acidification was delayed until immediately before titration in the laboratory. Biological oxygen demand Large plastic containers (9.5 or 18.9 liters) were used to collect water sajnples in the field for transport to the laboratory. For the river and lake surface samples, the mouth of the jug was held approximately 20 cm below the surface and allou-ed to fill. Deeper samples were collected with a 3-liter Van Dorn sampler and the water was drained into the plastic jugs. A subsample was usually collected and fixed at the field site for measuring the in situ dissolved oxygen concentration. In the afternoon or evening of the day of collection, the oxygen concentrations of the water in the plastic jugs were measured with a YSI oxygen meter (Model 51A) . If the concentration was below 5 mg O^/liter, the samples were aerated until they contained 6-8 mg/liter oxygen.

PAGE 163

145 After thorough inixing of the v\'ater to ensure hoinogeneity of organic matter and oxygen, samples were carefully siphoned into eight darkened BOD bottles. The first and the eighth bottles filled were immediately analyzed for oxygen. The other six BOD bottles were submerged in a darkened water bath to prevent air leaks and to minimize ambient fluctuations in temperature. The bottles were agitated daily to resuspend particulate matter that may have settled to the bottom. Duplicate bottles were analyzed for dissolved oxygen after one, three and seven days of incubation. Due to the limitation in number of BOD bottles, several trials were incubated for five days only, i.e., duplicates were analyzed on day zero and after five days and respiration rates were averaged on a daily basis. A directly comparable 5-day daily rate can be determined from the other procedure by averaging the 3-day and 7-day oxygen concentrations, subtracting this from the zero-day concentration, and dividing the difference by the number of days (5) . Respiration of bottom muds and their organic content Samples of bottom mud were collected during October 1972 with an 2 Ekman sampler (522.6 cm ). When each haul (2 per station) was lifted into tlie boat, efforts were made to disturb the structure as little as possible. The top leaves of the sampler were folded back to gain access to the mud and a hypodermic syringe was used to remove 5-ml subsamples from the upper 2 cm of the mud in the Ekman sampler. The orifice of the s>'^ringe was enlarged to 7-nmi diameter to allow the collection of particulate matter. Two 5-ml subsamples were placed in a 300-ml BOD bottle which was stoppered and returned to the laboratory. Three BOD bottles were filled from eacli liaul, for a total of six per station.

PAGE 164

146 In the laboratory, the BOD bottles were carefully filled by slowly siphoning lake water down the side of t!ic bottles to minimize disturbance of the mud. The water was initially turbid, but became clear after one hour with the settling of suspended mud particles. An initial bottle was filled for determination of the concentration in the lake water as well as three control bottles for determining the respiration of lake water only (without mud) . The bottles were incubated in darkness at ambient room temperature (ca. 25 C) . Subsamples from these BOD bottles were drawn from ca. 4 cm above the mud layer into 75-ml bottles with an aspirator. The volume of reagents used for oxygen concentration determination were adjusted to the size of the sample. The initial bottle (lake water) was analyzed immediately after all samples were prepared. Duplicate bottles with mud and one buttle witli lake water weie analyzed for OAygen after Lwo, f oui , and eight hours of incubation time. Respiration rates were calculated from the differences between these determinations. To convert the rate of change of oxygen concentration to respiration rate per unit area of mud, the following calculations were made: Volume of water mass = 0.300 liter (total) 0.010 liter (mud) = 0.290 Inside diameter of BOD bottle = 6.35 cm 2 -4 2 Cross sectional area 31.67 cm or 31.67 x 10 m mg 0^/liter 0.29 liter ^ X 37 — 7^ = 91.57 mg /mhr hr 31.67 X 10 m" The respiration rate (mg 0„/]iter hr) was multiplied by the constant 91.57 2 9 to yield mg /m" hr.

PAGE 165

147 The Ekman samples, from which the small subsamples were removed for the mud respiration experiment, were passed through a No. 40 mesh screen (U.4i7-mm aperture) to retain particulate matter and mud-dwelling organisms. After removal of the organisms for counting (see Section-Planktonic and Benthic Communities), the rem.aining samples were analyzed for organic matter content. Samples were oven dried (70 C) and weighed, then ignited at 550 C for 1.5 hours and reweighed. The weight loss by 2 ignition was calculated per unit area (m ) of mud surface and as percent of the total particulate matter retained by the screen (excluding the subsamples for respiration rates and the organisms). Light and dark bottle method Several attempts were made at the beginning of the study to estimate primary production by the use of diurnal changes in oxygen concentration of the water mass. These attempts were abandoned due to the erratic results from the error involved in the horizonal displacement of water masses of differing metabolic history. The oxygen light and dark bottle method (Vollcnweider 1969) was then used and stations for measurement were selected in the western, middle, and eastern areas of the lake (Figure 4). Water samples were collected with a three-liter Van Dorn bottle from the surface, and 1, 3, 6, and 11 m. After distributing each sample among an initial, light, and dark bottle (covered with black electrical tape and aluminum foil), the two latter bottles were returned to the depth from which they were collected for incubation. The initial bottles were fixed and placed in an insulated box to exclude sunlight and prevent excessive heating. The trials were usually duplicated at each station and the incubation time, in most cases, was for tliree hours, from about 0900 to 1200.

PAGE 166

148 V/hen posrible, relatively cloudless mornings were chosen for the measurements arid the three stations were measured on consecutive days. Total incoming radiation was measured in El Estor with a pyrheliometer (Solar Radiation Recorder 9-401, R.E. White Instruments, Inc.). Secchi disk values were recorded using a standard 20-cm diameter disk. Results and Discussion The bulk of the data for organic matter, measured as chemical oxygen demand (COD), is available from March through October 1972. Before March the procedure had not been modified sufficiently to yield reliable measurements of COD in moderate to less dilute concentrations; only measurements of net particulate organic matter are reported for December 1971, January and February 1972 in the Appendix (Table F-6) . 9 The results will be reported as mg COD/liter or g COD/m" of lake surface, or as rates, i.e., g COD/m day or g COD x 10 /month. The results vs'ill not be converted into organic matter equivalents until the section , Balance of the Organic Matter Budget, because (1) the technique measured COD directly and organic matter only indirectly, (2) comparisons among stations, collection dates, and with data in the literature (often reported as mg COD/liter) can be made directly without conversion, and (3) the conversion requires some rounding off of data. In a few cases the reported concentration of dissolved COD is higher than the total COD concentration. This is apparently due to variation in sampling or to error in the analysis of the samples. When further calculations were made, the higher of the two values was used as the total COD concentration.

PAGE 167

149 The concentrations of COD will be examined first, followed by the rates of input, that is, the concentrations multiplied by the discharge rates of the rivers. Concentrations of Chemical Oxygen Demand The concentrations of COD in all samples (as mg/liter) are reported in the Appendix, Tables F-lto F-6 . For simplification and clarity, the data are partitioned into water types recognized earlier in the discussion of water characteristics (p. 66). Swa mp waters The waters of the Rio OscurO; Amatillo, and El Padre Creek are coffee-colored, visibly suggesting the presence of dissolved organic matter. The highest concentrations of total and dissolved COD were measured from this water type (Figure 54a) . July represented the month of greatest total COD concentrations, with Rio Oscuro highest (63.13 mg/ liter), Amatillo next (54.05 mg/liter) and El Padre Creek third (46.30 mg/liter). These maxima coincide with the first rains and appear to be a result of initial flushing of the swamp forest. During the dry season (March-June), Rio Oscuro showed marked differences in COD concentration between the surface and 5 m (Appendix, Table F-1). This was due to the independent origin and stratification of the two water masses (as explained in the section-Hydrology and Water Characteristics) with the more concentrated and warmer surface waters originating from the swamp. The water column was thermally stratified at the Amatillo station also, but the two water masses were of the same origin and had similar

PAGE 168

Figure 34.Concentrations of particulate and dissolved COD (mg/liter) during the sampling period for (a) swamp waters, [h) Rio Polochic distributaries, (c) small rivers, and (d) lake stations (A, B, C, and San Felipe). March, April, and May are dry-season months; n.d. indicates that no data were collected.

PAGE 169

151 60 5040 30 2010A "^ _j DParticulate COD DDissolved COD MAMJJAO MAMJJAO Oscuro Amatillo M A M J J A O El RadreCr Q 20 o U 10 B R E — =aBs« iS(£. M A M J J A O Comercio MAMJJAO Cob&n I M A M J J A O Bujajal M A M J J A O San Marcos M A M J J A O Sajce M A M J J A O Manacas Cr 10 D MAMJJAO M A M J J Station A Station B A O 10 M A M J J A O Station C i«!3^ nd .^M A M J J A O San Felipe

PAGE 170

152 COD values. The slightly higner COD concentrations at 3 m (Appendix, Table F-1) in March and April may represent real differences (relative to surface) since the 3-m water was anaerobic. Inorganic reducing substances could have contributed to the COD at 3-m depth since the test does not distinguish between organic and inorganic reducing compounds. The generally higher dryseason measurements of COD at Amatillo as compared with Rio Oscuro and El Padre Creek were likely the result of intense localized phytoplankton production. In these stagnant waters current v\'as absent, whereas Rio Oscuro always had a slight dry-season flow. The high net particulate COD's from April through June support the visual observations that phytoplankton biomass was high. Rio Polochic distributaries The COD conceiitrations of all three distributaiies show an almost continuous monthly increase until June, and a subsequent decline through October (Figure 34b) . The differences between the distributaries for any one month were probably variations due to sampling. The COD concentrations of the July samples were predominately due to particulate matter, unlike the swamp waters of Figure 34a. The Rio Polochic samples were also quite high in the proportion of particulate COD in October. Small rivers Three small rivers were sampled for COD from March through October (Figure 34c). The range in total concentration of C0!1 (3.69-24.06 mg/ liter) was similar to the Rio Polochic distributaries (4.97-27.76 mg/ liter). The Manacas Creek samples showed the least variation; particulate COD never increased above one-fifth of the total COD (August 21). Since

PAGE 171

153 the only noticeable discharge at Manacas Creek was during July and August (the months of the two highest COD concentrations), samples from other months contained dissolved and suspended COD probably unrelated to runoff. The Rio Sauce had no detectable flow until June, accounting for the sharp increase in both particulate and dissolved COD at the beginning of the wet season. The July sample had an extremely high net particulate concentration (Appendix, Table F-3) which exceeded the sensitivity of the test used. The Rio San Marcos was in constant flow throughout the year and differed from the Rio Sauce in that the watershed was mostly deforested, possibly contributing to the visibly higher silt load. Total COD concentrations never reached the high values of the Rio Sauce. Lake stations and outlet Stations A, B, and C, and the San Felipe outlet demonstrated fewer differences in seasonal COD concentrations than all river stations (Figure 34d) . As a result, seasonal trends are harder to discern. However, by comparing dryseason with wet-season COD concentrations some differences can be noted. At Station A, March through June samples were below 9.00 mg total COD/ liter, while from July through October total COD values were above 10 mg/liter. At Station B the differences between the two seasons were not as great. Station C had no samples above 10 mg total COD/litcr. Only in June, July, and August did they exceed 9 mg/liter. At San Fciipe there was no evidence of seasonal trends in COD concentration and values fluctuated around a mean of 8.96 mg/liter.

