¡zwo

insftjuiveo? ! « e * í S « S ( á W ^ O R d a z o * IÄW j
te>î!lî$ètor Uï ii i s $ Ü i3 ¡'íL íS S ííS tf i
■■•; s . ^ O u ^ S B R G l i i ! » !

>3-3’ m r

DEVELOPMENTS IN HYDROBIOLOGY

The Ecology of Mangrove

and Related Ecosystems

edited by
Victor Jaccarini and Els Martens

Kluwer Academic Publishers

The Ecology of Mangrove and Related Ecosystems

izwo
Instituut voor Zeewetenschappelijk Onderzoek ( vzw)
Institute for Marine Scientific Research
VICTORIALAAN 3-B -8400 OOSTENDE BELGIUM
Tel.+32-(0) 59-321045— Fax:+32-(0) 59-321135
Developments in Hydrobiology 80

Series editor
H. J. Dumont
 The Ecology of Mangrove and
Related Ecosystems
Proceedings o f the International Symposium held at M ombasa, Kenya,
2 4 —3 0 Septem ber 1 9 9 0

Edited b y
V icto r Jaccarini & Els Martens

R eprinted from Hydrobiologia, vol. 2 4 7 (1 9 9 2 )

W
Kluwer Academic Publishers
Dordrecht / Boston / London
Library of Congress Cataloging-in-Publication Data

The E c o l o g y o f ma ng ro v e and r e l a t e d e c o s y s t e m s : p r o c e e d i n g s o f t h e
i n t e r n a t i o n a l s y m p o s i u m h e l d a t Mo mbas a, K e n y a , S e p t e m b e r 2 4 - 3 0 ,
1990 / e d i t e d by V i c t o r J a c c a r i n i & E l s M a r t e n s .
p. cm. — ( D e v e l o p m e n t s i n h y d r o b i o l o g y ; 8 0 )
" R e p r i n t e d f r o m H y d r o b i o 1o g i a v o l . 2 4 7 ( 1 9 9 2 ) . "
Includes indexes.
I S BN 0 - 7 9 2 3 - 2 0 4 9 - 2 ( h b : a c i d f r e e p a p e r )
1. M a n g r o v e s wa mp e c o l o g y — C o n g r e s s e s . 2. Mangrove f a u n a -
-Congresses. 3. Mangrove s w a m p s - - C o n g r e s s e s . I. Ja c c a rin i,
V ictor. II. M artens, Els. III. Series.
Q H 5 4 1 . 5 . M2 7 E 2 5 1992
5 7 4 . 5 ‘2 6 3 2 5 — dc 20 92-38659

ISBN 0-7923-2049-2

Published by Kluwer Academic Publishers,

P.O. Box 17, 3300 AA D ordrecht, The Netherlands.

Kluwer Academic Publishers incorporates

the publishing programmes of
D. Reidel, M artinus Nijhoff, D r W. Junk and M TP Press.

Sold and distributed in the U.S.A. and Canada

by Kluwer Academic Publishers,
101 Philip Drive, Norwell, MA 02061, U.S.A.

In all other countries, sold and distributed

by Kluwer Academic Publishers Group,
P.O. Box 322, 3300 AH D ordrecht, The Netherlands.

The publication of this volume was funded by the Commission of the European Communities,
and sponsored by the Hydrobiological Society of E astern Africa (HYSEA).

Printed on acid-free paper

All Rights Reserved

© 1992 Kluwer Academic Publishers
No part of the material protected by this copyright notice may be reproduced or utilized in any
form or by any m eans, electronic or mechanical, including photocopying, recording, or by any
information storage and retrieval system, without written permission from the copyright owner.

Printed in Belgium
 V

Contents

P re f a c e ......................................................................................................................................................................... vii

P i c t u r e ......................................................................................................................................................................... ix

List of p a rtic ip a n ts .................................................................................................................................................. x

S T A T U S , P R IM A R Y P R O D U C T IV IT Y A N D T A X O N O M Y O F M A N G R O V E S,
SEA G RA SSES A N D ALGA E
Developing m anagem ent plans for the m angrove forest reserves o f m ainland T anzania
by A. K.Semesi ............................................................................................................................................. 1
S tructure and species distribution in C oringa m angrove forest, G odavari D elta, A ndhra Pradesh,
India
by J. A zariah, H. A zariah, S. G unasekaran & V. Selvam ........................................................... 11
S tructural properties o f tw o types o f m angrove stands on the northw estern coast o f Sri Lanka
by M. D. A m arasinghe & S. B alasubram aniam ................................................................................ 17
The current status o f m angroves along the K enyan coast: a case study o f M ida Creek m angroves
based on rem ote sensing
by P. O. G ang & J. L. A g a ts iv a ............................................................................................................. 29
N et prim ary productivity o f tw o m angrove forest stands on the northw estern coast o f Sri L anka
by M. D. A m arasinghe & S. B alasubram aniam ................................................................................ 37
The effect o f coastline changes on m angrove com m unity structure, Portuguese Island, M ozam bique
by J. C. H atto n & A. L. C o u t o ................................................................................................................ 49
The seagrass and associated m acroalgal vegetation o f G azi Bay (K enya)
by E. C oppejans, H. Beeckm an & M. D e W i t .................................................................................. 59
M icrophytoplankton o f the P itchavaram m angals, southeast coast o f India: species com position
and population density
by L. K an n an & K. V asantha ................................................................................................................ 77

USERS O F TH E M ANGALS
The ecology o f Belizean m angrove-root fouling com m unities: patterns o f epibiont distribution and
abundance, and effects on root growth
by A. M. Ellison & E. J. F a r n s w o r th .................................................................................................... 87
The m eiobenthos o f five m angrove vegetation types in G azi Bay, K enya
by S. Vanhove, M. Vincx, D. Van G ansbeke, W. Gijselinck & D. S c h r a m ............................. 99
Ecom orpholy o f Crassostrea cucullata (Born, 1778) (O streidae) in a m angrove creek (G azi, Kenya)
by J. F. T ack, E. V anden Berghe & Ph. P o l k .................................................................................... 109
A note on the Zooplankton distribution and diversity in a tropical m angrove creek system, G azi,
K enya
by M . K. W. O sore .................................................................................................................................... 119
vi

CO RAL R EEF ECOSYSTEM S

P hase shifts in coral reef com m unities and their ecological significance
b y T . J. D one ................................................................................................................................................ 121
The incidence o f hurricanes on the north coast o f J am aica since 1870 : are the classic reef descriptions
atypical?
by J. D. W oodley ......................................................................................................................................... 133
Im portance o f the Lac-lagoon (Bonaire, N etherlands Antilles) for a selected num ber o f reef fish
species
by G . van der Velde, M. W. G orissen, C. den H artog, T. v a n ’t H o ff & G . J. M eijer . . . . 139

C H E M IC A L A N D P H Y S IC A L P R O C E S S E S
H ydrodynam ics o f m angrove sw am ps and their coastal w aters
by E. W o la n s k i............................................................................................................................................. 141
Intratidal fluctuations in stratification within a m angrove estuary
by R. J. U ncles, W .-K. G ong & J.-E . O n g ......................................................................................... 163
Physical processes in the Rufiji delta and their possible im plications on the m angrove ecosystem
by J. F r a n c i s .................................................................................................................................................. 173
D iurnal variation in physical-chem ical properties and prim ary production in the interconnected
m arine, m angrove and freshw ater biotopes o f K akinada coast, A n d h ra P radesh, India
by V. Selvam, J. A zariah & H. A zariah .............................................................................................. 181
M ethanogenic bacteria in m angrove sedim ents
by R. M ohanraju & R. N atarajan ......................................................................................................... 187

F IS H E R Y B IO L O G Y A N D M A R IC U L T U R E
M angroves as a habitat for fish and praw ns
by A. Sasekum ar, V. C. Chong, M. U. Leh & D ’C ruz .................................................................... 195
M ariculture developm ent in Kenya: alternatives to siting ponds in the m angrove ecosystem
by J. R asow o ................................................................................................................................................ 209
T he effect o f oocytic atresia on fecundity estim ates o f the rabbit fish Siganus sutor (Pisces:
Siganidae) o f K enyan m arine inshore w aters
by M. J. N tiba & V. Jaccarini ................................................................................................................ 215
Differences in biom ass production and carrageenan yields am ong four strains o f farm ed carrageeno-
phytes in N orthern Bohol, Philippines
by G . C. T rono & A. O. L lu is m a ........................................................................................................... 223

E X P L O IT A T IO N , C O N S E R V A T IO N A N D M A N A G E M E N T A S P E C T S
Status and exploitation on the m angrove and associated fishery resources in Z an zib ar
by M. A. K. Ngoile & J. P. Shunula ..................................................................................................... 229
The m anagem ent o f Insular C aribbean m angroves in relation to site location and com m unity type
by P. R. B acon & G . P. Alleng .............................................................................................................. 235
Regeneration strategy o f m angroves along the Kenyan coast: a first approach
by D. Van S p e y b ro e c k ................................................................................................................................ 243
Im pact o f p ast m anagem ent practices on the present status o f the M uthupet m angrove ecosystem
by J. A zariah, V. Selvam & S. G unasekaran ..................................................................................... 253

Subject index 261

Hydrobiologia 247: v ii-v iii, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. vii

Preface

T he international sym posium on The Ecology o f Mangrove and Related Ecosystems, M om basa, Kenya,
24-30 Septem ber, 1990, w as organized by the D epartm ent o f Zoology, University o f Nairobi, in col­
laboration with the K enya M arine and Fisheries R esearch Institute, and the Regional D ispatch Centre
o f the Regional C ooperation in Scientific Inform ation Exchange in the W est Indian Ocean, M om basa,
Kenya. It was officially opened by Professor F. N. O nyango, Principal, College o f Biological and Physical
Sciences, U niversity o f N airobi. The closing address was given by Professor M. Hyder, retired Principal
and Professor o f Zoology, C B PS, University o f Nairobi.
Scientific investigations o f the m angrove and coral reef ecosystem s in K enya have been going on at
least since the eartly 70’s and have been gathering m om entum with the involvem ent o f the University of
N airobi, K M F R I, and British, Belgian, D utch and Italian universities and scientific institutions. So the
tim e w as ripe to take stock o f the situation and to try to bring together w orkers in the same field from
as m any tropical regions as possible. A cursory glance at the list o f participants is sufficient proof o f the
w idespread interest in the them e o f the sym posium . C ontributions cam e from all the main tropical m arine
areas o f the globe, representing tw enty-three countries, in W est and E astern Africa, South and Southeast
A sia, C entral and South A m erica, and A ustralia. The thirty papers published in this volume were selected
from seventy presented at the symposium.
It is recognized th at the m angroves have intim ate and complex linkages with other coastal systems,
notably seagrass beds and coral reefs. So, starting from the m angrove forest and its users at the interface
betw een the terrestrial, freshw ater and m arine environm ents, the papers presented proceed to the fully
subm erged seagrasses and further out to the corals. The nature o f the coupling between the different
systems w as touched on at the symposium, but stands out as a serious gap in our knowledge. Next,
chem ical and physical processes which constitute the m echanism s o f linkage received attention. H um an
uses and abuses o f the ecosystem s form ed the last m ajor topic o f the symposium.
W orkshop discussions were held during the sym posium in preparation for the adoption in plenary
session o f two declarations: A Statem ent o f Priorities fo r Scientific Research and A Declaration on
Management Strategies.
The following priorities for basic research were agreed : the interlinkages between the mangrove, coral
reef and seagrass ecosystem s; the need for m ore inform ation on nearshore hydrological processes
operating at the small and m eso-scale and elaboration of predictive m odels ; m ore w ork on nutrient cycling
and energy flow including soil processes; biological processes and use o f bioindicators.
Priorities were also set for applied research. This, it w as agreed, should be directed tow ards m anage­
m ent o f the forest or aquatic product on a sustainable basis, taking into account the integrated nature
o f the ecosystem s. Em phasis should be placed on minimizing perturbations to the systems and the
m aintenance of ecological integrity. B road areas needing investigation were identified: m ethods for
efficient and ecologically sound land use in the w atershed adjacent to the m angal, seagrass beds and coral
reefs; effect o f m anagem ent o f the m angroves on off-shore fisheries; im pact of industralization on the
m angal and related ecosystem s; effect o f dom estic and industrial effluents and m anagem ent o f their
im pact; com parative studies on destructive and non-destructive aquacultural techniques with a view to
m inimizing negative effects; the effect of hum an activities on physical, chemical and biological soil
processes; research on m anagem ent practices appropriate to different m angrove forest types and
prom otion o f a biotechnological aproach in m angrove silviculture; socio-econom ic studies o f hum an
viii

coastal settlem ents; natural products o f the m angals; the im pact o f global clim ate change; and active
rehabilitation o f seagrass beds and coral reefs.
The Declaration on M anagement Strategies was positive, concluding th at m uch can be done to stop the
degradation o f dam aged m angals, seagrass beds and coral reefs and to rehabilitate them , but th at new
techniques m ust be w orked out to allow use on a sustainable basis for long term econom ic return and
for the wellbeing o f coastal hum an settlem ents and a healthy environm ent.
To achieve the above goals the Symposium recom m ended the following strategies.
T he establishm ent o f National Coastal Management Plans including the following com ponents : survey
o f both natural and hum an resources and their present uses ; identification o f the problem s affecting coastal
ecosystem s; adjustm ent o f legal fram ew orks with regard to the uses o f these ecosystem s and enforcem ent
o f existing m anagem ent regulations; provision o f technical assistance and training for the introduction
o f econom ically and ecologically viable activities in these ecosystem s; rehabilitation o f dam aged areas;
creation o f natural reserves both for tourism and for the recuperation o f natural plant and anim al
populations; education and creation o f public aw areness on the value and problem s o f these ecosystem s.
In form ulating and im plem enting each national P lan all concerned parties should be involved, i.e., the
inhabitants and the outside users o f these ecosystem s, the policy m akers, the scientists and technical
experts, and also the national and international funding agencies. U se should be m ade o f existing
institutions and netw orks. The scientists gathered in the sym posium urged the establishm ent or strength­
ening o f regional and international associations o f scientists to provide the scientific basis to national and
regional m anagem ent plans.
It is our pleasant duty as C hairm an and Secretary o f the Symposium Organizing C om m ittee to thank
the m any sponsors who generously provided financial support for the Sym posium : the African
Biosciences N etw ork for providing the seed funds; the U S N ational R esearch C ouncil; the C om m ission
o f the E uropean C om m unities ; the A ustralian International D evelopm ent A ssistance B ureau; the N ether­
lands M inistry o f Agriculture, N ature M anagem ent and Fisheries; U nited N ations Environm ent P ro ­
gramm e, Forestry D epartm ent; U N E S C O - R O S T A ; and the T hird W orld A cadem y o f Sciences.
Special thanks are due to the Com m ission o f the E uropean C om m unities for providing funds covering
the publication costs o f this volume. A debt o f gratitude is owed to Hydrobiologia (K luw er A cadem ic
Publishers) and the C hief Editor, Professor H. D um ont o f U niversity o f G h en t (Belgium), for agreeing
to publish the Proceedings as a special volume in their series D evelopm ents in Hydrobiology. W e would
also like to thank the m em bers o f the Executive Com m ittee for their help in the organization o f the
Symposium . Finally, we wish to thank the University o f N airobi in the persons o f its previous Vice-chan­
cellor, Professor P. M. M bithi, Professor F. N. O nyango, Principal, College o f Biological and Physical
Sciences, and Professor R. W. M wangi, D ean, Faculty o f Science, for their support. T hanks should also
go to Professor G . M. O. M aloiy who negotiated the initial funding with the A frican Biosciences N etw ork.

V. J a c c a r in i & E. M a rte n s
 1. D. Onyango 15. N. 0 . Isirimah 29. A. M. Ellison 43. W. M. Kudoja 57. M. A. Ngoile
2. J. Kasyi 16. M. Hemminga 30. A. Robertson 44. J. G. Mutwiri 58. M. M astaller
3. C. Amadou 17. P. O. Gang 31. A. H. Moallim 45. A. Yanez-Arancibia 59. N. Revisa
4. J. Francis 18. M. A. Amakiri 32. F. van Averm aat 46. M. D. Fortes 60. E. Okemwa
5. J. K athupotha 19. M. D. Amarasinghe 33. V. Selvam 47. L. K annan 61. E. M artens
6. G. C. Trono 20. R. W. Mwangi 34. A. G. Tejan 48. E. Coppejans 62. K. Delbeke
7. V. Jaccarini 21. M. Vannini 35. A. Sasekumar 49. P. Akiwumi 63. R. M ohanraju
8. C. A. M uhando 22. S. G unasekaran 36. J. K. Muraya 50. S. Abdulrahm an Issa 64. M. Vincx
9. A. Thorhaug 23. J. P. Shunula 37, U. Helbing' 51. G. van der Velde 65. T. Done
10. A.-F. Woitchik 24. M. Vuai 38. K. N askar 52. G. W amukoya 66. D. van Gansbeke
11. J. L. Agatsiva 25. K. Deges 39. L. Vergeer 53. E. Wolanski 67. J. Schrijvers
12. L. Solorzano 26. J. D. Woodley 40. D. Prasad 54. J. H atton 68. M. J. N tiba
13. J. Azariah 27. B. Musyoki 41. J. Rasowo 55. F. W. Fong
14. E. Slim 28. N. Visser 42. J. M utere 56. F. M. Chale

X V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems.
List of participants
A bdulrahm an i s s a , S., Commission for Lands & Environ­
m ent, P.O. Box 811, Z anzibar, Tanzania
A g a t s i v a , J. L., Remote Sensing, D epartm ent o f Resource
Surveys and Remote Sensing, P.O. Box 47146, Nairobi,
Kenya
A i k e n , K., D epartm ent of Biological Sciences, Florida Inter­
national University, Tamiami Campus, Miami,
FL 33199, USA
A k i w u m i , P. K., O ceans & C oastal Areas Programme, Ac­
tivity Centre (OCA/PAC) U nited N ations Environment
Programme, P.O. Box 30552, Nairobi, Kenya
A l l e n g , G., D epartm ent of Zoology University of the W est
Indies, M ona, Kingston 7, Jam aica
A m a d o u , C., U N D P/FA O Fuelwood Project c/o U N D P,
P.O. Box 1011, Freetown, Sierra Leone
A m a k i r i , M. A., Rivers State University Science & Techn.
Faculty of Agriculture, P.M.B. 5080, Port H arcourt,
Nigeria
A m a r a s i n g h e , M. D., National Aquatic Resources Agency,
Regional Research Centre, Kalpitiya, Sri Lanka
A n d e r s o n , M., D epartm ent o f Biological Sciences, Florida
International University, Mamiami Cam pus, Miami,
FL 33199, USA
A n y i e n i , Z. M., Perm anent Presidential Commission on Soil
Conservation & Afforestation, P.O. Box 46191, Nairobi,
Kenya
A z a r i a h , J., D epartm ent of Zoology University of M adras,
Guindy Campus, M adras - 600 025, India
A z a r i a h , H., CAS in Botany, University of M adras, Guindy
Campus, M adras - 600 025, India
B a c o n , P. R., D epartm ent of Zoology, University of the
W est Indies, M ona, Kingstone 7, Jam aica
B a l a s u b r a m a n ia m , S., D epartm ent of Botany, University
of Peradeniya, Peradeniya, Sri Lanka
B o r g e r , R., Ecology & Systematics Laboratory, Free U ni­
versity of Brussels, Pleinlaan 2, B-1050 Brussels, Belgium
C a r b y , B., D epartm ent of Biological Sciences, Florida Inter­
national University, Tamiami Campus, Miami,
FL 33199, USA
C h a l e , F. M., D ept, o f Zoology & Marine Biology, U ni­
versity of D ar Es Salaam, P.O. Box 35064, D ar Es
Salaam , Tanzania
C h o n g , V. C., D epartm ent o f Zoology, University of
M alaya, 59100 K uala Lumpur, Malaysia
C o p p e j a n s , E., State University of G hent, L aboratory of
M orphology, Systematics & Ecology of the Plants, K. L.
L edeganckstraat 35, B-9000 G hent, Belgium
C o u t o , M., Faculty of Biology, University Eduardo Mon-
diane, C.P. 257, M aputo, M ocambique
Hydrobiologia 247: x - x i i, 1992.
D ’C r u z , R., D epartm ent of Zoology, University of M alaya,
59100 K uala Lumpur, M alaysia
D a r o , M. H., Ecology & Systematics Laboratory, Free
U niversity of Brussels, Pleinlaan 251, B-1050 Brussels,
Belgium
D e g e o r g e s , A., Regional Environm ent Advisor, U.S. -
A ID , P.O. Box 30261, Nairobi, Kenya
D e g e s , K., State University of G hent, Institute of Zoology
- M arine Biology, K. L. Ledeganckstraat 35, B-9000
G hent, Belgium
D e l b e k e , K., Kenya-Belgium Project M arine Science
Kenya M arine & Fisheries Research Inst., P.O. Box
81651, M om basa, Kenya
D e n H a r t o g , C ., L aboratory of A quatic Ecology, Catholic
University of Nijmegen, Toernooiveld, 6525 ED Nijme­
gen, The N etherlands
D o n e , T., A ustralian Institute of M arine Science A IM S,
PMB 3 Townsville MC, Q ueensland 4810, A ustralia
E l l i s o n , A. M., D epartm ent of Biological Sciences, M ount
Holyoke College Clapp Laboratory, South Hadley,
MA 01075-1484, U SA
Farnsw orth, E. J., D epartm ent of Botany University of
Vermont, 120 M arsh Life Sciences Building, Burlington,
VT 05405, U SA
F o n g , F. W., D epartm ent ofZoology, University o f M alaya,
Field Studies Centre, 59100 K uala Lumpur, M alaysia
F o r t e s , M. D., M arine Science Institute, University of the
Philippines, Diliman, Quezon City 1101, Philippines
F r a n c i s , J., Institute of M arine Sciences, University of D ar
Es Salaam , P.O. Box 668, Z anzibar, T anzania
G a m b r e l , R. P., D epartm ent o f Soli Science, Rivers State
University of Science, P.M.B. 5080, Port H arcourt,
Nigeria
G a n g , P. O., Technical Officer, D epartm ent of Resource
Surveys and Remote Sensing, P.O. Box 47145, Nairobi,
Kenya
G e e t s , G., Dept. Zook, K enyatta Univ., P.O. Box 43844,
N airobi, Kenya
G h e r a r d i , F., D epartm ent Animal Biology & Genetics,
U niversity o f Florence, Via R om ana 17, Firenze 50125,
Italy
G ic h i r u , M., D epartm ent o f Soil Service, Rivers State
University o f Science, P.M.B. 5080, Port H arcourt,
Nigeria
G ie s e n , W., L aboratory o f Aquatic Ecology, University of
Nijmegen, Toernooiveld, 6525 ED Nijmegen, The N ether­
lands
G o r d o n , V., D epartm ent o f Biological Sciences, Florida
International University, Tamiami Campus, Miami,
FL 33199, USA

G o r i s s e n , M. W., Laboratory of A quatic Ecology, Catholic
U niversity of Nijmegen, Toernooiveld, 6525 ED Nijme­
gen, The N etherlands
G u n a s e k a r a n , S., D epartm ent o f Zoology, University of
M adras, Guindly Campus, M adras - 600 025, India
H a l l , K. C., U N D P/FA O Fuelwood Project c/o U N D P,
P.O. Box 1011, Freetown, Sierra Leone
H a r r y , F. O. C., D epartm ent of Soli Science, Rivers State
University of Science, P. M. B. 5080, Port H arcourt,
Nigeria
H a t t o n , J. C., Faculty of Biology, University Eduardo
M ondiane, C. P. 257, M aputo, Mozambique
H e i m s a t h , A., G erm an Agency for Technical Cooperation,
Kilifi W ater and Sanitation Project, P.O. Box 666, Kilifi,
Kenya
H e l b i n g , U., IBAMA, Av. Rio Branco 186, 490000 Ara-
caju-SE, Brazil
H e m m i n g a , M. A., D elta Institute for Hydrobiological Re­
search, V ierstraat 28, 4401 EA Yerseke, The N ether­
lands
H y d e r , M., D epartm ent of Zoology, University of Nairobi,
P.O. Box 30197, Nairobi, Kenya
I s i r i m a h , N. O., D epartm ent of Soli Science, Rivers State
University Science & Techn., P.M.B. 5080, Port
H arcourt, Nigeria
J a c c a r i n i , V., D epartm ent o f Zoology, University of
Nairobi, P.O. Box 30197, Nairobi, Kenya
K a n n a n , L., Annamalai University, Centre o f Adv. Studies
M arine Biology, Porto Novo - 508 502, Tamil N adu,
Indiua
K a s y i , J. N., D epartm ent o f Zoology, University o f Nairobi,
P. O. Box 30197, Nairobi. Kenya
K a t u p o t h a , J., D epartm ent of Geography, University of Sri
Jayew ardenepura, Gangodawilla, Nugegoda, Sri Lanka
K u d o j a , W. M., D epartm ent o f Zoology, University of
Nairobi, P.O. Box 30197, Nairobi, Kenya
L a r a - D o m i n g u e s , A. L., EPOM EX Programme, A uto­
nomous University of Campeche, Ap. Postal 520,
C lm peche 24030, Mexico
L a r e w , H., U SA ID - Office of the Science Advisor Agency
for International D evelopment, Room 320, SA-18,
W ashington D C 20523, USA
L e h , M. U., D epartm ent of Zoology, University of Malaya,
59100 K uala Lumpur, Malaysia
L e v y , I., D epartm ent of Botany, University of W est Indies,
K ingstone 7, Jam aica
L i t t a u , J., D epartm ent of Botany, University of West
Indies, M ona, Kingstone 7, Jam aica
L l u i s m a , A. O., M arine Science Institute, University o f the
Philippines, Diliman, Quezon City 1101, Philippines
M a r t e n s , E. E., D epartm ent of Zoology, University of
Nairobi, P.O. Box 30197, Nairobi, Kenya
M a s t a l l e r , M., Tournelystr. 16, D-8901 Stadtbergen,
Germany
M e i j e r , G. J., Stinapa, Bonaire, N etherlands Antilles
xi
M e n k i r , V., N ational Academy of Sciences BOSTID ,
H A 476, 2101 Constitution Ave., N.W., W ashington,
DC 20418, USA
M i c h e l i , F., D epartm ent Animal Biology & G enetics, Uni­
versity of Florence, Via R om ana 17, Firenze 50125, Italy
M il l e r , B., D epartm ent o f Biological Sciences, Florida
International University, Tamiami Campus, Miami,
FL 33199, USA
M o a l l im , A. H., D ata Collection & Storing Centre,
Ministry o f Fisheries and M arines Resources, P.O. Box
2589, M ogadishu, Somalia
M o h a n r a j u , R., Centre Advanced Study in M arine Biology,
Annamalai University, Porto Novo - 608 502, Tamil
N adu, India
M o n d e g o , Faculty of Biology, University Eduardo M ond­
iane, C.P. 257, M aputo, Mozambique
M u h a n d o , C. A., Institute of M arine Sciences, University of
D ar Es Salaam, P.O. Box 668, Z anzibar, T anzania
M u n g a , D., Kenya M arine & Fisheries Research Institute
(K M FR I), P.O. Box 81651, M om basa, Kenya
M u r a y a , J. K., D epartm ent of Botany, University of
Nairobi, P.O. Box 30197, Nairobi, Kenya
M u s y o k i , B. M., Kenya Wildlife Service Provincial Wildlife
Office, P.O. Box 82144, M om basa, Kenya
M u t e r e , J., Kenya M arine & Fisheries Research Institute
(K M FR I), P.O. Box 81651, M om basa, Kenya
M u t w i r i , J. G., Kenya Marine & Fisheries Research Insti­
tute (K M FR I), P.O. Box 81651, M om basa, Kenya
M w a n g i , R. W., D epartm ent o f Zoology, University of
Nairobi, P.O. Box 30197, Nairobi, Kenya
N a s k a r , K., Central Inland Capture Fisheries Research
Institute, B arrackpore - 743 101, W est Bengal, India
N a t a r a j a n , R., Ocean D ata Centre for W ater Resources,
Annamalai University, M adras 600 025, India
N g o i l e , M. A. K., Institute of M arine Sciences, University
of D ar Es Salaam, P.O. Box 668, Z anzibar, T anzania
N g u l i , M. M., Kenya M arine & Fisheries Research Insti­
tute (K M FR I), P.O. Box 81651, M om basa, Kenya
N t i b a , M. J., D epartm ent of Zoology, University of Nairobi,
P.O. Box 30197, Nairobi, Kenya
O j a n u g a , A. G., D epartm ent of Soil Science, Rivers State
University of Science, P.M.B. 5080, Port H arcourt,
Nigeria
O k e m w a , E., Kenya Marine & Fisheries Institute (K M FR I),
P.O. Box 81651, M om basa, Kenya
O l u o c h , A., D epartm ent of Zoology, University o f Nairobi,
P.O. Box 30197, Nairobi, Kenya
O n y a n g o , F., College Biological & Physical Sciences, Uni­
versity of Nairobi, P. O. Box 30197, Nairobi, Kenya
O n y a n g o , D., N airobi University, D epartm ent of Zoology,
P.O. Box 30197, Nairobi, Kenya
O s o r e , M. K. W., Kenya Marine and Fisheries Research
Institute, P.O. Box 81651, M om basa, Kenya
O y i e k e , H., Kenya M arine & Fisheries Research Institute
(K M FR I), P.O. Box 81651, M om basa, Kenya
xii
PiS S lE R S S E N S , P., RECOSCIX-W IO Project Kenya Marine
& Fisheries Research Inst., P.O. Box 95832, M om basa,
Kenya
P o l k , Ph., Kenya-Belgium Project M arine Sc, Ecology &
Systematics Laboratory, Free University o f Brussels,
Pleinlaan 2, B-1050 Brussels, Belgium
P r a s a d , D., D epartm ent o f Botany, University o f the W est
Indies, M ona, Kingston 7, Jam aica
R a s o l o f o , V. M., Centre N ational de Recherches O céano­
graphiques (C N R O ), B.P. 68, Nosy Be 207, M adagascar
R a s o w o , J., D epartm ent of Zoology, Moi University, P.O.
Box 3900, Eldoret, Kenya
R e e s e , R., D epartm ent of Biological Sciences, Florida Inter­
national University, Tamiami Campus, Miami,
FL 33199, USA
R e v i s , N., Ecology & Systematics Laboratory, Free Univer­
sity o f Brussels, Pleinlaan 2, B-1050 Brussels, Belgium
R o b e r t s o n , S. A., Coast Forest Survey, N ational Museums
of Kenya, .P.O. Box 162, Malindi, Kenya
R o d r i g u e z , M., D epartm ent o f Biological Sciences, Florida
International University, Tamiami Campus, Miami,
FL 33199, USA
R o h r s t e d , T., Delegation of the Commission of the
European Communities, P.O. Box 45119, Nairobi, Kenya
R u w a , R. K., Kenya M arine and Fisheries Institute
(K M FR I), P.O. Box 81651, M om basa, Kenya
S a s e k u m a r , A., D epartm ent o f Zoology, University of
M alaya, K uala Lumpur, Malaysia
S c h r i j v e r s , J., State Universtiy o f G hent, Institute of
Zoology - M arine Biology, Ledeganckstraat 35, B-9000
G hent, Belgium
S e l v a m , V., D epartm ent of Zoology, University o f M adras,
Guindy Campus, M adras 600 025, India
S e m e s i , A. K., D epartm ent of Botany, University o f D ar Es
Salaam , P.O. Box 35060, D ar Es Salaam, T anzania
S h a y , E. G., N ational Academy of Sciences BOSTID ,
HA 476, 2101 C onstitution Ave., N.W., W ashington,
D C 20418, USA
S h e t e , M. H. A ., E ast African Wildlife Society, P.O. Box
20110, Nairobi, Kenya
S h u n u l a , J. P., Institute of M arine Sciences, University of
D ar Es Salaam , P.O. Box 668, Z anzibar, T anzania
S i d r a k , G., D epartm ent o f Biological Sciences, Florida
International University, Tamiami Campus, Miami,
FL 33199, USA
S l i m , E., EEC - M angrove Project, D elta Institute for
Hydrobiological Research, Vierstraat 28, 4401 EA
Yerseke, The N etherlands
S o l o r z a n o , L., Instituto N acional de Pesca Letamendi 102
y La Rita, P.O. Box 5918, Guayaquil, Ecuador
T a c k , J. F., Ecology & Systematics Laboratory, Free Uni­
versity of Brussels, Pleinlaan 2, B-1050 Brussels, Belgium
T e a s , H., D epartm ent o f Biological Sciences, Florida Inter­
national University, Tamiami Campus, Miami,
FL 33199, USA
T e j a n , A. G .,U N D P/FA O Fuelwood Project c/o U N D P,
P.O. Box 1011, Freetown, Sierra Leone
T h o r h a u g , A., D epartm ent of Biological Sciences, Florida
International University, 555 N E, 15th St. P H H , Miami,
FL 33132, USA
T r o n o , G. C., Jr, M arine Sciences Institute University of the
Philippines, Diliman, Quezon City 1101, Philippines
V a n A v e r m a e t , F., State University of G hent, Institute of
Zoology-M arineB iology,K . L .Leganckstraat 35,B-9000
G hent, Belgium
V a n D e K e r k h o f , K., L aboratory o f Aquatic Ecology,
University of Nijmegen, Toernooiveld, 6525 ED Nijme­
gen, The N etherlands
V a n D e r V e l d e , G., Laboratory of A quatic Ecology,
Catholic University of Nijmegen, Toernooiveld, 6525 ED
Nijmegen, The N etherlands
V a n G a n s b e k e , D., State University of G hent, Institute of
Zoology - M arine Biology, K. L. Ledeganckstraat 35,
B-9000 G hent, Belgium
V a n H o v e , S., State University of G hent, Institute of
Zoology - M arine Biology, K. L. Ledeganckstraat 35,
B-9000 G hent, Belgium
V a n S p e y b r o e c k , D., D epartm ent of Botany, University of
Nairobi, P.O. Box 30197, N airobi, Kenya
V a n ’t H o f f , T., Stinapa, Bonaire, N etherlands Antilles
V a n d e n B e r g h e , E., D epartm ent of Zoology, University of
Nairobi, P.O. Box 30197, Nairobi, Kenya
V a n n i n i , M., University of Florence, Dept. Animal Biology
& G enetics, Via R om ana 17, 50125 Firenze, Italy
V a s a n t h a , K., Centre Adv. Studies M arine Biology, A n­
namalai University, Parangipettai - 608 502, Tamil
N adu, India
V e r g e e r , L. H. T., L aboratory o f Aquatic Ecology, Univer­
sity of Nijmegen, Toernooiveld, 6525 ED Nijmegen, The
N etherlands
V i n c x , M., State University of G hent, Institute of Zoology,
M arine Biology Section, K. L. Ledeganckstraat 35,
B-9000 G hent, Belgium
V i s s e r , N., Agricultural A ttache, Royal N ertherlands Em­
bassy, P.O. Box 41537, N airobi, Kenya
V u a i , M., Forestry D epartm ent, Z anzibar, T anzania
W a m u k o y a , G. M., KEN -U -TAN E X PO R TS, P.O. Box
84384, M om basa, Kenya
W i j n a n t , J., Kenya-Belgium Project M arine Sciences
Kenya M arine & Fisheries Research Inst., P.O. Box
81651, M om basa, Kenya
W o i t c h i k , A. F., EEC Project Mangrove Ecology, Free
University of Brussels A N C H , Pleinlaan 2, B-1050
Brussels, Belgium
W o l a n s k i , E., A ustralian Institute of M arine Science
(A IM S), PMB 3, Townsville M C, Queensland 4810,
A ustralia
W o o d l e y , J. D., Discovery Bay M arine Laboratory, P.O.
Box 35, Discovery Bay, Jam aica
Y a n e z -A r a n c ib ia , A., E PO M EX , Autonom ous University
of Campeche, Av. Agustín Melgar, Ap. Postal 520,
Campeche 24030, Mexico
Hydrobiologia 2 4 7 : 1-10, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 1
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Developing management plans for the mangrove forest reserves of

mainland Tanzania

A delaida K. Semesi
Botany Department, University o f Dar es Salaam, P.O. B ox 35060, Dar es Salaam, Tanzania

K ey words: m angroves, vegetation, distribution, m anagem ent, T anzania

Abstract

The m angroves o f T anzania are currently no t being m anaged although they are legally gazetted as for­
est reserves. The m anagem ent o f this overexploited and under estim ated natural resource rests with the
F orest and Beekeeping Division in the M inistry o f N atural R esources and Tourism . It is the D ivision’s
intention to conserve the m angroves o f T anzania, therefore steps are being taken to reach th at goal. So
far the Division, through its established m angrove project, has done an inventory o f all the m angroves
of m ainland T an zan ia with the assistance o f N O R A D . Aerial photography and ground checks were used
to assess the state o f all the m angrove reserves in the country. Literature search and a socio-econom ic
study w as also undertaken.
The inventory revealed th at the area covered by m angrove vegetation is 115901 ha while the previ­
ous records indicate about 80000 ha. If the bare saline areas, w ater bodies, and salt pans which are part
of the m angrove reserves are included, the area am ounts to 172879 h a for the T an zan ia m ainland. The
m angrove reserves have already been m apped in 30 sheets. Some sheets are m apped using a scale of
1:25000 while others are m apped in 1:50000 scale. The m aps show the vegetation types, the area o f each
vegetation type, and the stand density and height o f each com partm ent.
The p aper therefore presents the current status o f the m angroves o f T an zan ia in respect to area,
distribution and vegetation types. Included are the strategies being used to develop the m angrove
m anagem ent plans for T anzania.

Introduction Large areas have been converted to agricultural

M ost m angrove areas in the w orld have received
very little attention in respect to m anagem ent.
Therefore large areas o f m any m angroves are
being depleted and degraded at an alarm ing rate
(Linden & Jernelvöv, 1980; M ercer & H am ilton,
1984). An inadequate knowledge o f the m angrove
ecosystem , together with the intense pressure on
the resource is the cause o f the high rate o f de­
pletion o f the m angroves (Saenger eta l., 1983).
land, praw n and fish p onds (FA O , 1982; M ackin-
non e ta l., 1986; C hua T hia Eng & Paw, 1987),
salt pans (Sem esi, 1988) or to other uses.
T he m angrove forests o f T an zan ia are currently
neither rationally exploited nor effectively p ro ­
tected. The F orestry Division has identified the
form ulation o f a m anagem ent strategy for the sus­
tainable use and conservation o f this areas as a
prim e task.
2

Description of the area w aves w hich are variable in strength depending

on the position w hether in protected areas or in
Location bays or exposed sites. B each erosion is prom inent
in exposed sites.
T an zan ia lies ju st below the equator, between
I o S—11 ° 4 5 ' S and from 29° 21' E - 4 0 ° 2 5 ' E.
Ocean currents
It covers an area o f 938000 km 2 (size o f F rance,
G erm any and Belgium com bined or six times the
T an zan ia is w ashed by a northw ard flowing cur­
size o f England). T an zan ia em braces the Islands
rent know n as the E ast A frican coastal current.
of Pem ba, Z an zib ar and M afia and has a co a st­
T he surface w aters are w arm , 2 5 -2 7 ° C w ith only
line o f about 800 km.
small annual tem perature variation. T he w ater
bathing the continental shelves o f T an zan ia and
Climate K enya is poor in nutrients and has a salinity of
34.5 ppt.
The coastal areas are hot and hum id w ith an
average tem perature ranging from 22 ° C to
Continental shelves
30 °C . H um idity is high throughout the year, up
to 90% during the rainy season. The clim ate from
The continental shelf o f T an zan ia has a total area
Ju n e-S e p te m b e r is pleasant, w ith little rain, lower
o f 18 000 km 2 with an average w idth o f 15-25 km.
hum idity, and is the coolest time. The m ain rainy
T he shelf is narrow from P angani to the K enya
season is from M a rc h -M a y and in N ovem ber to
border but south of Pangani the shelf w idens and
D ecem ber the area gets interm ittent short rains.
encircles the island o f Z anzibar. South o f Z a n ­
Lindi and M tw ara have the low est rainfall, less
zibar the edge o f the shelf again draw s near to the
th an 500 m m per annum while M afia has the
shore, and is only a few kilom etres distant outside
highest. M afia, Rufiji, K isaraw e, Z anzibar,
D a r es Salaam . South o f Kisiju, the shelf expands
Pem ba, M uheza and T anga have m ore than 95%
to include M afia Island and the Rufiji D elta. C oral
chance o f 750 m m per annum . Kilwa, Bagam oyo
grows extensively in shallow sections w here there
and D a r es Salaam have 8 5 -9 5 % chance of
are also num erous outcrops. The sh elfs sedi­
750 m m rainfall.
m ents are generally sandy but becom e m uddy
The country is influenced by M onsoon winds.
near river m ouths.
The N E M onsoon w inds (K askazi) blow from
O c to b e r-M arch and the SE M onsoons (Kusi)
blow in April to O ctober. The m onths o f M arch/ Drainage Ihydrology
April and O ctober/N ovem ber are calm est be­
cause they are the time for M onsoon reversals. The coastal plains o f T an zan ia are traversed by
The average w ind speed during N E M onsoon is perm anent rivers, seasonal rivers and by num er­
5 m s “ 1 and during SE M onsoon average at ous creeks. The perm anent rivers flowing to the
8 m s“ \ Indian O cean are: Pangani, W am i, Ruvu, Rufiji,
M atandu, M bem kuru, Lukuledi, and Ruvum a.
Rufiji is the largest river and drains 20% o f the
Tides
country.

T he coast o f T an zan ia experiences mixed sem id­

iurnal tides. The average tidal range at spring Soils
tides is about 3.2 m and at neap tides it is about
1.1m . The tide is of a standing wave type. The The soils in the coastal area and islands are pre­
tides and the w inds bring about locally generated dom inantly sandy and coralline with poor mois-

ture holding capacity, extrem e alkalinity and hard
subsoil which results in poor drainage. Along river
basins alluvium deposits are found.
Population
The total population of T an zan ia is 23174336
(1988 census). T an zania m ainland has 22533 758
inhabitants and the islands of Z anzibar and
P em ba have 640578 people. The growth rate for
1978-1988 was 2.8% and the m ean population
density is 25.3 km 2 for T anzania m ainland and
260 km 2 for Z anzibar.
Materials and methods
An inventory o f the m angroves o f T an zan ia was
carried out using aerial photographs which were
taken in 1988/89. They were supplem ented by
ground reconnaissance and topographic m aps to
assess the m agnitude o f the forests and to clas­
sify them into different vegetation types. 30 m aps
were printed depicting 10 adm inistrative blocks.
A socio-econom ic study on coastal com m unities
and how they use their m angrove environm ent
was carried out (von Mitzlaff, 1990). Finally a
w orkshop on m angrove m anagem ent brought to ­
gether resource persons and decision-m akers in
order to arrive at a consensus on a future policy
concerning the m anagem ent o f this resource.
These activities were carried out w ith the support
o f the F orestry Division and N O R A D .
Results
Mangrove forest distribution
M angroves occur along the entire coastline of
T anzania, from the K enyan border in the N orth
to the border of M ozam bique in the South, and
around the islands off the coast (Fig. 1). The
m angroves of M ainland T anzania, including
those o f the island o f M afia, were found to cover
a total area o f about 115475 ha. The largest area
3
of m angroves w as found in the Rufiji D elta. Fairly
large areas were also found at Tanga, Kilwa, at
the m ouths o f the Ruvu, W am i, Pangani and Ru-
vum a rivers.
The m angroves are distributed in the different
adm inistrative regions/D istricts as shown in T a­
ble 1.
Mangrove vegetation
T here are 8 species o f m angroves trees th at occur
in m ainland T an zan ia belonging to six families:
R hizophoraceae, Sonneratiaceae, Avicenniaceae,
C om bretaceae, Sterculiaceae and M eliaceae as
has been previously reported by Semesi (1986).
M ost com m on are Rhizophora mucronata and
Avicennia marina (Table 2). M ost im portant in
term s o f pole production is Rhizophora mucronata.
The outer zone is m ore often covered by
Avicennia marina and som etim es by Ceriops tagal
or Lumnitzera racemosa. The seaw ard zone is o c­
cupied by Sonneratia alba or Avicennia marina.
M oving landw ards the Sonneratia alba zone gives
way to the Rhizophora mucronata zone which is
alm ost im penetrable because o f the arched aerial
roots.
D ue to the lim itation o f fresh w ater com ing into
the upper intertidal sites some locations like parts
of the coast o f Tanga, Bagam oyo, Kilwa, Lindi,
and M tw ara, the concentration o f salts in the soil
m ay exceed the physiological tolerance o f all the
plant species and thus a vegetation-free sandy
area is found before other non m angrove halo-
phytes like Sesuvium portulacastrum and Sueda
monoica are observed. This area is norm ally
flooded by seaw ater twice m onthly during the
highest spring tides.
Mangrove fauna
The crabs are the m ost conspicuous invertebrates
o f the m angrove. T he upper zone is inhabited by
m arsh crabs (Sesarma). In the Ceriops and Bru­
guiera zone, Uca spp (the fiddler crabs) are the
m ost dom inant. The creeks are inhabited by Uca
4
TANGA ANCr i
MUHEZA
d i s t r i c t s
( BLOCK I )

PEMBA

PANGANI
DIS TRIC'
Bl o c k i

ZANZIBAR

b a g a Mo t o
O IS T R I CT
( BLOCK 3 )

•
}-DA
SAL AAffy
REV"

^BL OCK 4 )

K I S AR AWE
( BLOCK 5 )

MAFIA'
MAFIA___
"b D IS T RI CT __
(BLOCK 6*T

I d ist r ic t
(BLOCK 7 )

k ilw a o istric

Block/D istrict ibloc*»'

B o u n d ary .
International Boundary
Mangrove F orest
O ce a n ,W a ter

S C AL E I-.IOOO 0 0 0
1 0 0 KM
(BLOCK

■m t w a r a
/d istrict
.^ B L O C K I O )

the coast of mainland Tanzania.

distribution of mangrove forests along
Fig• ^ • The
 5

Table 1. Mangrove forest area in the different districts in are im portant. B arnacles and the spiny oyster
Tanzania.
(Crassostrea cucullata) are very com m on on stems
District/region Area quoted in Area found o f m angroves especially on Sonneratia trees. On
forest records inventory (ha) in 1989 (ha) the leaves o f m angroves especially those o f Bru­
guiera, Sonneratia and Ceriops are seen species of
1. M uheza and Tanga 9,996.1 9,403.3 (12,931.3)*
2. Pangani 600.0 1,755.6 (3,035.4)* Littorina.
3. Bagamoyo 1,499.0 5,635.8 (9,183.0)* Bees are visitors to the flowers o f the m an­
4. D ar es Salaam - 2,168.2 (3,213.3)* groves, the honeydew o f which attracts the weaver
5. Kisarawe 3,498.0 3,858.3 (6,050.3)*
6. Rufiji 40,470.0 53,254.8 (67,612.2)* or tailor ants, Oecophyllia which weave the leaves
7. Mafia 404.7 3,472.9 (4,365.0)* together to form a n est and are m ost com m on in
8. Kilwa 13,994.5 22,438.7 (36,747.5)* the Ceriops zone in M uheza.
9. Lindi 1,416.0 4,546.5 (7,300.8)*
10. M twara 7,996.9 8,941.5 (13,249.1)* Several dem ersal fish species inhabit the creeks
and the m ud skipper lives in the m angrove mud.
Total 79,875.2 115,475.6(163,688.9)* M angroves are also the feeding and hunting
* The figure in the brackets include also the area of water in places for a num ber o f birds: herons, egrets, ibis,
the creeks, saline bare areas, and salt pans. sea eagles, and several species o f w aders and
Kingfishers. The insect and crab eating kingfisher,
and Scylla serrata. On sandy portions soldier Halcyon senegaloiides is restricted to the m an­
crabs (Dotilla spp) are seen. Dotilla fenestrata is groves while the other birds are also found in
restricted to the E ast A frican coast. other habitats. T he area between the Rufiji delta
The m olluscs Potamides decollatus and Littorina and M afia island contains large num bers o f w ad­
scabra are the m ost com m on species found in the ers and terns during the w inter in northern E u­
m angrove form ation north o f Tanga. The m an­ rope (D anish-T anzania IC B P expedition to T an ­
grove snails like Terebralia, Cerithidea and Nerita zania from Ja n u a ry -M a rc h 1988). The m ost

Table 2. The relative composition of Mangrove vegetation types found along the coast of Tanzania.

