Hierarchical control of coral reef ecosystems

MISSING IMAGE

Material Information

Title:
Hierarchical control of coral reef ecosystems
Physical Description:
xvi, 219 leaves : ill. ; 28 cm.
Language:
English
Creator:
McClanahan, Timothy Rice, 1957-
Publication Date:

Subjects

Subjects / Keywords:
Coral reefs and islands -- Africa   ( lcsh )
Environmental Engineering Sciences thesis Ph. D
Dissertations, Academic -- Environmental Engineering Sciences -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 206-218).
Statement of Responsibility:
by Timothy Rice McClanahan.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001659153
oclc - 24529517
notis - AHX0880
sobekcm - AA00004001_00001
System ID:
AA00004001:00001

Full Text









HIERARCHICAL CONTROL OF CORAL REEF ECOSYSTEMS


By


TIMOTHY RICE McCLANAHAN












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


UNIVERSITY OF FLORIDA


1990













ACKNOWLEDGEMENTS


Research presented in this dissertation included a 3 year
period of field data collection while working in East Africa and a 2-
year period of data analysis and modelling undertaken at the Center
for Wetlands, University of Florida. The research in East Africa was
supported financially from the East African Wildlife Society (EAWLS),
the Kenyan-Belgian Program in Marine Sciences and the
International Union for the Conservation of Nature and Natural
Resources (IUCN). Research clearance was provided by Kenya's Office
of the President. Logistical support was provided by the Kenya
Marine and Fisheries Research Institute, Friends World College, the
School for Field Studies, Stanford's Center for Conservation Biology
and the Kenyan-Belgian Program in Marine Ecology. Field assistance
was provided by numerous students and colleagues including L.
Anderson, G. Izard, J. Kurtis, J. Mutere, N. Muthiga, D. Obura and S.
Shafir. S. Shafir and J. Kurtis also provided simultaneous analysis of
some results and discussions which proved fruitful in detecting
errors and checking ideas. My wife, N. A. Muthiga, constantly
provided support, ideas and skepticism. During the data analysis and
modelling the Center for Wetlands provided logistical and technical
support. Dr. H.T. Odum and my committee members L.D. Harris, C.S.
Holling, F. Maturo and C. L. Montague provided the opportunity,








support, confidence and encouragement to see this research through
its final stages.














TABLE OF CONTENTS

page

ACKNOWLEDGEMENTS.................................................................................... ii

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

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

ABSTRACT........................................................................................................... xv

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

Questions of Hierarchical Control................................... .......... 1
Design of Self Organizing Systems................................................. 5
Emergy, Transformity and Hierarchy............................ ........ 7
Community Structure and Theories of Diversity...................... 8
Competition, Predation and Disturbance........................... 8
Structural Complexity.......................................... ................ 10
Population Regulation............................................... .................... 11
Role of Fishing.................................................................................. 1 5
Modelling............................................................................................. 16
The Coral Reef Environment....................................................... 1 7
Productivity.............................................................................. 19
Calcium Carbonate Structure.................................... ......... 20
Kenyan Study Sites........................................................................ 21


METHODS....................................................................................................... 24

Plan of Study...................................... ............................................... 24
Measurements of Community Structure.............................. ....... 25
Substrate Complexity........................................................... 25
Sea Urchin Populations.............................................................. 26
Sea Urchin Distributions...................................................... 26
Fish Populations...................................................................... 27







Community Structure Data Analysis................................ 28
Field Experim ents............................................................... .................. 29
Measurements of Predation................................................ 29
Tests of Interspecific Competitive Behavior.................... 3 1
Intraspecific Competition and Population Regulation
of Echinometra mathaei .......................................................... 33
Density Manipulation Experiments........................................ 33
Behavioral Studies................................................................. 35
Starvation Experiments....................................................... 36
Population Counts......................................... .......................... 37
Measurements of Ecological Processes........................... ......... 38
Bioerosion.............................. .................. ................................ 38
Calcium Carbonate Deposition............................... ............ 3 9
M odels................................................................................................... 4 0
Energy and Emergy Analysis of a Coral Reef.............................. 41

U LT S............................................................................................................ 4 3

Emergy Analysis................................................... 43
Seasonal Patterns of Energy Flow..................................... 49
Community Structure.................................................................... 5 1
The Fish Assemblage............................................................ 5 1
The Sea Urchin Assemblage.............................................. 62
Sea urchin coexistence............................................... 68
Competition and predation............................... ........ 76
Comparisons of predation on different reefs............ 80
Relationship Between Living Communities and
Substrate Cover and Complexity............................. ...... 80
Differences in Trophic Structure......................................... 94
Population Regulation of Echinometra mathaei....................... 94
Density Manipulation Experiments....................................... 94
Morphological and Physiological Measurements.............. 101
Long Term Population Studies............................................... 110
Calcium Carbonate Balance.................................................. 110
Model Results.................................................................................. 118
Simulation of Competition and Predation with
Minimodels............................... ............................................ 118
Effects of consumption, turnover and minimum
resource requirements on competitive ability
and harvesting............................................................... 129
Coral Reef Ecosystem-Fisheries Model.............................. 135








DISCUSSION .........................................................................................................


Disturbed Hierarchy on East African Reefs................................ 157
Emergy Signature and Dominant Components................ 157
Transformity and Hierarchical Position............................. 159
Territory, Turnover Time and Control Windows............. 160
Hierarchical role of calcareous reef structure.......... 162
Hierarchical Patterns of Space and Species Abundance 162
Mechanisms of Hierarchical Control.............................................. 164
Predator Control of Sea Urchins............................................. 164
Species Diversity Control by Selective Consumption
(Compensatory Mortality)..................................................... 166
Control of Competition Between Sea Urchins and
Herbivorous Fish.............................................................. 168
Control of Substrate Cover and Production....................... 170
Overgrazing and Starvation Without Hierarchical
C ontrol.............................................................................................. 17 1
Control by Recycle and Positive Reinforcement ............. 172
Feedbacks and Compensatory Mortality............................ 173
Resolution of Concepts of Population and Systems
Ecology of Reefs............................................................................ 175
Effects of Overfishing....................................................................... 17 6
Reorganized System of Overfished Reefs........................... 176
Response of Fisheries Models to Overfishing.................. 179
Hierarchical Role of Maximizing Productivity................. 180
Recommendations for Reef Management................................... 181
Sum mary............................................................................................ 183



APPENDIX A COMPUTER PROGRAM OF TURNOVER AND
YIELD .................................................................................................................. 185

APPENDIX B COMPUTER PROGRAM OF CONSUMPTION AND
YIELD ............................................................................................................... 189

APPENDIX C COMPUTER PROGRAM OF ALGAE AND CORAL
COM PE TI ON...................................................... ........................................... 193

APPENDIX D COMPUTER PROGRAM OF SEA URCHIN AND
HERBIVOROUS FISH COMPETITION....................................................... 1 9 6


157








APPENDIX E COMPUTER PROGRAM OF THE CORAL REEF
ECOSYSTEM MODEL.......................................................................................... 200

REFERENCES......................................................................................................... 206

BIOGRAPHICAL SKETCH...................................... 219













LIST OF TABLES


Table Page

1 Emergy analysis of coral reef................................................... 44

2 Emergy analysis of reef biological components................. 47

3 Fish population densities by family and locations............... 52

4 Fish population densities in protected and unprotected
reefs................................................................................................ 5 3

5 Sea urchin population densities by species and location... 63

6 Sea urchin population density and mortality in protected
and unprotected reefs........................................................... 64

7 Sea urchin length-weight relationships and average
w eights........................................................................................... 6 7

8 Morphometric and behavioral data of dominant sea
urchins........................................................................................... 7 3

9 Diurnal behavior patterns of dominant sea urchins............ 75

10 Intra and interspecific sea urchin competition
experiments.............................................................................. 77

11 Competition experiments between Diadema species.......... 78

12 Survival rates of the dominant sea urchin species............. 79

13 Substrate variables in the studied reef locations................. 84


viii







14 Substrate variable comparisons for protected and
unprotected reefs........................................... ....................... 8 5

1 5 Population densities of major sea urchin predators in
protected and unprotected reefs.................................. ..... 92

1 6 Survival of Echinometra mathaei on density doubling
experim ents............................................................................... 102

1 7 Burrow defense behavior of Echinometra mathaei........... 1 03

1 8 Respiration, gonad weights, and jaw weight of
Echinometra mathaei on different reefs and
experimental conditions.......................................................... 106

1 9 Population variables of Echinometra mathaei on three
studied reefs............................................... ................. 108

20 Comparison of substrate changes on Diani reef between
1985 and 1988... .......................................................................... 112

21 Equations used in food chain model.......................................... 120

22 Equations used in coral reef fisheries model......................... 138

23 Calculation of coefficients used in coral reef fisheries
m odel............................................................ .......... ...................... 14 0

24 Literature comparison of algal biomass under sea
urchin and fish dominated reefs................................ .... 169













LIST OF FIGURES


Figure EPAU

1 Overview energy-circuit diagram of coral reef and
coral reef fisheries......................................... .............. .......... 2

2 Hypothetical example of energy and pathways in an
ecosystem .............. .................................. ................................. 4

3 Designs of predator-prey interactions..................................... 6

4 Aerial and profile diagram of reef study sites.................... 18

5 Map of the southern Kenyan coast and study sites ......... 22

6 Seasonal patterns of energy inputs into the coral reef..... 46

7 Seasonal emergy inputs into the Kenyan coastline............ 50

8 Relative importance of studied fish families based on
absolute density basis ........................................ .......... 54

9 Relative importance of studied fish families based on
relative density................. ....... .... ... ............... ...... 55

10 Size class frequency distribution of the measured fish
families comparing protected and unprotected reefs..... 57

11 Cluster analysis and Principal Component Analysis of
studied sites.................................. ............. .... ....... 59

12 Absolute and relative abundance of major trophic
groups in protected and unprotected reefs....................... 60








13 Species-area curves for Kenyan reef lagoon sea urchins... 65

14 Cluster analysis of the sea urchin assemblage..................... 66

15 Estimated sea urchin biomass in six reef lagoons.................. 69

16 Species rank and abundance of the sea urchins................... 70

17 The relationship between sea urchin body size, density
and species rank................................................................... 7 1

18 Biomass distribution of the three sea urchin species
across the Kanamai reef lagoon............................ ............ 7 2

19 Test length-body volume (test + spines) relationships
for three sea urchin species................................... ............ 74

20 Total sea urchin and Echinometra mathaei densities
plotted against relative predation intensity..................... 81

21 Species diversity (Simpson's Index) and species
richness as a function of relative predation intensity.... 82

22 Species diversity and species richness plotted against
Echinometra mathaei density............................................... 83

23 Frequency distribution of topographic complexity
measurements in Kenyan reef lagoons .............................. 87

24 Coral cover versus topographic complexity........................... 88

25 Fish density, topographic complexity and coral cover........ 89

26 Topographic complexity and coral cover plotted against
sea urchin density...................................................................... 90

27 Sea urchin and the herbivorous fish density........................ 91

28 Triggerfish densities versus predation intensity on
Echinometra mathaei and total sea urchin density.......... 93

29 Trophic biomass pyramid for two alternate pathways....... 95








30 Echinometra mathaei densities after density
manipulation experiments...................................................... 96

31 Echinometra mathaei density changes versus initial
densities....................................................................................... 9 8

32 Percent Echinometra mathaei individuals remaining on
coral heads after doubling densities.................................. 99

33 Test size frequency distributions of Echinometra
mathaei before and after doubling and total
reduction experiments................................................................. 100

34 Respiration rates of feed and starved Echinometra
m athaei....................................................................................1 0 5

35 Echinometra mathaei Aristotle lantern allometric
relationships.............................................................................. 109

36 Historical population density changes of Echinometra
m athaei......................................................................................... 111