PAGE 172

154 Monthly Flows of COD The discharge rates of the rivers and watershed areas that are listed in Table 2 were used to calculate the rate of organic matter flow into Lake Izabal from March through October, the months for which COD measure3 ments were available. The concentration of COD (mg/liter or g/m ) multiplied by the monthly discharge (m x 10 ) yields the monthly flow of total COD (g X 10 ) . These calculations were performed on all COD fractions for the three Rio Polochic distributaries (Comercio, Coban, and Bujajal), Rio Oscuro, Rio Amatillo, Rio San Marcos, Rio Sauce and the remaining watershed areas to the north and south of the lake (Appendix, Table G) . The COD export at the San Felipe outlet was calculated from the outflow rates estimated by balancing the water budget (Table 5) and the COD concentrations measured from that station. These data are summarized in Table 11 on the basis of calculations in the Appendix, Table G. Dryand wet-season total inflows increased by approximately one order of magnitude between May and June, the transition to the wet season. Total COD inflow nearly doubled between June and July, and declined thereafter (Figure 35). The output at San Felipe was relatively constant and close to the total inflows for March, April, and May (dry season) . June, July, and August outflows from San Felipe lagged behind the total inflows, but by October, the flow values were approximately the same. The average monthly inflows for the wet season (June, July, August, and October) were approximately 11 times tlie monthly inflows for the three dry-season months. For the purpose of later calculations, the September flows, for whicli no data were obtained, will be defined as the average of the August and October values. Accordingly, total September

PAGE 173

155 w

PAGE 174

Figure 35.Rates of organic matter inflows and outflows of Lake Izabal for the lake as a whole (g COD x 10^/month) and for an average m^ of surface area (g COD/m" day) ,

PAGE 175

157 40 ^30 c o E % O ^ ?0 10 Inflows Outflow 2.0 1.5 ns T3 CvJ E 1.0 o D) 05 i 1 J 1 1 1 i 1 1 1 1 1 KD J FMAMJ JASON D^ MONTH

PAGE 176

158 inflows would be 2b, 202. 6 x 10 g COu, and the San Felipe outflow would be 22,729.8 x 10^ g COD. To evaluate the relative contribution of watershed areas to the total organic matter inflow, the Izabal Watershed can be divided into two runoff components. The Polochic Valley component is defined as those rivers which empty into the lake through the Polochic delta and include the three Rio Polochic distributaries, Rio Oscuro, and Rio Amatillo. These two latter "swamp rivers" are included because the Rio Polochic, during wet season flooding, spread over the delta and mixed with the waters of the Oscuro and Amatillo. The minor runoff component of the basin includes the Rio Sauce, Rio San Marcos, and the watersheds to the north and south of the lake. Table 12 compares the monthly percentage contribution of organic matter runoff from the Polochic Valley and the minor watershed components. The Polocliic Valley watershed contributed between 87.8 and 96.2 percent of the total organic matter for the months sampled. In summary, the flow of organic matter into Lake Izabal was strikingly seasonal, and not due only to increased wet-season runoff rates, but also to increased organic matter concentration of the runoff. The monthly percentage of total organic runoff during the sampling period averaged 1.7% for each of the three dry-season months (March-May) and 19.01. for each of the five wet-season montlis (June-October) . At the beginning of the wet season, organic matter outflow at San Felipe lagged behind upstream inflows for approximately three montlis; thereafter outflows and inflows did not differ widely. The Polochic Valley alone accounts for 80 percent of the watershed area but more than 90''
PAGE 177

159 Table 12.Relative contribution (percent) of organic matter runoff into Lake Izabal from the Rio Polochic Valley and from the minor watersheds Mar Apr May Jun Jul Aug Oct Polochic Valley Watershed 96.2 87.8 92.0 95.0 89.3 89.5 96.1 Minor watersheds 3.8 12.2 8.0 5.0 10.7 10.5 3.9

PAGE 178

160 Respiration Rates (BOD) The results of 48 experiments of respiration rates, 40 of which were 7-day incubation periods and 8 of which were 5-day incubations, are reported in Table H of the Appendix. These have been grouped according to water t)'pe and are summarized in Figure 36. The respiration rates of the swamp waters, represented by Rio Oscuro, Amatillo, El Padre Creek, and Lagartos stations were the most variable throughout the year and also had some of the highest rates of all stations. The three highest values (1.55, 1.63, and 1.56 mg /liter day) from Amatillo and El Padre Creek, were apparently due to the accumulation of organic matter resulting from stagnation and anaerobic conditions. The Rio Polochic samples resulted in generally lower and less variable rates than the swamp waters and ranged from a low rate of 0.02 mg 0^/liter day in December over a 5-day period to a high of 0.40 mg 0^/liter day in June during the first day of incubation. This latter value coincided with the beginning of high discharge rates initiated by upstream rainfall (see section--"Hydrology and Water Characteristics") . Likewise, samples from the small rivers were higher in June than other months. Of the lake station samples, the highest observed rate was 0.54 mg /liter day (May 10) while most values ranged between 0.10 and 0.30 mg /liter day. The daily rates of respiration for any single sample were greater during the first day of incubation (day 0-day 1) than between day 3 and day 7. Tlius the respiration rates of 1-day incubations (circles in Figure 36) were higher than the 5-day incubations (x's) wlicn compared on a daily basis. There were some exceptions, most notably where rates of resj)irat:on

PAGE 179

o o rH o &< o •H Pi w

PAGE 180

162 TF i>^ fd T> +-> I o <^ (D > O 0) asm I o l: in c in (13 03
PAGE 181

163 cf the first day were low (Comercio, January 25; Coban, June 17; San Felipe, July 29). This may have been due to sampling error, but it is conceivable that the inside surface of the bottle, which provides an artificial surface for bacterial colonization, may have resulted in increased densities of bacteria, thereby resulting in higher respiration rates after the first day of incubation (Pratt and Berkson 1959) . The nature of the progressive decrease in respiration rate over a 7-day period can be examined by plotting the oxygen concentrations on semi-logarithmic paper. If the curve is linear, then the process is exponential. Figure 37 illustrates the curves generated by the data in Table H of the Appendix. By inspection, it can be seen that the curves are approximately linear, thus representing exponentially decreasing rates of oxygen consumption throughout the incubation period. COD values were plotted against BOD rates to determine if there was any relationship [correlation) between organic matter concentration and respiration rate. No correlation was apparent between 1and 5-day respiration rates and total or dissolved COD. This lack of correlation may have been partially attributable to variation introduced by not always collecting COD and BOD samples on the same day. However, more important was the restricted range of COD concentrations throughout the sampling period. This was especially well illustrated by the lake station samples (Figure 58), wliich demonstrated a broad range of respiration rates within a narrow range of COD concentrations. These concentrations could have been manipulated aritficially througli dilution, but no experiment was designed for tliis purpose. Figure 38 also shows the degree to wliich tlie water types overlap based on the two parameters. Only the lake stations and swamp waters

PAGE 182

Figure 37.Dissolved oxygen concentrations of water samples during 7-day incubation periods. Linear curves plotted on semi-log paper represent exponentially decreasing rates of oxygen consumption throughout the incubation period.

PAGE 183

165 Swamp Waters Small Rivers 4) -^ o E 3 Oscuro 2 1 2 3 4 5 6 Days of Incubation 1 L Amatillo 1 2 3 Days of Incubation RTo Polochic Lake Stations tstatipnC 1 2 3 4 5 6 7 Days of Incubation Days of Incubation

PAGE 184

Figure 38.Relationship between oxygen consamption rates [BOD) and total COD concentrations for the four water types characterized.

PAGE 185

167 1.0 I s t i I I I I § g i I B i I f t Lake Stations f f 0,8 ^0.6 C\J O g^O.4 L" -c/ 0.2 0.0 i Swamp Waters Small Rivers 12 16 20 24 28 mg COD/liter

PAGE 186

168 are mutually exclusive while there is considerable overlap in the areas occupied by small rivers and the Rio Polochic. It is likely that the rate of respiration was more dependent on the quality of the organic matter, rather than the precise quantity. Respiration and Organic Content of the Bottom Muds Figure 59 illustrates the locations from which bottom mud samples were collected for respiration rate and organic content determinations. The results from the mud respiration experiments are presented graphically in Figure 40. In almost all cases the hourly respiration rate during the initial 2-hour incubation period was higher than during the second 2hour period or the following 4-hour period. This can be attributed to the disturbance of filling the bottles which initially suspended the mud. The controls containing only lake water showed little or no respiration, so no correction was made for the overlying water. The rates calculated after 4 and 8 hours were usually lower and more similar to each other than during the initial 2 hours. It was felt that the lower and more stable rates were more indicative of the respiration of undisturbed muds in the lake. Thus, the hourly rates, between 2 and 4 hours, and between 4 and 8 hours, were averaged to estimate the rate of in situ mud respiration. The daily values ranged be2 2 tween 0.258-0.452 g /m day with an overall average of 0.36 g /m day (Table 13) . These values compare favorably with results from the study by Hayes and MacAulay (1959) who used sophisticated coring devices to minimize disturbance of the mud structure. Most of their values for small Canadian 2 lakes fall within the range of 0.10 to 0.40 g Ojm' day (11 C incubation

PAGE 187

6 S o p +-• o Xi AS J S o u '-H
PAGE 188

170

PAGE 189

Figure 40.Respiration rates of mud samples from Lake Izabal. Arrows on the horizontal axes indicate midpoints between samples and the lengths of vertical bars represent differences between duplicate rate determinations.