Classification Area ha V
/o

Rhizophora dominant with occurrences of Avicennia, Ceriops, Sonneratia, Bruguiera, 55549.9 48

Heritiera and/or Xylocarpus
Sonneratia almost pure stands 1223.3 1
Sonneratia dominant with Avicennia, Bruguiera and/or Rhizophora 6123.2 5
Heritiera - almost pure stands 91.2 0
Heritiera dominant with Avicennia, Bruguiera and/or Rhizophora 8188.4 7
Avicennia dominant with Rhizophora, Bruguiera, Heritiera, Ceriops and/or Xylocarpus 17141.6 15
Avicennia - almost pure stands 1687.4 1
Mixture of Avicennia and Cerips 17432.7 15
Ceriops dominant with occurrences of Rhizophora, Avicennia and/or Bruguiera 8037.9 7

Total Mangrove 115475.6 100

W ater in creeks 24076.0

Clear cut areas 4435.0
Saline, bare areas 20740.0
Salt pans 3093.0
Non-mangrove forest 5069.3

Total reserve area 172888.9

num erous species counted were Curlew sandpiper
15000; Little Stint 8000; C rab plover 5400; and
R oseate Tern 5 500. Lesser B lack-backed Guii,
H erring Guii and C aspian Tern also spend the
w inter in the Rufiji delta.
H ippopotam us, m onkeys and wild pigs are
com m on m am m als in the m angroves. Crocodiles
are found in the m ajor rivers. H ow ever very lit­
tle inform ation is available on the fauna found in
the m angroves o f T an zan ia and as such we ca n ­
not identify their critical habitats within the re­
serves.
Uses o f mangroves
M angroves are an item o f trade and they are a
source o f em ploym ent and incom e for the coastal
com m unities: the w ood is used as fuel for com ­
m ercial production o f salt, lime and for the p ro ­
cessing o f fish. M angroves are cut for poles for
export and as local building m aterial. M angrove
areas are used for the construction o f solar evap­
o ration pans for the production o f salt and in the
Rufiji D elta they are turned into rice fields. Lish-
eries, and the praw n catches in particular, depend
on an intact m angrove ecosystem . M angroves are
the source o f dom estic fuelwood and o f raw m a­
terial for artisans and boat-builders; fishermen
m ake trap s and fish floats from the trees, some
species are used for anim al fodder, vegetables
and others for local m edicine. Rarely bees are
kept in these forests.
Potentials o f the mangrove resource
Presently m ost o f the pole cutting is done illegally
and the royalty fees collected in 1988, am ounting
to about 10 M illion T S hs, were only a fraction of
the actual am ount o f revenue due to the govern­
m ent. 15 M illion T S hs were collected as revenue
from the export o f praw ns in 1989. It has been
estim ated th at the present praw n fishing does not
net even a quarter o f the possible catch in T an ­
zan ia (Siegel, 1986). In term s o f revenue, bu t also
in term s o f incom e and em ploym ent, the m arine
resource has n o t been fully exploited.
Salt production by solar evaporation is taking
place in the m angrove areas but is not run effec­
tively and could be m uch higher. Similarly the
potential o f this resource for tourism , for com ­
m ercial honey and w ax production, and for any
kind o f aquaculture has no t yet been realized at
all in T anzania.
Problems and threats to the mangrove ecosystems
Ecological issues
Though the m ain th reat to the m angroves com es
from m an, there are also natural causes for their
death.
- River floods m ay cause considerable areas
of m angroves to die as a result o f alteration of
w ater level, ban k erosion and diversion o f w ater
courses. M angroves cannot tolerate long inunda­
tion and will die. This phenom enum can clearly
be observed in som e parts o f the Rufiji D elta.
- Sand deposition from sea and land can cut
off portions o f m angroves from salt w ater caus­
ing the m angroves to die. This problem is p ro ­
nounced in m ost p arts o f Bagam oyo D istrict.
- Sea level rise (because o f the global w arm ­
ing) would flood the present m angrove areas and
kill all m angroves (H ekstra, 1989). H ow ever they
m ay soon colonize new land.
Social and economic issues
It is m ainly com m ercial activities which are over-
exploiting the m angrove resource w hereas the
coastal village com m unities use their m angrove
environm ent m ostly in a sustainable way.
The destruction o f large areas o f m angrove for­
ests is caused by business people who usually
com e from inland regions. They either cut the
m angroves for boiling brine for the production of
salt (Tanga), for lime burning (Bagam oyo, Lindi,
M tw ara) or for fish processing (Pangani). They
use the resource w ithout caring about its survival
or reproduction, the com m unities are not enu­
m erated for the loss o f their resource and alm ost
none o f the profit gained from these enterprises
rem ains in the villages.
This is similarly true for the construction o f salt

pans for solar evaporation in Tanga, Bagam oyo
and M tw ara D istricts. They are often placed
within m angrove areas which entails the total
clearing of the forest, whilst topography and soil
are equally suitable for their location behind m an­
groves or onbare patches.
Pole cutting, if done in an unplanned way and
on a large scale, can reduce the genetic pool and
endangers reproduction.
The dem and for dom estic fuelwood endangers
the m angroves only in densely populated areas
around tow ns. If the inland resource is already
exhausted, people turn to m angroves as a second
choice for fuelwood.
Agricultural activities interfering with the m an­
grove environm ent take place in the Rufiji D elta
only and this quite recent developm ent has a p ar­
ticular historical background. The practice of
shifting cultivation for rice farm ing is becom ing a
serious th reat to the m angroves.
The m angroves are also endangered by oil pol­
lution and industrial pollution, by excessive silt-
ation because o f m an’s degradation o f catchm ent
areas, by herbicides com ing from inland via the
rivers, and by the construction o f dam s or m ajor
irrigation schemes.
The aw areness o f the overall im portance o f this
resource is low am ongst the people in general.
They value it rather for its direct m onetary return
than its m any indirect benefits.
Institutional issues
H um an activities threatening the m angrove envi­
ronm ent should be controlled by the respective
governm ent bodies. There are several institutional
problem s which ham per this control:
- N o overall authority exists which effectively
coordinates conflicting issues like salt licenses
and land titles concerning the Division o f F o r­
estry, the Division o f L ands and the M inistry
o f W ater, Energy and M inerals sim ultaneously.
- Lack of adequately trained personnel, very low
num bers of field personnel, lack o f tran sp o rt
and infrastructure renders actual control
through patrolling impossible.
- The knowledge on the m angrove ecosystem
and o f activities or technologies related to it is
7
inadequate and dispersed to different institu­
tions (e.g. Forestry, Fisheries). N o institution­
alized channels o f com m unication have been
established to overcom e this problem .
Policy issues
The m anagem ent policy which consist o f control
and prohibition only, w ithout having the m eans
and capability to carry it out (lack o f funds, per­
sonnel, tran sp o rt and infrastructure), has affected
the m anagem ent o f the m angroves in Tanzania.
O n top o f this it has no t considered traditional
interests o f people in their resource or their in­
volvem ent in its m anagem ent. The N ational ban
on cutting m angroves spelled out in 1987 has poi­
soned the relationship betw een people and gov­
ernm ent bodies even further.
Legal issues
D espite the ban, licenses for cutting poles and for
the construction o f salt pans are issued. C onse­
quently the users o f the m angrove resource, be it
villagers or business people, are confused about
the actual legal situation and their rights. Also
they lack any pow er to lobby for their interests.
The traditional rights o f the coastal village
com m unities to use their resource for their small
local needs have been disregarded although they
have never been a threat to the m angroves.
Limitations and constraints encountered during
management planning
The following lim itations and constraints have to
be recognized:
- N o national strategy for conservation exists
w hich could provide a fram ew ork for the for­
m ulation o f the m angrove m anagem ent plan.
- Lim ited knowledge on the resource concerning
ecological, botanical and zoological m atters is
a m ajor obstacle for planning.
- Low salaries, lack o f incentives, tran sp o rt and
equipm ent are om nipresent and inextricable
problem s which ham per m otivation and con­
tribute to low efficiency.
- Law s do no t necessarily guarantee protection
 and lack o f forest patrols and law enforcem ent
is the m ain hindrance in resource protection.
Identified management objectives and strategies
The purpose o f the m anagem ent plan is to iden­
tify a program m e and its objectives and strategies
in order to assist the governm ent in its endeavour
to rationally use and conserve the m angrove re­
source o f T anzania. It will be taken as a basis for
a plan o f operations. The overall developm ent
goal is to enhance the contribution o f m angrove
ecosystem s to the econom y o f the country and the
objective o f the m anagem ent program m e is the
rational utilization o f m angrove ecosystem s on a
sustainable basis.
T o achieve this objective the following short
term objectives have to be realized:
- to conserve m angrove areas with essential p ro ­
tective functions
- to m aintain h ab itat for fauna and flora o f this
unique ecosystem
- to optim ize the com bination o f direct and in­
direct uses o f m angrove areas
- to m eet needs o f com m unities living in and
depending on m angrove environm ent
- to conserve representative areas for biodiver­
sity, tourism , research and education
- to m aintain m angrove gene pools
- to im prove the institutional set up
- to support research activities
M anagement strategies
The strategies to achieve the above m entioned
objectives are geared tow ards the institutional as
well as the operational level. First and forem ost
the capabilities o f the F orest Division will be
strengthened and enlarged. A new section within
the F orest Division will have the overall author­
ity for the m anagem ent o f the m angrove resource.
The num ber o f field staff will be drastically in­
creased, (mostly to be recruited from coastal vil­
lages) and will be provided with appropriate
training in all m atters concerning m angroves
(m anagem ent, extension w ork, harvesting etc.).
Their equipm ent, tran sp o rt and other infrastruc­
ture will be im proved. Training for high level per­
sonnel will also be provided.
Lack o f coordination betw een the m any differ­
ent governm ent institutions directly or indirectly
involved in issues concerning m angroves h as been
identified as a m ajor hindrance to successful
m anagem ent.
T he governm ental institutions directly or indi­
rectly involved in issues concerning m angroves
(Division o f L ands, Fisheries Division, F orest
Division, M inistry o f W ater, Energy and M iner­
als, M inistry o f Local G overnm ent) and potential
resource persons will be represented in a steering
com m ittee which is directly linked to the D irec­
to r o f F orestry and the coordinator o f a N ational
M angrove M anagem ent Project. This steering
com m ittee, as the linking body to the m inistries,
will have to deal with the conflicting interests and
responsibilities of their institutions and agree on
a hierarchy o f decision-m aking and com m unica­
tion. Problem s arising from unclarified legal is­
sues have to be solved before any new policy can
be form ulated.
In order to find a consensus on such disputed
legal issues the respective governm ent bodies
concerned will be brought together by the D irec­
tor o f Forestry to discuss and decide on the nec­
essary steps to be taken. This decision will be
widely publicized in order to ensure the different
governm ent institutions as well as the users o f the
m angrove resource know about their rights and
responsibilities.
The participation o f the coastal people is of
p aram ount im portance for the success o f the
m anagem ent o f the m angrove resource. Only if it
is an asset to them they will defend it against il­
legal exploiters. Different approaches will be used
to seek their participation:
- local people will be involved actively in the
identification o f zones for different uses near
their villages. N o t only has this m erit as a kind
o f aw areness cam paign, but it will also give the
villagers a chance to voice their interests.
- certain areas will be allocated to villages for
their local use (firewood, raw m aterial).

- other areas will be allocated to individual vil­
lagers or groups for com m ercial use, i.e. they
will have preferential access to licenses for the
exploitation o f the resource (pole cutting, salt
production).
- training on the m anagem ent by rationally ex­
ploiting the resource will be available for those
who are engaged in it. This extension service
has to be provided by the M angrove F orest
Officers.
- villagers will be given support in developing
alternative econom ic opportunities. The Bee­
keeping Division will encourage and assist vil­
lage w om en and m en in the establishm ent o f a
honey and wax industry, in its processing and
m arketing. T he Fisheries Division will assist in
im proving the situation o f the artisanal fisher­
m en through the supply o f gear or through
loans, through training and through im proving
technologies o f processing fish (the latter con­
cerning w om en mainly).
- com plem entary m easures, like establishing fu­
elw ood plots for villages or individuals, will
receive every possible support.
- infrastructure im provem ents will benefit the
villages as well as the m angrove m anagem ent.
- villagers will be em ployed as M angrove F orest
G u ard s
- contracts will be m ade betw een the villagers or
groups o f villagers involved and the F orest D i­
vision with regard to their rights and obliga­
tions o f exploiting and protecting the resource.
U ltim ate authority lies with the F o rest D ivi­
sion who can w ithdraw these rights if the ob­
ligations are n o t met.
- through the establishm ent o f U sers’ A dvisory
B oards villagers and other com m ercial users
will be given an opportunity to voice their in­
terests.
Zoning as a management strategy
The m angrove forests are o f very different quali­
ties and occur under varying conditions. In order
to optim ize their direct and indirect uses they
have been categorized in four zones.
9
Zone I are all forests which will receive prime
protection.
Zone I I are all ecologically stable areas with
sufficient regeneration potential. They are consid­
ered productive forests suitable for controlled
harvesting.
Zone I I I are degraded areas which will be
closed for som e time to allow recovery and reha­
bilitation.
Zone I V are areas m arked for ‘development’ of
different kinds.
Conclusion
Ecological changes, uncontrolled exploitation by
m an, uncoordinated and inadequate institutional
m easures, deficiencies in policies and an unclear
overall policy have all contributed to the present
situation, where the survival o f the m angrove for­
ests o f T an zan ia is in doubt. T here is urgent need
for planned, coordinated action, there is need for
a new policy and there is need also for an orga­
nizational set-up which can m eet these tasks.
Acknowledgements
I w ould like to thank the N orw egian Agency for
International D evelopm ent (N O R A D ) for finan­
cial support and the U niversity o f D ar es Salaam
for granting m e perm ission to w ork on the m an­
grove project. M any thanks should also go to
m essrs A. M itoi, S. K ipangula, A. M tw eta for the
airphoto interpretation, to M r. H ak an Sjöholm,
C oordinator o f C atchm ent F o rest Project for his
valuable discussions. D r. Von M itzalaff, Dr. K.
Howell, D r. F. U rasa, D r. A. Kivaisi and M an ­
grove W orkshop P articipants m ust be thanked
for their critical com m ents.
References
Chua Thia. Eng & J. N. Paw, 1987. Aquaculture Develop­
ment and Coastal Zone Management in S.E. Asia: Con­
flicts and Complementary. Proc. of the Fifth Symp. on
Coastal and Ocean Management. 2: 2007-2019. Elsevier
Scientific Publishing Company, Amsterdam.
10
FAO, 1982. Management and Utilization of Mangroves in
Asia and the pacific. FAO Environment paper No. 3. Rome:
FAO.
Flekstra, G. P., 1989. Global warming and Rising Sea levels:
The Policy Implications. The ecologist, 19 (1): 4-15.
Linden, O. & A. Jernelov, 1980. The Mangrove Swamp - An
ecosystem in Danger. Ambio, 9: 81-88.
Mackinnon, J., K. Mackinnon, G. Child & J. Thorsell, 1986.
Managing Protected Areas in the Tropics. IU C N Switzer­
land.
Mercer, E. & L. S. Flamilton, 1984. Mangrove ecosystems:
Some economic and natural benefits. Mar. Resour.
U N ESC O , 20: 14-19.
Von Mitzalaff, U., 1990. Coastal communities in Tanzania
and their mangrove environment. A socio-economic study
prepared for N O R A D -D ar es Salaam, 60 pp.
Saenger, P., E. J. Hegerl & J. D. S. Davis (eds), 1983. G lo­
bal status o f mangrove ecosystems. Commission on Ecol­
ogy Papers Number 3. G land Switzerland: IU C N . The En­
vironmentalist 3. Supplement No. 3.
Semesi, A. K., 1986. Zonation and vegetation structure of
mangrove communities in Tanzania. In: Mainoya J. R. &
Siegel P. R. (eds), Proc. of a W orkshop on Save the M an­
grove Ecosystems in Tanzania. 21-22nd February, U ni­
versity o f D ar es Salaam. 174-181.
Semesi, A. K., 1988. Status and utilization of mangroves along
Tanga, Tanzania. In: Mainoya J. R. (ed.), Proc. of a Work­
shop on Ecology and Bioproductivity of the Marine Coastal
W aters of Eastern Africa. 18-20th January, University of
D ar es Salaam. 174-181.
Hydrobiologia 2 4 7 : 11-16, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 11
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Structure and species distribution in Coringa mangrove forest, Godavari

Delta, Andhra Pradesh, India

Jayapaul A zariah, H ilda A zariah,* S. G unasekaran & V. Selvam

Department o f Zoology, University o f Madras, Guindy Campus, M adras - 600 025, India; * C A S In
Botany, University o f Madras, Guindy Campus, M adras 600 025, India

K ey words: C oringa m angrove, structure, distribution, atm ospheric pollutants, industrial im pact

Abstract

Coringa m angrove forest is located in the G odavari delta, A ndhra P radesh, India. The m angrove
com m unity consisted o f m ore than 13 species o f m angrove and other plants in the present study area.
The following three dom inant m angrove plants, Avicennia marina, Excoecaria agallocha and Sonneratia
apetala were found to be present on the banks o f a m ajor channel o f the G odavari river running through
the forest. The structure and species distribution o f m angrove, in the C hannel N ag ath an a K alava has
been described. T he area behind the belt consisting o f Acanthus ilicifolius and Myriostachya wightiana is
generally colonized by E. agallocha and A . marina. The zone has been called the A vicennia and Excoecaria
zone. A djacent to this zone species like Aegiceras corniculatum and A . officinalis were the com m on
species. In the flat clayey soil, Suaeda maritima w as found to grow. In areas o f high elevation, devoid
of inundation o f tidal seaw ater during the high tidal period, species such as M . wightiana and Acanthus
were found to colonize both the banks o f the channels.
A n analysis o f species diversity, indicated a definite trend in the distribution o f m angrove from the
m outh o f the estuarine region to the inland waters.
The levels o f atm ospheric pollutants such as sulphur dioxide ( S 0 2), oxides o f nitrogen (N O x), am ­
m onia (N H 3) and suspended particle m atter (S P M ) were within the legal limits.

Introduction is located on the southern coast o f the K akinada

A survey o f the global status o f m angroves may
show th at vast areas are being destroyed either
intentionally or as a secondary result o f other
activities (Saenger et al., 1983). The m angrove of
In d ia also is no exception to such an uncontrolled
exploitation. The m angrove vegetation o f India
have an area o f approxim ately 356500 h a o f
which K rishna-G odavari m angrove com plex
com prises an area o f approxim ately 12 800 ha
(Sidhu, 1963). The C oringa m angrove forest,
which is a p art o f the G odavari m angrove forest,
tow n in A n d h ra Pradesh, where the present in­
vestigation w as carried out. T he K ak in ad a port is
a natural harbour and an all w eather port which
can provide protection from severe w eather co n ­
ditions. H ence, it is possible to handle ships for
m ore than 300 days per year. Currently, loading
and unloading are carried out around the clock.
Recently, the G overnm ent of A n d h ra P radesh has
developed a m aster plan to convert the K akinada
p o rt into a m ajor po rt so as to m ake it a gateway
for trade and export. The proposed developm en­
tal plan involves the construction o f three deep
12
w ater berths, extension o f break w ater to 3.35 km,
dredging o f access canal and provision o f ancil­
lary facilities such as road, approach bridges and
other navigational aids. T he project, costing R u­
pees 1335.8 million, is expected to be com pleted
by April 1995 (Indian Express, 1990).
A second m ajor developm ental activity is the
building o f the fertilizer com plexes namely, G o ­
davari Fertilizers and Chem icals L td (C apital cost
o f R upees 1080 million, w ork com pleted) and the
N ag arju n a Fertilizer and Chem icals Ltd (capital
cost o f R upees 6700 m illion-work nearing com ­
pletion). These tw o fertilizer factories are ex­
pected to m eet the fertilizer dem ands within a
200 km range from the factory. In view o f such a
rap id pace o f industrialization in the coastal zone
o f K akinada, a detailed study on the environm en­
tal im pact assessm ent was m ade on a m ultidis­
ciplinary basis. In an earlier paper the results of
the analysis o f the m angrove vegetation within a
15 km radius o f the fertilizer com plex have been
given (A zariah eta l., 1990). The present paper
examines the structure o f the m angrove com m u­
nity and its distribution in the C oringa forest,
w hich is located on one of the deltaic branches of
the northern zone o f the G odavari delta, so as to
provide quantitative baseline inform ation on the
current status o f the ecosystem together with a
survey o f the current levels o f atm ospheric pol­
lutants such as sulphur dioxide, oxides o f nitro­
gen, am m onia and suspended particulate m atter.
Description of site
M angrove forests are seen in places like Coringa,
Y anam , K andikuppa, K akinada and S acram ento
which are in the E ast G odavari delta. The C or­
inga river originates at Y anam which is about
6 km on the upper reach o f G odavari river. This
tributory along with G adeur form s a p art o f the
G o d av ari estuarine system. B oth the tributaries
open into the southern side o f the C oringa bay.
The bay is nam ed after C oringa (Koriangi, a well
know n ancient port). A group o f islets are located
in the vicinity o f C oringa bay. Am ong these, H ope
Island, B hairavapalend series and C oringa and
tributaries, G ad eru and V asavakalava are note­
w orthy (see Fig. 1).
M aterials and methods
T he present study w as carried out in the N agath-
ana K alva (canal) w hich extends for a distance of
10 km from the C oringa river. T he area is swam py
w ith knee deep m ud.
A series of quadrats o f 5 x 5 m (Sidhu, 1963)
were laid at an interval o f 1 km along the 10 km
long canal. The num ber o f trees in each quadrat
w as counted and the d ata were tabulated. U sing
certain indices o f species structure in com m unity,
d ata on the index o f dom inance, species diversity,
richness o f species were calculated. Besides, d ata
on the relative density, relative frequency and
relative dom inance were calculated. D etails o f the
form ulae used are given in an earlier paper (A z­
ariah et al., 1986). A few m icrom etereological as­
sessm ents were carried out to obtain baseline in­
form ation on the level o f sulphur dioxide, oxides
of nitrogen, am m onia and suspended particulate
m atter. The details o f the analysis are given in an
earlier paper (A zariah e ta l., 1990).
.1Mi le
Kakinada d Mangrove forest
Town
■ St udy area
Kaki nada
Bay
-16 55N
Fig. 1. M ap showing the location of the Coringa mangrove
forest and the study site.
 13

Results higher at K akinada during the post m onsoon sea­

Concentration o f atmospheric pollutants
A biweekly sam pling o f atm ospheric pollutants
were m ade using a high volum e sam pler. The
studies were carried out from D ecem ber 1988 -
M ay 1989 (Table 1).
The S 0 2 concentration varied from zero to
10 pg m ~ 3 over a period o f 6 m onths. The values
didn’t exceed the m axim um adm issible limit o f
80 pg m - 3 . The N O x values varied from zero to
98 ,ug m ~ 3. The maxim um value w as observed
during the m onth o f Septem ber 1988, which m ay
indicate th at there were occasional outburst of
oxides o f nitrogen. It m ay be seen th at the levels
of N O x exceed the m axim um adm issible limit
(80 / i g m ' 3). The concentration o f am m onia v ar­
ied from zero to 28 pg m ~ 3 against the adm issi­
ble limit o f 30 ,ug m ~ 3. It can be seen th at the
levels o f am m onia were nearing the critical co n ­
centration. The total suspended particulate m at­
ter (S P M ) varied betw een 31 and 1 3 7 p g m ~ 3,
which is well below the adm issible limit of
200 f i g m " 3. The dust fall is low, the value being
9.8 metric tonnes k m ~ 2 per m onth.
A com parison of d ata between Coring village
and K ak in ad a neighbourhood w as m ade. It can
be seen th at the levels o f the S 0 2 were higher in
the C oringa m angrove forest than at K akinada
neighbourhood during the sum m er season, i.e.,
M arch to M ay. It can be noted th at the level was
son (D ecem ber to February). A sim ilar pattern
m ay be seen with respect N O x. The concentration
of am m onia and SPM w as higher during both the
sum m er and post m onsoon m onths. A similar
pattern w as observed with respect to S PM (See
Fig. 2).
Nature of environmental pollutants in the vicinity
of Kakinada (Table 1)
Level o f sulphur dioxide
In an earlier study, it has been found th at the S 0 2
concentration did not exceed the limit throughout
the year (A zariah et a i, 1990). N otable variations
in its concentration were found between the sum ­
m er and N o rth E ast m onsoon (June to August)
seasons at K akinada tow nship with the values o f
1.9 and 4.4 pg m ~ 3 respectively. D uring the sum ­
m er the concentrations o f w as m inim al in the
entire vicinity o f K akinada, the value being
1.9 pg m ~ 3 (Table 1).
NO x
The concentration o f N O x exceeded the adm is­
sible limit in the K akinada tow n but it rem ained
within the adm issible limit in the coastal zone
during the sum m er (M arch to M ay) season. A
relationship between N O x and hum an activity

Table I. Concentration of atmospheric pollutants (/ig m 3) in Coringa forest and K akinada neighbourhood

Period so2 NOx NH3 SPM

Coringa:
Dec 88-F eb 89 1.3 ± 2.4 10.6 ±5.7 14.4 ±8.1 64.9 ±23.8
(0-3.0) (2-20.2) (0-28) (31-97)
M ar 89-M ay 89 3.6 ±2.9 36.0 ±27.0 13.0 + 4.7 80.5 ±31.6
(0-10) (0-98) (0-19) (38-137)
Kakinada neighbourhood:
Dec 88-Feb 89 2.76 ±2.83 46.9 ±30.5 31.0 ± 17.3 104.03 ±52.8
(0-20) (0-330) (0-180) (24-294)
M ar 89-M ay 89 1.9 ± 2.26 45.1 ±36.2 9.3 + 9.46 106.86 ±56.8
(0-12) (0-276) (0-81) (28-302)
Maximum admissible limit 80 /tm m 3 80 fig m ""3 30¿ig m 3 200 /tg m ~ 3
14

0 DEC - FEB A CORINGA nh 3

KAKINADA SPM
The am m onia concentration exceeded the adm is­
90
sible limit, at all times, except in the coastal zone
during the sum m er season. The m axim um con­
centration, 1 8 0 /¿g m _3, w as found during the
E p ost m onsoon season in the K akinada and lower
C71
levels (0 to 8.2 /tg m _ 3) were recorded during the
nh3 sum m er season in the coastal zone.

SPM

A B
A B A B
Fig. 2. Concentration of atmospheric pollutants in the Kaki­
nada town and Coringa mangrove forest.
w as found. The concentration was higher in
places where there were m ore hum an activities in
term s o f handling nitrogen based fertilizers (load­
ing and unloading). D uring the sum m er m onths,
the coastal zone registered a value ranging be­
tween zero and 40 /ig m “ 3 and it varied from 2
to 330 f t g m “ 3 during the post m onsoon season
i.e. D ecem ber to February.
The concentration o f suspended particulate m at­
A B
ter exhibited a direct relationship with hum an ac­
tivity. It’s level w as high both during the post
m onsoon and sum m er seasons, the values being
24 to 294 /ig m -3 and 28 to 302 /ig m ~ 3 respec­
tively.
Analysis o f mangrove ecosystem
The num ber o f plants recorded in 10 quadrats
along the distance o f 10 km are given in Table 2.
The vegetation in the m angrove sw am p consisted

Table 2. Structure of Coringa Mangrove Community, East G odavari

SI. Species Quadrat number Total Frequency Relative Relative

no. number density frequency
1 2 3 4 5 6 7 8 9 10 of (% ) (% ) (% )
individuals
N umber of individuals

1. Excoecaria aquallocha 25 9 38 28 8 17 22 24 8 9 188 100 30.62 17.54

2. Avicennia marina 6 27 7 14 4 11 4 12 11 6 102 100 16.61 17.54
3. Avicennia officinalis - - - - 2 6 7 3 9 7 34 60 4.4 8.77
4. Aegiceras corniculatum - 14 3 4 - 2 - - 4 - 27 50 5.54 10.53
5. Acanthus ilicifolius 6 17 17 23 63 40 10.26 7.02
6. Sonnaretia apetala 4 8 9 21 30 3.42 5.26
7. Derris trifoliata 4 6 2 3 15 40 2.44 7.02
8. Ipomea tuba 3 4 3 6 16 40 2.61 7.02
9. Cyperus rotundus 33 22 - - 55 20 8.96 3.51
10. Clerodendrum inerme 2 3 5 20 8.96 3.51
11. Ceasalpinia nuga 1 3 8 20 0.81 3.51
12. Myriostachya wightiana 4 2 6 20 1.30 5.26
13. Suaeda maritima 62 12 74 20 12.05 3.51

Total 93 62 48 46 14 36 79 93 71 72 614

of 13 species o f plants out o f which 11 species
belonged to the m angroves. Secondly, their dis­
tribution varied in a distinct pattern from the
m outh o f river tow ards the inland water.
The distribution o f Avicennia officinalis Lam .
w as restricted tow ards inland w aters whereas
Excoecaria agallocha Lam . and Avicennia marina
(F orsk.) Vierh. were found along the entire stretch
of the river i.e. from the coast to the inland w a­
ters. Between E. agallocha and A . marina the
dom inance o f the form er w as seen throughout the
study area. H owever, E. agallocha was m ore
thickly distributed tow ards the m outh o f the river.
On the other han d for the latter species, although
distributed throughout the study area, the m ag­
nitude o f colonization was relatively less than
E. agallocha.
The zone o f m angrove can be divided into two
areas. First, the Avicennia area which is located
near the inland w ater and the second region dom ­
inated by E. agallocha. This region w as found
m ore tow ards the ocean front. A feature o f inter­
est is that, when the num ber o f E. agallocha in­
creased the num ber o f Avicennia decreased. It is
also interesting to note th at Avicennia officinalis
was n o t distributed within the area o f 4 km from
the m outh o f the river. It is likely th at this species
may prefer a inland habitat with lesser m agniture
of salinity variation. W hereas A. marina did not
exhibit such a relationship. A non m angrove spe­
cies Ipomaea tuba Linn, which w as closely asso­
ciated with m angroves show ed a discontinuous
distribution. N ine species (item N o. 5 to 13 in
Table 2) showed a salinity related distribution in
th at they were represented in regions the river,
where there m ay be wide fluctuations in salinity,
and were recorded only in quadrat num bers 6 to
12. T he q u adrats 1 to 3 were located near the
m outh which w as characterised by the presence
of neritic w aters throughout the year due to its
proximity to the sea w hereas quadrat num bers 7
to 12 were located in the inland area where fresh
w ater condition existed during the m onsoon sea­
son and brackish w ater regime found during the
post m onsoon and sum m er seasons (S arm a &
G anapti, 1971). Myriostachya wightiana H ook was
found in habitats which were located m ore to ­
15
w ard the inland w aters. O n the other hand,
Suaeda maritima D um ort w as recorded in the first
tw o quadrats w hich were laid very close to the sea
coast.
In general, the following three distinct zones of
distribution can be m ade.
Z one I: Excoecaria agallocha - Suaeda maritima
zone, which extends 1 to 2 km from the river
m outh.
Z one II: A m iddle zone in which a mixed com ­
munity of Excoecaria, Avicennia and Aegiceras
corniculatum L. m ay be found. This zone extends
upto about 6 km from the river m outh.
Z one III: W as represented by Excoecaria, Avicen­
nia and Cyperus which were found at a distance
o f 7 -8 km from the sea. In this zone a mixed
com m unity o f Excoecaria, Avicennia and A can­
thus m ay be recognized.
It is interesting to note th at Suaeda maritima
was found only near the swam py regions o f the
coastal zone along with Excoecaria and Avicen­
nia. A nother interesting feature is the occurrence
of the Rhizophora apiculata Blume and Lumnitzera
racemosa Willd. B ut these were no t present within
the 10 quadrats. Rhizophora apiculata was repre­
sented by a single tree, which w as located in the
m iddle stretch o f the study area. W hereas L. race­
mosa w as found to be sparingly distributed.
The species diversity varied from 0.92 to 3.13
with the highest level tow ards the inland waters.
Similarly, an ascending value in species richness
w as recognized from the m outh o f the river to ­
w ards the inland water. W here Excoecaria and
Avicennia dom inated, as in the case o f first 3 km,
the richness index was p oor ( 1.02 to 1.67) w hereas
it ranged between 4.85 and 5.4 in the region of
freshw ater regime. The species evenness index
w as relatively high in areas o f low salinity (T a­
ble 3).
Discussion
Figure 2 com pares the concentration of atm o­
spheric pollution between K akinada tow n and its
neighbourhood and the C oringa forest. The vari­
ation in the atm ospheric pollution is a function of
16
Table 3. Species diversity, species richness and species eveness values in different quadrats
Indices Q uadrat number
1 2 3 4
Coast line
Species diversity 1.15 1.81 0.92 1.26
Species richness 1.02 1.67 1.19 1.20
Species eveness 0.73 0.90 0.58 0.76
the influence o f m onsoon accom panied by the
direction o f wind speed. The behaviour o f S 0 2
and N O x w as similar in th at the concentration of
these com ponents w ere higher in C oringa during
the sum m er season w hen the wind w as blowing
in the SE and SW directions. It is likely th at w ind
borne S 0 2 and N O x m ay be concentrated at the
Coringa reserve forest. O n the other hand, the
concentration N H 3 and S P M w as higher at the
K ak in ad a tow n which is due to wide spread
hum an activity and the use o f nitrogenous fertil­
izers.
The species structure in the m angroves o f
C oringa appears to vary from site to site. R ao
et al. (1963) and V enkatesan (1966) reported
the occurrence of 26 species in the neighbour­
hood o f K akinada. O f which Suaeda maritima,
Heliotrophium curassavicum L., Fimbristylis
cymosa R. Br. and Aeluropus lagopoides (L.) Trin.
occur in considerable num bers. In the C oringa
forest, B lasco (1975) recorded the occurrence of
Excoecaria agallocha, Avicennia officinalis, Phoenix
paludosa, Derris sp., Dalbergio sp. and Ceasalpinia
sp. L ater Sidhu (1963) recognized three m ajor
forest com m unities. They are tidal ever green,
semi-evergreen and tidal decidous. H e recognized
pure Avicennia com m unity, mixed Avicennia com ­
m unity and Aegiceras mixed com m unity. H e also
recognized a total num ber o f 20 species in his
studies carried out the at Coringa C oup N o. 2
and 4 (coup = area alloted by forest authorities
for clear felling) and various region within the
C oringa reserves such as G undhuhalla, Varava-
calea and P adavakpalam . A n interesting and
5 6 7 8 9 10
Inland water
1.42 1.69 2.52 2.69 3.13 2.95
1.75 1.93 3.16 4.05 5.40 4.85
0.90 0.84 0.90 0.85 0.90 0.89
contradictory feature betw een the earlier and the
present study is the variation and the dom inance
o f species within the com m unity. In the above
areas o f study, Avicennia w as the dom inant spe­
cies, w hereas in the present study, Excoecaria was
found to be the dom inant species and no pure
stand o f Avicennia w as recognized at any place.
A nother feature o f difference is the association of
Avicennia w ith Acanthus.
References
Azariah, J., H. Azariah, P. Pandion & S. Gunasekaran, 1990.
Coastal wetlands of Kakinada, A ndhra Pradesh. A case
study in environmental impact assessment. In W etland
ecology and management, Bharatpur (in press).
Azariah, J., P. Banth, H. Azariah & V. Selvam, 1986. Impact
of urbanization on the status of mangrove swamps of M a­
dras. In D. D. H ook etal., (eds), The Ecology and M an­
agement of W etlands, Croom Helm, London: 225-233.
Blasco, F., 1975. Mangroves of India. Inst. Fr. Pondicherry.
Tran. Sect. Sei. Tech., 15, 175 pp.
Indian Express, 1990. K akinada Port-Gateway for trade and
industry. March, 2: 1-15.
Rao, T. A., K. R. Agerwal & A. K. Mukherje, 1963. An eco­
logical account of the vegetation of Rameswaram Island.
Bull. Bot. Surv. India, 5: 301-323.
Saenger, P., E. J. Hegal & J. D. S. Davie, 1983. Global sta­
tus of Mangrove ecosystems. Commission on Ecology,
Paper No. 3, IU C N Gland. 88 pp.
Sarma, D. V. R. & P. N. G anapti, 1971. Hydrography of the
Coringa river of the G autam i-G odavari estuarina system. J.
mar. biol. Ass. India. 13: 234-248.
Sidhu, S. S., 1963. Studies on the mangrove of India. I. East
G odavari region. Indian Forestar. 89: 337-351.
Venkatesan, K. R., 1966. The mangroves of M adras state.
Indian Forestar, 92: 27-34.
Hydrobiologia 2 4 7 : 17-27, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 17
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Structural properties of two types o f mangrove stands on the

northwestern coast o f Sri Lanka

M . D. A m arasin g h e1 & S. B alasubram aniam 2

1National Aquatic Resources Agency, Crow Island, Colombo 15, Sri Lanka; 2Department o f Botany,
University o f Peradeniya, Peradeniya, Sri L anka

K ey words: m angrove ecosystem s, com m unity structure, floristic com position, complexity index,
estuarine m angals, fringing m angals

Abstract

M angrove stands in P uttalam lagoon and D utch bay, tw o interconnected lagoons situated on the
northw estern co ast o f Sri L anka have been classified broadly into tw o groups, i.e., estuarine and
island/m ainland-fringing stands. S tructural diversity o f six m angrove stands, representing the tw o types
was studied in term s o f floristic com position, density, basal area, m ean stand diam eter, tree height,
standing above-ground biom ass and leaf-area index. Rhizophora mucronata and Avicennia marina were
the dom inant species. Higher m ean stand diam eters for the m angrove stands in P uttalam lagoon indi­
cated greater m aturity than the estuarine m angrove stands in D utch bay. N evertheless, estuarine stands
in D utch bay were structurally m ore com plex (complexity indices 8.11-22.7) than the island/m ainland-
fringing m angrove stands (complexity indices 1.38-6.78). H igher num ber o f species present in the es­
tuarine m angrove stands is the m ajor elem ent th at contributes to the higher values for the complexity
indices for those stands. This appears to m ask the contribution o f stand-age to the com plexity o f a
m angrove stand. Therefore complexity indices alone m ay no t be used to explain adequately the struc­
tural diversity am ong m angrove stands.

Introduction such as tides, nutrients, hydro-period and stresses

M angrove ecosystem s are open system s, which
exchange m atter and energy with the adjacent
m arine and terrestrial ecosystem s. How ever, they
are found to differ in their ‘energy signature’ or the
sum o f all forces which dictates the types o f o r­
ganism s th at will survive and the speed o f eco­
logical processes (C intron e ta l., 1985). C intron
eta l. (1978), Lugo eta l. (1973) and C intron &
Schaefer-Novelli (1984) have show n th at the
m agnitude and periodicities o f forcing functions
such as cyclones, drought, salt accum ulation and
frost m ay largely determ ine the ‘energy signature’
o f a m angrove stand and therefore the floristic
and faunistic com position as well as the com m u­
nity structure.
In the present study, the structural character­
istics o f m ajor types o f m angrove stands in n o rth ­
w estern Sri L anka have been quantified with a
view to assessing w hether they can be used for
classification o f m angrove com m unities in the Sri
L ankan context.
18

DUTCH BAY

Ka l a V
INDIAN 1 SRI
LANKA
OCEAN Erumathivu

IV

100 200

Serakkuli/
Karathivu

Eul l upi di

Pallivasalthurei

PUTTALAM

LAGOON

LEGEND

Estuarine
mangals
Fr i n g i n g
mangals
Salt
marshes

Puttalam

Fig. 1. Distribution of mangroves in the study area and the location of the study sites. I - Serakkuli; II - Pullupidi; III -
Pallivasalthurai; IV - Erumathivu; V and VI - Kala Oya.

Study area
P uttalam lagoon and D utch bay, together with
Portugal bay (location - 8° 15' N, 79° 50' E;
Fig. 1) form the largest brackish w ater body in Sri
L anka. T otal w ater surface is approxim ately
46000 ha. M angroves cover nearly 1200 h a and
2100 h a o f inter-tidal land around P uttalam la­
goon and D utch bay respectively (K anakaratne
eta l., 1983).
Am ong the few river and creek systems asso­
ciated with these lagoons, K ala O ya is the long­
est with the largest river basin. A sem i-diurnal
tidal p attern occurs in the lagoon and the m axi­
m um tidal range is 79 cm (P erera & Siriw ardena,
1982). Surface salinities vary between 0.00 pp t (at
K ala O ya estuary) and 45.00 ppt.
A dry clim ate prevails (Fig. 2) and the average
annual rainfall (100-110 cm) is seasonal with two
distinct dry seasons, i.e., Ja n u a ry -M a rc h and
Ju ly -S ep tem b er. M axim um rainfall is in the p e­
riod from O ctober to D ecem ber by inter-m on-
soonal convectional rains.
J F M A M J j a s o n d j
Fig. 2. Climatic diagram of Puttalam area, Sri Lanka.
19
M aterials and methods
U sing auxiliary inform ation on physiognom y such
as species present, vigor of the stand, relative
location etc., m angals (the area covered with
m angrove plants) o f the study area have been
classified according to Lugo & Snedaker (1975)
and the m ethod m odified by C intron et al. (1985).
R epresenting the tw o types o f m angals in the area
i.e., estuarine and island/m ainland fringing (Am-
arasinghe & Perera, 1984), six study sites were
selected (Fig. 3).
Floristic composition
M angrove areas on the m ainland were visited
over-land while the islands and K ala O ya estuary
were accessed by boat. Specim ens o f the species
encountered were identified using descriptive lit­
erature and taxonom ic keys (W atson, 1928; D as-
sanayake & Fosberg, 1980; Blasco, 1984) as well
as by com paring them with the specim ens in the
N ational H erbarium at Peradeniya, Sri Lanka.
E num erated species were categorized into true
m angrove and m angrove associated species
(Saenger eta l., 1983).
Community structure
Belt transects o f 10 m in w idth and 50 m in length,
subdivided into 10 m x 10 m plots were laid per­
pendicular to the shoreline to collect structural
features of the m angals. Except at K ala O ya es­
tuary and Erum athivu island (Fig. 3), only one
transect in each site w as em ployed for the p u r­
pose.
Vertical distribution o f each stand w as deter­
m ined and profiles were prepared.
M ethods described in detail by C intron &
Schaefer-Novelli (1984) for the m easurem ent of
stand density, basal area, m ean stand diam eter,
tree height and leaf-area index were adopted.
Trees which are equal to or exceeding 2.5 cm in
diam eter at breast height (dbh) were m easured for
the purpose.
20

Pallivasalthurei
Serakkuli/ Kàrathivu

Am
8-

t Am

4-
Arrw

Erumathivu
Pul l upidi

8-

4-

Rm,

Ka l a Oya Kala Oya

8-

, Am ■Ami
4-

0-

0 20 40 0 20 40 60

DISTANCE FROM SHORE (m )

Fig. 3. Vegetation profiles of study sites. Am - Avicennia marina; Rm - Rhizophora mucronata; Lr - Lumnitzera racemosa; Ct -
Ceriops tagal.

Relationship between dbh and tree biomass for
Rhizophora m ucronata and A vicennia marina
As R . mucronata and A . marina are the dom inant
species in the study area these tw o species were
chosen to determ ine dbh and tree biom ass. Thus,
the relationships were obtained using the m ethod
described by C intron & Schaefer-Novelli (1984)
and D ay et al. (1987). G irth at breast height (gbh)
of the trees in the four perm anent transects
(10 X 50 m 2) in the m angals at Erum athivu island
and K ala O ya estuary (2 transects per site) were
m easured in order to determ ine the range o f gbh
for R. mucronata and A. marina in these tw o m an ­
gals. These two sites were particularly chosen for
biom ass studies as litterfall and gbh increm ent
m easurem ents (to estim ate the net prim ary p ro ­
ductivity, which will be discussed in a subsequent
paper) were obtained only from these tw o m an ­
gals which were presum ed to represent estuarine
and fringing m angals. Ten trees o f R . mucronata
(representing the whole range o f dbh) from each
study site (total o f 20 trees), nineteen trees of
A . marina from Erum athivu and ten trees o f the
sam e species from K ala O ya m angals were se­
lected for harvesting. These trees were cut as close
above-ground as possible. The canopy i.e.,
branches, twigs and leaves (flowers and fruits
were absent at the time o f harvesting), was
stripped and the weight o f each com ponent w as
obtained. Sam ples from each com ponent (about
500 g were collected and the fresh weight o f each
sam ple w as m easured to the nearest 0.1 g in the
field, using a triple beam balance. Tree trunks
were cut further into convenient lengths o f about
2 m, and the weight o f each length w as obtained
with a spring-balance. A few sections from a few
lengths were cut as sam ples and fresh weights
were obtained. F resh weight o f the prop roots of
harvested R . mucronata trees were obtained and
weighed sam ples were brought to the laboratory
to determ ine the dry weight to wet weight ration
(conversion factor). W et weights m easured in the
field were converted to dry weights using this con­
version factor. These estim ated dry weights for
each com ponent were sum m ed up to obtain dry
weight o f trees. Slopes o f the regression relation­
21
ships betw een dbh and total biom ass and the
biom ass o f other com ponents o f the plant for
each species obtained from the tw o sites were
com pared using S tudent’s t-test and found to have
no significant difference at 5% level between the
two sites, thus all the d ata for a species in the both
sites were pooled and the relationship between
dbh and biom ass was com puted for each species.
Allometric relationships betw een dbh and bio­
m ass were no t calculated for other species en­
countered in the transects, as they often occur in
low densities. A conservative estim ate o f biom ass
for such species w as obtained by assigning the
lowest biom ass to the corresponding dbh. The
lowest biom ass values were selected out o f the
biom ass values calculated for R . mucronata and
A. marina for the corresponding dbh.
Standing above-ground biomass
Standing above-ground biom ass o f the stands
were calculated by determ ining the frequency dis­
tribution o f the diam eters in each stand and es­
tim ating the biom ass o f each diam eter class sep­
arately, using the com puted regression equation
for each species. T he average biom ass o f a diam ­
eter class obtained thus, w as m ultiplied by the
num ber o f stem s in th at class and the results for
all the diam eter classes were sum m ed to obtain
the total biom ass o f the trees in the transect.
Table 1. List of mangrove species enumerated in Puttalam
lagoon and D utch bay mangals.
1. Acanthus ilicifolius L.
2. Aegiceras corniculatum (L.) Blanco
3. Avicennia marina (Forsk.) Vierh.
4. Bruguiera cylindrica (L.) Blume
5. Bruguiera gymnorhiza (L.) Lam
6. Ceriops tagal (Perrot.) C.B. Rob.
7. Cynometra ¡ripa Kostei.
8. Exoecaria agallocha L.
9. Lumnitzera racemosa Wiild.
10. Rhizophora apiculata Blume.
11. Rhizophora mucronata Lam.
12. Scyphiphora hydrophyllaceae Gaertn.
13. Sonneratia alba J. Smith
14. Xylocarpus granatum Koenig.
22

Table 2. Relative density (Rde), dominance (Rdo), frequency (Rfr) and importance value (Imv) for the major mangrove species
(Based on the trees with diam eter>2.5 cm). Roman numerals in the parentheses indicate the location in Fig. 3. A.m. - Avicennia
marina; R.m. - Rhizophora mucronata', L.r. - Lumnitzera racemosa', C.t. - Ceriops tagal.

Mangal type/location A.m. R.m. L.r. C.t. Other

Riverine mangals
(Dutch bay)
Kala Oya (V) Rde 65.70 25.90 1.40 2.10 4.90
Rdo 88.53 8.85 0.39 0.69 1.54
Rfr 33.33 25.00 8.33 16.67 16.67
Imv 187.56 59.75 10.12 19.46 23.11
Kala Oya (VI) Rde 13.50 52.30 28.80 4.50 0.90
Rdo 27.18 62.75 2.96 7.00 0.11
Rfr 36.36 27.27 9.09 18.19 9.09
Imv 77.04 142.32 40.85 29.69 10.10
Fringe mangals
(Puttalam lagoon)
Karathivu/Serakkuli (I) Rde 100.00 - - - -
Rdo 100.00 - - - -
Rfr 100.00 - - - -
Imv 300.00 - - - -
Pallivasalthurai (III) Rde 100.00 - - - -
Rdo 100.00 - - - -
Rfr 100.00 - - - -
Imv 300.00 - - - -
Pullupidi (II) Rde 89.20 6.50 4.30 - -
Rdo 99.42 0.24 0.34
Rfr 60.00 20.00 20.00 - -
Imv 248.62 26.74 24.64 - -
(Dutch bay)
Erumathivu island (IV) Rde 36.20 63.80 - - -
Rdo 36.20 63.80 - - -
Rfr 50.00 50.00 - - -
Imv 122.40 177.60 - - -

Table 3. Regression coefficients used to estimate the above­

Complexity index ground standing biomass. The equations are in the form,
loge7 = A + B logeX with Y = dry weight (g), X = dbh (cm) and
r 2 = coefficient o f determination.
Complexity index (CI, H oldridge et al., 1971) was
calculated as follows: Species Plant A B r2
component
C l = Num ber of species x stand density
X stand basal area x stand height 10 - 5
R. mucronata Leaves 5.408 1.241 0.597
Branches 4.262 2.103 0.781
Results Trunk 4.339 2.520 0.956
Prop roots 5.359 2.024 0.863
Floristic composition Total biomass 6.247 2.064 0.945
A. marina Leaves 3.984 1.825 0.816
O ut o f the 14 true m angrove species encountered Branches 4.074 2.299 0.807
Trunk 5.148 2.044 0.906
in the study area, R . mucronata and A. marina
Total biomass 5.551 2.153 0.922
were the dom inant. True m angroves, m angrove
 23

( A)

91.0 12 - 0 *
y= 6-18 + 2-11 x
r= 0-973 ( P < 0 0 5 )

co.
o
X
Ul

V, 47-0 10 - 0 -
CtJ
E
o
CQ

«3
O

3-0 80- —r~

2-0 6 '62 11-24 0-6 1-5 2-4
dbh cm log dbh
e
Fig. 4. (A) Curvilinear and (B) log-log relationships between total biomass and dbh for R. mucronata in D utch bay.

(A) (B)
87-0' 11.5-
V= 5-55 -t- 2-15 x
°
r = 0-960 ( P < 0 05

oo~ U
U ll
o CD
X E
O) o
in
cn
C/) CD
£ 44-0- 9-5-
o
bo
O)
o

1-a 7-5
2-0 6-8 11-6 0-6 3-4
dbh (cm) lo g dbh

Fig. 5. (A) Curvilinear and (B) log-log relationships between total biomass and dbh for A. marina in D utch bay.
24

3 (D "tí
tí > S tí I
títí °) o3 ¡ S
X m r-~
oo un
o en ÍN ,—i
t í ON t-~ r-
<! bß x w íN
£
o
o<
X 3 ■a
X
T3 X)
j
r - OO oo
00 en r -
un
ON
u° .a
"2 íN oo
ÍN
ÍN 1-H NO

tí; un íN en en ON
tí tí r- Ö t í en

ON o o r- <N
t í r- tí o oo
oo NO as NO oo
tí íN ÍN r"H

tí "Ö _ O n r - íN OO
X g S tí. O n r- CN tí
+2 tí ü
C/3 X c3
O n en en ON t í
t í <N en
ex
tí
íN t í NO NO NO un
On r- t í NO en en
ON íN íN OO oo en
t í CN CN

f - t '­ NO ÍN un
O en en O ON -—<
3
tí O o r-H r-H un oo
CQ
Q
-o
§
tío T3 ."tí
H
o § o o
o o
o un
O O tí
O un en
o
o
SP 55 "O w NO t-~
e n íN
OO t í un CN
(N

T3 un o O un r - O
tí S íN o o íN NO <N
CX un ÍN NO en r - ÍN

o O t— o r - O
un en NO o NO en
NO íN NO en OO ÍN

bß
tí un O o un r - O
r- un o CN NO OO
'5b o un íN r- ON
tí en ÍN

bß
b ÍN
o o
o NO
en OO IN­
en en NO
O
c O o On oo t í
ÍN
tí
A t í tí ÍN tí

O
r - un *—
h en íN

tí
bß
tí
tí T3
¿a .s tí
■SI S •tí
C/5
tí 2h ■M
3 c
3O 1 > > ^ co ;ss
tí tí tí bOß rtí| tí>
t í tí bß

en ex . ?n tí
tí —1T3tí ^ X 5 3
tí S-o o 'e x 2
^ *«, <% tí tí tí
1 3 ”tí Db ß J5 33 JÜ 2 tí
^“-CD 13 tí

Ö oi-H
^ *¡
•s s a £ tí tí
S E fe ex Q

associated and salt m arsh species enum erated in
Puttalam lagoon and D utch bay are listed in T a ­
ble 1. Vegetation profiles o f the study sites are
depicted in Fig. 3. In P uttalam lagoon m angals,
A . marina had a higher im portance value (C intron
& Schaefer-Novelli, 1984) than in the D utch bay
m angals and the reverse w as true for R . mucro­
nata (Table 2).
Community structure
S tructural features are presented in Table 4.
Stand densities o f the m angals in D utch bay
are considerably higher than those in P uttalam
lagoon and a reverse trend occurs with regard
to stand basal area and standing above-ground
biom ass. Com plexity indices for estuarine m an­
gals are greater th an those for the fringing m an­
gals.
R elationship between dbh and biom ass for
R . mucronata and A . marina are graphically p re­
sented in Figs 4 & 5. Regression coefficients used
to estim ate above-ground biom ass are presented
in Table 3.
Discussion
Floristic composition and zonation
As in m any other m angals in the dry coastal re­
gions in Sri Lanka, R . mucronata and A. marina
are the m ajor constituent species of the m angals
of P uttalam lagoon and D utch bay. A . marina is
the m ost dom inant species in the m angals in P u t­
talam lagoon with relative dom inance values
ranging from 60 to 100 and im portance values
ranging from 248.62 to 300. Ability o f A . marina
to tolerate a wide range o f environm ental condi­
tions (W ells, 1982) such as salinity and air tem ­
perature, m ay favour its distribution, no t only in
frequently inundated areas but also in infrequently
flooded upper inter-tidal areas. R. mucronata is
m ore dom inant th an A . marina in the estuarine
habitats o f K ala Oya.
M o st o f the fringing habitats, especially those
25
in P uttalam lagoon are occupied by a single spe­
cies, frequently by A. marina. A ccording to C in­
tro n et al. (1985), the m ost co n stan t conditions of
salinity and nutrients as well as the exposure to
expectedly higher kinetic energy regimes which
are relatively unfavourable to support a large
num ber o f species m ay restrict the num ber of
species occurring in fringing habitats. In the es­
tuarine m angals, nutrients and salinity m ay oscil­
late widely and the kinetic energy o f w ater flow
m ay anticipated to be lower than th at of fringing
m angals. M oreover, during periods o f flood, con­
siderable tidal and w ater flow energies m ay be
dissipated in the riverine (estuarine) sites (C in­
tron e ta l., 1985). A lthough such periods are rel­
atively short, they m ay favour m angrove estab­
lishm ent m ore than the co n stan t tidal m otion in
the fringing m angrove habitats, especially in a
low-tidal environm ent. Thus as a result o f the
oscillating conditions and m ore favourable nutri­
ent balances, estuarine m angals m ay be able to
support m ore species than fringing m angals.
G eneral zonation o f species within a m angal
reflects the gradient in flooding and soil charac­
teristics o f the m angrove species. Kenneally
(1982) listed frequency o f tidal flushing, soil type,
soil salinity, drainage plant and anim al interac­
tions as the m ajor factors th at appear to deter­
m ine the survival o f individual species and thus
zonation w ithin a m angal. In the study area where
a dry clim ate and a small tidal am plitude prevail,
the frequency o f tidal flushing in the upslope may
be significantly low and as the terrain is flat, soils
are poorly drained thus creating a soil salinity
gradient inland. M ore salinity tolerant m angrove
species such as A. marina and Excoecaria
agallocha as well as the salt m arsh species occupy
such landw ard portions while know n less tolerant
R . mucronata occupies the w ater-front zones of
the inter-tidal land. Seasonal precipitation and
high evapo-transpiration rates m ay have largely
contributed to the developm ent o f salt m arshes
and salt flats which characteristically occur
around P uttalam lagoon where freshw ater influx
is apparently low.
26
Structural parameters
Tree densities o f estuarine m angals are higher
th an those o f the fringe m angals, owing to the
higher num ber o f smaller trees (diam eter less than
2.5 cm ) in the estuarine habitats. Fringing m an ­
gals on the m ainland (around P uttalam lagoon)
are the least densed while those on the islands
record interm ediate values. These trends indicate
th at the stands in P uttalam lagoon are m ore m a­
tured than those in D utch bay.
Stand basal area, m ean stand diam eter, stand
height and standing above-ground biom ass val­
ues calculated for these m angrove stands also
m anifest a similar trend. T he m ainland-fringing
m angrove stands in P uttalam lagoon seem to be
the m ost m atured, followed by the island fringing
m angals in the sam e area. The least m atured
stands are the estuarine m angals at K ala Oya
estuary. In general, the presence o f fewer num ber
o f small trees and higher num ber o f large trees in
the m angals in P uttalâm lagoon imply th at these
m angals are m ore m atured than those in D utch
bay. This m ay imply th at the coastal areas in
P uttalam lagoon are géom orphologie ally m ore
stable than in D utch bay.
E stuarine m angals m anifest a greater structural
complexity (complexity indices 8.11-22.7) than
the m ainland and island fringing m angals (com ­
plexity indices 1.38-6.78). This m ay be attributed
to the prevailing favourable flushing patterns as
well as the nutrient and salinity regimes which
contribute to the ‘energy signature’ (C intron et al.,
1985) in the estuary.
How ever, the low complexity index-values for
the m angrove stands in P uttalam lagoon, which
have resulted m ainly due to the low num ber of
species present, appear to m ask their relative eco­
logical im portance. Therefore complexity index
m ay n o t adequately indicate the structural com ­
plexity and functional capacity o f a m angal com ­
prised o f a few species.
Acknowledgements
This study w as supported by the N ational A quatic
R esources Agency and the assistance given by the
staff o f the Regional R esearch C entre at K alpitiya
is gratefully acknowledged. O ur thanks are also
extended to D r. Y ara Shaefer-N ovelli at Univer-
sidade D e Sao Paulo, Brazil for her critical com ­
m ents on the m anuscript.
References
Amarasinghe, M. D. & W. K. T. Perera, 1984. A preliminary
survey of the peripheral vegetation communities around
Puttalam lagoon and Dutch bay, Sri Lanka. Proc. 41st
Ann. Sess. SLAAS (abstract only).
Blasco, F., 1984. Taxonomic considerations of the mangrove
species. In S. C. Snedaker & J. C. Snedaker (eds), The
mangrove ecosystems: research methods. U N ESC O , Paris;
France: 81-90.
Cintron, G., A. E. Lugo, D. J. Pool & G. Morris, 1978. M an­
groves of arid environments in Puerto Rico and adjacent
islands. Biotrop. 10: 110-121.
Cintron, G. & S. Y. Schaefer-Novelli, 1984. M ethods for
studying mangrove structure. In S. C. Snedaker & J. C.
Snedaker (eds), The mangrove ecosystem: research meth­
ods. U N ESC O , Paris; France: 91-113.
Cintron, G., A. E. Lugo & R. Martinez, 1985. Structural and
functional properties of mangrove forests. In W. G. D ’Arcy
& A. M. D. Correa (eds), The botany and natural history
of Panama: La botanica e historia natural de Panam a IV
series, M onographs in systematic botany, V 10, Missouri
Botanical G ardens, St. Louis, Mo; USA: 53-56.
Dassanayake, M. D. & F. R. Fosberg, 1980. A revised hand­
book to the flora of Ceylon, Vol. I. Amarind Publishers,
New Delhi; India: 508 pp.
Day, J. W. Jr., W. H. Corner, F. Ley-lou, R. H. Day & A. M.
Navarro, 1987. The productivity and composition of man­
grove forests, Laguna de Terminos, Mexico. Aquat. Bot.
27: 267-284.
Holdridge, L. R., W. C. Grenko, W. H. Hatheway, T. Liang
& J. Tosi, 1971. Forest environment in tropical life zones;
a pilot study. Peragamon Press, New York: 747 pp.
K anakaratne, M. D., W. K. T. Perera & B. U. S. Fernando,
1983. An attempt at determining the mangrove coverage in
Puttalam lagoon, Dutch bay and Portugal bay, Sri Lanka,
using remote sensing techniques. Proc. 4th Asian Conf.
Remote Sensing, Colombo; Sri Lanka; P-15-1-9.
Kenneally, K. F., 1982. Mangroves of Western Australia. In
B. F. Clough (ed.), Mangrove ecosystems in Australia:
structure, function and management, Australian Institute of
Marine Sciences; Australia: 95-110.
Lugo, A. E., G. Evink, M. M. Brinson, A. Broce & S. C.
Snedaker, 1973. Diurnal rates of photosynthesis, respira­
tion and transpiration in mangrove forests in south Florida.
In: The role of mangrove ecosystems in the maintenance of
environmental quality and high productivity of desirable
fisheries. Centre for Aquatic Sciences.

