Title: Modeling the response of mangrove ecosystems to herbicide spraying, hurricanes, nutrient enrichment and economic development /
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00098864/00001
 Material Information
Title: Modeling the response of mangrove ecosystems to herbicide spraying, hurricanes, nutrient enrichment and economic development /
Physical Description: xx, 390 leaves : diagrs. ; 28 cm.
Language: English
Creator: Sell, Maurice George, 1941-
Publication Date: 1977
Copyright Date: 1977
 Subjects
Subject: Mangrove swamps -- Florida   ( lcsh )
Coastal zone management -- Florida   ( lcsh )
Environmental Engineering Sciences thesis Ph. D
Dissertations, Academic -- Environmental Engineering Sciences -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 382-389.
Statement of Responsibility: by Maurice George Sell, Jr.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00098864
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000187484
oclc - 03406984
notis - AAV4084

Downloads

This item has the following downloads:

PDF ( 14 MBs ) ( PDF )


Full Text
aw-'


MODELING THE RESPONSE OF MANGROVE ECOSYSTEMS
TO HERBICIDE SPRAYING, HURRICANES, NUTRIENT
ENRICHMENT AND ECONOMIC DEVELOPMENT










By

MAURICE GEORGE SELL, JR.


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






UNIVERSITY OF FLORIDA
























"Let everyone who is concerned with interference of naturally
occurring mangrove formations on-exposed coasts take heed."

Fosberg (1971)















ACKNOWLEDGMENTS


I would like to thank Dr. Howard T. Odum, chairman of my

research committee, for his guidance in bringing this dissertation

to completion. Many helpful suggestions were also received from my

committee: Dr. Suzanne Bailey, Dr. Jackson L. Fox, Dr. Samuel C.

Snedaker, and Dr. Ariel E. Lugo.

Tim Ahlstrom, Joan Browder, Bruce Heinly, Sanit Aksornkoe,

and Mark Homer assisted with field measurements in an environment

where they had to endure the relentless pursuit of sand flies and

mosquitoes. Dr. Howard Teas and Jerry Kelly of the University of

Miami initiated the herbicide experiments on Marco Island.

My wife, Jackie, also assisted in field measurements and

exhibited great patience with me during the writing of this disserta-

tion.

The research on the effect of herbicides on mangroves in South

Vietnam was supported through a contract between the National Academy

of Sciences and the systems ecology group of the Department of

Environmental Engineering Sciences, H.T. Odum, principal investigator.

Other support was provided by the South Florida study of the United

States Department of Interior on contract with the Center for

Wetlands, University of Florida.
















TABLE OF CONTENTS


ACKNOWLEDGMENTS . . ...


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

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

ABSTRACT .. ... .. . . . . .........


Page

. . . iii


xiii

xviii


INTRODUCTION . . . . .
Study Plan and Objectives.
Description of Study Areas
South Vietnam Study
Florida Study Areas


Area.....


REVIEW OF LITERATURE .. . .......
Ecosystem Modeling . . . . .
History of Herbicide Use in Mangroves.
Sewage Disposal in Wetlands .. ..
Effects of Hurricanes .. . ....
Occurrence and Frequency of Hurricanes
Economic Development in Mangroves. .


and Typhoons.


METHODS .. ....... ..........................
Modeling and Simulation Techniques .. . ......
Development of Models and Equations .. ....
Aggregation . . . . . . . . .
Calculation of Rate Coefficients .. . ...
Scaling of Equations .. . . .......
Energy Calculations . . . . ... . . .
Energy Quality . . . . .. . . .
Energy Investment Ratio .. . .......
Fie'd Studies . . . . . . . . . . .
Tree Growth . . . . . . . . .
Litter Fall .. .................
Total Phosphorus in Water Column .. . ...
Phosphorus in Leaves and Wood ....
Numbers of Green Leaves, Yellow Leaves, Live
Seedlings, Dead Seedlings, Snails, and
Crabholes .................
Number of R. mangle Seedlings on Trees ....









TABLE OF CONTENTS (continued)


Page
RESULTS . . . . . . ....... . . .. 48
Evaluation and Simulation of Models. . . . ... ... 48
Model of Herbicides and Mangroves in South
Vietnam . . . . .. .... .. . .48
Model of Mangroves, Nutrients, and Hurricanes .. 75
Model of Mangroves and Economic Development . .. 110
Energy Calculation Results .. . .. .... ... . 135
Calculation of Energy Flows .. . .. . .. 135
Calculation of Investment Ratios. . . . . ... 141
Field Study Results. . . . ... .. . . . 141
Sites Treated with Herbicide. . . . .. 142
Sites Enriched with Sewage Effluent .... . . 155

DISCUSSION . . . . ... . . . . .. . 198
Herbicide and Recolonization of Mangroves in South Vietnam 198
Herbicide and Mangrove Succession in Florida . . ... 209
Hurricanes and Mangrove Structure . . .. . . . 214
Undeveloped Area of Mangroves .. . ..... 214
Developed Area. . ... .. . .. . . . 216
Economic Development and Mangroves . . .. .. . 218
Role of Nutrients in Mangroves .. . ... . .. 225
Role of Mangrove Litter Fall . . . . . .. 236
Role of Mangrove Detritus. . ... . . .. . . 236
Energy Quality, Herbicide, Hurricane, and Nutrient Enrich-
ment . . . . . . . . . . .. 240
Mangroves and the Balance of Productive Potentials and
Stresses . .. . .. . . . . . 241
Summarizing Guidelines for Management of Mangroves .... .242

APPENDICES

A ENERGY LANGUAGE SYMBOLS USED IN DEVELOPMENT OF MODELS. . 246

B PROCEDURE TO ANALYZE FOR CONCENTRATION OF TOTAL PHOSPHORUS
IN WATER SAMPLES .. . . .. . . . . . 250

C SUPPLEMENTARY DATA FOR SIMULATION OF MODEL OF HERBICIDES
AND MANGROVES SHOWN IN FIGURE 4. .. . ....... .253

D SUPPLEMENTARY DATA FOR SIMULATION OF MODEL OF HURRICANES
AND MANGROVES SHOWN IN FIGURE 15 .. . . ... ... 274











TABLE OF CONTENTS (continued)

APPENDICES Page


E SUPPLEMENTARY DATA USED IN SIMULATION OF THE MODEL OF
ECONOMIC DEVELOPMENT AND MANGROVES SHOWN IN FIGURE 28. . 291

F DETAILED DATA COLLECTED AT THE SPRAYED PLOTS ON MARCO
ISLAND, FLORIDA .... .... ..... . . . 311

G DETAILED DATA COLLECTED ON RESPONSE OF MANGROVE FOREST
TO NUTRIENT ENRICHMENT .... . .... .... ... 333

BIBLIOGRAPHY .. . . . .. .. .. .. ... .. .. 382

BIOGRAPHICAL SKETCH. . . .. . . . . . ... . .390
















LIST OF TABLES


Table Page

1 Energy quality factors of several ecosystem work
processes. . . . . . . . . . .... . 42

2 Initial and steady state simulated woodcutting rates
for the mangrove forest of the Rung-Sat district in
South Vietnam at two rates of primary production . . 62

3 Energy values for hypothetical cases of steady state
residential development in a mangrove forest consisting
of 1000 hectares . . . . . . .. .. . . 136

4 Density of live and dead trees in the herbicide study
sites on Marco Island, Florida, at the time of spraying
and 2 years after spraying ..... . . . .. 144

5 Mean tree diameter at breast height of each species
of mangrove in the herbicide study sites at Marco
Island, Florida. .. .. . . ... . . ... 145

6 Mean basal area per tree of each species of mangrove
in the herbicide study sites at Marco Island, Florida. 148

7 Density of trees in the sites receiving sewage effluent
and at the control sites . .. ... ..... .. 156

8 Density of Rhizophora mangle seedlings on parent trees
at the sites receiving sewage effluent and at the control
sites. .. .... .. .. ... . . . . . 156

9 Mean tree diameter at breast height of each species of
mangrove in the sites receiving sewage effluent and
in the control sites . . . . . . . . 158

10 Mean tree diameter at breast height of each species of
mangrove in the sites receiving sewage effluent and in
the control sites . . . . . . . . . . 161











LIST OF TABLES (continued)


Table Page

11 Mean basal area per tree of each species of mangrove in
the sites receiving sewage effluent and in the control
sites. ......... ... ... .. . ... 163

12 Mean basal area per tree of each species of mangrove in
the sites receiving sewage effluent and in the control
sites. ... ............. ... . .. 165

13 Mean tree diameter at breast height, growth, and percent
growth for September 1973 and 1974 in the mangrove forest
receiving sewage effluent at Naples, Florida . ... 167

14 Mean tree diameter at breast height, growth, and percent
growth for September 1973 and 1974 in the control forest
at Naples, Florida ....... ......... 169

15 Mean tree diameters at breast height, growth, and percent
growth for September 1973 and 1974 in the mangrove forest
receiving sewage effluent at Everglades City, Florida.. 171

16 Mean tree diameter at breast height, growth, and percent
growth for September 1973 and 1974 in the control forest
at Everglades City, Florida. . ..... . . . 173

17 Mean tree basal area, growth, and percent growth for
September 1973 and 1974 in the mangrove forest receiving
sewage effluent at Naples, Florida . . .. .. 176

18 Mean tree basal area, growth, and percent growth for
September 1973 and 1974 in the control forest at
Naples, Florida . . ... . . . . 178

19 Mean tree basal area, growth, and percent growth for
September 1973 and 1974 in the mangrove forest receiving
sewage effluent at Everglades City, Florida. ......180

20 Mean tree basal area, growth, and percent growth for
September 1973 and 1974 in the control forest at
Everglades City, Florida . . . . . .... 182

21 Concentration of phosphorus in mangrove wood and leaves
at the Naples and Everglades City study sites. ..... 197

22 Time in years required for recolonization of land at
five rates of spraying and three rates of mangrove
primary productivity . . . . . . . . 202









LIST OF TABLES (continued)


Table Page

23 Estimated biomass and growth of trees in the mangrove for-
ests at the Naples and Everglades City study sites for
September 1973, September 1974, and September 1976. .... .228

24 Comparison of estimated wood growth rates and biomass with
those predicted by the model of Figure 15 and those found
in the literature ... .. .. .... .. .. . 231

25 Rate of uptake of phosphorus in wood for the Naples and
Everglades City study sites . ... .. .. ........ 234

26 Comparison of measured annual litter fall at the Naples and
Everglades City study sites with rates found in the
literature. . . . .. . ... . . . . .. 237

27 Comparison of detritus amounts simulated by the model in
Figure 15 with amounts reported in the literature .. .. 239

C-l Monthly averages of solar radiation data for Saigon from
January 1964 to October 1967 . ... ..... . ... .254

C-2 Descriptions and values for the outside driving forces, state
variables, and pathways of the herbicide and mangrove model 255

C-3 Data on density and weight of Rhizophora mangle seedlings at
Rookery Bay, Florida. . . . . . . . . 263

C-4 Calculations of rate coefficients for the model
simulating the effect of herbicide on the Rung-Sat mangroves
in South Vietnam . . . . . . . . .... 264

C-5 Scaled differential equations for the model simulating the
effect of herbicide on the Rung-Sat mangroves in South
Vietnam . . . . . . . .. . . . . 268

D-1 Descriptions and values for the driving forces, state vari-
ables, and pathways of the model simulating the relationship
among hurricanes, nutrients, and mangroves. . . ... .275

D-2 Calculation of rate coefficients for the model simulating the
relationship among hurricanes, nutrients, and mangroves . 281

D-3 Scaled differential equations for the model simulating the
relationship among hurricanes, nutrients, and mangroves . 286

E-l Descriptions and values for the driving forces, state vari-
ables, and pathways of the model simulating the relationship
between economic development and mangroves. .. . ... .... .292










LIST OF TABLES (continued)

Table Page

E-2 Calculation of rate coefficients for the model
simulating the relationship between economic develop-
ment and mangroves ..... . ... .. . . 300