PAGE 190

172 1 2 3 4 5 6 Hours

PAGE 191

173 Table 13.Hourly and daily respiration rates of Lake Izabal bottom muds

PAGE 192

174 temperature) . Results are about an order of magnitude higher than this for river muds where stirring is used in the incubation chambers. 2 McDonnell and Hall (1969) reported hourly rates of ca. 0.22 g /m for a "mildly polluted eutrophic stream," and similar values (0.10-0.20 g 2 /m hr) were reported for river muds in England by Edwards and Rolley (1965). The central, nearly fiat basin of Lake Izabal is a very uniform clay-gyttja of moderately high organic content, while inshore regions contain considerable sand and gravel of alluvial origin (Tsukada and Deevey 1967) . The moderately high organic content of the clay-gyttja (8.7-12.9% loss on ignition) is mostly carbonized plant fragments. Tsukada 's cores, for which these analyses were reported, are well below the surface of the mud (to 3.2-m depth) and have long been unavailable to potential detritus feeders and decomposers. To estimate the quantity and organic composition of the mud surface, especially the particulate detritus, Ekman samples were collected and treated as reported in the "Methods." Figure 59 illustrates the results obtained from samples collected throughout the lake basin. Samples from inshore stations near river outfalls and in bays and coves generally contained larger quantities and percentages of combustible particulate organic matter than did offshore samples. Offshore stations below 10.5-m depth (Nos. 6, 9, 10, 14, 17, 19), excluding stations where large pieces of plant material were collected (Nos. 15 and IS), yielded between 14.6 and 41.0 percent loss on ignition (x = 25.8%) for particulate organic matter. These values, all being proportionately higlier in organic matter than the deeper muds, possibly reflect a bias toward higher

PAGE 193

175 organic content of the particulate fraction, but also imply that the mud surface contained substantial amounts of organic matter potentially available for detritus consumers. Measurements of Primary Productivity One of the chief difficulties of the light and dark bottle method for primary productivity measurements is expressing the results obtained during a 3-hour measurement on a daily basis. To do this the following calculations were made as follows: (1) The dissolved oxygen concentrations at the beginning of the experiment and those after incubation in the light and dark bottles were plotted on graph paper with oxygen (mg/litcr) as the abscissa and depth (m) of incubation as the ordinate. The area between the light and dark bottle curves was integrated by planimetry, the averages were calculated for the duplicate experiments, and the value was defined as gross primary production (Pg) for the incubation period. The area between the dark bottle and initial readings was the respiration (R) (Figure 41) . (2) The Pg and R during the incubation period were divided by the number of hours of incubation to yield hourly gross primary productivity (Pg/hr) and hourly nighttime respiration (R/hrJ . (3) The difference between the Pg/hr and R/hr was the net primary productivity (Pn/hr) . All these values are listed in the first three columns of Table 14. (4) Next, the hourly radiation received during incubation, calculated from the solar radiation recorder readings, was divided into the monthly average of daily radiation received. The result is the number of hours of effective radiation per day. Tlie assumption is that daytime

PAGE 194

Figure 41.Example of curves generated from light and dark bottle experiments from which metabolism is determined planimetrically. The area between light and dark bottle curves is gross primary productivity (Pg) , and the area between initial and dark bottle curves is respiration (R) for the incubation period. Net primary productivity is the arithmetic difference between Pg and R.

PAGE 195

177 Oxygen (mg/ liter) 8 18June72 Station C 0912-1212 Q) 6 8 "--x Initial Concn. <^-oLight Bottle •— *Dark Bottle 10 Comoensation Depth

PAGE 196

178 primary production maintains the same proporLiojial relationsliip to total daily light as hourly primary production during incubation does to hourly light received during incubation. It was believed to be more realistic to use a monthly average of daily radiation rather than radiation determined on a single day for extrapolating daily metabolism over longer periods. A hypothetical example of this calculation is given in Figure 42. (5) Both Pn/hr and Pg/hr are multiplied by the effective hours of light per day to yield Pn/day and Pg/day respectively. (6) The ratio of gross community metabolism (Pg) to 24-hour respiration (R-j.) was calculated from the following formula: Pg/day _ (Pg/hr) (hrs of effective light) R^^ ' (R/hr) (24 hrs) This is the P/R ratio used by Odum (1956) and Margalef (1963) as an indicator of biomass accumulation (P/R>1) or consumption (P/Ps
PAGE 197

Figure 42.Method for calculating the number of hours of effective light per day.

PAGE 198

180 0600 1200 . ' . Hour incubation period 1600 50 kcal , 2 1cm 50 kcal/m^ 2 2 Total Radiation = 64.07 cmVday x 50 kcal/m cm = 3,202.5 kcal/m day Radiation 2 2 during Incubation = 25.66 cm^/3hr x 50 kcal/m^cm 427.7 kcal/m hr Total Radiation ^ 3 ,203.5 kcal/m day ^ y ^^^ hours of effective Incubation Radiation 427.7 kcal/m-hr ' light/day

PAGE 199

181 Table 14 Metabolism calculations of light and dark bottle experiments for Stations A, B, and C during 1972. Rates of gross primary productivity (Pg) , net primary productivity (Pn) , and respiration (R) are in g 02/m^

PAGE 200

182 Data for Stations B and C are available for only May through Scptem2 ber. At Station B the highest Pg (6.41 g /m day] occurred on June 5 2 decreasing to a low of 2.42 g /m day in September. Respiration was 2 2 lowest on June 5 (1.85 g /m day) and highest in August (8.58 g /in day) The highest Pg/R24 ratio was on June 5, the day of highest Pg; the lowest ratio (0.56) was measured in September, the day of lowest Pg. 2 At Station C, Pg was highest August 7 (3.44 g /m day) and lowest 9 September 25 (1.22 g /m" day). Overall, respiration rates were less 2 variable than those of other stations; a low of 2.52 g /m day was mea2 sured May 5 and a high of 5.91 g /m day on June 18. Ratios of Pg/R ranged between the May 5 high (1.21) to the September low of 0.32. The Pg, R_., and the Pg/R_. ratios are summarized graphically in Figure 43. Because eight months of data are available for Station A, the seasonal trend of metabolism is especially well developed. The increase in Pg during the dry season (February-May) was followed by an increase in R . This resulted in a change in P/R ratio from greater than one in April to less than one in May. Respiration continued to exceed gross primary production throughout the wet season except for the August reading. At Station B excess Pg over R continued until July, and thereafter the ratio was less than unity. At Station C metabolism rates were generally lower and less variable than at Station A and B. The highest rates of Pg at Station C were approximately one-half the values of the highest rates at Stations A and B. To summarize. Stations A and B appeared to have trends of metabolism which roughly corresponded to the change from the wet to the dry season and had periods during which the P/R ratio was well above unity. At Station C

PAGE 201

w rt O •H t/) +-» O rt <+^ •H -P O +J CO Cj (/) 5h ^< T3 O O (N U oi m O W) '^ M -O (H -H Q) O 3 -P •H -P +J W Oj rt r-i !~i O Pi 3 -H t/5 O C (D 03 ^ (1> M ^1 -P O O t+H r-l ^ c o CnI O -P •H !/l 0) ^ C ^ M (1) P Cu -P (U 4-) p o e •H .C -H > H P •H W +-> 0) o • 3 U O -a -p O T3 w X 3 • rt
PAGE 202

184 u c o n 1/5 _l I I I<

PAGE 203

similar seasonal trends were lacking; the P/R ratio was more commonly less than unity, and during no period did the P/R deviate far above one. Thus, the western and central parts of the lake (Stations A and B) seem to be areas where production seasonally exceeded respiration, whereas the eastern part of the lake (Station C) demonstrated less of a seasonal pulse in metabolism. Efficiency of gross primary productivity The efficiency of primary productivity will be defined as the fraction or percentage of incoming radiant energy (available to photosynthesis) that is converted into chemical energy by gross primary productivity. The energy of the spectral region available for photosvnthesis (400-770 my) was assuii'.ed to be 50% of the energy recorded by the pyrheliometer. Total solar radiation, recorded almost continually from October 1971 through November 1972, is reported in Table 15. Daily readings were averaged on a weekly and monthly basis. The monthly averages were used to calculate efficiencies of Pg/day, and the hourly radiation, determined during the incubation period, was used to calculate efficiency of Pg/hr (Table 16). Since solar radiation was expressed as kcal/time, the productivity had to be converted from grams oxygen to equivalent energy and time units. The choice of 4 kcal/g lies between the value for one of the products of photosynthesis (glucose) and the final products of biomass (plankton) . If glucose were the only product of photosynthesis, approximately US kcal of solar radiation would be required per mole of oxygen, or 3.7 kcal/g 0^. If biomass were the product of photosynthesis, the 4.8-5.0 kcal /g biomass (or equivalent) represents the energy content of plankton (Maciolek

PAGE 204

186 Table 15.Total daily incoming radiation averaged on weekly and monthly basis from October 1971 through October 1972 Dates of Measurement