37 Echinometra mathaei gut and calcium carbonate
evacuation......................................................................................... 1 13

38 Percent calcium carbonate in the gut content of
Echinometra mathaei as a function of their
population density................................................ 115

39 Gross reef accretion, bioerosion and net growth of six
studied reef lagoons................................................................... 11 6

40 Calcium carbonate balance of Kenyan reef lagoons
comparing protected and unprotected reefs..................1 17

41 Variations on the food chain model...................................... 119

42 Predator-prey model with a feedback.................................... 121

43 Relationship between reef productivity and
herbivorous fish biomass in the absence of predators.. 123







44 Relationship between yield of prey and predation level
for different prey consumption levels................................124

45 Relationship between yield of prey and predation level
for different prey turnover and yields.............................. 126

46 Effect of the minimum food level on consumer biomass.. 127

47 Effect of feedback on producer and consumer.................... 128

48 Competition between two consumers with the same
coefficients................................................................................ 130

49 Competition between competitors with different
consumption and different turnover rates...................... 131

50 Competition between competitors with different lower
food resource limits................................................................. 1 33

51 Competition between competitors with different lower
food resources, different turnovers and consumption
rates .................................................................................................... 13 4

52 Competition between two competitors where the
superior competitor is harvested.............................. ..... 136

53 Energy-circuit diagram of coral reef ecosystem-
fisheries model........................... ............................................. 1 37

54 Competition between algae and coral and sea urchins
and herbivorous fish.................................................................... 144

55 Coral reef fisheries ecosystem model output with
piscivores, herbivorous fish and invertivores fished.... 146

56 Coral reef fisheries model output with only piscivores
fished.................................................................................................. 14 9

57 Coral reef fisheries model with piscivores and
herbivorous fish fished....................................................... 150

58 Coral reef ecosystem model with feedbacks........................ 152


xiii








59 Coral reef ecosystem model with feedback from
herbivorous fish and sea urchins.................................... 154

60 Territory and replacement time of coral reef
com ponents........................................... ... .. ..................... 16 1

61 Feedback effect of reef structure on production............... 163

62 Illustration of sea urchin coexistence patterns................. 167


xiv













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


HIERARCHICAL CONTROL OF CORAL REEF ECOSYSTEMS


by


TIMOTHY RICE McCLANAHAN


December 1990

Chairman: Howard T. Odum
Major Department: Environmental Engineering Science


Patterns of hierarchy and control were studied in coral reefs of
East Africa with field experiments, simulation models, and energy
analysis. The affects of reef control by the larger fishes and
calcareous structures were determined by comparing overfished
reefs with reefs protected from fishing. Reefs without normal fish
populations had surges of urchin population growth, more
competition, destructive erosion of calcareous reef structure, and loss
of diversity. Predators controlled many population and inter-
relationships. In particular, triggerfish regulated sea urchin
populations which indirectly controlled their grazing and over-








grazing of corals and algae. Unfished reefs, which had a high density
of triggerfish, had low sea urchin densities ( fish abundance, high reef accretion rates and reef topographic
complexity. Fished reefs, with fewer triggerfish, were dominated by
herbivorous sea urchins (5-20/m2) which eroded reefs and reduced
coral reef complexity. Sea urchins competitively excluded
herbivorous fish by reducing the abundance of algae below levels
necessary to maintain herbivorous fish. The diversity of the sea
urchin guild (about 10 species) was affected by triggerfish through
preferential predation on the dominant urchin Echinometra mathaei.
Simulation models developed from eclectic coral reef data
sources produced similar results to those found in field studies.
Model results suggest that the intensity of fishing and the removal of
high-level consumers effects unfished components and processes.
Model results suggest that the transition from a fish-dominated to
sea urchin-dominated ecosystem is rapid: an example of multi-
equilibria controlled by human fishing. Leaving predators of
invertebrates (i.e. triggerfish) unfished results in the highest
fisheries yields and most intense fishing.
By using the emergy concept (spelled with an "m"), several
kinds of energy that contribute to the reef ecosystem were expressed
on a comparable basis in equivalent units of one kind of energy
(solar emjoules/yr). Total annual emergy indicated that wave
energy was the largest component. Evaluating the position of main
reef components in the energy hierarchy using transformities
(emergy per unit energy) indicated that fish and reef structure were
highest in the energy hierarchy.
xvi













INTRODUCTION


Ecosystems are physical, chemical and biological systems that
are maintained far from thermodynamic equilibrium. Consequently,
the role of producers and consumers and the design of ecosystem
structure that maintains their stability and production away from
thermodynamic equilibrium remains a central focus of ecosystem
science. Coral reefs are among the most complex ecosystems and are
among those with the greatest diversity and abundance of consumer
organisms. The role of consumers in the maintenance of ecosystem
structure and processes may be important but has been poorly
studied. This dissertation explores the factors that control East
African coral reef community structure, diversity and some
ecological processes through field studies, experimentation and
simulation models. Specifically, the study focuses on reef building
corals, algae, sea urchins, herbivorous fish, and piscivorous fish (Fig.
1, Odum 1983).




Questions of Hierarchical Control


Ecosystems are composed of interacting biological, chemical and
physical components that are organized into interacting webs (Pimm















































o-
C








ea











o






0
0)
0
3




&oF~


1 __








1982, Odum 1983). Energy and nutrients transformed or taken up at
the base of the food web can often follow multiple pathways before
being recycled or lost from the ecosystem (Fig. 2). Energetic support
for the food web is provided by the primary producers but energy
may follow multiple pathways dependent on consumer choices. On
the long term, choices may be evolutionarily determined by
predator-prey interactions (Janzen 1980). On the short term, choices
may be more flexible and based on optimal foraging considerations
(Schoener 1971).
As energy is transferred between trophic components the
amount of energy decreases. Yet, energy embodied in each
component increases with increasing trophic level (Fig. 2). Schoener
(1989) evaluated food webs and found, on average, that each trophic
"species" predator has 2 trophic "species" prey. Consequently, the
number of embodied connections or pathways increases with
increasing trophic position. The potential for control may increase
with increasing trophic level due to the increase in embodied
pathways and the greater choices and flexibility of top-level
consumers. A hypothesis of ecosystem organization is that as energy
embodied per trophic level or species increases so does the potential
for control of the ecosystem (Odum 1983). Decisions about prey
choice and consumer feedback at the top of the trophic pyramid can
influence the abundance of specific organisms and interactions
between and within trophic elements. These choices may have
surprising outcomes due to the complexity of interactions beneath
the trophic elements. Multiple pathways within ecosystems permit



































L0
0
I.


A BC













0 A B C
ConneEnergy Production
n=11


o n=3

o A B C
Connections


B C
Emergy
n=23


A B C
Embodied
Connections


A B C
Transformity
n=23




n=10

n=3.67



A B C

Embodied
Connections/ Element


Fig. 2. Hypothetical food web with 3 trophic levels (A, B, C).
The energetic and connections (nonheatsink) for each level are
analized. Energetic histograms show the energy, embodied
energy and transformity (emergy/energy) of each component.
Connections shows total connections, embodied connections and
connections per element.


-I








ecosystems to have multiple states dependent on control processes
(Holling 1973).




Design of Self Organizing Systems


In ecology, as in the science of general systems, major
unsettled questions concern the systems designs that emerge from
the self organizational process. Answers to these questions about
organization come from study of the parts, their inter-relationships,
field experiments where manipulations result in changed
relationships or through computer simulation experiments of
different designs and intensity of interactions. Many of the
controversies in ecology about population regulation and ecosystem
design can be summarized by systems diagrams which avoid
definitional or semantic arguments.
Figure 3 presents four possible designs of consumers and
producers that may occur in ecosystems. The simplest and perhaps
most frequently described interaction (i.e. Lotka-Volterra predator-
prey interactions) is a consumer that simply removes a certain
fraction of its prey (Fig 3a). Yet, consumers may have different
effects on their producers that may feedback on production
processes. Consumers may preferentially select different species,
guilds or parts of an organism (i.e. leaves versus stems) which gives
a competitive preference to unselected component (Fig. 3b). This
relationship can be described as compensatory as it gives one
component a competitive advantage over another which can






















Ji h ~.
I, E ~4)
~m, .
"Uc
91C0

II 1' Yc

dCu



t 4) )I -L:


~3o o


6) U





o~ Y


hi4)o e )

0. ) o

I' U
v 4) 4) hi
Cu~ Cu h ..








compensate for its competitive inferiority in the absence of
consumers. Additionally, consumers can affect production through
mineral recycling (Fig. 3c) or through direct effects on producer
production processes (Fig. 3d). All of these designs may result from
self-organization but their relative importance requires further field
studies of ecosystem design and experimentation.




Emergy. Transformity and Hierarchy


One method to describe the hierarchical structure of an
ecosystem is by calculating the energy used in generating an
ecosystem element. This measure has been given the name energy
which is a measure of the energy used to generate an element,
usually expressed in emjoules (embodied joules). Transformity is the
ratio of emergy divided by the actual energy of a component and is a
measure of a component's hierarchical relationship with other
elements. The higher the transformity the greater the energy used
to generate the element and the higher the element in the hierarchy.
Calculating the transformity of components allows one to determine
the hierarchical structure of the ecosystem. As in Figure 1, diagrams
are organized from left to right with low transformity elements on
the left and high transformity elements on the right.
In order to obtain a broad overview of the physical forces and
the hierarchical arrangement of the coral reef, an energy/emergy
analysis was performed using data from the East African region
(McClanahan 1988). The main energy flows in the coastal marine








environment: sunlight, wind, rain (physical), currents, tides and
waves make the high concentration of reef organisms possible. Their
actions support and effect circulation, photosynthesis, and calcium
carbonate deposition. Adey (1987) has emphasized the importance
and the synergistic effect of physical forces in maintaining high reef
production in the low nutrient environment of coral reefs.




Community Structure and Theories of Diversity


Competition. Predation. and Disturbance


The causes of high species diversity in coral reefs has
stimulated vigorous debate (Sale 1980). Early research suggested
that coral reefs, over a long and stable evolutionary period, evolved
resource partitioning mechanisms (Smith and Tyler 1972). Early
studies implicated energy as an important determinant of diversity
as high and stable energy irputs in coral reefs differentiated coral
reefs from other less diverse ecosystems (Connell and Orias 1964)
which may allow greater diversification and specialization of
production tasks (Odum 1963). Subsequent work has challenged this
view and suggested that high species diversity is maintained by
frequent perturbations which keep species from monopolizing
resources and causing competitive exclusion (Sale 1977, Connell
1978). Disillusionment with the competitive resource hypothesis
resulted from the difficulty of measuring interspecific competition
between closely related fish species (Sale 1980). If competition has








created observed species diversity it is difficult to measure at
present. Connell (1980) suggested that utilizing the competitive
resource partitioning hypothesis to explain observed patterns was to
resurrect "the ghost of competition past." Yet, subsequent work in
favor of a "nonequilibrium" view has not been rigorous and is largely
based on observed hurricane induced mortality (Leviten and Kohn
1980), patterns of species richness in calm and disturbed habitats
(Abele 1976) and turnover rates of fish populations (Talbot et al.
1978, Sale 1979). Consequently, factors contributing to coral reef
species diversity are still open to debate.
Whereas competition has been difficult to measure in coral
reefs (although, see Williams 1981, Hay and Taylor 1985, Robertson
and Gaines 1986) predation and herbivory are readily observed.
Biomass and species composition of algae may be greatly affected by
the abundance of herbivores (Hay et al. 1983, Hay 1984, Lewis 1985,
Lewis 1986). Carnivore control of biomass and species composition
has been shown for coral-fish and sea urchin-coral interactions
(Neudecker 1979, Sammarco 1980, Carpenter 1981, Wellington
1982), coral-sea star interactions (Moran 1986), fish predation on sea
urchins (McClanahan and Muthiga 1989) and on gastropods
(McClanahan 1989). Predation may control coral reefs and indirectly
regulate competition through keystone species (Paine 1966) or
compensatory mortality interactions (Connell 1978). Shifts in species
composition have the potential to influence many ecological
processes such as productivity, nitrogen fixation and calcium
carbonate deposition.