E-3 Scaled differential equations for the model simulating
the relationship between economic development and
mangroves. .. .. .... .. ... .. . . . 306

F-l Number of fallen green leaves per m2 occurring in the
control, cleared, and sprayed areas near Marco Island,
Florida, over a period of 26 months after aerial appli-
cation of herbicide in December 1972 . . . . 313

F-2 Number of fallen yellow leaves per m2 occurring in the
control, cleared, and sprayed areas near Marco Island,
Florida, over a period of 26 months after aerial appli-
cation of herbicide in December 1972 . ....... . 314

F-3 Number of dead seedlings per m2 occurring in the control,
cleared, and sprayed areas near Marco Island, Florida,
over a period of 26 months after aerial application of
herbicide in December 1972 ... . . . . . 315

F-4 Number of live seedlings per m2 occurring in the herbi-
cide study sites near Marco Island, Florida, over a
period of 26 months after aerial application of herbi-
cide in December 1972 . ...... . ...... 317

F-5 Number of Melampus coffeus per m2 occurring in the
herbicide study sites near Marco Island, Florida, over
a period of 26 months after aerial application of
herbicide in December 1972 ... .. .. . .... 319

F-6 Number of crabholes per m2 occurring in the herbicide
study sites near Marco Island, Florida, over a period
of 26 months after aerial application of herbicide
in December 1972 .... . ..... ......... 320

G-1 Evaluation of species and populations of mangroves in
two sites receiving sewage effluent and also two sites
not receiving sewage effluent. ... ........ 336

G-2 Population density of Rhizophora mangle seedlings on
the parent tree at the two sites receiving sewage
effluent and the two control sites .... . .... 338










LIST OF TABLES (continued)

Table Page

G-3 Diameter, height, basal area and volume of individual
trees in the mangrove forest receiving sewage effluent
from the Naples, Florida, sewage treatment plant . 340

G-4 Diameter, height, basal area, and volume of individual
trees in the control mangrove forest at Naples,
Florida ...... .. .. . . ....... 350

G-5 Diameter, height, basal area, and volume of individual
trees in the mangrove forest receiving sewage effluent
from the Everglades City, Florida, sewage treatment
plant. . .. . .. . . . . . . . 352

G-6 Diameter, height, basal area, and volume of individual
trees in the control mangrove forest at Everglades City,
Florida . . . . . .... . . . . 361

G-7 Leaf, wood, and seed fall during a 55-week period from
September 6, 1973,to September 27, 1974, in the mangrove
forest that occasionally received sewage effluent from
the Naples, Florida, sewage treatment plant; and from
June 21 to September 26, 1974, in the control mangrove
forest . ... ...... .. . . . .. . 366

G-8 Leaf, wood, and seed fall during a 55-week period from
September 4, 1973, to September 27, 1974, in the mangrove
forest that occasionally received sewage effluent from
the Everglades City, Florida, sewage treatment plant . 368

G-9 Rate of leaf, wood, seed, and total litter fall during a
55-week period from September 6, 1973, to September 27,
1974, in the mangrove forest that occasionally received
sewage effluent from the Naples, Florida, sewage treat-
ment plant; and from June 21 to September 27 in the
control mangrove forest. ... . . . . . ... 370

G-10 Rate of leaf, wood, seed, and total litter fall during a
55-week period from September 4, 1973, to September 27,
1974, in the mangrove forest that occasionally received
sewage effluent from the Everglades City, Florida,
sewage treatment plant . . . . . . . . 372










LIST OF TABLES (continued)


Table Pa e

G-11 Concentrations of total phosphorus found in the Gordon
River at Naples, Florida, at various sampling times
from September 1973 to February 1975 . . . 374

G-12 Concentration of total phosphorus in the tidal canal
at Everglades City, Florida, at various sampling times
from September 1973 to February 1975 . . ... 376

G-13 Calculation of mangrove forest biomass for the Naples
and Everglades City, Florida, study sites .. . . 378
















LIST OF FIGURES


igCure Page

1 Model of the mangrove ecosystem . . . ... ... 4

2 Map of the Rung-Sat district in South Vietnam showing
sprayed and unsprayed areas . ... .. .... . 10

3 Map of southwest Florida showing the sites treated
with sewage, the site treated with herbicide, and the
sites where mangrove productivity and biomass were
measured by Lugo and Snedaker (1974a) . .... 14

4 Model of the mangrove ecosystem of the Rung-Sat
district of South Vietnam showing the interaction of
herbicide, land, mangrove biomass, woodcutters, and
mangrove seedlings. . . . . . . . . 50

5 Model of the mangrove ecosystem of the Rung-Sat
district in South Vietnam with numerical values for
the state variables, outside driving forces, and
pathways. .. ... .. . . . . . . .... 54

6 Variation of simulated herbicide application rates
during a five-year period of spraying ...... .. 58

7 Steady state levels of mangrove biomass attained by the
Rung-Sat mangrove forest in South Vietnam at four
simulated rates of primary production .. . .. 58

8 Simulated effect of woodcutting on the mangrove forest
of the Rung-Sat district in South Vietnam .. . 61

9 Simulated effect of herbicide spraying on mangrove
land and mangrove biomass in the Rung-Sat district in
South Vietnam . . . . . . . . . . 65

10 Simulated effect of herbicide spraying on mangrove land
ard mangrove biomass in the Rung-Sat district in
South Vietnam . . . . . . . . . . 67










LIST OF FIGURES (continued)


Figure Page

11 Simulated effect of herbicide spraying on mangrove
land and mangrove biomass in the Rung-Sat district
in South Vietnam. . . . . . . . . . 69

12 Simulated effect of herbicide spraying on mangrove land
and mangrove biomass in the Rung-Sat district in South
Vietnam . . . . . . . . . . . 72

13 Simulated effect of seedling availability on the rate
of recolonization by the mangroves of the Rung-Sat
district in South Vietnam . . . . . . . 74

14 Simulated effect of seedling planting by man on the rate
of recolonization by the mangroves of the Rung-Sat
district in South Vietnam . . . . . . . 77

15 Simplified model of a mangrove forest in Florida show-
ing the interactions between the forest and nutrients,
hurricanes, and tides . . . . . . . 79

16 Relationship between tangential wind velocity of a
hurricane and the surface shear stress that results . 83

17 Simplified model of a mangrove forest in Florida showing
the values for the driving forces, compartments,and
pathways . . . . . . . . . . . 85

18 Simulation of the effect of altering the nutrient flux
into the mangrove forest when nutrient storage was
initially high and runoff to sea was low .. .... 88

19 Simulation of the effect of altering the nutrient flux
into the mangrove forest when nutrient storage was
initially low and runoff to sea was high .. .... 90

20 Simulation of effect of tidal exchange between the
estuary and the mangrove forest when nutrient storage
was initially high . . . . . . . . .

21 Simulation of the effect of tidal exchange between the
estuary and mangrove forest when nutrient storage was
initially low . . . . . . . . . . 95










LIST OF FIGURES (continued)


Figure Page

22 Simulation of effect of hurricane on the mangrove
forest when nutrient storage was initially high . 97

23 Simulation of effect of hurricane on the mangrove
when nutrient storage was initially high. ....... 99

24 Simulation of the effect of hurricanes on the
mangrove forest when nutrient storage was initially
high. . . . . . . . . . . . . 101

25 Simulation of effect of hurricanes on the mangrove
forest when nutrient storage was initially low. ... 104

26 Simulation of effect of hurricane on the mangrove
forest when nutrient storage was initially low. ... 106

27 Simulation of the effect of hurricanes on the mangrove
forest when nutrient storage was initially low. ... 108

28 Model of the interactions between a mangrove forest
and economic development within the forest. . . ... 112

29 Model of the interactions between a mangrove forest
and economic development. .. . . . . 116

30 Simulation of the impact of the amount of land develop-
ment on mangrove forest biomass and economic structure. 119

31 Simulation of the impact of economic development on
mangrove productivity and flow of goods and services.. 121

32 Simulation of the impact of changing the price index on
economic structure. . . . . . . . . 125

33 Simulation of the impact of one hurricane on mangrove
biomass and economic structure. ......... 127

34 Simulation of the impact of one hurricane on mangrove
biomass and economic structure. . ....... ... 129

35 Simulation of the impact of two hurricanes within 10
years on mangrove biomass and economic structure .. 132










LIST OF FIGURES (continued)


Figure Page

36 Simulation of the impact of two hurricanes within 10
years on mangrove biomass and economic structure. . 134

37 Undeveloped and developed steady state energy flows for
a hypothetical development of 1000 hectares of mangrove
forest. . . . . . . . . ...... . 140

38 Variation with time in number of fallen green leaves;
number of fallen yellow leaves; number of dead seed-
lings; number of live seedlings; number of snails and
number of crabholes for the control site, the cleared
site and the average of the three sprayed sites
located in a mangrove forest on Marco Islandi,
Florida . . . . . . . . . . . . 151

39 Litter fall from September 1973 to September 1974 in a
mangrove forest receiving sewage effluent from the sewage
treatment plant in Naples, Florida. .. . .. .. 185

40 Litter fall from September 1973 to September 1974 in a
mangrove forest receiving sewage effluent from the
sewage treatment plant in Everglades City, Florida. 187

41 Map of the study sites at Naples showing the average total
phosphorus concentration in the waters of the Gordon
River at the point of discharge of sewage effluent and
other selected locations along the river. ....... 192

42 Map of the study sites at Everglades City showing the
average total phosphorus concentration in the waters
of a tidal canal at the point of discharge of sewage
effluent and other selected locations along the canal 195

43 Simulation of effect of seedling availability on the
recovery of sprayed mangroves in the Rung-Sat district
and in the Camau peninsula of South Vietnam .. . 206

44 Model of the interaction between herbicide and the
mangrove forest on Marco, Islando, Florida . . 212

45 Variation in steady state economic structure and flow
of purchased energies as the percent of land de-
veloped is increased from 0 to 100 percent ...... 221










LIST OF FIGURES (continued)


Figure Page

46 Variation in steady state economic structure as the
price index is varied from 0 to 1 . .... . . 224

C-1 Annual variation in solar radiation for Saigon, South
Vietnam . . . . . .. . . . . . 271

C-2 Analog circuit diagram for the model simulating the
effect of herbicides on the mangroves of the Rung-Sat
district in South Vietnam .. . . . .. 273

D-1 Analog circuit diagram for simulated model of nutrients,
hurricanes and mangroves .. . . ..... ... 290

E-l Analog circuit diagram of mangrove swamp and economic
development system model ......... . . 310

F-l Number of fallen green leaves per m2 of study area in
the Marco Island herbicide spraying experiment. ... . 322

F-2 Number of fallen yellow leaves per m2 of study area in
the Marco Island herbicide spraying experiment. ... 324

F-3 Number of dead seedlings per m2 of study area in the
Marco Island herbicide spraying experiment . . .. 326

F-4 Number of live seedlings per m2 of study area in the
Marco Island herbicide spraying experiment. . . . 328

F-5 Number of snails per m2 of study area in the Marco
Island herbicide spraying experiment. . . ... ... 330

F-6 Number of crabholes per m2 of study area in the Marco
Island herbicide spraying experiment. .. .. ...... 332










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



MODELING THE RESPONSE OF MANGROVE ECOSYSTEMS
TO HERBICIDE SPRAYING, HURRICANES, NUTRIENT
ENRICHMENT AND ECONOMIC DEVELOPMENT

By

Maurice George Sell, Jr.

June 1977

Major Department: Environmental Engineering Sciences
Chairman: Dr. H. T. Odum


Models of energy flow through mangrove forests were developed

and simulated to assess the impact of tropical storms, herbicides,

nutrient enrichment, and economic development.

Simulations of the effect of herbicide spraying on the mangrove

forests in the Rung-Sat district of South Vietnam suggested that com-

plete mangrove recolonization of sprayed areas may take 55 years to more

than 100 years. Rates of recolonization were strongly dependent on the

availability of seedlings and amount of woodcutting. Reforestation can

be accelerated by planting seedlings.