PAGE 205

187 Table 15.continued 2 2 Dates of No. Days kcal/m day kcal/m day Measurement Measured (weekly) (monthly) Jul.

PAGE 206

188 Table 16.Daily efficiencies of energy conversion froiri visible solar energy to energy fixed by gross primary productivity (Pg) . The energetic equivalent of one gram of oxygen was assumed to be 4 kcal and visible solar radiation was assumed to be 50% of total incoming radiant energy , 50% of Total ^^'^^y Radiations Percent g Oo/m^ kcal/m2 (kcal/m^ day) Efficiency 1,737 0.95 1,737 0.81 1,761 0.26 2,275 1.07 2,283 1.28 2,281 0.85 2,281 1,03 2,200 0.78 2,243 0.80 2,086 0.33 1,859 0.57 0.79 Ave. Station

PAGE 207

189 1962) . Since neither glucose nor highly structured plankton biomass are the sole products represented by 0^ evolution in the experiments, the intermediate value of 4 kcal/g seems justified. The daily efficiencies of gross primary productivity reported in Table 16 ranged between 0.23-1.28°6. Solar radiation was more constant that primary productivity which resulted in efficiencies being more proportional to productivity than to radiation. Efficiencies greater than \% were exceeded only three times at Station A and twice at Station B 2 when the primary productivity was greater than 20 kcal/m day. Light penetration and compensation depth At least twice monthly, Secchi disk transparency measurements (except July) were recorded at Stations A, B, and C (Figure 44) . There was a general trend of decreased transparency as the dry season progressed; the least transparent readings were during July (1.5-1.7 m.) . Thereafter, a trend toward greater transparency increased values to 3.1-4.2 m in October. A submarine pliotometer was not available at Lake Izabal to measure the percent extinction at the depth of Secchi disk transparency. The extinction coefficients were calculated from Secchi disk transparencies from the relationship 1.4 where ri is the extinction coefficient and z is the depth in meters of the Secchi disk transparency. The choice of the constant 1.4 is explained in the following paragraphs. The light intensity at the limit of Secchi disk transparency is about 15°o of the subsurface intensity according to several authors (Poole and

PAGE 208

rt

PAGE 209

191 u u LlI en u c o *-> CO X u u LlI to c o 03 -t-' LO I u u UJ LO < c o +-' 05 to c ro O JL U I c o to o £ a E C o in 03 E 03 E c o u c o to 03 E 03 E c\j n '^ Lf) ( uu ) A V 1 1 q ! s I A (D :z O

PAGE 210

192 Atkins 1929; Clarke 1941; Ichimura 1956; Beeton 1957). According to Beer's Law of light extinction Iz = I e"'^ ^m o where I represents the light intensity at depth z, I the incident light z o impinging upon the water surface, and n the extinction coefficient. If I is 15% of I , then, z o 15 = 100 e"^ ^m n 100 -

PAGE 211

193 in 24.7% of the surface intensity reaching the depth of Secchi disk visibility. Table 17 presents the data for Secchi disk transparencies and calculations which relate to primary productivity experiments. Compensation depths were determined from the graphs used for planimetric determination of primary productivity and represent that depth where the light bottle concentration at the end of the incubation period crosses the initial 0_ concentration (Figure 41) . Extinction coefficients ranged between 0.412 and 0.966. The percent of surface intensity reaching the compensation depth ranged between 1.68 and 3.29, which is slightly greater than the 1% level generally considered to be the compensation intensity (Ruttner 1963). The disagreement is probably due to two factors: (1) an error in the assumed light intensity of Secchi disk readings and (2) the lack of good resolution in determining the compensation depth due to bottles not being used at 1-m intervals in depth. In spite of these factors, the observed compensation depths (3.75-8.93 m) overlapped to a large degree with the calculated depths of 1% light penetration (4.94-11.18 m) . B alance of the Organic Matter Budget The results presented in the preceding sections represent most of the principal flows and storages of organic matter in the lake. In order to evaluate these data, a simplified model was used as a tool for comparing the relative importance and magnitude of the organic matter flows in relationship to the seasonal regime (Figure 45). The energy language of Odum (]971) was used to illustrate these flows and storages. The central pool or storage represents total organic matter (living and dead) whicli cither gains or loses organic matter depending on the relative magnitudes

PAGE 212

194 >s

PAGE 213

-c

PAGE 214

196 fH o w o H CO

PAGE 215

197 of the inflows and outflows from the pool. Inflows are represented by importation of organic matter from the watershed and by the contribution of gross primary production. Losses from the pool include exports of organic matter, loss by planktonic respiration, and by benthic respiration. The components of the model parallel measurements that were made in the field over an eight-month period, except for benthic respiration (one month of data) . It must be remembered, however, that these were not measurements for every day of the month, but rather one to several measurements during a single month at a limited number of sampling sites. The eightmonth sampling period included three months of dry season (March-May) and four months of wet season (June-October). Since the Izabal Basin was characterized by a longer wet than dry season, the balance of the sampling period was toward the wet season, thus closely representing the hydrological regime over a full year. Also the sampling period included what could be considered the maxima of the two seasons, i.e., the dryest part of the dry season, and the wettest part of the rainy season. Table 18 summarizes the organic matter flow data according to the flows defined in the model. The values are daily rates and are assumed to be the average for each day in the month. The calculations on which Table 18 is based are given in the Appendix, Tables I and J. Oxygen production and consiunption values were assumed to be equivalent to organic matter (OM) values. Some interesting ecosystem cliaracteristics become apparent upon examination of Table IS. For example, Pg remained relatively constant from Marcli-August (above 4 g OM/m" day) and then declined to below 2 g 2 OM/m day be October. However, when Pn was exajnined, there was a net

PAGE 216

198 u o M o o X

PAGE 217

199 positive gain only during the dry months (March-May) and values became increasingly negative throughout the wet months (June-October) . The excess production by photosynthesis during the dry-season maximum was ca. 2 g 2 OM/m day, a value close to the maximum consumption over production in October. The organic detritus imports and exports can be evaluated best by examining their net gains and losses. In June and July there were net 2 gains of 0.501 and 1.045 g OM/m day respectively, which compensated for the net loss from respiration exceeding gross primary production during those months. For all other months, the net gains or losses of organic detritus were small when compared to other flows. Consumption of organic matter by benthic respiration was assumed 2 constant at 0.36 g OM/m day for all montlis. It was clearly a flow of importance to the overall balance of organic matter, for it represented a relatively large percentage of the total energy drain from the ecosystem when compared with other flows. The balance of ecosystem gains or losses reflected the same general trend as did Pn with some modifications. For example, the May value was slightly negative, owing mostly to benthic respiration cancelling the gain due to Pn. June and July losses were somewhat ameliorated by the large positive gains of organic detritus during those months. The average values for each of the organic matter flows were calculated in order to arrive at an overall evaluation of potential and realized ecosystem metabolism. Likens (1972) used the term "ecosystem source carbon" to include all reduced carbon compounds that can provide energy for consumers (and presumably decomposers). In Lake Izabal where

PAGE 218

200 horizontal displacement was of considerable magnitude due to the short residence time of the water mass (6.6 months). Likens' term would be equivalent to all the organic matter that flowed through or was metabolized under a unit area of ecosystem. This would be represented by the Pg and OM imports. As calculated in Table 18 (last column) this 2 2 value averaged 4.362 g OM/m day or an annual total of 1,592 g OM/m . This value can be compared with the production of other tropical and temperate lakes listed in Table 19. Lake Izabal ranks below that of several other tropical lakes, but above the range in annual rates for eutrophic temperate lakes. The overall balance during the sampling period was a negative average 2 of -0.281 g OM/m day and would imply that there was a rather large deficit in the annual energy budget for the lake. The negative balance may be attributed to a series of factors, such as the failure to account for measurements during the remaining four months of the year. Another possibility is failure to measure the energy flows with exact precision. Measurements that may have slightly underestimated organic sources and overestimated organic losses would have the combined effect of producing a deficit in the final estimate of the budget. However, the purpose of the organic matter budget was not so much to prove that the ecosystem was or was not at steady state, but rather to compare the relative magnitude of organic flows and to demonstrate the seasonal activity of metabolism and organic balance. This it clearly does, and the positive dry-season gain of "surplus" organic matter corresponded closely with an increase in the activity of consumers in the lake, especially fishes.

PAGE 219

201 c V) oi o •H +-> I V) U o CL, P •H > •H •P O 3 O X •H rt ^1 —I Oh to 03 O J-i ^( <+H Oi U W 0) &0 C P. C3 0) T3 -H C -P •H S ct5 O Q O o> 1—1 O 3 O $-1 c a. 2 c o < I— I CM 05 a, "-. Q U o o 2 Q 03 Oj(N D < ^ 3 ^ •P E U O § ^ •H H 4-> CO (U O AS O 03 C-, E

PAGE 220

202 Summary Statement In the Izabal Watershed, the transfer or flow of organic matter from the terrestrial to the aquatic ecosystem was tightly linked to the hydrologic regime. However, the flow did not perfectly parallel the movement of water because of the higher organic matter concentration that occurred at the beginning of the wet season. This initial "flushing out" of organic matter from the watershed occurred during June for the Rio Polochic, but a month later for the colored swamp waters of the delta. During these months, the waters contained far higher proportions of particulate organic matter than during the remainder of the sampling period. During this initial flushing pulse, the inflow of organic matter exceeded the outflow at the San Felipe outlet. By August, the rates of inflow and outflow began to converge and both continued to decline at the end of the sampling period. Organic matter that discharged into the western end of the lake from the Polochic Delta accounted for approximately 90"o of the organic detritus contributed by runoff. No correlation was observed between organic matter (COD) concentration and respiration rates (BOD) of any of the inflowing or lake waters. Gross primary productivity (Pg) by phytoplankton represented the major source of organic matter for the lake, and the efficiency of Pg to total solar radiant energy ranged between 0.23 and 1.28"o. IVliereas Pg remained relatively constant throughout most of the sampling period, respiration in the water column was relatively low during the dry-season months compared to respiration rates that followed the onset of the wet season. The resulting balance of Pg and respiration (e.g. net primaiy production) was positive during the dry season, and became increasingly negative during tlic wet season.