To determine the importance of biological control in
maintaining coral reef species diversity, an intensive study was
undertaken on three common sea urchins, Diadema setosum, D.
savignyi and Echinometra mathaei that inhabit reef lagoons.
McClanahan and Muthiga (1988) hypothesized that: 1) E. mathaei is
the top competitor of this guild, 2) E. mathaei is the species most
susceptible to predation, and. 3) that the 3 species inhabit different
microspatial locations in the reef that are maintained by differential
predation on the three species. In the absence of predation the
competitive dominant, E. mathaei, should undergo population
increases resulting in the competitive exclusion of subordinate
species.




Structural Complexity


The reef's topographic complexity.is a notable attribute of coral
reefs that has been suggested to control the abundance and diversity
of coral reef organisms. The combination of coral calcium carbonate
deposition and physical and biological erosion create a sculptured
and complex physical environment. Kohn (1967) suggested that reef
complexity allowed for spatial resource partitioning of species which
maintained species diversity. Subsequent work on fish also indicates
that fish abundance and diversity appear to be affected by reef
complexity (Luckhurst and Luckhurst 1978, Bell and Galzin 1984).
Reef complexity could potentially allow spatial resource partitioning
based on competitive interactions or may simply provide predator








refuge which allows more individuals and species to persist in the
face of intense predation. In order to maintain reef complexity over
the long term, calcium carbonate accretion must exceed erosion.
In this study, rates of bioerosion and the topographic
complexity of reef lagoons were measured in reefs with various sea
urchin densities to test the hypotheses that 1) individual sea urchin
reef erosion rates should increase with increasing sea urchin density
and 2) that reef complexity should decrease with increasing sea
urchin density due to increased reef erosion by sea urchins.




Population Regulation


Population regulation can affect the maintenance of diversity
by the impact of one population on another. In addition, grazing
intensity and prey selection may affect ecological processes such as
productivity and ecosystem structure (i.e. the calcium carbonate
storage of coral reefs). Population regulation of coral reef organisms
can occur due to low levels of reproduction, starvation or predation
during planktonic stages (Doherty 1983), benthic predators
(Wellington 1982, Shulman 1985, McClanahan 1989, McClanahan
and Muthiga 1989), intra- and interspecific competition (Robertson
and Gaines 1986, McClanahan and Shafir 1990), disease (Lessios et al.
1984) or environmental factors such as intense storms (Connell 1978,
Leviten and Kohn, 1980). The model of Shulman and Ogden (1987)
indicates that pre-settlement mortality in benthic habitats is an
important population control only if post-settlement mortality is low.








Yet, what regulates coral reef organisms when predator densities and
postsettlement mortality are low and settlement is high? Do coral
reef organisms have density-dependent population regulation
mechanisms that result in a balance between populations and
resources?
Recent marine benthic population work has suggested that
many species are "recruitment limited" (Doherty 1983, Roughgarden
et al. 1988, Hughes 1990, Karlson and Levitan 1990); which means
that populations are 1) not in equilibrium with their resources (Sale
et al. 1984), or 2) that density-dependent mortality does not occur
(Doherty 1983). In contrast, more traditional population models
suggest that density-dependent mortality due to aggressive
behavior, intra-specific competition, and subsequent density-
dependent mortality can regulate populations (Pearl and Parker
1922, Wynne-Edwards 1965). This mechanism has been suggested
for birds (Lack 1966, Klomp 1972) and other organisms (Wynne-
Edwards 1965, Bustard 1970, Lamincki 1988).
Experimental density manipulations of adult coral reef
organisms (Sale 1976, Williams 1978) have resulted in population
changes. But, Doherty (1983) suggests that population density
changes represent a redistribution of individuals, not mortality or
population regulation at the larger scale, which requires
experimentation focusing on density-dependent juvenile-adult
interactions. Work on juvenile-adult interactions suggests species-
specific density-dependent interactions as experimental results have
shown both positive, negative and no interactions between juveniles
and adults (Doherty 1983, Shulman et al. 1983, Sweatman 1985,








Jones 1987). Research on benthic marine invertebrates also suggests
that recruitment may or may not limit populations depending on the
abundance of larval settlement (Connell 1985, Roughgarden et al.
1988). Levitan (1989) found that size, growth and species biomass
of a common Caribbean sea urchin Diadema antillarum was affected
by the abundance of food. Population densities of D. antillarum may
be regulated by periodic diseases (Lessios et al. 1984), which may or
may not be density dependent, rather than strictly density-
dependent mortality (Karlson and Levitan 1990) due to predation or
competition.
As predation intensity is reduced, does intra- and interspecific
competition become progressively more important in regulating
populations? Are the competitively subordinate species populations
regulated by competitively dominant populations? What will control
competitively dominant species populations when predator
populations are absent or reduced? Echinometra differs from
Diadema as it variably exhibits inter- and intraspecific aggressive
behavior and has the ability to force other organisms out of its
burrows (Grunbaum et al. 1978, Tsuchiya and Nishihara 1985, Neill
1988). The intraspecific aggressive behavior of E. mathaei may
regulate its population densities below the level at which food
resources are limiting.
A number of possible population and biomass regulation
mechanisms are plausible. Perhaps the aggressive behavior of E.
mathaei maintains a constant population density through increased
predation or food limitations on recruits, or perhaps population
density increases in proportion to planktonic settlement such that








food eventually limits individual growth. This would suggest that
biomass is regulated by energetic limitations rather than density-
dependent population control.
In order to test these hypotheses I performed a series of
measurements and experiments to determine the relative
importance of density-dependent population regulation and food
resource limitations. A series of short-term experiments that added
or reduced populations were performed on the reef to determine 1)
the ability for intraspecific behavior to regulate localized densities
and behavior, and 2) the relative rates of predation on recruits
compared to established individuals. Behavioral studies were
undertaken to determine the relationship between population
density and the frequency of aggressive behavior. Additionally,
long-term population and recruitment patterns were measured over
a four-year period.
Reduced food availability and competition for food should
result in feeding adaptations and changes in body condition.
Therefore, feeding and gut evacuation experiments were undertaken
to determine if consumption rates of organic and mineral (calcium
carbonate) matter were density-dependent. Measurements on
respiration, gonad, body size and mouthpart size were made to
determine the effects of body size and population density on
individual morphology and physiology.









Role of Fishing


Despite the importance of coral reef fish as food in many
tropical countries, there are few studies of coral reef fisheries. The
complexity of coral reefs does not allow for simple fishing and
harvesting schemes and as a result most coral reef fisheries are
dominated by small-scale enterprises without research budgets.
Additionally, the diversity of the coral reef fish assemblage has not
attracted traditional fisheries studies based on single-species models.
Finally, the incidence of toxic fish (i.e. ciguatera) makes fishing
enterprises vulnerable to legal action. This discourages large-scale
organizations that are capable of research.
Research suggests that the abundance of preferred game fish
and particularly larger fish may be absent on fished reefs (Stevenson
and Marshall 1974, Bohnsack 1982, Goeden 1982). Additionally,
catch rates may decline (Munro 1983, Koslow et al. 1988), and shifts
in species composition may occur (Koslow et al. 1988). Koslow et al.
(1988) suggest that the complexity of coral reef fisheries may make
them less stable and more vulnerable to overfishing. Sea urchin
(Hay 1984, Hay and Taylor 1985, McClanahan and Muthiga 1989)
and gastropod (McClanahan 1989) population densities may increase
in fished reefs due to predator and competitor reductions. Mass
mortality of D. antillarum in the Caribbean resulted in population
increases of herbivorous finfish (i.e. parrot and surgeonfish) in areas
previously dominated by sea urchins (Morrison 1988, Carpenter
1990). This suggests that sea urchins compete with herbivorous fish
for algal food resources.








The following hypotheses about fishing are tested in this
dissertation: 1) fishing increases sea urchin population density
through a reduction in predators and competitors, 2) sea urchin
population increases result in reduced herbivorous fish abundance
beyond that which is attributable to fishing alone and 3) bioerosion
is greater and reef accretion is lower in fished than unfished reefs


Modelling




To further understand coral reef processes, and to develop a
fisheries management tool, simulation models were developed.
Prior to the development of a coral reef fisheries model, which would
attempt to describe the behavior of the coral reef under different
fishing regimes, some simpler models of predation and competition
were developed to determine the effect of different configurations
and coefficients on the outcome of proposed interactions. Most
models were simple predator-prey or competition models and were
simplifications of the more complex coral reef model.
Models simulated included 1) the effects of a population's
consumption and turnover rates on biomass, and subsequent yields
to the next higher trophic level, 2) the effects of a population's lower
food resource limit on the population's biomass 3) the effects of
population turnover and consumption rates on competitive ability,
and 4) the effects of harvesting on competitive ability. Each of these
models helped in calibrating the larger model, and for understanding








dynamics of competition and predation and the impacts of fishing on
these dynamics.




The Coral Reef Environment


Coral reefs are major marine ecosystems of tropical latitudes.
They harbor one of the world's most diverse species assemblages
(Anderson et al. 1981), have among the highest productivities of
either domestic or pristine ecosystems (Larkum 1983), fix significant
quantities of nitrogen (Wilkinson et al. 1984), deposit 50% of the
ocean's calcium carbonate and associated carbon dioxide (Smith
1978) and contribute 10% of the world's fisheries production (Smith
1978). Coral reefs are frequently located in near-shore
environments between the 2.0oC isotherms in regions without
upwelling and large river discharges. Water temperatures are warm,
water clarity is high and light penetrates to the benthos where most
(> 95%) productivity occurs (Larkum 1983). Physical factors such as
waves, currents, hurricanes, tidal range and temperature fluctuations
are variable among and within regions and sites.
Reefs are frequently divided into three zones; the reef crest (or
edge), the reef flat, and the reef lagoon (Fig. 4). The reef crest is the
most seaward location and is exposed to the full force of waves,
currents and hurricanes. Reef flats are frequently exposed to lesser
extremes of waves and currents but are often exposed to air
depending on their height and the region's tidal range. Reef lagoons
are largely protected from physical forces, but they may experience















































Fig. 4. Aerial (a) and profile (b) view of a typical Kenyan
fringing reef, study sites and transect placement.








temperature and salinity changes if isolated from the ocean during
low tides. Kinsey (1977) suggests that reduced water movements in
reef lagoons reduces productivity and the rate of calcium carbonate
deposition. Reduced impacts of physical forces in reef lagoons
suggest that biological forces may be important controls. Research
reported in this work was undertaken in reef lagoons.




Productivity


Coral reefs are highly productive and near the theoretical limits
of primary production (Larkum 1983). Gross production varies from
2 to 12 gC/m2/day but averages around 8 gC/m2/day (Kinsey 1983,
Larkum 1983). Most studies indicate P/R ratios near 1 although
values both greater and less than 1 have been reported (Kinsey
1983). After coral and algal respiration, the majority (>90%) of net
production is consumed by grazers (Carpenter 1986, Polunin 1988).
The balance of gross production and total community respiration
(P/R=1) may be changed by disturbances such as overfishing.
Measurements of nutrient contents in water flowing over coral
reefs indicates no net uptake of phosphorus (Pilson and Betzer 1973),
but net exports of nitrogen (Wilkinson et al. 1984). Nutrient
enhancement studies indicate some production response to
fertilization but only marginal increases. This suggests that despite
the low nutrient concentrations of ambient waters, coral reefs are not






20


severely nutrient limited (Kinsey and Davies 1979, Williams and
Carpenter 1988). Wilkinson et al. (1984) suggests that grazing fishes
enhance nitrogen-fixing algae by selectively removing nonnitrogen-
fixing forms. Grazing sea urchins have nitrogen-fixing gut symbionts
that increase the nitrogen content of their feces and may, in turn,
enhance primary production (Williams and Carpenter 1988). Hatcher
(1983) suggests that reef productivity may be limited by grazers
rather than any chemical or physical factor.