Spraying of a mangrove forest on Marco Island, Florida, showed

that white mangrove (Laguncularia racemosa) was the most susceptible

of the three species present followed by red mangrove (Rhizophora

mangle) and black mangrove (Avicennia erminans). The sprayed sites

seemed to be recolonized at a slower rate than the harvested sites.

Detritus-feeding snails were reduced in numbers after the spraying but


xviii










crab populations were little changed. Simulations of the effects of

hurricanes on mangrove forests suggested that the growth cycle was

adapted to recovery from major hurricanes that occur every 15-25 years

Rate of recovery was dependent on the concentration of nutrients

available to the mangroves. Even when buildings are not directly

damaged by hurricanes, the economy of a developed coastal area may be

affected by damage to the mangrove forests that acted as a storm

buffer, an input to fishery nurseries, and in other indirect roles of

environmental support. The simulations suggested that destruction of

mangrove vegetation by hurricanes might result in a reduction in the

inflow of purchased energy, goods, and services to the developed area.

Simulations of the relationship between economic development and man-

groves showed a peak of growth returning to a lower steady state. At

1973 prices steady state economic structure was highest when 50%

of the land was developed in the form of condominiums or residential

finger canal estates. Half of 1% of the land could be developed

in the form of condominiums at a density of 44 units/ha or only 1.2%

of the land could be developed in the form of residential finger canal

estates at a density of 10 units/ha.

Simulations of the effect of altering the nutrient flow into

mangrove forests suggested roles For tidal exchange and freshwater run-

off in developing biomass of mangroves. Field studies at Naples and

Everglades City, Florida, indicated mangroves grew faster when bathed

by tidal waters enriched with nutrients from sewage effluent.

Estimated growth rates of wood during the three-year study were










4.6 g/m2 day at the Naples sewage-enriched site, 2.8 g/m2 day at the

Naples control site, 2.8 g/m2 day at the Everglades City sewage-

2
enriched site, and 1.3 g/in2-day at the Everglades City control site.

The management of mangroves for maximum contribution to the

combined economy of nature and man requires maintaining access to nutri-

ents and tidal exchange, abundant seedling supply after a killing

stress, and a high ratio of mangrove land with the kind of economic

developments that are attracted by the energy values of the estuarine

zone.
















INTRODUCTION


This dissertation is concerned with the effects of various

perturbations on the structure of mangrove forests. Studies were

made on the impact of herbicides, hurricanes, nutrient enrichment,

and economic development on mangrove forest. The study included

three models and results of their simulation. Data for models and

for validating simulation results were obtained with selected field

studies.

The mangrove forest is a marine ecosystem in which the trees

have evolved special adaptations enabling them to grow and prosper

in salty or brackish water. These forests are very prevalent along

the coastlines of the world between 25"N and 25S latitude. The geo-

graphical distribution of mangroves is limited to those coastal areas

that are not subjected to frequent hard frosts. The mangrove eco-

system absorbs the energies of tides, wind, and waves, especially

those that result from tropical storms. In this way mangrove forests

function as a natural buffer between land and sea. The mangrove

forest is a highly productive ecosystem, which aids in supporting the

neighboring estuarine communities. Food and shelter are provided

for numerous coastal and marine animals and a nursery area is avail-

able for many species of marine organisms of commercial importance.










The mangrove forest also captures the nutrients in runoff and concen-

trates nutrients from tidal waters.

Figure 1 is a model of a mangrove ecosystem showing the coupling

of the estuarine system with the ocean and upland areas. Fresh water

is shown flowing into the estuary and carrying with it nutrients,

organic matter, toxins, and silt. A two-way exchange of water, nutri-

ents, organic matter, toxins, and salt occurs between the ocean and

estuary. Silt in the estuary shades out phytoplankton and periphyton

and may also interact with mangrove biomass to increase land area.

Silt also settles and becomes part of the bottom sediment. Toxins in

the estuary are shown as a flow to interstitial water in the sediments

where they may be fixed by the sediments or taken up by the mangrove

roots. Toxins such as heavy metals and oil spills are shown in the

model as a stress on mangrove vegetation. Nutrients such as phosphates

and nitrates are utilized by mangroves, periphyton, and phytoplankton.

Nutrients are also fixed in the sediments. Filter feeders circulate

estuarine water with its mix of ingredients and remove some of the

organic matter and toxins. Saline concentrations as reported by

Carter et al. (1973) act to decrease mangrove productivity. Heald

(1971) noted that the rate of degradation of red mangrove leaves was

faster in brackish water than in fresh water and that the slowest

rate of degradation was under dry conditions.

Total mangrove productivity is influenced by sunlight, avail-

able land area, salinity, nutrient availability, evapotranspiration,
















mangrove roots, leaves, and wood. Mangrove biomass contributes to

litter fall, which may be degraded further into particulate detritus

or grazed upon by crabs, snails, and other consumers. Some litter

may also be incorporated into the sediments as peat. Roots also form

a portion of the peaty sediments. Detritus in this model includes

particulate fragments of litter fall and subsequent stages of degrada-

tion. Detritus is consumed by detritus-feeding organisms and filter

feeders and is exported into the open estuary under tidal influence.

In the model hurricanes and herbicide spraying are shown as

stresses on the mangrove forest. It was also felt that hurricanes

play an important role in the shaping of the mangrove system. The

effect of spraying herbicide is illustrated by the conversion of

undisturbed land to bare land as described by the National Academy

of Sciences (1974). Recolonization can occur from both local and

outside seed sources.

In many areas of the tropics and subtropics the mangrove forests

have been subjected to intense pressures of development from man-

related activities. The concern here is related to whether these

types of development are able to achieve maximum value from a mangrove

forest. In southeast Asia the wood of the mangrove species, Rhizophora

mucronata is made into a high grade charcoal. In West Africa many

hectares of mangrove forests have been cleared for rice production.

In the United States, Australia, and probably other parts of the

world, mangrove forests are cleared and homes are constructed on the









cleared sites. Real estate development in the model occurs when

both money and fossil fuel are available. The adjacent estuary can

also be affected by impoundment or channelization of upland streams

or by impoundment or dredge and fill of estuarine areas.



Study Plan and Objectives


Modeling and simulation methods were used to help in under-

standing mangrove ecosystems and their responses to management alterna-

tives. Knowledge from previous studies about important processes,

components, and causal factors for mangroves was organized into gen-

eral energy and material flow models (Figure 1). Then simplified

models were developed to include only those features of the mangrove

ecosystem believed-to be major determinants in the situations studied.

Equivalent mathematical equations were written and simulations were

run on an analog computer. The simulation results were compared with

results from field studies included as part of this dissertation and

from results of field studies reported in the literature.

Since mangrove forests occur in locations that are desirable

for economic development, management of land with both uses needs to

be optimized. The development of predictive models which consider

changes over time in the mangrove forest may aid the decision making

process involving mangrove management. Such models will be needed

if the impact of increasing human influence on mangroves is to be

predicted. With this in mind, the objectives of this dissertation

were










(1) To demonstrate the usefulness of aggregated simplistic

models to simulate the sensitivity of mangrove forests

to environmental factors such as herbicide spraying,

hurricanes, nutrient enrichment, and economic development.

(2) To ascertain through computer simulations those variables

that have a significant effect on the rate of recoloniza-

tion of a mangrove forest that has been sprayed with

herbicide.

(3) To determine through computer simulations the response

and rate of recovery of a mangrove forest subjected to

hurricanes of varying intensity and frequency.

(4) To assess the response of a mangrove forest to nutrient

enrichment through computer simulations.

(5) To determine through computer simulations the ratio of

natural to developed land that will provide the maximum

value to the region under development.

(6) To observe through field studies the response of a

mangrove forest to nutrient enrichment from sewage efflu-

ent and spraying of herbicide.

Development designs in the past have not always been beneficial

to the mangrove ecosystem. What are the best ways to develop human

settlements in an estuarine area? Should the mangrove ecosystem remain

in a pristine condition or can the wastes and energy flow of man

blend into the forest without destroying it? Should a mangrove










forest be used as a disposal area for domestic sewage effluent?

What are the long-term effects on estuarine productivity when large

areas of mangrove are destroyed by herbicide spraying or hurricanes?



Description of Study Areas


Study sites were located in South Vietnam and in southwest

Florida. The investigation was concerned with processes that occur

in the mangrove forests of these regions.



South Vietnam Study Area


The area investigated in South Vietnam was the Rung-Sat district

located about 40 kilometers south-southeast of Saigon in Gia Dinh

Province (Figure 2). The Rung-Sat is an extensive delta of 105,000 ha

formed by alluvial deposits from three rivers--the Saigon, Dong Nai,

and Thi Vai. Vu Van Cuong (1964) listed the mangrove area of the

Rung-Sat as 40,000 ha. Analysis of aerial photography taken by the

National Academy of Sciences (1974) revealed that prior to herbicide

application, the land use distribution was as follows: mangrove vege-

tation covered 51% of the Rung-Sat, open water 23%, cultivated land

8%, abandoned land 6%, bush 5%, bare ground 5, and urban land 2'.

Portions of the Rung-Sat are flooded twice daily by the tides and a

normal high tide of 3.3 m will cover 85% of the area. When the highest

tides occur in June-July and December-January, the entire Rung-Sat is

inundated.































Figure 2. Map of the Rung-Sat district in South Vietnam showing
sprayed and unsprayed areas. Sprayed areas are white,
unsprayed areas are dotted, water areas are black.








10































Song



S LAOS
R7LL L .A 0


Spr II- [


SOUTH
CAMBODIA


-- VIETNAM




Rang Sot


SCmu PPoto
Conou Point










According to Chapman (1974) more than 20 species of plants were

classified as mangroves in the Rung-Sat. Vu Van Cuong (1964) listed

84 species of plants found in mangrove associations in the Rung-Sat.

Much of the mangrove vegetation formerly in the Rung-Sat was a secondary

forest because of frequent cuttings. The dominant species occurring

in the higher salinity outer zones of the Rung-Sat was Rhizophora

mucronata which was harvested by the South Vietnamese for firewood

or charcoal. Avicennmia sp. dominated the inner zone which is influ-

enced more by fresh water influx. Other species present included

Sonneratia alba (thought to be a pioneer on newly formed land),

Ceriops tagal, Bruguera parviflora, R. apiculata, S. caseolaris,

and a palm, Nypa fruticans.

During the period from 1965 to 1970, approximately 57% of the

Rung-Sat district was subjected to extensive spraying with two herbi-

cide mixtures known as "Agent White" and "Agent Orange." Agent White

is a 4:1 mixture of the tri-iso-propanolamine salts of 2,4-dichloro-

phenoxyacetic acid (2,4-D) and 4-amino-3,5,6-trichoropicolinic acid

(picloram) in water. Agent Orange is a 1:1 mixture of the n-butyl

esters of 2,4-D and 2,4,5-trichlorophenoxyacetic acid (2,4,5-T).

According to the National Academy of Sciences (1974), 3,730,000 liters

of Agent Orange and 1,580,000 liters of Agent White were sprayed on the

mangrove forests of South Vietnam. Spraying was halted in 1971 and

the sprayed areas are slowly being recolonized by mangrove vegetation.










Florida Study Areas


Three areas were chosen for study along the southwest coast of

Florida (Figure 3). In one of the areas herbicide similar to the

mixture known as "Agent White" in South Vietnam was applied by heli-

copter in an experiment designed and reported by Teas and Kelly (1974).

In two other areas, the mangrove forest was covered with diluted

sewage effluent at high tide.

Two of the three study areas were located in a region known

as Ten Thousand Islands. In the Ten Thousand Islands region the man-

grove strip is two to three miles wide with a mile wide lagoon behind

the mangrove strip. Tanner, Evans, and Holmes (1963) reported that

in southwest Florida the shoreline north of Cape Romano was a quartz-

sand beach while the shoreline to the south of Cape Romano was a

mangrove tree cluster barrier. The Ten Thousand Islands area seems

to be one of low wave energy but relatively large tidal activity due

to the shape of the shoreline. A concave shoreline such as the

southern shore of southwest Florida causes waves to be refracted

and tides to be funneled into the bay. Tidal range is from 2-4

feet in southwest Florida.