PAGE 221

203 Respiration by the bottom muds was approximately one-tenth the rate of planktonic respiration per unit area of lake. Thus it represented an important sink for organic matter, as might be expected in a shallow lake. The particulate component of benthic organic matter m.ay have been an important source of energy for activity in the muds. Over half of the organic matter contributed by the Rio Polochic in June was of the particulate fraction, yet it did not appear to measureably increase the particulate fraction in the lake samples during June or the months that followed Figure 34d). The disappearance of particulate matter would have been, by inference, from the water column to the bottom muds through sedimentation. The profile-bound density current of the Rio Polochic outfall may have served to isolate the particulate matter from the overlying waters and thus facilitated its sedimentation. The primary productivity of Lake Izabal was within the range of polluted eutrophic temperate lakes, but below the productivity of many tropical lakes. The replacement of a large portion of the lake water by high-turbidity runoff water possibly reduced transparency sufficiently such that the primary productivity did not approach the maximum theoretically possible for planktonic systems. Data are lacking to help determine the direct causal effects which resulted in a net positive balance of energy during the dry season and its apparent consumption during the following wet season. However, these seasonal pulses of production and consumption suggest a mechanism by which the Lake Izabal ecosystem, and perhaps other similarly structured ecosystems, can achieve a steady-state balance of organic matter storage.

PAGE 222

CONCLUSIONS In the preceding chapters it has become clearly apparent that Lake Izabal is not an isolated ecosystem, but a unit of a larger ecosystem by virtue of flows that couple it with the upstream watershed and the downstream marine environment. The overriding force that controls this coupling is the hydrologic regime, whose seasonality is responsible for the pulses and oscillations characteristic of the ecosystem. By determining the relationship between different components (subsystems) of the Izabal Watershed, one can assign distinguishing features, or attributes, that make this ecosystem distinct from others. It follows lIiaL any implicaLluiis of the study for ecusystem managcjaeiit inay be applicable to other ecosystems with similar attributes and vice versa. The following are attributes that seem to be of importance for characterizing the Izabal Watershed: (1) The ecosystem possesses a seasonal pulse of organic detritus movement . (2) Tlie lacustrine component of tlie watershed experiences oscillations in food concentration and consumer activity. (3) The connection of the lake to a marine environment provides easy access for euryhaline marine fishes that can adapt to freshwater conditions . These attributes are illustrated in Figure 46 as storages and flows of energ)' and matter between the subsystems of the Izabal Watershed. The 204

PAGE 223

NO u •H

PAGE 224

206

PAGE 225

207 energy diagram is representative of the regional situation before the additional influences of modern agricultural and industrial man. The four main subsystems are (1) the terrestrial ecosystem (watershed), (2) the lagoons and coves in the Polochic Delta, (3) the main basin of Lake Izabal, and (4) the coastal marine ecosystem. The forcing functions that provide the energy for ecosystem power, maintenance, and intersubsystem coupling are (1) sun and wind which drive primary productivity and evapotranspiration, (2) rainfall, which in connection with seasonality, controls the hydrologic regime and water movement, and (3) tide, gravity, and wind whose seasonal regimes act with water level to control fish migration and salt water movement from the coastal marine ecosystem. Terrestrial ecosystem material exports to the lagoons and coves and the lake basin are carried by the energy provided by the do^mhill flow of water. Exports from the lagoons and coves to the lake basin include organic matter originating in the swamp forest of the delta region as well as organic matter from algal production. These exports occur in proportion to the water level of the lagoons and coves. The lake basin not only receives inputs from upstream ecosystems, but also receives salts and fish from the coastal marine ecosystems. Many of the internal flows and storages of the lake have already been discussed. Fishermen take advantage of the seasonal migration of fish which is possibly stimulated by factors such as salt-water movement, water levels, and other seasonal phenomena, especially the food concentration resulting from dry-season net productivity of phytoplankton in the coves and lagoons. These interactions and ecosystem attributes will be discussed in relationship to their ecological implications for the aquatic ecosystem.

PAGE 226

208 Seasonal Pulse of Organic Detritus Movement It was shown that nearly half of the total organic detritus movement from the watershed to the lake occurred during 3 months, or just one-fourth of the year. The timing of this event coincided with the transition from a positive balance of gross primary production and respiration, to a negative balance. Although there was no direct evidence to demonstrate that this transition was a result of increased organic matter input to the lake, the change from an autotrophic (P/R > 1) to a heterotrophic (P/R < 1) balance suggests that it was at least a seasonal phenomenon associated v;ith the hydrologic regime. Rates of gross primary productivity remained nearly constant during the dry season and continued at nearly a constant rate during the initial months of the wet season. Thus the change from a P/R ratio of greater than unity to a P/R ratio of less than one was due to an increase in respiration rate. The wet-season inflow of allochthonous organic matter represents an auxiliary energy source for the maintenance of high respiration rates of the lacustrine ecosystem. Similarly, increased respiration rates have been documented for estuarine ecosystems during periods of increased fresh-water inflows which contain organic detritus (Cooper and Copeland 1973; Odiim 1967). Lake Izabal thus has metabolic characteristics similar to shallow estuaries, although physical features such as depth and salinity may be different. Movement of Organic Matter to Site of Consumption The swamp forest in the Polochic Delta is the source of some of the organic detritus that is discharged into the lake. This represents a

PAGE 227

209 displacement from anaerobic conditions (the swamp forest) where organic matter consumption occurs slowly, to an aerobic environment (the lake) where the consumption of organic matter is more rapid and complete. Thus the physical flushing of the swamp forest is a mechanism that prevents large accumulations and storages of energy. The movement of organic matter from other regions of the watershed occurs in two phases. First, there is the initial flushing at the beginning of the wet season, when a high percentage of the organic detritus arrives at the lake in particulate form. Following this, the composition of influent detritus is predominately in dissolved form, and represents a more steady leakage from the terrestrial ecosystem. It is not clear whether this movement is from an environment of relatively slow organic matter consumption to a more rapid one, or to what extent metabolism that occurs in the rivers is responsible for altering the quality and quantity of detritus before it reaches the lake. In the case of the particulate detritus, the bottom muds of the lake and its associated fauna may provide conditions for the effective consumption of particulate organic matter. By dividing the total annual OM runoff from the Izabal Watershed (163,051.25 X 10 g OM) by the area of the watershed (6,860 km ) a value 2 of 23.8 g OM/m is obtained. This is well above the annual runoff of 7 5.3 g OM/m for the Hubbard Brook Experimental Forest (Likens 1972). The higher value for the Izabal Watershed might be due to increased OM runoff from a partially deforested ecosystem. However, it is likely tliat terrestrial ecosystems in the humid tropics may experience more OM leakage than their temperate counterparts. Until more data on OM runoff from a broader spectrum of latitudes and rainfall regimes are available, this

PAGE 228

210 supposition is tentative. However, based on what is known about ecosystem strategy toward the conservation of mineral nutrients, it is unlikely that OM would be regarded a "scarce" material because of its abundance in relatively productive terrestrial ecosystems. There is evidence in this study and others (Nelson and Scott 1962; Bormann et al. 1969) that in a given watershed OM runoff increases in greater proportion than hydrologic runoff. Thus, in an assemblage of watersheds with a gradient from low to high hydrologic runoff on an annual basis, O.M runoff could be expected to increase in greater proportion than hydrologic runoff. Assuming that the OM produced in one part of a watershed is utilized or consumed in another region leads to the possibility that regional coupling mechanisms are more predominant in areas of higher hydrologic runoff, i.e. higher rainfall. The conceptual image that emerges is a region composed of a mosaic of subsystems where upstream subsystems export potential energy in the form of organic matter to the more predominately heterotrophic sybsystems do^^mstream. The export of this energy is subsidized upstream by the weathering of rocks which yields nutrients for primary production. Mechanisms for Steady-State Balance According to the measurements during the sampling period, the organic matter budget shows a deficit, but his should not be interpreted to mean that the lake is a strictly "heterotrophic" ecosystem. Interpretation of the budget needs some qualification with respect to tlie year of study and adequacy of sampling. The hydrologic regime during 1972 was unusual due to greater than average rainfall in the region of the lake. Not only was rainfall intensity greater during the wet season, but the dry season

PAGE 229

211 (<]00 nim per month), which typically begins in December, did not start until March 1972. Thus, the 1972 dry season was only three months duration, while five months constitutes the average dry season (Snedaker 1970). The balance of organic production to consumption is positive during the dry season and negative during the wet season as the data clearly demonstrate (Table 18). A year of atypical wetness, such as that experienced during the year of study, may be responsible for the "heterotrophic" character of the lake. Measurements would be required on more t>i)ical years in order to establish if the lake is characteristically heterotrophic. Regardless of the nature of the annual balance, positive gains of organic matter during the dry season are followed by net losses during the wet season. This may be a mechanism by which the lake achieves steady state with respect to organic matter loading. Flows or sinks which prevent organic matter overloading are respiration (planktonic and benthic) and export at the lake's outlet. Lesser flows which were not measured would include the fossil sink for organic matter in the sediment of the lake and emigration of fish to the marine environment. Positive gains in net production were probably underestimated because the intense dry-season primary productivity of the deltaic lagoons (which are some of the highest daily rates for the region) (Brinson 1973), were not calculated as part of the budget. Although these lagoons represent only a small percentage of the total surface area of the lake, their regional importance is magnified by attributes found in no other part of the Izabal Watcrslicd.

PAGE 230

212 Oscillations in Food Concentration for Consumers As mentioned in the description of the study area, the shallow coves and lagoons of the Polochic Delta are popular areas for the activities of fishermen and fish-eating birds during the dry season. The high rates of net primary productivity and high densities of planktonic standing crop provide conditions attractive to fishes. The shallow depth relative to that of the lake basin offers the distinct advantage of reducing energy expended in feeding by concentrating the food in a compressed column of water. Dry-season activities of the consumers coincide with cloudless days which provide optimal conditions for photosynthesis and low or negligible rates of flushing during the hydrological m.inimum. Cichlasoma gutulatum that spawn in these shallow waters have an adaptive life cycle that entails breeding during the dry season when high planktonic densities provide a concentrated food source for their offspring. The seasonal availability of a concentrated food source in fresh-water systems could also be of selective advantage to those marine euyhaline species that can reach this energy source before runoff and flushing of the lagoons and coves dilute the accumulated organic matter by the time it reaches the sea. A similar account of seasonal food exploitation is the tismiche described by Gilbert and Kelso (1971) in the estuary at Tortuguero, Costa Rica. Following these dry-season activities, wetseason rains flush the lagoons of their plankton with anaerobic black waters from the surrounding swamp forest, and the majority of the fish leave to become dispersed throughout the lake. The pulse is somewliat analogous to management practices used in fish pond culture, whereby ponds are periodically drained before a new crop of fish is cultivated.