Calcium Carbonate Structure


Coral, coralline and calcareous algae, molluscs, sea urchins and
other plants and animals remove calcium and carbonate ions from
seawater and deposit calcium carbonate in their skeletons (Chalker
1983). Coral are the most important source of calcium carbonate
deposition in most reefs. Average calcium carbonate deposition rates
lie between 1.0 and 1.2 kg/m2/yr which translates into a vertical
accretion rate of around 0.5 mm/yr (Smith 1983). This calcium
carbonate forms the reef matrix which is bound together by algae,
sponges, bryzoans and other organisms. The most speciose
organismic assemblages are associated with hard substrate formed
from calcium carbonate deposition and its subsequent binding.
Many organisms, both plant and animals, burrow and excavate
the calcium carbonate substrate (Hutchings 1986). The most
important bioeroders include parrotfish, sea urchins and in some
locations sipunculan and polychaete worms (Ogden 1977, Hutchings
1986, Birkeland 1988). Comparisons between sea urchin and








parrotfish bioerosion indicate that sea urchins erode the substrate at
rates 1 to 2 orders of magnitude greater than parrotfish (Ogden
1977, Birkeland 1988). Comparisons of sea urchin bioerosion range
from 0.07 to 1.4 g/urchin/day. This large range may, in part, be due
to differences in measurement techniques. The highest reported
rates by Downing and El-Zahr (1987) used a superior method. At low
latitudes, reefs with low sea urchin densities should have net
accretion rates.




Kenyan Study Sites


The Kenyan coast south of Malindi (Fig. 5) is bordered by a
nearly continuous fringing reef which lies between 100 m and 3 km
offshore. Most areas have a shallow (0.5 to 5 m deep) reef lagoon
that lies between the shore and the reef platform (Fig. 4). Lagoons
are depositional environments dominated by sand and seagrass
ecosystems, but also contain, hard substrate and coral outcrops.
These coral outcrop areas harbor the greatest faunal diversity and
field research was performed in these areas. Physical conditions in
the reef lagoons are generally calm, particularly during low tides
which makes field work relatively easy.
The Kenyan coastline has a variable human population density,
but fishing is common in most locations. Two sections of the reef,
Malindi and Watamu, have been designated as Marine National Parks
(MNP) and have received complete protection since 1968. Fishing in
reef lagoons includes beat-seining, spearfishing, traps and line
















































Fig. 5. Map of the southern Kenyan coast and study sites.








fishing. Fishing is largely part of a subsistence economy with
fishermen selling part of their catch to neighbors and local markets.
Many fishermen travel by foot to fishing locations, most fishermen
lack boats and motorboats are very rare. For comparative purposes
2 protected reef sites (Malindi and Watamu) and 4 unprotected sites
(Vipingo, Kanamai, Bamburi and Diani) were chosen. All research
was done in the reef lagoons in areas dominated by coral outcrops,
usually of the genera Porites or Pavona. Subsequent to this field
work Bamburi was designated as a Marine National Park. All sites
were chosen for their similarity in reef structure, in having shallow
water (<2 m deep), calm conditions and hard substrate.













METHODS


Plan of Study


Research included field work along the Kenyan coast, spanning
3 years, in which basic field measurements of the coral reef
community were made (i.e. substrate, fish, invertebrates) on 6 reefs
which were a priori believed to represent different levels of fishing
intensity. Field experiments were conducted to test hypotheses
generated from field measurements and observations. During 2
years at the Center for Wetlands, University of Florida, an ecosystem
model was developed based on field measurements and a literature
review. Minimodels were developed to simulate the effects of
different configurations of competition, predation, and control, where
calibrations were from coral reef conditions. The effect of coefficient
changes on yield rates to a high-level consumer was a major focus of
simulation studies. Models were used to determine the effects of
changing model coefficients and for calibration of the larger Coral
Reef Fisheries Model. The Coral Reef Fisheries Model is an ecosystem
model which was developed to test the impacts of fishing on major
reef components and ecological processes.









Measurements of Community Structure


Within each of the six studied reef lagoons measurements were
made on substrate cover and complexity, sea urchin species densities
and fish population densities and sizes. Substrate cover and sea
urchin population variables were measured in 1 to 3 randomly
chosen locations per reef lagoon (a total of 14 sites). Each individual
site covered a 30 m x 30 m area. Three parallel nylon lines
separated by ten meters were established in each site (Fig. 4). The
two ends and the middle of each line acted as foci for substrate and
sea urchin density measurements. Consequently, each site consisted
of 9 measurements except one site in Malindi and Bamburi where a
single line was lost, thus reducing the sample size to 6.




Substrate Complexity


Substrate was sampled by a line transect method. A 1 cm by
10 m flexible nylon line was laid perpendicular to and bisecting the
transect line. The distance covered by each category: hard coral, soft
coral, algal turf (microscopic filamentous algae), calcareous algae (i.e.
Halimeda), macroalgae (i.e., Turbinaria, Padina and Sargassum),
coralline algae, coral sand, seagrass and sponge were measured to the
nearest 1 cm and percent cover calculated.
Topographic complexity was calculated using the rugosity
measurement (straight line distance/contour distance). The 10 m
line was pressed against the substrate and allowed to follow the









bottom's contour for the 10 m distance. The straight-line distance
which this 10 m line travelled was measured and rugosity was then
calculated by dividing the straight line distance by the contour
distance.




Sea Urchin Populations


Sea urchin population densities were sampled by
circumscribing a nylon line of known length around the foci of the
lines. Individuals encountered in the circle created by the
circumscribed line were identified (Clark and Rowe 1971) and
counted. Population densities varied by four orders of magnitude
and therefore variable size quadrats of 2, 10 and 25 m2 were used.
Data were all adjusted to 10 m2 areas (the area commonly reported
in other coral reef sea urchin studies) for calculations of density, and
diversity. Nonparametric rank order tests (i.e. Mann-Whitney U-test
and Kruskal-Wallis test) statistics were used. These tests do not rely
on measures of variance which are affected by the above
normalization procedure (Sokal and Rolf 1981).




Sea Urchin Distributions


Within Kanamai and Vipingo a more extensive study was
undertaken of the distribution and coexistence patterns between the
3 dominant sea urchin species Echinometra mathaei, Diadema








savignyi and D. setosum. Measurements included recruitment rates,
distribution patterns and body morphology. Eight 100 m lines
marked at 5 m intervals were laid parallel to shore at 50 m intervals
and were visited and counted at low tides during the day. D.
setosum, D. savignyi and E. mathaei were counted, and their
frequency of occurrence in crevices, burrows, social groups and social
group size were recorded in eighty 25 m2 quadrats for Diadema and
10 m2 quadrats for E. mathaei. Within each quadrat, body lengths,
length of the longest primary spine, and the shortest length of the
inhabited crevice, of up to 5 randomly selected individuals, were
measured with calipers to the nearest half millimeter. Because E.
mathaei has an elliptical shape, both the short and long axis were
measured and the average used in calculations. Body lengths,
weights and volumes were calculated. Body volume was calculated
using the following equation for a half perfect sphere:


body volume = 2/3 x (test length/2 + spine length)3


Fish Populations


Fish populations were sampled by visual counts of individuals
within a 5 m band between the swimmer and the line.along a 100 m
line placed across the reef (Fig. 4). Three to five transects were
made in each lagoon in the general vicinity of the transects made for
sea urchin and substrate cover. The observer swam slowly (20 to 30
minutes/transect), counted all observed fish greater than 3 cm in
length, assigned fish to 5 size categories (3 to 10 cm, 10 to 20 cm, 20








to 30 cm, 30 to 40 cm and > 40 cm), and 10 fish families (families
listed in Result section) or an "others" category if the fish were not
members of the preselected families. The 10 families were selected
from all potential families because: 1) a priori they appeared to be
the most common families based on density observations, 2) they
were important algal grazer families, or 3) they were members of the
sea-urchin predator guild (Randall 1967). Coral outcrops
encountered in transects were circumnavigated in order to count and
identify fish otherwise hidden from view. No fish < 3 cm were
counted in order to reduce errors in density comparisons (Bellwood
and Alcala 1988). Data were analyzed by comparing size-frequency
distributions and densities of protected and unprotected reefs.




Community Structure Data Analysis


Densities, diversity, cluster analysis and Principal Component
Analysis (PCA) were calculated on each set of population data.
Diversity was calculated using a modification of the Simpson's Index
(D) (Simpson 1949) with the following formula:


D = 1 i (ni/Nt)2


where ni is the number of individuals in a species and Nt the total
number of individuals in all species combined. This index results in
a number between 0 and 1, zero being the lowest and 1 the highest
possible diversity. This index was chosen over other diversity









indices as it is easy to calculate, it gives a bounded range of diversity
(0 to 1) and is less sensitive to error in small samples (Routeledge
1979).
Species-area curves were also calculated for sea urchins and
the total number of species was estimated from these curves. Cluster
analysis and PCA were performed to determine the similarity of the
species assemblages in the different study sites. Cluster analysis
used the Bray-Curtis (1957) measure of similarity and average
between-group linkages (Ludwig and Reynolds 1988). Additionally,
scatter-plots and correlations were performed between substrate,
sea urchin and fish variables.




Field Experiments


Measurements of Predation


Relative rates of predation were determined for comparisons
among the common species and among sites. Predation was
measured in each reef lagoon by attaching threaded sea urchins to
nylon transect lines. The tethering technique was introduced by
(Ebert 1965) and developed more fully by McClanahan and Muthiga
(1989). Sea urchins were pierced with a large hypodermic needle
(60mm x 2 mm) and threaded with monofilament line. Urchins were
then tethered to nylon transect line, visited daily and the urchins'
presence or absence then recorded. McClanahan and Muthiga (1989)
found that tagging induced less than 1% mortality; all other mortality









being attributable to predation. Since the technique restricts urchins
outside burrows and affects their normal predator avoidance
behavior, measures of predation have to be considered relative to
the treatments.
Ten urchins were attached at 2 m intervals to each 30 m line
for a total of 30 urchins per site. Sites were visited daily for three
days, removal rates were recorded and the last day an urchin was
alive was used as a measure of survival. Relative predation rates (P)
were calculated with the following formula:


P = (t-x)/t


where x is the average survival in days and t is the total length of
the experiment (3 days). This measure results in a value between 0
and 1 with 1 being the maximum rate of predation. A value of zero
indicates that none of the urchins were eaten while a value of 1
indicates that all urchins were eaten.
Within Kanamai and Vipingo reef lagoons, a series of
experiments were conducted to determine 1) differences in
predation between the three dominant species (E. mathaei, D.
savignyi and D. setosum), 2) the effect of burrow habitation on
survival of E. mathaei, and 3) the effect of social behavior on survival
of Diadema. To determine predation rates between species, threaded
individuals were attached to nylon lines and species were alternated
along the line. A total of 30 individuals per species were tied to lines
in each reef and visited for 5 consecutive days.








The effect of burrow habitation on E. mathaei survival was
determined by placing one group of tethered sea urchins in existing
burrows and another group outside burrows. Both groups were
fastened to the substrate and allowed 15 cm of free line. Sociality
tests were made by tying groups of 2 to 5 individuals per species
together on the same nylon line and allowing each individual 20 cm
of line. Solitary individuals (15 to 29 individuals/reef) alternated
with groups (24 to 42 individuals/reef).