Lugo and Snedaker (1974b) listed five types of mangrove eco-

systems as determined by local tidal patterns and upland surface

water runoff. The fringe forest occurs along slightly sloping, pro-

tected shorelines of the mainland and larger islands. Tidal water

movement is restricted to an in-out pattern. The riverine forest































Figure 3. Map of southwest Florida showing the sites treated with
sewage, the site treated with herbicide, and the sites
where mangrove productivity and biomass were measured
by Lugo and Snedaker (1974a).

(a) Map of the state of Florida.
(b) Map of south Florida showing locations of study areas.
(c) Map showing location of Naples study plots.
(d) Map showing location of Everglades City study plots.
(e) Map showing location of Marco Island study plots.







14








ATLANTIC
S( \ OCEAN







GULF OF

MEXICO





STUDY---
SITES


a









NAPLES-
c) ^ ROOKERY
BAY
F L O R I D A
FLORID A
MARCO
ISLAND IA

TEN MIAMI
THOUSAND D S
ISLANDS

EVERGLADES
CITY
(dl
LUIF TURKEY PT
OF *
MEXICO ATL ANrlT

0 OCEA /


SCALE
0 S o 300 oMILES
0 a 17 25 as o KILOMETERS



























































Figure 3 (continued)










is situated on the mainland in strand-like formations along river and

creek drainages. Tidal flushing can be daily and flooding is caused

by fresh water runoff and/or high tides. The overwash forest occurs

on small islands and finger-like projections of larger land areas.

Daily tidal water movement is through the forest and velocities are

high enough to remove most of the debris. The basin forest occurs

behind the fringe forest. The land is flat, water movement is slow

and tidal flushing may be infrequent. Another type may be the dwarf

mangrove forest, which occurs in flat coastal fringes where trees are

< 2 Im tall and may be nutrient limited or salt stressed. The mangrove

hanmmock is a tree island surrounded by dwarf mangroves or sawgrass

wetlands. Water flow is around the island and gives the hammock a

tear-drop shape.

The forest sprayed with herbicide was located between the towns

of Goodland and Marco Island along State Road 92 (Figure 3e). The

sites nearest the road (farthest from tidal creek) were dominated by

A. germinans while sites nearest the tidal creek were dominated by

R. mangle. Scattered trees of the species Laguncularia racemosa were

also located in the area. A small succulent, Batis maritima, was also

present. Forests of -he riverine and the basin type were found here.

The herbicide used on the sprayed sites was Tordon 101R manu-

factured by Dow Chemical Co. Tordon 101R is similar to Agent White

and consists of a mixture of 0.24 kg/l of 2,4-D and 0.65 kg/l of

picloram on an acid equivalent basis. The sprayed and control sites










measured 20 m by 40 m and the cleared sites measured 40 m by

40 m.

The study sites receiving nutrients from treated sewage effluent

were located in Naples and Everglades City. In each site sewage

effluent was not discharged directly into a mangrove forest. The

contact of mangroves with sewage was limited to tidal flooding at high

tide. The Naples study site (Figure 3c) was located along the Gordon

River which meanders lazily through the town of Naples on its way to

the Gulf of Mexico. The nutrient enriched study site was located

on an island in the Gordon River about 100 m from the point of dis-

charge of sewage effluent from the Naples sewage treatment plant.

The plant was designed in 1966 to treat 9500 m3/day (Candeub, Fleissig,

and Associates, 1973). This design capacity was increased to 19,000

m3/day in March 1975. Treatment consisted of contact stabilization,

chlorination, and then retention in a stabilization pond before dis-

charge into the Gordon River. The experimental study site consisted

of an interior composed almost entirely of L. racemosa ringed by R.

mangle growing at the water's edge. A control site was selected up-

stream from this site away from the influence of the sewage treatment

plant. The dominant tree in the control site was L. racemosa with a

few trees of R. mangle. These sites were usually inundated by tides

twice each day. During the dry season, however, there were days when

the tides did not reach this level. These forests are examples of

the riverine type forest.










The Everglades City study sites (Figure 3d) were located along

a tidal canal which discharges into Chokoloskee Bay. The experi-

mental sites were located immediately upstream and downstream from

the point where the sewage effluent is discharged into the canal.

The treatment plant was designed to treat 100,000 gallons per day of

sewage. Treatment consisted of extended aeration, chlorination and

retention in a stabilization pond before discharge into the canal.

A control site was selected farther upstream along a small tidal

creek that emptied into the canal. The sewage site was composed of

L. racemosa and R. mangle with a few A. germinans while the control

site was dominated by R. mangle with a few L. racemosa and A. ger-

minans. These forests are examples of the riverine type forest.
















REVIEW OF LITERATURE


This review will briefly discuss previous publications that

relate to this dissertation. These include the problems of ecosystem

modeling, modeling of mangrove forests, past use of herbicides on man-

groves, effects of sewage effluent, effects of hurricanes on man-

groves, and interactions with economic development in mangrove forests.



Ecosystem Modeling


Modeling at the ecosystem level generally involves a simpli-

fication of the extremely complex processes and interactions within

ecosystems. One approach is to try and model all details so as to

emulate the real system. An analysis by O'Neill (1971) indicated that

increased model complexity does not always bring about more accurate

results that are more comparable to the real system. According to

O'Neill, adding more parameters to an ecosystem model makes it necessary

to quantify these additional parameters in field and laboratory experi-

ments which have errors of measurement. Other approaches to under-

standing a system may deal with studying its parts--a reductionist

approach that can have adverse effects on conclusions about the system.

A third approach is to simplify by aggregating, a method used in this

dissertation. Aggregating is a process of combining the details of a










base model into a less complex model so that simulations can be made.

Combining species or parameters with comparable turnover times is one

means of aggregating ecosystem models.

Burgess and Kern (1973) felt that the mathematical rigor of eco-

system modeling should result in greater efficiency and more precise

structure in research programs. Modeling can also formalize insights

and graphically highlight areas where progress was slowed due to missing

data. Limitations to progress in modeling are the understanding of

mechanisms and the ability to measure relevant parameters. Models pro-

vide a framework for synthesizing and comparing data. Ecosystem model-

ing may be a problem because of the uniqueness of each ecosystem.

Levins (1966) pointed out that if this is true, data obtained from one

ecosystem cannot be applied to another ecosystem.

Very little modeling has been done of mangrove swamps. Sell,

in Odum et al. (1974), reported on preliminary simulations of man-

groves recovery following herbicide spraying in South Vietnam. The

complete results are in this dissertation. Miller, Ehleringer,

Hynum, and Stoner (1974) have simulated the reforestation problem

of the red mangroves in Vietnam by combining climatological,geo-

logical, and physiological characteristics of the Rung-Sat district

in South Vietnam in order to understand the processes and interactions

influencing regeneration of the mangrove forest. They tested two hy-

potheses as to the cause of slow mangrove regrowth: (1) low propagule

availability and (2) growth of propagules and seedlings is stopped










because defoliation has increased temperature and salinity. Their

model indicated that high temperatures and low availability of propa-

gules are the major factors influencing regrowth and combined may

play the major role in reforestation of mangroves on the Rung-Sat.

Zieman et al. (1972) developed a model to study natural suc-

cession in the mangrove swamps of Vietnam. Waterlogging of the soil,

soil water salinity, frequency of tidal inundation, and shade were

programmed as limiting factors in the model. The model looked at suc-

cession in drained soils and in waterlogged soils. Available nutrients

and establishment of seeds were not considered because no data were

available. In waterlogged soils S. alba grew first followed by R.

mucronata at year 17 which peaked at year 59 with a biomass of about
2
23,000 g/m R. apiculata began growing at year 59 and reached a

steady state biomass of about 39,000 g/m2 at year 130. In drained

soils, A. marina was the pioneer, followed by B. cylindrica in year

27, which was followed by B. gymnorhiza in year 53. B. gymnorhiza

reached steady state biomass of about 72,400 g/m2 in year 200. Increas-

ing productivity did not have a large effect on biomass but did increase

the rate of succession. These biomass values seem much too high be-

cause of the extremely high rates of net primary productivity used in

this model. These authors also neglected hurricanes and woodcutting

effects on biomass.

Miller (1972) also simulated a model of primary production and

transpiration for various levels of a mangrove forest canopy in Florida.

The simulation made use of an energy-budget equation of individual










leaves. For a red mangrove (R. mangle) forest, Miller (1972) calcu-

lated net primary production to be an average of 3.4 g/m2.day in June

and 2.2 g/m2 day in January for a yearly average of 2.8 g/m 2day.

The model showed that primary production decreased with increasing

air temperature and humidity. Increasing solar radiation increased

primary production but at high levels, production was decreased.

Transpiration was calculated to be 20% of the total water loss of the

forest with the rest coming from the forest floor. Miller concluded

that the canopy of the red mangrove forest was adapted to maximize net

primary production under conditions of saturated water supply.

Lugo, Sell and Snedaker (1976) simulated a model of the overwash

mangrove forest type in order to study the effects of upland runoff

and tidal flushing on processes occurring within a mangrove system.

They found that tidal flushing affected the storage of detritus in the

forest and nutrient availability affected the productivity of the

forest. Upland runoff reduction lowered the productivity of the man-

grove forest. The model in Figure 15 of this dissertation is similar

to that model but the stress of the occasional hurricane has been added.



History of Herbicide Use in Mangroves


The control of mangrove vegetation by the use of herbicides

has been going on for about 20 years. Ivens (1957) reported that by

applying either 2,4-D or 2,4,5-T at concentrations of 4-20' in diesel

oil Rhizophora and Avicennia species could be killed. The Avicennia










species was found to be slightly more resistant than the Rhizophora

species. Dixon (1959) reported that a 10% solution of 2,4,5-T in

diesel oil applied with a brush to the bark of trees of B. parviflora

in Malaya resulted in a 90% kill after three months. More desirable

Rhizophora seedlings were planted immediately after the treatment

and grew quite well. Truman (1961) found in Australia that A. marina

was defoliated when solutions of 1,2, or 4;, of 2,4,5-T were sprayed

on the bark or leaves of trees of that species. He concluded that

this mangrove species was very susceptible to these herbicides.

Nielands et al. (1972) reported that herbicide experiments

made in South Vietnam during 1961 and 1962 were generally successful

in killing most species of vegetation found in Vietnam. Minarik and

Bertram (1962) reported that in September and October of 1962 the

United States Army conducted a series of spraying operations on mangroves

in Vietnam. These experiments were performed along several canals

and roads in Camau peninsula served as transportation arteries and

needed to remain open. Spraying was successful to the extent that

95% of the mangrove vegetation was defoliated at a spray concentration

of three gallons of "Agent Purple" per acre. "Agent Purple" consisted

of a 50:30:20 mixture of the n-butyl ester of 2,4-D and the n-butyl and

isobutyl esters of 2,4,5-T. Its use was discontinued in 1964. Minarik

and Bertram (1962) concluded that the spraying operation proved that

defoliation of large areas of mangrove vegetation was technically and

operationally possible.










Teas and Kelly (1974) assessed the impact of the herbicide,

"Agent White" by spraying several small areas of mangroves near Marco

Island, Florida. Defoliation was 100% for L. racemosa within five

weeks after treatment. R. mangle was 90% defoliated within six weeks

while A. germinans was only 25% defoliated after 16 weeks and was

nearly recovered in 16 months. The trees of L. racemosa began to die

at 24 weeks and the trees of R. mangle started dying at 30 weeks after

spraying.

Lugo and Snedaker (1974b) manually defoliated R. mangle trees

in South Florida. New leaves were put out by these trees even when the

defoliation was repeated, suggesting that loss of leaves in this manner

does not result in mortality of the trees.