PAGE 231

213 Consequences of the Connection to a Marine Ecosystem The dominance of the lake's fishery by marine fishes provides clear evidence that the Rio Dulce-El Golfete waterway has profound effects on the faunal composition of the lake. The extent to which a brackish water gradient to the lake regulates migration or facilitates the adaptation of euryhaline marine species to fresh water is probably dependent on the inherent physiological mechanisms for osmoregulation possessed by each species. However, the dry-season penetration of brackish water into the lake may serve to stimulate migration. The periodicity and intensity of migrations is an obvious research need before the fishery can be rationally managed. There is substantial evidence for control of the Na:Cl ratio in the lake by seasonal inflow of brackish water from the sea. Regardless, the lake water did not approach oligohaline concentrations during the study period, but there is evidence that suggests salinities are higher during some years. Because of the short residence time of the water in the lake [<1 year), it would be difficult for higher salinities to escape being diluted and flushed out during the ensuing wet season. Therefore, those euryhaline species that do populate the lake must either be extremely well adapted to fresh water, or have behavioral patterns that facilitate extensive and frequent migrations to conditions of higher salinity. Ecosystem Management In the "Introduction," I suggested that watersheds, because of their well-defined boundaries, make conceptually attractive ecological units for demonstrating regional relationships and coupling. The I-abal Watershed

PAGE 232

214 has been described a<; one that fits this criterion and many of the mechanisms by which this ecosystem sustains or manages itself have been discussed, Wherever man lives, he manages or manipulates ecosystems to some degree and the Izabal Watershed is no exception. There has been a long history of moderate human activity there, but the recent rate of increased development could conceivably put a strain on the free services the ecosystem provides to man without man-made reciprocal investments of energy or materials (feedback rewards) . The apparent decline of the fishery exemplifies this notion. Fishery Management In the section--"Regional Setting"--the status of the lake's fisheries was; briefly described and the need for management was emphasized. In addition to the extensive management of the natural fish populations through fishing controls and regulations, intensive management methods, by means of aquaculture, deserve consideration. Where aquacultural techniques are employed successfully in other humid tropical regions (Hickling 1961), they are usually closely associated with cultural diet patterns and often accompany wetland rice cultivation. Raising fish in cages has been suggested for Lake Izabal (T.C. Dorris, personal communication), cind rearing of Cichlasoma gutulatum fingerlings to marketable size has been tried in an enclosure near the lake's edge by one individual. The question is not whether aquacultural tecimiques will work in Lake Izabal, but whetlier the energetic subsidies required for fish production ^ See H.T. Odum (1971) for a full discussion of man's partnership with nature.

PAGE 233

215 can be provided. The present fish production in Lake Izabal is provided as a free service by nature. Raising fish in cages necessitates additional energetic subsidies in equipment, feed, and human labor, not to mention possible disease control measures. Projects in which aquaculture is attempted should account for these extra costs, and balance them with the marketable value of the fish produced. At the same time these costs should be compared with the cost of extensive management techniques that would be necessary for maintaining a maximum sustained yield for the freeliving fish populations. Modern Agricultural and Industrial Man The energy diagram in Figure 47 illustrates the possible influences and modifications that industrial and agricultural development could have in the Izabal Watershed. Additional forcing functions not included in the previous diagram (Figure 46) include fossil fuel and outside energy and economy. Supplies from outside the ecosystem, such as nylon gill nets, outboard motors, and gasoline have already increased the rate of fish removal from the aquatic ecosystem. Preliminary mining activities have stimulated immigration of people into the region and have increased the exchange of money, goods, and services. The eventual export of mining products would presiunably provide capital inputs for more land development. Since the hydrologic pattern of the watershed is the overriding feature that regulates the attributes associated with special characteristics of this ecosystem, schemes that involve alteration of this pattern should receive careful study and consideration. Two schemes for development that could potentially alter hydrologic patterns are (1) impoundment of water in the Polochic Valley for hydropowcr

PAGE 234

13

PAGE 235

217

PAGE 236

218 and C2) channeling the shallow El Golfete to facilitate access to the lake by ships with deep draft. The first alteration might have downstream effects on the Polochic Delta. Some regions of the delta are currently growing toward the lake as determined by comparing old maps with present conditions. Entrapment of the sediment load of rivers by upstream impoundments would reduce the alluvium available for delta growth. The small erabayments protected by the levees of the Polochic distributaries are the areas where high rates of primary production during the dry season provide a concentrated food source for consumers. The continued maintenance of these coves depends upon the degree to which the protective levees are dependent on sediment loads received during the wet-season flood stage of the Polochic distributaries. Also the upstream impoundments would tend to dampen oscillations between wet and dry season discharge rates. These pulses currently function as the seasonal controls on consumer activity in the lagoons and coves. Channeling the relatively shallov\? El Golfete (4.5-m depth) that provides a formidable barrier to brackish water penetration may aid in dry -season movement of saline water to the lake. Water impoundment in the Polochic Valley may deter the upstream penetration of brackish water by discharging wet-season storages during the dry season. However, the effectiveness of a fresh-water current from the lake in displacing brackish water must be weighed against the opposing forces of tides, winds, and gravity that facilitate the upstream penetration of brackish water. In the event that channeling El Golfete facilitates the upstream movement of coastal waters, one can only speculate on the consequences of massive inputs of saline water into the lake. Some clianges surely could be expected.

PAGE 237

219 and Lake Maracaibo, Venezuela may be a good model upon which to judge these effects. Some of the basic data obtained in this study may be of use in determining the consequences of large scale hydrologic changes. Questions raised by this investigation, especially with respect to fisheries management, will require specific data for answers. However, the unknown periodicity of shallow tropical lakes referred to by Tailing (1969) has been established for at least one of these lakes by demonstrating that the watershed is closely coupled with activities in the lake.

PAGE 238

APPENDICES

PAGE 239

221

PAGE 240

222 u

PAGE 241

223 Table CCross-sectional areas, velocities, and monthly discharges for the major influent rivers sampled CrossMaximum Maximum Sampling Sectional Velocity Discharge Date Area (m^) fm/secl fm-^/sec) Comercio (Polochic) 4 Nov 71 119 0.538 64.02 124 0.699 86.68 119 0.447 53.19 120 0.458 54.96 114 0.263 29.98 118 0.169 19.94 116 0.222 25.75 120 0.685 82.20 127 0.736 93.47 147 0.237 54.84 131 0.658 86.20 23 Oct 72 127 0.729 92.58 Coban (Polochic) 30

PAGE 242

224 Table C. continued Sampling Date CrossSectional Area (m^] Maximum Velocity (m/sec) Maximum. Discharge (m-^/sec) Average Discharge (m /sec) Monthly Discharge (m^xlO^) Rio Oscuro 3

PAGE 243

225 ^ c

PAGE 244

226 a rt

PAGE 245

227 in I o I o LO o LO I o I o IT) o c o o I u X o^ o •H +J (-' CO CM.— Ii— (r-H.-Hr\li— IrH I OOOOOOOO I o o o o coo LO LO O SsssssissSSSssssssss S vD M ^£) -I vi^ O O vO VD C7i vO t-vO ^_ 00 r<^, ^, f^. ddoo^ddoooodoodocM'-icorsj VV V VVVVV V vv LOOLOLOLOOOl/^lOLOLOLOLOOOLOLOO SSoodr-i(MOOOOOO<^_o-|OOOr-H dddddddddddddooooooo o CM -^ o rr-J 't '-I '^ o \o LO -^^ t-o "tt "* -H o o _;^^'^^^'^^<^d-^LO^i>j-^fO^£)-s-oo V V VO r-l to to to to O I I o \0 I o to to o I 1 I I I I ^ ^ ^_^ ^ -^ ^ o LO r^ LO LO vc vD to "* LO to to r-^ ^ ooooooioooooooooinoooLO ,^ oi ^* to' r-j -t' CO -^ ^ 'j t^ to •=3-I ^ o CO CO r-i vD cr>cooojo(MLOtoo]Lor-~o\Otot--t^t-~vooo ^(>]tO(^J(NfNr^tototo!>itotototototocsiorto ^ LO LO LO vO(>J.-Jvr)t--vOOI^(MtOvDtO'vf.-l(NLOrNj(NO^O dr^r-ddt^di-^ocJr^i-ir^r^tor^Lntoiqtoo lOLOOLOOlOOOOJOLOLOr-lOLOLOLOLOOOO dLOLodddi-^rvJi-^r-Jd-^-HLOr-^w^cnvo^ o bO C •H fH P< CO ?J S 5 LO --^ 2 (C| .0 (^ ^ (D O CO coi-HCO'-icO'-< ^ "Sin ^ WOfHOJ DQ .Su-S -I oc;ccr:cc!^-;^t^rt3D^ M O C C ••-> C^3 M -H -H -H .H -H •S «, O CO -H Q cjc:;ujucS-5Jwcocococotococos:
PAGE 246

228 60

PAGE 247

229 m 3 •H ?H +-i !/) 0) 03

PAGE 248

230 in (-> O cn ^< > S o c O s to I .a

PAGE 249

231

PAGE 250

232 Table F-5.Monthly concentrations of organic matter (mg COD/liter) for the San Felipe outflow mg COD/ liter Total Dissol. Part. Ne;^ 8.63 7.38 1.25 .023 9.00 8.60 0.40 .030 9.62 8.09 1.53 .021 .047 8.67 7.78 0.89 .081 10.37 8.81 1.56 .058 7.55 7.27 0.28 .057 ^ Where two depths are indicated but only one set of determinations reported, analysis was made on a mixture of samples from two depths. The four fractions represent Tctal COD, Dissolved COD (<0.80y). Particulate COD (>0.80y), and Net Particulate COD (>76y) .