Tests of Interspecific Competitive Behavior


The behavior of urchins that were competing for space was
studied with field experiments on E. mathaei, D. savignyi and D.
setosum. Experiments used an artificial shelter (crevice) constructed
by bending a 50 x 16 cm rectangular piece of sheet metal into a
semi-circle with a 15 cm radius and placing the convex site upward
so that an opening occurred beneath and at both ends. Undersides of
the sheet metal were painted with black polyurethane paint. Since
the species are negatively phototaxic and positively thigmotaxic
(Pearse and Arch 1969) I hypothesized that the species would
compete for space beneath these shelters. Within Kanamai, 10
crevices were randomly placed throughout the reef lagoon. A species
was chosen and placed within each of the shelters, allowed to
acclimate for not less than 5 minutes at which time an additional
animal was placed in 5 randomly selected shelters. Additional
animals were either of the same or of different species. After 15








minutes, crevices were visited and the presence or absence of the
initial and supplemental animals was recorded. The experiment was
replicated at least three times for each possible inter- and
intraspecific interaction. A G-test, which is an improvement on the
Chi-squared test (Sokal and Rohlf 1981) was used to test for
differences between control and experimental crevices.
A series of additional experiments was undertaken to
determine the more subtle interactions that occurred between the
two Diadema species where crevice space was reduced. In order to
test the effect of crevice size on competitive behavior, the above
crevices were halved, experimental individuals were placed
equidistant, but on opposite sides of the crevice entrances, tapped on
the spines in the direction of the crevice and allowed to equilibrate
their positions in the crevice for 3 minutes. After 3 minutes the
amount of test hidden under the crevice for each individual was
measured as well as each individual's test size. Individuals were
considered to "win" competitions if they had a greater percentage of
test beneath the crevice. Experiments included 1) interspecific
competition between randomly chosen D. savignyi and D. setosum, 2)
intraspecific competition between randomly selected D. savignyi,
and 3) preferentially selected D. savignyi with larger test length and
body sizes than D. setosum.






33

Intraspecific Competition and Population Regulation of Echinometra
mathaei


In order to determine the population regulation ability of E.
mathaei, a series of field experiments included short term density
manipulation experiments, long term (about 4 years) population
counts, and density-dependent recruitment, behavioral,
morphological, and physiological studies. Studies were undertaken
on three reef lagoons (Vipingo, Kanamai and Diani) that have
different population densities and represent a continuum of inter-
related factors of predation intensity, frequency of burrow
habitation, and reef topographic complexity. Diani has the highest sea
urchin density, followed by Kanamai and Vipingo. Vipingo and
Kanamai lagoons were dominated by coral heads and E. mathaei most
frequently inhabits crevices within these coral heads whereas Diani
was dominated by coral rubble and E. mathaei are often found
exposed (Muthiga and McClanahan 1987).




Density Manipulation Experiments


Population density experiments were done by adding and
removing urchins from various reef lagoons and monitoring
populations. Populations were monitored by counting individuals on
the tops of small discrete Porites clusters. Within reefs at Kanamai
and Vipingo the tops of circular Porites were randomly chosen and
randomly allocated for either the addition of urchins, the removal of

/








urchins, or left as controls. Individual E. mathaei were counted, the
short and long axes of the "coral head" were measured, circular area
was estimated and population densities were calculated. E. mathaei
were counted daily and followed until density was nearly constant
(i.e. <2% change between consecutive days). The body size of E.
mathaei (long + short axis/2) on haphazardly chosen individuals was
measured before and at the end of experiments to look for body size
changes resulting from the experimental manipulation. Within the
reef at Diani, where coral heads were scarce, 1 m2 areas were
marked off and a population doubling experiment performed.
In a second experiment, individual urchins marked with acrylic
paint (nail polish) were added to coral heads to compare population
changes of new individuals and original inhabitants. Experimental
individuals were collected, dried in the sun (5 to 20 minutes), given
two coats of paint to their spines, dried and added to the
experimental area, doubling the density of marked coral heads. On
control heads, eight individuals were removed and replaced with
marked individuals to maintain original densities. The number of
tagged and untagged individuals in experimental and control
categories were counted on consecutive days. Most (>90%) markings
lasted 3 days before they began to wear off.
To determine the effect of predation on density changes that
occurred during the first two experiments, experimental and control
heads were chosen. Individuals were tagged on each head by
piercing them with a hypodermic needle, threading them with
monofilament and replacing them in their burrows (McClanahan and
Muthiga 1989). On the following day seven more tagged individuals








were added to experimental heads along with untagged individuals
until the density was doubled. Initial inhabitants (on control and
experimental heads) and the additional individuals were strung
together with nylon line, using differently colored lines to distinguish
original urchins from those that were added. Approximately 0.5 m
of line was allowed between individuals. Dead or missing
individuals from each category were recorded daily for 3 days.
Average survival rates were calculated and statistically compared.
Attachment to lines may have decreased survival probability but
mortality due to predation was distinguishable from other mortality
by the animal's test condition. Predator-induced mortality and
mortality attributable to other causes were distinguished in the
analysis. Tethering techniques were used instead of caging because
E. mathaei escapes from cages and Kenya's large tidal range (4m)
damaged cages.




Behavioral Studies


Behavioral studies were undertaken on reefs at Diani and
Vipingo to determine the frequency and types of aggressive behavior
which might be attributable to different urchin population densities.
A modification of the technique of Grunbaum et al. (1978) was used.
Intruder individuals were placed at a host's burrow entrance and the
result of the interaction recorded. Fights were followed for no more
than 20 minutes. Two individuals remaining in the same burrow
beyond 20 minutes were classified as coexistent, although they may









have been fighting beyond the time limit. This occurred infrequently
(n=2) and only at the reef in Vipingo. Within the Diani reef, few
individuals were found in burrows or crevices. Only individuals in
crevices were used in experiments. Between-reef comparisons of
behavioral categories were made with a G-test (Sokal and Rohlf
1981).




Starvation Experiments


To ascertain the response of individuals to starvation, 20 E.
mathaei were placed in an aerated seawater aquarium. Seawater
was replaced every 4 to 6 days and cabbage was supplied for food.
Respiration was measured periodically over 9 days by oxygen uptake
(Winkler titrations; Strickland and Parsons 1972) using rubber-
sealed 1-liter mason jars for 1 hour (6 experimental jars and 1
control jar). After 9 days, the aquarium was divided, and half the
individuals were starved. Periodic measurements of respiration
were continued for 25 days. Cabbage was replaced daily and air-dry
weights were measured before and after immersion. Additional
control cabbage was placed apart from sea urchins. Daily
consumption rates were calculated. At the end of the experiment,
the individuals wet body and gonads were weighed. Respiration
rates in the field, and gonad weights were measured for
haphazardly-selected individuals from reefs at Diani and Vipingo for
two different time periods (same method as above). To obtain
averages for animals that might have differences on a lunar cycle








measurements were made at different stages of the moon. Field and
laboratory respiration rates were measured at temperatures
between 24 and 250C. Test lengths (long+short axis/2), wet body
weights, and lengths and weights of Aristotle lanterns (exposed tooth
+ lantern (jaw)) were measured. Gonad and lantern indices were
calculated (lantern or gonad weight/body weight x 100). The dry
weights of various body components and organic matter of cabbage
were determined from drying at 600C for 3 days. Ashed weights
were measured after 3 hours of combustion at 5500C.




Population Counts


Periodic population density and test size measurements were
made on Kanamai and Diani reefs between 1985 and 1988 (Muthiga
and McClanahan 1987, McClanahan and Muthiga 1988). Wet weight
was calculated from test length measurements with a length-weight
correlation (Muthiga and McClanahan 1987). Dry organic matter
weight was estimated using the above described combustion
procedure data. In July 1988 the density of adult and recruits (test
lengths < 1.5cm) were counted in haphazardly placed 1 m2 quadrats
on the three reefs. Recruits were small and hard to find and despite
thorough searching, counts were undoubtedly underestimated
although relative numbers are probably comparable. Within Diani,
the cover of sand, hard substrate (dead coral), and seagrass cover
was estimated for statistical comparison with 1985 data (Muthiga
and McClanahan 1987) in 90 1 m2 quadrats. Statistical comparisons








used running averages (i.e. x = (x + X(d-1) + x(d+1))/3; d=distance on
transect) of the 90 quadrats. Running averages reduces the variance
between quadrats and was able to detect within reef trends in
density more accurately than data without running averages.




Measurements of Ecological Processes


Bioerosion


Substrate bioerosion by sea urchins may affect reef complexity.
Previous researchers have measured bioerosion rates (reviewed by
Ogden 1977, Hutching 1986, Birkeland 1988), but recently Downing
and El-Zahr (1987) developed a new technique that was used on 3
Kenyan reefs (Vipingo, Kanamai and Diani) to determine rates of
substrate bioerosion.
Sea urchins were removed from their burrows and placed in a
predator exclusion cage (lifted above the substrate) from which 10
urchins were removed every 2 hours for eight hours. Removed sea
urchins were sacrificed, their gut contents removed, dried and
weighed on a triple beam balance. The gut content was then soaked
in 1.13 N HCI to dissolve the calcium carbonate fraction, dried and
weighed again. Plots of time since removal from the substrate and
the remaining gut content give a rate of gut evacuation for both the
organic and inorganic fractions of the gut. A daily substrate erosion
rate was calculated assuming ingestion equals defecation. These








experiments were completed twice in Vipingo and Diani and 3 times
in Kanamai.




Calcium Carbonate Deposition


Net rates of calcium carbonate deposition have been
determined by an alkalinity depression technique (Smith 1978a)
which measures carbonates deposited and/or dissolved from
solution. A rough estimate of gross calcium carbonate deposition
was made from coral cover measurements, known rates of vertical
increase, and coral skeleton porosity. Gross calcium carbonate
deposition was estimated with the following formula.


Deposition = Coral cover x linear growth rate x (1-porosity) x calcium
carbonate density


The linear growth rate of massive heads (more than 80% of the coral
forms found in the study sites) is about 1 cm/yr (McClanahan and
Muthiga unpublished data). Aragonite has a density of 2.9 g/cm3
and coral has an average porosity of 50% (Smith 1983). Therefore,
from the above equation, multiplying coral cover by a factor of 14.5
will give a measure of gross calcium carbonate deposition in the units
of kg/m2/yr.






40


Models


Main features and mechanisms of the coral reef that were
believed most important were combined using energy systems
diagramming. Then, equations implied by the relationships and
connections were written. To calibrate models quantitatively, a
coefficient was calculated for each pathway and listed in the tables of
coefficients. For calibration a value of each storage (state variable)
and each pathway flow was estimated either from field
measurements, a literature measurement from a similar system, or
calculated by difference to make any unknown flows consistent with
measured ones. The mathematical expression for each pathway was
set equal to the pathway flow and then solved for the coefficient. For
example, if


K1*A*B = 20 when the state variables A=2 and B=5


then

kl=20/(A*B) = 20/(2*5) = 2


My method of calculating coefficients assumed that each component
had an upper maximum biomass (with predators absent), at which
point gains and losses were equal. The component's (state variable)
resource was then some fraction of its maximum. Consumption rates
were usually estimated as some percentage of the maximum biomass
as literature values are frequently reported this way. Turnover
rates were assumed to be some daily percentage of maximum








biomass with gains equal to losses at this maximum level. After
initial calibration, graphs of simulated variables were compared with
real data. Models were frequently run by changing coefficients and
then running the model to steady state for a variety of consumer
levels. Steady state values were then plotted for different
coefficients and consumer levels. More complete explanations of this
methodology are given elsewhere (Odum 1983).