Walsh, Barrett, Cook, and Hollister (1973) investigated in the

laboratory the effects of a commercial formulation of 2,4-D and picloram

on the seedlings of the red mangrove, R. mangle. Seedlings were

treated with herbicide concentrations of 1.12, 11.2, and 112 kg/ha.

A dosage of 1.12 kg/ha appeared to have very little effect on growth

but at the higher dosages, seedlings died within 40 days.

The mangrove vegetation of the Rung-Sat has been very slow to

recolonize bare areas that resulted from spraying. This slow recoloni-

zation has been noted in other mangrove swamps of the world. Macnae

(1968) has referred to the general problem of mangrove recolonization

in cleared mangrove areas and wondered why recently cleared areas often

become deserts suitable only for salt production in regions with low

or seasonal rainfall.










Meselson, Westing, Constable, and Cook (1970) reported that

mangrove species in Vietnam showed no signs of recolonization of

areas sprayed three or more years prior to inspection.

Westing (1971) noted that spraying mangrove vegetation with

herbicide not only caused defoliation but also seemed to prevent re-

establishment of any new plant community. Westing observed that six

years after spraying there was no clear evidence of regeneration

of mangroves.

Orians and Pfeiffer (1970) observed that the areas they visited

inthe Rung-Sat were still barren although spraying had occurred several

years earlier. They concluded that re-establishment of mangroves

may take more than 20 years because of the problems of seed dispersal

into the sprayed areas and the possibility of herbicide residues in

the soil. In order to determine if there was a herbicide residue

problem the National Academy of Sciences (1974) conducted seedling

planting experiments near VungTau in South Vietnam in 1972 to determine

whether the presence of herbicide in the soil has any effect on seed-

ling survival. The study indicated that the presence of herbicides in

the soil made no difference in seedling survival. Tschirley (1969)

estimated that an area of mangroves sprayed in 1962 in the Rung-Sat

may reasonably be expected to take 20 years to return to prespraying

conditions based on the regeneration he observed at that time.

Natural regeneration of mangroves may depend upon a number of

factors, the most important being availability of adequate numbers of

seedlings and seeds for dispersal into bare areas. If seedlings are









present, other factors such as pH of soil, salinity, nutrient levels,

soil temperature, and predation by animals such as crabs may hinder

regeneration of mangroves. Debris may prevent seeds and seedlings

from reaching inland areas. Woodcutters also significantly deter the

full recovery of mangroves by cutting the mangroves before they bear

fruit. In Thailand Banijbatana (1957) felt that illegal cutting had a

serious impact on natural regeneration of mangroves. Krishnamurthy

(1974) estimated that it may take 80 years for a mangrove community

to recover from destruction by man. Dixon (1959) noted that in

Malaya clear cut areas required seven years for natural regeneration,

while Wadsworth (1959) reported that after a mangrove forest in Puerto

Rico was clearfelled, natural regeneration occurred within two years.

The studies of herbicide use reviewed here have indicated that

the mangrove forest is very susceptible to herbicide and that complete

recovery of the mangrove forests may be a long time coming in Vietnam. In

fieldwork by H. T. Odum in VungTau in 1972 many cut stems were ob-

served. Woodcutting of mangroves was the basis of a charcoal industry

with export sales to Japan in the early 1970s.



Sewage Disposal in Wetlands


Nedwell (1974) has suggested the disposal of sewage effluent

on mangrove land based on the ability of bacteria in the anaerobic

mangrove sediments to reduce nitrate to gaseous nitrogen. No










previous studies have been located on the effects of nutrient enrich-

ment on tree growth and litter fall in mangrove forests.

A few studies have been made on the response of salt marshes to

treated municipal sewage effluent. Marshall (1970) assessed the effects

of added nutrients on a Spartina salt marsh along the North Carolina

coast. The discharge rate was 1900 m3/day of sewage into a tidal creek.

The marsh receiving treated sewage effluent generally produced signifi-

cantly greater weights of live Spartina than the control marsh. The

number of steins per square meter was not significantly greater and the

greater biomass in the sewage creek was accounted for by the increase in

stem diameter. Phosphorus concentrations in July 1969 were 0.93 ng/l for

the sewage creek and 0.078 mg/l for the control creek. The sewage marsh

also had a higher net productivity than the control marsh as measured

by the difference between the maximum and minimum standing crops dur-

ing the growing season.

Valiela and Teal (1974) designed experiments to measure the re-

sponse of salt marsh communities in Massachusetts to phosphorus and

nitrogen enrichment. They found that nitrogen enriched plots had in-

creased biomass of salt marsh plants while phosphate enrichment had

negligible effect on the biomass as compared to control marshes. The

conclusion was that levels of nitrogen were critical to the productivity

of vegetation in their study marsh in Massachusetts. Valiela and Teal

(1974) also found that after 3 years of applying sewage sludge to the

salt marsh, primary productivity, decomposition and secondary produc-

tivity were higher than in a control.










Spangler, Sloey, and Fetter (1976) investigated the response of

a fresh water marsh in Wisconsin to the discharge of municipal sewage

effluent. The city of Brillion sewage treatment plant discharged ef-

fluent into a creek that flows into the marsh. Discharge of effluent

ranged from 750 to 1500 i3/day. They found that the experimental sewage

marsh produced lower standing crops than the control marshes. Measure-

ments also revealed that with a multiple harvest during the year 0.83-

1.62 g P/m2 were removed from the sewage marsh while a single harvest

removed only 0.6 g P/m2 from the control marsh.

Odum, Ewel, Mitsch, and Ordway (1975) reported on the recycling

of treated sewage through cypress wetlands in Florida and observed that

at a loading rate of 10 cm/week only 4% of the phosphorus was removed

but at a rate of 4 cm/week 75% of the phosphorus was removed from the

sewage effluent. Mitsch (1975) used ecosystem modeling to help under-

stand the impact on a cypress dome in Florida of high nutrient inputs

from sewage. A computer simulation of the addition of sewage to a

cypress dome indicated that in a 30-year period the production of cy-

press wood increased 25% at a loading rate of 2.5 cm/week and 50% when

the loading rate was increased to 5 cm/week. Mitsch recommended that

a loading rate of 3-5 cm/week would be best if cypress domes were to

be used for disposal of sewage effluent.

Boyt (1976) reported on the responses of a mixed hardwood swamp

to 20 years of receiving sewage effluent. Measurements indicated that

98% of the phosphorus influx was retained in the swamps. Growth rates

for cypress trees during the 20-year period were 5.5 cm in the sewage










swamp and 3.9 cm in the control swamp. Ash trees had a 20% higher net

productivity in the sewage swamp than in the control swamp. Understory

vegetation in the sewage swamp was six times that of the control.

Wentz (1975) reported that wetlands in Michigan showed no sig-

nificant changes in growth, productivity, or nitrogen and phosphorus

concentrations following application of simulated sewage effluents and

concluded that as nutrients accumulate, some species may show increased

growth while others may show decreased growth.



Effects of Hurricanes


Winds and wind-produced tides and waves cause great damage when

a hurricane passes, often covering roots with a deep layer of mud. Craig-

head (1971) noted that Hurricane Donna in 1960 did extensive damage to

the mangrove forests in the Ten Thousand Islands coastal area. Many

trees were sheared off completely while others lost many branches.

Nearly all trees were completely defoliated. However, within four weeks

new leaves appeared but six weeks later, the lowland mangroves were dead

while those at slightly higher elevations continued to live. Craighead

(1971) attributed this to a lack of gas exchange in the roots because

they were covered by an impervious mud layer anywhere from 3-15 cm thick.

Sauer (1962) reported that in 1960 two strong cyclones struck the

tiny island of Mauritius off the east coast of Africa. The first one

passed near the island with winds gusting to 160-200 km/hr while a month

later the second cyclone was a direct hit with winds of 240 km/hr. Al-

though flooded to a depth of 2 meters the mangroves were only slightly

damaged. Some older trees and seedlings were uprooted.










Glynn, Almodovar, and Gonzalez (1964) reported that Hurricane

Edith passed by La Parguera in Puerto Rico but only slight defoliation

of the mangroves occurred. Winds were not high enough to uproot or

shear the mangroves along the coast.

Major hurricanes can do extensive damage when they occur, but

the frequency of these great tropical storms is low and the total

effects are relatively small when averaged over the years of frequency

of hurricanes. Tanner (1961) estimated that the impact of Hurricane

Donna in 1960 was equivalent to the work of 100 years of ordinary

processes of moving sediments. The mangrove ecosystem appears to be

adapted to and dependent on the infrequent occurrence of well-developed

tropical storms.



Occurrence and Frequency of Hurricanes and Typhoons


Ludlum (1963) gave an account of early American hurricanes and

stated that a hurricane in 1835 was probably the first major hurricane

with documentation to pass through the Florida Keys. The storm was

believed to have passed along the Florida west coast and may have

affected the Ten Thousand Islands mangroves. In 1846, the worst hurri-

cane since 1821 struck Havana, Cuba and then moved along the west coast

of Florida. Riehl (1972) reported that a hurricane passed through the

Ten Thousand Islands in 1876. Tannehill (1938) reported that hurricanes

may have hit the Ten Thousand Islands in 1891, 1910, 1924, and 1935.

The last major hurricane to hit this area was Hurricane Donna in 1960.










This brief history revealed that since 1835 possibly 8 major hurri-

canes have passed through the mangrove region known as the Ten Thousand

Islands, a frequency of 1 major hurricane every 18 years. Lugo, Sell,

and Snedaker (1976) reported a frequency of one hurricane every 20

years for Florida and every 24 years for Puerto Rico.

Hurricane records for Jamaica date back to 1689 and Fowler

(1952) felt that 17 could be considered major which gave a frequency

of one major hurricane every 15 years. Shellard (1971) estimated from

wind speed data that a wind speed of 120 mi/hr (54 m/sec) would be

expected to occur in Jamaica only once every 50 years. He also said

that a well-developed hurricane with winds of 150-160 mi/hr (67-71

m/sec) should not be expected to occur again for more than 50 years.

Typhoons are frequent in Southeast Asia. Ramage (1971) gave

a typhoon frequency for the Rung-Sat of about one every ten years.

In 1948 a typhoon passed just north of the Rung-Sat. In 1950 and

1956 typhoons passed to the north of the mangroves in the Camau Peninsula.

The intensity of these storms was not given.



Economic Development in Mangroves


Development of coastal mangrove areas has been occurring through-

out the world's mangrove zone. Baines (1974) discussed the uncontrolled

development of mangrove lands in Australia. Intense pressures were

applied to convert the mangrove swamps into residential canal develop-

ments as is being done in the United States. Recent awareness of their

value as a recreational and commercial fishery was evidenced in the










recommendation by the Australian Conservation Foundation (1972) that

the relevant authorities should carefully study mangrove areas before

allowing development and consider their value beside the proposed

land use values.

Odu, (1971) presented evidence that mangrove swamps are valu-

able, productive regions and indicated the importance of mangrove

detritus in the estuarine food web. Removal of mangrove swamps for

the purpose of economic development would in the long run reduce the

estuarine productivity that originally attracted the development.

The effect of development on a mangrove forest will largely

depend on the nature of the development. Tabb and Heald (1973) have

proposed the use of an interceptor canal that would be constructed

between the coastal mangroves and the upland development. Veri

et al. (1973) proposed a resource buffer between upland development

and the coastal mangroves. These developments were designed to pre-

serve the coastal mangrove swamps. Finger canal developments have

caused extensive damage to estuarine areas of Florida.

Steller (1976) evaluated alternative residential development

plans in a coastal mangrove area of southwest Florida. Energy in-

vestment ratios (ratio of fossil fuel amplifying energies to resident

natural energy flows) were evaluated for interceptor canal, resource

buffer, finger canal, special treatment, and high density developments.

Since the calculated investment ratios exceeded those for Florida

and the United States, Steller (1976) concluded that these






33



developments may not be competitive. In order to be competitive,

the land area developed needed to be much less for each develop-

ment.