PAGE 251

233 Table F-6.Net particulate organic matter concentrations (mg COD/liter) for December 1971, January 1972, and February 1972, when other values were not obtained Net Particulate (mg COD/liter) Station Dec Jan Feb Oscuro

PAGE 252

234 in

PAGE 253

235 4-' C o \D Q O U to a. rH O CTi t~^ r^ ^0 <-H tn o 00 o CTi LO CO O .— I >d(N rt 00 o I— I t^ 00 1— I >— ' 00 LO VD 00 O 00 \o 'gr^ r-CTi o o ^ o t^ .— I 00 \o o o cr> (Nj -^ LO r^
PAGE 254

236 +-> c o E \0 Q O u 5-1 a. M O Eo 00 CTi O vD Irt ^ >— 1 cr> r— I so CTi •^

PAGE 255

237 T3 O 3 C • H +-> c o a o u OS +J 0) Cl, P o o o u 6 P Cl, P — I LO r^j 00 vo O "^ O O ^ CN CTl o o o o o -^ o vD Cl CTi ^O o r-1 »* CN CM r-~ r-~ o CM r-~ »^ -tihO I— ( CO oi i-l t-o LO CO r-r^ O t^ ^ 00 C^ CTl C7^ t^ r^ o t-o r^ Tj1— I K) CO r~LO r^ i-H \D CTi to 00 — I vo o o t— I r^ r^ vo t^ o CO o cN CM CO r( LO vD t^ vO CM •=a,_^

PAGE 256

258 •p

PAGE 257

239 Table i-1 Results of respiration experiments (BOD) in which duplicate samples were sacrificed after one, three, and seven days of incubation. Values represent oxygen concentrations Cmg 02/liter) and values in parentheses are the rates (mg 02/liter day) between sampling dates^ Collection Date

PAGE 258

240 Table H.continued Collection Date

PAGE 259

241 Table H.continued Collection Date Initial Day 1 Day 5 Day 7 5-Day BOD Station B (Lake) continued 8 Oct 72 7.36 (0.08) 7.28 (0.11) 7.06 (0.05) 6.86 (0.08) (surf) 8 Oct 72 6.88 (0.17) 6.71 (0.10) 6.51 (0.07) 6.25 (0.10) (15 m) Station C (Lake) 19 May 72 7.56 (0.12) (surf) 5 Oct 72 6.53 (0.15) 6.59 (0.07) 6.26 (0.06) 6.00 (0.08) San Felipe (Lake) 3

PAGE 260

242 t/) T3

PAGE 261

243 00

PAGE 262

LITERATURE CITED American Public Health Association. 1965. Standard Methods for the Examination of Water and Wastewater, 12th ed. APHA. 769 p. Anderson, G.C. 1958. Seasonal characteristics of two saline lakes in Washington. Limnol. Oceanogr. 3:51-68. Barlow, J. P., C.J. Lorenzen, and R.T. Myren. 1963. Eutrophication of a tidal estuary. Limnol. Oceanogr. 3:251-262. Beeton, A.M. 1957. Relationship between Secchi disk readings and light penetration in Lake Huron. Trans. Amer. Fish. Soc. 87:73-79. Bormann, F.H., G.E. Likens, and J.S. Eaton. 1969. Biotic regulation of particulate and solution losses from a forest ecosystem. Bio-Science 19:600-610. Bre7onik, P L. , W.M. Morgan, E.E. Shannon, and H.D. Putnajr;. 1969. Eutrophication factors in north central Florida lakes. Engineering and Industrial Experiment Station. Bull. Ser. No. 134. 101 p. Brinson, L.G. 1973. A comparison of seasonal plankton abundance and primary productivity in five tropical fresh-water deltaic embayments. M.S. thesis. University of Florida, Gainesville, Fla. Brooks, H.K. 1969. A preliminary report to the Organization for Tropical Studies, Inc. on Lake Izabal, Geology and Hydrology. 19 p. Carr, A.F., III. 1971. The commercial snook ( Centropomus undecimalis) fishery of Lake Izabal, Guatemala. M.S. thesis. University of Florida, Gainesville, Fla. 105 p. Carter, W.E. 1969. New Lands and Old Traditions: Kekclii Cultivators in the Guatemalan Lowlands. Latin American Monographs: No. 6. University of Florida Press, Gainesville. 155 p. Clarke, G.L. 1941. Observations on the transparency in the SouthWestern section of the North Atlantic Ocean. J. Mar, Res, 4:221-230. 744

PAGE 263

245 Col" G A 1966. The yVT.erican Southwest and Middle America, p. 393^' ' ' 434. In: D.G. Frey (Ed.), Limnology in North America. Univ. Wisconsin Press, Madison. Cooper, D.C. and B.J. Copeland. 1973. Responses of continuousseries estuarine microecosystems to point-source input variations. Ecol. Monogr. 43:213-236. Darnell, R.M. 1961. Trophic spectrum of an estuarine community, based on studies of Lake Pontchartrain, Louisiana. Ecology 42:553-568. 1967 Organic detritus in relation to the estuarine Ecosystem, p. 376-382. In: G. Lauff (Ed.), Estuaries. Am. Ass. Advmt. Sci. Publ. No. 83. Debski, K. 1966. Continental hydrology. Vol. II. Physics of _ Water, Atmospheric Precipitation and Evaporation. Scientific Publ. Foreign Coop. Center. 606 p. Deevey, E.S. 1941. Limnological studies in Connecticut. VI. The quantity and composition of the bottom fauna in thirtysix Connecticut and New York lakes. Ecol. Monogr. 11: 413-455. 1957. Limnological studies in Middle America, with "a chapter on Aztec limnology. Trans. Conn. Acad. Arts Sci. 39:213-328. Dengo, G., and 0. Bohnenberger . 1969. Structural development of northern Central America. Bull. Amer. Assoc. Petrol. Geol., Mem. 11:203-220. Dickinson, J.C, III. (in press) Fisheries of Lake Izabal, Guatemala. Dorris T.C. 1972. La ecologia y la pesca del Lago de Atitlan. Publicacion Especial No. 1. El Centro de Investigacioncs de Embalses de La Universidad Estatal de Oklahoma. 34 p. Dunn, I.e., J.J. Burgis, G.G. Ganf, L.M. McGowan, and A.B. Viner. 1969. Lake George, Uganda: A limnological survey. Verb. Internat. Verein. Limnol. 17:284-288. Edmondson, W.T. (Ed.). 1959. H.B. Ward and G.C. Whipple Freshwater Biology, 2nd ed. John Wiley & Sons, New \ork and London. 1248 p. Edwards, R.W. , and H.L.J. Rollcy. 1965. Oxygen consumption of river muds. J. Ecol. 53:1-19.

PAGE 264

246 Forbes, S.A. 1887. The lake as a microcosm. Reprinted 1925. Bull. 111. Nat. Surv. 15:537-550. Frcy, D.G. 1959. A limnolcgical reconnaissance of Lake Lanao. Int. Ver. Theor, Angew. Limnol. Verh. 17:1090-1102. Gibbs, R.J. 1967. The geochemistry of the Amazon river system: Part I. The factors that control the salinity and the composition and concentration of the suspended solids. Geol. Soc. Amer. Bull. 78:1203-1232. Gilbert, C.R., and D.P. Kelso. 1971. Fishes of the Tortuguero Area, Caribbean Costa Rica. Bull. Florida State Mus., Biol. Sci. 16:1-54. Golley, F.B., J.T. McGinnis, R.G. Clements, G.I. Child, and M.J. Duever. 1971. Mineral Cycling in a Tropical Moist Forest Ecosystem. Athens, Georgia. (1971). Golterman, H.L. (Ed.). 1969. Methods for Chemical Analysis of Fresh Waters. IBP Handbook No. 8. Blackwell Scientific Publications, Oxford. 166 p. Grove, A.T. 1972. The dissolved and solid load carried by some West African Rivers: Senegal, Niger, Benue, and Shari. J. Hydrcl, Harbeck, G.E., G.E. Koberg, and G.H. Hughes. 1959. The effect of the addition of heat on the thermal structure and evaporation of Lake Colorado City, Texas. U.S. Geol. Survey Prof. Paper 272-B. Hayes, F.R., and M.A. MacAulay. 1959. Lake water and sediment. V. Oxygen consumed in water over sediment cores. Limnol. Oceanogr. 4:291-298. Heald, E.J. 1971. The production of organic detritus in a south Florida estuary. Sea Grant Tech. Bull. No. 6. Univ. of Miami. 110 p. Hickling, C.F. 1961. Tropical Inland Fisheries. John Wiley § Sons, New York. 287 p. Holdridge, L.R. 1967. Life Zone Ecology. Tropical Science Center, San Jose, Costa Rica. 206 p. Holloway, A.D. 1948. Recommendations for the Development of the Fisliery Resources of Guatemala. U.S. Fish and Wildlife Service, Washington. 114 p.