Energy and Emergy Analysis of a Coral Reef


Main energy flows used by an East African coral reef were
estimated and compared in a table. Data on sunlight, rainfall and
wave height on a monthly basis were previously summarized
(McClanahan 1988). Tidal data were derived from tide table (Kenya
Ports Authority 1988). Data were available on a monthly and annual
basis.
Energy flows were also expressed in units of Emergy (spelled
with an "m"), which compares each type of energy as equivalents of
one energy type or as the energy of the one type required to produce
the other type. Solar emergy is expressed in solar emjoules, the solar
insolation energy required directly and indirectly to produce a flow
or storage.
Solar transformity is the solar emergy per unit energy
expressed in solar emjoules per Joule. Solar transformities used
were previously calculated from the energy web of the biosphere






42

(Odum 1987). In the emergy analysis table solar emergy was
calculated by multiplying each energy flow its solar transformity.













RESULTS


Coral Reef Emergy Analysis


Emergy Analysis


Results of calculating emergy use by a square meter of coral
reef are given in Table 1. Although there is more solar energy
received than other kinds of energy, on an Emergy basis the
hydrodynamic contribution is much larger. Emergy calculations
using presently known transformities indicate that the emergy
comes from waves. Although there are large currents nearby (Table
1), current emergy absorbed in the ecosystem is less than the wave
energy absorbed. Total solar emergy (item 6, Table 6) from
independent sources was 2.82 E13 sej/m2/year.
With hydrodynamic emergy in excess, direct solar energy may
be limiting as high quality physical energy requires low quality solar
matching emergy for its full effect. The seasonal pattern of emergy
flows in waves and currents (Fig. 6). indicates that peak energy
occurs during the southeast monsoon when wind and wind-driven
processes are prevalent.
An energy/emergy analysis of the reefs main biological
components (Table 2) shows that algae have the highest production
followed by coral, sea urchins, herbivorous fish, piscivores, and









Table 1. Annual Emergy use by a square meter of coral reef
community. Equations used in calculations given in Odum et al.
(1987).


Note Energy Type Actual Energy Transformity Solar
Joules/m2/yr sej/j Emergy
E+10sej/yr

1 Solar Energy 7.12E+09 1 .71
2 Waves absorbed 9.93E+08 25889 2570.00
3 Rain, physical 3.08E+04 15423 .05
4 Tides 1.07E+08 23564 252.00
5 Currents, kinetic 6.62E+08 5981 396.00
6 Total Emergy Used, item 2+4 1.43E+09 2822.00


1. Data from (McClanahan 1988) based on monthly averages from
1963 to 1980
Average Insolation= 170 Kcal/cm2/yr
SI = 170 Kcal/cm2/yr x 10000 cm2/m2 x 4186
Joules/Kcal

2. Data from McClanahan (1988) based on significant wave heights
Average wave height= 1.39 m
Average wave period= 8 seconds
Average reef width= 100 m
WE = 1/8 x 1.025 g/cm3 x 980 cm/s x height2 x 10000 cm2/m2 x
2.38E-11 erg/Kcal x 9.9 m/s x 0.125 s/wave x 3.15E+7 s/year / 100
m/reef

3. Kinetic energy of rainfall, data from McClanahan (1988) based on data
from 1946 to 1980
Average rainfall= 1.06E+03 mm/yr
K.E.= rainfall mm/yr x 0.5 x 1 g/cm3 x 762 cm/s x 2.38E-11 Kcal/erg
x 4186 J/Kcal x 10000 m2/cm2

4. Tidal energy based on 1988 Kenyan tide tables (KPA 1988)
Average tidal range= 2.25 m








Tides are semidiurnal= 706 tides/yr
T.E. = 0.5 x 706 tides/yr x .05 x 1.025 g/cm3 x 980 cm/s x
2.25 (cm)2 x 2.38E-11 Kcal/erg x 4186 J/Kcal x 10000 cm2/m2

5. Current energy, data from Deutsches Hyrdrographisches Institut (1960)
adjusted for near shore region by factor of 0.145 x Ocean current
(McClanahan, unpublished data).
Average current speed =14 cm/s
Average depth= 1.5 m

C.E.=0.5 x 1g/cm3 x 14 cm/s2 x 2.38E-11 erg/cal x 4186 J/Kcal x
depth (cm) x 14 cm/s x 3.1536 E7 s/yr x 1002 cm2/m2
An absorption of 10% of the above energy was used as an estimate of
the amount of kinetic energy absorbed by the reef bottom.

6. Rain, waves, sunlight and currents are all part of the same
geobiospheric process. Therefore, in order to avoid double counting
energy sources only the item with the largest value is included. Tides
are part of another process and therefore tides and waves were
summed in order to calculate the total emergy.


















7e+1 2
S CURRENTS
6e+12 --- WAVES


S 5e+12
U)
4e+12-

LU
S 3e+12
LU
2e+12


le+12


00+0' ,
J F M A M J J A S O N D
MONTH

Fig. 6. Seasonal emergy inputs into the Kenyan coastline.
Waves and currents are the two largest values, are very similar
in magniture, and within the margin of error associated with
transformities. The southeast monsoon occurs between April
and September. SEJ=solar equivalent joules or the amount of
solar energy required to produce this product.









Table 2. Solar transformities of major biological reef components.
Based on 2.57E+13 sej/m2/yr from Table 3.


Energy Type Energy Transformity
Joules/m2/yr sej/j


Algal Production
Coral Production
Herbivorous Fish
Sea Urchin Production
Triggerfish Production
Piscivorous Fish
Fisheries Production


2.14E+08
2.09E+07
1.36E+06
5.76E+06
4.40E+04
1.57E+05
4.19E+04


SeIl


1.20E+05
1.23E+06
1.89E+07
4.46E+06
5.84E+08
1.63E+08
6.13E+08


1. Reef gross production varies between 2 and 12 gC/m2/day with an
average production of 8gC/m2/day (Larkum 1983) and 70% algal
cover
Production= 8 gC x 2.5 gC/gdw x 4 Kcal/gdw x 4186 J/Kcal x
2.5 gdw/g wet weight x 70% cover x 365 days/yr


2. Coral gross production about 5000 kcal/m2/yr (Lewis
1981)


3. Maximum herbivorous fish biomass is 500 kg/ha (Goldman and
Talbot 1976) and a gross P/B ratio of 6.5 (Chartock 1983).
Production=500 kg/ha x 1000 g/kg x .00001 ha/m2 x 1 Kcal/g
x 4186 J/Kcal x 6.5 kg/kg

4. Maximum sea urchin biomass is 500 g/m2 (Muthiga and
McClanahan 1987)
having a P/B ratio of 2.75 (Hawkins and Lewis 1982)
Production= 500 g/m2 x 2.75 g/g x 1 Kcal/g x 4186 J/Kcal

5. Maximum triggerfish biomass is about 70 kg/ha with a P/B ratio of
1.5 (Polovina et al. 1984).


- -7 t- "






48
Production= 70 kg/ha x 1000 g/kg x 0.00001 m2/ha x 1 Kcal/g
x 4186 J/Kcal x 1.5 kg/kg

6. Maximum piscivore production is about 250 kg/ha with P/B ratio
of 1.5 (Goldman and Talbot 1976).
Production= 250 kg/ha x 1000 g/kg x 0.00001 m2/ha x 1 Kcal/g
x 4186 J/Kcal x 1.5 kg/kg

7. Fisheries production approximately 100 kg/ha (Smith 1978)
Production= 100 kg/ha x 1000 g/kg x 0.00001 m2/ha x 1 Kcal/g
x 4186 J/Kcal







49

finally triggerfish. The transformity for reef fisheries production is
6.13E+8 (SEJ/J) solar equivalent joules/joule.




Seasonal Patterns of Energy Flow


The Kenyan coast is dominated by seasonal patterns of the
Intertropical Convergence Zone (ITCZ) and its seasonal migration
back and forth across the equator (McClanahan 1988). Two distinct
seasons are created by this migration, the northeast monsoon (NEM)
that occurs between October and March and the southeast monsoon
(SEM) between March and October. Currents and sunlight are the
largest energy sources followed by waves, tides, and the kinetic
energy in rainfall, which is negligible (Fig. 6). During the southeast
monsoon, currents, waves and rain are more important than during
the northeast monsoon when sunlight is the major energy source.
The northeast monsoon is typified by clear skies when the ITCZ has
migrated away from the equator. Tidal energy is highest during the
two intermonsoon periods. Total emergy (Fig. 7) is highest during
the southeast monsoon and dominated by waves and current kinetic
energy.























70+3
6e .+3 RAIN, kinetic
5o+3*
403-
30*3-
29+3-
19+3 -

J F A J J ASOND


2.0.48
WAVES




1.0J* F A J




O.Oe+O--r--r-i--r-r-r-i-i-i-i
JFMAMJJ ASOND


< ^ -- ^ "


1.0**.9

8.0.4"

L).4

4.0..8


2A 6 4 Li- i-f -I-I -.-I- .-
J FMAM J J ASOND
MONTH


1.0*7
TIDES

9.0G+6 "



8.0+6


7.0.+6 .
J PFMAM JASOND


7.0"


6.0e+8 *


5.0+8"


4.0018 F-- -
J F M A M J JASON D



TOTAL
1.60+9

1.4e9 -

1.2*+9 -

1.0+9 -

8.0" -IMI8 TI I
J FMAM J A SOUND
MONTH


Fig. 7. Seasonal patterns of energy inputs into the coral reef
environment of the Kenyan coastline. For source of data and
calculations see Table 3. Current emergy flux is only partially
absorbed (i.e. <10%) where as much larger fractions of the other
energy sources are absorbed.










Community Structure


The Fish Assemblage


The absolute density of all studied fish families were
significantly different between sites with the exception of the
Diodontidae (Porcupine fish) and the Lagocephalidae (Puffers) which
existed at low (<0.13 fish/100 m2) densities at all sites (Table 3).
The Pomacentridae (Damselfish) was the most abundant family but
the abundance of other families depends upon the protected versus
unprotected dichotomy. Reefs protected from fishing had
significantly denser populations of all fish families with the exception
of the Diodontidae and Lagocephalidae (Table 4). In terms of
density, protected reefs had a predominance of the Pomacentridae,
Acanthuridae (Surgeonfish), Labridae (Wrasses) and Scaridae
(Parrotfish). The Pomacentridae and Labridae were most common in
unprotected reefs; where the herbivorous Acanthuridae and Scaridae
densities are greatly reduced. Data analyzed on a percentage basis
resulted in similar patterns (Figs. 8 and 9) except in Bamburi and
Diani where other families were also important. From observations,
the Holocentridae were common in Bamburi and the Ostraciidae in
Diani. Wilcoxon signed-rank tests comparing family densities
between protected and unprotected reefs indicates rank differences
in families for absolute density values (z= -2.67, p<0.01) but not for
relative densities (z=-1.16).

















* ) *
* Z Z


NC
00

m >
as v

E

C )













on







oc
S0
u





C.6
.2 C


Eu



























acc
'









0S






K *S






SSra








25 S
CO U






v>e*

111 E


Cc



M



c u
cc

S.