METHODS


Modeling and simulation techniques were used in this disserta-

tion as a means of developing an understanding of the responses of

mangrove ecosystems to perturbations by man and nature. Energy flow

calculations were used in the model of economic development and man-

groves to assess the relationship between natural and fossil fuel

energy flows. Field studies were used to provide additional infor-

mation and to validate the simulated responses.



Modeling and Simulation Techniques


Ecosystem models were conceptualized and simulated on analog

computers to assess the effects of such factors as hurricanes, herbi-

cides, economic development, and nutrients on the mangrove ecosystem.

This section describes the procedure for developing a model, the

writing of the equations that describe the model in mathematical

terms, the use of model aggregation, the calculation of rate coeffi-

cients for the pathways of energy or material flow and the scaling

of the equations.










Development of Models and Equations


The modeling procedures in this dissertation involved deter-

mining the energy flows from outside the ecosystem being studied. Also,

an assessment was made of those storage or compartments thought to be

important within the ecosystem. Finally, the pathways, interactions

and processes that are important in the ecosystem being studied were

also determined. A qualitative diagram was then drawn of the simpli-

fied ecosystem. The symbols used in diagramming the models in this

dissertation were those developed by Odum (1971). Each of the symbols

is shown in Appendix A along with a description and also a mathematical

equivalent where appropriate.

After a model was developed, the next step was to write the

differential equations that describe each of the state variables or

compartments of the model. The differential equation described the

time rate of change in a compartment or state variable as it is af-

fected by the flows entering and leaving the compartment. Each com-

partment in the model could thus be described by a differential equa-

tion. Use of these differential equations made it possible to simulate

the model on computers. The output graphs show the trends with time

that resulted for a particular model and its set of parameters.


Aggregation


For the prediction of the overall response of an ecosystem model

to various interactions, aggregation of parameters was an important










means of simplifying the model for better understanding. If many of

the parameters were constant or small in their effect on the system,

many of the internal details occurring in the mangrove ecosystem

could be omitted. For example, the photosynthetic process produced

organic matter used by the mangrove trees for their growth and meta-

bolic processes. The process of photosynthesis actually included

numerous steps or chemical reactions that eventually result in the

production of this organic matter. Details of this nature were gen-

erally unnecessary in these simulation models since the primary concerns

of the study involved events on a larger scale. The technique of

aggregation or lumping is probably valid whenever the overall result

of some process is desired rather than the intricate details. Whether

a process is important or insignificant can be studied by trial and

error or by estimating the relative magnitude of its effect on energy

flows.



Calculation of Rate Coefficients


Each line or pathway that appears in the model diagram has a

set of terms associated with the pathway along with a rate coefficient.

If the value of this pathway was known and also the state variables

or driving forces describing the pathway, then the rate coefficient

could be calculated. The calculation of the rate coefficient involved

setting the observed flow equal to the algebraic term for that flow

where the rate coefficient was the only unknown parameter. An example









of the calculation is given below for the flow of organic matter

into mangrove biomass (see Fig. 15). The equation for the flow is

klIRQ1Q3 where


k1 = rate coefficient

IR = solar radiation available for photosynthesis

= 8.0 x 105 kcal/m2-year

Q1 = initial amount of mangrove bioiiass = 10,500 g C/m2

Q3 = initial amount of nutrients = 1600 g/m2


and the value for the flow is 2820 g C/m2 year. Therefore,



kiIRQ Q3 = 2820 g C/m2-year



or


k 2820 g C/m2.year
S (8.0 x 105)(10,500)(1600)(kcal/m2 year)(g C/m2)(g/m2



-10 4
kI = 2.10 x 1O 10 m /g-kcal



Calculations must be made for the rate coefficients of each pathway in

the model. In this simulation method the assumption was made that

these rate coefficients did not change during a simulation run.

Although this may not be true in many instances, the lack of data









showing a change with time usually requires the researcher to make

the above assumption.


Scaling of Equations


Once the rate coefficients have been determined, the equations

describing each compartment need to be scaled if an analog computer

is used in the simulations. Scaling is the expression of equations as

percent of full scale on output graphs. Scaling was necessary because

the output from an analog computer is in terms of voltages. If a

maximum voltage is exceeded, the results will be in error. Each

variable was replaced by a quotient made by dividing by the maximum

value expected. In this way the quotient was kept less than unity

(full scale) and thus within the maximum scale chosen and the voltage

range of the analog computer. Suppose we had an equation such as that

given for the rate of change of mangrove biomass in Figure 15 where



dQl 2
dt= k IRQ3 k31 k2Q1 k16 1H (1)



In order to scale this equation, one needs to know the maximum

values that should be expected for the compartments (Ql' Q3) and the

driving forces (IR, H). In the model shown in Figure 15, Q1 was

assigned a maximum value of 3 x 104, Q3 a maximum value of 10 IR a

maximum value of 2.43 x 106 and H a maximum value of 13. The next

step was to divide each compartment and driving force in the equation









by its maximum value and to multiply at the same time as shown
below.


dQ I
d k1 (2.43 x 106) [ 6 (3 x
2.43 x 10


k3 (3 x 104)2 [__-1 2
3 x 10


104) 1 4] (104) E4]
3 x 10 10


k2 (3 x 104) [ 4]
3 x 10


- k6 (3 x 10 1 l4] (13) [1H
3 x 10 13


Next, divide both sides of the equation by the maximum value of 3 x 104
for Q1. If the differential equation were for Q2, then the maximum
value for Q2 would be used. Equation 2 was then rewritten so that


dQl
dt
3 x 104


0 'R %1 4 zi 2
k (2.43 x 1010)[ IR Q1 -k(3x10 4Q 2
2.43 x 10 3 x 1 10 3x10
2.43 x 106] [ x 10 10 3xlO


QQ Q
- k2 k6 1 (13)
3 x 10 3 x 10'


The calculated values for each rate coefficient were then substituted
into equation 3 and the resulting numerical values represented the
pot settings for the model pathways on the analog computer.










Energy Calculations


Energy flow calculations were made in conjunction with the model

assessing the relationship between economic activity and the mangrove

ecosystem. The concepts of energy quality and investment ratio were

used in the energy calculations and are discussed next.



Energy Quality


The value of a pathway within a system will vary according to

the amount of useful work the pathway does for the good of the system.

An energy flow low in heat content but high in the ability to feedback

and to do useful work is said to be of high quality while an energy

flow high in heat content but low in the ability to do useful work

is said to be of low quality. Sunlight is of low quality because of

the small amount of useful work performed per calorie of heat content.

Electricity is of high quality because of the large amount of useful

work performed per calorie of heat content. In order to compare the

different abilities of pathways to do work for the system, the heat

content energies must all be converted to a common energy quality

denominator. In this dissertation the coal equivalent (CE) was used

as the base. This means that for higher quality energy flows, the

energy quality per calorie will be X coal equivalents where X is

greater than one. Electricity is an example of this because of the

requirement of 4 units of coal energy to produce 1 unit of electricity

thus giving electricity an energy quality of four CE. For lower










quality energy flows, the energy quality per calorie will be Y coal

equivalents where Y is less than one. Sunlight is an example of this

because 2000 kcal of sunlight are required to make one CE, thus giving

sunlight an energy quality factor of 0.0005. Energy quality might

also be calculated from the energy needed to develop a flow without

waste. Table 1 is a listing of the energy quality factors used in

energy flow calculations contained in this dissertation. The refer-

ence source explaining the calculations is also given in Table 1.

Economic flows are often expressed as dollar flows. Odum

and Odum (1976) calculated the energy basis of the United States for

1973 in terms of the ratio of kcal of coal equivalence to dollars

spent. The value was 25,000 kcal/$ and was used in this dissertation

to calculate the contribution of goods and services and fuels to the

development.

Tables were prepared showing the main energy flows of man and

nature affecting kcal of heat equivalence and kcal of coal equivalence.

Coal equivalent flows were obtained by multiplying the energy quality

factor times the heat equivalent flows.



Energy Investment Ratio


Energy investment ratio was defined by Odum and Odum (1976) as

the ratio of high quality feedback energy flow investment to lower

quality natural energy flow already present where both flows are

expressed in kcal of coal equivalents. This theory suggests that









42



5-
S-
ro


=3 C
u
00
00 .-0 C


Z) 1- m

5 3 >,C) 2U -n -

cu
o C C o C







u 44 1 i-I r, s
51 */')A C)01 C
aj ~ i- '- _r

0 1 -J E C5 00









CU t^ Q. C C 54 D
S- C C C O





o CC








a)
--
0










u,~
C5 -4-1


cu, C
QJ
U +- CD)
o 0l












Ou
- C1 04
- Sn O -s

5- .>1 0 0 0 0 ro C)- 0S
O a) co 10

















ZCTCL
0 W
3 : 0 05

C) ( nJ
s, h O
=uS 0

0
U
u





n3n
5-
a)




o -i-
U 0
Co -
5- 0-

o U
4- 0 C)

>5 C.
4-SJ / C
o -0 '-. 4-

S -g Ca C)


01i C)51 3: r O
cr ^ n o

Sa ^h~ E
5-. C V "r- S~ i C)


C- 4- C.<1 = t4 /
L2 C) >i >5 0 3: C_
C 4--' Ca C

-_ .- >5 5)1 C)" O i U -

.0 I C 0 C 0 0 0 1 C o 0 -
C) I; ^ .'- .- 0 0 0 '- 0 0 5-)
I- I ^ Q- S 5-- W 5-2 Wl 05 C) Q










systems better able to match the high quality feedback energies with

more of the lower quality natural energies can charge lower prices

for their exports. If an investment ratio for a system is higher

than the ratios for surrounding systems, that system may not be able

to successfully compete with the other systems. This theory probably

applies only for those economies that have stopped growing. When

energy is plentiful, a growing system may be able to out compete other

systems even though the growing system has a higher investment ratio.

High investment ratios suggest that a region is dependent on large

amounts of energy from other areas. A low investment ratio suggests

that natural energies account for a significant part of the total

energy flow within the system.

Energy investment ratios were calculated for the hypothetical

cases involving economic development within a mangrove forest. These

results were compared to the average investment ratio for the United

States given by Odum and Odum (1976).



Field Studies


Data were collected in the field to supply missing information

needed for the models and also to validate results of the simulations.

Field measurements included tree growth, litter fall, total phosphorus

in water and in mangrove wood and leaves, number of live seedlings,

number of dead seedlings, number of snails, number of crabholes,

number of green leaves on the ground, number of yellow leaves on the

ground, and number of seedlings on trees.










Tree Growth


The growth of mangrove trees was measured at the Naples,

Everglades City,and Marco Island study sites. This was accomplished

by measuring the changes in trunk diameter over a period of time.

At the time of initial diameter measurements the trees were marked

with aluminum tags so that future identification would be possible.

The usual procedure of measuring tree diameter at breast height (about

1.5 m above the forest floor) was followed for L. racemosa and A.

germinans. However, for R. mangle the measurements were made at a

level about 15 cm above the highest prop root. The level of the

highest prop root was variable and ranged from less than 30 cm to

greater than 2 m above the forest floor. Measurements were made with

a diameter tape calibrated in metric units for those trees with diameters

greater than 4 cm and calipers were used for trees having diameters

less than 4 cm. Precision of measurement was to the nearest millimeter

with the diameter tape and to the nearest 0.1 millimeter with the

calipers. Diameters were measured in January 1973 and February 1974

at the Marco Island site and in September 1973, 1974, and 1976 at the

Naples and Everglades City sites.



Litter Fall


Baskets were placed at both the Naples and Everglades City
study site. Each basket had an area of 0.25 m. Ten baskets were
study sites. Each basket had an area of 0.25 m Ten baskets were









placed at each site. All of the baskets used at the Naples study

site were initially located on the mangrove island near the Naples

sewage treatment plant. Five of these baskets were located on the

upstream side of the island close to the point of discharge of sewage

effluent into the Gordon River. These remained in place for the dura-

tion of the study. The remaining five baskets were located on the

downstream side of the island furthest from the discharge of sewage

effluent. These baskets were removed after nine months and placed

in the control site for the remaining time of the 12-month study.