Lugo, A. E. & S. C. Snedaker, 1975. Properties of a mangrove
forest in southern Florida. In G. E. Walsh, S. C. Snedaker
& H. J. Teas (eds), Proc. International Symposium on Bi­
ology and Management of Mangroves, IFAS, University of
Florida; USA: 170-212.
Perera, W. K. T. & P. P. G. S. N. Siriwardena, 1982. Topog­
raphy of Puttalam lagoon. J. Inland Fish. (Sri Lanka) 1:
97-104.
Saenger, P., E. J. Flegerl & J. D. S. Davie, 1983. Global sta­
27
tus of mangroves. IU C N Commission on Ecology, Pap.
No. 3.
Watson, J. D., 1928. Mangrove forests of the Malay penin­
sula. Malaya Forest Records 6: 1-275.
Wells, A. G., 1982. Mangrove vegetation of northern A ustra­
lia. In B. F. Clough (ed.), Mangrove ecosystems in A ustra­
lia: Structure, Function and Management. Australian In­
stitute of Marine Sciences: 57-78.
g ft
Hydrobiologia 2 4 7 : 29-36, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 29
© 1992 Kluwer Academic Publishers. Printed in Belgium.

The current status o f mangroves along the Kenyan coast: a case study of
M ida Creek mangroves based on remote sensing

P. O. G ang & J. L. Agatsiva

Department o f Resource Surveys and Rem ote Sensing, Ministry o f Planning and National Development,
P.O. B ox 47146, Nairobi, Kenya

K ey words: m angroves, K enya, rem ote sensing, species com position and distribution

Abstract

M angroves form a unique ecosystem o f limited extent covering an area o f about 53000 h a along the
K enyan co ast which need protection from overexploitation. B ackground inform ation on the areal ex­
ten t and status o f these m angroves is limited and m akes their protection and m anagem ent difficult.
A m odel study has been carried out on the M ida Creek m angroves based on a double sam pling ap ­
proach starting with S PO T m ultispectral satellite imagery followed by ground checks to provide infor­
m ation on species com position, density and distribution o f m angroves. The utilization o f m angroves and
uses around and within the m angroves were determ ined.
D uring this study, seven m angrove species were identified in M ida Creek, namely: Avicennia marina,
Bruguiera gymnorrhiza, Ceriops tagal, Rhizophora mucronata, Sonneratia alba, Xylocarpus granatum and
Lumnitzera racemosa. This report briefly outlines the present utilisation o f m angrove species. It is rec­
om m ended th at a m ultidisciplinary m anagem ent plan be developed in order to conserve and m anage the
m angroves o f K enya on a sustainable yield basis.

Introduction ucts. These include poles for building, vegetable

M angroves are a group o f highly evolved halo-
phytes occupying the intertidal zone in estuaries,
lagoons and coastal m ud flats in tropical and sub­
tropical zones. While they have proven ability to
function in poorly oxygenated water-logged, sa­
line habitats, a seasonal flushing with freshw ater
is necessary for their continued survival. There
are some which prefer a daily tidal w ash while
others find their optim um conditions in areas sub­
ject to only seasonal high tides (U N D P /
U N E S C O , 1984). This determ ines the distribu­
tion o f m angrove species within a given
com m unity. M angroves have been exploited by
coastal people for their num erous natural p ro d ­
tannins, pulp, fuelw ood and m edicine (U N E P ,
1988).
The K enyan coastline has a total o f about
5 2 9 8 0 h a o f m angrove forest (D oute et al., 1981).
T he bulk o f these forests occur in the intertidal
areas where subm arine ground w ater discharge
(S G D ) or seepage occur rather than in estuaries
(R uw a & Polk, 1986).
Background and problem
T he lack o f appraisal o f the status o f K enyan
m angrove forest has been a limitation on the for­
m ulation o f an adequate m anagem ent and con­
servation policy. Quite often, the exploitation of
30
m angroves has occurred in the p ast unnoticed or
unquantified, as has been previously reported by
D oute eta l. (1981) who based their study on the
use of L an d sat M ultispectral Scanner (M S S ) im­
agery. The study is now under review (Agatsiva
et al., 1990) to up-date it at the sam e time giving
m ore details which were never achieved in the
1981 study.
The lack o f up-to-date and accurate d ata poses
planning problem s on conservation and m anage­
m ent of such m angrove species. It has been noted
th at m angroves and their related ecosystem s have
in the p ast undergone drastic changes w ithout
m uch attention being paid to their ultim ate fate.
W hen exploiting a renew able resource, the con­
cept o f sustainable developm ent should be borne
in m ind. Conflicting land-uses such as fishing,
cutting o f m angroves for different uses, salt m in­
ing and various m arine activities are likely to ex­
haust the m angrove resources.
The M ida Creek m angroves (G ang, 1988) have
been considered in an effort to design an appro­
priate m ethodology for having the K enyan m an­
groves rapidly and cheaply assessed. The recom ­
m endations m ade m ay enable them to be properly
conserved, m anaged and where exploited, to be
appropriately harvested on a sustainable yield
basis. M ida Creek w as preferred because the area
has a variety o f m angrove types as well as differ­
ent regeneration stages and som e dense old
stands. There are also several land use practices
in and around the creek w hich m ay also influence
the m angrove ecosystem along the coast. The
bro ad recom m endations m ade on this area are
therefore likely to represent a w ider spectrum of
inform ation concerning all m angroves along the
K enyan coast.
The objectives o f the study are therefore
(i) To docum ent the extent and status o f the
m angroves of M ida Creek.
(ii) T o assess the suitability o f S PO T m ultispec­
tral satellite imagery in providing baseline
d ata for the m onitoring and m anagem ent of
m angrove areas.
(iii) To prepare a checklist o f plants growing in
close proximity to the m angroves both within
and w ithout the m angrove area in M ida
Creek.
(iv) To give possible suggestions and recom m en­
dations th at m ay be applied for sustainable
utilisation o f the m angroves.
The study area
Location o f the study area
M ida Creek is located in Kilifi district on the
north coast o f K enya and is approxim ately 25 km
south o f M alindi tow n. The area lies within ap­
proxim ate bounds o f longitude 39° 58' E and lat­
itude 3° 20' S (Fig. 1).
Climate
The rainfall pattern is bim odal. The long rains fall
in April to June, with a peak in M ay. The short
rains fall from O ctober to D ecem ber. T he area
receives an annual rainfall o f about 900-1100 mm
(D istrict D evelopm ent P lant 1989/83).
Soil types
Soils o f the m angrove sw am ps are poorly drained,
deep and excessively saline, olive to greenish grey
loam to clay and often w ith sulfidic m aterial (D is­
trict Environm ental A ssessm ent R eport, 1985).
L and use
Agriculture is one o f the land use types in the
area. Tree crops occupy m ore than one third o f
the area under cultivation. The m ost im portant of
which are cashew nuts, m angoes and coconuts.
C oconuts are very m uch valued due to the role
they play in providing daily requirem ents such as
cooking fat, woodfuel, building m aterials, local
beer and hair oil. C ashew nuts are sold to the
processing plant based in Kilifi. M angoes are sold
to the village, hotel industry and traders in M o­
m basa.
Fishing is another im portant activity in the
study area. M ost o f the catch is harvested by
individual fisherm en using their own boats and
gear following traditional m ethods. They are often
seen in their small boats in the creek which is
surrounded by m angroves.

M I D A
Fig. 1. Study area map.
Methodology
T he m ethodology used w as based on a double
sam pling approach which include the following:
(i) U se o f rem ote sensing to provide initial strat­
ification o f the area into m ajor cover types
and ground-truthing o f the satellite image.
(ii) U se o f ground checks for species identifica­
tion and categorisation into various cover
types.
Rem ote sensing
Rem ote Sensing is broadly defined as a process
o f gathering inform ation about objects from a dis­
tance based on the electrom agnetic energy re­
flected or em itted by th at object. Satellite rem ote
sensing d ata can provide natural resource m an­
31
C R E E K
All w e a t h e r r o a d
L oo se s u rf a c e
M a in t r a c k s
N P /G a m e r e s e r v e b o u n d a r y
H uts
4 0 * 0^ E o f G re e n w ic h
agers and planners with periodic inform ation on
the extent, location and status o f the resources.
Such inform ation types are used in the day to day
decision m aking on the im plem entation o f m an ­
agem ent policies such as conservation and re­
source allocation.
A whole range o f rem ote sensing techniques
could have been used to m ap the area bu t w eather
conditions did not allow all these techniques to be
used and their results com pared for the area of
interest. The region is under cloud cover for m uch
of the year. How ever, S P O T 1 m ultispectral im ­
agery was obtained for the area from Toulouse -
F rance. S PO T m ultispectral imagery K J 146 357
(scale 1:50000) w hich has a spatial resolution of
20 by 20 m w as the ideal imagery for m apping
cover types because different m angrove species
and zonation could be easily identified from the
32
image. This w as done by use o f a tran sp aren t
overlay put over the image and placed on a light
table where all the distinguishable features were
delineated. The differentiation w as based on co­
lour intensity, chrom a and hue in addition to other
image characteristics and interpretation elements
including pattern and texture am ong others as
described by Lillesand & Keifer (1979).
A topographic m ap at the scale o f 1:50000
(sheets N o 193/1 and 193/3 from survey of
K enya) w as used to associate various m anifesta­
tions in the image with contents o f the m ap re­
ferring to land use and also for orientation and
navigation during field surveys.
Ground measurements
In each stratum obtained, representative stands
w ere sam pled to obtain inform ation regarding the
m angroves on:
(i) M angrove Species
(ii) Height class
(iii) C row n cover
(iv) D ensity o f species
(v) D iam eter at breast height (dbh).
R epresentative stands were selected within each
stratum , using 50 m by 50 m quadrats. T he total
count m ethod w as used as described in M uchoki
(1988). All m angrove trees 0.7 m high and above
were m easured and quantitative attributes such
as height, dbh and crow n cover were recorded.
Visual observation was also done to determ ine
the species being exploited. This included actual
counting o f stum ps o f the already cut m angroves
Table 1. Mangrove species found in the study area.
Mangrove species
Avicennia marina (Forsk.) Vierh
Bruguiera gymnorrhiza (L.) Lam
Ceriops tagal (perr.) C. B. Robinson
Lumnitzera racemosa Wild.
Rhizophora mucronata Lam.
Sonneratia alba Sm.
Xylocarpus granatum Koen.
Xylocarpus sp.
and counting the dhow s transporting m angrove
poles to a collection yard at U yom bo. Finally an
attem pt w as m ade to prepare a vegetation check­
list o f plants growing in close proxim ity to m an­
groves. T he m ethods used were those used by
K enya R angeland Ecological M onitoring U nit
(K R E M U ) in its R angeland M onitoring attem pts
(K uchar, 1979) and those used by botanists else­
where in com piling lists o f plants by occurrence
in different habitats.
Results and discussion
T he above double sam pling approach yielded
good results on the m angrove species distribution
and uses. Seven m ajor m angrove species were
identified (Table 1).
The S P O T image delineated h ad the following
cover classes (Fig. 2).
A vicennia m arina
This is found in the landw ard fringe o f the Creek.
The underground is mainly covered by pneum ato-
phores. This covers 160 h a (2.9% ).
A vicennia/C eriops/R hizophora
This is characterised by species which have dis-
continous canopy form ations and generally in
poor state o f health, covering 145 h a (2.6% ).
R hizophora/C eriops
D om inated m ainly by Rhizophora mucronata and
Sonneratia alba at the edge o f the creek and
Ceriops tagal w ith few Bruguiera gymnorrhiza and
Xylocarpus spp. The undercanopy consists of
Standard name in Swahili Family
Mchu or Mtu Avicenniacea
Muia Rhizophoraceae
M kanda Rhizophoraceae
Kikanda or M kanda-M witu Combretaceae
Mkoko Rhizophoraceae
Mlilana or Mpia Sonneratiacea
Mkomafi Meliaceae
Meliaceae
 33

L E GE ND
Avicennia

n Avicennia / R hizophora / C eriops

E S R hizophora / C eriops

1-4=1 A vicenia / R hizophora / C e rio p s / B ruguiera /

S o n n e ra tia / X ylocarpus
C erio p s /B ru g u ie ra / Rhizophora!

I c I Mixed tr e e A griculture

1 * ^ 1 O p e n S w a m p s - Muddy

Kitilepwc

Uyombo

Fig. 2. Classification of M ida Creek mangroves and surrounding cover types based on SPOT imagery.

Rhizophora and Ceriops seedlings growing vigor­ C eriops/B ruguiera/R hizophora

ously. Cutting in this zone is selective and rather This is dom inated by Ceriops tagal with very high
intensive. This category covers 67.5 h a (1.2% ). regeneration o f both Rhizophora and Ceriops
on the outer fringes. F rom the land seaw ards,
A vicennia/R hizophora/C eriops/B ruguiera/ this zone has few Avicennia marina while
S onneratia/X ylocarpus Rhizophora mucronata m ixed with Sonneratia alba
This category is characterised by very m ature dom inate the edge o f the creek. This covers 890 ha
stands o f m ajority o f the species. T here are very- (15.9% ).
tall Bruguiera. Cutting o f Ceriops and Xylocarpus
is m ost intensive probably due to their straight M ixed tree agriculture
poles. Avicennia is very scarce and least utilised. This occurs in the hinterland surrounding the
This unit covers 237.5 h a (4.2% ). creek, Sudi and Kilelepwa islands found in M ida
34

creek. Tree crops found in this zone are m angoes, Table 3. Plants growing in close Proximity to mangroves in
M ida Creek.
coconuts and cashew nuts. There is also natural
vegetation growing together with tree crops. This Species Family
includes Thespesia danis Olive and Azadrachta
indica A. Juss. This class covers 3805 h a (68% ). Aeva lanata Amaranthaceae
Anisotes parviflora Acanthaceae
Aristogeitonia monophylla Euphobiaecae
Open swamps ¡muddy areas
Asparagus africanus Lam. Liliaceae
These are sw am py areas th at no longer receive Azima tetracantha Lam. Salvadoraceae
flooding except during high tides, im mediately Cremaspora trifloa Rubiceae
within the vicinity o f these areas, small halophytic Cyperus tuberosus Rottb. Cyperaceae
plants such as Salicornia sp. and Sueda sp. are Canaralia carthartica Thouars Papilionaceae
Dactyloctineum gem inatum Hack Gramineae
dom inant. The category covers 292.5 h a (5.2% ).
Derris spp. Papilionaceae
Satellite imagery (S P O T image) proved to be Digitaria milanjiana (Randle) Stapt Gramineae
very useful in identifying m angrove areas and Eleaodendron schweinfurthianum Loes Celastraceae
other land use types in close proximity. Different Enteropogon macrostachysus (A. Rich) Gramineae
m angrove zonations are easily identifiable using Benth
Euclea schimperi (A. D. C.) Dandy Ebenaceae
tonal differences in the image. It is no t possible to
Flacourtia indica (burn, f) Merr. Flacourtiaceae
identify m angroves to species level in the image Fimbristylis triflora (L.) K. Schum Cyperaceae
but tonal differences interpreted were confirmed Fuirena ciliaris (L.) Roxb. Cyperaceae
to represent different m angrove types during field Grewia vaughanii exell. Tilliaceae
surveys. It was n o t possible also to detect the Hyphaene compressa H. Wendl. Palmae
Ipomoea pes caprae (L.) R. Br. Convolvulaceae
cutting o f the m angroves from the satellite im ag­
Mormodica trifoliata Hook F. Cucurbitaceae
ery because the cutting is selectively done and it Panicum infestum Anderss. Gramineae
is n o t concentrated at one spot. A ctual counting Panicum piniplium Gramineae
in the sam pling plots indicated a cutting intensity Phoenix reclinata Jacq. Palmae
o f thirty poles per 50 X 50 m plot. Table 2 shows Phragnites communis Gramineae
Premna spp. Verbanceae
som e o f the m angroves species identified in the
Psilotrichum sericeum (Roxb.) Dal 2 Amaranthaceae
M ida Creek using the total count approach. The Saliconia sp. Chenopodiaceae
rest o f the plants in the surrounding areas are Salvadora persica L Salvadoraceae
indicated in Table 3. The cutting intensity de­ Sesuvium portulacastrum (L.) L Aizoaceae
creases away from the landw ard fringes probably Sideroxylon inerme L Sapotaceae
due to accessibility and convenience. The forest
departm ent controls the cutting by issuing licenses
and ensures th at poles are cut according to the
quantity specified, and suspends further cutting in
Table 2. Total number of species obtained through the total som e areas until regeneration is restored. There
count method from five sampling plots randomly chosen in
are no afforestation activities in the creek but
M ida Creek.
rather the m angroves regenerate naturally, grow ­
Mangrove Plots ing into exploitable poles after 3 -6 years. N o
quantitative m easurem ents were carried out on
1 2 3 4 5 Total
regeneration but it w as observed during field sur­
Ceriops tagal 46 107 7 47 2 209 veys th at species like Ceriops tagal and Rhizophora
Xylocarpus granatum 7 - 12 1 - 20 mucronata have several seedlings actively grow ­
Bruguiera gymnorrhiza 2 1 - 2 1 6 ing. This m ay suggest th at the forest departm ent
Avicennia marina 1 - - - 82 83 need to regulate the m ode o f harvesting in order
Rhizophora mucronata - 32 27 23 1 83
to ensure sustainable exploitation o f the m an ­
Sonneratia alba - - - - 5 5
groves in this creek.

The licensing policy o f the forest departm ent
should be reviewed and strengthened to ensure
th at licenses issued to licensees are consistent in
order to prevent illegal cutting o f m angrove spe­
cies and types which are not originally included
in the license. This requirem ent should control
the current practices in M ida creek where m an­
groves are cut to the edge o f the creek. The cut­
ting o f Rhizophora sp. at the edge allows siltation
since the stilt roots which hold sand leads to the
death o f the tree. This activity m ay interfere with
the future o f the creek since siltation m ay in the
long run turn the creek into the beach.
M ida Creek m angroves are not m uch exploited
for firewood. A ran dom survey indicates th at peo­
ple living in the periphery o f the creek prefer to
collect their fuelwood from the neighbouring Ar-
abuko Sokoke forest. Firew ood is m ainly ob­
tained from the dead branches o f com m on tree
species including: Brachystegia speciformis,
Cynometra webberi, Afzelia quanzensis, Croton
pseudopulchellus and maytenus undata. N everthe­
less, M ida Creek m angroves (Table 1) are heavily
exploited for building and fencing poles. A visit to
a collection yard at U yom bo, confirms the export
of m angrove poles to the m ainland. They are piled
in the different classes depending on the diam e­
ter at bu tt end. These range from 3.8 cm to 14 cm
in diam eter.
Com plem entary studies on w oodfuel/tim ber
dem and and supply m ay be essential to ensure
sustainable utilisation o f the w oody biom ass.
These studies should be able to com pute further,
the attributes such as the ones shown in the T a ­
ble 4 to precisely calculate the am ount o f biom ass
available to the entire creek.
Table 4. Summary of mangrove data in plot No. 2 near D abaso in M ida Creek.
Name of plant Total above
ground biomass
kg/ha
Rhizophora mucronata 11832.0
Bruguiera gymnorrhiza 1029.0
Ceriops tagal 24178.0
Xylocarpus granatum 514.0
35
Conclusion
The superior resolution o f S PO T d ata m ade it
possible to identify different cover types in the
study area. W ith the com plem ent o f ground sur­
vey, species identification w as possible in the area.
F rom these results, S P O T d ata prom ise a new
dim ension in the use o f Rem ote Sensing partic­
ularly in this region w here rapid acquisition o f
d ata is necessary for developm ent planning.
Hopefully, with further research into the use of
satellite imagery for coastal resource planning and
m anagem ent, it will be possible to answ er some
o f the questions arising from such studies and
refine w hat am ounts to a very useful and time
saving m ethod o f m onitoring m angrove areas.
Q uestions such as how m uch area o f kenya is
currently under m angroves, their status and ex­
ploitation patterns will be answered.
Suggestions and recommendations
Use o f further satellite imagery in monitoring m an­
grove areas
H aving proved the ability o f satellite and recog­
nizing it as a powerful aid to fast field d ata col­
lection, it would be worthwhile to apply this tool
in docum enting the status o f the m angroves in the
entire K enyan coast. A good sam pling strategy
incorporating aerial photography on a sam ple
basis m ay rapidly enhance the results for m an­
grove m anagem ent.
Further studies into the parameters controlling the
distribution o f mangroves in a given habitat
F urther studies should be carried out in order to
Utilizable % of trees Density Frequency
woody biomass per species
kg/ha
9727.0 32.0 46.0 23
846.0 3.0 4.0 2
19877.0 64.0 94.0 47
423.0 1.0 2.0 1
36
gain in-depth understanding of the param eters
controlling the distribution o f each m angrove spe­
cies within a given habitat. Effects o f such p a­
ram eters as land use activities and erosion in the
hinterland, salinity, tem perature and depth of
substrate type are im portant for designing advi­
sory and protection schemes for the proper inter­
pretation o f the results from m onitoring p ro ­
gramm es.
Use o f computers and digital mapping techniques
The capability to digitize S PO T d ata should be
enhanced as well as the use of com puters and
digital m apping techniques in the region in order
to derive d ata from the imagery. Certainly for any
tim e series this would ease the problem o f iden­
tifying changes to m angrove cover.
Increased awareness o f mangroves values
The efforts to conserve the m angroves w ould ben­
efit from a greater aw areness in G overnm ental,
N o n G overnm ental organisations and am ongst
the general public o f the im portance o f m an­
groves. To this end, an educational cam paign
should be initiated w hich em phasizes the values
o f the m angroves and their links to other elements
in the ecosystem , principally seagrass beds, coral
reefs and fish populations.
Silvicultural m easures could potentially en­
h ance the productivity o f the m angrove forests,
but at present there is very limited inform ation
available on w hich to base such m easures. F u r­
ther research should therefore be carried out to
docum ent the correct silvicultural practices th at
m ay be useful for the K enyan coast.
A N ational inventory o f the m angroves of
K enya should be m ade. The Status o f the entire
m angrove ecosystem should be highlighted in this
inventory report. It is after this th at a national
m angrove m anagem ent scheme can be fram ed
and prepared by a multidisciplinary group.
Acknowledgements
G ratitude is due to M r. D. K. A ndere, D ep art­
m ent o f R esource Surveys and Rem ote Sensing,
D .R .S .R .S ., N airobi, for his support and encour­
agement. I also extend my gratitude to African
Wildlife foundation, regional office in N airobi,
through the fund for conservation endeavour for
providing p art o f the funding for the survey ac­
tivities in M ida Creek. I am also grateful to m em ­
bers o f the E ast A frican H erbarium for their help
with plant identification and finally to technical
staff, D .R .S .R .S . for criticizing the text.
References
Agatsiva, J. L., J. P. Delsol, N. Ochanda, M. Shimba & J. M.
Terresa. (unpublished results). A remote sensing approach
to mapping and monitoring forest cover in Kenya.
Doute, R., N. O chanda & H. Epp, 1981. Forest cover map­
ping using remote sensing techniques Kenya Rangeland
Ecological Monitoring Unit, P.O. Box 47146, Nairobi,
technical report No. 30, 72 pp.
Gang, P. O., 1988. Application o f Remote Sensing in
Landuse/Landcover assessment of Coastal Environments
(Dissertation submitted to the Kenya Polytechnic, P.O. Box
52428, Nairobi, Kenya).
Gang, P. O., 1988. The distribution of mangroves forest of
Kenya. A case study in M ida Creek, Sabaki and Ngomeni
areas, Kenya Rangeland Ecological Monitoring Unit, Box
47146, Nairobi (unpublished results).
Kuchar, P., 1979. A methodology for sampling in order to
monitor conditions, trend and plant biomass. Kenya
Rangeland Ecological Monitoring Unit, Box 47146,
Nairobi.
Lillesand, T. M. & R. W. Kiefer, 1979. Remote sensing and
image interpretation. John Wiley and Sons Inc., New York:
528-590.
Kenya Government, Ministry of Planning and National D e­
velopment 1989, Kilifi district Development plan. 1989/93,
pp. 1-3.
Muchoki, C. H. K., 1988. A preliminary survey of efficiency
of three methods used in studying rangeland vegetation:
The quadrat method, PCQ and line transect methods,
Kenya Rangeland Ecological Monitoring Unit, Box 47146,
Nairobi. Technical Report No. 132. pp. 3-10.
Kenya Government, National Environmental Secretariat,
1985. District environmental assessment report of Kwale
district, 6 p. (6 pp.)
Ruwa, R. K. & P. Polk, 1986. Additional information on the
mangrove distribution in Kenya: some observation and re­
marks - Kenya J. Sei. Ser. B7: 41-45.
U N E P, 1988. Ambio, A Journal o f the human environment.
Royal Swedish Academy of Sciences/Pergamon Press, vol.
xvii No. 3 1988: 214-216.
U N D P /U N E S C O , 1984, Regional Mangrove Project RA/79/
002 in Asia and the Pacific.
Hydrobiologia 2 4 7 : 37-47, 1992.
V. Jaccarini & E. Martens (eds). The Ecology o f Mangrove and Related Ecosystems. 37
© 1992 Kluwer Academic Publishers. Printed in Belgium.

N et primary productivity o f two mangrove forest stands on the

northwestern coast of Sri Lanka

M . D. A m arasin g he1 & S. B alasubram aniam 2

1National Aquatic Resources Agency, Crow Island, Colombo 15, Sri Lanka; 2Department o f Botany,
University o f Peradeniya, Peradeniya, Sri L anka

K ey words: m angrove ecosystem s, estuarine m angals, fringing m angals, litterfall, prim ary productivity

Abstract

Productivity studies were carried out from Septem ber, 1985 to August, 1987 in tw o m angrove stands,
i.e. estuarine and island fringing, in D utch bay, a lagoon situated on the northw estern coast o f Sri Lanka.
N et above-ground prim ary productivity w as m easured by m onitoring litterfall and above-ground biom ass
increm ent. The average annual rate o f litterfall in the estuarine and island-fringing m angrove stands are
588.14 g m -2 (approxim ately 6 t h a _1) and 407.33 g m -2 (approxim ately 4 t h a _1) respectively. The
average annual rates o f above ground w oody growth are 614.74 g m ' 2 (approxim ately 6 t h a - 1) in the
estuarine stands and 286.8 g m -2 (approxim ately 3 t h a _1) in the island-fringing m angrove stands.
H ence estuarine m angrove stands record a higher annual rate o f above-ground net prim ary production
(N P P ; 1207.88 g m -2 or approxim ately 1 2 t h a _ 1 ) than the fringing m angrove stands (694.22 g m - 2 ;
approxim ately 7 t h a - 2). The annual rate o f N P P in the w ater front zones o f the stands (1300.47 g m “ 2
in the estuarine stands and 874.56 g m _ 2 in the fringing stands) are greater than those in the back-
m angrove zones (115.28 g m “ 2 in the estuarine stands and 513.88 g m -2 in the island-fringing stands).
T hese variations m ay be attributed to the differences in tidal flushing and influence o f freshw ater in the
two localities.

Introduction (below -ground N P P ). Shedding o f plant parts,

Tropical forest ecosystem s especially m angrove
forests, are am ong the m ost productive in the
w orld in term s o f gross prim ary productivity
(G P P ), (R odin, 1975) and it has been shown th at
considerable variation in G P P exists am ong the
various m angrove areas (Lugo & Snedaker, 1974;
Snedaker & Brown, 1982). An im portant portion
o f this G P P is the net prim ary productivity (N P P )
which represents the portion o f the G P P that
rem ains after carbon loss through respiration is
deducted and accum ulated as biom ass, primarily
in w ood, leaves (above-ground N P P ) and roots
i.e. leaves, floral parts, fruits, twigs, branches and
bark, know n as litterfall, is an im portant source
o f organic detritus, which support economically
im portant detrital m arine food webs (O dum &
H eald, 1975). Indicating this ecological role of
m angals, a positive relationship betw een the ex­
tent o f m angals and m arine produce (e.g. praw ns)
has been observed in a num ber o f tropical areas
(M artosubroto & N aam in, 1977; T urner, 1977).
As tidal flushing, salinity and nutrient regimes
as well as substrate characteristics appear to in­
fluence prim ary productivity (C arter et al., 1973),
its m agnitude m ay vary from one m angrove area
38
to another. As such, present study w as u n d er­
taken to u n d erstan d the productivity o f estuarine
and fringing m angrove ecosystem s associated
with a dry coastal zone lagoon in N o rth W estern
Sri L an k a where a low tidal am plitude prevails.
Study area
D utch bay is a p art o f the largest brackish w ater
body in Sri L an k a (L ocation - 8° 15' N ,
79° 50' E). D escription o f the study area, m ap of
D utch bay and the location o f study sites are
given in A m arasinghe & B alasubram aniam (this
volume).
M angroves cover nearly 2100 h a o f the inter­
tidal areas o f D utch bay and the m angals at K ala
O ya estuary are the m ost extensive (nearly
1800 ha; K an ak aratn e et al., 1983). A dry clim ate
prevails and the average annual rainfall (1000-
1100 m m ) is seasonal with two distinct dry sea­
sons i.e., J a n u a ry -M a rc h and July-S eptem ber,
coinciding with the north east and south w est
m onsoons. M axim um rainfall is in the period from
O ctober to D ecem ber by inter-m onsoonal con-
vectional rains (A m arasinghe & B alasubram a­
niam , this volume). Surface w ater salinity at the
sam pling sites in D utch bay varies from 0.00%o to
39.00%o (Fig. 1).
S u r f a c e - w a t e r salin ity a t E rum athivu
S u r f a c e - w a t e r salinity a t Kala O ya
e s tu a r y
86
Fig. 1. Monthly variation in salinity from December 1985 -
January 1987 at the study sites.
M aterials and methods
Study sites
Study sites (Fig. 2) were selected so as to repre­
sent the m ajor types o f m angals i.e., estuarine
m angals (K ala O ya estuary), fringing m angals
(Erum athivu island). A t both sites, Avicennia
marina and Rhizophora mucronata were the dom ­
inant m angrove species and the latter is found
occupying the w ater-front zone and the flood-
ways within the m angals while the form er species
occupies a land-w ard position (back-m angrove
areas). K ala O ya m angals are floristically richer
and structurally m ore com plex than the fringing
m angals at Erum athivu island (A m arasinghe &
B alasubram anium , this volume). Ceriops tagal,
Lumnitzera racemosa and Aegiceras corniculatum
occur sporadically in the back-m angrove areas in
the estuary.
Litterfall
U sing auxiliary inform ation, i.e., presence o f dif­
ferent m angrove species and associations, it was
possible to divide the m angals into sub-popula­
tions or zones (strata), each o f which w as as­
sum ed hom ogeneous internally. In both the study
sites, tw o m ajor zones or strata were distinguish­
able. N um ber o f trap s to collect litter, for each
stratum or zone w as determ ined on the approx­
im ate proportion o f ground area o f each zone and
a total o f 38 trap s, 20 at K ala O ya (7 in the
w ater-front R . mucronata zone and 13 in the
mixed back-m angrove zone) and 18 at E ru­
m athivu (8 in the w ater-front R . mucronata zone
and 10 in the A . marina back-m angrove zone)
were used for the purpose.
Traps o f 1 m 2 in area, m ade out of w ooden
fram es and nylon m esh were placed random ly in
each zone, along the 10 m wide belt transects
which have been laid perpendicular to the
shoreline/riverbank across the m ajor zones o f the
m angals upto 50 m inland, at the tw o study sites
87 (2 such transects per site). T raps were installed at
a level higher than the highest tide level, using
stakes/poles and nylon ropes.
 39

. t ot al litterfall

CD
"O
CN

E
O) eaves

misc. + tw lg s

85 86 87
Fig. 2. Variation in total litterfall and components in the Rhizophora mucronata zone of the riverine/estuarine mangals at Kala Oya
estuary.

Litter (trapped) w as collected fortnightly, for a
period o f two years from Septem ber 1985 to A u­
gust 1987. Collected litter w as brought into the
laboratory in plastic bags, sorted into species and
litter o f each species into plant com ponents, i.e.,
leaves, stipules, floral parts, fruits/seedlings,
bran ch w ood and m iscellaneous item s and they
were dried at 70 ° C to co n stan t weight.
Above-ground woody biomass production
biom ass of the trees in the plots were estim ated
using the following regression equations (A m a­
rasinghe & B alasubram anium , this volume).
F o r R . mucronata
log,, (total biom ass) = 6.247 + 2.64 log,, (dbh)
F o r A. marina
log,, (total biom ass) = 5.551 + 2.153 log,, (dbh)

D iam eter at b reast height (dbh) o f all the trees Results

with a dbh equal to or exceeding 2.5 cm, in the
10 m X 10 m subplots of the four perm anent belt
transects, were m easured in A ugust 1985 and re­
m easured in O ctober 1987 at Erum athivu and
M arch 1988 at K ala Oya. The initial and final
Litterfall
The average annual rates o f com ponent and total
litterfall in the tw o m angals are presented in Ta-
40

Table 1. Averages litterfall rates ± standard deviation (g m 2 yr ') by component in the major zones of the mangals at Kala
Oya.

Leaves Stipules Repro, parts Branches Debris Total

R. mucronata zone
1985-1986 535.63 57.57 14.30 22.49 14.24 644.23
± 18.24 ±2.24 ± 2.11 ±2.25 ±0.85 ± 18.49
1986-1987 505.81 52.11 13.25 18.42 13.75 603.34
± 17.14 ±2.36 ± 1.33 ± 2.02 ±0.87 ± 18.23
Whole period 19.22 54.84 13.78 20.46 14.00 623.79
+ 17.36 + 2.26 + 1.73 + 2.10 + 0.84 + 18.03

Mixed mangrove zone

1985-1986 428.92 11.78 33.03 27.80 37.62 539.15
± 12.32 ±0.58 ±3.29 ±3.19 ± 1.73 ± 11.83
1986-1987 445.38 13.37 51.76 17.28 38.01 565.80
± 11.01 ±0.69 ±3.33 ± 1.51 ± 1.15 ± 10.03
Whole period 437.15 12.58 42.40 22.54 37.82 552.48
+ 11.45 + 0.63 + 3.31 + 2.48 + 1.44 + 10.78

Table 2. Average litterfall rates + standard deviation (g m 2 yr ') by component in the major zones of the mangals at Erumathivu.

Leaves Stipules Repro, parts Branches Debris Total

R. mucronata zone
1985-1986 408.54 46.95 11.16 29.43 16.80 513.38
± 19.06 ±2.11 ± 1.68 ±4.99 ±1.26 ± 22.58
1986-1987 273.68 44.61 20.59 13.30 16.35 368.53
±9.25 ± 1.78 ±1.93 ± 1.45 ± 1.08 + 8.45
Whole period 341.11 45.78 16.13 21.37 16.58 440.96
± 15.86 ± 1.91 ±1.81 ±3.67 ± 1.15 ± 17.77

A. marina zone
1985-1986 304.77 - 34.10 20.13 34.80 363.80
± 16.68 - ±4.20 ±2.38 ±1.43 ± 15.87
1986-1987 300.00 - 10.17 14.48 28.92 353.57
± 17.29 - + 1.17 ± 1.49 ± 1.29 ± 15.90
Whole period 302.39 - 22.14 17.31 31.86 373.69
± 16.61 _ ±3.19 ±1.95 ± 1.35 ±15.63

bles 1 and 2 and the percentage contribution of
com ponent litterfall to the total litterfall in the
m ajor zones o f the tw o study sites are show n in
Table 3. The average annual litterfall rates in the
tw o zones at K ala Oya estuary are higher than
those in the tw o m ajor zones at Erum athivu m an­
gals. In both the m angals, the average litterfall
rates in the w aterfront (R . mucronata) zones
were greater th an those in the back-m angrove
zones.
L eaf litter com prised the largest portion (7 7 -
84% ) o f the total annual litterfall. N on-leaf plants
(m angrove) litter accounts only for 2 -1 0 % o f the
total litterfall in these m angrove ecosystem s.
Four-weekly totals for the litterfall indicate a
m arked seasonality in leaf, reproductive struc­
tures and total litterfalls in the tw o m angals. In all
the zones, the period o f high litterfall was during
M arch -A u g u st (with m inor variations), which
coincides with the south w est m onsoon, generally
 41

Table 3. Percentage contribution o f component litterfall to the total litterfall in the major zones of the mangals at Kala Oya es­
tuary and Erumathivu island (R.m. - R. mucronata', A.m. - A. marina)',

Component Percentage contribution

Kala Oya Erumathivu

R.m. zone A.m. zone R.m. zone A .m .zone

Leaves 83.5 79.1 77.4 80.0

Stipules 8.8 2.3 10.4 -
Reproductive parts 2.2 7.7 3.7 5.9
Branches 3.3 4.1 4.8 4.6
Debris (Miscellaneous) 2.2 6.8 3.2 8.5

the dry season for this area. T he period o f low
litterfall coincides with the north east m onsoon
and the inter-m onsoon m onths are the wet
m onths o f the year (Figs 2, 3, 4 and 5).
V ariations in total and com ponent litterfall o f
R . mucronata and A . marina in the tw o sites are
presented in Figs 6 and 7. In both the sites, leaf
fall follows a similar pattern with m inor differ­
ences in the m onth o f m axim um leaf fall. The
m agnitude o f litterfall o f the species is higher at
K ala Oya than in the m angals at Erum athivu.
Annual above-ground woody biomass production
Average basal area increm ent, average woody
grow th per stem per year and total annual growth

total l i t t erf al l

eaves
OJ 1

e o r o a u c tiv
Parts
misc. + twigs

S O N D J F M A M J J A S O N D J F M A M J J A
85 86 87

Fig. 3. Variation in total litterfall and components in the mixed mangrove (back-mangrove) zone of the riverine/estuarine mangals
at Kala Oya estuary.
42

t o t a l litterfall

2
■>
CO
■O
CM
'E
Ul

£
n
’0 1
eaves

>
Q

repro.
Iparts I

m is c .+ tw ig s

85 86 87
Fig. 4. Variation in total litterfall and components in the in the Rhizophora mucronata zone of the fringing mangals at Erumathivu
island.

o f m angrove trees m easured in the m angals at
K ala O ya estuary and Erum athivu island are
given in Table 4.
T he average rate o f basal area increase and net
w oody above-ground production per stem for the
tw o m ajor m angrove species, i.e., R . mucronata
and A . marina were greater at K ala O ya than at
Erum athivu. In both sites, average w oody growth
o f the trees in the w ater front zone was higher
than th at in the back-m angrove zone. A t K ala

Table 4. Average basal area increment (A) in cm2 per stem per year, average woody growth (B) in kg per stem per year and total
annual growth (C )in g m -2 y ~ 1 of mangrove trees measured at Kala Oya estuary (I) from July 1985 to M arch 1988 and Erumathivu
island (II) from July 1985 to October 1987.

Species A B C

I II I II I II

Water-front zone
R. mucronata 1.68 0.89 1.33 0.82 676.68 433.60

Back-mangrove zone
R. mucronata 1.66 - 1.24 - 210.35 -
A. marina 1.56 1.10 0.86 0.29 253.42 140.19
Other species 3.32 - 1.52 - 99.02 -
 43

to ta l litterfall

2-

■>
CD
cm"0
'E
O)
-C
O)
'0

>
L_
Û

r e p r o . p<

misc. + twigs _
1 i 1----- 1-------- 1----- 1 j 1------- 1-------- 1------1------1-------1-------1---------1----- 1----- 1— “ i------- 1-------- 1----- 1-------- 1— i-----r
S O N D J F M A M J J A S O N D I F M A M J J A
85 86 87
Fig. 5. Variation in total litterfall and components in Avicennia marina zone of the fringing mangals at Erumathivu island.

Oya, A . marina had a smaller basal area increase
th an R . mucronata. A t Erum athivu, the trend was
reversed. In both sites, increase in w oody biom ­
ass o f R . mucronata w as greater than th at for
A . marina trees. Although R . mucronata in the
mixed m angrove zone at K ala O ya h ad a higher
basal area increase and average w oody growth
th an A . marina in the sam e zone, its contribution
to total w oody grow th (210.35 g y r “ *) w as less
than th at by A . marina.
The total w oody growth o f the trees in the
w ater-front zones at K ala O ya and Erum athivu
were higher th an those in the back-m angrove
zones o f the tw o m angals.
Table 5. The N P P values for the tw o zones at
K ala O ya were greater th an those values calcu­
lated for the tw o zones at Erum athivu. A t each
site, the N P P for the w ater-front zone w as higher
than th at for the back m angrove zone. Except in
the back-m angrove zone at Erum athivu, where
w oody growth accounted only for about 27%
of the N P P , in all other zones litterfall com ­
prised nearly 50% o f the total N P P o f those m an­
gals.
Discussion and conclusions
Litterfall

Total above-ground NPP
T otal above-ground N P P values calculated for
the m ajor zones o f the estuarine (K ala Oya) and
fringing m angals (Erum athivu) are presented in
Am ong the different types o f m angals, the high­
est litterfall values have been recorded from
riverine/estuarine m angals (average 1100—
1200 g m “ 2 y r “ % and fringing m angals are re­
ported to have a value o f 906 + 89 g m “ 2 y r “ 1
44

Leaves

b Reproductive Structures
-g 0-05 b. Stipules Ü 0-051
CM
'E

? 0-05 c. Reproductive Stru ct ures 0-05-, c Branch Wood

S N J M M J S N J M M J

d. Branch Wood 85 86 87
0-05-
Fig. 7. Variations in litterfall by components of Avicennia
marina at Kala Oya (--------- ) and Erumathivu (............ ).

S N J M M J S N J M
85 86 87
Fig. 6. Variations in litterfall by components of Rhizophora
mucronata at Kala Oya (--------- ) and Erumathivu ( - -.........).
(Twilley, 1982 cf. D ay et al., 1987). The average
litterfall value obtained (average for both the
zones), for the estuarine m angals at K ala Oya
(about 588 g n C 2 y r _1) is m uch lower than the
average range o f values reported for the estuarine
m angals and this value is even lower than th at
reported for the fringe m angals. Litterfall rate in
the fringing m angals at Erum athivu, i.e.,
407 g m “ 2 yr “ 1 is m uch below the average range
reported for fringe m angals (C intron et al., 1985).
N evertheless the total litterfall rate estim ated for
K ala O ya estuarine m angals (annual rainfall 100-
110 cm ) is com parable with th at o f the m angals
at Siar river estuary (annual rainfall 426 cm ) in
Saraw ak, E ast M alaysia (O thm an, 1988). The
M J
notable similarity o f these tw o m angrove areas is
th at they are dom inated by R . mucronata. F u r­
therm ore, the annual litterfall rates for A . marina
(at K ala Oya, about 356 g m -2 y r “ 1 and at E ru­
m athivu, nearly 342 g m ~ 2 yr - x) are m ore com ­
parable with those for the d w a r f s , marina stands
in N ew Z ealand (W oodroffe, 1982) and A . nitida
stands in South F lorida (S nedaker & Brown,
1982), w hich are less than 1 m tail, in contrast to
3 -4 m tail stands in this study area. This could
partly be due to the differences in freshw ater input
and tidal am plitude. Pool et al. (1975) have sug­
gested th at the rate o f litterfall in m angroves is
related to increasing turnover o f freshw ater. Be­
sides bringing nutrients required for continuous
productivity o f m angroves, within the m angal it­
self, w ater is im portant in bringing nutrients and
oxygen to the ro o t system and rem oving C 0 2,
toxic w astes and organic debris as well as in con­
tinual m aintenance o f the soil-salt balance
 45

Table 5. Estimated above-ground net primary production of mangrove species in the major zones of the estuarine mangals at Kala
Oya and the fringing mangals at Erumathivu island (percentage contribution of each species is given in parentheses).
A - R. mucronata zone; B - Mixed mangrove zone; C - A. marina zone.