E

















co
E
Cu
















03










2
C




































-Cu
"3






>-,


* *
4 4 4


V)
co
o








co
r'



v
Nn
0













o a
o


(S l
V*)
0







o
o


N

o -







0 N
IS"







0





v, +1
It)









(S
E
11-i
0 1 cS


SE-
*o" Cu

22
Cu <
C) U
U,
Cu Cuo


o o d







+1 +1 -
0 0 C ;


1 o N o
0 0 0















0









o o






+1 +1 +1

; o a





o oe 0
4+144141
oo o ". o






c a 0




t 0O 0 0
o O
4141414
00 ~


+H -1




6 w
10

+1 41 +1









+1 +1 -H










N o -













41 +1 +1





c*c





CP

u
- u



CL UC5











Table 4. Density (#/100m2, x + S.D.) comparisons of the major fish
families found within the two protected marine parks (Malindi
and Watamu, n=8) and the four unprotected reefs (Vipingo,
Kanamai, Bamuburi and Diani, n=18). Mann-Whitney U-test of
significance includes U value and level of significance. NS = not
significant


Fish Protected Unprotected Mann-
Family Whitney
U-test

Acanthuridae 14.4 4.6 0.81 0.85 144 p<0.001

Balistidae 0.70 + 0.37 0.07 + 0.12 140 p<0.001

Chaetodontidae 2.03 + 1.69 0.51 + 0.54 123 p<0.005

Diodontidae 0.025 + 0.066 .0.044 0.083 83.5 NS

Labridae 11.4 4.4 8.9 8.5 105 p<0.05*

Lagocephalidae 0.05 0.13 0.06 0.16 73 NS

Lutjanidae 3.2 + 5.1 0.01 + 0.04 144 p<0.001

Pomacanthidae 0.38 0.35 0.04 0.08 117 p<0.01

Pomacentridae 68.7 22.6 19.1 9.2 144 p<0.001

Scaridae 9.3 + 4.9 2.1 + 3.3 132 p<0.001

Others 11.4 + 6.8 .3.8 + 2.3 127 p<0.001

Total 121.6 + 31.4 35.4 + 19.4 144 p<0.001

* Significant only for a one-tailed test







54





0u -
Sa 0 4E






000\0 \ "\\ l *
ALooo V'/ 4'p Q .

cE

.c
000000A a
CO
0**00000
>?a /6 / \/ I\ ^\ /\ \\/J \ %




********* \ \ i i
ee@@OeO 4. r


0000," l* IL/ -



09 2) q, .C-W
I ALEm8



III l ll l ii '
I -t
; *'/;/: /s ^ /^/N^/ .^ /v^ >^ /. / *-'' rW : :' ^
\~~Q %I 'ou
00 ,,.. .. oo'
i4~~~~~~~ ~~~~~~ '''' /- ^ ^<'\ y \ / \ ^ ***'' ^
[H w~ ^/^/^ ^l ^:- S .f:



00 gT 8 8
Allh I 3ALLVT?






























0
a)
U


C)





O
*,
I-







0
Uc







cc



o
GO

C CO


0







Cr
o


4*o
u *
>


AIISN30 3AI1V13&






56

With the exception of the Lagocephalidae there were fewer fish
on the unprotected reef, particularly in the larger size classes (Fig.
10). The Labridae which were only slightly more densely populated
in protected reefs, had smaller individuals in unprotected reefs (G-
test, G=9.94, p<0.05). Other (i.e. unclassified) families were most
important in protected reefs (Tables 3 and 4) and included the
Siganidae (Rabbitfish), Serranidae (Groupers), Lethrinidae
(Snappers), Aulostomidae (Trumpetfish), Holocentridae
(Squirrelfishes) and Apogonidae (Cardinalfishes). Other families in
unprotected reefs included the Holocentridae, Ostraciidae
(Trunkfishes) and Apogonidae. Although a species level analysis was
not quantified, there appeared to be greater species diversity in
protected reefs.
Both Principal Component Analysis (PCA) and cluster analysis
of fish families (Fig. 11) indicate distinct differences between
protected and unprotected reef communities. The first axis of the
PCA separated the protected from the unprotected reefs whereas the
second axis separated the two protected reefs. This indicates that
protected and unprotected reef family composition differ based on
this management distinction. Eighty-nine percent of the variation was
accounted for by these two axes. Cluster analysis also indicates 90%
similarity between protected reefs but 60% similarity between
protected and unprotected reefs. Among unprotected reefs Kanamai
and Vipingo were similar (90%). Diani was the least similar site
among the unprotected reefs.
Fish populations in all trophic levels were less abundant in
unprotected reefs (Fig 12a). The analysis based on relative density

















a 80.4
Acanthuridae Balistidae

E 6 G=1431
p.<0.001 0.3- GO68.1
a p<0.001
4- 0.2-

2- o.1


0 I 0 0.0 -
3-10 10-20 20-30 30-40 >40 3-10 10-20 20-30 30-40 >40

M 1.0 0.03
Chaetodontidae Diodontidae
E 0.8- G=70.1 G=0.8
0 p<0.001 0.02- NS
o 0.6

0.a
0.4 0.01

c 0.2 -

0.0 0.00 -
3-10 10-20 20-30 30-40 >40 3-10 10-20 20-30 30-40 >40


A 8 0.04
E Labridae Lagocephalidae
E G=837 0.03 G=0.1
o p<0.001 NS

4- 0.02-

2 0.01
C

of0E09 0.00
3-10 10-20 20-30 30-40 >40 3-10 10-20 20-30 30-40 >40
Size Class, cm Size Class, cm



Fig. 10. Size class frequency distribution of the measured fish
families in protected versus unprotected reefs. Size class I = 3
to 10 cm, II = 10 to 20 cm, III = 20 to 30 cm, IV = 30 to 40 cm,
V = >40 cm. Dark-shaded bars refer to the protected reefs and
light-shaded bars to unprotected reefs.






58





M 2 0.2
Lutjanidae Pomacanthidae
E Gs810 G=24.1
0 p<0.001 I p<0.001
o
1 0.1"

a-






G=6019 3. G=1989
S40- p<0.001 p<0.001
30 2-






5. n Size Class, cm
1 Others
E 4 G=1136
0 4 p<0.001
'-
C 1



0





"P.


Figure 10. continued


3-10 10-20 20-30 30-40
Size Class, cm
















BAMBUd
0
KANAMAI
0 VIPINGO


-1 -0.6 -0.2 0
3DIANI -0.2

-0.6


-1*


O PROTECTED
0 UNPROTECTED


0.6 1


WATAMU O


1 65%


1.4 1.8


MALINDI 0


24%


PRINCIPAL COMPONENT AXIS I


DIANI

BAMBURI

KANAMAI

VIPINGO

WATAMU

MALNDI


1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

SIMILARITY


Fig. 11. Cluster analysis and Principal Component Analysis
(showing the first two principal components) of the six studied
sites ( o = protected and = unprotected sites) based on family
densities. Cluster analysis uses the Bray-Curtis (1957) measure
of similarity and average between-group linkages.


|






















Fig. 12. Absolute (a) and relative abundance (b) of major
trophic groups comparing protected and unprotected reefs.
Herbivores include the Acanthuridae and Scaridae, Omnivores
include the Pomacentridae, Pomacanthidae and Chaetodontidae,
Carnivores are principally invertebrate-feeders and include the
Balistidae, Diodontidae, Labridae, Lagocephalidae and
Lutjanidae.












60-


40-


0; PROTECTED REEFS
O UNPROTECTED REEFS









^.:.-:


7


a- 4.i I


F/


HERBIVORES


OMNIVORES


CARNIVORES


PROTECTED REEFS


'I--
6 5
.I-
U.
UJ
w
c-


UNPROTECTED REEFS


OTHERS


'''~~


I .. .. I


I I IL


I








indicates fewer herbivores fish and more invertebrate carnivores
(Fig. 12b). Most of the carnivore group feed on invertebrates. The
unclassified group was about 10% by density and mostly composed
of carnivores with the exception of a few Siganidae, Blenniidae
(Blennies) and Gobiidae (Gobies).




The Sea Urchin Assemblage


Density and diversity measurements of sea urchins indicate
large differences between the six studied reefs (Table 5) and
between protected and unprotected reefs (Table 6). Protected reefs
had low sea urchin population densities and diversity (Fig. 13).
Species-area curves for the protected reefs included only the three
species Echinostrephus molaris, Echinometra mathaei and Echinothrix
calamaris. E. mathaei in protected reefs were very small (i.e. <1.5 cm
test width). In contrast, 10 species of sea urchin were found in
unprotected reefs and principally included Diadema savignyi, D.
setosum, Tripneustes gratilla and Echinothrix diadema not found in
protected reefs. Cluster analysis (Fig. 14) of the sea urchin
assemblage showed protected and unprotected reefs as two distinct
groups.
Biomass of dominant sea urchins on each reef was calculated
from average lengths and length-weight correlations (Table 7);
weights of the rarer species were estimated. Much more urchin
biomass (3 orders of magnitude) was found on all the unprotected







63


Table 5. Sea urchin densities (x + S.D.) per 10 m2, the number of sites, sample
sizes and area sampled in the six locations. Kruskal-Wallis test of difference
between reefs and level of significance provided. Survival given days for a
possible maximum of 3 days. NS = not significant.

Kruskal-
Malindi Watamu Vipingo KanamaiBamburi Diani Wallis


Sites n=
Quadrats, n=
Area sampled, m2


2 1
15 9
375 225


3 2
24 18
420 450


12.9 3.4 6.3 0.2
7.6 4.0 6.3 0.5

3.7 2.3 0.5 0.02
3.7 1.9 0.8 0.9


Echinometra
mathaei

Echinostrephus
molaris

Echinothrix
calamaris

Echinothrix
diadema

Prinocidaris
sp.

Stomopneustes
variolaris

Toxopneustes
pileolus

Tripneustes
gratilla


0.2 0.2
0.3 0.4

0.3 0.1
0.4 0.3


0.03
0.10


0.0 0.04
0.13

0.0 0.0


32.6 80.6 10.8 135.0 p < 0.001
20.7 53.9 7.1 123.0


1.1 0.2 10.7 0.5
1.3 0.3 8.1 0.7

0.6 0.7 11.1 0.3
0.9 0.7 9.1 0.6


0.1
0.3

0.0


0.04
0.19

0.2
0.4


0.0


0.03
0.11

0.0


0.4 0.02
0.5 0.1

0.0 0.0


0.1
0.3


0.2 0.02
0.4 0.1


8.2 8.9 1.3 5.9
5.7 8.2 1.4 8.2


p < 0.001


p < 0.001


p < 0.001


NS


NS


p < 0.005


p < 0.001


0.8 0.4 59.4 60.0 41.4 141.4 p < 0.001
1.2 0.6 25.8 57.4 23.5 120.4

0.36 0.48 1.07 2.31 1.25 2.32 p < 0.001
0.10 0.15 0.13 0.12 0.13 0.15


-------- ----------------~---------- -


Diadema
savignyi

Diadema
setosum


p < 0.001


p < 0.001


Total


Survival







64























u


C C
o) Il

C >




CI 0
cai













U00
-. 8 8

Cg .S 2$^ v v

.v )-I +I +1





o.- 2 3
C o >







w = C-
II f a c
'5 e 6


< u a

DE o I U a
I^ I sk 4







65









12

UNPROTECTED REEFS
10- ln(y)=0.98+0.171n(x) o o
r=0.99 0

8 oo
Soco
0 o
PU
IL 6 0
(6

4 PROTECTED REEFS
In(y)=-0.85+0.351n(x)
e r.0.87
2-


0
0 1000 2000
AREA, m2


Fig. 13. Species-area curves for Kenyan reef lagoon sea urchins
comparing protected and unprotected reefs.





















BAMBURI

VIPINGO

KANAMAI

DIANI

MALINDI

WATAMU


I I I I I I I I *


0.5


1.0


SIMILARITY COEFFICIENT




Fig. 14. Cluster analysis of the sea urchin assemblage using the
Bray-Curtis (1957) measure of similarity and average
between-group linkages.




























C

OoC


0 0






2^
'? E r
E e S








Eo~
E-g
M c8 g



a 0 ..





00Cm
.^ 0
'.C!
0 eo e




0
vM "




,i"E
s -


W.C
I S00


CC

CO U
Cs ta
a _>ee



aeoQ
o 'p -

s.. a- 1


SOW
.a 2 a




*; N
S&. C t9

to c *-

60 ** ..