At the Everglades City study site five litter baskets were placed up-

stream from the point of sewage effluent discharge and five baskets

were placed downstream from the point of discharge.

Litter fall was collected from these baskets at 4, 13, 21, 33

40, 48, and 55 weeks after the baskets had been put out in the study

sites in September 1973. Litter from each basket was brought back to

the laboratory and separated into leaves, wood and seeds. The

separated portions were dried to constant weight at 700C (usually

about 48 hours) and weighed. Rates of leaf, wood and seed fall were

expressed in g/m2.day.



Total Phosphorus in Water Column


Water samples were collected from the Naples and Everglades

City sites and analyzed for total phosphorus. The center point of

sampling in each case was the sewage effluent discharge point and










other sampling areas were located at increasing distances from this

point. Water samples were collected in duplicate at each station in

volumes of 25 ml. These samples were then brought to the laboratory

for analytical determination of phosphorus. The procedure used

follows that outlined by Menzel and Corwin (1965) which involves a

molybdenum blue color with spectrophotometer readings at 900mu using

adsorption cells of 4 cm length. The procedure for determining total

phosphorus in water is given as Appendix B.



Phosphorus in Leaves and Wood


Phosphorus concentration was measured in mangrove leaves and

wood using the method for phosphorus determinations in the water

column. Leaf and wood samples were collected from the Naples and

Everglades City sites, dried, ground,and dissolved in hydrochloric

acid prior to analysis for phosphorus.



Numbers of Green Leaves, Yellow Leaves,
Live Seedlinq., Dead Seedlings,
Snails, and Crabholes


At the mangrove sites near Marco Island the number of green

leaves and yellow leaves on the ground, live seedlings, dead seedlings,

snails, and crabholes was determined at the control site, one of the

cleared sites, and the three sprayed sites. These parameters were

selected because changes in population density could be easily detected

by simple measurements.










The initial measurements were made about 1 month after the

sites were sprayed with Agent White on December 18, 1972. Subsequent

data were collected 5, 8, 12, 20, and 26 months after spraying. The

first set of data was collected using a circular hoop to enclose

sampling areas of 0.4 m2. All data were collected after that first

time in sampling areas enclosed by a circular hoop with an area of 0.77

m2. Ten areas were sampled in each study site and counts were made

of each of the above mentioned parameters. These counts were then

averaged and divided by the hoop area to give the values on the basis

of numbers per m2. The ten areas in each site were selected at ap-

proximately 5 m intervals moving in a zigzag manner through the plot.

The sampled areas were different for each sampling period.



Number of R. mangle Seedlings on Trees


At Naples and Everglades City, the number of R. mangle seedlings

on parent trees was measured using the 0.77 m2 circular hoop. Only

portions of the tree over the waterway were considered in these

measurements. This technique was also used in Rookery Bay to obtain

seedling data that was applied to the simulation model of Vietnam.















RESULTS


Evaluation and Simulation of Models


Three models of mangrove ecosystems are presented with

numerical evaluations of the outside energy sources, the compart-

ments, and the pathways. The models were conceived as a means of

showing the feasibility of modeling as a descriptive and to some

extent, also a predictive tool in the management of mangrove

ecosystems. Simulations looked at the effects of herbicides, the

effects of hurricanes, the effects of nutrients,and the effects of

economic development in mangroves. The results are presented in

graphical form as families of curves that hopefully cover the range

of conditions that could occur in a mangrove forest affected by

management decisions or natural perturbations.



Model of Herbicides and Mangroves in South Vietnam


Using the symbols described in Appendix A, a model (Figure 4)

was constructed showing the relationship between herbicides, wood-

cutting, land, mangrove biomass,and mangrove seedlings in the Rung-

Sat district of South Vietnam. In the Rung-Sat R. muconata was the

dominant species prior to spraying because of its value for charcoal.

Therefore, the results of this model apply to this species.
























0 CU

4-E C





-C)



uO
.-C-

















5o
r- O
5 OC








) E
C
















-> C





S0
o
-C C





+-' U-














., C
0 u





































CCC-
4- O













C EU
*-' =-








+ ci rr













&-4
-S0 --














CU- C
0C--
LJ"
m4iC i} 4
'r 0 '
>-L t *d


S CD

C- cr


II l- I i






C .,


0 '-

c-i IC 4-

ci Cr Cr O



II ii II II


-Cr Cr -Cr -D



































Ld
I-



C-
0






51



In Figure 4 the sun was an outside driving force that inter-

acted with the amount of land covered by the mangrove trees. In

this model the amount of solar energy available to the mangroves was

50% of the incoming solar radiation. Plants utilized the short-

wave radiation portion of the energy spectrum in photosynthesis and

shortwave radiation accounted for 50% of the total energy in electro-

magnetic radiation. The flow of the sun-land interaction was the

primary production that contributed to measurable aboveground biomass.

This pathway of organic matter production flowed into mangrove bio-

mass and increased as more land was colonized by mangroves. The

organic matter produced through photosynthesis was used by mangrove

vegetation in growth and metabolic activities. Woodcutters once played

an important role in the mangrove forests of the Rung-Sat. In the

model woodcutters cut mangrove wood at a rate proportional to the amount

of mangrove biomass and the number of woodcutters.

The flow of organic matter to the production of mangrove seed-

lings was simulated as a seasonal occurrence controlled by the sun

(see Table C-1 and Figure C-1). The time of the year for seed produc-

tion and seedfall was taken from Vu Van Cuong (1964). In May of each

year a switch in the model was closed and the mangrove trees began to

use some of the primary production to grow seedlings on the trees.

The number of seedlings growing on that portion of the mangrove tree

extending out over the waterway was chosen as a state variable.

Growth of seedlings continued until about October when the growth

switch was opened and another switch closed and the seedlings began










to fall from the trees. The dropping of seedlings was estimated to

take about 60 days. When seedlings fell,some remained beneath the

parent tree and the remainder were carried away by the tide. The

number washed away varied with the tidal energy and the position of

the tide at the time the seedlings fell. In the model seedlings

carried by tidal or river currents to other areas were considered

a state variable and were labeled as seedlings in the water. Some

of these seedlings may eventually colonize an area devoid of man-

groves. This process of colonizing was shown as an interaction be-

tween seedlings in the water and bare land to give a flow to land

covered by mangroves. Seedlings from other areas were shown as a

pathway into the storage of seedlings in water.

Extensive spraying with herbicide was shown as an outside

driving force acting as a stress to convert land covered by mangroves

to bare land because most or all of the mangrove vegetation was killed

by the herbicide. Normally, the dead trees would remain standing and

decay would gradually occur. However, in the Rung-Sat woodcutters

generally harvested the dead trees and hauled them away, thus creating

a landscape dotted with tree stumps. A pathway was also put in the

model that represented planting of mangroves by man to accelerate the

recovery of mangrove trees in the Rung-Sat. Figure 5 is identical to Fig-

ure 4 but includes the initial values used for the state variables,

outside driving forces, and pathways of the model. The data and calcu-

lations are given in Table C-2. The values represent totals for the

Rung-Sat. The model and its calibration were done as part of the


























'ci
U
I-











E


4-'
'ci







4-' 4-'







4-'


sUo

4-' 4-
4-
-U Y
V)







C')







>Cl)
m
a5--
a -o

ci)l












Ci) -d


0
4-'
00

4-






Us-

a)d

-a)

tC)4-

Ea',



C-z






-ci

owa


zn


C)

cci
5-
0)







54








o,



2 G

7;, V)
to (o2



4) x

LD




ccc:
cy on O D


C: co)

~D 1 L" V)v



x 0 0-

C r



to




0,0




Ol

VU)










work of the committee contracted by the National Academy of Sciences

to study the effect of herbicides in South Vietnam.

At the time this modeling was being done in the summer of 1972,

the problem of security made it virtually impossible to obtain the

field data in South Vietnam that was necessary for the simulation of

this model. Therefore, values were obtained from previous mangrove

studies and from the abbreviated field studies.

Mangrove biomass and productivity values needed for the simu-

lation were obtained from research studies by Golley, Odum,and

Wilson (1962) in Puerto Rico, since these forests seemed to be similar

to the heavily managed Rung-Sat mangrove forests. Solar radiation,

sprayed land area, water areas,and seasonal events of seedlings were

available for the Rung-Sat district. Number and biomass values for

seedlings in trees were measured in Rookery Bay, Florida, for this

dissertation. Table C-2 presents the data used to compute the average

values used in the herbicide and mangroves model (Figure 4).

The differential equations that describe the model are shown in

the legend of Figure 4. The calculations of the rate coefficients are

presented in Table C-4. The scaled differential equations are given

in able C-5. The analog circuit diagram for the simulation model

is given i;i Figure C-2. The symbols will not be explained since many

analog computer texts are available. Simulations were run on an

Electronic Associates, Inc. analog computer, the EAI 680.

The impacts of variations in pathways on the amount of mangrove

biomass were simulated. The pathways varied were intensity of










herbicide spraying, productivity, woodcutting rate, seedling

availability,and reseeding rate.



Intensity of herbicide spraying


The effects of herbicide spraying were simulated using the

forcing functions in Figure 6. Five intensities of herbicide spray-

ing were simulated for a five-year period, approximately equivalent

to the number of years of extensive spraying of the mangrove forests

of South Vietnam (period from 1965 to 1970). Curve A simulated the

highest intensity of spraying. At this level the maximum simulated

rate of application for a given year was 1.1 million liters with 3.5

million liters over the entire five-year period. Other curves in

Figure 6 illustrate lesser intensities of herbicide spraying. The

actual amounts of herbicide sprayed on the mangroves in the Rung-Sat

during each of the years were calculated and are plotted as circles

in Figure 6. The highest actual rate of application for any one year

was 1.58 million liters in 1968 and the total sprayed on the Rung-Sat

was 3.9 million liters. Although curve A approximates the situation

for the Rung-Sat, other curves of Figure 6 may approximate responses

of other areas of mangroves in South Vietnam where lower amounts of

herbicide were applied.




















Figure 6. Variation of simulated herbicide application rates during
a five-year period of spraying.


















Figure 7. Steady state levels of mangrove biomass attained by the
Rung-Sat mangrove forest in South Vietnam at four simu-
lated rates of primary production (initial rate of wood-
cutting was 170 g/m2.year and initial mangrove biomass
was 5000 g/m2).


Gross
Gross
Gross
Gross


primary
primary
pri mary
primary


productivity was
productivity was
productivity was
productivity was


2
3.5 g/m 2day
7 g/m day
14 gq/m2day
19 g/m2.day






58












0


r- A 3x105
Q I01x106

S- 2x105
0G B
m .5xl06 '
rw 0c 1x105
D E
0 1 --- 10
0 5 10 15 20
(1960) TIME,years (1980.






cO
< cl a
240
SE IOxlO4
10,000 -
a D
> C
oE
r o 5x104 c
D 5,000 B 0

S L__ A

0 O 0
0 25 50 75 100
TIME,years










Steady state levels


Figure 7 shows steady state biomass values attained by the man-

groves without spraying and with woodcutters initially harvesting 3%

of the primary production per year. Curves are given for four differ-

ent rates of gross primary productivity (3.5, 7, 14,and 19 g/m2 day).
2
From an initial biomass of 5000 g/m2, steady state was reached within

12 years for the rates of gross primary productivity used in the

simulations.

Figure 8 shows the effect on biomass values of changing the

woodcutting rates. In Figure 8a a high rate of primary production

was used and curves are given for five woodcutting rates. Without

cutting mangrove biomass reached a steady state value of 7200 g/m2

Increasing initial cutting rates reduced the steady state biomass

values. The highest rate of cutting reduced the biomass value to

1500 g/m2 (Figure 7a curve E). In Figure 8b a low rate of primary

production was used and without cutting biomass decreased to a steady

state value of 4000 g/m2. The highest cutting rate reduced the

steady state mangrove biomass value to 900 g/m2. Given in Table 2

are the initial and steady state cutting rates. Note that high cutting

at the start was not sustained due to the decline in mangrove bio-

mass.





