Component Location and zone

K ala Oya estuary Erumathivu island

A B A C

Litterfall (g m 2 yr ’)
R. mucronata 609.79 (46.9) 127.72(11.5) 424.38 (48.5) -
A. marina - 355.96 (31.9) - 341.83 (66.5)
Other species - 30.99 (2.8) - -
Miscellaneous 14.00(1.1) 37.82 (3.3) 16.58(1.9) 31.86 (6.2)
Total 623.79(48.0) 552.49 (49.5) 440.96 (50.4) 373.69 (72.7)
Average for mangal 588.14 407.33

Woody growth (g m 2 yr ')

R. mucronata 676.68 (52.0) 210.35 (18.9) 433.60 (49.6) -
A. marina - 253.42 (22.7) - 140.19(27.3)
Other species - 99.02 (8.9) - -
Total 676.68 (52.0) 562.79 (50.5) 433.60 (49.6) 140.19(27.3)
Average for mangal 619.74 286.89

Annual above-ground net primary production (N PP) (g m -2 )

R. mucronata 1286.47 (98.9) 388.07 (30.3) 857.98 (98.1) -
A. marina - 609.38 (54.6) - 482.02 (93.8)
Other species - 130.01 (11.7) - -
Total 1300.47 1115.28 874.56 513.88
Average for mangal 1207.88 694.22

N PP : Litterfall 2.08 2.02 1.98 1.34

(Snedaker, 1978). K ala O ya river system is ex­
tensively utilized in the upstream areas for irriga-
tional purposes. Eight reservoirs have been built
in the upstream by dam m ing the river and the
w ater stored in K ala W ewa, the largest reservoir
in the system, is diverted to a num ber o f smaller
tan k s in the N o rth C entral Province to irrigate
paddy land. H ence the volum e o f freshw ater dis­
charged by the river m ay inevitably be low. Al­
though the nutrients in this have no t been qu an ­
tified, due to low freshw ater inputs and associated
p o o r sedim ent tran sp o rtatio n into the estuary,
it is plausible th at the nutrient influx is also
low. Fringing m angals at Erum athivu, situated
farthest from the rivers m ay receive even lower
input o f freshw ater and nutrients. Although the
elevation is sufficiently low, low tidal am plitude
ensue infrequent and insufficient tidal flushing
m ay also contribute to the decreased productiv­
ity o f these m angals. This also m ay explain the
lower productivity in the back-m angrove zones
than the w ater-front zones which subject to daily
flushing.
Leaves form the largest portion o f the total
litterfall and hence constitutes a significant input
o f organic m atter in both the m angals (Table 3).
A lthough woodfall in these m angals appears to be
unim portant, it m ay no t be the reality as dead
trunk fall has no t been taken into account in the
present study.
A general trend for peak litterfall to occur d u r­
ing the rainy season has been reported by a num ­
ber o f m angrove studies (H eald, 1971; Pool et al.,
1975; Leach & Bürgin, 1985; D ay et al., 1987).
46
H ow ever a reverse trend w as true for the litterfall
in the m angals in D utch bay. Lugo & Snedaker
(1975) also have reported higher leaf fall rates
during periods o f extrem e drought. This m ay be
attributed to increased energetic cost o f m aintain­
ing photo synthetic tissues in high-salinity condi­
tions during droughts, thus causing shedding of
leaves.
D uke et al. (1984) have shown th at the abscis­
sion o f a pair o f stipules in association with the
emergence o f a pair o f leaves, to be an accurate
indicator o f leaf emergence. L eaf production as
m easured by stipule fall o f R. mucronata (at K ala
O ya and E rum athivu) w as found to be continu­
ous throughout the study period, which suggests
th at the environm ental conditions are favourable
for leaf emergence all year round and the stress
does n o t appear to limit the leaf production o f this
species. T he period o f slightly increased stipule
fall (rates) during M arch -A u g u st at K ala O ya
m angals coincides with higher leaf fall period.
Therefore it could be assum ed th at the photosyn­
thetic rates m ay rem ain co n stan t as shown by
Pool et al. (1975) and Lugo & Snedaker (1975).
Thus this study suggests th at the physiology of
the trees m ay have an effect on the seasonality in
litter production.
Above-ground woody growth o f mangroves
Differences in above-ground w oody grow th of
m angroves in the m ajor zones at K ala O ya and
E rum athivu m ay be attributed to the lower salin­
ities (w ater and soil) and greater input of fresh­
w ater and hence nutrients in the estuarine m an­
gals at K ala Oya. As D ay et al. (1987) have
shown, the low er rates o f total w oody grow th in
the back-m angrove zones than in the w ater-front
zones, especially at Erum athivu may reflect the
effect o f stresses such as salinity and anaerobic
conditions resulted from infrequent tidal flushing
and also the effect o f hydrogen sulphide on the
biom ass production o f the trees in this zone. Thus
m ore o f the available energy from prim ary p ro ­
duction m ay be used to deal with stress and less
will be available to build the structure.
Above-ground NPP
R eported values o f total above-ground N P P in
m angrove areas are very few and they range from
380 g m “ 2 yr~ 1 for a scrub m angal in F lorida
(Teas, 1979) to 5475 g m -2 y r “ 1 for a mixed for­
est in F lorida (C arter et al., 1973). The N P P val­
ues com puted for K ala O ya estuarine m angals
(1 2 0 8 g m _ 2 y r _1) and E rum athivu fringing m an ­
gals (694 g m “ 2 yr “ 1) fall within this range. D ay
eta!. (1987) reported values o f above-ground
N P P o f 2458 g m -2 y r “ 1 and 1 6 0 6 g m ~ 2 y r -1
for riverine (estuarine) and fringing m angals re­
spectively in M exico. T he assum ption th a t total
above-ground N P P for m angroves is three times
the total litterfall (Teas, 1979) is no t supported by
the results o f the present study, in w hich N P P
w as approxim ately twice the am ount o f litterfall.
D ay et al. (1987) also have reported sim ilar re­
sults for m angals in M exico.
Acknowledgements
This study w as supported by the N ational A quatic
R esources Agency (N A R A ) o f Sri Lanka. Valu­
able assistance by D r. U . S. A m arasinghe, Senior
R esearch Officer, N A R A in statistical analyses is
gratefully acknow ledged. U ntiring assistance in
the field by the staff at N A R A R egional R esearch
C entre at Kalpitiya, Sri L an k a is appreciated.
References
Amarasinghe, M. D. & S. Balasubramaniam, 1992. Struc­
tural properties of two types of mangrove stands on the
northwestern coast o f Sri Lanka. In: V. Jaccarini & E.
M artens (eds), The Ecology of Mangrove and Related Eco­
systems. Developments in Hydrobiology 80. Kluwer Aca­
demic Publishers, Dordrecht: 17-27. Reprinted from Hy-
drobiologia 247.
Carter, M. R., L. A. Burns, K. R. Carvinder, P. L. Dugger,
D. B. Fore, D. B. Hicks, H. L. Revells, T. W. Schmidt &
R. Farley, 1973. Ecosystem analysis of the big Cypress
swamp and estuaries. U S Environmental Protection Au­
thority.
Cintron, G., A. E. Lugo & R. Martinez, 1985. Structural and
functional properties of mangrove forests. In W. G. D ’Arcy

& M. D. Correa (eds), The botany and natural history of
Panama: La botanica e historia natural de Panam a IV se­
ries, Monographs in systematic botany, V 10, Missouri Bo­
tanical Gardens, St. Louis, Mo; USA: 53-56.
Day, J. W. Jr., W. H. Conner, F. Ley-lou, R. H. Day & N a­
varro, 1987. The productivity and composition of mangrove
forests, Laguna de Terminos, Mexico. Aquat. Bot. 27: 267-
284.
Duke, N. C., J. S. Bunt & W. T. Williams, 1984. Observa­
tions on the floral and vegetative phenologies of North-
Eastern Australia mangroves. Aust. J. Bot. 32: 87-99.
Heald, E. J., 1971. The production of detritus in a South
Florida estuary. Uni. Miami Sea G rant Bulletin, No. 6.
Kanakaratne, M. D., W. K. T. Perera & B. U. S. Fernando,
1983. An attempt at determining mangrove coverage in
Puttalam lagoon, D utch bay and Portugal bay, Sri Lanka,
through remote sensing techniques. Proc. 4th Asian Con­
ference on Remote Sensing, Colombo: 1-9.
Leach, G. J. & S. Bürgin, 1985. Litter production and sea­
sonality of mangroves in Papua New Guinea. Aquat. Bot.
23: 215-224.
Lugo, A. E. & S. C. Snedaker, 1974. The ecology of man­
groves. Annu. Rev. Ecol. Syst. 5: 39-64.
Lugo, A. E. & S. C. Snedaker, 1975. Properties of mangrove
forests in Southern Florida. In G. E. Walsh, S. C. Snedaker
& H. J. Teas (eds), Proc. Int. Symp. Biol. Management of
Mangroves, IFAS, Univ. Florida: 170-212.
Martosubroto, P. & N. Naamin, 1977. Relationship between
tidal forests (mangroves) and commercial shrimp produc­
tion in Indonesia. Marine Research in Indonesia, No. 18:
81-86.
47
Odum, W. E. & E. J. Heald, 1975. The detritus based food
webs of an estuarine mangrove community. Estuar. Res., 1:
265-286.
Othman, S., 1988. The structure of vegetation and rate of litter
production in a mangrove forest at Siar Beach, Lundu,
Sarawak, East Malaysia. A paper presented at the Sym­
posium on Mangrove Management; Its Ecological and
Economic Considerations, Bogor, Indonesia.
Pool, D. J., A. E. Lugo & S. C. Snedaker, 1975. Litter pro­
duction in mangrove forests in Southern Florida and Pu­
erto Rico. In G. E. Walsh, S. C. Snedaker & H. J. Teas
(eds), Proc. Int. Symp. Biol, and Management of M an­
groves, IFAS, Univ. Florida: 213-237.
Rodin, L., N. I. Bazilavich & N. N. Rozov, 1975. Productiv­
ity of the worlds main ecosystems, National Academy of
Science, Washington D.C.
Snedaker, S. C., 1978. Mangroves: their value and perpetu­
ation. Nat. Res. 14: 6-13.
Snedaker, S. C. & M. S. Brown, 1982. Primary productivity
of mangroves. In C. C. Black & A. Mitsui (eds), CRC
H andbook of Biosolar Resources, 1: 57-70.
Teas, H. J., 1979. Silviculture with saline water. In A. Hol-
laender, J. C. Aller, E. Esptein, A. E. Pietro & O. R. Za-
borsky (eds), The Biosaline Concept. Plenum Press, New
York: 117-161.
Turner, R. E., 1977. Intertidal vegetation and commercial
yields of penaeid shrimp. Trans, am. Fish. Soc. 106: 411-
416.
Woodroffe, C. D., 1982. Litter production and decomposition
in the New Zealand mangrove Avicennia marina var. resin­
ifera. New Zealand J. mar. Freshwat. Res. 16: 179-188.
r<4¡i 1»

* z
Hydrobiologia 2 4 7 : 49-57, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 49
© 1992 Kluwer Academic Publishers. Printed in Belgium.

The effect of coastline changes on mangrove community structure,

Portuguese Island, M ozambique

J. C. H a tto n & A. L. C outo

P.O. B o x 257, Faculty o f Biology, University o f Eduardo Mondlane, Maputo, Mozambique

K ey words: coastline configuration, m angrove, selective death, M ozam bique

Abstract

C hanges in coastline configuration and sand m ovem ents on Portuguese Island, a small island lying close
to In h aca Island 35 km east o f M aputo, M ozam bique have resulted in large areas o f the Island being
eroded aw ay and consequent restricted tidal recharge o f the m angrove. M easurem ents m ade during 1989
show th at flooding o f the m angrove occurred on only three occasions. Percentage m ortality o f the
d om inant species, Ceriops tagal (Perr.) C. B. R obinson., is high. How ever, at the tim e o f the study, the
co-dom inant, Bruguiera gymnorrhiza (L.) Lam ., show ed no signs o f stress. H alophytic conditions are
indicated by an invasive spread o f salt-tolerant herbs. All species of crustaceans and m olluscs have
disappeared from the m angrove.
A lthough m ortality o f C. tagal is high, the regeneration status o f this species is also high, com para­
ble to th at in a regularly recharged m angrove on In h aca Island. It is suggested th at C. tagal propagules
were shed under conditions when seedling establishm ent w as favorable i.e. flooding. These seedlings m ay
represent the last cohort which m ay also suffer stress and death unless tidal recharge is re-established.

Introduction few published d a ta are available. G eneral descrip­

M ozam bique has the third largest coastline in
Africa, extending m ore than 2700 km from the
T an zan ian b order in the north (10° 27' S) to the
South A frican b order in the south (26° 57' S).
Only Som alia and South Africa have a greater
extent o f coastline. Extensive m angrove com m u­
nities are located at several sites along the coast
in sheltered bays and estuaries. It is estim ated
th at m angrove com m unities occupy some
850 km 2 in M ozam bique, second in im portance
only to T an zan ia (960 km 2) along the E ast Afri­
can coast (B ennessalah, 1988).
A lthough m angrove com m unities are an im ­
p o rtan t feature o f the M ozam bican coastline very
tions o f flora and associated fauna have been
given for selected sites along the coast (C ott, 1929;
Jacques, 1975; K alk, 1959; Sim, 1910). These
w orks are o f a descriptive n ature or provide an
inventory o f plant and anim al species. In contrast
to the m angroves along m ost o f M ozam bique’s
coastline, the m angroves o f In h aca and P o rtu ­
guese Islands are relatively well-docum ented.
M acN ae & K alk (1962 and 1969) provide de­
tailed descriptions o f the taxonom y, distribution,
zonation patterns and faunal associations o f the
Inhaca Island m angroves. M ogg (1963) carried
out a prelim inary investigation into the relation­
ship betw een salinity and distribution o f m an­
grove tree species on the Island. H ow ever, no
50
investigations related to ecophysiology, m angrove
dynam ics or utilization patterns o f the M o zam ­
bican m angroves have been carried out.
T he present study represents one o f several
research projects related to m angrove ecology on
In h aca Island. A prelim inary survey carried out
by the authors in 1988 noted the presence o f m or­
ibund and dead trees in the m angrove on nearby
Portuguese Island w hich appeared to be related
to im peded tidal recharge due to closure o f the
canal feeding into the m angrove. A com parison of
aerial photographs (1958 to 1985) confirmed th at
m arked changes in coastline configuration had
taken place probably leading to this sequence of
events. T he present paper represents the results
o f a m ore detailed investigation into the changes
in coastline configuration around Portuguese Is­
land and the effects o f these changes on the m an­
grove com m unity.
Study area
In h aca Island and nearby Portuguese Island are
located 35 km east o f the city o f M aputo (see
m ap, Fig. 1). Tides at In h aca are sem i-diurnal
with an average am plitude o f 3.3 m for springs
and 1.5 m for neaps. Average annual tem perature
is 22.9 °C with m axim um m ean m onthly tem per­
ature in F ebruary (26.2 °C ) and m ean m onthly
m inim um in July (19.2 °C ). Average annual rain ­
fall is 874 m m w ith a distinct wet season during
O ctober to April and a dry season during M ay to
Septem ber. E vaporation exceeds rainfall for all
m onths except January and February.
F our broad vegetation types m ay be distin­
guished: in tact forest, agricultural landscape with
scattered trees, sw am ps and m angroves (Fig. 2).
The m ain study w as carried out in the m angrove
on Portuguese Island with a com parative study
(‘no n -stressed ’ m angrove) being m ade in the
m angrove at P o n ta R asa, In h aca Island.
The P o n ta R asa m angrove is a typical creek
m angrove approxim ately 1 km in length running
in a north-south direction. The banks o f the canal
leading into the m angrove are lined w ith a dense
stand o f Rhizophora mucronata Lam . The fiddler
crab Uca urvillei (M . Edw .) is com m on along the
b anks o f the canal. The central core o f the m an­
grove is covered by a dense, mixed stand of
R. mucronata and Bruguiera gymnorrhiza (L.)
Lam . In this zone the soils have a relatively high
organic m atter content (5 -6 .5 % ) and sesarm id
crabs are ab u ndant (Sesarma meinerti de M an.
and S. guttata M. Edw.). T he landw ard edge of
the m angrove com prises a less dense stand of
smaller growing Ceriops tagal and Avicennia m a­
rina (F orsk.) Vierh. trees w ith occasional scat­
tered individuals o f Lumnitzera racemosa Willd.
on the outerm ost fringes. T he soils o f this zone
are drier and sandy with relatively low levels of
organic m atter (0 .1 -1 % ). F iddler crabs {Uca
inversa (H offm an), U. annulipes (M . Edw .) and
U. gaimardi (M . Edw .) are com m on faunal asso­
ciates in this zone (M acN ae & K alk, 1962).
The m angrove on Portuguese Island w as also,
originally, a creek m angrove. This m angrove is
m uch smaller than the P o n ta R asa m angrove and
zonation patterns are less distinct. A dense stand
of R . mucronata lines the original course o f the
canal. T he outerm ost landw ard zone o f the m an­
grove com prises a dense stand o f C. tagal and
B. gymnorrhiza. Only a few scattered individuals
of A. marina are found in the latter zone whilst
L . racemosa is absent altogether. The diverse
m angrove fauna o f crustaceans and m olluscs
found in the m angrove o f P o n ta R asa are absent
from the P ortuguese Island m angrove although
rem ains o f these anim als are present providing
evidence o f their recent disappearance.
Methods
Changes in coastline configuration were deter­
m ined using aerial photographs (1:15 000, D ep art­
m ent o f G eographic Surveys, M aputo) for 1958,
1967, 1985 and 1989. Changes in area o f m an­
grove and associated vegetation types were cal­
culated from enlarged ( x 4) images o f these pho­
tographs. Extensive ground truthing to support
the photo-interpretation w as carried out in 1989.
T he frequency and extent o f the tidal recharge
during 1989 for the Portuguese Island m angrove
 51

30° 40° 50°E

20 "

500 IOCOKm
Indian Ocean

15Km

MAPUTO PORTUGUESE
ISLAND

INHACA r
ISLAND /

M aputo B ay

Fig. 1. The geographical setting of Inhaca Island and Portuguese Island. Inset shows the location of M aputo in southern Africa.

w as determ ined by daily m onitoring. W hen flood­ D ensity (num ber o f individuals per ha.), % fre­
ing occurred, the extent w as m easured by m ap­ quency, % m ortality and regeneration (num ber of
ping on the ground with reference to perm anent seedlings per m 2) o f tree species were m easured
‘m arker’ trees in the m angrove. in nine 5 x 5 m replicate quadrats in flooded and
 COMMUNITY AREA (ha)
Ponta Torres
ED Forest 945.8

CHAgricultral land 1792.0 M B S - Marine Biological S tation

f/^l Swam p 642.5

0 1 2 3 A 5 Km
1------ 1---------- 1------- 1_____i_____ i
■ Mangrove 258.6

Fig. 2. Location of study sites (arrows) on Portuguese Island and Inhaca Island (Ponta Rasa). Also shown are the distribution
of the four main vegetation types.
 53

2 Km

1958 1967

1985 1989

Fig. 3. Changes in coastline configuration, Portuguese Island between 1958 and 1989 (continuous line). The original 1958 con­
tour is shown for comparison for 1967, 1985 and 1989 (dotted line). Shaded areas indicate mangrove. Arrow (1958) indicates
original entrance to canal. Arrows (1989) indicate areas where new sand banks are forming.

non-flooded areas o f the Portuguese Island m an ­ Results

grove. H erbaceous cover (% ) was determ ined in
1 m 2 quad rats in each o f the larger quadrats.
C om parative studies were m ade in Bruguiera and
Ceriops zones in the P o n ta R asa m angrove on
In h aca Island with six 5 X 5 m q uadrats in each
o f the zones. Vegetation d a ta were analyzed sta­
tistically by two-way t-test.
Island and mangrove area changes
Aerial photographs for 1958, 1967, 1985 and 1989
show dram atic changes in the coastline configu­
ration o f Portuguese Island (Figs 3a, 3b, 3c and
3d). In 31 years m uch o f the northern coastline
has been eroded away forming a broad embay-
m ent. In the process, m uch o f the m angrove has
54

Table 1. Surface area changes (m2) of Portuguese Island and
mangrove.
A rea (m2)
Portuguese Island
1958 3027000
1967 2619000
1985 1519000
1989 2161000
also been eroded away, the canal leading into the
m angrove h as been closed and the rem aining
m angrove has becom e isolated from direct access
to the sea by the form ation o f beach ridges.
In co n trast to these erosion patterns, a new
peninsula h ad form ed in the north east o f the
Islan d by 1985, the process continuing until 1989.
The prograding nature o f this peninsula is high­
lighted by the form ation o f sand banks in an area
around the northern shore, the erosion o f this
shore apparently having halted by 1989 (Fig. 3d).
Table 1 shows changes in area o f Portuguese
Island and associated m angrove.
Tidal exchange
D uring 1989 the Portuguese Island m angrove ex­
perienced only three periods o f tidal flooding:
three days in February, five days in M arch and
four days in Septem ber. A t these times (corre­
sponding approxim ately to the equinoxes) the
Area (m2)
tidal am plitudes were sufficient to breach the
Mangrove newly-form ed beach ridges on the northern shore
separating the m angrove from the sea.
110150 A lthough tidal recharge o f the m angrove took
73860
place, the reduced strength o f the tidal w ash re­
28430
26240 sulted in only partial im m ersion o f the m angrove.
F looded and non-flooded areas o f the m angrove
corresponded to 72% and 28% respectively.
Vegetation
The density (num ber o f individuals per ha.), fre­
quency (% ), m ortality (% ) and regeneration
(num ber o f seedlings per m 2) for Ceriops tagal,
Bruguiera gymnorrhiza, Avicennia marina and
Rhizophora mucronata in the Portuguese Island
m angrove is given in Table 2. A com parison is
m ade between flooded and non-flooded areas of
the m angrove.
C. tagal and B. gymnorrhiza are by far the m ost
im portant tree species o f the m angrove. A . marina
and R . mucronata occur at low densities and fre­
quencies in both flooded and non-flooded areas.
M ortality o f C. tagal is highest with a significantly
higher m ortality o f individuals o f this species in
the non-flooded areas. M ortality o f the other spe­
cies is low. H ow ever, regeneration o f C. tagal is

Table 2. Density, frequency, mortality and regeneration of Ceriops tagal, Brugieria gymnorrhiza, Avicennia marina and Rhizophora
mucronata in flooded and non-flooded areas in the mangrove o f Portuguese Island. Standard errors about the mean are given in
brackets. Differences of significance are between flooded and non-flooded areas.

Density Frequency (% ) M ortality (% ) Regeneration

(no. of individuals ha - *) (seedlings m 2)

Flooded Non-flooded Sign Flooded Non-flooded Sign Flooded Non-flooded Sign Flooded Non-flooded Sign

C. tagal 22444 19777 n.s. 76.2 72.5 n.s. 17.9 56.0 *** 42 28 n.s.
(± 3 2 3 0 ) (± 6387) (± 5 .4 ) (± 5 .8 ) (± 4 .9 ) (± 6 .2 ) (± 8 ) (± 6)
B. gymnorrhiza 5933 4022 n.s. 19.8 14.6 n.s. 0.7 0 n.s. 12 12 n.s.
( ± 1869) (± 1126) (± 5 .4 ) (± 4 .4 ) (± 2 .1 6 ) (± 8 ) (± 3 )
A. marina 133 555 n.s. 0.1 9.8 n.s. 0 5.8 n.s. 0 0 n.s.
(± 9 5 ) ( ± 385) (± 0 .7 ) ( ± 8 .1 ) (± 5 .8 )

R. mucronata 622 311 n.s. 3.04 0.55 n.s. 0 7.9 n.s. 0 0 n.s.
( ± 327) (± 3 1 1 ) (± 1 -5 ) (± 0 .5 5 ) (± 7 .9 )

n.s. = n o t significant; ***/7<0.001; n = 9.

 55

Table 3. Percentage cover of herbaceous species in the Por­ Sporobolus virginicus (L.) K unth. C over d ata (% )
tuguese Island mangrove. Differences of significace are be­ for these three species in flooded and non-flooded
tween floded and non-flooded areas (*** = p < 0.001). Stan­
areas is given in Table 3. F o r all three species
dard errors about the mean are given in brackets (n = 9).
percent cover w as significantly higher in non-
Cover (% ) Sign. flooded areas than in flooded areas.
A com parison o f m ortality and the regenera­
Flooded Non-flooded
tion status o f the tw o dom inant species, C. tagal
Sesuvium portulacrastum 5.5 ( + 3.6) 34.4 (± 7 .6 ) Hi** and B. gymnorrhiza, in the ‘stressed’ (restricted
Sporobolus virginicus 5.5 (± 5 .5 ) 33.3 (± 5 .8 ) *** tidal recharge) m angrove o f Portuguese Island
Chenolea diffusa 0 19.4 (± 5 .5 ) *** and ‘non-stressed’ (diurnal tidal recharge) m an­
grove on In h aca Island (P onta R asa) is given in
Figs. 4a and b. F o r the P o n ta R asa m angrove
highest followed by B. gymnorrhiza; no differences d ata is given for both a Ceriops zone and a Bru­
between flooded and non-flooded areas could be guiera zone. Although, as noted, m ortality of
detected. C. tagal is higher in the non-flooded area as com ­
Three invasive herbaceous species were noted pared to the flooded areas in the m angrove o f
in the Portuguese Island mangrove: Sesuvium Portuguese Island (Table 2, Fig. 4a), m ortality in
portulacastrum L., Chenolea diffusa Thunb. and b oth areas is higher than in the Ceriops or Bru­
guiera zones o f the m angrove at P o n ta R asa. The
regeneration status o f C. tagal on Portuguese Is­
land is similar to th at found in the Ceriops zone
at P o n ta R asa.

jC eriops ta g a l

Discussion
B ruguiera g y m n o rrh iz a
£30
1
2 20 It is evident th at between 1958 and 1985 a strong
a ) M o rtality
current moving in an arc ground the north of
10- Portuguese Island led to the erosion o f m uch of
J± L this shoreline. In the process large areas o f m an­
grove have been w ashed aw ay with a loss o f an
area m ore than 75% o f the 1958 cover (Fig. 3,
Table 1). Between 1985 and 1989 new tidal forces
b) Regeneration appear to be operating resulting in deposition o f
sand at various locations around Portuguese Is­
land. O ne o f the consequences o f this deposition
S1^ of sand is th at the Portuguese Island m angrove is
20
isolated by beach ridges which are only rarely
breached by tidal wash. Tinley (1985) has pro­
posed th at a continental shelf counter-current,
moving in a direction opposite to the deep w ater
Non flo o d e d C eriops z o n e B ru g u iera z o n e

P o rtu g u e s e Islan d
Fig. 4. Percentage mortality (Fig. 4a) and regeneration
(Fig. 4b) of C. tagal and B. gymnorrhiza in flooded and non-
flooded areas of the Portuguese Island mangrove and in Ce­
riops and Bruguiera zones of the Ponta Rasa mangrove (In­
haca Island). Standard errors about the mean are shown.
P o n ta R a s a
M ozam bique C urrent, tran sp o rts sand n o rth ­
w ards along m uch o f the M ozam bican coast. It
also appears from a study o f aerial photographs
th at the northern-m ost point o f Inhaca Island
acts as a fulcrum, sweeping sand around this point
w estw ards tow ards Portuguese Island. A sand
56
b ar now extends betw een the tw o Islands and
around the northern shore o f Portuguese Island
(Fig. 3d). These processes have probably resulted
in the prograding beaches presently being form ed
on Portuguese Island and consequent im pound­
m ent o f the m angrove.
W hatever the m echanism s leading to im pound­
m ent, by 1989 the m angrove was only occasion­
ally flooded by tidal w ater and, because o f re­
duced tidal flow, only partial im m ersion o f the
m angrove occurred. Similar im poundm ents and
restricted tidal exchanges have been reported by
Breen & Hill (1968) for the K osi Estuary in South
Africa, 120 km south o f In h aca Island, and by
G o rd o n (1987) for the arid D am pier-C ossack re­
gion o f W estern A ustralia leading, in both cases,
to m ass m ortality o f m angrove trees.
In the case o f the Portuguese Island m angrove
the dom inant species, Ceriops tagal, appears to be
m ost affected by restricted tidal exchange. M o r­
tality rates o f C. tagal are m uch higher than indi­
viduals o f the sam e species receiving regular ex­
changes o f tidal w ash at the P o n ta R asa m angrove
on In h aca Island (Fig. 4a). Individuals o f
Bruguiera gymnorrhiza, a codom inant, do not ap ­
pear to suffer the sam e consequences. N o exact
d ata as to the im poundm ent o f the m angrove is
available, but it probably occurred some time dur­
ing 1988. C. tagal, therefore, appears to be m ost
susceptible to the new adverse conditions. This is
further supported by the fact th at extrem e and
less-extrem e conditions exist in the m angrove i.e.
non-flooded and occasionally-flooded areas.
D eath o f C. tagal is significantly higher in the
non-flooded areas (Table 2). The susceptibility of
C. tagal m ay be due to several factors. Along the
E ast African coast, C. tagal is alm ost at its
southern-m ost distributional limit (C hapm an,
1976) and any environm ental shift w ould have a
proportionally larger effect on the growth o f this
species. H utchings & Saenger (1987) report that
the root system o f C. tagal is particularly suscep­
tible to sedim entation by sand, a process which
m ay be occurring in the m angrove. The w ork of
Breen & Hill (1968) at K osi Bay, where C. tagal
is rare, showed th at all m angrove species, includ­
ing C. tagal suffered death after short term restric­
tion o f tidal exchange and im poundm ent of fresh
w ater. N o im poundm ent o f fresh w ater w as o b ­
served in the Portuguese Island m angrove al­
though it is possible th at the m angrove suffered a
short-term im poundm ent o f fresh w ater due to
rainfall w hich m ay have led to the death o f C. ta­
gal. A nother cause o f die-back could be due to
increasing salinity in the m angrove. The hydro-
logical and clim atic conditions on Portuguese Is ­
land (little fresh w ater runoff, low rainfall and
high evaporation) w ould probably prom ote salin­
ization o f the soils. A n increase in salinity is fur­
ther indicated by the invasive spread o f halophytic
herbs in areas o f the m angrove no t flooded by the
occasional tidal recharge. G o rd o n (1988) reports
th at salinity levels o f 90%o were associated with
death o f C. tagal and other species in W estern
A ustralia. Sm ith (1988), in laboratory grow th ex­
perim ents, dem onstrated th at salinity levels of
60%o resulted in death o f C. tagal. T he salinity
and ground w ater conditions in the Portuguese
Island m angrove are presently being studied. The
disappearance o f the once diverse fauna m ay also
play a role; M acN ae & K alk (1969) point out the
close association betw een sesarm id crabs and the
roots o f C. tagal which enhance aeration o f the
ro o t system.
Som ew hat contradictory is the observation
that, although C. tagal has the highest m ortality
rates, regeneration status is com parable with that
o f the P o n ta R asa m angrove on In h aca Island
(Fig. 4.b). This m ay be explained by C. tagal
shedding its propagules at a time w hen conditions
in the im pounded m angrove were suitable for
generation viz. sufficient w ater for rooting and
germ ination. Propagules o f C. tagal and B. gym ­
norrhiza on In h aca Island begin germ ination in
January and are shed in late January and F ebru­
ary at about the sam e tim e as one o f the last tidal
floodings o f the Portuguese Island m angrove
(F ebruary and M arch 1989). D evelopm ent of
these seedlings is now being studied and the re­
sults will be presented in a later paper.
The future o f the Portuguese Island m angrove
will depend on the build-up o f beach ridges
around the m angrove. The rapid death o f C. tagal
m ay be followed by increased m ortality o f the

co-dom inant, B. gymnorrhiza. The new ly-estab­
lished seedlings o f these tw o species will certainly
require a regular exchange of sea w ater if they are
to flourish. W ard ei'al. (1986) have show n that
re-establishm ent o f all m angrove species is pos­
sible even after m ass m ortality due to im pound­
m ent providing th at physical factors preventing
tidal recharge are rem oved (in this case a sand bar
blocking the m outh o f the K osi Bay estuary south
o f In h aca Island). If no natural tidal recharge of
the Portuguese Island m angrove occurs in the
n ear future and deleterious trends continue, then
m anagem ent options such as artificially creating
a canal m ay have to be considered.
Acknowledgements
The authors gratefully acknowledge financial sup­
p o rt by the N orw egian O rganization for R esearch
and D evelopm ent for this project. W e thank all
m em bers o f staff o f the M arine Biological R e­
search Station, In h aca Island for their logistic
support throughout the period o f this work.
References
Bennessalah, D., 1988. Manual on Mapping and Inventory of
Mangroves. FAO, Rome.
Breen, C. M. & B. J. Hill, 1968. A mass mortality of man­
57
groves in the Kosi Estuary. Trans, r. Soc. S. Afr., 38:
285-303.
Cott, H. B., 1929. Observation on the natural history of the
land crabs Sesarma meinerti, with special reference to the
theory of warning colours. Proc. Zool. Soc. Lond. 18: 679-
692.
Chapman, V. J., 1976. Mangrove Vegetation. Cramer, Vaduz.
Hutchings, P. & P. Saenger, 1987. Mangrove Ecology. St.
Lucia, Queensland: University of Queensland Press.
Jacques, J. M., 1975. Aperçu sur un biocerose cohere; la man­
grove. N at. Belg. 56: 45-50.
Gordon, D. M., 1988. Disturbance to mangroves in tropical-
arid Western Australia: hypersalinity and restricted tidal
exchanges as factors leading to mortality. J. Arid Envir. 15:
117-145.
Kalk, M., 1959. The fauna of the shores of Northern M ocam­
bique. Rev. Biol. Lisbon, 2, 1.
M acNae, W. & M. Kalk, 1962. The ecology of the mangrove
swamps at Inhaca Island, Mocambique. J. Ecol. 50: 19-34.
MacNae, W. & M. Kalk, 1969. A Natural History of Inhaca
Island, Mocambique. W itwatersrand University Press, Jo ­
hannesburg.
Mogg, A., 1963. A preliminary investigation of the signifi­
cance of salinity in the zonation of species in salt-marsh and
mangrove swamp associations. S. Afr. J. Bot. 59: 81-86.
Sim, T. R., 1910. The flora of Portuguese East Africa. S. Afr.
J. Sei. 6: 294-300.
Smith, T. J., 1988. Differential distribution between subspe­
cies of the mangrove Ceriops tagal: competitive interactions
along a salinity gradient. Aquat. Bot. 32: 78-89.
Tinley, K., 1985. Coastal Dunes of South Africa. South Af­
rican National Scientific Programmes Report No 133.
CSIR, Pretoria, 193 pp.
W ard, C. J., T. D. Steinke & M. C. W ard, 1986. Mangroves
of the Kosi System, South Africa: Their re-establishment
since mass mortality in 1965/66. S. Afr. J. Bot. 52: 501-
512.
.
Hydrobiologia 2 4 7 : 59-75, 1992.
V. Mccanni & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 59
© 1992 Kluwer Academic Publishers. Printed in Belgium.

The seagrass and associated macroalgal vegetation o f Gazi Bay (Kenya)

Eric C o p p ejan s,1 H ans B eeckm an2 & M ie D e Wit

1Laboratorium voor Morfologie, System atiek & Ecologie van de Planten, R .U .G ., K .L . Ledeganckstraat
35, B-9000 Gent, Belgium; 2Koninklijk M useum voor M idden-Afrika, B-3080 Tervuren, Belgium

K ey words: seagrasses, m acroalgae, K enya

Abstract

The seagrass and m acroalgal vegetation o f G azi Bay (at approxim ately 50 km south o f M om basa) have
been studied by m eans o f 88 relevés along 7 transects. C orrelation betw een the distribution o f the
seagrasses and som e abiotic factors (particle size fractions, chem ical com position o f the substrate) is
not well m arked. N evertheless a general zonation and succession o f seagrasses could be established:

1. a transition zone between the m angal and the seagrass beds is covered by Boodleopsis pusilla',
2. the pioneer association Halophila ovalis + Halodule wrightii form s low sandy bum ps at the upper limit
o f the seagrass beds, but also occurs in the whole m idlittoral w here sandlayers have recently been
accum ulated (e.g. on coral platform s);
3. the climax vegetation o f the intertidal zone seems to be Thalassia hemprichii which som etim es is
associated with Cymodocea rotundata and C. serrulata, certainly in deeper pools and close to low
w ater m ark; Halimeda opuntia, Gracilaria salicornia and G. corticata are also frequent in this vegeta­
tion type;
4. from low w ater at neap tide dow nw ards patches o f m onospecific Enhalus acoroides vegetation can
also occur;
5. from m ean low w ater dow n to approxim ately - 1 m mixed m eadow s o f Thalassia, C. serrulata, C. ro­
tundata and Halodule uninervis are well developed; the seaweeds Halimeda macroloba and Avrainvillea
obscura are also typical for this zone; locally patches o f Syringodium isoetifolium grow on small bum ps
and Halophila stipulacea grows as a pioneer on bare sand;
6. from - 1 m dow nw ards the whole lagoon is covered by hom ogeneous, m onospecific Thalassodendron
ciliatum m eadow s, locally replaced by E. acoroides.

Introduction m ents and im proving w ater clarity; as prim ary

The im portance o f the seagrass ecosystem has
only been recognized recently. N ow it is even
supposed to be one o f the m ost productive eco­
system s (M e. Roy & M e M illan, 1977; Z iem an &
W etzel, 1980). T he ecological roles o f seagrass
beds are (den H artog, 1967; D aw es, 1981: 481—
484): sedim ent traps, stabilizing bottom sedi­
producers they have high rates o f production; they
are a direct food source for m any anim als; they
are im portant habitats and shelters for a num ber
o f anim al species; they are im portant substrata
for attachm ent o f epiphytes (H arlin, 1980) which
also provide im portant food sources for small fish
and invertebrates; they are highly efficient in re­
moving nutrients (nutrient stripping) from m arine
60
w aters and surface sedim ents and are therefore
critical plants in the control o f w ater quality of
shallow w aters (Patriquin, 1972).
In G azi Bay the large quantities o f drifting sea­
grasses (especially o f Thalassodendron ciliatum)
and seaweeds (Sargassum sp.) brought into the
m angrove at each high tide (certainly in some
periods of the year) might account for a m ajor
im port o f organic m aterial in th at ecosystem.
N ext to these ecological roles o f the seagrass
ecosystem , the econom ical uses o f these m arine
angiosperm s are em phasized by different authors:
production o f paper because o f the high cellulose/
low lignin co ntent o f the leaves; young rhizom es
as well as fruits o f some species are eaten in dif­
ferent parts o f the w orld; dried seagrass leaves
mixed with terrestrial plant m aterial form an ex­
cellent fodder; seeds o f Zostera are ground as
flour in Sonora, M exico; seagrasses like Thalassia
are excellent for green m anuring.
In G azi Bay several biological research p ro ­
gram s are carried out: Floristics, Faunistics and
Ecology o f K enyan coastal biotopes (F K F O
2.0043.88: 1988-1991), D ynam ics and assess­
m ent o f K enyan m angrove ecosystem s (E E C
project TS2-0240-C (G D F ): 1989-1992), Sys-
tem atics, Ecology and Biogeography o f M arine
O rganism s in the Indian O cean (F K F O
2.0009.92: 1992-1995).
The purpose o f this paper is to give a descrip­
tion o f the seagrass vegetation. It is hoped to be
a m ain fram ework for m ore detailed studies of
this ecosystem.
Material and methods
Description o f the site
A m ap o f K enya with the localisation o f G azi is
given in C oppejans & Beeckm an (1989: 382); a
general description and m ap o f G azi Bay ( = M af-
tah a Bay), approxim ately 50 km south o f M om ­
basa, in C oppejans & G allin (1989: 4 8-49,
Fig. 1); clim atologieal d ata and m ean salinity val­
ues are sum m arized in G allin et al. (1989: 198).
Except for some restricted seagrass grow th in
m angal tide pools the seagrass beds are best de­
veloped from the lower m argin o f the m angal
dow n to the subtidal zone. They cover the whole
lower basin (lagoon) o f G azi Bay, stretching up
to the fringing reef. Only a few sandy patches
(G azi Beach) or fossil coral outcrops do not have
seagrasses. T he com plete area is subhorizontal,
slightly sloping tow ards the tide channel and the
shallow lagoon. A t neap low tide m ost o f the area
rem ains inundated; at spring low tide vast areas
( = the infralittoral fringe) are em ergent except for
the deepest p a rt o f the tide channel and the la­
goon.
The substrate is very diverse because of sedi­
m entation p atterns (C oppejans & Gallin, 1989:
49; O nyango, 1989): locally it is rather coarse
sand, on other places one sinks knee-deep in silty
sand. Large areas are characterized by closely
packed cone-like bum ps, ± 0,5 m high, + 1 m in
diam eter, enclosing small tide pools. The bum ps
are built up by still unknow n anim als which reg­
ularly extrude sedim ents from a central crater.
The solid sandy bum ps are devoid o f seagrasses,
the pools contain a mixed seagrass/seaw eed veg­
etation. This diversity in environm ental condi­
tions results in a zonation and patchiness of the
seagrass vegetation.
Sampling
Field w ork was concentrated in July and August
1987 (D e W it, 1988). A dditional observations
have been m ade in August 1989. Seven transects
through the seagrass beds have been studied d u r­
ing spring low tides (Fig. 1). Five o f them (3 -7 )
are at right angles to the coastline, starting at the
low erm ost pneum atophores o f Sonneratia
(C oppejans & Gallin, 1989; G allin eta l., 1989)
going dow n to the rather hom ogeneous subtidal
m eadow s. The vegetation o f the tide channel was
not studied here. A detailed study o f it has been
carried out by D e P auw (1990). Each transect
w as subdivided in hom ogeneous zones defined by
the presence/absence of seagrass species, visible
changes o f substrate an d /o r relief. In total 88
relevés (1 x 1 m 2) were m ade in the different
zones.
 61

fie ld la b

rV man made c a n a i 5 0 0

h a rb o u r'

m an g ro v e
^ w a te r l e v e l a t
••sand e x tr e m e low t i d e
’bankyv'V-.
/Ya muddy s a n d

d ead c o r a l p la tfo rm

sandy i n t e r t i d a l zone
la r g e ly covered by se a g ra sse s

W
-vCU iiv
<
f i s h e r m e n 1s ^ ~ V« •*> ------- -
hut r¿ VC ~ ^rC'r-
W - ~ ~ ~ ~ rJ ~ ^

^ ~ ~ t" ~ ~ ~ ~
V» ~ ~ ~ iV
„ , _. ^«j rv/
~^
<v « j «J cu cv ™

Vy, ^ ru fvy ^vy~ “ “ ^„ Z r

V~ ~ ~ A ~ rv»^ rv^ nj~<^~j
It 'VJ 'Vy ^ ^
— ÍL.^ ^ , ' (v - v K / ~ ~ ^ ' v
fj /\/ ev\ c\y !~J n-J r*i <\J
/V ~ -v (. -V -V ~ ~ A
¿ ^ ru
Kj ^
^ nj
~ ~ryj rv rv
~~ ~
^ ~^
/ , /V A / ~ ~ ~ ~ ~
/ ^ ^ ~ ^
6 ^ j rxjr\j f\j la g o o n
~ ^
^ ^
^
ru /v/
~ ^ ~ ~ " rss -
'f r y j ' V s v ' V ' V ' ^ n J

4 /V Ay c~u c\j r o 'v* rj rJ AJ A-/ A-

s\s /\J S>J <\j ^ ^
Fjg. 1. Map of the study site at Gazi, with position of the transects 1-7.
62
The presence and cover o f the seagrasses as
well as o f the m acroalgae were recorded using the
Londo-scale (L ondo, 1984). Specim ens which
were unidentifiable on the field (mainly sea­
w eeds) were collected and either pickled in 4%
form alin/seaw ater or prepared as herbarium
specim ens which are deposited in the H erbarium
U niversitatis G andavensis (G E N T ) under H E C -
num bers (H erbarium Eric C oppejans); duplicates
are present in the K enya M arine and Fisheries
R esearch Institute (K .M .F .R .I., M om basa,
Kenya).
Salinity and w ater tem perature o f the tidal
pools along the transects 1, 3 and 4 were m ea­
sured on the field with a com bined conductivity/
therm om eter. Sedim ent sam ples were also taken
along som e o f the transects.
Analysis o f the samples
The unidentified seaweed specim ens were an a­
lyzed in the laboratory for P lant M orphology,
System atics and Ecology. T he chem ical soil fea­
tures were determ ined in the L aboratory for A n ­
alytical and Agrochem istry: the organic substance
by the m ethod o f W alkley and Black (1934); the
concentration o f C a, K and N a by flame p ho to m ­
etry, the concentration o f M g by atom ic absorp­
tion and the concentration o f P by the colori­
metric m ethod o f Scheel (C ottenie e ta l., 1982:
43). The texture o f the soil w as analyzed by the
pipette m ethod in the L aboratory for Agronom ic
Soil Science (for m ore details: D e Wit, 1988: 2 4 -
29).
Numerical analysis
The d ata on benthic plant species and on abiotic
factors were subm itted to m ultidim ensional n u ­
merical analysis (Legendre P. & L., 1984a, b).
The d ata were checked for possible clusterings
betw een seagrasses an d /o r seaweeds and for co r­
relations betw een biotic and abiotic com ponents.
Therefore a m atrix o f 37 species and 88 relevés
w as subm itted to the F O R T R A N 77 program
T W IN S P A N (Hill, 1979). T he tw o-w ay table has
been m anually rearranged in order to show the
m ain gradient in the vegetation. A p art o f the
m atrix, consisting o f 32 taxa and 45 relevés, is
com pared to another m atrix with inform ation on
15 environm ental variables for the sam e 45 sam ­
ple units by the algorithm C anonical C orrespon­
dence A nalysis (C C A ) (C A N O C O , ter B raak,
1988).
The involved environm ental param eters are:
the distance (m) o f the plots to the starting point
o f each tran sect ( = low w ater level), the concen­
trations o f N a, K, C a, M g and P (m g/kg sub­
strate), the quantity o f organic m aterial (% ), the
p H o f the substrate, the proportion o f seven p ar­
ticle size fractions (% ).
C anonical C orrespondence Analysis can be
considered as a m ultidim ensional direct gradient
analysis into which regression and ordination
have been integrated (ter B raak, 1986, 1987;
B eeckm an, 1989). C anonical or constrained or­
dination is based on H utch in so n ’s concept o f the
niche (H utchinson, 1944, 1957, 1978; Schoener,
1989), which w as first introduced in a footnote to
a 1944 paper in Ecology on phytoplankton peri­
odicity: ‘the term niche is here defined as the sum
o f all the environm ental factors acting on the o r­
ganism ; the niche thus defined is a region o f an
n-dim ensional hyper-space, com parable to the
phase-space o f statistical m echanics’ (H utchin­
son, 1944: 20). U sing this concept, C anonical
C orrespondence A nalysis is an appropriate
m ethod to study ecological complexity and inter­
actions (B eeckm an & Lem eur, 1989). T he general
aim o f C CA , w hen used in an exploratory way, is
to turn out an ordination diagram o f sam ples,
species and environm ental variables w hich opti­
mally displays how com m unity com position v ar­
ies as a function o f environm ent. Biplot scores of
environm ental variables give the coordinates of
the heads o f arrow s in the ordination diagram .
The rules for constructing and interpreting
species-environm ent biplots are the sam e as those
given in ter B raak (1987: 127-129) for P C A bi­
plots: the direction o f the arrow indicates the di­
rection in w hich the am ount o f the corresponding
environm ental variable increases m ost; the length

o f the arrow equals the rate o f change in th at
direction; the angle betw een arrow s o f a pair of
variables (environm ental, species or sam ples)
provides an approxim ation o f their pair-wise cor­
relation. The biplot is constructed m ost easily by
draw ing separate plots o f species, sam ples and
environm ental variables on tran sp a ran t paper,
each one with its own scaling. How ever, within
each plot the scale units o f the axes m ust have
equal physical length. The biplot is obtained by
superim posing the plots with the axes aligned and
the origins o f the coordinate systems coinciding
(ter B raak, 1988).
Results
Species list o f the plants
Species o f plants listed here are exclusively from
the transects through the intertidal and subtidal
seagrass beds and partly sand-covered coral p lat­
form. Large intertidal pools with coral boulders
an d /o r dead tree trunks as well as the tide chan­
nel have a m uch m ore diversified algal flora; these
species are n o t included here.
Seagrasses (Helotiales)
D escriptions and illustrations are to be found in:
B rouns & Heys (1986), den H artog (1970), Isaac
F. (1968), Phillips & Mefiez (1988), M oorjani &
Sim pson (1988). They have been sum m arized in
D e W it (1 9 8 8 :4 6 -55). Eleven o f the twelve know n
seagrass species from K enya (Isaac F., 1968) have
been found in the study area: Cymodocea
rotundata A sch. & Schweig. (H E C 8477), C. ser­
rulata (B raun) Asch. & M agnus (H E C 8478),
Enhalus acoroides (L.f.) Royle (H E C 8506),
Halodule uninervis (F orssk.) A sch. (H E C 8476),
H. wrightii A sch. (H E C 8475), Halophila minor
(Zoll.) den H artog (H E C 8471b), H. ovalis
(B raun) H ooker (H E C 8503), H. stipulacea
(F o rssk .) Asch. (H E C 8472) Syringodium
isoetifolium (A sch.) D andy (H E C 8502), Thalassia
hemprichii (E hrenb.) Asch. (H E C 8479),
Thalassodendron ciliatum (F orssk.) den H artog
(H E C 8480). Only Zostera capensis Setch., col­
63
lected in G azi Bay by Isaac (1968: 32) has not
been observed again.
Following the m orphological classification sys­
tem o f the seagrasses o f den H artog (1967, 1977:
92) the seagrasses from G azi Bay can be classi­
fied in six categories o f grow th form s which can
be linked with differences in ecology.
The parvozosterids (e.g. Halodule) with fine,
linear leaves; they can be found in all kinds o f
habitats, from coarse sand to alm ost liquid m ud,
from the intertidal zone dow n to considerable
depth, from open sea to the estuaries and m an­
grove swam ps.
T he m agnozosterids (e.g. Cymodocea, Thalas­
sia) with wide linear leaves; they also grow on
various substrata, bu t they are typically subtidal.
The syringodiids (e.g. Syringodium) with long sub­
ulate leaves; they do tolerate some em ersion, but
are usually restricted to the upper sublittoral.
The enhalids (e.g. Enhalus) with leathery strap ­
shaped leaves; they are restricted to the ‘upper’
p art o f the sublittoral and inhabit sandy and rocky
substrate, being rarely found on m ud.
The halophilids (e.g. Halophila) with delicate,
fast growing rhizom es and elliptic, ovate or lan­
ceolate delicate leaves; as with the parvo­
zosterids they have a very wide ecological range.
T he am phibolids (e.g. Thalassodendron) with
sym podially branched vegetative system, the
leaves being distichously arranged along the up­
right stem s; as the enhalids they are strictly sub­
tidal, euhaline, inhabiting sandy and rocky sub­
strates (dow n to - 3 0 m for T. ciliatum.
Seaweeds
D escriptions and illustrations o f m ost o f the spe­
cies can be found in Jaa su n d (1976) and in M o o r­
jani & Sim pson (1988).
Chlorophyta
Avrainvillea obscura J. Ag. (H E C 6069),
Boodleopsis pusilla (Coli.) Taylor (H E C 6741),
Bryopsis hypnoides (H E C 6903), Caulerpa
sertularioides (G m el.) H ow e (H E C 6076),
Chaetomorpha crassa (Ag.) Kiltz., Cladophora
patentiramea (M ont.) Ktitz. f. longiarticulata
Reinb. (H E C 6840), Enteromorpha clathrata
64

(R oth) Ag., E. flexuosa (W ulf.) J. Ag. (H E C Abiotic variables

5707), E. kylinii Bliding (H E C 6696), E. ramulosa
(Sm ith) H ook., Halimeda macroloba D ecsne. The salinity along transect 1 is significantly lower
(H E C 5709, 6080), H. opuntia (L.) Lam our., (variance analysis, unbalanced designs, STAT-
Nereodictyon imitans G erloff (H E C 6075), Udotea G R A P H IC S : D e Wit, 1988: 15-18) than along
orientalis A. & E. S. G epp (H E C 7377), Ulva transects 3 and 4, varying from 24 to 27%0 in 1,
lactuca L., Ulva reticulata F orssk. from 46 to 56%0 in 3 and from 47 to 52 in 4. The
low salinities in 1 might be due to the nearby
Phaeophyta presence o f the outlet o f the Kidogow eni river
Padina boryana Thivy (H E C 6907), Rosenvingea com bined with seepage which is clearly visible at
orientalis (Ag.) Borgs. (H E C 6695), Sphacelaria some places. T he high salinities in 3 and 4 are due
rigidula K ütz. (H E C 6705a), S. tribuloides M e- to the evaporation from the pools at low tide.
negh. (H E C 6705b). The tem perature o f the tidal pools, m easured
along transects 1 -4 from 8.45 to 12.40 h on a
Rhodophyta sunny day varies from 25,5 at 8.45 to 34,0 °C at
Acanthophora spicifera (Vahl) Borgs. (H E C 5711, 11.30 and 12.04. In a single shallow pool (covered
6700), Amphiroa fragilissima (L.) Lam our., A . ri­ w ith seagrasses) the tem perature w ent up 5 °C in
gida Lam our., Ceramium strictum (K ütz.) H arv. l l/ i hours (from 10 to 11.30). In February 1988
(H E C 5710), Gracilaria corticata Ag. (H E C 5705), a m axim um tem perature o f 42 ° C w as m easured
G. salicornia (Ag.) D aw son (H E C 5704), G. ver­ at noon in one o f these tidal pools.
rucosa (H uds.) Papenf. (H E C 5703), Hypnea M ean values o f N a, K, C a, Mg, P, organic
[identifications within this genus should be re­ m aterial, pH are given in Table 1. A ccording to
garded with care] H. cornuta Lam our. ex Ag. Cox (1959) the relative am ounts o f som e ions in
(H E C 6073), H. hamulosa (Turn.) M ont. (H E C seaw ater, com pared to K are: M g 3, C a 1, N a 28.
5708), H. nidifica Ag., H. valentiae (Turn.) M ont., In the soil potassium is fixed to clay m inerals and
Laurencia papillosa (F orssk.) Grev. generally only represents less than 1 % o f the total
These seaweeds can also been classified ac­ exchangeable fraction. As clay m inerals are ab ­
cording to the substrate where they grow. sent from the soil sam ples from G azi, the K
M ost o f the m acroalgae are growing on hard
substrate: on rock and coral substrate: Cla­
Table 1. Mean values, standard deviation and inflation fac­
dophora, Ulva lactuca, Padina, m ost o f the red
tors of environmental variables.
algae m entioned above; on shells and coral frag­
m ents: Enteromorpha, some Hypnea, Gracilaria Variables Mean Stand, dev. Infl. fact.
verrucosa; epiphytic: Bryopsis, Rosenvingea,
Sphacelaria, Ceramium, Gracilaria corticata. Dist low 102.73 59.50 3.04
N a (mg/kg) 3122.58 1006.33 51.27
Some only occur in or on loose substrate: the
K (mg/kg) 244.74 129.47 91.30
filam entous Boodleopsis largely grows in the silty- Ca (mg/kg) 35901.72 17724.70 55.26
sandy substrate w hereas the upright fronds of Mg (mg/kg) 1891.16 774.61 28.72
Avrainvillea obscura and Halimeda macroloba are P (mg/kg) 22.01 6.68 5.69
fixed in the sand by a pseudobulbous holdfast; Org m at (% ) 1.39 1.45 16.35
pH 8.37 0.23 2.90
Caulerpa sertularioides, Udotea orientalis and
2000-1000 pm 0.34 0.48 17.61
Halimeda opuntia on the other hand develop n u ­ 1000-500 pm 3.09 3.83 55.81
m erous groups o f rhizoids for their fixation. 500-200 pm 26.89 18.80 1088.62
A third group includes seaweeds which are not 200-100 pm 49.52 16.11 930.14
really attached to a substratum but are entangled 100-50 pm 20.29 13.92 636.84
50-2 pm 0.03 0.07 11.05
to other algae or seagrasses: Chaetomorpha crassa,
2 -0 pm 0.08 0.09 17.93
Ulva reticulata, Nereodictyon.

present in them should be com ing from the sea­
w ater. B ased on the K /M g relation it can be con­
cluded (D e Wit, 1988; Table 10 p. 34) th at in
transects 1 and 2 the quantity o f M g is also the
sam e as in seawater. In transects 3, 4, 5 there is
a m arked enrichm ent, up to 5 times.
The am ount of calcium is considerably higher
than in the seaw ater (up to 275 tim es) in the
sam ples o f all transects. A ccording to Patriquin
(1975) the C aC o 3 fraction o f seagrass substrate
is from biotic origin, com ing from the calcified
Halimeda (Chlorophyta), the Corallinaceae
(R hodophyta), corals, foram inifera and gastro­
pods.
The clay and loam fractions o f the soil sam ples
along the transects 1 -5 are generally less than
0,5% . The am ount o f very coarse sand is also
very limited. The proportion o f coarse, medium
coarse, m edium fine and fine sand varies from one
sam pling spot to the other (D e W it, 1988, Table 8,
p. 32), even within a transect. In transects 1-2
they mainly belong to the 500-200 and 2 0 0 -
100 p.m fractions, in transects 3 -4 to the (5 0 0 -
B oo d le op sis pus i 1 la 01
011
Enteromorpha spp.
01011
010111 010110
Laurenci a pap i 11 osa Halophila oval i s
Halodule w r ig h t ii
Fig. 2. TW INSPAN-dendrogram of the relevés indicating the subsequent splitting levels of the TW IN-groups and the indicator
species.
65
200), 200-100 and 100-50 p m fractions, in
transect 5 to the 500-200, 200-100 and 100—
50 p m fractions.
A ccording to the texture triangle o f soil taxon­
omy (FA O 1977) the sam ples can therefore be
described as sand. G rady (1981) found 82-9 9 %
sand in seagrass flat sedim ents in Florida.
T W IN S P A N classification
The 88 relevés are split up in 8 groups:
- T W IN -group 1 with Boodleopsis pusilla as in­
dicator species, characteristic for the high in­
tertidal zone (between the highest seagrasses
and the low erm ost m angroves);
- TW IN -group 2 with Enteromorpha flexuosa
and E. ramulosa as indicator species, charac­
teristic on m uddy substrates in the intertidal
zone;
- T W IN -group 3 with Laurencia papillosa as in­
dicator species, characteristic for thinly sand-
covered coral platform in the intertidal zone;
00
8
010
H alophila s t i p u l a c e a
0101 0100
01010
Thalassondendron c il i a t u m
010101 010100
T h a la s sia hemprichi i Syringodium i s o e t i f o l i u m
Cymodocea rotundata
66

Table 2. Two way ordered TW INSPAN-table after manual rearrangement. The significance of the name codes is given after ta­
ble 7. [1 = < 1%; 2 = l-2 % ;3 = 2 -5 % ; 4 = 5 -10% ; 6 = 30-50% ; 7 = 50-70% ; 8 = 70-90% ; 9 = > 9 0 % ].