Q) 015
o.- S~ o



- >m
E* 5 a ...2


C.
D g
0 0

1)1)

O
04-
eonc
2ME
v> c
> 0
< -3


o 0
- 0


*e 0
cO IO


I- ir


en o
r-- tn


as
I tQ
0 2
- 5
*<2
Y ar






68

reefs than protected reefs (Fig. 15). Among unprotected reefs
biomass was similar.
An analysis of the sea urchin's species composition in
unprotected reefs by the rank-abundance technique (Fig. 16)
indicates a geometric or logseries (Odum et al. 1960, May 1975)
distribution. There was a positive relationship between the sea
urchin's adult body size and it species rank (Fig. 17a) and a left-
skewed distributional relationship between body size and abundance
(Fig. 17b).


Sea urchin coexistence


An analysis of coexistence between the three most abundant
species within Kanamai reef indicates similar across-reef distribution
patterns (Fig. 18) with E. mathaei being the most abundant species
followed by D. savignyi and lastly D. setosum. Despite these
similarities, it is clear that each species has microspatial preferences
within the reef (Table 8). E. mathaei inhabited small crevices, D.
savignyi inhabits intermediate crevices and occasionally small social
groups, and D. setosum was occasionally found in large crevices but
most frequently in the open in social groups. Crevice habitation and
sociality appear to be closely related to body size and morphology
(Table 8 and Fig. 19), the smaller the species the more frequently
they were found in crevices. The smaller their spine length, the less
their sociality. A day-night comparison (Table 9) indicates that D.
setosum leaves crevices and reduces its social group size at night
presumably because of nocturnal grazing. The other two species
















SEA URCHIN BIOMASS IN SIX KENYAN REEF LAGOONS


MALINDI WATAMU VPINGO KANAMAI BAMBURI


DIANI


PROTECTED SITES


UNPROTECTED SITES


Fig. 15. Estimated sea urchin biomass in the six studied Kenyan
reef lagoons. Estimates based on average weights (Table 7) and
densities of each species.


sooo




6000-




4000-




2000-






70









04 6
E y = 4.62-0.87x
0 4 r = 0.95
\ p<0.01
U 2- V
*
S0-



-4- 0
O *


o
-J
0 2 4 6 8 10 12
RANK

Fig. 16. Species rank plotted against the log abundance of the
sea urchins found within unprotected Kenyan reef lagoons.
Data taken from unprotected reefs in Table 5.














C4
E
o






w
UJ
















Z



u

LU


0 1i0
0 100


200


300


400


500 600


500


600


BODY SIZE, g

Fig. 17. The relationship between sea urchin body size and (a)
density and (b) species rank from unprotected Kenyan reef
lagoons.


0 100 200 300 400


0

S

S

S

0

S

S

(B)
















4572
1400
--* D. setosum
---- D. savignyi
E E. mathaei
In 1050-
I/

S-700

0 -oo

=so- ^ / / I,
". .


^- T- -- ---
0 100 200 300
landward seaward
DISTANCE ALONG TRANSECT

Fig. 18. Biomass distribution of the three sea urchin species
across the Kanamai reef lagoon. Bars represent + 1 standard
error of the mean (s.e.m.)







73












cu
CA




ou
4)

o4 e n




*E 0
co




+1 0
+1 +1 +1 +1 +1 +1



00 -



2 0 0


w 00






E u




1.5 W E c
cc u I















10


D. setosum
S8- y = 0.87 + 2.26x
E r = 0.98 *
o n = 35
LL mF
D 6
.j
0
D. savignyi
o 4 y = -0.72 + 2.09x
S r = 0.96 A
(3 n = 37
O A
_j E. mathaei
2 y = 4.59 + 2.54x
A r = 0.98
n = 36
AA
0 1 1
2 3 4 5
LOG TEST LENGTH, mm



Fig. 19.Test length-body volume (test + spines) relationships for
the three sea urchin species.

























II


eu
uo *
cc > cc









o .M N
.0 00
W) 0 i_
4 6 )
CI u
o c5













C C









a
E 9





= O "U
)O 0


o V







E- "
I- d







.1) 0*
0 ) 0


VI
z







0








zO


0



























L
f^



0,
























cC'








showed less diurnal changes. Apparently, nocturnal foraging by D.
setosum reduced the frequency of D. setosum in the D. savignyi
groups and therefore the average nocturnal group size.


Competition and predation


Competition experiments for shelter and space indicate that E.
mathaei exhibited intra- and interspecific agonistic behavior (Table
10). In competition for shelter "wins" consistently went to initial E.
mathaei inhabitants. Interspecific competition experiments indicated
that E. mathaei consistently won interspecific competition with
Diadema regardless of the initial or addition sequence.
Competitive interactions among and between Diadema species
were clarified by the experiment with reduced shelter sizes. When
shelter sizes were halved, larger individuals acquired better shelter
positions (Table 11). In interspecific competition experiments using
randomly selected individuals, D. setosum was the top competitor but
probably because D. setosum is larger on average than D. savignyi.
Where D. savignyi had similar or larger body lengths or volumes
than D. setosum, D. savignyi was the superior competitor in
interspecific interactions.
Comparisons of predator susceptibility suggest that E. mathaei
is most susceptible to predation when it is outside its burrows, but
burrow habitation greatly improves its survival probability (Table
12). Most Diadema survived the experiment's duration. Differences
in survival of Diadema species were not statistically significant.
Sociality may have improved D. setosum survival in Kanamai

















_E 0




cc8 Ea

*^ aP .z




0 0 :E
O





caa
.-E E
t o Cu






Qx r

o *S '- *

oE o U
i 0






oS .s o A

. o3
, a1 5 1. e
















- 2S > e





) -*s Cu



0 0 e 5
t-c U ^ CD
Cu0w C, c
g cO bO .)


.












w-0

.<












o
0)













U
C


- % O V0 0 (*


- %O r= 0% -



- o to % 2 o








0 tO O so 0> '0 ( tO
-- -- --c-
Ia r to a s O s
--N----


E E

f3~,
1 11
OC S S
Q~ = hC Q" Q


U
CI
-r
CI
L

q





























o


C6u
S)J l^E &




Ca
C o c cc a


aE.- IS a











cc cY c l-.I-
'.- 0. V
) 0. a




s E a





: g^&l VirE






o o


at' *E o |2
3 bv









o & "D .0 u .0 .






5 >'5 >t -o &
ia. U O3.) -
H0 va s a t ~


z
















.-
5,
E


*


+1 +

0m f
kn en


*4


+1+1

en O


I ,'
00


4
*
room
+1+1
0 -
C.,',


C4 -


c5




,


ci


W5
'> h
5 |
Si s 5
** Q &ib
2S "
































o C


t8 2 is
J g


VmCu
s.-siC


" E



u m




cca
7Vt
S-








Cu
V,















.cd M *B
e o'E
53c5
60

u
0,

'*5j 0'






-n4
cS-3





'- U 'a




u 3u
Vcui


w T3



3 I- ^ I
Cu u

V ~"


.0o,

&. ca ;


(V g
(S >^^-

D -c >,
.^1
r" l


+ + ++1








COO



00'~

+I
00 +










+1+1
r\o -
d5ri'-


I-
0


C3









Sel
i3


o
0




*
1





C "


<



on











.c
0



SII




.< 3
"5 e
o >









U 0
V

=





^ u
ne i
S<>M


+I- -I
C.-

^I^















I,
+1 + 1


vr




.4.4
*^ 0:


*





s4-















C4
* *


. 0..
. O <
LUcQ U


c~ei4









4?
C- 4 .
~ .

L~LZIZUX


0s


pmm
lOi
* S30






80
(p<0.10), but not in Vipingo. In Vipingo, although tied together,
Diadema were occasionally observed inhabiting different crevices.
Therefore, the experiment may not have consistently measured the
effect of sociality on predation rates at this site.




Comparisons of predation on different reefs


Survival of tethered E. mathaei in the 14 study sites indicated
that predation was correlated with differences in sea urchin
abundance and species composition. Population densities of sea
urchins were negatively correlated with relative predation rate (Fig.
20). Species richness and diversity (Fig. 21) were highest with
intermediate predation rates and with low to intermediate E. mathaei
density. Where E. mathaei were dense, diversity and species
richness were greatly reduced (Fig. 22). Where predation rates was
reduced the absolute and relative importance of E. mathaei
increased.



Relationship Between Living Communities and Substrate Cover and
Complexity


Table 13 and 14 present the results of measuring substrate and
cover diversity in different reefs. Large variance of measured
variables indicates patchy distributions. Coral cover varied from
4.6% in Diani to 30% in Watamu. Protected and unprotected reefs







81














200.

E. mathaei
Sqrt(y) = 13.8-15.13x
r = -0.91
p < 0.001
CM 150I
E All Sea Urchins
0 y = 153.65-171.59x
r : -0.88
m p < 0.001
100oo

z



50



*


0.0 0.2 0.4 0.6 0.8 1.0

RELATIVE PREDATION



Fig. 20. Total sea urchin and Echinometra mathaei densities
plotted against relative predation intensity at 14 sites.






82

0.8

(A) *

S0.6 -


I-C
w
LU

0.
0.2 -



0.0 i ,
0.0 0.2 0.4 0.6 0.8 1.0

12

(B)
c) 10
8o *
*
z 8

6 *
Cl *



V) 2


0.0 0.2 0.4 0.6 0.8 1.0

RELATIVE PREDATION

Fig. 21. Scatter-plots of species diversity (a) (Simpson's Index)
and species richness (b) as a function of relative predation
intensity on Echinometra mathaei.















0.6 4


0.4




0.2-


0.0 I-I I -. I I I
0 40 80 120 160
12

(B)
10 *
*S *
8 *
*S *

6- *


4-


24


0 40
Echinometra


I I I *
80 120 160
mathael DENSITY, #/10 m2


Fig. 22. (a) species diversity (Simpson's Index) and (b) species
richness plotted against Echinometra mathaei density.


,** (A)

*

*

*






*







84


Table 13. Water depth at low tide, topographic complexity
(rugosity), diversity of benthic organisms (Simpson's Index), and the
percent cover (x S.D.), in the six locations. The significance level
of a Kruskal-Wallis test of difference between reefs is provided.


Malindi Watamu VipingoKanamaiBamburi Diani


Kruskal-
Wallis


Depth, m 0.8 1.1 0.4 0.4 1.1 0.6
0.3 0.2 0.2 0.1 0.3 0.1

Topographic 1.37 1.45 1.30 1.24 1.23 1.15 p < 0.001
complexity 0.33 0.20 0.16 0.18 0.15 0.12

Substrate 0.82 0.78 0.59 0.67 0.53 0.60 p = 0.07
diversity 0.0 0.1 0.0 0.1 0.0

Percent Cover


Hard coral


Algal turf


Calcareous
algae

Macro-algae


Seagrass


Coralline
algae

Soft coral


Coral sand


Sponge


17.6 30.0 18.1 12.7 8.9 4.6
12.8 24.7 13.6 9.5 6.9 4.6


p < 0.001


27.6 19.7 58.6 44.3 66.3 54.7 p < 0.001
20.5 17.5 19.8 15.4 14.8 22.2

8.3 10.3 0.6 0.0 0.0 1.4 p < 0.001
6.3 10.3 1.5 0.0 0.0 3.1


2.4 2.5 2.8 0.9 2.4 1.5
3.2 2.9 4.2 2.4 3.0 1.9


p < 0.05


18.1 28.1 13.8 31.1 9.2 31.9 p < 0.001
22.6 32.9 14.6 18.9 15.6 23.0

11.6 0.2 0.04 0.1 2.2 0.3 p < 0.001
14.0 0.7 0.2 0.4 2.8 0.7

1.4 0.0 1.9 1.8 3.0 0.8 p < 0.006
2.4 0.0 3.0 3.6 3.8 1.2


14.6 9.1 4.3 10.1 8.2 4.7
17.0 9.6 7.3 10.1 10.9 8.4


p < 0.02


0.0 0.0 0.6 0.1 0.5 0.04 p < 0.008
0.0 0.0 1.1 0.4 1.1 0.2


- ----------- --