PAGE 265

247 Hutchinson, G.E. 1957. A Treatise on Limnology. Vol. 1. Geography, Physics, and Chemistry. John Wiley 5 Sons, Inc. New York. 1015 p. . 1967. A Treatise on Limnology. Vol. 2. Introduction to lake l>'iology and the limnoplankton. John Wiley 5 Sons, Inc. New York. 1115 p. Ichimura, S. 1956. On the ecological meaning of transparency for the production of matter in phytoplankton community of lake. Bot. Mag., Tokyo 69:219-226. Institute Geografico Nacional, 1966. Atlas preliminar de Guatemala, Tercera edicion. Litografia, Guatemala. 25 p. Kaushik, N.K., and H.B.N. Hynes. 1968. Experimental study on the role of autumn-shed leaves in aquatic environments. J. Ecol. 56:229-243. Lawacz, W. 1969. The characteristics of sinking materials and the formation of bottom deposits in a eutrophic lake, p. 319331. In: D.G. Frey (Ed.), Symposium on paleolimnology . Int. Assoc. Limnol., Commun. no. 17. 448 p. Lefevre, M. , H. Jacob, and M. Nisbet. 1952. Autoet heteroantagonisme chez les algues d'eau douce. Annls St. Cant. Hydrobiol. appl. 4:5-197. Lewis, W. 1975. Factors controlling primary production in a tropical lake. Paper presented to: II International Symposium on Tropical Ecology, February 8, 1973, at Caracas, Venezuela. Likens, G.E. 1972. Eutrophication and aquatic ecosystems, p. 3-13. In: G.E. Likens (Ed.), Nutrients and Eutrophication: The limiting-nutrient controversy. Amer. Soc. Limnol. Oceanogr. Special Symposia, Vol. I. 328 p. , F.H. Bormann, N.M. Johnson, D.W. Fisher, and R.S. Pierce. 1970. Effects of forest cutting and herbicide treatment on nutrient budgets in the Hubbard Brook Watershed-Ecosystem. Ecol. Monogr. 40:23-47. Lund, J.W.G. 1950. Studies on Asterionella formosa Mass. II. Nutrient depletion and the spring maximum. J. Ecol. 58:1-35. Maciolek, J. A. 1962. Lininological organic analysis by quantitative dichromate oxidation. Res. Report No. 60 of Bur. Sport. Fish, and Wildlife, U.S. Fish and Wildlife Service. 61 p. Margalef, R. 1963. On certain unifying principles in ecology. Amer. Nat. 97:357-574.

PAGE 266

248 McConnell, W.J. 1963. Primary productivity and fish harvest in a small desert impoundment. Trans. Amer. Fish. Soc. 92:1-12. McDonnell, A.J., and S.D. Hall. 1969. Effect of environmental factors on benthal oxygen uptake. Jour. WPCF. Res. Supp. 41:353-363. Nelson, D.J., and D.C. Scott. 1962. Role of detritus in the productivity of a rock-outcrop community in a Piedmont stream. Limnol. Oceanogr. 7:396-413. Nordlie, F.G. 1970. Lake Izabal -Energetics. Final report to the Organization for Tropical Studies, Inc. 22 p. Odum, E.P. 1971. Fundamentals of Ecology, 3rd ed. W.B. Saunders Company, Philadelphia. 574 p. , and A. A. de la Cruz. 1967. Particulate organic detritus in a Georgia salt marsh-estuarine ecosystem, p. 383-388. In: G. Lauff (Ed.), Estuaries. Am. Ass. Advmt. Sci. Publ. No. 83. Odum, H.T. 1956. Primary production in flowing waters. Limnol. Oceanogr. 1:102-117. . 1967. Biological circuits and the marine systems of Texas, p. 99-158. In: T.A. Olson and F.J. Burgess (Eds.), Pollution and Marine Ecology. Interscience Publ., New York. . 1971. Environment, Power, and Society. Wiley-Interscience, New York. 331 p. Odum, W.E. 1971. Pathways of energy flow in a south Florida estuary. Sea Grant Tech. Bull. No. 7. Univ. of Miami. 162 p. Pennak, R.W. 1953. Freshwater Invertebrates of the United States. Ronald Press Co., Nev\f York. 769 p. Pescod, M.B. 1969. Photosynthetic oxygen production in a polluted tropical estuary. Jour. WPCF. Res. Supp. 41:309-320. Poole, H.H. , and W.R.G. Atkins. 1929. Photoelectric measurements of submarine illumination throughout the year. J. Mar. Biol. Ass. U.K. 16:297-324. Popenoe, li.L. 1960. Effects of shifting cultivation on natural soil constituents in Central America. Ph.D. dissertation, University of Florida, Gainesville. 158 p. Pratt, D.M., and U. Berkson. 1959. Two sources of error in the oxygen light and dark bottle method. Limnol. Oceanogr. 4:328-334. Pride, R.W., F.W. Meyer, and R.M. Cherry. 1966. Hydrology of Green Swamp area in central Florida. Florida Gcol. Survey Rept. Inv. 42. 106 ]).

PAGE 267

249 Ray, C.L. 1931. Evaporation in the Eastern Caribbean. Mon. Weather Rev. 59:192-193. Reiser, CO. 1969. Analysis of evaporation control system of the Sea of Galilee. Water Resour. Res. 5:380-394. Remane, A., and C. Schlieper. 1971, Biology of Brackish Water. John Wiley q Sons, New York. 372 p. Roberts, R.J., and E.M. Irving. 1957. Mineral deposits of Central America. U.S. Geol. Survey Bull. 103'i. 205 p. Rodhe, W. 1969. Crystallization of eutrophication concepts in northern Europe, p. 50-64. In: Eutrophication: Causes, consequences and correctives. Nat. Acad. Sci./Nat. Res. Council, Publ. 1700. Ruttner, F. 1952. Planktonstudien der Deutschen Limnologischen. Sundan-Expedition. Arch. Hydrobiol. Suppl. 21 : 1-274. . 1963. Fundamentals of Limnology. 3rd ed. Translated by D.G. Frey and F.E.J. Fry. Univ. Toronto Press. 295 p. Seki, H., J. Skelding, and T.R. Parson. 1968. Observations on the decomposition of a marine sediment. Limnol. Oceanogr. 15.440-447. Simmons, C.S., J.M. Tarano T., and J.H. Pinto Z. 1959. Clasificacion de Reconocimiento de los Suelos de la Republica de Guatemala. Editorial del Ministerio de Educacion Publica, Guatemala. 1,000 p. Smith, G.M. 1950. The Fresh-water Algae of the United States, 2nd ed. McGraw-Hill Book Co., New York. 719 p. Snedaker, S.C. 1970. Ecological studies on tropical moist forest succession in eastern lowland Guatemala. Ph.D. dissertation. University of Florida, Gainesville. 133 p. Szcicz, G. , and G. Endrodi. 1969. Aerodynamic and surface factors in evaporation. IVater Resour. Res. 5:380-394. Tailing, J.F. 1965. Comparative problems of phytoplankton production and photosyntlietic productivity in a tropical and temperate lake, p. 401-424. In: C.R. Goldman (Ed.), Primary Productivity in Aquatic Environments. Mem. 1st. Ital. Idrobiol., IS Suppl., University of California Press, Berkeley. . 1966. The annual cycle of stratification and phytoplankton growth in Lake Victoria (East Africa). Int. Revue ges. Hydrobiol. 51:545-621.

PAGE 268

250 1969. The incidence of vertical mixing, and some biological and chemical consequences, in tropical African lakes. Verh. Internat. Verein. Limnol. 17:998-1012. Teal, J.M. 1962. Energy flow in the salt marsh ecosystem of Georgia. Ecology 45:614-624. Tergas, L.E. 1965. Correlation of nutrient availability in soil and uptake by native vegetation in the humid tropics. M.S. thesis, University of Florida, Gainesville. 64 p. Thorson, T.B., CM. Cowan, and D.E. Watson. 1966. Sharks and sawfish in the Lake Izabal--Rio Dulce System, Guatemala. Copeia 3: 620-622. Tsukada, M. , and E.S. Deevey. 1967. Pollen analysis from four lakes in southeastern Maya area of Guatemala and El Salvador, p. 303-331. In: E.J. Gushing and H.E. Wright (Eds.), Quaternary Paleoecology. Yale University Press, New Haven, Turner, J.F., Jr. 1966. Evaporation study in a humid region. Lake Michie, North Carolina. U.S. Geol. Survey. Prof. Paper 272-G. Vollenweider , R.A. fEd.). 1969. Primary Production in Aquatic Environm.ents. IBP Handbook No. 12. Rlackwell Scientific Publications, Oxford. 213 p. Voorhies, B. 1969. A prehistoric Maya hinterland: the Izabal zone of eastern Guatemala. Paper presented to: Amer. Anthropol. Ass., November 20, 1969, at New Orleans, La. Walper, J.L. 1960. Geology of Coban-Purulha area, Alta Verapaz, Guatemala. Bull. Amer. Assoc. Petrol. Geol. 44:1273-1513. Walter, H. and H. Leith. 1960-67. Klima diagram -welt atlas. V.B. Gustav Fisher Vcrlag; Jena. Welch, P.S. 1948. Limnological Methods. Blakiston Co., Philadelphia. 581 p.

PAGE 269

BIOGRAPHICAL SKETCH Mark McClellan Brinson was born October 6, 1943, in Shelby, Ohio. He lived there with his parents. Glen and Geneva, until graduation from Shelby High School in 1961. He graduated with the degree Bachelor of Science in biology at Heidelberg College in June, 1965, where he was awarded membership to Who's Who in American Colleges and Universities and Tower Men, an academic honorary. In August, 1965, he entered graduate studies at The University of Michigan where he received the Newcombe Fellowship in botany. Upon receipt of a Master of Science in Botany in May, 1967, he v;ent to Central America as a Peace Corps Volunteer. There he worked as a fisheries biologist for two years in Turrialba, Costa Rica. He entered graduate school at the University of Florida in September, 1969, where he is presently a candidate for the degree Doctor of Philosophy While there he was awarded a Research Fellowship from the Program for Latin America and the Caribbean sector of the Foreign Area Fellowship Program. He is a member of the American Society of Limnology and Oceanography, Phi Sigma Society, The Ecological Society of America, and the American Institute of Biological Sciences. 251

PAGE 270

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. J:^4^ Ariel E. Lugo, Chairman Assistant Professor of Botany I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. <=^. I J.^ .a-iDavid S. Anthony Professor of Botany I certify that I have read tliis study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for "lie degree of Doctor of Phiiojoptiy ma G. Griffin, 1^1/ associate Professor of Botany I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Frank G. Nordlie Associate Professor of Zoology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Pliilosoph}-. Hugh L. Por Professor iioe Soils

PAGE 271

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ?r ^f.<^ Z:^ U\ Leland Shanor Professor of Botany This dissertation was submitted to the Dean of the College of Agriculture and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December, 1975 OaJIvs n. College of Agriculture Dean, Graduate School