12 7 5 6 6 7 7 7 7 1 1 1 1 1 2 2 2 2 3 4 4 4 4 4 4 4 5 5 6 6 6 6 7 7 7 7 8
9 0 12 3 0 9 8 9 1 2 3 5 0 3 4 6 7 1 2 5 9 3 0 1 2 3 4 5 8 7 8 0 3 6 7 4 6 7 8 0 1 2 3 4 5 6 7

BOO PUS 2 3
ENT FLE . 1 . 673
ENT RAM 9 6 7 .
ENT CLA . . . 3
EN T SPE .11.
ULV LAC . . . 3
LAU PAP 1 . 4 5 5
PAD BOR 5 5 . 2 .
ACA SPI 1 2 . . .
CHA CRA 1. 1
GRA SAL 1
GRA ED U
HYP N ID
CLA PAT
HAL OVA 2. 3. 2 2 2 2 3 2 2 3 . 2 3 5 2 . 2. 2 2 2 . 2 3 5 5 . . 2 2 5 . 255
HAL WRI 2 2 2 2 2 2 2 3 3 2 3 5 5 2 3 2 3 3 2 3 2 3 5 3 3 3 4 3 3 5 6 3 7 6 7 5 6 6 6 5 5 5 5 6
THA HEM
CYM ROT
HAL OPU
ROS ORI
HAL UNI
CER TRU
AMP FRA
HAL M IN
CYM SER
CAU SER
SYR ISO
ENH ACO
HAL MAC
HAL STI
THA CIL

EN T KYL
HYP CO R 1. 1
ULV RET 11
NER IM I
AM P R IG
HYP VAL

m m 0

T W IN -group 4 with Halophila ovalis and
Halodule wrightii as indicator species, which
can be considered as a pioneer com m unity
in the intertidal zone;
T W IN -group 5 with Thalassia hemprichii and
Cymodocea rotundata as indicator species,
which can be considered as the climax com ­
m unity in shallow intertidal pools and flats;
T W IN -group 6 with Syringodium isoetifolium
as indicator species, which can be consid­
ered as the climax com m unity betw een low
w ater m ark at neap tide and at spring tide;
including the relevés w ith Enhalus acoroides',
- T W IN -group 7 with Thalassodendron cilia­
tum as indicator species w hich can be con­
sidered as the climax com m unity o f the sub­
tidal zone;
- T W IN -group 8 w ith Halophila stipulacea as
indicator species, w hich can be considered
as an infralittoral pioneer com m unity.
 67

1 1 1 2 2 2 3 3 3 3 3 4 4 4 5 6 6 6 6 7 3 3 3 3 5 5 5 5 2 2 5 5
4 5 6 2 5 8 6 7 8 0 1 2 4 9 6 7 9 0 1 2 4 5 9 5 6 7 8 1 2 5 6 3 4 3 4

BOO PUS
EN T FLE
ENT RAM 2 3 2 5
E N T CLA
ENT SPE
ULV LAC
LAU PAP
PAD BOR
ACA SPI
CHA CRA
GRA SAL
GR A E D U
HYP N ID
CLA PAT
HAL OVA 2 . . . . . 3 ............................ 3 . . .
HAL WRI 2 2 . 2 . . . . 2 . 2 2 3 . 5
. . . . 2 2 2
T H A H EM . 5 . 5 3 3 5 5 4 3 6 3 3 5 5. . 5 2 5 4 5 3 2 5 5 6 6 5 6 3
CYM ROT 2 5 2 5 3 4 5 5 4 5 5 5 5 3 3 5 3 4 4 4 2 5 . . . 5 3 . 2 5
HAL O PU . 3 . 2 1 2. 5 2 4 4 . . . 2
ROS ORI 1 . 2 . 1 1 . . . . 1 1 . .
HAL U N I 2 2 3
CER TRU 3 5 5
AM P FRA
HAL M IN
CYM SER
CA U SER 2 1
SYR ISO 5 2 4 3 5 5 5 3
E N H ' ACO . 5
HAL MAC 1
HAL STI 3 3
TH A CIL 8 9

ENT KYL
HYP COR 1 1
ULV RET
NER IM I
AM P RIG
HYP VAL

0 0

Correlations between the vegetation and abiotic fa c ­ cal significance for explaining the variability are:
tors (CCA) distance to low w ater, P, organic m aterial, pH ,
fractions 2000-1000, 5 0 -2 , 2 -0 /im . W ithin the
Intercorrelations between environmental factors latter, the C anonical C orrespondence Analysis
Some variables are strongly correlated with oth­ was applied again.
ers (inflation factor > 2 0 ) and therefore do not The correlation m atrix specifies the intercorre­
have an univocal contribution to the explanation lations between some o f the variables o f the ex­
of the variation in vegetation. They are: N a, K, cluded group: N a: strong correlation with K and
Ca, Mg, fractions 1000-500, 500-200, 200-100, small particle fractions; K: strong correlation with
100-50 /um. P aram eters which do have an univo­ organic m aterial and small fractions; Ca: strong
 68

correlation with M g and P, strong negative co r­

relation with the coarse fractions; Mg: strong co r­
relation with C a and P, negative correlation with
the coarse fractions; particle fraction 1000-
500 ¿un: positive correlation with other coarse
fractions, distance to low w ater, negative co r­
o VO ©
relation with C a, Mg, P and small fractions; frac­
p p cq
Ö Ö © tion 500-200 n m: positive correlation w ith coarse
1
fraction, negative with small fractions, Ca, M g
O IO Os CO SO and P; fraction 200-100 ¿un: negative correlation
O tq SO p 04
Ö Ö Ö © with distance to low w ater, Ca, Mg, P and coarse
1 1 1
fractions; fraction 100-50 ¿un: negative correla­
o o CO Os os oo tion with coarse fractions, positive correlation
p t> tq cq ©
o ö © © with C a and Mg.
1 1 1

o VO 04 04 IO 04 CO Canonical ordination axes

p OO IO vq 04 © 04
o Ö Ö O Ö © To evaluate the reaction o f the species to the
1 1 1 1
current environm ental variables, a M onte Carlo
o 04 r- IO O' VO perm utation-test has been calculated. The first
p cq 04 04 O ©
Ö o Ö Ö Ö © © eigenvector calculated from the d ata - as o b ­
1 1 1 1 1
served - is larger than m ost o f the test statistics
o 04 m r- o CO Os 04 calculated from the random d ata (P = 0.13), which
p cq p VO tq
© Ö o o Ö Ö © © indicates th at the species do react to the current
1 1 1 1
environm ental variables.
o co co Os VO Os CO CO IO) F o r the evaluation o f the contribution o f the
o 04 o vq p 'vf cq cq © ©
r—
<Ö1 Ö Ö Ö o Ö Ö © © rem aining variables in the variance o f the d a ta ­
1 1 1
set, the length o f the vectors corresponding with
O OO TT o rT oo VO each variable has been calculated in a four­
p vq p 04 vq rq 00 SO vo CO
,—lÖ o Ö o o Ö Ö Ö © © dim ensional space.
1 1 1 1
In decreasing im portance the variables are: P,
o C-" co oo Os co Os VO VO CO fraction coarse sand (1000-2000 /jm), distance to
p OO VO 04 •o vq oo so so 04 ©
Ö O Ö © © Ö Ö Ö o © © low w ater, organic m aterial, fraction 5 0 -2 ¿un,
1 1 1 1 1 1
fraction 2 -0 ¿¿m.
O ■■vi­ o co o OS VO io CO VO VO Axes 1 -4 explain 85.7% of the total variance
O co Os 04 p O p oo OO
© Ö o Ö © Ö Ö Ö o Ö © © o f the d ata set.
1 1 1 1 1
T he correlation o f environm ental variables with
o 04 Os io Os 04 to io o- Os 04 so the axes is defined by Table 6.
© oo IO O vq p 04 04 tq p p OO OO
Ö © Ö Ö Ö Ö Ö Ö Ö Ö Ö © © Axis 1: m ost o f the variables are positively cor­
1 1 1 1
related with this axis; only pH is negatively cor­
Table 3. Correlation matrix.

o co VO «O oo t" Os VO OO O' O' O' oo VO related. Axis 1 is mainly defined by the am ount of
o © o 04 04 p 04 '3- 04 so 04 ©
r-î Ö © Ö Ö Ö Ö Ö Ö Ö Ö Ö o © © organic m aterial and the particle fraction 5 0 -2 ¿un
in positive direction and p H in negative direction
(higher pH values corresponding with smaller co­
3. GC C
'S ^ ^ 'S 0 =3. =3. 5S. ordinates).
go 2*¡ bß O
o
O oO oO o 3. Sa. a.
o
E
6Û\ M 5? ^ cd - H i n M r H (N) O
Axis 2: the defining variables are the distance
S J g Ë »G I I I I I I I
O O O O O O 04 to low w ater m ark in positive direction and P in
'S A" 00
^ ff E O
ô öO O O O UO
> Q Z, *4 u S a* o o.S 2 IO (N -H negative direction. C oarse sedim ents are also
 69

Table 4. Biplot scores of environmental variables. (Ax = Axis).

Ax 1 Ax 2 Ax 3 Ax 4 Length vectors Rank

(4-dimensional)

Dist low 81 357 -7 8 -9 374 3

P (mg/kg) 86 -260 -3 0 8 -8 5 421 1
Org mat (% ) 277 -9 1 93 167 349 4
pH -2 0 3 70 - 134 -1 5 2 295 5
2000-1000 ¿im 60 218 290 - 164 403 2
50-2 gm 215 -5 1 43 87 241 6
2 -0 gm 54 - 190 -2 1 116 230 7

Table 5. Percentage variance accounted for by first 4 axes. the particle size fraction 2000-1000 p m , organic
m aterial and particle size fraction 5 0 -2 p m , P
Axis Percentage Eigenvalue
and particle size fractions 2 -0 )im; negative cor­
1 37.6 0.67 relation between pH and organic m aterial or 5 0 -
2 22.5 0.40 2 fim size fraction.
3 14.2 0.25 Well m arked correlations betw een environm en­
4 11.4 0.20
tal factors and species (diagram 2) are no t n u ­
m erous: Boodleopsis pusilla (5), Enteromorpha
Table 6. Inter set correlations of environmental variables with flexuosa (17) and E. ramulosa (14) are positively
axes. (Multiplier 100).
correlated with high values o f organic m aterial
Variables Axis 1 Axis 2 Axis 3 Axis 4 and particle size fraction 5 0 -2 p m and with low
pH values. In the field these seaweed species are
D ist low 157 518 - 128 - 15 typically found (in and) on the m ud flats devoid
P (mg/kg) 166 -3 7 7 -5 0 7 - 137 o f seagrasses. Boodleopsis pusilla is a tiny filamen­
Org mat (% ) 535 - 133 154 270
tous algae, growing in dense m ats, the vertically
pH -3 9 1 102 -2 2 1 -2 4 7
2000-1000 115 316 478 -2 6 6 growing filam ents fixing enorm ous am ounts o f ex­
50-2 414 -7 4 72 141 tremely small particles.
2 -0 104 -2 7 6 -3 5 189 C orrelations between the T W IN S P A N relevé
clusters and environm ental variables are ex­
pressed in diagram 3: the T W IN -groups 1, 2, 3,
positively correlated, small sedim ents rather neg­ 4 + 5 are distributed along the axis distance to low
atively. w ater m ark [from subtidal (1, 2) over infralittoral
Axis 3: mainly defined by P in negative direc­ fringe (3) to m idlittoral (4,5)]; the T W IN -
tion and the fraction very coarse sand in positive groups 4 + 5, 6, 7, 8 (all from the m idlittoral zone)
direction. are distributed along the axes organic m aterial
Axis 4: m ost positive correlation with the and particle size fraction 5 0 -2 p m suggesting that
am ount of organic m aterial and m ost negative the distribution o f the seagrasses within the m id­
correlation with the fraction very coarse sand. littoral zone is due to these environm ental varia­
bles.
Ordination diagrams based on canonical correspon­
dence analysis Conclusions and discussion
The m ost evident correlations appear from the
diagram o f the first and second canonical ordi­ A zonation pattern o f seagrasses and associated
nation axes. C orrelations am ongst environm ental m acroalgae can generally be recognized in G azi
factors are illustrated in diagram 1: positive cor­ Bay. This has been corroborated by the T W IN ­
relation between distance to low w ater m ark and S PA N and C CA -analysis.
 70

AXIS 1
-540 0 540
CD
i_n

dist low

2000-1000
AXIS 2

CD
5 0 -2 ;jm
org mat

2-0 y CD
CD

CD
LTD
I

CO
CD
X

CD
CD

CD
CD

-SO
AXIS
AXIS 2

CD A OA

CD
CD
I
-GO GO
AXIS 1
 71

Table 7. Ranked species coordinates for the 4 canonical ordination axes.

Axis 1 Axis 2 Axis 3 Axis 4

Eig. val. = 0.669 Eig. val. = 0.401 Eig. val. = 0.253 Eig. val. = 0.204

ENT SP5 47 ENT SP3 39 HAL M IN 82 ENT SP2 106

ENT RAM 44 ENT KYL 34 CYM C ER 58 ENT SP3 46
BOO PUS 41 CER cTR 31 AMP R IG 42 PAD SPI 41
SPH SPE 22 ACA SPI 26 AMP FRA 31 CER cTR 38
ENT cCR 20 ENT SP2 26 ENT KYL 29 CAU SER 35
AMP FRA 13 PAD SPI 19 HAL STI 26 AM P R IG 19
PAD SP1 11 SPH SPE 16 HAL WRI 24 SPH SPE 19
ACA SPI 10 HAL OVA 13 PAD SPI 22 GRA SAL 17
CYM SER 9 ENT cCR 11 GRA SAL 19 ENH ACO 15
HAL UNI 8 HAL WRI 10 CER cTR 18 HAL O PU 11
ROS O RI 0 THA H EM 9 CAU SER 16 HAL STI 9
HAL MAC - 1 HYP COR 7 HAL O PU 9 HAL UNI 8
CYM ROT - 2 CYM ROT 6 HAL UNI 9 ACA SPI 6
THA H EM - 2 AMP FRA 5 SPH SPE 4 AM P FRA 6
HYP COR - 3 CAU SER 3 ENT RAM 2 ENT RAM 6
GRA SAL - 7 HAL UNI 1 THA CIL 0 THA CIL 3
SYR ISO - 7 ENH ACO 0 HAL OVA - 1 AMP SPI 1
THA CIL - 7 HAL O PU - 1 HAL MAC - 3 CYM SER - 2
AMP R IG - 8 ROS ORI - 1 ENT SP2 - 4 HAL OVA - 2
HAL OVA - 8 ENT RAM - 2 CYM ROT - 5 CYM ROT - 3
HAL WRI - 9 CYM SER - 3 THA H EM - 5 BOO PUS - 6
CAU SER - 1 2 SYR ISO - 4 ENT SP3 - 6 THA HEM - 7
AMP SPI - 1 3 GRA SAL - 7 ROS ORI - 8 SYR ISO - 8
HAL M IN - 1 3 BOO PUS - 8 BPP PU S - 9 ROS ORI - 10
CER cTR - 14 HAL M IN - 9 SYR ISO - 9 ENT cCR - 14
HAL O PU - 1 6 AMP SPI - 1 2 ENH ACO - 1 1 ENT SP5 - 1 5
HAL STI - 1 8 HAL MAC - 1 5 ENT cCR - 1 3 HYP COR - 1 5
ENT KYL -2 1 ENT SP5 - 17 ENT SP5 - 1 3 HAL WRI - 1 7
ENH ACO - 2 7 AMP R IG - 1 9 AMP SPI - 2 0 ENT KYL - 2 1
ENT SP3 - 4 0 THA CIL - 2 1 HYP COR - 2 2 HAL MAC - 2 3
ENT SP2 - 4 2 HAL STI - 3 6 ACA SPI - 3 6 HAL M IN - 3 3

The numbering of species used for the canonical ordination diagrams are added between [ ].
ACA SPI = A canthophora spicifera [1], A M P FRA = Amphiroa fragilis [2], AM P R IG = A. rigida [3], AM P SPE = A. species
[4], BOO PU S = Boodleopsis pusilla [5], CAU SER = Caulerpa serrulata [6], CER cTR = Ceramium cf. truncatum [7],
CYM ROT = Cymodocea rotundata [8], CYM SER = C. serrulata [9], E N H ACO = Enhalus acoroides [10],
ENT cCL = Enteromorpha cf. clathrata [11], ENT KYL = E. kylinii [13], ENT RAM = E. ramulosa [ 14], ENT SP2, EN T SP3,
ENT SP5 = E. species 2 [15], 3 [16], 5 [17], GRA SAL = Gracilaria salicornia [18], HAL MAC = Halimeda macroloba [20],
H AL M IN = Halophila minima [20], HAL O PU = Halimeda opuntia [19], HAL OVA = Halophila ovata [24], Halophila min­
ima [20], HAL STI = H. stipulacea [23], H AL U N I = Halodule univervis [21], HAL W RI = H. wrightii [22],
HYP COR = Hypnea cornuta [26], PAD SP1 = Padina species [27], ROS O RI = Rosenvingea orientalis [28],
SPH SPE = Sphacelaria species [29], SYR ISO = Syringodium isoetifolium [30], TH A H EM = Thalassia hemprichii [31],
THA CIL = Thalassodendron ciliatum [32].

Fig. 3. Ordination diagrams based on canonical correspondance analysis: 1. Environmental factors (dist low = distance to low
water mark, org mat = organic material, P = phosphorus, 2000-1000 /im, 50-2 /im, 2 -0 /im = particle size fractions); 2. Species
(numbers corresponding with those given after Table 7); 3. Relevé clusters belonging to the TW IN-groups 1: 2: ♦ , 3: A, 4: 0 ,
5: A , 6: □ , 7: ■ 8: O .
72
The intertidal zone
The relevés with the green alga Boodleopsis pusilla,
forming group 1 in the T W IN S PA N -analysis, re­
present the interm ediate zone between the sea­
grass beds and the mangal.
Extrem e m uddy substrates in the intertidal
zone o f G azi Bay are generally devoid o f sea­
grasses: they are covered by Enteromorpha spp.,
which form group 2 in the T W IN S P A N analysis.
A t some places (fisherm en’s harbour) the pioneer
seagrass vegetation Halophila ovalis + Halodule
wrightii starts developing on the mud.
A ccording to the C CA -analysis these T W IN -
groups (1, 2) seem to be correlated to the small
particle size fraction (5 0 -2 p m ) and to the am ount
o f organic m aterial.
The lower p art o f the intertidal coral platform s
are generally covered by Laurencia papillosa
(group 3 o f the T W IN S P A N analysis) com bined
with Padina boryana, Hypnea cornuta and
Acanthophora spicifera. This vegetation enhances
sand retention on which Halophila ovalis and
Halodule wrightii develop as pioneer seagrass veg­
etation, possibly evolving into the climax Thalas­
sia vegetation. This is a similar phenom enon to
the one described by M olinier & Picard (1951,
1952) in the M editerranean sea: seaweeds Padina
and Jania on rocks, silting up, colonization by the
seagrass Cymodocea nodosa (U cria) A sch., devel­
opm ent o f the climax Posidonia oceanica (L.) D e­
file seagrass vegetation.
The com bination o f Halophila ovalis and
Halodule wrightii (group 4 o f the T W IN S P A N -
analysis) occurs in the whole seagrass covered
p art of the intertidal zone. They form the pioneer
com m unity in the intertidal zone, growing as a
rather open cover on sandy patches within the
Thalassia vegetation, as well in shallow pools as
emerged at low tide. They frequently form low
sandy bum ps at the upper fringe o f the seagrass
beds.
M ost o f the intertidal sand flats from M L W N
(m ean low w ater at neap tide) dow nw ards as well
as the highest p arts o f the subtidal zone [dow n to
- 1 m under M L W S (m ean low w ater at spring
tide)] are covered by a mixed seagrass vegetation
m ainly com posed o f Thalassia hemprichii and
Cymodocea rotundata. These m agnozosterids
which form group 5 in the T W IN S P A N analysis
can be considered as the climax vegetation o f this
zone. Thalassia especially is a slow growing spe­
cies which is very firmly anchored in the substra­
tum by well developed rhizom es. O n subhorizon­
tal stabilized surfaces o f sand-covered coral
rubble Thalassia form s an extremely dense, short,
alm ost m onospecific vegetation. After spring ebb
tides a m arked degree o f ‘burning’ can be seen.
In slightly deeper pools with sandy substrate
and on the banks o f the tidal channel Thalassia is
mixed with variable quantities o f Cymodocea
rotundata an d /o r C. serrulata. The latter spp. be­
com e dom inant in pools where the substrate is
regularly shifted (f. ex. the pools between the
cone-like structures). O n places with strong sed­
im entation Halodule uninervis develops as a sec­
ondary pioneer species and sedim ent accum ator
{Halophila stipulacea being the prim ary one). The
relevés including this species do no t appear as a
separate group in the T W IN S P A N analysis. On
the banks o f the tidal channel which are often
eroded, the bare areas are first colonized by the
quick growing Halophila stipulacea, followed by
Halodule uninervis, restabilizing the bottom , giving
the opportunity to Cymodocea rotundata and
C. serrulata to develop. Finally Thalassia ‘defi­
nitely’ stabilizes the substrate. Newly form ed sand
bum ps in the Thalassia bed will then be covered
by Syringodium.
The subtidal zone
Bare surfaces such as ‘potholes’, originating ei­
ther naturally (e.g. grazing, storm ) or from hum an
activities, and lower p arts o f tidal channel banks
are colonized by Halophila stipulacea, a halophilid
which form s group 8 in the T W IN S P A N -an aly ­
sis. H. stipulacea can therefore be considered as
the sublittoral colonizing species.
The whole lagoon (from - 1 m under M L W S )
o f G azi Bay is covered by dense, hom ogeneous
Thalassodendron ciliatum m eadow s. This euhaline
am phibolid which form s group 7 in the T W IN ­
SPA N -analysis represents the climax vegetation
o f the subtidal zone, but this species can also
 73

develop in deep rocky intertidal pools. The leaves M id midlittoral:

and stem s of Thalassodendron are richly covered silty substrate:
by epiphytic algae (which have also been studied - Boodleopsis pusilla
in detail by D e Wit, 1988). Some rhizophytic algae - Enteromorpha spp.
grow characteristically between the T. ciliatum locally evolving to Halophila ovalis + Halodule
stolons: Udotea orientalis, Halimeda macroloba. wrightii
Both o f these show a m arked patchiness, being coral platform :
num erous on some surfaces and alm ost totally - Laurencia papillosa + other algae sand cov­
absent at other places. Avrainvillea obscura ering— >
grows as isolated specim ens all over the lagoon. - Halophila ovalis + Halodule wrightii— >
Halimeda opuntia together with Amphiroa - Thalassia hemprichii + Cymodocea rotundata
fragilissima form scattered brittle cushions on + Halimeda opuntia, Hypnea cornuta,
the bottom , b u t they also occur in the intertidal Chaetomorpha crassa
pools. D eep (rock) pools with sandy bottom :
The loosened stolons o f Thalassodendron at the - Thalassodendron ciliatum on the bottom , Cy­
m argins o f the ‘potholes’ are covered by the brow n modocea on the pool m argins
alga Rosenvingea orientalis. Low sandy bum ps are
generally covered by Syringodium isoetifolium, es­ Infralittoral fringe (betw een M L W N and
pecially close to M LW S, emerging only at ex­ M LW S):
trem e springtides. This syringodiid form s group 6 - Halophila ovalis + Halodule wrightii— >
in the T W IN S PA N -analysis and som etimes - Halodule uninervis— >
grows as an interrupted belt between the m idlit­ - Cymodocea rotundata + C. serrulata +
toral m agnozosterids and the subtidal amphibol- Thalassia hemprichii
ids. - Syringodium isoetifolium
Enhalus acoroides locally grows as m onospe­ - Enhalus acoroides
cific stands in the Thalassodendron beds. This + Halimeda macroloba, Avrainvillea obscura,
subm erged vegetation forms open com m unities, Cladophoropsis vaucheriaeformis (A resch.)
the leathery leaves being vertically directed, the Cribb, Gracilaria corticata, G. salicornia,
eroded apices reaching the w ater surface at Hypnea cornuta
M LW S. Patches of Enhalus vegetation also occur
in the lowest intertidal zone, between M LW N Infralittoral dow n to - 1 m:
and M LW S. A t low tide the leaves lie partly ex­ - Halophila stipulacea— >
posed, partly subm erged on the bottom o f shal­ - Halodule uninervis
low pools or in sandy grooves, all directed sea­ - Thalassia hemprichii + Cymodocea serrulata
w ard by the flow off o f the water. ( + C. rotundata)
- Syringodium isoetifolium
Synthesis o f the plant zonation in Gazi Bay - Enhalus acoroides
+ Rosenvingea orientalis (on seagrass stolons),
High midlittoral: Halimeda macroloba, H. opuntia, Avrainvillea
mangal: obscura, Udotea orientalis, Amphiroa
- Lumnitzera racemosa Van Steenis fragilissima, Caulerpa sertularioides
- Avicennia marina (F orsk.) Vierh.
- Ceriops tagal (Perr.) Rob. + Bruguiera Infralittoral, deeper than - l m :
gymnorrhiza Lam arck + Xylocarpus grana­ - Thalassodendron ciliatum ( + num erous epi­
tum Koenig phytes) (very locally replaced by Enhalus
- Sonneratia alba J. Smith + Rhizophora acoroides).
mucronata Lam arck
74
A ccording to our field observations this zonation
scheme can be generalized for the K enyan coast,
except for Enhalus w hich seems to be a rather rare
species in this region. It also corresponds well
with the one described by B rouns & Heijs (1985)
and Heijs (1987) for S. P ap u a N ew G uinea (P ort
M oresby area) except for the absence o f
Thalassodendron ciliatum in the subtidal zone
around P o rt M oresby; it is there replaced by
Thalassia hemprichii which either form s subtidal
m onospecific stands or is growing together with
other species, forming extensive mixed seagrass
m eadow s.
Acknowledgements
This study w as sponsored by N .F .W .O . (N atio ­
naal F o n d s voor W etenschappelijk O nderzoek),
F .K .F .O . (F o n d s voor K ollektief en F undam en­
teel O nderzoek) and E E C (E uropean Econom ie
Com m unity). W e are also grateful to Ph. P olk for
giving us the opportunity to do botanical research
in K enya within the fram ew ork o f the K enyan
Belgian R esearch Project, to M r. Alíela for p u t­
ting the infrastructure o f the K enya M arine and
Fisheries R esearch Institute at our disposal. O ur
gratitude also goes to I. W olfs, E. M artens and N.
Revis for the local organization o f and T. Beeck-
m an for the help w ith the field w ork and labora­
tory research.
References
Beeckman, H., 1989. Mathematische verwerking van phy-
tosociologische gegevens. Groene Band 76: 1-36.
Beeckman, H. & R. Lemeur, 1989. Ecology, science of com­
plexity and interactions. FT I 1989 Project Sheet Interface
R U G 9: 3 pp.
Brouns, J. & F. Heijs, 1985. Tropical seagrass ecosystems in
Papua New Guinea. A general account of the environment,
marine flora and fauna. Mar. Biol. 88: 13-50.
Brouns, J. & F. Heijs, 1986. Structural and functional aspects
of seagrass communities and associated algae from the
tropical West-Pacific. Ph.D. thesis Nijmegen (Nether­
lands). 431 pp.
Coppejans, E. & T. Beeckman, 1989. Caulerpa section Se-
doideae (Chlorophyta, Caulerpales) from the Kenyan coast.
N ova Hedwigia 49: 381-393.
Coppejans, E. & E. Gallin, 1989. Macroalgae associated with
the mangrove vegetation of G azi Bay (Kenya). Bull. Soc.
r. Bot. Belg. 122: 46-60.
Cottenie, A., M. Verloo, L. Kiekens, G. Velghe & R. Cam-
erlynck, 1982. Chemical analysis of Plants and Soils.
I.W .O.N.L. Brussel. 63 pp.
Dawes, C. J., 1981. Marine Botany. John Wiley & Sons, New
York.
den Hartog, C., 1967. The structural aspect in the ecology of
sea-grass communities. Helgol. Wiss. Meeresunters. 15:
648-659.
den Hartog, C., 1970. The sea-grasses of the world. Verh.
Kon. Ned. Ak. Wetensch., Natuurk. Reeks 2, 59: 1-275.
den Hartog, C., 1977. Structure, function, and classification
in sea-grass communities. In: McRoy & Helfferich (eds),
Seagrass Ecosystems: a scientific perspective. M. Dekker
Inc., New York: 89-121.
De Pauw, K., 1990. De vegetatie van de getijdengeul van Gazi
Bay (Kenia). M. Sc. thesis State University G hent (Bel­
gium): 142 + tab.
De Wit, M., 1988. Vegetatie-ecologische studie van de zee­
grasgemeenschap (incl. wiercomponent) in G azi Bay,
Kenia. M. Sc. thesis State University G hent (Belgium):
189 +81 pp.
FAO, 1977. FAO-guidelines for soil profile description. Rome.
66 pp.
Gallin, E., E. Coppejans & H. Beeckman, 1989. The man­
grove vegetation of Gazi Bay (Kenya). Bull. Soc. r. Bot.
Belg. 122: 197-207.
Grady, J. R., 1981. Properties o f sea grass and sand flat sedi­
ment from the intertidal zone of St. Andrew Bay, Florida.
Estuaries 4: 335-344.
Harlin, M. M., 1980. Seagrass epiphytes. In Phillips R. C. &
McRoy C. P. (eds), H andbook of seagrass biology. An eco­
system perspective. G arland STPM Press, New York:
147-152.
Heijs, F., 1987. Community structure and seasonality of macro­
algae in some mixed seagrass meadows from Papua New
Guinea. Aquat. Bot. 27: 139-158.
Hill, M. O., 1979. Twinspan - a FO RTRA N program - for
arranging multivariate data in an ordered two-way table by
classification of the individuals and attribute: 90 pp. Cornell
Univ. Ithaca, New York.
Hutchinson, G. E., 1944. Limnological studies in Connecticut
VII. A critical examination of the supposed relationship
between phytoplankton periodicity and chemical changes
in lake waters. Ecology 25: 3-26.
Hutchinson, G. E., 1957. Concluding remarks. Cold Spring
Harbor Symp. Quant. Biol. 22: 415-427.
Hutchinson, G. E., 1978. An introduction to population eco­
logy. Yale Univ. Press, New Haven, Connecticut.
Isaac, F., 1968. Marine Botany of the Kenya coast. 4. An-
giosperms. J. E. Afr. N at. Hist. Soc. 27: 29-47.
Jaasund, E., 1976. Intertidal seaweeds in Tanzania. Univ.
Tromso. 160 pp.
Legendre, P. & L. Legendre, 1984a. Ecologie numérique. 1.

Le traitement multiple de données écologiques. Masson.
Paris. 260 pp.
Legendre, P. & L. Legendre, 1984b. Ecologie numérique. 2.
La structure des données écologiques. M asson. Paris.
335 pp.
Londo, G., 1984. The decimal scale for relevés of permanent
quadrats. In R. Knapp (ed.), Handbook of vegetation
science 4: 45-50.
McRoy, C. P. & C. McMillan, 1977. Production ecology and
physiology of seagrasses. In: McRoy C. P. & Helfferich C.
(eds), Seagrass ecosystems: a scientific perspective. M. De­
kker Inc., New York: 53-88.
Molinier, R. & J. Picard, 1951. Biologie des herbiers de
Zostéracées des côtes françaises de la Méditerranée. C. r.
Acad. Sei., Paris 233: 1212-1214.
Molinier, R. & J. Picard, 1952. Recherches sur les herbiers de
phanérogames marines du littoral Méditerranéen français.
Ann. Inst. Océanogr. 27: 157-234.
Moorjani, S. & B. Simpson, 1988. Seaweeds of the Kenya
coast. Oxford Univ. Press. Nairobi. 134 pp.
Onyango, H. B. A., 1989. Numerical modelling of Gazi Creek
(South coast, Kenya): 1. Hydrodynamics model; 2.
Descriptive ecological model. M. Sc. thesis FAM E. VU
Brussel.
Patriquin, D. G., 1972. The origin of nitrogen and phospho­
rus for growth of the marine angiosperm Thalassia
testudinum. Mar. Biol. 15: 35-46.
Patriquin, D. G., 1975. ‘Migration’ of blowouts in seagrass
75
beds at Barbados and Cariacon, W est Indies and its
ecological and geological implications. Aquat. Bot. 1: 163—
189.
Phillips, R. C. & E. G. Meflez, 1988. Seagrasses. Smiths.
Contrib. Mar. Sei. 34: 104 pp.
Schoener, T. W., 1989. The ecological niche. In: J. M. Cher-
rett (ed.), Ecological Concepts. Blackwell Scientific Publi­
cations, Oxford, 385 pp.
ter Braak, C. J. F., 1986. Canonical correspondance analysis:
a new eigenvector technique for multivariate direct gradient
analysis. Ecology 67: 1167-1179.
ter Braak, C. J. F., 1987. Ordination. In: Jongman R. H.,
C. J. F. ter Braak & O. F. R. Van Tongeren (eds), D ata
Analysis in community and landscape ecology. Pudoc
Wageningen. 301 pp.
ter Braak, C. J. F., 1988. CANOCO - A FO RTRA N pro­
gram for canonical community ordination by partial de­
trended canonical correspondance analysis, principal com­
ponents analysis and redundancy analysis (version 2.1.).
Agrie. Math. Group. Wageningen. 95 pp.
Walkley, A. & E. A. Black, 1934. An examination of the
Degtjareff method for determining soil organic matter and
proposed modification of the chromic acid titration method.
Soil Sei. 37: 29.
Zieman, J. C. & R. G. Wetzel, 1980. Productivity in
seagrasses: methods and rates. In: Phillips R. C. & McRoy
C. P. (eds), Handbook of seagrass biology. An ecosystem
perspective. Garland STPM Press, New York: 87-116.
 'jp w iy p
"jiUpj»)!)' Mjj
Hydrobiologia 2 4 7 : 77-86, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 77
© 1992 Kluwer Academie Publishers. Printed in Belgium.

M icrophytoplankton of the Pitchavaram mangals, southeast coast of

India: species composition and population density

L. K an n an & K. V asantha
C A S in Marine Biology, Annam alai University, Parangipettai - 608 502, India

K ey words: M icrophytoplankton, P itchavaram m angals, species com position, population density

Abstract

Species richness o f phytoplankton o f the P itchavaram m angals was high. T here were 82 species con­
stituted by 67 species o f diatom s, 12 species of dinoflagellates and 3 species o f bluegreen algae. The
diatom s form ed the bulk with 72% followed by the dinoflagellates with 15% , in which there were au­
tochthonous (tem porary and perm anent) and allochthonous forms. P hytoplankton population density
exhibited a wide seasonal fluctuation with the m inim um during m onsoon and the m axim um during
sum m er, suggesting the possible differential influence o f various environm ental factors.

Introduction position and population density o f phytoplankton

The phytoplankton constitute 95% o f the total
m arine production (Steem ann Nielsen, 1975) so
they form a vital source of energy at the first
trophic tier and also serve as a direct source of
food to m any aquatic anim als. R eports on the
species com position, diversity, distribution, suc­
cession, seasonal, tidal and diurnal variations and
decom position o f phytoplankton are available
from various parts o f the world.
In India, studies on m arine phytoplankton sys-
tem atics, species com position, com m unity struc­
ture, distribution and productivity in relation to
environm ental param eters were carried out ex­
tensively from both the east and w est coasts (eg.
Q asim eta l., 1974; G opinathan, 1975; Vijayalak-
shmi & V enugopalan, 1975; S anthanam , 1976;
S undararaj, 1978; Abidi eta l., 1983; C handran,
1985; Jayalakshm i et al., 1986; R am an and Phani
P rak ash , 1989; Sum itra V ijayaraghavan & Krish-
nakum ari, 1989). The present investigation was
carried out to focus attention on the species com ­
of the P itchavaram m angals and to establish the
relationship betw een phytoplankton and various
environm ental param eters including particulate
organic carbon (P O C ) and m icrobiota.
M aterials and methods
The study area (P itchavaram m angals) is shown
in Fig. 1 along with the sam pling stations. The
study area is located 10 km south o f P orto N ovo
(Lat. 11° 29' N ; Long. 79° 46' E) in the Vellar-
Coleroon estuarine complex. Station 1 is located
in the P itchavaram m angal region where a luxu­
riant m angrove vegetation exists with the dom i­
nance o f Avicennia marina and Rhizophora
apiculata. Station 2 is situated at a distance of
2 km north o f first station and this area is ch ar­
acterized by the rich grow th o f the seaweeds and
seagrasses.
Sam ples were collected at fortnightly intervals
from the sam pling stations during January to D e­
78
India
79*46'E
-n
St. z
SM
• Mangrove
Fig. 1. Map of the study area showing the location of sam­
pling stations.
cem ber 1985 at high tide regimes. M onthly rain­
fall d ata were obtained from the m eteorological
section o f the C entre o f A dvanced Study in M a ­
rine Biology, P orto N ovo. Surface w ater tem per­
ature was recorded using a standard m ercury
therm om eter o f + 0 .5 °C accuracy. Light pene­
tration w as m easured with a Secchi disc and the
light extinction coefficient (K ) w as calculated
using Pool & A tkins (1929) formula. Salinity and
dissolved oxygen were m easured following the
m ethods described by Strickland & P arsons
(1972). p H w as m easured w ith a p H m eter (Elico-
M odel L I-10). D issolved fractions o f the nutri­
ents and P O C were m easured adopting the p ro ­
cedures described by P arsons et al. (1984). T otal
heterotrophic bacteria were estim ated adopting
serial dilution plating technique using ZoB ell’s
2216 e agar m edium (A aronson, 1970).
P hytoplankton w as collected only from the sur­
face w aters from both the stations. F o r qualitative
analysis, a plankton net, N o. 30 (Bolting silk:
m esh size - 48 ¿um) w as used for the collection.
F o r assessing the phytoplankton density, surface
w ater sam ples were collected in 1 litre polythene
bottles to which a few drops o f 5% neutralized
form alin were added as preservative. The sam ­
ples were allowed to settle for 48 hrs. T he super­
n atan t w as discarded and the sedim ented portion
w as used for the quantitative analysis. N um erical
plankton studies were carried out using U ter-
m öhl’s inverted plankton m icroscope. W henever
necessary, sufficient dilutions o f the sam ples were
m ade and the num ber o f dilutions in all the in­
stances were noted and the values m ultiplied ac­
cordingly. F o r each sam ple, countings were m ade
five times and the m ean w as taken (expressed
as cells r 1) as the final value. In the present in­
vestigation, only larger phytoplankton w as con­
sidered and this study provides only a partial co n ­
tribution to the total phytoplankton biom ass
and prim ary production o f the P itchavaram m an ­
gals.
F o r the sake o f convenience and interpretation,
the calendar year w as divided into four seasons
based on the n o rtheast m onsoon w hich alone
brings bulk rainfall to the southeast coast o f India
where the present study area is situated. T he sea­
sons are: M onsoon (O ctober-D ecem ber), P o st­
m onsoon (Ja n u ary -M a rch ), Sum m er (A pril-
June) and P rem onsoon (July-S eptem ber).
Results and discussion
Environmental parameters
M onthly variations o f different m eteorological,
physico-chem ical and biological param eters re­
corded at stations 1 and 2 are shown in Figs 2 to
8.
Species composition
Species com position o f phytoplankton revealed a
total num ber o f 82 species from both the stations.
A checklist o f phytoplankton recorded from both

MONTHS
800-
Fig. 2. Monthly mean of rainfall recorded at Porto Novo (common to stations 1 & 2).
the stations during the study period is given in
T able 1.
Population density
The fluctuations in phytoplankton population
density were wide in different m onths o f the
present study period (Fig. 9). In general, the m ax­
imum density w as recorded during sum m er (M ay)
and the m inim um , during m onsoon and p ostm on­
soon seasons. A t station 1, the phytoplankton
density ranged from 360 (D ecem ber) to 685 260
cells 1“ 1 (M ay) and at station 2, the density var­
ied betw een 420 (January) and 784320 cells 1“ 1
(M ay).
The phytoplankton com prised 67 species o f di­
atom s (Bacillariophyceae), 12 species o f di­
noflagellates (D inophyceae) and 3 species of
bluegreens (C yanophyceae). A total num ber o f 76
species o f phytoplankton w as recorded from sta­
tion 1 which included 84.21% o f diatom s, 13.15%
o f dinoflagellates and 2.64% o f bluegreens. A t
station 2, a total num ber o f 80 species o f phy­
79
1985
toplankton was recorded which included 83.75%
o f diatom s, 13.75% o f dinoflagellates and 2.50%
o f bluegreens. H ence diatom s were the dom inant
phytoplankton group in the study areas. Khalil
(1990) also reported th at diatom s were the dom ­
inant group com prising 52 to 90% in lake M an-
zala (Egypt) owing to their relative tolerance to
salinity.
Generally, the diatom s ranked first in species
com position followed by dinoflagellates and blue­
greens. This tendency w as observed throughout
the study period at both the stations except d u r­
ing the freshw ater flow in m onsoon season when
the dom inance o f freshw ater phytoplankton con­
stituting green algae such as Pediastrum simplex
w as observed.
The perm anent autochthonous diatom species
described from here by Sivakum ar (1982) viz.
Asterionella glacialis, Bacillaria paradoxa,
Chaetoceros affinis, Coscinodiscus marginatus,
Odontella mobiliensis, Thalassiothrix frauenfeldii
and the dinoflagellates, Ceratium furca and Pro­
toperidinium sp. were recorded from station 1 dur-
80

Station I St at i on 2

Sur f ac e w a t e r
35

33

31

29

27

25
35 Atmospheric

33

31

29

J J F F M M A A M M J J J y J y A A S S O O N N D D
MONT HS 1985
Fig. 3. Monthly variations in atmospheric and surface water temperatures.

Table 1. Checklist of microphytoplankton recorded from sta- Table 1. (Continued)

tions 1 and 2. -----------------------------
Serial no. Species
Serial no. Species ----------------------------
18 C. oculus-iridis Ehr.
B A CILARIOPHY CEAE 19 C. rothii (Ehr.) Grun.
1 Actinoptychus undulatus (Bailey) Ralfs 20 C. sub-lineatus (Grun.) Ratt.
2 Amphiprora gigantea Grun. 21 C. thorii Pavill.
3 Asterionella glacialis Castracane 22 Diploneis sp.
4 Bacillaria paradoxa Gmelin 23 Ditylum brightwellii (West) Grun.
5 Bacteriastrum comosum Pavill. 24 D. sol (Grun.) De Toni
6 B. hyalinum Lauder 25 Eucampia cornuta (Cl.) Grun.
7 Chaetoceros affinis Lauder 26 Fragilaria intermedia (Grun.) Grun.
8 C. curvisetus Cl. 27 Guinardia flaccida (Castr.) Perag. H.
9 C. lascinosus Schutt non G ran 28 Gyrosigma balticum (Ehr.) Rabenh.
10 C. lorenzianus Grun. 29 Hemidiscus hardmannianus (Grev.) Mann.
11 C. messanensis Castr. 30 Lauderia annulata (Grev.) Cl.
12 Climacosphenia moniligera Ehr. 31 Leptocylindrus danicus Cl.
13 Coscinodiscus centralis Ehr. 32 L. minimus Gran.
14 C. eccentricus Ehr. ( = Thalassiosira eccentrica) 33 Lithodesmium undulatum Ehr.
15 C. gigas Ehr. 34 Melosira granulata (Ehr.) Ralfs
16 C. lineatus Ehr. 35 M. sulcata (Ehr.) Kutz.
17 C. marginatus Ehr. 36 Navicula sp.
 81

16
o—o S t a t i o n 1 Station 2

12

36

32

28

20

16

J J F F M M A A M M J J J y J y A A S S O O N N D D
MONTHS 1985
Fig. 4. Monthly variations in light extinction coefficient and salinity.

Table 1. (Continued) Table 1. (Continued)

Serial no. Species Serial no. Species

37 Navicula sp. 56 R. styliformis Btw.

38 Nitzschia closterium (Ehr.) Sm. 57 Skeletonema costatum (Grev.) Cl.
39 N. longissima (Breb.) Ralfs 58 Stephanopyxis palmeriana (Grev.) Grun.
40 N. seriata Cl. 59 5. turris (Grev.) Ralfs
41 N. sigma (kutz.) W. Sm. 60 Synedra formosa H antzsch
42 N. vitrea Norman 61 5. ulna (Kutz.) Ehr.
43 Odontella mobiliensis (Bail.) Grun. 62 Thalassionema nitzschioides Grun.
44 0 . rhombus (Ehr.) Kutz. 63 Thalassiosira subtilis (Ostenf.) Gran.
45 O. sinensis (Grev.) Grun. 64 Thalassiothrix frauenfeldii Grun. non Cl.
46 Planktoniella sol (Wallich) Schutt 65 T. longissima Cl. & Grun.
47 Pleurosigma aestuarii Breb. W. Sm. 66 Triceratium favus Ehr.
48 P. directum Grun. 67 T. reticultum Ehr.
49 P. elongatum W. Sm.
50 P. normanii Ralfs D INO PHY CEA E
51 Rhizosolenia alata Btw. 68 Ceratium furca (Ehr.)/Clap. & Lachm.
52 R. imbricata Btw. 69 C. fusus (Ehr.) Dujardin
53 R. robusta Norman 70 C. massiliense (Gourret) Karsten
54 R. setigera Btw. 71 C. trichoceros (Ehr.) Kofoid
55 R. stolteiforthii Perag. H. 72 C. tripos (O.F. Mull.) Nitzsch
82

Table I. (Continued) pos were recorded as perm anent autochthonous

species.
Serial no. Species
The tem porary autochthonous diatom species
73 Dinophysis caudata Sav. Bacteriastrum comosum, Fragilaria intermedia,
74 Gymnodinium sp. Planktoniella sol and Stephanopyxis palmeriana
75 Prorocentrum micans Ehr. and the dinoflagellate species Dinophysis caudata
76 Protoperidinium sp. and Gymnodinium sp. were recorded from sta­
77 P. depressum (Bail.) Bai.
tion 1. F rom station 2, the tem porary autochtho­
78 P. obtusum (Karsten) Parke & Dodge
79 P. oceanicum (Van.) Bai. nous diatom species Guinardia flaccida, Lauderia
annulata, Lithodesmium undulatum and Synedra
CYANOPHYCEAE form osa and the dinoflagellate species
80 Anabaena ambigua Rao, C.B. Prorocentrum micans were recorded.
81 Oscillatoria subbrevis Schmidle
A t both stations, the im portant allochthonous
82 Trichodesmium erythraeum Ehrenberg ex
Gomont constituents w ere Amphiprora gigantea, Ditylum
brightwellii and Rhizosolenia setigera am ong dia­
tom s and Ceratium fusus and C. massiliense,
ing the present study period. A t station 2, the am ong dinoflagellates. T he com m on freshw ater
diatom s, Bacteriastrum comosum, Odontella form s recorded from the tw o stations particularly
sinensis, Melosira granulata, Rhizosolenia alata and during periods o f low salinity were Anabaena
the dinoflagellates, Ceratium trichoceros and C. tri­ ambigua and Pediastrum simplex.

Station I Station 2

o
o

>

J J F F M M A A M M J J J y J y A A S S O O N N D D
M O N T H S 1985
Fig. 5. Monthly variations in dissolved oxygen and hydrogen-ion concentrations.

S tation 2.
o -o S tat ion I
08
20
ce. 0 8
300
•r 2 5 0
3200
150
□ 100
50
MONTHS 1985
Fig. 6. Monthly variations in phosphate, nitrate and silicate
concentrations.
The im portance o f tem perature as a factor in
controlling the grow th o f phytoplankton w as em ­
phasized by m any w orkers (Williams & M urdoch,
1976; G oldm an, 1977). In the present study, m ax­
imum biom ass o f phytoplankton w as observed
when the tem perature o f surface w ater was high.
The peaks o f phytoplankton density (685260
cells 1“ 1 at station 1 and 784320 cells 1“ 1 at sta­
tion 2) were observed when the surface w ater
tem perature (31.0 °C at station 1 and 33.1 °C at
station 2) w as high. H owever, a low phytoplank­
to n density (360 cells 1 ' w as also observed in
D ecem ber when the surface w ater tem perature
w as high (30.2 °C ) at station 1. Sivakum ar (1982)
also noticed the coincidence o f low tem perature
with low phytoplankton density in the backw a­
ters o f P o rto N ovo. H ence influence o f surface
w ater tem perature on phytoplankton density is
rather obscure in tropical m arine environs.
Salinity acts as a m ajor ecological factor co n ­
trolling the distribution and seasonal succession
83
o f phytoplankton (K etchum , 1964; K rishnam ur-
thy et al., 1974; R am an et al., 1990). In the present
study, a low density o f phytoplankton (360
cells 1“ 1 at station 1 and 384 cells 1" 1 at station 2)
w as noticed during late m onsoon and early p o st­
m onsoon seasons respectively w hen the salinity
w as low (20.23%o at station 1 and 21.21%0 at sta­
tion 2) due to the inflow o f freshw ater from the
upper reaches causing greater turbidity. R ecord­
ing o f a low er density o f phytoplankton during
m onsoon m onths was due to the heavy freshw a­
ter inflow which flushed m ost o f the native or­
ganism s including phytoplankton from study
areas into the sea. C adee (1984) also reported
th at excessive run-off might lead to dilution o f the
phytoplankton and produce ‘w ash o u t’.
Subsequently the population density o f phy­
toplankton gradually increased during postm on­
soon season (8 490 cells 1“ 1 at station 1 and
19450 cells D 1 at station 2) and reached the peak
(685260 cells I -1 at station 1 and 784320
cells D 1 at station 2) during sum m er and contin­
ued till early prem onsoon season (102970
cells 1” 1 at station 1 and 127450 cells D 1 at sta­
tion 2). The sum m er peak w as aided by the inflow
o f clear neritic w aters coupled with penetration
o f m ore light into the w ater colum n and flourish­
ing grow th o f m ostly the net phytoplankton
{Bacillaria paradoxa, Chaetoceros affinis, Odontella
mobiliensis, Rhizosolenia alata and Thalassiothrix
frauenfeldii), tolerating a higher salinity (32.32%0
at station 1 and 34.85%0 at station 2). Such a bi-
m odal oscillation o f phytoplankton recording high
density during sum m er and prem onsoon w as ob­
served by R am adhas (1977) in the Vellar estuary.
In general, low concentrations o f nitrate and
silicate recorded at stations 1 and 2 during p o st­
m onsoon to prem onsoon seasons suggest the
possible utilization o f these nutrients in large
am ount by the abundantly occurring phytoplank­
ton. H owever, these tw o nutrients were not lim­
iting as revealed by the good population density
o f phytoplankton especially diatom s though sili­
cate could becom e a limiting nutrient for diatom s
(W etsteyn eta l., 1990).
High standing crop of phytoplankton (685260
cells 1“ 1 at station 1 and 784320 cells D 1 at sta-
84

0-15
□ Station 1
0.13 ■ St a t i o n z

0.11

0-09

ct 0-07
E

0-05
o
CL
0-03

0-01

J J F F M M A A M M J J JyJyA A S S O O N N D D
MONTHS 1985
Fig. 7. Monthly variations in particulate organic carbon.