Figure 8. Simulated effect of woodcutting on the mangrove forest of
the Rung-Sat district in South Vietnam; (a) primary pro-
duction was 14 g/m -day; (b) primary production was 7 g/m2.
day.

A. Initial rate of woodcutting was zero
B. Initial rate of woodcutting was 170 g/m2-year
C. Initial rate of woodcutting was 1700 g/m2-year
D. Initial rate of woodcutting was 3400 g/m2.year
E. Initial rcte of woodcutting was 17,000 g/mL year





















15,000



E
0 10,000-
m

o E
0 5
o 5,000-
z


0
C


U)
c)
m 3

> V"
o E
crco

c[
z


A
f B
C
D
K. E


25 50
TIME,years


10,000-



5,000- A


DE
O-
0 25 50 75 10
TIME,years


-0Oxl04 P
0
0
a)


-5xi04 c
CL


lOxiO4




-5xl04


I






















C
4-
















- Z
o,














4-
r
















4--'


0
3








































45
0


































4L-
0*
Os
0 L

4- n
0 0
4-,






4) c
S5- T












0








4r-
E'-












-4-'
0 3













-r' 0


+ 0-
-0



















0
I-US


4-



C)

7












O
u



o -













F-













n n














0*

CM
4-
or-



0
u-














0






o m
3

0 -
5-
0-









4-







4- E

3 0n
O0

5-
0_


0 0 0 0 0C
*t 0 0 0
CM C) C) 0C










0 ro o
D CM CD


































0o o
r 03 (O 0
,- CM























(D CD 0C






















0 0C 0 CD 0








-0 CO 0 W









QJ (D (1 a) H)

S- S- 5- S- S-










Effects of herbicide spraying


Figures 9, 10,and 11 show the effect of herbicide spraying on

mangrove land areas and mangrove biomass at three rates of primary

production. These rates were: Figure 9, 3.5 g/m2 day; Figure 10,

7 g/m 2day; Figure 11, 14 g/m 2day. As expected, when spraying

occurred, land covered with mangroves was converted to bare land.

The extent of destruction varied with spraying intensity. With

low intensity spraying the amount of bare land produced during five

years of spraying was small. Reestablishment of the mangroves on bare

land was much more rapid at the less intensive spraying rates and

higher rates of primary production. Recovery was extremely slow at

high spraying intensity and low primary production.

Figure 9a suggests that when rates of primary production were

low, high intensity herbicide spraying (curve F) resulted in almost

total devastation of the mangrove land. Ninety years after spraying

stopped, 45,000 ha still had not recovered. Figure 10 shows that where

mangrove primary production is twice the rate of Figure 9, the maximum

spraying rate destroys 90% of the land, but recovery was complete 90

years after the halt of spraying. In Figure 11, primary production

was doubled again causing the amount of destruction to remain the same

but the speed of recovery was faster. Similar trends were noted for

mangrove biomass as shown in Figures 9b, 10b, and llb. For example,

at the highest herbicide spraying intensity, recovery of mangrove

biomass at a low primary production rate required more than 100 years

(Figure 9b, curve F).






























Figure 9. Simulated effect of herbicide spraying on (a) mangrove land
and (b) mangrove biomass in the Rung-Sat district in South
Vietnam (initial rate of woodcutting was 170-g/mL year and
rate of primary production was 3.5 g/m2.day).

A. No spraying
B. Total herbicide sprayed was 0.16 x 106 liters
C. Total herbicide sprayed was 0.45 x 106 liters
D. Total herbicide sprayed was 0.89 x 10 liters
E. Total herbicide sprayed was 1.66 x 106 liters
F. Total herbicide sprayed was 3.50 x 10 liters



































0 25 50
TIME, years


(a)


25 50
TIME, years

(b)


225,000



150,000



75,000


75 100


JOx 104





5r 104
a.


75 100


4
SE 10,000



03
LaJ


I 5000




























Figure 10. Simulated effect of herbicide spraying on (a) mangrove land
and (b) mangrove biomass in the Rung-Sat district in South
Vietnam (initial rate of woodcutting was 170 g/im year and
rate of primary production was 7 g/m2-day).

A. No spraying
B. Total herbicide sprayed was 0.16 x 106 liters
C. Total herbicide sprayed was 0.45 x 10 liters
D. Total herbicide sprayed was 0.89 x 10 liters
E. Total herbicide sprayed was 1.66 x 10 liters
F. Total herbicide sprayed was 3.50 x 10 liters







67











Q Herbicide
S 900 pulse -225,000




S 600 150,000





I II
S300 / / / 75,000





0 25 50 75 100
TIME, years


(a)







' a- iOx104
s u
o 10,000
SHerbicide
L pulse

E A B C 5x0l40
S5000
IIa




0 25 50 75 100
TIME, years





























Figure 11. Simulated effect of herbicide spraying on (a) mangrove land
and mangrove biomass in the Rung-Sat district in South
Vietnam (initial rate of woodcutting was 170 g/m -year and
rate of primary production was 14 g/m *day).

A. No spraying
B. Total herbicide sprayed was 0.16 x 10 liters
C. Total herbicide sprayed was 0.45 x 106 liters
D. Total herbicide sprayed was 0.89 x 106 liters
E. Total herbicide sprayed was 1.66 x 106 liters
F. Total herbicide sprayed was 3.50 x 10 liters





















O pulse
S1 900 -p 225,000
A

B
S 600 150,000


o 0
z 300 75,000

F E


0 25 50 75 100
TIME, yeors

(a)








SHerbicide
pulse
S- 10xo104
SOxlO
SE 10,000
m A B C



z 5,000 S O

E
F

0 25 50 75 100
TIME, years










Effect of woodcutting


Figure 12 shows the effect of a higher rate at which wood-

cutters harvest the mangrove trees using a high rate of primary pro-

duction. The cutting effect on recolonizing the land area was not

very large but mangrove biomass was reduced. Mangrove biomass values

were lower at steady state and the time to return to steady state

after spraying was halted was 20 years longer when woodcutting was
2
increased from 170 to 1700 g/m2 year (Figure 12b, curve F compared

with Figure l1b, curve F).



Effects of seedling availability


Figure 13 shows the effect of seedling availability on the

rate of recovery at high intensity herbicide spraying and high rate

of primary production. The time required for mangrove recovery was

increased as seedlings were made less available during a given year.

Figure 14 shows the effect of artificial planting of seedlings

on the recovery of sprayed mangrove land. Given are successful plant-

ing rates of 37 and 185 seedlings/ha-year in addition to natural

regeneration. As expected, the model indicated that when mangroves

were planted the recovery of the mangrove land to its original state

was accelerated. Success in planting of mangrove seedlings was

estimated to be 10% of the seedlings planted survived the first year

after planting.





























Figure 12. Simulated effect of herbicide spraying on (a) mangrove land
and (b) mangrove biomass in the Rung-Sat district in South
Vietnam (initial rate of woodcutting was 1700 g/m2iyear and
rate of primary production was 14 g/m6-day).

A. No spraying
B. Total herbicide sprayed was 0.16 x 16 liters
C. Total herbicide sprayed was 0.45 x 10 liters
D. Total herbicide sprayed was 0.89 x 10 liters
E. Total herbicide sprayed was 1.66 x 10l liters
F. Total herbicide sprayed was 3.50 x 10 liters



































0 25 50
TIME, yeors


(a)


25 50
TIME, years

(b)


225,000



150,000



75,000


75 100


10 104





5Sa0


75 100


S10,000
CE



0E
| 5,000
z





























Figure 13. Simulated effect of seedling availability on the rate of
recolonization by the mangroves of the Rung-Sat district
in South Vietnam (rate of primary production was 14 g/mL*
day and total herbicide applied was 3.5 million liters).

A. Normal seedling availability
B. Seedling availability reduced 50%
C. Seedling availability reduced 75%












































V)

o 10,000


> a
0
5 5,000
z
<


25 50

TIME, years


10.104





5 104


75 100









In Figure 14a the recovery time was reduced from 80 to 12 years

by using a successful planting rate of 185 seedlings/ha-year. In

Figure 14b, recovery time was reduced from 40 to 12 years with a

successful planting rate of 185 seedlings/ha-year. In Figure 14c,

at an increased woodcutting rate, the recovery time is reduced from

55 to 12 years with a successful planting rate of 185 seedlings/ha-

year.



Model of Mangroves, Nutrients, and Hurricanes


A simplified model depicting the interactions of a mangrove

swamp with tides, nutrients, and hurricanes is given in Figure 15.

In the model solar radiation interacts with nutrients and mangrove

biomass to produce organic matter for use by the mangrove trees.

The rate of gross primary production was expressed in this model in

units of g C/m2-year.

The organic matter produced is partly respired by the trees

and partly stored as biomass in the form of leaves, wood, flowers,

seeds, and roots. Some of the biomass is shown depositing on the

forest floor as detritus. In the model detritus was generated from

mangrove litterfall. Detritus was also affected by export from the

forest floor into the estuary due to tidal exchange, by grazing, by

decomposition in situ or by accumulation as mangrove peat. Detrital

decomposition occurred under both dry and wet conditions. Nitrogen

and phosphorus were the nutrients included in the model. The major




























Figure 14. Simulated effect of seedling planting by man on the rate of
recolonization by the mangroves of the Rung-Sat district in
South Vietnam; (a) initial rate of woodcutting was 170 g/m2.
year, rate of primary production was 7 g/m -day and total
herbicide applied was 3.5 million liters; (b) same as (a)
except that the rate of primary production was increased to
14 g/m2.day; (c) same as (b) except that the initial rate
of woodcutting was increased to 1700 g/m2-year.

A. No planting
B. Successful planting rate of 37 seedlings/ha-year
C. Successful planting rate of 185 seedlings/ha-year




















10,000



5,000


E 10,000


0
g 5000
z


10xO4
Herbicide
pulse

5x104 |
C 0-




0 25 50 75 100
TIME, years


Herbicide
pulse

C



TA


IOx 104




5xi0 0
a.


I I I I

0 25 50 75 100
TIME, years


0
o
E 10,003



5 o 5,000
r


Herbicide O ~x
- pulse


C --1 5xl04 1?





25 50 75 100
TIME, years


I


I






















C







r-














C
0
5-

























I,
4 1


Cv -
-c
4-,


























=)i CM
C a 0

On a
f In



C-4 C
O z






E p
on I











a m CY
L I C I 0C
O- -C\J













41 c
--I I I 3 .




SCM Ca
4- a





ca-




0 U-
L C '
raC ^-c
E t Co' C-
ro y -













sources of these nutrients are terrestrial runoff, tidal exchanges

twice a day, rainfall,and sediments. In the model these sources were

all grouped as a single nutrient source. Some of the nutrients were

not used by the mangroves because they were quickly bound to the soil

either in the mangrove forest or in the open estuary. A portion of

the remaining available nutrients was used during photosynthesis.

An important influence on the mangrove forest has been the

periodic occurrence of the hurricane which affects the mangrove eco-

system through its often destructive winds and its high wave and tidal

energies. High wind speeds strip leaves from the trees and may

break or uproot trees. Hurricane tides inundate the prop roots making

gas exchange between the root system and the atmosphere impossible

and therefore killing the trees. Craighead and Gilbert (1962) reported

that the effects of Hurricane Donna on the mangroves of south Florida

were variable. In some areas 50-75% of the mature mangroves were

killed. A forest that had sprouted in a mangrove area previously

devastated by a hurricane in 1935 was wiped out. In the Ten Thousand

Islands, the outer mangroves were the most severely damaged by Hurri-

cane Donna. In the model a hurricane pulse was initiated one or more

times during the simulation run. The modeled effect of the hurricane

was to convert mangrove biomass into detritus and to export detritus into

the open estuary.

The mangrove research project of Lugo and Snedaker (1974a)

supplied productivity, biomass, and litter fallvalues. A study by




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

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