MONTHS 1985

Fig. 8. Monthly distribution of heterotrophic bacterial population in water.


6 -
- 5
) i F FM M A A M M J J J y J y A A S S O O N N D D
1985
Fig. 9. Monthly variations in phytoplankton population density.
tion 2) dom inated by large diatom s were recorded
during sum m er in which season high contents
(0.115 mg C l -1 at station 1 and 0.148 mg C 1 " 1
at station 2) of P O C and low bacterial counts
(1.213 X IO4 C F U m l -1 at station 1 and
1.643 X IO4 C F U m l -1 at station 2) were re­
corded. This revealed the im portance of phy­
toplankton for the production o f P O M (S arm a &
N agesw ara R ao, 1989) and retardation of het-
erotrophic bacterial grow th owing to their extra­
cellular release o f antibacterial substances
(K ogure eta l., 1979; Sieburth, 1979).
Summary
P itchavaram m angals are rich in m icrophy­
toplankton resources with diatom s, dinoflagel­
lates and blue greens. Both autochthonous and
allochthonous form s are com m on. P hytoplank­
ton population density varies betw een 360 and
85
784320 cells I -1 recording low density during
m onsoon and high density during summer. Sum ­
m er crop o f m icrophytoplankton contributes to
the production o f P O M and retardation o f het-
erotrophic bacterial growth.
Acknowledgements
W e th an k the D irector o f the CAS in M arine
Biology and the authorities o f the A nnam alai
U niversity for facilities and encouragem ent. F i­
nancial help by the U G C , N ew D elhi, is gratefully
acknowledged.
References
Aaronson, S., 1970. Experimental microbial ecology. Aca­
demic press, New York. 236 pp.
Abidi, S. A. H., B. N. Desai & M. Jiyalal Ram, 1983. Stud­
ies on the hydrography and plankton o f waters off Akarpati
 near Navapur, west coast of India during February-M arch.
Mahasagar-Bull. nata. Inst. Oceanogr. 16: 91-94.
Cadee, G. C., 1984. Particulate and dissolved organic carbon
and chlorophyll a in the Zaire river estuary and Plume.
Neth. J. Sea Res. 17: 426-440.
Chandran, R., 1985. Seasonal and tidal variations of phy­
toplankton in the gradient zone of Vellar estuary. M ahasa­
gar-Bull. nata. Inst. Oceanogr. 18: 37-48.
Goldman, J. C., 1977. Temperature effects of phytoplankton
growth in continuous culture. Limnol. Oceanogr. 22: 932-
936.
G opinathan, C. P., 1975. Studies on the estuarine diatoms of
India. Bull. Dept. mar. Sei. Univ. Cochin. 7: 995-1000.
Jayalakshmi, K. V., S. Kumaran & M. Vijayan, 1986. Phy­
toplankton distribution in Cochin backwaters - a seasonal
study. Mahasagar-Bull. nata. Inst. Oceanogr. 19: 29-38.
Ketchum, B. H., 1964. Relation between circulation and
planktonic populations in estuaries. Ecology 35: 191—
200 .
Khalil, M. T., 1990. Plankton and primary productivity in
lake Manzala, Egypt. Hydrobiologia 196: 201-207.
Kogure, K., U. Simidu & N. Taga, 1979. Effect o f Skeletonema
costatum (Grev.) Cleve on the growth of marine bacteria. J.
exp. mar. Biol. Ecol. 36: 201-215.
Krishnamurthy, K., R. Santhanam & V. Sundararaj, 1974.
Species distribution in aquatic environment. Indian J. mar.
Sei. 3: 135-138.
Parsons, T. R., Y. M aita & C. M. Lalli, 1984. A manual of
chemical and biological methods for seawater analysis. Per­
gamon press, Oxford. 173 pp.
Pool, H. H. & W. R. G. Atkins, 1929. Photoelectric measure­
ments of submarine illumination throughout the year. J.
mar. biol. Ass. U.K. 16: 297-324.
Qasim, S. Z., Sumitra Vijayaraghavan, K. J. Joseph & V. K.
Balachandran, 1974. Contribution of microplankton and
nannoplankton in the waters of a tropical estuary. Indian
J. mar. Sei. 3: 146-149.
Ramadhas, V., 1977. Studies of phytoplankton, nutrients and
some trace elements in Porto Novo waters. Ph.D. thesis,
Dept. Mar. Biol. Annamalai University. 235 pp.
Raman, A. V. & K. Phani Prakash, 1989. Phytoplankton in
relation to pollution in Visakhapatnam harbour, east coast
of India. Indian J. mar. Sei. 18: 33-36.
Raman, A. V., Ch. Satyanarayana, K. Adiseshasai & K. Phani
Prakash, 1990. Phytoplankton characteristics of Chilka
lake, a brackish water lagoon along east coast of India.
Indian J. mar. Sei. 19: 274-277.
Santhanam, R., 1976. Hydrobiological studies in Porto Novo
waters - Ecology o f phytoplankton in the different biotopes.
Ph.D. thesis, Dept. Mar. Biol. Annamalai University.
101 pp.
Sarma, N. S. & I. Nageswara Rao, 1989. Distribution of chlo­
rophyll a, organic carbon, nitrogen and carbohydrate of
particulate m atter of Visakhapatnam harbour waters, east
coast of India. Indian J. mar. Sei. 18: 24-29.
Sieburth, J., 1979. Sea microbes. Oxford University Press,
New York. 491 pp.
Sivakumar, V., 1982. An environmental inventory of the tidal
zone of the Vellar estuary. Ph.D. thesis, Dept. Mar. Biol.
Annamalai University. 178 pp.
Steemann Nielsen, E., 1975. Marine photosynthesis with spe­
cial emphasis on ecological aspects. Elsevier Oceanography
series 13, Amsterdam.
Strickland, J. D. H. & T. R. Parsons, 1972. A practical hand­
book of seawater analysis. Fish. Res. Bd. Can. Bull. 167:
1-311.
Sumitra Vijayaraghavan & L. Krishna Kumari, 1989. Primary
production in the southern Arabian Sea during southwest
monsoon. Indian J. mar. Sei. 18: 30-32.
Sundararaj, V., 1978. Hydrobiological studies in the Vellar-
Coleroon estuarine system. Ph.D. thesis, Dept. Mar. Biol.,
Annamalai University. I l l pp.
Vijayalakshmi, G. S. & V. K. Venugopalan, 1975. Studies on
diversity of plankton species, pigments and succession with
a note on primary production at a tidal zone in the Vellar
estuary. In R. N atarajan (ed.), Recent researches in estu­
arine biology. H industan Publishing Corporation, Delhi:
237-246.
Wetsteyn, L. P. M. J., J. C. H. Peters, R. N. M. Duin, F. Veg-
ter & P. R. M. de Visscher, 1990. Phytoplankton primary
production and nutrients in the Oosterschelde (The N eth­
erlands) during the pre barrier period 1980-1984. H ydro­
biologia 195: 163-167.
Wilhams, R. B. & M. B. Murdoch, 1976. Phytoplankton pro­
duction and chlorophyll concentrations in the Beaufort
channel, N orth Carolina. Limnol. Oceanogr. 11: 73-82.
Hydrobiologia 247: 87-98, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 87
© 1992 Kluwer Academic Publishers. Printed in Belgium.

The ecology of Belizean mangrove-root fouling communities: patterns of

epibiont distribution and abundance, and effects on root growth

A aron M . Ellison & Elizabeth J. F arnsw orth 1

Department o f Biological Sciences, Clapp Laboratory, Mount Holyoke College, South Hadley,
Massachusetts 01075-1484, USA; Field Naturalist Program, Department o f Botany, 120 Marsh Life
Sciences Building, University o f Vermont, Burlington, Vermont 05405-0086, USA; 1Current address:
Department o f Organismic and Evolutionary Biology, Biological Laboratories, Harvard University, 16 Divinity
Ave., Cambridge, M A 02138, USA

K ey words: Belize, epibionts, isopods, m angroves, Melongena, Rhizophora mangle

Abstract

The aerial prop roots o f the neotropical red m angrove, Rhizophora mangle L., begin growing well above
highest high w ater (HE1W) and often extend well below lowest low w ater (LLW ) before rooting in the
benthic substratum . In Belize, C entral Am erica, prop roots growing below LLW are colonized by di­
verse assem blages o f organism s, including m acroalgae, hydrozoans, ascidians, sponges, anem ones, hard
corals, and isopod crustaceans. M angroves, root-fouling epibionts, root herbivores, and benthic pred­
ators engage in com plex interactions th at are m ajor determ inants o f m angrove growth and production.
Species richness o f root epibionts increases with distance from the m ainland and with proxim ity to the
barrier reef. Species richness decreases with variability in w ater tem perature and salinity. A scidians and
sponges transplanted from L ark Cay into the coastal Placencia Lagoon failed to survive, but anem o­
nes from L ark Cay survived in Placencia Lagoon. R eciprocal transplants survived off-shore. The gas­
tro p o d predator, Melongena melongena L., present only in m ainland estuaries, reduced local barnacle
abundance and epibiont species richness in Placencia Lagoon. Isopod species richness also increases
with distance from shore, but the num ber o f roots bored by these species decreases. These isopods can
reduce root relative grow th rate (R G R root) by 55% . On off-shore cays, sponges and ascidians amelio­
rate negative effects o f isopods. In m ainland estuaries where epibionts are less com m on, isopod dam ­
age to roots is m ore severe. Experim ental studies in m angrove sw am ps throughout the w orld would
clarify the im portance o f plant-anim al interactions in these w idespread tropical ecosystems.

Introduction and invertebrate epibionts (Riitzler, 1969). Al­

E pibionts are organism s th at colonize and grow
on virtually any living, solid, exposed surface in
the m arine environm ent (reviewed by W ahl,
1989). In m angal, littoral m angrove roots are often
the only local substratum not subject to heavy
sedim entation and provide ideal habitat for algal
though epibionts m ay significantly im pact their
host (and vice versa', W ahl, 1989), m angrove epi­
bionts have received only brief m ention in even
the m ost com prehensive reviews (M aenae, 1968;
C hapm an, 1976; Tom linson, 1986; W ahl, 1989;
but see Riitzler, 1969). Except for species inven­
tories (e.g. M attox, 1949; M aenae & Kalk, 1962;
M aenae, 1968; Ríitzler, 1969; H utchings & R e­
cher, 1974; Sasekum ar, 1974; P into & W ignara-
jah , 1980; Sutherland, 1980; Alvarez I., 1989),
investigators have ignored epibionts (and mobile
anim als) because they have been thought to co n ­
tribute insignificantly to direct m atter and energy
flow in m angrove ecosystem s (e.g. Golley eta l.,
1962; O dum & H eald, 1972; Lugo & Snedaker,
1974), and because the m angrove fauna is not
considered host-specific (e.g. M aenae, 1968;
T om linson, 1986).
T hroughout the neotropics, prop roots o f
R. mangle extending below LLW are colonized
by num erous algal and invertebrate epibionts
(Rtitzler, 1969; Sutherland, 1980; Taylor et al.,
1986; Alvarez I., 1989; Ellison & F arnsw orth,
1990). Intertidal portions o f roots are often col­
onized by barnacles and red algae (Rützler, 1969;
M . M . Littler et al., 1985; Perry, 1988; D. S. Lit-
tler et al., 1989). B oth subtidal and intertidal roots
are bored by isopod crustaceans (S phaerom a­
tidae and Lim noriidae: R ehm & H um m , 1973;
Snedaker, 1973; Estevez & Simon, 1975; Sim-
berloS e t a l ., 1978; Ribi, 1982; Perry, 1988; Ken-
sley & Schotte, 1989; Ellison & F arnsw orth,
1990).
R ecent experim ental studies have show n th at
epibionts directly and indirectly affect ro o t grow th
and production (Perry, 1988; Ellison & F a rn s­
w orth, 1990) and total energy flow (W ada &
W ow or, 1989; R odriguez & Stoner, 1990), and
th at herbivorous crabs can influence m angrove
zonation (Sm ith, 1987; Smith eta l., 1989). Here,
we describe patterns o f epibiont distribution and
abundance on m angrove roots in Belize, C entral
A m erica; discuss some o f the apparent causes of
the observed patterns; docum ent interactions be­
tween various epibionts; and sum m arize the
know n effects o f these epibionts on root and tree
growth.
Study sites
Belize, located in northeastern C entral Am erica,
includes the largest continuous barrier reef in the
w estern hem isphere (R ützler & M acintyre, 1982).
Tidal am plitude throughout Belize is « 30 cm
(Kjerfve et al., 1982). The m angroves Rhizophora
mangle L., Avicennia germinans (L.) Stearn,
Laguncularia racemosa (L.) G aertn. f., and
Conocarpus erectus L. (taxonom y follows Tom lin­
son, 1986) are found along rivers, in estuaries,
and on num erous m angrove cays in the lagoon
shorew ard o f the barrier reef (H a rtsh o rn eta l.,
1984; S to d d art eta l., 1982). All four species co­
occur from H H W to m ean w ater (M W ). Below
M W , R . mangle form s m onospecific stands ex­
tending seaw ard to below LLW .
W e studied m angrove-root epibionts at six lo­
cations in southern Belize (Fig. 1). These sites are
representative o f m ainland estuaries (Placencia
L agoon and Big Creek), near-shore cays (Lark
and Bugle Cays), and cays close to the barrier reef
(N ortheast C ay and Twin Cays). D etailed de­
scriptions o f these study sites are given in Ellison
& F arnsw orth (1990) and F arnsw orth & Ellison
(1991).
Belize
Dangrigae
16° 50’
miles
10
km
Reef
BuC
Fig. 1. M ap of southern Belize, showing the six study sites.
P L : Placencia Lagoon; BgO. Big Creek; LO. Lark Cay; BuO.
Bugle Cay; NO. N ortheast Cay; TO. Twin Cays. Names and
locations follow British Survey Sheets of Belize, nos. 32
(1966), 36 (1976), and 40 (1982).

P a t t e r n s o f e p ib io n t d is tr ib u tio n
Methods
At all six sites (Fig. 1), we sam pled a m inim um of
100 roots for epibionts. R oots were selected h a p ­
hazardly along a 100 m transect parallel to shore.
W e distinguished roots as aerial or ground (rooted
in the substratum ). W e catalogued all epibiont
species encountered. A t Placencia Lagoon, Lark
Clay, and Twin Cays (hereafter referred to as
Placencia, Lark, and Twin, respectively), we m ea­
sured percent cover and vertical zonation o f epi­
bionts on roots. P ercent cover w as assessed by
laying a tran sp aren t 5 x 20 cm sheet o f acetate
divided into grids o f 100 l x l cm squares over
the seaw ard-facing side o f the root. The num ber
o f squares under which each epibiont occurred
w as used as an estim ate o f percent cover (after
correction for actual root diam eter). W e used ad ­
ditional grids to m easure epibiont cover over the
entire root. To assess epibiont zonation, we re­
corded presence o f each species in 2 cm intervals
along each root from 1 cm above the root cap to
the upper limit of epibiont occurrence. Species
occurring at the ro o t tip (up to 1 cm above the
ro o t cap) were recorded separately.
Results
Species encountered at the six sites are listed in
T able 1. T hroughout, we will refer to all these
species generically, except where such reference
would be am biguous. F o r all m ajor taxa, species
richness increased from the m ainland tow ards the
barrier reef (Table 1). A scidians and cnidarians
were absent from the Placencia estuary, and only
one species in each o f these tw o taxa w as present
in Big Creek. Twin and N ortheast Cay, each only
4 km from the barrier reef, had the highest species
richness o f algae, sponges, crustaceans, and as­
cidians.
Epibiont species richness at L ark w as m uch
higher th an at Bugle Cay, although both islands
had fewer species than Twin and N ortheast Cay
(Table 1). Lark and Bugle Cay are both 7 km off
89
shore; Bugle C ay is 2 km SE o f Lark. How ever,
L ark is protected from extrem e wave action by a
series o f cays to the east, while Bugle Cay is m ore
wave-swept. W e suspect th at the low num ber o f
species present at Bugle Cay reflects increased
wave exposure there. Differences in root fauna
observed betw een L ark and Bugle Cay are com ­
parable to differences observed betw een exposed
and sheltered roots at Twin (A. M . Ellison &
E. J. F arnsw orth, pers. obs.), and we are cur­
rently exploring relationships betw een epibiont
patterns and wave exposure.
Species com position and percent cover differed
between aerial and ground roots (Fig. 2), and
these patterns varied with location. A cross all
sites filam entous algae, fleshy algae, and balanoid
barnacles were generally m ore abundant on aerial
roots (Fig. 2B, 2C). C rustose coralline algae and
chtham aloid barnacles were m ore abundant on
ground roots (Fig. 2A). A t Lark, ascidians other
than Diplosoma were m ore ab u ndant on ground
roots (Fig. 2B). A t Twin, however, only
Didemnum conchyliatum was found on both aerial
and ground roots; all other ascidians were found
only on aerial roots (Fig. 2C). Sponges were
found on both aerial and ground roots at all sites.
A t Lark, all sponges other than Haliclona
curacaoensis were m ore ab u ndant on ground roots
(Fig. 2B), but at Twin, sponges were generally
equally ab u ndant on aerial and ground roots.
A t all sites, isopods attacked a m uch higher
percentage o f aerial roots than ground roots
(Fig. 3). W e encountered only Sphaeroma at P la­
cencia and only Phycolimnoria at Lark. T he latter
species accounted for 95% o f the isopod individ­
uals found at Twin (Ellison & F arnsw orth, 1990),
where we also found tw o other sphaerom atids
(Table 1).
T aken over all roots surveyed, epibiont species
richness tended to be higher on aerial roots than
on ground roots. At Placencia, all five epibiont
species occurred on b oth aerial and ground roots.
A t Lark, 33 species w ere found on aerial roots,
and 26 occurred on ground roots. A t Twin, 47
species were found on aerial roots, while only 28
were found on ground roots. Ephem eral cnidaria
(Halocordyle, Halecium, and Sertularia) occurred
90

Table 1. Taxa encountered on mangrove roots in various locations o f coastal Belize (Site abbreviations as in Fig. 1).

Taxa PL BgC LC BuC NC TC

Cyanobacteria1
Scytonem a polycystum Barnet & Flahault y y
Chlorophyta2
Acetabularia crenulata Lamour. V y y y
Caulerpa mexicana (Sond.) J. Ag. y y
Caulerpa racemosa (Forssk.) J. Ag. y y y
Caulerpa sertulariodes (S. G. Gmelin) Howe y
Caulerpa verticillata J. Ag. v/ v/ y y
Derbesia osterhoutii (L. & A. H. Blinks) Page y y
Halimeda spp. y y y y
Valonia macrophysa Ktitz. y y
Ventricaria ventricosa J. Ag. y y
Phaeophyta2
Dictyota spp. y y y y y
Lobophora variegata (Lamour.) Worm. y y y y
Padina sanctae-crucis Borg. y y y y y
Rhodophyta2
Amphiroa rigida var. antillana Borg. y
Bostrychia spp. y y y y y y
Corallinaceae y y y y
Wrangelia argus (Montagne) Montagne y y y
Porifera3
Amphimedon viridis (D. & M.) y y y y y
Chondrilla nucula Schmidt y
Halichondria melanadocia de Laubenfels y
Haliclona implexiformis (Hechtel) y y y y
Haliclona curacaoensis (van Soest) y y y
Haliclona pseudom olitba de Weerdt, Riitzler, and Smith y
Lissodendoryx isodictyalis (Carter) y y y
Mycale microsigmatosa Arndt y
Mycale sp. 2 fide Riitzler y y y
Pellina carbonaria (Lamarck) y y
Tedania ignis (D. & M.) y y y y y
Ulosa ruetzleri Wiedenmayer y y y y
C nidaria1
Aiptasia pallida (Verrili) y y y y
Bartholomea annulata (Lesueur) y y
Condylactis gigantea (Weinland) y y y
Halocordyle disticha (Goldfuss) y y y y
Halecium spp. y y y
Millepora alicornis L. y y
Palythoa caribaea D. & M. y y y
Sertularia turbinata (Lamour.) y y y
Annelida
Sabellidae y y y y
Serpulidae y y y y
 91

Table 1. (Continued)

Taxa PL BgC LC BuC NC TC

Crustacea (Cirripedia)1
Balanus eburneus Gould V y V y y y
Chthamalus angustitergum thompsoni (Henry) y y
Crustacea (Isopoda)4
Dynamenella cf. angulata (Richardson) y
Paracerceis caudata Say y
Phycolimnoria clarkae Kensley & Schotte J y
Sphaeroma terebrans Bate y V
Mollusca (Bivalvia)5
Crassostrea rhizophorae Guilding J y y
Isognomon alatus Gmelin y d d y y
Chordata (Tunicata)6
Ascidia nigra (Savigny) s/ y
Botrylloides nigrum Herdman J y y
Botrylloides cf. perspicuum Herdman J y y
Clavelina picta (Verrili) y
Didemnum conchyliatum (Sluiter) d y y
Didemnum psammathodes Sluiter y
Diplosoma glandulosum (Milne Edwards) J y y
Distaplia corolla Monniot s/ y y
Ecteinascidia minuta (Berrill) y
Ecteinascidia turbinata Herdman y y y
Eudistoma olivaceum (van Name) d s/ y y
Eudistoma obscuratum (van Name) y
Perophora formosana (Oka) y
Perophora regina Goodbody & Cole y
Perophora viridis (Verrili) ■J y y
Styela partita (Stimpson) J y
1 Nomenclature follows Sterrer (1986).
2 Nomenclature follows Littler et al. (1989); algae identified by D.S. Littler, Smithsonian Institution.
3 Sponges identified by K. Riitzler, Smithsonian Institution. Haliclona nomenclature follows de W eerdtetal. (1991).
4 Nomenclature follows Kensley & Schotte (1989); isopods identified by B. Kensley, Smithsonian Insitution.
5 Nomenclature follows McLean (1951).
6 Tunicates identified by I. Goodbody, University of the W est Indies.

only on aerial roots, while other cnidarians (Pa­
lythoa, Millepora, and Condylactis) occurred only
on ground roots. A scidian species richness also
w as m uch higher on aerial roots at these tw o sites
(Lark: 5 species only on aerial roots; Twin: 11
species only on aerial roots). At Lark and Twin,
isopods only attacked aerial roots.
Few clear patterns in vertical zonation o f epi­
bionts on roots were observed. Isopods were
found m ost com m only in the ro o t tip ( F < 0 .0 1 ,
G -test), and barnacles only occurred on the in­
tertidal portions o f the roots (P < 0.001, G -test).
Bostrychia w as m ost com m on at the low w ater
line f P < 0.05, G -test). All o f the other species
occurred only below LLW , and were found with
equal likelihood at all locations along the subtidal
portion o f the root (F > 0 .5 0 , G -test).
92

% o f sample % o f cover in 100

Q I ¡A e ria l r o o ts
50 100 o
100 50 0 Oh G ro u n d r o o ts
o
in
B o stry c h ia 75
>H
BaJanus oa
Chthamalus Q
W
P e11 ina g 50
-I <
H
Chlorophyta H
in 25
* Phaeophyta H
O
Fleshy Rhodophyta o
K
/f C orallinaceae PL BgC LC TC
* Hydrozoa LOCATION

Diplosoma Fig. 3. Percent of aerial and ground roots attacked by iso­

* Other Tunicata
Jf Ulosa
H aliclo n a sp. 1
★ Other P orifera
★ A ip ta s ia
\ Chlorophyta
Phaeophyta
ir Fleshy Rhodophyta
♦ C orallinaceae
Hydrozoa
* Serpulidae
Didemnum
D is ta p lia
Perophoridae
Other Tunicata
H aliclo n a sp. 1
★ H. im periform is
Tedania
Other P orifera
★ A ip ta s ia
Fig. 2. Distribution and abundance o f common taxa at Pla­
cencia (A), Lark (B), and Twin (C). Only taxa that occurred
on at least 10% of the roots sampled (% of sample) are
presented. Mean % cover ( + 1 SE) of each taxa for roots on
which it occurs is given for aerial (open bars) and ground
(solid bars) roots. Stars indicate significant differences in %
cover between aerial and ground roots (P < 0.05,2-tailed /-test
on angularly transformed data [Sokal & Rohlf, 1981]).
pods. Site names as in Fig. 1.
Factors affecting epibiont distribution
Methods - effects o f salinity and water temperature
on epibionts
O ur survey revealed th at several epibionts com ­
m on on the cays, including the colonial ascidian
Diplosoma, the sponge Amphimedon, and the
anem one Aiptasia were absent from Placencia L a­
goon. To determ ine if site-specific salinity and
w ater tem perature affected epibiont distribution,
we reciprocally tran sp lan ted several epibionts be­
tween L ark and Placencia. O n 24 N ovem ber
1988, colonies o f Diplosoma, Amphimedon, and
Aiptasia w ere tran sp lan ted from L ark into P la­
cencia. Pellina, a sponge only found in the m ain­
land estuaries, w as transplanted from Placencia
to Lark. F o r each species, 20 replicate aerial roots
w ith at least 50% cover o f the epibiont were col­
lected at each site. W e tran sp lan ted ten o f the
roots in situ (transplant controls), and tra n s­
planted 10 to the other site. W e m easured percent
cover o f epibionts on the roots, then tied the
transplants onto intact hanging roots with plastic
cable ties. W e tran sp o rted roots in covered buck­
ets betw een sites, and all roots were transplanted
within 1 h o f collection. T ransplants were exam ­
ined every 2 d for 28 d. Salinity (with a refracto-
m eter) and tem perature were m easured concur-
 93

rently. After 1 m onth, transplanted roots had 40 40

begun to decay and the experim ent w as term i­
nated.
30

Results - effects o f salinity and water temperature

20
on epibionts

Diplosoma and Amphimedon died within 4 d of

transplanting into Placencia. At Lark, after 28 d,
50% and 65% respectively of the Diplosoma and
Amphimedon transplant controls rem ained
24 29
(Fig. 4). Aiptasia from L ark survived equally well N ov. D ec .
in situ and w hen transplanted into Placencia
Fig. 5. Salinity (%0) and water temperature (°C ) at Placencia
(Fig. 4). Pellina transplanted to L ark survived u n ­
and Lark during the transplant experiment.
changed through 4 wks (Fig. 4). D uring the first
4 d, salinity at Placencia fluctuated between 18
and 26%0, and w ater tem perature increased by changed as m uch as 6%0 over 48 hr, with com m on
2 °C . Over the 4 wks o f the experim ent, w ater diurnal variation o f l-3 % 0 (Fig. 5).
tem perature averaged 27 ° C at L ark and 28 ° C at
Placencia w ith low variance. Salinity at L ark was
a co n stan t 34%0, but at Placencia ranged from Methods - effects o f predation on epibionts
18%0 to 31%0 (Fig. 5). Salinity at Placencia
W e also studied the effects o f the predaceous gas­
tropod Melongena melongena on epibionts at
125 Placencia. A t this site, Melongena is the m ost
|H i^ i t.h in - s it e tr a n s p la n t
o 1 IB e tw e e n - s ite t r a n s p l a n t com m on benthic predator, and we have ob­
z
z 100 served it preying on barnacles on aerial and
< ground roots at Placencia. Sea urchins, thaid
s
H snails, and herm it crabs, which affect epibiont
K 75
H distributions in Belize and C osta R ica (Taylor
Z
< eta l., 1986; Perry, 1988), are absent at Placencia.
h-) 50
Ph W e random ly chose 16 small (1.0-1.5 m ht)
in
Z R. mangle trees growing between LW and LLW
<!
K 25 for m anipulation. O n 1 F ebruary 1988, we co n ­
E-1
K structed snail exclosure fences (1 .5 x 1.5x 1.0 m,
plastic m esh cages sunk 30 cm into the m ud)
Dg Av Pc Ap
SP E C IE S TRANSPLANTED around 6 o f these trees. O f the rem aining trees,
4 were designated cage controls (cages with
Fig. 4. Success of epibiont transplants between Placencia and
Lark. Diplosoma (Dg), Amphimedon (Av), and Aiptasia (Ap)
only tw o parallel sides) and six were unm ani­
were transplanted from Lark to Placencia (and in situ at Lark). pulated. W e sam pled percent cover o f all epi­
Pellina (Pc) was transplanted from Placencia to Lark (and bionts on all roots o f each tree (> 3 0 0 ro o ts/
in situ at Placencia). Bars show the mean % (cover) of the treatm ent) on 2 February, 22 M arch, 6 N ovem ber,
transplant remaining after 28 d when transplanted in situ (solid 26 N ovem ber, and 14 D ecem ber 1988. A t each
bars) or between sites (open bars). A * above a pair of bars
indicates a significant difference between in situ and between-
sam pling date, we counted the num ber o f Melon­
site transplants (P < 0.05, 2-tailed t-test on angularly trans­ gena in a 100 X 20 m transect th at included these
formed data). trees within it. W e sim ultaneously m easured shell
94

100 length and w idth ( + 1 m m ) o f the snails with ver­

# Cage (s n a ils e x c lu d e d )
nier calipers.
A Cage c o n t r o l
75 O C o n tro l

50 Results - effects o f predation on epibionts

o~.

25 A bundance o f algae (Bostrychia spp. and Caulerpa

verticillata) fluctuated throughout the year on both
T
aerial and ground roots, but there w as no trea t­
100 m ent effect on these species (Figs. 6A, 6B). B ar­
nacles (Balanus and Chthamalus) were m ore
75 ab u ndant on aerial roots (Fig. 6C) than on ground
roots (Fig. 6D ) and barnacle abundance re­
50 m ained relatively co n stan t until N ovem ber. A t
the beginning o f N ovem ber, snail abundance in­
25 creased dram atically (Fig. 7) and barnacle abun­
dance declined on unprotected aerial roots
(Fig. 6C). Sim ultaneously, barnacle abundance
increased on ground roots o f caged plants and
100
decreased in the caged controls (Fig. 6D). The
COVER

barnacle abundance d a ta (Figs. 6C, 6D ) and the

75
snail size d ata (Fig. 7) both indicate possible re­
50
cruitm ent o f these species in late N ovem ber.
PERCENT

However, we have no quantitative d ata on re­

25
cruitm ent in these species. Snails had no effects
on Pellina abundance on roots in any treatm ent
0* (Fig. 6E).

too
Effects of epibionts on mangroves
75
Interactions with M elongena
50
T o determ ine the effects o f these faunal interac­
25

100
ft=- A tions on the trees them selves, we m easured tree
height and stem diam eter (‘d b h ’: above the high­
est prop root) in the Melongena experim ent
(above). T here w as a trend tow ards increased
height and ‘d b h ’ in control and cage-control trees
75 where snails foraged on barnacles, relative to trees

50
Fig. 6. Results of the snail exclosure experiment on epibiont
25 % cover (x + 1 SE) at Placencia. A: Algae (Bostrychia spp.
and Caulerpa verticillata) on aerial roots; B: Algae on ground
roots; C: Barnacles (Balanus and Chthamalus) on aerial roots;
D : Barnacles on ground roots; E : Sponges (Pellina) on ground
0 50 100 150 200 250 300
roots. D ay 0 is 1 February 1988.
DAYS

io
5 SH E L L e
LEN GTH
(m m )
01
0 50 100 150 200 250
DAYS
Fig. 7. Density (m ~ 2) and size distributions of Melongena
during the snail exclosure experiment. Day 0 is 1 February
1988. The inset histograms show snail shell length distribu­
tions on 2 February, 26 November, and 14 December 1988.
Shell length and width are allometrically correlated
(width = 0.43-length ^ = 0.90).
in snail exclosures (P = 0.07, Kruskal-W allis non-
param etric ANO VA). This result indicates some
negative effects on tree grow th by barnacles and
possible am elioration o f these effects by snails.
Interactions with isopods
Isopods (S. terebrans) were rare in these trees, but
in another experim ent at Placencia (Ellison &
Farnsw orth, in prep.), we found th at S. terebrans
reduced R G R root by 60% relative to unattacked
roots. A t Twin Cays, Ellison & F arnsw orth (1990)
found th at the root-boring isopod Phycolimnoria
reduced R G R root ( c m c m _ 1 d _1) o f R . mangle
by 55% . In th at sam e study, we then asked
if sponges (Haliclona curacaoensis and Tedania),
ascidians (Perophora form osana and Didemnum
conchyliatum), anem ones (Aiptasia) and cyano­
bacteria affect isopod colonization o f R. mangle
roots. The sponges and ascidians inhibited iso­
pod colonization, thus indirectly facilitating root
grow th (Ellison & F arnsw orth, 1990). R G R root of
sponge-covered roots grew 83% faster than roots
where we rem oved sponges in situ at Twin (Elli­
son & F arnsw orth, 1990) and 50% faster than
95
roots from which we excluded all fauna (Ellison
& F arnsw orth, in prep.).
Discussion
Together, w ater conditions, wave exposure, root
50 in
status (aerial or ground), proxim ity o f roots to the
barrier reef, and predation pressure influence epi­
biont distribution and diversity on Belizean m an­
grove roots. Epibiont species richness increases
with increasing distance from the m ainland (Ta­
ble 1), and decreasing wave exposure and vari­
300 ability in w ater salinity and tem perature (Figs. 4,
5; B acon, 1971).
Epibiont species com position differs between
aerial and ground roots. Species richness is higher
on aerial roots than on ground roots. Although
one might expect th at ground roots would be m ore
species-rich because of their greater age and size,
we observed the opposite patterns. A scidians and
ephem eral hydrozoans are m ore speciose on
aerial roots. These species m ay be outcom peted
by those sponges and colonial anem ones (as they
are in Venezuela [Alvarez, I., 1989]) th at are m ore
abundant on older ground roots (Riitzler, 1969;
S utherland, 1980). Y oung aerial roots m ay there­
fore serve as refugia for these weakly com petitive
ascidians and hydrozoans. W e also observed a
decline in fleshly algae abundance on ground roots
relative to aerial roots. Aerial roots m ay provide
fleshy algae an escape from herbivorous sea u r­
chins (Taylor eta l., 1986). O ther differences in
species com position between aerial and ground
roots appear to be site-specific (Fig. 2), and may
reflect local variability in recruitm ent dynam ics or
epibiont succession (Sutherland, 1980; Riitzler,
1987; Alvarez I., 1989).
M any ro o t epibionts affect root growth and
production directly (Perry, 1988; Ellison & F a rn ­
sworth, in prep.) and indirectly by ameliorating
the negative effects o f root-boring isopods (Elli­
son & F arnsw orth, 1990). Isopods colonize only
young, aerial, uncorticated roots (R ehm &
H um m , 1973; Perry, 1988; Ellison & F arnsw orth,
1990). At Twin, isopods significantly reduce
R G R root, but the presence o f dense epibiont cover
96
prevents isopod colonization o f roots (Ellison &
F arnsw orth, 1990). In C osta Rica, isopods
(Sphaeromaperuvianum R ichardson) also reduced
ro o t grow th rate by 50% and net root production
by 62% (Perry, 1988). Perry’s experim ental re­
sults also indicated some reduction in the effects
of isopods in roots exposed to w ater colum n
predators, but she w as unable to identify the iso­
pod predator.
In Belize, predation m ay also am eliorate p o ­
tentially negative effects o f epibionts on tree
growth. In the m ainland estuary at Placencia,
Melongena predation reduces barnacle cover on
b o th aerial and ground roots (Fig. 6). T hose trees
on which Melongena reduced barnacle cover grew
m ore rapidly than trees with heavy barnacle cover.
In the only other study o f the effects o f barnacles
on m angrove roots, Perry (1988) found th at Bal­
anus spp. reduced R . mangle root grow th rate
(c m m o n th “ 1) by 30% and net root production
( g r o o t- 1 m o n th “ % by 52% along the Pacific
coast o f C o sta Rica. Perry (1988) found th at b ar­
nacles were preyed upon by snails ( Thais
kiosquiformis D uclus and Morula lugubris A dam s)
and herm it crabs (Clibanarius panamensis Stimp-
son) and proposed th at predators indirectly fa­
cilitate tree growth.
Epibionts are found w orld-wide on m angrove
roots (M attox, 1949; M aenae & Kalk, 1962;
M aenae, 1968; Riitzler, 1969; H utchings & Re-
cher, 1974; Sasekum ar, 1974; Pinto & W ignara-
jah , 1980; Sutherland, 1980; A lvarez I., 1989),
b u t studies o f their roles in the m angal have only
ju st begun (Perry, 1988; W ada & W ow or, 1989;
Ellison & F arnsw orth, 1990; R odriguez & Stoner,
1990). Historically, ecological studies o f m an­
groves have em phasized floristic inventories (re­
viewed in C hapm an, 1976; Tom linson, 1986) and
descriptions o f energy flow and nutrient cycling
(Golley e ta l., 1962; O dum & H eald, 1972; Lugo
& Snedaker, 1974). Although an oft-stated goal
o f faunistic studies in m angal is the determ ination
o f the trophic structure and energy budget o f a
given m angrove swam p (Sasekum ar, 1984), this
goal can n o t be achieved w ithout detailed exper­
im ental studies. Intensive m anipulative experi­
m ents in the m arine rocky intertidal have revealed
trophic com plexity and determ inants o f com m u­
nity structure no t predicted by observation alone
(e.g. Connell, 1961; Connell, 1975; Paine, 1966;
Paine, 1988; M enge eta l., 1986; M enge & S uth­
erland, 1987). Similar studies in tem perature zone
salt m arshes have docum ented m ajor effects of
benthic invertebrates on plant production (Bert-
ness, 1984; B ertness, 1985). In m angrove
sw am ps, recent studies have uncovered unex­
pected interactions betw een herbivorous craps
and plant zonation (Sm ith, 1987; Smith eta l.,
1989), and betw een epibionts and root p roduc­
tion (Perry, 1988; Ellison & F arnsw orth, 1990;
R odriguez & Stoner, 1990). Long-term experi­
m ental studies on plant-anim al interactions in
m angroves throughout the w orld w ould likely
yield new insights into com m unity structure,
functioning, and complexity.
Acknowledgements
E A R T H W A T C H (C enter for Field R esearch),
the N ational G eographic Society, and the Sm ith­
sonian Institution N M N H (with additional sup­
port from E X X O N ) supported this research. W e
th an k six team s o f dedicated E A R T H W A T C H
volunteers for field assistance, and Chris, Skip,
Jessica, and M eredith W hite for friendship and
logistical support at L ark and Placencia.
M eredith W hite collected the salinity d a ta at P la­
cencia. This is contribution no. 303, C aribbean
C oral R eef Ecosystem s (C C R E ) Program , Sm ith­
sonian Institution, and paper 2 in the series ‘The
ecology o f Belizean m angrove-root fouling com ­
m unities’.
References
Alvarez, I. A., 1989. Establecimiento, desarrollo y manten­
imiento de una comunidad epibentonica tropical. Ph.D.
Thesis, Universidad Central de Venezuela, Facultad de
Ciencias. 150 pp.
Bacon, P. R., 1971. The maintenance of a resident population
of Balanus eburneus (Gould) in relation to salinity fluctua­
tions in a Trinidad mangrove swamp. J. exp. mar. Biol.
Ecol. 6: 187-198.
Bertness, M. D., 1984. Ribbed mussels and Spartina

alterniflora production in a New England salt marsh. Ecol­
ogy 65: 1794-1807.
Bertness, M. D., 1985. Fiddler crab regulation of Spartina
alterniflora production on a New England salt marsh. Ecol­
ogy 66: 1042-1055.
Chapman, V. J., 1976. Mangrove vegetation. J. Cramer,
Vaduz, Lichtenstein, 447 pp.
Connell, J. H., 1961. Effects of competition, predation by
Thais lapillus and other factors on natural populations of the
barnacle Balanus balanoides. Ecol. Monogr. 31: 61-104.
Connell, J. H., 1975. Some mechanisms producing structure
in natural communities: a model and evidence from field
experiments. In M. L. Cody & J. Diamond (eds), Ecology
and evolution of communities. Belknap, Cambridge, MA,
USA: 460-490.
Ellison, A. M. & E. J. Farnsworth, 1990. The ecology of Be­
lizean mangrove-root fouling communities: I. Epibenthic
fauna are barriers to isopod attach of red mangrove roots.
J. exp. mar. Biol. Ecol. 142: 91-104.
Estevez, E. D. & J. L. Simon, 1975. Systematics and ecology
of Sphaeroma (Crustacea: Isopoda) in the mangrove habi­
tats of Florida. In C. Walsh, S. Snedaker & H. Teas (eds),
Proceedings of international symposium on biology and
management of mangroves. University of Florida, Gaines­
ville, FL: 286-304.
Farnsworth, E. J. & A. M. Ellison, 1991. Patterns of her-
bivory in Belizean mangrove swamps. Biotropica 23: 555-
567.
Golley, F., H. T. Odum & R. F. Wilson, 1962. The structure
and metabolism of a Puerto Rican red mangrove forest in
May. Ecology 43: 9-19.
Hartshorn, G., L. Nicolait, L. H artshorn, G. Bevier, R.
Brightman, J. Cal, A. Cawich, W. Davidson, R. DuBois, C.
Dyer, J. Gibson, W. Hawley, J. Leonard, R. Nicolait, D.
Weyer, H. White & C. Wright, 1984. Belize country envi­
ronmental profile. Trejos H nos Sucs. S.A., San Jose, Costa
Rica, 152 pp.
Hutchings, P. A. & H. F. Recher, 1974. The fauna of Careel
Bay with comments on the ecology of mangrove and sea-
grass communities. Aust. Zool. 18: 99-128.
Kensley, B. & M. Schotte, 1989. Guide to the marine isopod
crustaceans of the Caribbean. Smithsonian Institution
Press, Washington, D.C., 308 pp.
Kjerfve, B., K. Riitzler & G. H. Kierspe, 1982. Tides at Carrie
Bow Cay, Belize. In K. Riitzler & I. G. Macintyre (eds),
The Atlantic barrier reef ecosystem at Carrie Bow Cay,
Belize, I. Structure and communities. Smithsonian Inst.
Press, Washington, D.C.: 47-51.
Littler, D. S., M. M. Littler, K. E. Bucher & J. E. Norris,
1989. Marine plants of the Caribbean: a field guide from
Florida to Brazil. Smithsonian Inst. Press, Washington,
D.C., 263 pp.
Littler, M. M., P. R. Taylor, D. S. Littler, R. H. Sims & J. N.
Norris, 1985. The distribution, abundance, and primary
productivity of submerged macrophytes in a Belize barrier-
reef mangrove system. Atoll Res. Bull. 282: 1-18.
97
Lugo, A. E. & S. C. Snedaker, 1974. The ecology of man­
groves. Ann. Rev. Ecol. Syst. 5: 39-64.
M aenae, W., 1968. A general account of the fauna and flora
of mangrove swamps and forests in the Indo-West-Pacific
region. Adv. mar. Biol. 6: 73-270.
M aenae, W. & M. Kalk, 1962. The ecology of the man­
grove swamps at Inhaca island, Mozambique. J. Ecol. 50:
19-34.
Mattox, N. T., 1949. Studies on the biology of the edible
oyster, Ostrea rhizophorae Guilding in Puerto Rico. Ecol.
Mono. 19: 339-356.
M cLean, R. A., 1951. The Pelecypoda or bivalve mollusks of
Porto Rico and the Virgin Islands. In Scientific survey of
Porto Rico and the Virgin Islands. N.Y. Acad. Sei. 17:
1-183.
Menge, B. A., J. Lubchenco, S. D. Gaines & L. R. Ashkenas,
1986. A test of the Menge-Sutherland model of community
organization in a tropical rocky intertidal food web. Oeco-
logia (Beri.) 71: 75-89.
Menge, B. A. & J. P. Sutherland, 1987. Community regula­
tion: variation in disturbance, competition, and predation
in relation to environmental stress and recruitment. Am.
Nat. 130: 730-757.
Odum, W. E. & E. J. Held, 1972. Trophic analyses of an
estuarine mangrove community. Bull. mar. Sei. 22: 671—
738.
Paine, R. T., 1966. Food web complexity and species diver­
sity. Am. N at. 100: 65-75.
Paine, R. T., 1988. Food webs: road maps of interactions or
grist for theoretical development. Ecology 69: 1648-1654.
Perry, D. M., 1988. Effects of associated fauna on growth
and productivity in the red mangrove. Ecology 69: 1064-
1075.
Pinto, L. & S. Wignarajah, 1980. Some ecological aspects of
the edible oyster Crassostrea cucullata (Born) occurring in
association with mangroves in Negombo lagoon, Sri Lanka.
Hydrobiologia 69: 11-19.
Rehm, A. & H. J. Humm, 1973. Sphaeroma terebrans', a threat
to the mangroves of southwestern Florida. Science 182:
173-174.
Ribi, G., 1982. Does the wood-boring isopod Sphaeroma
terebrans benefit red mangroves (Rhizophora manglej! Bull,
mar. Sei. 31: 925-928.
Rodriguez, C. & A. W. Stoner, 1990. The epiphyte commu­
nity of mangrove roots in a tropical estuary: distribution
and biomass. Aquat. Bot. 36: 117-126.
Riitzler, K., 1969. The mangrove community, aspects of its
structure, faunistics and ecology. In Lagunas Costeras, un
Simposio. U N A M -U N ESC O , Mexico: 515-535.
Riitzler, K., 1987. Recruitment of sponges and their space
competitors to settling plates in Twin Cay mangrove hab­
itats. In K. Rützler (ed.), Caribbean Coral Reef Ecosys­
tems: progress report 1987. National Museum of Natural
History (Dept, of Echinoderms & Lower Invertebrates),
Smithsonian Inst., Washington, D.C.: 23-24.
Riitzler, K. & I. G. Macintyre (eds), 1982. The Atlantic bar­
98
rier reef ecosystem at Carrie Bow Cay, Belize, I. Structure
and communities. Smithsonian Institution Press, Washing­
ton, D.C., 539 pp.
Sasekumar, A., 1974. Distribution of macrofauna on a M a­
layan mangrove shore. J. anim. Ecol. 43: 51-69.
Sasekumar, A., 1984. M ethods for the study of mangrove
fauna. In S. C. Snedaker & J. G. Snedaker (eds), The m an­
grove ecosystem: research methods. U N ESC O , Paris,
France: 145-161.
Simberloff, D., B. J. Brown & S. Lowrie, 1978. Isopod and
insect root borers may benefit Florida mangroves. Science
201: 630-632.
Smith, T. J. III, 1987. Seed predation in relation to tree dom ­
inance and distribution in mangrove forests. Ecology 68:
266-273.
Smith, T. J. III, H. T. Chan, C. C. M clvor & M. B. Robblee,
1989. Comparisons of seed predation in tropical tidal for­
ests from three continents. Ecology 70: 146-151.
Snedaker, S. C., 1973. Mangroves, isopods, and the ecosys­
tem. Science 183: 1036-1037.
Sokal, R. R. & F. J. Rohlf, 1981. Biometry, second edition.
W. H. Freeman & Co., New York, N.Y., 859 pp.
Sterrer, W. (ed.), 1986. M arine fauna and flora o f Bermuda.
John Wiley & Sons, New York, N.Y., 742 pp.
Stoddart, D. R., F. R. Fosberg & D. L. Spellman, 1982. Cays
of the Belize barrier reef and lagoon. Atoll Res. Bull. 256:
1-76.
Sutherland, J. P., 1980. Dynamics of the epibenthic commu­
nity on roots of the mangrove Rhizophora mangle at Bahia
de Buche, Venezuela. Mar. Biol. 58: 75-84.
Taylor, P. R., M. M. Littler & D. S. Littler, 1986. Escapes
from herbivory in relation to the structure of mangrove is­
land macroalgal communities. Oecologia (Beri.) 69: 481-
490.
Tomlinson, P. B., 1986. The botany of mangroves. Cambridge
University Press, Cambridge, U .K., 413 pp.
W ada, K. & D. Wowor, 1989. Foraging on mangrove pneu-
matophores by ocypodid crabs. J. exp. mar. Biol. Ecol. 134:
89-100.
Wahl, M., 1989. Marine epibiosis. I. Fouling and antifouling:
some basic aspects. Mar. Ecol. Prog. Ser. 58: 175-189.
W eerdt, W. H. de, K. Riitzler & K. P; Smith, 1991. The
Chalinidae (Porifera) of Twin Cays, Belize, and adjacent
waters. Proc. Biol. Soc. Wash. 104: 189-205.
Hydrobiologia 247: 99-108, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 99
© 1992 Kluwer Academic Publishers. Printed in Belgium.

The meiobenthos of five mangrove vegetation types in Gazi Bay, Kenya

S. Vanhove, M. Vincx, D. Van G ansbeke, W. Gijselinck & D. Schram

University o f Ghent, Zoology Institute, Marine Biology Section, K .L . Ledeganckstraat 35, B-9000 Gent,
Belgium

K ey words: m eiobenthos, m angroves, distribution, K enya

Abstract

T he vertical distribution o f m eiofauna in the sedim ents o f Avicennia marina, Bruguiera gymnorrhiza,
Ceriops tagal, Rhizophora mucronata and Sonneratia alba at G azi Bay (Kenya), is described. Seventeen
taxa were observed, with highest densities in the sedim ents o f Bruguiera (6707 ind. 10 c m ' 2), followed
by Rhizophora (3998 ind. 10 c m ~ 2), Avicennia (3442 ind. 10 c m - 2 ), Sonneratia (2889 ind. 10 c m - 2 ) and
Ceriops (1976 ind. 10 c m - 2 ). N em atodes accounted for up to 95% o f total densities; other com m on taxa
were copepods, turbellarians, oligochaetes, polychaetes, ostracods and rotifers. High densities occurred
to about 20 cm depth in the sedim ent. Especially Ceriops sedim ents show still high densities o f nem a­
todes (342 ind. 10 c m - 2 ) and copepods (11 ind. 10 c m - 2 ) in the deepest layer (1 5 -2 2 cm). Particle size
and oxygen conditions were m ajor factors influencing m eiobenthic distribution; Uca burrow s h ad a m ajor
im pact on distribution and abundance o f m eiofauna.

Introduction pilot study o f m eiobenthos in an east African

M ost publications on the distribution of m eio­
benthos in m angrove sedim ents focus on the
upper layers o f the sedim ent, while studies on
vertical distribution are scarce (G erlach, 1958;
D ecraem er & C oom ans, 1978; H o d d a & N icho­
las, 1985; Dye & Lasiak, 1986; L alana-R ueda &
G osselck, 1986; Alongi, 1987a, b). Only two
studies examine the m eiobenthos densities in
deeper layers (Dye, 1983a; N icholas et al., 1991).
The bulk o f m angrove w ork concerns prim ary
production and decom position o f litter (O dum ,
1988; B oaden & Seed, 1985; Lugo & Snedaker,
1974; Boto e t a l , 1984; Twilley eta l., 1986;
A ndersen & H argrave, 1984; B enner & H odson,
1985; Cundell eta l., 1979; Alongi eta l., 1989;
W oodroffe, 1987; Steinke & W ard, 1988). This
m angrove swam p, deals with vertical distribution
within the sedim ent with em phasis on the role of
sedim ent structure on m eiofauna densities.
Description of sites
Studies were carried out at G azi Bay circa 50 km
south of M om basa (Fig. 1). The mixed m angrove,
described by G allin etal. (1989) was chosen, be­
cause it is undisturbed and relatively easy to
reach. M oreover, studies on vegetation patterns
and on the ecology o f the oyster Crassostrea
cucullata o f th at area are available (G allin eta l.,
1989; O kem w a et al., 1986; R uw a & Polk, 1986;
Ruw a, 1990).
Sam ples were collected from the sedim ent
100

MOMBASA
t 1 .....
X j \ r
F IS H E R M E N 'S H A R B O U R 1 : B ru g u iera s i t e
P / I S 2 : A v ic e n n ia s i t e
' 2 3 : a: C e rio p s s i t e
yiHlIlTillb \ b: R h i z o p h o r a s i t e
a 3b 3c c: S o n n e ra tia s i t e

VILLAGE GAZI BAY

* * * A v ice n n ia m a rin a
* * * * * SCHO OL
* * * L u m n itz e ra r a c e m o s a
* *
& R h iz o p h o ra m u c ro n a ta
* * %
* % S o n n e r a tia a lb a
* * *
* * * % m ix ed m a n g a l
* * * * *
* ■è * C o c o s -p la n ta tio n
* * *
* *
F IS H E R M E N 'S H UT
* C o rn
* * ■H*
* * * * dune
* *
* % * * pool
* * * *
* ta r ro a d
dry w e a th e r ro a d
m o to ra b le tra ck
p a th

50 0 m

Fig. 1. Study area: vegetation map of the mangrove in G azi Bay (after Gallin et al., 1989).

around five m angrove species (Avicennia marina, Material and methods

Bruguiera gymnorrhiza, Ceriops tagal, Rhizophora
mucronata, Sonneratia alba) during spring tide, in
July 1989.
Meiofauna
H andcores (3.6 cm diam eter, 30 cm length) were
taken to a depth o f 20 cm and subdivided into
1 cm portions for the upper 5 cm, 2.5 and 5 cm
 101

intervals for deeper layers. Each core was ing a m ercury therm om eter into the sedim ent. S a­
preserved in a h o t (60 °C ) 4% form aldehyde linity w as m easured with a refractom eter.
solution. L aboratory treatm ent consisted of ex­
tracting the m eiofauna by a com bination o f sed­ Results
im entation, sieving (38 /rm), centrifugation, and
staining with R ose Bengal. R oots and organic Sedim ent tem perature reached 2 4 -2 5 °C and sa­
detritus were n o t separated, w hat m ade counting linity 32—35%0; a small input o f fresh w ater by the
o f the m ajor tax a very tim e consum ing. small rivers K idogoweni and M kurum u had no
m easurable influence on the m angrove. p H was
Sedim ent analysis 7.15-7.85.
Figures 2 -6 show for each station, (a) the
O ne core w as used for granulom etric analysis. m eiofauna taxa in the whole core, (b) the meio­
Prelim inary drying at 100 °C was necessary to fauna taxa, excluding nem atodes, (c) depth dis­
rem ove peat. A m echanical shaker was used to tributions o f nem atodes and copepods, (d) depth
p ass the sand fraction through a graded series of profiles o f m edian grain size and (e) o f m ud co n ­
stan d ard sieves (B uchanan, 1984). tent, (f) depth distribution o f total meiofaunal
Sedim ent tem perature w as m easured by p ush­ density.

depth (cm)
a m p h ip o d a

oo em b o a
.W W W W Tÿl

gastrotricha
<
c
kinorhyncha
n em atoda
w
cc without nematodes
oligochaeta CO
X
ostraco d a CO
polychaeta VI
rotifera \ \ l
tardigrada □ nematodes
turbellaria ]
■ copepods
larvae of insects
nauplii 3
S3
0 500 1,000 1,500 2,000 2,500 o 50 100 150 200 250 300 350 0 200 400 600 800
absolute density absolute density density
lenght of total core : 20 cm A B total density copepods : 215 C
Avicennia total density nematodes : 2329
depth (cm) D depth (cm) E depth (cm)
67.67% of total meiofauna
F

\\W \W

4-5
WWWWWWN
5-7.5

7.5-10

10-15

15-inf 15-inf 15-inf.

0 0.1 0.2 0.3 0.4 0.5 0.6 o 1 2 3 4 5 6 200 400 600 800 1,000 1,200
median (mm) % mud density
meiofauna from all depths : 3442

Fig. 2. Avicennia site (a): density of all meiofauna taxa in the whole core (ind./10 cm2), (b): density of the meiofauna, excluding
the nematodes (ind./IO cm2) (c): depth distribution patterns of densities of nematodes and copepods (d): depth profiles of median
grain size of the sand fraction (in mm) (e): depth profile o f the percentage of mud (% ) (f): depth distribution of the total densi­
ties of meiofauna.
102

d epth (cm)
a m phipoda
calanoidea 0-1
cladocera
collem bola 1-2
copepoda \ \ \ 1
g astro p o d a 2-3
gastrotricha S3
halacaroidea 1 3-4
isopoda
kinorhyncha 4-5
nem atoda without nematodes
o ligochaeta 3 5-7.5
ostraco d a
polychaeta ] 7.5-10
rotifera
ta rdigrada
10-15
H nematodes
turbellaria
larvae of insects ■ copepods
nauplii 3

1,000 2,000 3,000 4,¿00 1,000 1,200 0 200 400 600 800 1,000 1,200 1,400
absolute density absolute density density ^
A B total density copepods : 1113
lenght of total core : 15 cm Bruguiera total density nematodes : 4205
62.69 % of total meiofauna p
depth (cm) D depth (cm) E depth (cm)

0-1 0-1

1-2 1-2

2-3 2-3

3-4 3-4

4-5 4-5
4-5
5-7.5 5-7.5

7.5-10 7.5-10
5-7.5
^ \\\\\\^ ^
7.5-10
10-15 10-15

15-inf 15-inf 10-15

x^W N W N
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 500 1,000 1,500
median (mm) % mud density
meiofauna from all depths : 6707

Fig. 3. Bruguiera site (a): density of all meiofauna taxa in the whole core (ind./10 cm2), (b): density of the meiofauna, excluding
the nematodes (ind./10 cm2) (c): depth distribution patterns of densities o f nematodes and copepods (d): depth profiles of median
grain size of the sand fraction (in mm) (e): depth profile of the percentage of mud (% ) (f): depth distribution of the total densi­
ties of meiofauna.

Meiofauna Taxa
Seventeen m eiofaunal taxa are recorded. N em a­
Total densities todes dom inate: 77.8% (Avicennia), 53.5% (Bru­
T otal densities are high in all five stations: guiera), 85.5% (Ceriops), 89.0% (Rhizophora)
6707 ind. 10 c m “ 2 in Bruguiera, 3998 ind. and 95.1% (Sonneratia). C opepod densities, p ar­
10 c m “ 2 for Rhizophora, 3442 ind. 10 c m “ 2 for ticularly harpacticoids, are low in com parison
Avicennia, 2889 ind. 10 c m “ 2 in Sonneratia and with nem atodes (less than 1 % o f total density).
1976 ind. 10 c m “ 2 in Ceriops. O ther com m on taxa are turbellarians, oligo-
Except for Bruguiera, the density o f m eiofauna chaetes, polychaetes, ostracods and rotifers which
is highest near the sedim ent surface, and espe­ occur in high num bers in som e areas. Few gas-
cially in Sonneratia where the upper centim eter trotrichs, kinorhynchs and tardigrades are re­
contains 50% o f the total (Fig. 6F). In Ceriops corded.
(Fig. 4F ) and Rhizophora (Fig. 5F), total m eio­ In Figs 2 -6 (A, B) no distinction is m ade be­
fauna fluctuates with depth, and in Bruguiera tw een adults and larvae o f polychaetes, though
(Fig. 3F) a m axim um is reached at 4 -5 cm. In high num bers were found for the larvae in A vi­
Avicennia (Fig. 2F), again, higher densities occur cennia and Bruguiera'. 93% and 78% respectively.
in the top centim eter. The other sedim ents show rem arkable low er d ata
 103

depth (cm)
a m phipoda
calanoidea 0-1
clad o cera
collem bola 1-2
copepoda g
gastro p o d a 2-3
gastrotricha
halacaroidea 3-4
iso p o d a
kinorhyncha 4-5
nem ato d a without nematodes
oligochaeta 3 5-7.5
o straco d a xa
polychaeta
rotifera
S3 7.5-10 Q nematodes
tardigrada 3 10-15
■ copepods
turbellaria
larvae of Insects
\\w x
15-inf
nauplii
XX
500 1,000 1,500 0 100 200 300 400
absolute density absolute density density
lenght of total core 22.5 cm A B total density copepods : 61 C
Ceriops total density nematodes : 1709
86.49% of total meiofauna
depth (cm)
D depth (cm)
E depth (cm) F
0-1 j - j I 0-1 \\\\\\\\\\\\\\\)
1-2 , 2 -......... ‘j 1-2
XWWWWN
2-3 2-3
J 2-3 XXXXN
3-4 3-4 3-4 X \\\\\\\\\\N
4-5 4-5 4-5 X \\\\\\\\\\\\3
5-7.5 5-7.5 5-7.5 \\\\ \ \\ \\ \\ \\ \1
7.5-10 ............................................ 1 7.5-10 7.5-10 NXXXXXXI
10-15
: ...........:...... ................................................................... i 10-15 - :w
.... ........................
10-15
x \\\\\\X
lö-inf '' 1 15-inf ■ i 15-Inf.
x \ \\\\\ \ \\ \ \ \\ \ \ \ \\ \ \ 3
........................................................................ . . . .
100 200 300 4

median (mm) % mud density

meiofauna from all depths : 1976

Fig. 4. Ceriops site (a): density of all meiofauna taxa in the whole core (ind./IO cm2), (b): density of the meiofauna, excluding the
nematodes (ind./10 cm2) (c): depth distribution patterns of densities of nematodes and copepods (d): depth profiles of median grain
size of the sand fraction (in mm) (e): depth profile of the percentage o f mud (% ) (f): depth distribution of the total densities of
meiofauna.

(as an extreme 2 o f the 85 polychaetes of Sonner­ drop. F rom 2 -3 cm on, only nem atodes and
atia are larvae). copepods rem ain abundant, respectively 83 and
Although no attem pt is m ade to detail the dif­ 7% . At 7 .5 -1 0 cm copepods reach a m axim um of
ferent taxa, a lot o f epsilonem atid nem atodes are 150 ind. (abundance: 37% ), while nem atodes fol­
counted, especially in Bruguiera sedim ents where low a decreasing trend (223 ind.; 55% ). Densities
they m ake up about 70% o f nem atode num bers. o f total m eiofauna quickly drop with depth
(Fig. 2f). Nevertheless there are still 161 nem a­
Depth profiles todes in the sedim ent beneath 10 cm.

Avicennia (Figs 2 C, F) Bruguiera (Figs 3 C, F)

The upper layer (0 -1 cm ) shows a high abun­ A bundances o f nem atodes are high in com pari­
dance of nem atodes (682 ind.; 67 % o f total m eio­ son with other taxa (m axim um value o f 1241 ind.
fauna at this interval). C opepods cover only 1 % at the 4 -5 cm interval); only the first cm shows
o f total m eiofauna, with nauplii four times m ore a low percentage (3% ), mainly due to high
ab u n d an t than adults. This is in contrast w ith the am ounts o f gastrotrichs (87% ). C opepod d ata
subsequent depths where adult copepods keep are very low in the first 2 cm (6 and 9 ind.), but
alm ost the sam e num bers, but num bers o f larvae increase deeper in the sedim ent, reaching a max-
104

depth (cm)
am phipoda
calanoidea
cladocera
collembola SXXXXXXX3
co p e p o d a
g astro p o d a
gastrotricha
halacaroidea <
isopoda c
kinorhyncha
co
nem atoda cc without nematodes
o ligochaeta CC XVÑ]_______
polychaeta
rotifera
tardigrada H nematodes
turbellaria
larvae of insects H copepods
nauplii 15-inf.

1,000 2,000 3,000 o 20 40 60 80 100 120 140 o 200 400 600 800
absolute density absolute density density
lenght of total core : 17 cm A total density copepods : 83 C
Rhizophora total density nematods : 3500

depth (cm)
D depth (cm) depth (cm)
89.27 % of total meiofauna
F
0-1 - ..................I 0-1 : .......... j 0-1
wwvww wwn
1-2 I 1-2 I 1-2 \W \W \N
2-3 2-3 : _________________________ _ J 2-3 ^ x \\\\\\\%
3-4 3-4 3-4
N \\\\\\\\\\\\\\\N
4-5 4-5 4-5
wwwwwwx
5-7.5 5-7.5 5-7.5
x \\\\\\\\\\\b
7.5-10 7.5-10 7.5-10
1_________________________ : ......- . . . . - . . V J
10-15
........... I
..............................................................................
10-15 10-15
\\W \N
15-Inf
------1------L .. 1 .1 I 1 : 1 , 1 ,
15-inf
i : i !. i i
15-inf.
3
0.15 0.2 0.25 0-35 0 1 2 3 4 5 6 0 200 400 600 800 1,000
median (mm) % mud density
meiofauna from all depths : 3998

Fig. 5. Rhizophora site (a): density of all meiofauna taxa in the whole core (ind./10 cm2), (b): density of the meiofauna, exclud­
ing the nematodes (ind./IO cm2) (c): depth distribution patterns of densities of nematodes and copepods (d): depth profiles of
median grain size of the send fraction (in mm) (e): depth profile of the percentage of mud (% ) (f): depth distribution of the total
densities of meiofauna.

imum of 325 ind. at 3 -4 cm. N em atodes and Rhizophora (Figs 5 C, F)

copepods are still present in high num bers in the T he general aspect o f high nem atode abundances
deepest layer (342 and 11 ind.), though other taxa is also present (m ean value: 89% ). C opepods are
(oligochaetes, polychaetes and rotifers) also re­ only ab u ndant in the first sedim ent layer (47 adults
m ain present and account for 17% at this depth and 20 nauplii), but this num ber is low in com ­
where a total o f 759 individuals is reported. parison with the 371 nem atodes at this depth. The
rem aining 32% o f the total m eiofaunal density is
Ceriops (Fig. 4 C, F) mainly m ade up by ostracods (19% ). In sedim ent
The entire core is characterized by lower num ­ layers deeper than 2 cm, nem atodes constitute
bers. N em atodes dom inate and are m ost frequent m ore than 90% o f the total m eiofauna in each
in the upperm ost and deepest core layers; 342 interval, and even 98% at 4 -5 cm. In the deep­
n em atodes are still present at a depth of est layer total density drops a m inim um o f 37
15-22.5 cm. C opepods only reach 4% o f nem a­ individuals.
tode num bers. Sam pling deeper than the present
depth w ould give a better idea o f how density Sonneratia (Figs 6 C, F)
evolves. A typical profile o f total density, w hich is again
similar to the figure o f nem atode distribution,
 105

depth (cm)
a m phipoda
calanoidea
cladocera
collem bola
copepoda \_\X_3
gastrotricha
halacaroidea
isopoda
kinorhyncha
n em atoda without nematodes
oligochaeta
o straco d a 3
polychaeta 7.5-10
rotifera
tardigrada S nematodes
turbellaria X]
larvae of insects copepods
nauplii :
1,000 2,000 3,000 100 150 200 400 600 800 1,000 1,200
500 1,500 2,500 absolute density density
absolute density
A B total density copepods : 51
lenght of total core : + 15cm Sonneratia total density nematodes : 2555

depth (cm) D depth (mm) E depth (cm)

88.44 % of total meiofauna p

0-1 0-1
XW W W W W W W W W ^
1-2 1-2

2-3 2-3

3-4 3-4 3-4

4-5 4-5 4-5

5-7.5 5-7.5 5-7.5

.5-10 .5-10 7.5-10

10-15 10-15 10-15

15-inf 15-inf 15-inf.

0 0.1 0.2 0.3 0.4 0.5 0.6 0 2 3 4 5 6 400 800 1,200 1,6C
median (mm) >mud 600 1,000 1,400
density
meiofauna from all dephts : 2889

Fig. 6. Sonneratia site (a): density of all meiofauna taxa in the whole core (ind./10 cm2), (b): density of the meiofauna, excluding
the nematodes (ind./10 cm2) (c): depth distribution patterns of densities of nematodes and copepods (d): depth profiles of median
grain size of the sand fraction (in mm) (e): depth profile o f the percentage of mud (% ) (f): depth distribution of the total densi­
ties of meiofauna.

shows high densities in the upper layer (1443 ind.; Sedim ent
50% o f total density), but a very quick drop
deeper in the sedim ent. C opepods only occur in Figures 2 - 6 (D , E) show the vertical profile of
the first layer (48 ind.). m edian sedim ent grain size and m ud content. The
m ean m edian particle size o f each core fluctuates
In conclusion. Although densities tended to be betw een 0.266 m m (Rhizophora) and 0.468 mm
higher in the upper layers, several taxa were dis­ (Avicennia); consequently all sedim ents are clas­
tributed in high densities in deeper sedim ent in­ sified as m edium sand. A co n stan t m edian grain
tervals as well. Regarding the deepest interval size vertical profile is shown for Avicennia sedi­
Sonneratia has the low est am ount o f individuals m ents (except for the upperm ost and deepest in­
(2% ), followed by Avicennia (3.2% ), Rhizophora tervals), Bruguiera and Ceriops', Rhizophora and
(6.5% ), Ceriops (8 % ) and Bruguiera (11% ). Both, Sonneratia have an increasing trend in m edian
nem atodes and copepods live at sedim ent depths grain size o f the sand fraction with depth. The
deeper than the one examined. Especially Ceriops m ud percentage o f Avicennia and Sonneratia is
show s a high nem atode and copepod num ber in higher in the top centim eter in com parison with
the deepest layer (15-22.5 cm: 342 nem atodes, 11 deeper intervals. U nlike the stable m edian grain
copepods). size, m ud content show s larger variation, with
106
highest values at 0 -1 cm and 15-inf. cm (Sonne­
ratia). H igher m ud percentages are also reported
in the deepest interval o f Ceriops. A clear increase
up to a depth o f 7 .5 -1 0 cm (with 5.41 % m ud), is
followed by a decrease o f the m ud content in
deeper layers for Rhizophora m ud content.
Distribution o f meiofauna with regard to sediment
type
The taxon com position o f the m eiofauna in the
different vegetation sites is m ostly related with
sedim ent grain size, and probably to som e extent
w ith oxygen conditions. Sedim ents with higher
m ud content are m ainly inhabited by burrow ers
such as oligochaetes, nem atodes and turbellari-
ans. T he sedim ent o f Bruguiera contains a low
m ud percentage, and consequently shows a higher
diversity o f taxa. Rhizophora and Sonneratia-
sam ples, with finer sedim ents, are mainly com ­
posed of nem atodes. Small m eiofauna such as
rotifers, gastrotrichs, nauplii and larvae o f poly­
chaetes are not linked with sedim ent character­
istics. A pparently the activity o f currents is res­
ponsible for the distribution o f these anim als.
C opepods and epsilonem atid nem atodes do not
w ithstand high silt fractions.
W ith the exception o f Bruguiera no evident re­
lationship with depth distribution is detected. The
decreasing density w ith depth in Sonneratia is
probably governed by low oxygen conditions.
Discussion
Sheltered regions with m uddy sedim ents rich in
detritus generally are characterised by low diver­
sity, b u t high densities (H eip eta l., 1985; Coull,
1988; Giere eta l., 1988). In the present study
m ud percentages reached 5 % and total densities
up to 6707 ind. are recorded, though a high num ­
ber o f taxa (17) is found. Table 1 gives a sum m ary
o f densities recorded in several studies on the
m eiobenthos o f m angroves (range is between 36
and 12000 ind. 10 c m - 2 ); values o f G azi are in
between the recorded values.
N em atodes were also the m ost ab u ndant taxon
Table 1. D e n s itie s o f m e io fa u n a in m a n g ro v e s (in d ./IO c m 2).
Density Author
S. Cuba 36-245 Lalana-R ueda & Gosselck, 1980
S.E. Cuba 341 H errera & del Valle, 1980
N.E. Australia 57-2,454 Alongi, 1987 (a + b)
S.E. Australia 64-12,058 H odda & Nicholas, 1985
S. Africa 840-5,300 Dye, 1983 (b)
E. Africa 1,976-6,707 This study
in the surveys o f A ustralian m angroves (H o d d a
and N icholas, 1985; N icholas e ta l., 1991) and
South African m angroves (Dye, 1983b). N em a­
todes seem to be less im portant in C uban m an­
groves (L alana-R ueda & G osselck, 1986: 35.3
and 61.1% ).
Alongi (1987a, b), H o d d a & N icholas (1985)
and Dye (1983a, b) noted a low copepod density
(respectively 1-22 ind. 10 c m - 2 , under 10% and
3 0 -8 0 ind. 10 c m - 2 ). L alana-R ueda & G osselck
(1986) in contrast found the copepods to be the
second m ost com m on taxon in term s o f indivi­
dual dom inance.
U nlike m ost other studies, turbellarians were
the m ost abundant m eiobenthic group in the d ata
o f Alongi (1987a), probably due to the used ex­
traction technique. H o d d a & N icholas (1985)
only found the oligochaetes as abundant group
next to nem atodes and copepods, while Dye
(1983a, b) recorded 6.4% ciliates, followed by oli­
gochaetes (4.5% ), turbellarians (3% ) and ki-
norchychs (2.7% ). C opepods, polychaetes, gas­
trotrichs, larvae o f crustaceans and insects only
appeared sporadically. The five m ost im portant
taxa in the T oledo Lagoon were nem atodes,
copepods, T anaidacea, oligochaetes and poly­
chaetes and in the B asto Lagoon nem atodes,
copepods, ostracods, am phipods and poly­
chaetes (L alana-R ueda & G osselck, 1986).
Am ong the different sam pling sites, Rhizophora
and Sonneratia sedim ents contained the highest
abundance (respectively 89 and 95% ) o f nem a­
todes, but also h ad the highest m ud content (re­
spectively 4 and 3% ). Low copepod percentages,
both in G azi and the study areas cited above,
suggest th at this taxon is m ore related to coarser
grain texture, and th at nem atodes show a broad

tolerance to diverse sedim ents. Interstitial groups,
e.g., G astrotricha, T ardigrada are typically ex­
cluded from the m uddy sedim ent. Sedim ents with
a m edian particle size smaller than 125 mm are
usually dom inated by burrow ers (Coull, 1988).
In m uddy sedim ents, alm ost no life persist
under a depth o f 8 cm (M cIntyre & W arwick,
1984). T he present study, however, shows that
high densities, especially o f nem atodes, occur be­
neath this level.
In a study o f vertical distribution o f m eiofauna
in m angrove sedim ents in Transkei, Dye (1983a)
found also th at m eiofauna occurs to considerable
depths, although highest densities occur in the
upper 10 cm o f the sedim ent. H e calculated that
when one restricts m eiofauna sam pling to the top
5 cm o f the sedim ent only 40% o f the total den­
sity can be obtained.
A recent study by N icholas etal. (1991) co n ­
firms those findings: about 85% o f the nem atodes
occur in the top cm o f soft m ud, but 5 -7 species
inhabit the deeper anoxic m ud dow n to 10 cm.
Low oxygen conditions are supposed to be an
im portant regulator for the low densities versus
depth o f Sonneratia. Coull (1988) asserts th at ver­
tical distribution patterns typically are controlled
by the depth o f the redox potential discontinuity
(R P D ) level, i.e., the boundary between aerobic
and anaerobic sedim ents, which is prim ary deter­
m ined by the ‘super variable’ oxygen. In this con­
text W ieser et aí. (1974) claim harpacticoid cope­
pods are the m ost sensitive m eiobenthic taxon to
decreased oxygen (this in com bination with the
previously discussed sedim ent factor). N em a­
todes seem to be m ore tolerant to reducing cir­
cum stances (B oaden, 1975 in H o d d a & N icholas,
1985).
Fiddler crab Uca spp. are a prevalent aspect of
the K enyan m angroves (personal observation of
the authors). D eP atra & Levin (1989) found a
fundam ental influence o f the Uca burrow s on the
abundance and distribution o f several m eiofaunal
tax a in a N orth C arolina salt m arsh. B oth by
passive deposition and enhancem ent o f the micro
environm ent, e.g., food resources, stagnation o f
w ater (which provides suitable habitats for the
developm ent o f young stages o f several taxa),
107
densities are higher in these ‘biogenetic depres­
sions’. On the other hand abundances tend to be
lower due to the accum ulation o f toxic com po­
nents and the fouraging activities o f the fiddler
crabs themselves. A study on their influence on
energy flow in the m angrove ecosystem is no t yet
analysed, but a similar phenom enon is present in
the m angroves o f northeastern A ustralia, where
the sesarm id crab Sesarma messa has a high im­
pact on the weight loss from leaves by their leaf
burying and consum ing activities (R obertson,
1986), and in the m angroves o f Q ueensland,
where the crab Sesarma erythrodactyla enhances
the decom position o f m angrove litter (Camilleri,
1989).
This pilot study o f m eiobenthic com position in
the m angroves o f G azi Bay will be followed by a
m ore intensive study investigating the quantity (in
the light of tidal regime and seasons) and quality
(identification o f the m ost im portant taxa and
trophic groups) o f som e ecological concepts like
predation, consum ption and interactions between
m acrofauna and m eiobenthos.
Acknowledgements
T he authors acknowledge the K enyan M arine
Fisheries and R esearch Institute in M om basa
(K enya) for providing logistic facilities for the field
w ork. D r A. Vanreusel and D r P. A erts are
thanked for the sam pling assistance. This study
is financed by the Belgian N ational Science
F oundation (F K F O 2.0043.88) and the EEC
project TS2-0240-C (G D F ).
References
Alongi, D. M., 1987a. Intertidal zonation and seasonality of
meiobenthos in tropical mangrove estuaries. Mar. Biol. 95:
447-458.
Alongi, D. M., 1987b. Inter-estuary variations and intertidal
zonation of free-living nematode communities in tropical
mangrove systems. Mar. Ecol. Prog. Ser. 40: 103-114.
Alongi, D. M., K. G. Boto & F. Tirendi, 1989. Effects of
exported mangrove litter on bacterial productivity and dis­
solved organic carbon fluxes in adjacent tropical nearshore
sediments. Mar. Ecol. Prog. Ser. 56: 133-144.
108
Andersen, F. O. & B. T. Hargrave, 1984. Effects o f Spartina
detritus enrichment on aerobic/anaerobic benthic metabo­
lism in an intertidal sediment. Mar. Ecol. Prog. Ser. 16:
161-171.
Benner, R. & R. E. Hodson, 1985. Microbial degradation of
the leachable and lignocellulosic components of leaves and
wood from Rhizophora mangle in a tropical mangrove
swamp. Mar. Ecol. Prog. Ser. 23: 221-230.
Boaden, P. J. S., 1975. Anaerobiosis, meiofauna and early
metazoan evolution. Zool. Ser. 4: 21-4.
Boaden, P. J. S. & R. Seed, 1985. An introduction to Coastal
Ecology, Chapter 7, Blackie USA: Chapman & Hall, N.Y.,
218 pp.
Boto, K. G., J. S. Bunt & J. T. Wellington, 1984. Variations
in mangrove forest productivity in northern Australia and
Papua New Guinea. Estuar. coast. Shelf Sei. 19: 321-329.
Buchanan, J. B., 1984. Sediment analysis. In N. A. Holme &
McIntyre (eds), M ethods for the Study of Marine Benthos.
IBP No 16, 2nd edn. Blackwell Scientific Publications,
387 pp.
Camilleri, J., 1989. Leaf choice by crustaceans in a mangrove
forest in Queensland. Mar. Biol. 102: 453-459.
Coull, B. C., 1988. Ecology of the marine meiofauna. In R. P.
Higgins & H. Thiel (eds), Introduction to the study of meio­
fauna: 18-38.
Cundell, A. M., M. S. Brown, R. Stanford & R. Mitchell,
1979. Microbial degradation of Rhizophora mangle leaves
immersed in the sea. Estuar. coast, mar. Sei. 9: 281-286.
Decraemer, W. & A. Coomans, 1978. Scientific report on the
belgian expedition to the G reat Barrier Reef in 1967. Nem­
atodes X II. Ecological Notes on the Nematode Fauna in
and around Lizard Island. Aust. J. mar. Freshwat. Res. 29:
497-508.
DePatra, K. D. & L. A. Levin, 1989. Evidence of the passive
deposition of meiofauna into fiddler crab burrows. J. exp.
mar. Biol. Ecol. 125: 173-192.
Dye, A. H., 1983a. Vertical and horizontal distribution of
meiofauna in mangrove sediments in Transkei, Southern
Africa. Estuar. coast. Shelf Sei. 16: 591-598.
Dye, A. H., 1983b. Composition and seasonal fluctuations on
meiofauna in a southern african mangrove estuary. Mar.
Biol. 73: 165-177.
Dye, A. H. & T. A. Lasiak, 1986. Microbenthos, meiobenthos
and fiddler crabs: trophic interactions in a typical mangrove
sediment. Mar. Ecol. Prog. Ser. 32: 259-264.
Gallin, E., E. Coppejans & H. Beeckman, 1989. The man­
grove vegetation o f Gazi Bay (Kenya). Bull. Soc. r. Bot.
Belg. 122: 197-207.
Gerlach, S. A., 1958. Die Mangroveregion tropischer Küsten
als Lebensraum. Z. Morph. Ökol. Tiere 46: 636-730.
Giere, O., A. Eleftheriou & D. J. Murison, 1988. Abiotic fac­
tors. In R. P. Higgins & H. Thiel (eds), Introduction to the
Study of Meiofauna: 61-78.
Heip, C., M. Vincx & G. Vranken, 1985. The ecology of
marine nematodes. Oceanogr. Mar. Biol. ann. Rev. 23:
399-489.
Herrera, A. & R. del Valle, 1980. Características de la meio­
fauna bentösica en la laguna y zona costera de Guayabal
en relación con el grado de contaminación. Ciencias Bio­
lógicas 5: 29-45.
H odda, M. & W. L. Nicholas, 1985. Meiofauna associated
with mangroves in the Hunter River Estuary and Fullerton
Cove, south-eastern Australia. Aust. J. mar. Freshwat. Res.
36: 41-50.
Lalana-Rueda, R. & F. Gosselck, 1986. Investigations of the
Benthos of Mangrove Coastal Lagoons in Southern Cuba.
Int. Revue ges. Hydrobiol. 71: 779-794.
Lugo, E. L. & S. C. Snedaker, 1974. The ecology of m an­
groves. Annu. Rev. Ecol. Syst. 5: 39-64.
McIntyre, A. D. & R. M. Warwick, 1984. Meiofauna tech­
niques. In N. A. Holme & McIntyre (eds), M ethods for the
study of marine benthos, Blackwell Scientific Publications:
217-244.
Nicholas, W. L., J. A. Elek, A. C. Stewart & T. G. Marples,
1991. The nematode fauna of a temperate Australian man­
grove mudflat; its population density, diversity and distri­
bution. Hydrobiologia 209: 13-27.
Odum, W. E., 1988. Comparative ecology on tidal fresh water
and salt marshes. Annu. Rev. Ecol. Syst. 19: 147-176.
Okemwa, E. N., R. K. R uw a& P. Polk, 1986. The autecology
of the edible oyster Crassostrea cucullata Born, 1778: size
related vertical distribution at Mkomani, M ombasa. Kenya
J. Sei. B 7: 9-14.
Robertson, A. I., 1986. Leaf-burying crabs: their influence on
energy flow and export from mixed mangrove forests
(Rhizophora spp.) in northeastern Australia. J. exp. mar.
Biol. Ecol. 102: 237-248.
Ruwa, R. K., 1990. The effects of habitat complexities created
by mangroves on macrofaunal composition in brackishwa-
ter intertidal zones at the Kenyan coast. Discovery and
Innovation 2: 49-55.
Ruwa, R. K. & P. Polk, 1986. Short communication. Addi­
tional information on mangrove distribution in Kenya: some
observations and remarks. Kenya J. Sei. B 7: 41-45.
Steinke, T. D. & C. J. Ward, 1988. Litter production by man­
groves. II. St. Lucia and Richards Bay. S. Afr. J. Bot. 54:
445-454.
Twilley, R. R., A. E. Snedaker & C. Pattersen-Zucca, 1986.
Litter production and turnover in basin mangrove forests in
southwest Florida. Ecology 67: 670-683.
Wieser, W., J. Ott, F. Schiemer & E. Gnaiger, 1974. An
ecophysiological study of some meiofauna inhabiting a
sandy beach at Bermuda. Mar. Biol. 26: 235-249.
Woodroffe, C. D., 1987. Pacific Island mangroves: distribu­
tion and environmental settings. Pacific Science 41: 1-4.
Hydrobiologia 2 4 7 : 109-117, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 109
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Ecomorphology o f C r a s s o s t r e a c u c u lla ta (Born, 1778) (Ostreidae) in a

mangrove creek (Gazi, Kenya)

J. F. T ac k ,1 E. V anden B erghe2 & Ph. P o lk 1

1Free University o f Brussels, Laboratory o f Ecology and Systematics, Pleinlaan 2, 1050 Brussels 5,
Belgium; 2 University o f Nairobi, Dept, o f Zoology, P.O. B ox 30197, Nairobi, Kenya

K ey words: Crassostrea cucullata, Fourier analysis, shell form, shell length, regression analysis,
environm ental factors

A b s tra c t

V ariations o f shell form and shell length were studied for oysters growing in the m angroves o f G azi Creek,
Kenya, and related to different environm ental factors.
F o r the study o f the form, Fourier analysis w as perform ed on the circum ference o f 85 oysters. The
resulting coefficients were com pared am ong specim ens using cluster analysis. The correspondence b e­
tween this classification and substrate diam eter is virtually perfect (only one m isclassification out o f 85
oysters). F o r the clusters based on height above chart datum , 9 specim ens, all on interm ediate height
levels, were misclassified. O rientation with respect to tidal current h ad 13 m isclassifications. M angrove
species seemed to influence form only m arginally, if at all.
In the study o f the size o f Crassostrea cucullata, the length o f 956 oysters, growing along tw o transects
were m easured, and correlated with several environm ental factors. O yster length w as no t related to
substrate diam eter or its orientation with respect to the m ain current. Length w as no t influenced by
density up to a cover o f 70% . F o r densities higher than 70% , there w as a fairly strong negative corre­
lation (r2 = 0.634, n = 217). Length w as no t correlated with height above bottom (base o f the tree) for
heights lower th an 20 cm, while oysters growing closer to the bottom were smaller. The correlation with
height above ch art datum w as negative but very low (r2 = 0.060, n = 957). How ever, if all m easurem ents
o f oysters closer than 20 cm to the bottom , and all from a density o f m ore than 70% cover are deleted
from the d ata set, the correlation with height increased dram atically, the slope still being negative
(r2 = 0.859, n = 543).

In tro d u c tio n w estern part o f the Indian O cean. Ecological

Crassostrea cucullata (Born, 1778) is an edible
oyster found in the upper littoral zone, following
the term inology o f Lewis (1964) and H artnoll
(1976). It is found on the trunks and the stilt roots
o f m angrove plants and rocky substrata in b rack ­
ish m arine environm ents. D ay (1974) noted the
zoogeographical distribution o f this species in the
studies o f this species have been carried out over
the entire region: Seychelles (Taylor, 1968), Al­
d ab ra (Taylor, 1970), T an zan ia (H artnoll, 1976),
Som alia (C helazzi & Vannini, 1980) and K enya
(Ruw a, 1984; O kem w a eta l., 1986).
Previous research already suggested a relation­
ship betw een the form o f this oyster and the sub­
strate. K esteven (1941) stated th at the form o f the
110
oyster is partly determ ined by the form o f the
substrate. Van Som eren & W hitehead (1961) dis­
tinguished different categories o f form o f oysters,
and correlated these form s with different types of
habitats, and w ith different types o f substrate.
These observations were however, qualitative.
The first quantitative w ork w as done by the above
authors. They com pared oysters growing on dif­
ferent artificial substrates by Fourier analysis.
Their conclusion w as th at the diam eter o f the
substrate has an influence on the form o f the shell
o f Crassostrea cucullata.
V arious studies show intraspecific differences
in size of m olluscs in relation to environm ental
param eters (Com fort, 1946; Ebling e ta l., 1962;
Vermeij, 1972, 1973; R uw a & Brakel, 1981).
O kem w a eta l. (1986) showed th at such differ­
ences also exist in the m angrove oyster
Crassostrea cucullata, growing on a rocky substra­
tum . The present investigation w as undertaken to
study the effects o f different environm ental fac­
tors on the size o f this oyster growing in a m an ­
grove ecosystem . The results o f this study m ay be
o f im m ediate im portance to the culture of
Crassostrea cucullata.
M a t e r i a l a n d m e th o d s
O bservations were m ade between July and Sep­
tem ber 1989 in a small creek near G azi, a small
village 49 km south o f M om basa (Kenya). Seven
species o f m angrove trees occur at G azi (G allin
eta l., 1989): Avicennia marina (F orsk.) Vierh.,
Bruguiera gymnorrhiza (L.) Lam ., Ceriops tagal
(Perrotet) C. B. R obinson, Lumnitzera racemosa
W illd., Rhizophora mucronata Lam ., Sonneratia
alba J. Smith and Xylocarpus granatum Koenig.
A n elaborate description o f this creek can be
found in T ack (1990).
The environmental factors
O rientation with respect to the m ain tidal current
w as tak en as either perpendicular or parallel. P er­
pendicular w as further subdivided into facing the
inflowing current or facing the outflowing current.
T he diam eter o f the bran ch serving as a substrate
w as m easured w ith calipers to the n earest mm.
T he height above bottom w as m easured to the
nearest cm.
To m easure approxim ate density, we m ade ver­
tical transects on each o f the 10 cm vertical bands.
These transects were m ade in four directions (fac­
ing N orth, E ast, South and W est). T he approxi­
m ate density w as expressed as the percentage of
the vertical transect th at w as covered by oysters.
H eight above chart datum w as determ ined fol­
lowing the m ethod o f O kem w a eta l. (1986). In
this m ethod, the highest level the w ater reaches
during the tidal cycle is m arked. The height o f this
level is taken to be the sam e as the one given in
the tide table for the sam e tidal cycle, and used
as an absolute reference for the rest o f the m ea­
surem ents. F rom this reference, height above
chart datum w as determ ined to the nearest cm.
Shell form
Sampling
85 oysters were collected in a restricted area of
10 X 10 m. This area w as chosen for the availabil­
ity o f all environm ental conditions of interest.
‘C ollection’ w as done by m aking a photograph of
the oyster. These photographs were taken so th at
the oyster filled only l¡9 o f the surface o f the image,
to avoid distortion.
The sam ple taken w as stratified with respect to
environm ental param eters. Only oysters th at were
growing isolated (i.e. not touching any other oys­
ter) were included in the sample. In this way,
crow ding did n o t interfere w ith the analysis: all
oysters included in the sam ple h ad the o p p o rtu ­
nity to grow to their full size.
The environm ental param eters studied were
(1) species o f the m angrove tree serving as a sub­
strate, (2) diam eter o f the branch or pneum ato-
phore on which the oyster w as growing, (3) height
above chart datum and (4) orientation w ith re­
spect to the m ain tidal current.
F our species o f m angrove trees were present in
the sam pling site. T he num ber o f oysters taken

from these species were m ore or less equal: 21
from A . marina, 22 from B. gymnorrhiza, 23 from
R. mucronata and 19 from S. alba. 28 oysters
were growing parallel with the current, 29 were
facing the inflowing current, 28 the outflowing
current.
The diam eter o f the substrate, a continuous
varíate, w as dichotom ised by taking oysters only
growing either on small branches (less than
30 m m diam eter) or large branches (m ore than
100 mm diam eter). The sam ple included 58 oys­
ters o f small branches, and 27 oysters o f large
branches.
H eight above chart datum varied between 2.20
and 3.40 m; all heights are represented in the
sam ple (Fig. 1). It was not possible to dichoto­
m ise this varíate, as was done for the diam eter o f
the substrate. N o t enough oysters were present
w ith the right com bination of environm ental p a ­
ram eters.
The photographs o f the oysters were magnified
10 times, and redraw n on paper. These drawings
were digitised using a C om plot series 7000 digi­
tiser, coupled to an IB M AT. The resolution of
the digitiser w as 0.025 mm; points were taken at
intervals o f 0.8 m m around the perim eter. The
average length o f the digitised circum ference was
about 1 m.
3-
2.2 2.6 3.0 3.4
Height above datum ( m )
Fig. 1.
Ill
Fourier analysis
T he form o f closed contours can be quantified
using Fourier analysis, by considering the polar
coordinates o f the points o f the contour, with the
origin o f the polar coordinates chosen within the
contour (e.g. A nstey & Delm et, 1972, 1973;
C hristophers & W aters, 1974; Ehrlich & W ein­
berg, 1970; Healy-W illiams & Williams, 1981;
K aesler & W aters, 1972; Y ounker & Ehrlich,
1977). This m ethod w as also used to describe the
form o f oysters by the authors.
A lim itation of the m ethod involving polar co ­
ordinates is th at only those form s where there is
only one intersection between the boundary and
a ray from the origin o f the polar coordinates can
be analysed. The form o f m ost o f the oysters is
quite irregular, and does no t satisfy this condi­
tion. H ence another m ethod, the elliptical m ethod,
has been used in this study. F o r this m ethod, both
X and y coordinates o f the boundary are inter­
preted as a function o f distance along the bound­
ary. The tw o resulting functions are periodical,
with a period given by the length o f the boundary,
and can be separately analysed using the Fourier
transform (F erson eta l., 1985; Kuhl & G iardini,
1982).
The analysis w as perform ed on a IB M AT.
Before perform ing the transform ation, a total of
64 equally spaced points are interpolated between
the digitised points. F or each o f the forms to be
analysed, 32 harm onics are calculated. Prelimi­
nary tests showed that, with this num ber o f h ar­
m onics, all features o f the form o f the circum fer­
ence could be described accurately.
Cluster analysis
The coefficients resulting from the Fourier tra n s­
form were entered in a cluster analysis, to detect
groups o f oysters o f similar form.
W ard’s m ethod (W ard, 1963; W ard & H ook,
1963) w as used for clustering. In this m ethod, the
distance m easure is the squared Euclidean dis­
tance. The m ethod o f aggregation is a m odifica­
tion o f the weighted average m ethod; the disper­
sion o f the elem ents o f a cluster around its centre
of gravity is taken into account w hen determ ining
which clusters are to be com bined.
112
T he cluster analysis w as perform ed on a M a ­
cintosh com puter, using the program S P S S /P C .
The distances betw een different groups are given
in arbitrary units. Since only relative distances
are im p o rtan t here, this does n o t interfere w ith the
interpretation o f the results.
Shell length
O bservations were m ade on C. cucullata growing
on the trunks and stilt roots o f m angrove trees
covering tw o areas (4 x 20 m ) on opposite sides
o f the m ain channel. 20 trees were growing in our
study area: 12 R . mucronata, 2 A . marina and 6
S. alba.
E ach tru n k and stilt ro o t w as divided into co n ­
secutive 10 cm vertical bands starting from the
b ottom o f the tree trunk. In each vertical ban d we
collected, if present, an oyster growing on the
n orthern, the eastern, the southern and the w est­
ern side o f the tru n k or stilt root.
The shell length o f all collected oysters was
m easured with calipers to the nearest 0.1m m .
Length w as defined as the distance from the cen­
tre o f the hinge to the farthest point o f the o p p o ­
site end. F o r each o f the m easurem ents, a num ­
ber o f environm ental variables were recorded:
(1) species o f the m angrove tree and diam eter of
the bran ch serving as a substrate, (2) approxi­
m ate density o f the oysters, (3) height above the
bottom (base o f the tree trunk), (4) height above
ch art datum , and (5) orientation with respect to
the m ain tidal current.
A regression analysis w as perform ed to try to
correlate length with each o f these variables sep­
arately. These variables were taken as the m ain
factors to study in relation to oyster length.
Results and discussion
Shell form
T he dendrogram resulting from the cluster anal­
ysis is shown in Fig. 2. In this figure, each o f the
oysters is represented by a four-letter code, indi­
cating the environm ental conditions in w hich the
oyster w as found, as explained in the caption to
this figure. O ysters less than one unit distance
apart are displayed on one line.
In this dendrogram , a first division o f the sam ­
ple into tw o groups o f m ore or less equal size
occurred on the distance level 25. These two
groups coincided with oysters th at occurred high
above datum or low above datum . O ysters grow ­
ing in interm ediate levels (betw een 3.05 and
3.13 m above datum ) occurred in bo th groups of
the cluster analysis. If the limit for discrim inating
‘H igh’ and ‘L ow ’ is set at 3.07 for oysters in the
field, 9 m isclassifications occurred (i.e. oysters
th at w ere actually growing low in the tidal zone,
but were classified w ith the oysters growing high
in the tidal zone by the cluster analysis). A con­
tingency table w as constructed, where one o f the
nom inal variables is the actual characteristic of
the specim en, and the other the predicted char­
acteristic based on the m em bership to a group in
the cluster analysis. This procedure perm itted the
calculation o f a Chi square value, indicating the
significance o f the association betw een observed
and predicted characteristics. T he Chi square
value and its probability are given in Table 1.
T able 1 also gives the value for L am bda,
G oodm an and K ruskal’s m easure o f predictive
ability (Everitt, 1977). This coefficient is a m ea­
sure o f association betw een tw o nom inal varia­
bles, and ranges from 0 to 1 (0 indicating no
association, 1 a perfect association). T he signifi­
cance o f L am bda is the sam e as the significance
o f the contingency table on w hich it is based.
T ablet. Statistics of the number of misclassifications. The
first three tests are based on a 2 x 2 table of predicted and
actual group membership. The last two lines are based on
respectively a 16 x 4 and a 16 x 2 table of frequency of occur­
rence of the different values for the tree species in the 16
different groups of Fig. 2.
Environmental parameter Chi square P Lambda
Height above datum 49.69 <0.001 0.84
Substrate diameter 75.99 <0.001 0.96
Orientation 31.69 <0.001 0.49
Mangrove species 64.6 <0.029 0.21
Avicennia or not 34.0 <0.003 0.11
 113

Gr oup S p e c i me n Co d e s Distances
Character is t ic
O 5 IO 15 20 25

zs 1 z s i r zs 1r zs 1r zs Ha Bs 1b zs !s zs 1b —\
as 1 bs i fo fos 1s a s 1r as i § s s Hs bs !b Zs 1r — \
ZS 1 zs i b zs i s As 1s zs l a zs 1s z s 1b - / i — -—\
as 1 a s 1r bs 1r Zs l a bs i r bsHs a s 1b a s Hr as Sb s sla —-/ ¡
t! ■- \
\
a 11 a 1Hs bi I r bl Ir - \ I 1
a 1i b! i b a l Ir 4- . -- / !
z!1 i.- 1
! 1s z 1! s - 1 1
a i1 a l l a Z 1 l a a l l a fol 1b b! l a - / {
i
a Ih fol hr Z 1ha bl ha b! ha a 1ha - \ 1
i îh z 1ha z ! h s Bl hs - +-------- — \ i
alh a ! hr fol hb a 1La a 1ha Z 1La a 1ha - / 1 i
! ---------- /
zsh Zs hs a s hr - \ }
ash a s b b a s L b aLhs Z s h r bs h b fos h b 3 S hr b S Lb --------- y ¡
zsh z s h s Bs hr z s h r ZS hr z s h b — \ ! - ■ ------------ ----- /
ash bshs bs ha b s h a a s h s ashfo Z s h r a 3 hr a s hb fos h b D S hb bshs /
ash Z s ha a s h b a s hr fosha _ //

1st Character 2nd C h a r a c t e r 3rd Ch ara ct er 4th Character

Cor i e n t â t i on) {diameter) (he ight) (tree species)

Parallel' s ; s mal l I ow a Av i c e n n i a
Perpendicular I : large ■i¡ oh b Br u gu i e r a
P e r p e n d f cu 1a r Rh i z o p h o r a
Sonneratia
Fig. 2. Results of cluster analysis. Distance units are arbitrary. Each four-character code represents one oyster. G roup charac­
teristics were determined on the basis of majority, a and b are considered together, and written as ‘a’ in the group characteris­
tics. Upper case characters are misclassifications (not for tree species). Oysters on one line are less than 1 unit distance apart.

Each o f the tw o above m entioned groups were
separated into tw o subgroups, on distance levels
o f 9 and 14 respectively. These tw o divisions co ­
incided with the oysters growing on small or on
large diam eters o f substrate. Only one misclassi-
fication occurred for this criterion. Statistics for
this division (Chi square, p and L am bda) are also
given in Table 1.
The interpretation for the tw o other environ­
m ental param eters w as no t as straightforw ard, in
the sense th at there w as no strict dichotom y sub­
dividing each o f the groups on levels lower than
the ones discussed above. How ever, it w as still
possible to assign a group characteristic to each
of the groups o f the cluster analysis with respect
to orientation, if only tw o categories are consid­
ered (parallel or perpendicular). If the group ch a r­
acteristic is taken to be the value o f the m ajority
in the group, 13 m isclassifications occurred. As
can be seen from Table 1, this is still a highly
significant result, though the value for the index
of predictive ability w as m uch lower than in
the tw o cases discussed above. The two values
for perpendicular (facing outflowing or facing in­
flowing current) seemed to have no influence
whatsoever: these values occurred together in
m ost o f the groups.
F o r the last environm ental param eter, the spe-
114

cies o f the m angrove tree, no group characteris­ SM ALL LARGE

tic could be determ ined, so it w as not possible to
construct a contingency table as before. If the
random ness o f the distribution o f the actual value □
for the m angrove species over the different groups
is tested using a 4 x 16 table, the value o f Chi x
o
square turns out to be ju st significant. However, X

in view o f the fact th at m ost o f the expected fre­

quencies are very low, no t too m uch im portance
should be attached to this value.
O yster form is non-random ly distributed with
regard to A vicennia, i.e., if a group o f related forms
contained an oyster growing on Avicennia, the
other oysters o f the group h ad a higher probability 5
o
—I
o f having been collected on Avicennia as well.
This can be tested by constructing a 2 x 16 co n ­
tingency table, w here the tw o categories are re­
spectively n o n -Avicennia and Avicennia. The Chi
Fig. 3. Outlines of shells of oysters growing in different envi­
square value for this table is highly significant.
ronmental conditions. 1: small diameter substrate, high above
It is clear from these results th at the form o f the datum; 2: large diameter substrate, high above datum; 3: small
oysters is influenced by both height above datum diameter substrate, low above datum; 4: large diameter sub­
and diam eter o f the substrate. Height above strate, low above datum.
datum h ad the m ost im portant influence, as re­
flected in the fact th at the two groups were sep­ distribution over the different groups o f the clus­
arated at the highest level of distance. Differences ter analysis w as clearly non-random . This w ould
o f form resulting from differences o f diam eter of imply a different form for the oysters growing on
the substrate were smaller. T he num ber o f m is­ Avicennia, as com pared to the other species. This
classifications for this last environm ental p aram ­ does not com e as a com plete surprise: the density
eter w as less in this study, but this w as caused by o f oysters w as m uch lower on this species o f m an ­
the way in which the sam pling w as done (contin­ grove tree. T he m echanism o f this phenom enon
uous sam pling throughout the range for height rem ains to be investigated, but it is n o t at all
above datum , versus sam pling on the extremes im possible th a t the sam e factor w ould also influ­
for substrate diam eter). ence the form o f the oysters. This factor could be
The influence o f substrate diam eter was already
dem onstrated by a prelim inary study, using an a­
PARALLEL PE R PE N D IC U L A R
logue m ethods on a small sam ple of oysters grow ­
ing on artificial substrates. T he influence o f the
height above datum on size was dem onstrated
earlier (O kem w a eta!., 1986), but no discussion
o f the influence o f this param eter on the form had
been described up to now.
O rientation with respect to m ain tidal current
seem ed to have a weaker, but still highly signifi­
can t influence on the form. The species o f m an­
grove tree had no apparent effect on the form of Fig. 4. Outlines of shells of oysters growing in different envi­
the oysters, if the four species were taken sepa­ ronmental conditions. Left: parallel to tidal current; right:
rately. If Avicennia w as considered separately, the perpendicular to tidal current.
 115

a mild oyster repellent, o f the texture of the bark

o f Avicennia, th at is sm oother than th at o f the
£ 45-
other m angrove species. £

35-
Shell length

The diam eter o f the branch serving as substrate

and the orientation with respect to the m ain tidal
current appeared to be o f no im portance for shell
length. Shell length was no t related to density up
to a cover o f 70 %. F o r densities higher than 70 %
there was a fairly strong negative correlation
(r2 = 0.634, n = 217), as one w ould expect (Fig. 5).
Length does n o t seem to be correlated with
height above bottom for heights higher than
20 cm. A t a height lower than 20 cm above the
bottom , the oysters are clearly shorter than at
higher levels.
The relationship betw een shell length o f C. cu­
cullata and the height above chart datum is shown
in Fig. 6. F rom this figure it is clear th at when
shell length is plotted against height above datum
the y values are n ot independently and norm ally
distributed for any given value o f x. However,
according to R icker (1973) regression analysis is
very robust to deviations from its underlying as­
sum ptions. N oting the fact th at we have to inter­
pret the following results with the necessary cau­
tion we can still calculate the regression
param eters. The correlation between oyster length
and height above chart datum is negative and
very low (r2 = 0.060; n = 957). If all m easurem ents
o f oysters closer than 20 cm to the bottom , and
~T---1---1--- 1--- 1---1---1-
1.4 1.6 1.8 2 2.2 2.4 2.6 2.8
height above datum (m )
Fig. 6. Relationship between height above chart datum (m)
and oyster length (mm). All measurements are included in this
figure.
all from a density o f m ore than 70% cover are
deleted from the d ata set, the negative correlation
increased dram atically (r2 = 0.850; « = 543;
Fig. 7).
The fairly strong negative correlation coeffi­
cients (r2) dem onstrate the influence o f density on
oyster length for densities higher than 70% cover.
O ther authors have dem onstrated th at density
m ay affect grow th rate and size in some littoral
bivalves (Trevaillon e ta l., 1970). This relation­
ship (between size and density) is not found by
O kem w a etal. (1986). They m easured the shell
length o f Crassostrea cucullata on a rocky sub­
stratum and concluded th at size is only a function
of height above chart datum . How ever, in
O kem w a et al.’s study, oyster density w as smaller
than 70% cover throughout. C om bined with our
results, this indicates th at oyster density does not
52.5-1

47.5-1

x: 42.5-
42.5-

I 37.5-
•c 37.5-

o 32.5-
32.5-

ô 27.5- 27.5-
1.6 1.8 2 2.2 2.4 2.6 2.8 3

22.5- height above datum level ( m)

100 105
Fig. 7. Relationship between height above chart datum (m)
density ( % )
and oyster length (mm) (r2 = 0.850; « = 543). Measurements
Fig. 5. Relationship between density (% cover) and oyster on oysters growing lower than 20 cm above the bottom and/or
length (mm) for densities higher than 70% ( ^ = 0.634; in areas with an oyster density higher than 70% cover are
n = 217). excluded from this data set.
116
affect oyster size for densities lower than 70%
cover.
Shell lengths o f C. cucullata are small at a height
lower th an 20 cm above the bottom as com pared
to oysters growing higher. This m ay be due to the
higher sedim ent load closer to the m uddy bottom .
B ecause oysters are filter feeders, they m ay have
to expend m ore energy in filtering the w ater at
lower levels to get an equal am ount of food as do
the oysters growing at higher levels.
The correlation betw een oyster length and the
height above d atum is negative and very low.
H ow ever, w hen we rem oved all oysters growing
lower th an 20 cm above the bottom , and all oys­
ters growing on su bstrata with density greater
th an 70% cover, the correlation becom es very
strong (r2 = 0.850; n = 543). This high negative
correlation coefficient dem onstrates th at
Crassostrea cucullata exhibits size-related patterns
in its vertical distribution with an up-shore reduc­
tion in shell length.
Conclusions
The form o f the oysters is clearly related to both
height above datum and diam eter o f the substrate.
Changes o f form associated with differences in
height above ch art datum were the m ost im por­
tan t, as reflected in the fact th at the tw o groups
were separated at the highest level o f distance.
Differences o f form resulting from differences of
diam eter were smaller. The num ber o f m isclassi­
fications for this last environm ental factor w as
less in this study, b u t this can be attributed to the
way in which the sam pling w as done (continuous
sam pling for height above datum , as opposed to
dichotom ised sam pling for substrate diam eter).
The distinction betw een these four groups can
be related to the appearance o f the oysters
(Fig. 3). O ysters growing on small diam eter sub­
strates are m ore elongated than the ones growing
on large diam eter substrates. This is w hat one
w ould expect: 30 mm is small com pared to the
size o f the oyster. F or the oyster to grow as wide
on this type o f substrate as on the larger diam e­
ter substrate, the oyster should grow with a large
proportion o f its shell detached from the sub­
strate. This w ould interfere w ith the structural
strength o f the shell.
The outline o f oysters growing higher above
chart datum is ‘w avier’ than the low er growing
oysters: their outline h ad m ore lobes, and the
lobes were m ore pronounced. It is no t clear w hat
the reason for this difference is. M aybe it is re­
lated to the tim e available to the oysters for feed­
ing: the higher above chart datum the oyster is
growing, the shorter the tim e it will be subm erged.
A wavier outline will increase the size o f the open­
ing via w hich w ater can be taken in or expelled,
and hence possibly increase the filtering rate. The
wavier outline o f the higher oysters could be a
com pensation for the shorter feeding time.
O rientation w ith respect to m ain tidal current
seemed to have a w eaker, but still highly signifi­
cant correlation w ith form. O nce again, this dis­
tinction is also apparent from a visual inspection
of the outline o f the oysters (Fig. 4). O ysters
growing perpendicular to the tidal current often
have lobes all around their outline, as opposed to
oysters growing parallel to the current, th at have
lobes m ainly on one side.
H eight above chart datum , and thus percentage
o f tim e im m ersed, seems to be the prim ary factor
determ ining the shell length o f the oysters. This
relationship is obscured by the influence o f
crow ding and proxim ity to the bottom . C om pe­
tition for space is the obvious m echanism by
which crow ding can influence the length o f the
oyster shell. T he way in w hich proxim ity to the
bottom is influencing oyster grow th rem ains to be
investigated, but a probable m echanism is the in­
terference o f high sedim ent load with the filter
feeding o f the oysters.
A s with the correlation betw een form and
height above chart datum , it seems logical to look
for an explanation o f the correlation betw een shell
length and height in the tim e an oyster is sub­
m erged, and hence feeding time. O ysters growing
lower have a longer proportion o f the day avail­
able for feeding, hence can be expected to grow
faster. If the m axim um size an oyster can reach
is dictated by age, rather th an by size itself, an
upshore reduction in size is the expected trend.
This study has only d em onstrated the existence

o f the relationship betw een environm ental param ­
eters on one hand, and form and size of the oys­
ters on the other. A following step in this research
should be the construction of a functional m odel,
to investigate the causal relationship between
m orphology and the environm ental param eters,
and to investigate the hypotheses form ulated
above.
Acknowledgements
This research w as conducted in the fram ew ork of
the Kenyan-Belgian Project in M arine Sciences,
at the K enya M arine and Fisheries R esearch In ­
stitute (K M F R I) in M om basa. W e wish to thank
b oth M r. S. O. Alíela and Dr. E. Okem w a, re­
spectively form er and current director, o f this in­
stitute. W e also th an k M r. R. K R uw a for his
cooperation.
References
Anstey, R. L. & D. A. Delmet, 1972. Genetic meaning of
zooecial chamber shapes. Science 177: 1000-1002.
Anstey, R. L. & D. A. Delmet, 1973. Fourier analysis of zoo­
ecial shapes in fossil tubular bryozoans. Geol. Soc. Am.
Bull. 84: 1753-1764.
Chelazzi, G. & M. Vannini, 1980. Zonation of intertidal mol­
luscs on rocky shores of southern Somali. Estuar. coast,
mar. Sei. 10: 569-583.
Christopher, R. A. & J. A. W aters, 1974. Fourier series as a
quantitative descriptor of miospore shape. J. Paleontol. 48:
697-709.
Comfort, A., 1946. Patella vulgata L. relation of habitat to
form. J. Conch. 22: 216-217.
Day, J. H., 1974. A guide to marine life of South African
shores. Balkema. Cape Town, 300 pp.
Ebling, F. J., J. F. Sloane & H. M. Davies, 1962. The ecol­
ogy of Lough Ine. XII. The distribution and characteristics
of Patella species. J. anim. Ecol. 31: 457-470.
Ehrlich, R. & B. Weinberg, 1970. An exact method for char­
acterisation of grain shape. J. Sed. Petrol. 40: 205-212.
Everitt, B. S., 1977. The analysis of contingency tables. Chap­
man and Haii. London, 128 pp.
Ferson, S., F. J. Rohlf & R. K. Koehn, 1985. Measuring shape
variation of twodimensional outlines. Syst. Zool. 34: 5 9-
68.
Gallin, E., E. Coppejans & H. Beeckman, 1989. The man­
grove vegetation of Gazi Bay (Kenya). Bull. r. Soc. Belg.
122: 197-207.
Hartnoll, R. G., 1976. The ecology of some rocky shores
117
in tropical East Africa. Estuar. coast, mar. Sei. 4: 1-21.
Healy-Williams, N. & D. F. Williams, 1981. Fourier analysis
of test shape of planktonic foraminifera. Nature 289: 485-
487.
Kaesler, R. L. & J. A. W aters, 1972. Fourier analysis of the
ostracode margin. Geol. Soc. Am. Bull. 83: 1169-1178.
Kesteven, G. L., 1941. The biology and cultivation of oysters
in Australia. I. Some economic aspects. Commonwealth of
Australia. Melbourne, 32 pp.
Kuhl, F. P. & C. R. Giardini, 1982. Elliptic Fourier features
of a closed contour. Computer Graphics and Image Pro­
cessing 18: 236-258.
Lewis, J. R., 1964. The ecology of rocky shores. H odder &
Stoughton, London, 323 pp.
Okemwa, E. R., R. K. Ruwa & P. Polk, 1986. The autecol-
ogyofthe edible oyster Crassostrea cucullata Bom, 1778: size-
related vertical distribution at Mkomani, Mombasa. Kenya
J. Sei. Ser. B 7: 9-14.
Ricker, W. E., 1973. Linear regressions in fishery research. J.
Fish. Res. Bd Can. 30: 409-434.
Ruwa, R. K., 1984. Invertebrate faunal zonation on rocky
shores around M ombasa, Kenya. Kenya J. Sei. Ser. B 5:
49-65.
Ruwa, R. K. & W. H. Brakel, 1981. Tidal periodicity and
size-related variations in the zonation o f the gastropod
Nerita plicata on East African rocky shore. Kenya J. Sei.
Ser. B 2: 61-67.
Tack, J. F., 1990. Ecomorfologie van de mangroveoester
Crassostrea cucullata (Born, 1778). Unpublished M. Sc.
thesis, Free university of Brussels, 99 pp.
Taylor, J. D., 1968. Coral reef associated invertebrate com­
munities (mainly molluscan) around Mahe, Seychelles. Phil.
Trans, r. Soc., Lond. Ser. B 254: 129-206.
Taylor, J. D., 1970. Intertidal zonation at Aldabra atoll. Phil.
Trans, r. Soc., Lond. Ser. B 260: 173-213.
Trevaillon, A., R. R. C. Edwards & J. H. Steel, 1970. Dy­
namics o f a benthic bivalve. In J. H. Steel (ed.), Marine
Food Chains. University of California Press, Los Angeles:
285-295.
Van Someren, V. D. & J. P. Whitehead, 1961. An investiga­
tion of the biology and culture of an East African oyster,
Crassostrea cucullata. Coi. Off. Fish. Pub. 14: 1-36.
Vermeij, G. J., 1972. Intraspecific shore-level size gradients in
intertidal molluscs. Ecology 53: 693-700.
Vermeij, G. J., 1973. Morphological patterns in high intertidal
gastropods: adaptive strategies and their limitations. Mar.
Biol. 20: 319-346.
W ard, J. H., 1963. Hierarchical grouping to optimise an ob­
jective function. J. am. Stat. Ass. 58: 236-244.
Ward, J. H. & M. E. Hook, 1963. Application of an hierar­
chical grouping procedure to a problem of grouping pro­
files. Educational and Psychological Measurement 23: 6 9-
81.
Younker, J. L. & R. Ehrlich, 1977. Fourier biometrics: har­
monic amplitudes as multivariate shape descriptors. Syst.
Zool. 26: 336-342.
.
Hydrobiologia 247: 119-120, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 119
© 1992 Kluwer Academic Publishers. Printed in Belgium.

A note on the Zooplankton distribution and diversity in a tropical

mangrove creek system , Gazi, Kenya

M elckzedeck K. W. O sore
Kenya Marine and Fisheries Research Institute, P.O. B ox 81651, M ombasa

Abstract

G azi Creek is situated at some 40 km south o f M om basa (K enya 4° 25' S, 39° 50' E) sam pling stations
are located in the creek m outh (stn 1), in the inner creek (stn 3) and interm ediate (stn 2). Sam pling is
done twice a m onth; it starts from stn 1 through stn 2 up to stn 3. A 335 ¡im M esh size net is tow ed
in n ear surface w ater for 5 m inutes and the collected sam ple preserved in 5% form aldehyde. H ydro-
graphic param eters are recorded during sampling.
The w ork at G azi is undertaken to survey the Z ooplankton com position both qualitatively and
quantitatively. 22 im portant taxa have been studied to determ ine their seasonal variation in abundance
and distribution. Z ooplankton population are highest in M arch (374 anim als m _3). The abundance
gradually falls during the S.E. m onsoon period (M ay -S ep tem b er) to the lowest value in August (30
anim als m _ 3) (see Fig. 1). C opepoda is the m ost abundant taxon throughout (4 8 .5 -9 2 .4 % ) (see Table 1).
T his group is found throughout the creek. N evertheless, it is evident th at the creek m outh has a higher
diversity o f copepoda com pared to the inner creek. U sing the M argalef Index a consistently higher value
h as been observed in stn 1 com pared to stn 3 (see Table 2). The surface w ater tem perature decreases
during the S.E. m onsoon (28.0 to 25.5 °C ) and it rises during the N .E . m onsoon (29.0 to 35.5 °C). The
tren d in tem perature variation corresponds broadly w ith Zooplankton abundance (see Fig. 1), suggest­
ing th at Z ooplankton thrive best in w arm er w ater. High Z ooplankton counts were also observed around
M ay (326 anim als m 3) during the long rains; probably due to high am ounts o f nutrients input. G en ­
erally, average m onthly pH values vary only slightly but the p H up the creek is alm ost always lower than
at the creek m outh. Salinity is quite co n stan t at 35%0.

5 100

Mar Apr May Jun Jul Aug Sep Oct Nov Dec
M o n th s

Fig. 1. Mean monthly total Zooplankton variation and mean monthly surface water temperature for the three stations (average)
at Gazi during M arch to December 1990.
120

Table 1. Percentage composition of important Zooplankton groups as average of the three stations at Gazi during M arch to
December 1990.

Group 1990

M ar Apr May Jun Jul Aug Sept Oct Nov Dec

Copepoda 83.36 79.10 62.56 76.49 75.05 76.85 48.49 92.42 _ 84.26
Medusae 0.18 0.16 0.31 0.09 0.00 0.00 0.00 0.04 - 0.00
Brachyura zoea 0.62 11.51 18.85 1.16 5.04 2.50 26.87 0.91 - 2.02
Caridea 0.28 0.06 1.91 0.68 3.86 2.26 6.33 0.76 - 0.35
Chaetognatha 0.17 0.86 2.84 0.42 3.24 0.99 0.71 1.19 - 0.75
O stracoda 0.13 0.58 0.16 0.09 0.28 0.82 0.63 0.19 - 0.27
Mollusca 0.29 1.55 1.68 5.08 4.86 8.41 3.10 0.70 - 1.33
Fish larvae 0.11 0.00 0.19 0.16 0.77 0.38 1.05 0.18 - 0.23
Nauplii 8.59 1.34 0.28 0.16 0.95 4.65 1.83 0.99 - 6.62
Amphipods 0.21 0.43 0.33 1.75 0.68 1.68 1.31 0.38 - 0.48
Appendicularia 0.29 2.08 0.85 0.43 0.68 0.00 0.00 0.63 - 0.11
Polychaeta 0.13 0.84 0.58 0.25 0.94 1.37 0.63 0.16 - 0.45
Isopoda 0.11 0.00 0.24 0.93 0.27 1.03 0.63 0.34 - 0.71
Fish eggs 0.10 0.41 8.06 0.43 1.92 0.82 6.14 0.18 - 1.18
Cumacea 0.16 0.06 0.00 0.34 0.50 0.72 0.26 0.25 - 0.38
Stomatopoda 0.02 0.00 0.06 0.00 0.00 0.27 0.71 0.16 - 0.00
Penaeidae 0.02 0.00 0.06 0.00 0.21 0.03 0.00 0.05 - 0.18
Cirripede nauplii 0.02 0.34 0.00 0.00 0.00 0.03 0.00 0.05 - 0.15
Brachyura megalopa 0.00 0.10 0.54 0.09 0.30 1.03 0.47 0.11 - 0.29
Siphonophora 0.00 0.06 0.64 0.09 0.18 0.00 0.00 0.00 - 0.00
Euphausiacea 0.00 0.00 1.19 0.00 0.00 0.00 0.00 0.00 - 0.00
Sergestidae 0.00 0.00 0.0 0.00 0.09 0.00 0.00 0.00 - 0.00
Others 0.05 0.32 0.00 0.00 0.00 0.03 0.54 0.18 - 1.50

Table 2. Monthly variation in diversity o f Copepods in the

three stations at Gazi during M arch to December 1990. The
Margalef Index is used.

Month Diversity

Stn 1 Stn 2 Stn 3

M arch 1.25 0.80 0.46

April 1.29 1.17 0.97
May 1.79 2.02 1.77
June 1.12 1.87 0.82
July 1.99 1.87 1.85
August 2.82 2.30 2.04
September 1.86 2.05 1.66
October 1.73 1.22 0.84
November - - -

December 1.64 1.43 1.63

Hydrobiologie 247: 121-132, 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 12 1
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Phase shifts in coral reef communities and their ecological significance

T. J. D one
Australian Institute o f Marine Science, PM B 3, Townsville M C , Queensland, 4810, Australia

K ey words: coral reef, degradation, m anagem ent, m acroalgae, phase-shift

Abstract

M any coral reefs around the w orld have degraded to a degree th at their present intrinsic value and utility
are greatly reduced: (m ass coral m ortality followed by algal invasions; local depletions o f reef fisheries;
deficit o f reef accretion com pared to physical and biological erosion). Though we can som etim es iden­
tify proxim al causes (outbreaks o f coral predators and eroders; over-fishing; habitat destruction), we do
n o t have a good understanding o f how population, com m unity and ecosystem structure and function
differ in degraded from un-degraded reefs. T he deficiencies in our understanding limit our ability to
interpret the long-term significance o f reef degradation, and therefore to develop scientifically based plans
for conservation and m anagem ent o f reefs.
A particular course o f action, or lack o f action, based on uncritical acceptance o f any o f the various
views o f tem poral variability can lead to further deterioration o f specific reefs. N one o f these views -
th at reefs are either inherently robust, inherently fragile, or inherently resilient - is true over all tim e-space
scales. This presentation reviews various m odels and case studies which suggest th at reefs can be
knocked precipitously or move slowly from one phase (coral-dom inated) to another (coral-depleted
an d /o r algal dom inated). T ransitions in the other direction (‘recovery’) involve changes (e.g. succession)
in populations and com m unities (of all reef-associated biota, no t ju st sessile benthos), and in reef function
(e.g. com m unity m etabolism , trophodynam ics) which are o f great intrinsic interest but only poorly u n ­
derstood.

Introduction long duration, frequently anthropogenic. H atcher

W e often hear coral reefs described in w ords
som ething like these: oases o f high diversity and
productivity in oceanic deserts; builders o f their
own habitat; objects o f great beauty. A nd yet ju st
as frequently, it seems, we hear o f yet another reef
which has becom e ‘degraded’ in som e way (e.g.
Brown, 1987; Salvat, 1987). R eef degradation has
m any causes, b oth episodic disturbances o f short
duration (e.g. storm s, outbreaks o f coral p re d a­
tors; brief exposure to toxic pollutants; siltation
caused by dredging), and chronic disturbances of
eta l. (1989) reviewed anthropogenic effects on
coral reefs under 9 categories: sedim entation;
sewage pollution, therm al pollution, radioactive
pollution; hydrodynam ic influences; physical dis­
turbance; extractive activities; introductions;
tourism . The scale o f reef degradation ranges from
the im m ediate vicinity o f a m arina developm ent
or sewage outfall, to a significant proportion of
the reefs within large systems such as the G reat
Barrier R eef (G B R ) or the F lorida Keys. F or
example, populations o f the coral eating crown-
of-thorns starfish Acanthaster planci seeded sue-
122
cessive reefs in a n o rth /so u th sequence over the
decade 1980-90, causing serious dam age to 65%
o f reefs surveyed in a 500 km section o f the
20 0 0 k m long G B R (M oran e ta l., 1988). By the
time those in the south were attacked, there had
been substantial coral recovery in the north (D one
e t a l , in prep.).
In this presentation, I discuss w hat degrada­
tion m eans. W hat are its sym ptom s and w hat is
its significance in relation to ‘norm al’ tem poral
variability in coral reef ecosystem s?
The one universal sym ptom o f coral reef deg­
radation is m ass coral m ortality followed by an
invasion o f coral skeletons by algae. This is the
‘phase shift’ referred to in the title o f this presen­
tation. A ssociated with such a change, there m ay
be significant depletions o f reef fisheries, and a
deficit o f reef accretion com pared to physical and
biological erosion. Thus the reef is degraded in
term s of its physical attractiveness, its fisheries
yield and its ability to retain its integrity as a
breakw ater protecting adjacent shorelines.
Clearly, the potential social and economic impli­
cations o f reef degradation will depend on how
long the reef rem ains degraded.
The following tw o quotations define some of
the im portant ecological issues and points o f dis­
agreem ent about the significance o f phase shifts
in coral reefs. Referring to the entire G reat B ar­
rier R eef system (som e 2900 reefs stretched over
2000 km ), D avies (1988) says:
‘Its uniqueness i s related to its robustness
and its ability always to regenerate after ca ta s­
tro p h es’ (italics added).
Referring to algal invasions o f previously coral
dom inated areas o f individual reefs, H atcher et al.
(1989) state:
‘Evidence is accum ulating th at coral reef
benthic com m unity structure can assum e at
least tw o stable forms: one coral-dom inated
and one m acroalgae-dom inated’.
A switch to m acroalgal dom inance which lasts
years to decades constitutes a fundam ental
change in the benthic com m unity structure o f the
reef with im plications for the trophic structure of
the reef and its resource value to hum an popula­
tions (D ahl, 1974). A m ore protracted switch (de­
cades to centuries) has m ajor im plications for reef
m aintenance and growth, w henever calcium ca r­
bonate losses to physical and biological erosion
are not replaced by calcifying organism s such as
corals and coralline algae.
Ecological and human consequences of phase
change
T he im plications th at extensive changes from
coral to algal dom inance have for the re e fs trophic
structure and resource value to hum an popula­
tions have received little system atic study. M ost
or all the increased algal production may be dis­
lodged by storm s, thence being exported from the
reef (C arpenter, 1990a) or entering benthic m i­
crobial com m unities on the reef (M oriarty et al.,
1985; H ansen eta l., 1987). In poorly flushed
areas, accum ulation o f algal detritus could lead to
excess biological oxygen dem and and anoxic con­
ditions inimical to survival o f some higher reef
biota (Johannes & Betzer, 1975). N o t surpris­
ingly, the w orst reported com m unity calcification
rates on reef flats were recorded on coral/algal
reef flats dom inated by m acroalgae - viz. at
Tulear, M adagascar, and M oorea (Pichón &
M orrisey, 1985; Sournia, 1976; both cited in K in­
sey, 1985).
Increased m acroalgal production m ay affect
harvestable secondary production o f herbivorous
fish and invertebrates, and their fish predators.
F ish production can be increased, either as an
increase in population size (C arpenter, 1990b) or
fish grow th rates, in situations w here food was
previously limiting. H ow ever it m ay be unaffected
in at least tw o circum stances: first, where food
chains are based m uch m ore on turfs an d /o r p e­
lagic Zooplankton w ashed onto the re e fs w ind­
w ard m argin (H am ner et al., 1988), than on m ac­
roalgae: and second, w here fish com m unity
structure is strongly determ ined by recruitm ent
and com paratively weakly determ ined by food
availability (D oherty & W illiams, 1989). In either
case, the significance o f the loss o f live coral may

relate m uch m ore to changes in habitat structure
(Bell & G alzin, 1984) than to trophodynam ic
considerations. Effects on motile reef inverte­
brates m ay be equally unpredictable, and for the
sam e sorts o f reasons.
Tim e scales o f decades to centuries are crucial
in the context o f a re e fs ability to m aintain its
structural and biological integrity over geological
time (B uddem eier & Hopley, 1988). It is true th at
coral reefs around the w orld have developed as
and where they are, no t despite, but because of,
w hat D avies {op. cit.) refers to as catastrophes:
i.e. continental drift, subsidence and sea-level
change. It is im possible no t to be im pressed with
these facts, and it is tem pting to believe th at coral
reefs will persist despite anything hum an-kind
m ay do to tropical seas. H ow ever our opinions -
w hether optim istic or pessim istic - should no t be
based only on the p ast ‘track-record’ o f coral
reefs. They should be based on an understanding
o f the ecological processes which determ ine
w hether in fact coral reefs, rather than algal-
covered carb o n ate deposits derived from ‘ex­
reefs’, can exist in environm ents created by an
exponentially increasing hum an population.
In the rem ainder o f this presentation, I refer to
com m only used ecological ‘m odels’ relevant to
the determ ination o f benthic com m unity structure
on coral reefs. I then present som e case studies
which show how local ecological idiosyncracies
o f particular reefs - physical; nutritional; biolog­
ical - can influence the course o f benthic com ­
m unity developm ent.
M odels for community structure in reef corals
and algae
Some characteristics o f minimally disturbed coral
com m unities on coral reefs are illustrated in
Fig. 1. As tim e passes, the largest corals get larger;
coral cover increases; ß species diversity (scale of
hectares) rises and asym ptotes. These net in­
creases occur despite localised colony m ortality
and recolonisation, which cause fluctuations in
percent coral cover and ocdiversity (scale o f square
m etres to decam etres). The reviews o f Connell
123
High Coral cover
/ Size of
largest corals
Low
High /h a
/s q . m.
Species d ive rsity
Low
years decades centuries
Time since colonisation
Fig. 1. Changes in some coral community attributes over
whole reef scale versus increasing time since colonisation and
without major disturbance. Partly after Connell (1978).
(1978) and C onnell & Keogh (1985) discuss the
im plications o f differences in scale, intensity and
frequency o f disturbance in relation to different
life history strategies o f corals.
U seful m odels for algal production and com ­
m unity structure on coral reefs have been pro­
vided by Littler & Littler (1985) and Steneck
(1988). Spatial patterns o f algal com position and
production are strongly influenced by patterns of
grazing, disturbance, and com petition with cor­
als, the potential upper levels o f production being
determ ined by am bient nutrient levels (Littler and
Littler, op. cit.). T here appears to be a system atic
causative relationship between the intensity of
grazing from scraping and denuding herbivores,
and the structure o f algal assem blages (Hixon,
1983; see also Fig. 2). It follows from Fig. 2 that
if pressure o f fishing substantially changes graz­
ing intensity, a consequence o f th at fishing m ay be
changes in biom ass and com m unity com position
o f the algal assemblage.
The types o f changes w hich have been observed
in coral reef benthic com m unities are shown in
Fig. 3 (after B azzaz, 1983). Reefs can be knocked
124

A precipitously or m ove slowly from one phase

(coral-dom inated) to another (coral-depleted
îMacroalgaeH!
an d /o r algal dom inated). F o r a period o f time,
en
en I (AG 4-6) ¡ü
< algae becom e the dom inant form o f benthic cover.
O fp p llll
m TURF ALGAE D isturbed areas can recolonise w ithout passing
CRUSTOSE ALGAE
through a m acroalgal phase, or they m ay enter a
(AG 7)
m acroalgal phase which they retain indefinitely,
INCREASING GRAZING FROM SCRAPING AND DENUDING HERBIVORES
either because ongoing conditions favour algal
grow th over coral grow th (Lighty, 1982; Lapointe,
Fig. 2. Herbivore-induced changes in algal community struc­ 1989) or because o f a relatively brief period in
ture, deduced from exclusion studies and natural herbivore
mortalities. ‘A G ’ refers to algal functional groups defined by
which algal grow th is enhanced, and is then sus­
Steneck (1988) as follows: 1 microalgae; 2 filamentous (sim­ tained long after any m easurable change in envi­
ple); 3 foliose or sheet; 4 corticated or coarsely branched; ronm ent exists (D ollar & Grigg, 1981; H atcher,
5 leathery; 6 articulated or jointed calcareous; 7 crustose. The 1984). H atch er {op. cit.) lists the following exam ­
m ost common herbivore-mediated shifts are between mac­
ples o f algal enhancing circum stances: (1) exclu­
roalgae and turfs (marked ‘A’) and between turfs and crusts
(marked ‘B’). Reproduced from Steneck (1988) with permis­
sion o f grazers; (2) increases in availability o f p o ­
sion of the organizing committee of the Sixth International tentially limiting nutrients; (3) reduction of
Coral Reef Symposium.

Macroalgae
Algal turfs ?High
on standing nutrients
coral skeletons

? Low
nutrients
Coral larvae
Major disturbance
? High
herbiv. settlement

Intermediate growth
disturbance

Minor or Corals Recolonisation by corals

no disturbance
Fig. 3. Effects of different intensities of disturbance on coral reef benthic community structure (after Bazzaz, 1983). Coral com­
munities which are undisturbed or subjected to minor disturbance (bottom left) undergo types of changes indicated in Fig. 1. After
intermediate disturbance, the coral community nevertheless retains coral dominance, whereas major disturbance, by definition,
causes a ‘phase-shift’ from coral dominance to algal dominance. The broken arrow indicates that reversion to coral dominance
will not necessarily take place, and the question marks indicate location-dependant variables which will influence the trajectory
of the benthic community.

com petition with other benthic organism s (e.g.
corals) by inhibiting their grow th or killing them ;
(4) clearing or m odification o f large areas o f sub­
stratum in a m anner, or at a time which favoured
colonisation by a formerly rare alga. H um ans may
be directly involved, either by affecting am bient
nutrient or sedim ent loads, or by changing fish
and invertebrate com m unity structure by fishing.
Coral to algal phase transitions and the extent of
coral recovery
T he following six examples describe coral-algal(-
coral) phase changes in H awaii, Jam aica, R e­
union, M oorea, and A ustralia. In some, the
causes o f the coral to algal shift are clear while in
others, they are m ore conjectural. The examples
show th at there are no a priori reasons to expect
th at corals will necessarily re-establish popula­
tions in areas they dom inated previously. R ecov­
ery o f coral dom inance requires a substantial sup­
ply o f propagules (larvae, fragm ents), followed by
prolific settlem ent and growth. Any given reef or
p art o f a reef m ay be ‘starved’ o f w ater-borne
propagules as a consequence o f a general scarcity
o f larvae in the w ater column (say as a result of
isolation from source reefs - e.g. G lynn, 1985) or
of topographic reef features which m ake it u n ­
likely th at larvae will be carried to the site in any
given year (e.g. Black, 1988). M oreover, even
w hen physical and nutritional conditions (includ­
ing irradiance) are optim al for coral growth, the
establishm ent o f populations o f corals m ay be
essentially dictated by the presence, absence or
abundance o f ‘other organism s’, for example sea
urchins an d /o r fish (see below).
Kaneohe Bay, Hawaii
In K aneohe Bay, H aw aii, particularly in its poorly
flushed southern basin, cause and effect have been
clearly established. Several decades of siltation
(due to dredging and land clearing) and eutro­
phication (due to sewage pollution) killed off m ost
o f the coral by the early 1970s (Sm ith e ta l., 1981).
125
The reefs becam e colonised by a non-reef-building
assem blage dom inated by m acroalgae (Dictyo­
sphaeria cavernosa) and a variety o f suspension
and filter feeding organism s (Sm ith et al., op. cit.).
In 1977-78, the sewage w as diverted aw ay from
the bay and by 1982 there was some decline in
Dictyosphaeria and some increase both in num ­
bers o f colonies and percent cover o f corals (M a-
ragos eta l., 1985). This example dem onstrates
the capacity o f corals to re-establish populations
once a m ajor perturbation is removed.
Discovery Bay, Jamaica
A t Discovery Bay, Jam aica, a shift from coral to
algal dom inance has been well docum ented. The
transitions which have taken place reflect both
the nature and the chronology o f recent disturb­
ances, and an effect of over-fishing. In 1981, H u r­
ricane Allen stripped off and sm ashed dow n m ost
branching and m uch encrusting coral (W oodley
et al., 1981). T u rf algae biom ass w as kept low by
a dense population o f the om nivorous sea urchin
Diadema antillarum on the reef, and some recruit­
m ent o f corals took place in 1980-83 (Hughes
eta l., 1987). How ever, in 1982 there was m ass
m ortality am ong the D. antillarum (H ughes et al.,
op. cit.), and a dense carpet o f m acroalgae dom ­
inated by Dictyota, Padina and Halimeda species
established itself, sm othering the recruits and
other small corals. U p to 1990, this algal carpet
effectively excluded further coral settlem ent
(H ughes, 1989 and pers. com m .).
T he urchin density m ay have been high, and
thus prone to the disease which finally decim ated
their populations (C arpenter, 1990a), because of
overfishing o f the fish predators an d /o r com pet­
itors of the urchins (W oodley, 1979). This hypo­
thesis has been supported by w ork on K enyan
reefs, where urchin (Echinometra mathaei) densi­
ties are lowest in m arine reserves protected from
fishing for the last 2 decades (M uthiga & M c-
C lanahan, 1987). Fish predation rates upon the
urchin are highest (M cC lanahan & M uthiga,
1989) in m arine parks. H ow ever Samoilys (1988)
considered th at siltation and dynam ite fishing also
126
directly affected the reefs and their fish popula­
tions.
There is still m uch to be learnt about the degree
to which overfishing can change the abundance
and com position o f herbivorous fish, predators of
herbivorous fish or invertebrates, let alone dow n­
stream effects on benthic com m unity structure
(H ay, 1984; M cC lanahan & M uthiga, 1989).
M unro & Smith (1984) attributed m ajor changes
in total fish catch and com position in the C arib­
bean, including changes in the proportion o f her­
bivorous fish, to sustained, intense and relatively
indiscrim inate fishing. In principle, fishing-in­
duced changes in fish com m unity structure m ay
be expected to have significant effects on benthic
com m unities. H ow ever fishing pressure is thought
unlikely to affect reef fish com m unity structure in
systems with strong recruitm ent lim itation o f fish
populations (M unro & Williams, 1985; D oherty
& Williams, 1989).
L a Saline Reef, Reunion
Overfishing is also thought to have played a part
in the degradation o f the reef at L a Saline in
Reunion. A t th at reef, chronic pollution and over­
fishing have been blam ed for a m arked coral-algal
shift. There h as been a proliferation o f filam en­
tous and fleshy algae Vaughaniella and Gracilaria
crassa and the bioeroding sponge Cliona
inconstans in previously coral-dom inated areas
(C uet eta l., 1988). The shift to algal dom inance
is believed to result partly from chronic nutrient
pollution o f groundw ater, but also from under­
grazing by urchins, which were rare, and by fish,
also rare as a result o f daily netting o f the reef.
The algae killed corals by overgrowth, and it was
predicted th at it would also inhibit settlem ent of
coral larvae.
Moorea, French Polynesia
A t M oorea (French Polynesia), nutrients and
overfishing have again been im plicated in reef
degradation. In addition, the im portance to coral
recovery o f hydrodynam ics and o f other popula­
tions (viz. urchins and dam selfish) have also been
recognised. In the period 1971-1981, the m ac­
roalgae Boodlea siamensis (C hlorophyta), Sargas­
sum sp. and Turbinaria ornata (P haeophyta) ex­
panded their ranges from the fringing reef to the
coral dom inated inner and outer barrier reef
(Payri & N aim , 1982). This took place during a
period o f dredging o f the coral reef for building
m aterial (G abrie eta l., 1985) and rapid develop­
m ent o f agriculture and tourist accom m odation
adjacent to the shore (Salvat, 1987). It had p er­
sisted until 1989 (D one et al., 1991) at which time
m uch o f the substratum occupied by m acroalgae
com prised standing skeletons o f corals killed by
Acanthaster planci around 1982 (B ouchon, 1985;
F aure, 1989).
It is very easy to point the finger o f blam e at the
coastal developm ent and dredging, but difficult to
prove the connections. T here w as no detectable
elevation o f nutrient levels in the strongly flushed
lagoonal and inner barrier reef w aters in a set of
sam ples collected in 1976 (R icard, 1982). Levels
o f nitrate + nitrite and reactive phosphate were
m uch less than occur in m oderately flushed la­
goons o f som e G B R reefs 50 km and m ore from
shore (F urnas eta l., 1990). This suggests th at if
nutrient enhancem ent is necessary to sustain the
algal biom ass at M oorea, it m ay occur in pulses,
such as during dredging o f storm s (resuspension
o f reef sedim ents, perhaps enriched by septic tank
runoff and groundw ater), and wet season runoff
o f silt-laden w ater, (which retains its integrity to
the outer reef passes (R. Steger, pers. com m .).
Alternatively, or in addition, there m ay be high
nutrient fluxes, despite low am bient concentra­
tions, as a result o f very rapid uptake by benthic
algae and phytoplankton (A tkinson, 1988).
C oral recolonisation at M oorea w as very
patchy by 1987-1989 (D one eta l., 1991). It was
high on the outside o f the northern barrier reef
and low on the inner barrier reef. In order to
understand coral recovery on the inner barrier
reef, one needs to u nderstand som ething o f the
hydrodynam ics o f the area and the n atural history
and population dynam ics o f some key species,
notably sea urchins and dam selfish. O n some

areas o f the reef platform , sand abrasion is a
m ajor problem for small corals. Only m assive
Porites corals were abundant in 1987-9, both b e­
cause they were no t eaten by the starfish, and
because they can colonise the sand-blasted areas
by fragm entation (Highsm ith, 1980).
P oor coral settlem ent and survival were also
n o ted in patches with dense urchin populations
(especially Diadema setosum and Echinometra
mathaei). H ow ever there are refuges from urchin
grazing. These are dam selfish (Stegastes nigricans)
territories which cover about 40% o f the habitat.
C orals (notably Acropora and Pocillopora spp.)
were m uch m ore abundant and diverse inside than
outside the territories, probably in large p art due
to S. nigricans’ active exclusion o f urchins and
herbivorous fish likely to graze off small corals.
This finding is precisely the opposite to the ob­
servations on other dam selfish species (e.g. Potts,
1977; Lobel, 1980; K aufm an, 1977), where there
is a net detrim ental effect on coral survival inside
territories.
Nearshore and fringing reefs, Great Barrier R e e f
N earshore and island fringing reefs on the G reat
B arrier R eef (G B R ) are dom inated over m uch of
their shallow, w ave-sw ept slopes and flats by
brow n m acroalgae, m ainly Sargassum spp. The
significance o f this prolific grow th is unclear. In
the C aribbean, where Sargassum-dom inated as­
semblages occupy a similar habitat, heavy wave-
action is thought to favour Sargassum through
exclusion o f grazing sea urchins and fish (Adey
et al., 1977). W ithin recent years to decades, Sar­
gassum dom inated algal assem blages have both
invaded large areas (hectares) o f G re at B arrier
R eef fringing reefs previously occupied by living
coral, and have them selves been displaced by co r­
als in other areas (T. J. D one & K. F. N avin,
unpubl.). The annual life cycle o f Sargassum gives
it an advantage in the w ave-sw ept rubble banks,
but puts it at a disadvantage when the alga is in
direct com petition for space against corals. F ast
lateral grow th o f encrusting Montipora and ta b u ­
late Acropora corals enable them to overgrow Sar­
127
gassum holdfasts during their seasonal (sum m er)
decline (D one & N avin, op. cit.).
High m acroalgal biom ass on G B R nearshore
reefs m ay be indicative o f high am bient nutrient
concentrations an d /o r low levels o f herbivory
(R uss, 1984; W illiams & H atcher, 1983). A case
has been m ade for eutrophication o f G B R lagoon
w aters (Bell eta l., 1989), but it is also possible
th at current standing crops o f m acroalgae may
not require an anthropogenic subsidy o f nutri­
ents. In a shallow ( < 10 m) nearshore G B R sta­
tion, soluble nutrients (nitrate + nitrite, reactive
phosphate) were strongly correlated with previ­
ous day’s w ind speed, suggesting the nutrients
were released from bottom sedim ents disturbed
by waves (W alker & O ’D onnell, 1981). H ow ever
the degree to which runoff and sewage outfalls
have enriched the sedim ent pore-w ater has not
been investigated, even though W alker and
O ’D onnell’s station is within a few kilom etres of
the sewage outfall o f the largest city adjacent to
the G reat B arrier Reef.
Offshore reefs, Great Barrier R e e f
O n the G reat B arrier Reef, coral to m acroalgal
phase shifts have been caused by m ajor episodic
disturbances such as Acanthaster planci and cy­
clones. A . planci infestations killed close to 100%
o f a high cover o f corals over large areas early in
the 1980s (E ndean & C am eron, 1985; Done,
1985). Algal biom ass per hectare increased greatly
as a result o f increased area for settlem ent (viz.
skeletons o f dead corals). U nfortunately, there
has been no large scale system atic study o f algal
com m unity dynam ics following A. planci dam age.
H ow ever it appears th at fish grazing, which usu­
ally keeps algal biom ass low on these reefs (R uss,
1984; W illiams & H atcher, 1983) w as no longer
able to do so, and th at there w as an increase in
abundance o f fleshy m acrophytes following the
outbreak (D one, pers. obs.).
N evertheless, there has been significant rever­
sion tow ards coral dom inance in the 6 -1 0 years
since the outbreak. P ercent coral cover and col­
ony size frequency distributions determ ined in
128
1990 (D one et al., in prep.) are consistent with
good settlem ent and grow th o f corals each year
since the A . planci outbreak, and a high coral
cover is predicted by the year 2000 (D one et a l,
1988). H ow ever on the inner lagoon o f one reef
(only 20 0 -3 0 0 m from the well colonised outer
slope), there h ad been a very poor recovery (D one
eta l., in prep.), possibly due to low densities of
planulae arriving in the lagoon, poor settlem ent,
an d /o r post-settlem ent survival (see B abcock,
1988).
Sum m ary o f examples
O f all the examples, the outer slopes o f offshore
G B R reefs m ay provide the best ecological cir­
cum stances for coral recovery. There is an ade­
quate supply o f planktonic larvae from upstream
reefs and com m unities (Oliver & Willis, 1987;
Williams et al., 1984), and sufficiently good early
coral grow th and survival throughout the first
6 -1 0 years following the disturbance. By co n ­
trast, som e reefs or sections o f reefs m ay rarely be
reached by dispersive coral larvae due to their
rem oteness from source reefs (G lynn, 1985),
or unfavourable reef-scale circulation patterns
(Black, 1988; cf. Sam m arco & A ndrew s, 1988).
In several other examples (e.g. Discovery Bay;
M oorea) it w as shown th at it m ay be necessary
to u n d erstan d som ething o f the natural history
and population dynam ics o f other reef biota (e.g.
urchins and dam selfish), the direct effect o f fish­
ing on fish com m unity structure, and indirect ef­
fects on benthic com m unity structure.
Temporal perspectives on phase change
W hile degradation m ay seriously affect the im me­
diate am enity o f a coral reef, it m ay not be sig­
nificant when viewed from a geological perspec­
tive.
‘Events th at loom large in the eyes o f contem ­
porary observes....m ay be undetectable in the
re e fs sedim entary record and trivial in the
re e fs geological history if the killed areas are
prom ptly recolonised’ (B uddem eier & Hopley,
1988).
This is a perceptive statem ent, both acknow ledg­
ing th at ‘prom pt recolonisation’ will no t neces­
sarily follow disturbance, and at the sam e time
counselling conservationists, biologists and m an­
agers to seek a longer-term view o f the im plica­
tions o f reef change. In som e circum stances, cor­
als o f any type are very slow to recolonise
dam aged areas, there being little recovery after a
decade or m ore, e.g. following coral mining at
M aldives (Brown & D unne, 1988), and following
coral death at M oorea (D one e ta l., 1991). H ow ­
ever there are other exam ples where dam aged
areas have been prom ptly recolonised by reef
building corals (P earson, 1981; Colgan, 1987;
D one eta l., 1988), and it is apparently on the
basis o f the latter type o f exam ple th at som e au­
thors have been lured to suggest th at corals are
m ost likely to quickly (decades) establish dom i­
nance in any suitable surface which becom es
available (e.g. H opley & Kinsey, 1988 - writing
in reference to inundation o f coral reef tops with
rising sea level).
H ow ever, the definition o f th at point in time
when recovery m ay be regarded as ‘com plete’ v ar­
ies. R estitution o f pre-disturbance percent coral
cover is the m ost cited (see Pearson, 1981) but
least stringent criterion o f recovery. The m ost
stringent - ‘total recovery’ by both ecological and
geological stan d ard s - w ould be the restitution of
prior diversity, mix o f colony grow th form s (if not
species), colony size frequency distributions and
structural extent and com plexity o f reef fram e­
w ork (after Johannes, 1975). These can take con­
siderably longer to recover th an percent coral
coverage (see Fig. 1). F o r example, D one (1988a)
and E ndean et al. (1988) suggest th at recovery o f
size-frequency distributions is partially killed and
injured populations o f slow growing corals (genus
Porites and other m assive corals) can take a cen­
tury or m ore in the sam e habitats in which per­
cent coral cover has recovered in the first 1-2
decades (D one e ta l., 1988).

Problems and scale of management
O ne m anagem ent option which is widely accept­
able is the creation o f m arine preservation areas
including coral reefs. A reas which are protected
from hum an exploitation can act as reservoirs for
unprotected reefs dow nstream , where species are
depleted by fishing, collecting etc. H ow ever the
sam e current systems which carry the larvae of
corals and fish also carry pollutants, sedim ents,
and the larvae o f destructive organism s such as
Acanthaster planci and Drupella spp. (coral-eating
gastropods). It is therefore prudent from a reef
conservation perspective to create as m any re­
serves as possible, and to m ake them as large as
possible. Experience on the G reat Barrier R eef is
th at m ajor coral kills can take place on reefs scat­
tered over hundreds o f kilom etres o f the reef tract,
equivalent to the length o f the entire Kenyan
coast.
Conclusions
While the sym ptom s o f w idespread coral death
and algal invasion are usually obvious, the causes
are frequently m ore a m atter for conjecture than
direct observation. W here a cause is definitely
identified - say, overfishing, siltation caused by
dredging or land clearing, sewage pollution, toxic
w astes - it m ay be in a country’s best im m ediate
econom ic interests to prevent th at im pact. As one
example, H odgson & D ixon (1988) found the rev­
enue loss to dive tourism caused by siltation o f a
Philippine reef area w ould be greater than reve­
nue gained by the logging w atersheds w hich would
cause the siltation in the first place. As a second
example, D one (1988b) suggested in relation to
M auritian fringing reefs, th at one m ust weigh the
cost o f protective m anagem ent against the poten­
tial for reef erosion and dam age to the shore and
structures.
A scientist offering advice to governm ent on
the best course o f action in relation to coral reef
m anagem ent walks a precarious path. Knowledge
o f general reef character, history, structure and
processes - as can be gleaned from personal ex­
129
perience and from the literature - is ju st the be­
ginning. The uncertainties associated with inter­
preting w hat has happened to a specific reef in the
past, and predicting how it will behave in the
future, m ust also be confronted. On the time and
space scales relevant to coral reef m anagem ent, it
is not appropriate to m ake decisions based on a
presum ption th at a specific reef or system o f reefs
has ‘ability always to regenerate after ca ta stro ­
phes’ (D avies, 1988). N either is it correct to as­
sume th at a degraded reef will rem ain so forever,
or th at those reefs which do regenerate will do so
in time scales which are convenient to hum an
users o f the reef.
This presentation has touched lightly on the
m any issues - physical, chem ical and biological
- which m ust be addressed in trying to under­
stand the causes and consequences o f coral to
algal phase shifts in coral reefs. Clearly, a multi­
disciplinary approach is desirable for in-depth
understanding o f all complex natural systems.
T he initiatives o f U N E S C O - C O M A R A F in
creating ‘The Regional Project for Training and
R esearch on A frican C oastal System s’, and the
U niversity o f N airobi in organizing this Sym po­
sium are both welcome and timely.
Acknowledgements
I am grateful to the sponsors, executive com m it­
tee and co-ordinators of the International Sym ­
posium on the Ecology o f M angrove and Related
Ecosystem s, 2 4 -3 0 Septem ber 1990, M om basa,
Kenya, for providing the forum for this presen­
tation. Els M artens’ tireless efforts throughout all
stages o f organisation and publication are espe­
cially appreciated. C om m ents on the m anuscript
by Jerem y W oodley and an anonym ous reviewer
were m uch appreciated. This is A ustralian Insti­
tute o f M arine Science C ontribution No. 536.
References
Adey, W. H., P. J. Adey, R. Burke & L. Kaufman, 1977. The
Holocene reef systems of Eastern Martinique, French West
Indies. Atoll Res. Bull. 218: 1-40.
130
Atkinson, M. J., 1988. Are coral reefs nutrient limited? Proc.
6th int. Coral Reef Symp. 1: 157-166.
Babcock, R. C., 1988. Fine-scale spatial and temporal pat­
terns in coral settlement. Proc. 6th int. Coral Reef Symp. 2:
635-639.
Bazzaz, F. A., 1983. Characteristics of populations in relation
to disturbance in natural and man-modified ecosystems. In
Ft. A. Mooney & M. G odron (eds), Disturbance and Ec­
osystems. Components of Response. Springer-Verlag, Ber­
lin: 258-275.
Bell, J. & R. Galzin, 1984. Influence of live coral cover on
coral reef fish communities. Mar. Ecol. Prog. Ser. 15: 265—
274.
Bell, P. R. F., P. F. Greenfield, D. Hawker & D. Connell,
1989. The impact of waste discharge on coral reef systems.
Wat. Sei. Tech. 21: 121-130.
Black, K. P., 1988. The relationship of reef hydrodynamics to
variations in numbers o f planktonic larvae on and around
coral reefs. Proc. 6th int. Coral Reef Symp. 2: 125-130.
Bouchon, C., 1985. Quantitative study of scleractinian coral
communities of Tiahura Reef, (Moorea, French Polynesia).
Proc. 5th int. Coral Reef Symp. 6: 279-290.
Brown, B. E., 1987. Worldwide death of corals: natural cyclic
events or man-made pollution? Mar. Pollut. Bull. 18: 9-13.
Brown, B. E., 1988. The environmental impact of coral min­
ing on coral reefs in the Maldives. Env. Cons. 15: 159-165.
Buddemeier, R. W. & D. Hopley, 1988. Turn-on and turn­
offs: causes and mechanisms o f the initiation and termina­
tion of coral reef growth. Proc. 6th int. Coral Reef Symp. 1:
253-261.
Carpenter, R. C., 1990a. Mass mortality of Diadema
antillarum. 1. Long-term effects on sea urchin population-
dynamics and coral reef algal communities. Mar. Biol. 104:
67-77.
Carpenter, R. C., 1990b. Mass mortality of Diadema
antillarum. 2. Effects on population densities and grazing
intensity of parrotfishes and surgeonfishes. Mar. Biol. 104:
79-86.
Colgan, M., 1987. Coral reef recovery of G uam (Micronesia)
after catastrophic predation by Acanthaster planci. Ecology
68: 1592-1605.
Connell, J. H., 1978. Diversity in tropical rainforests and coral
reefs. Science 199: 1302-1310.
Connell, J. H. & M. J. Keogh, 1985. Disturbance and patch
dynamics of subtidal marine animals on hard substrata. In
S. T. A. Pickett & P. S. White (eds), The Ecology of N at­
ural Disturbance and Patch Dynamics. Academic Press
Inc., Orlando: 125-151.
Cuet, P., O. Naim, G. Faure & J. Y. Conan, 1988. Nutrient-
rich groundwater impact on benthic communities on la Sa­
line reef (Reunion Island, Indian Ocean): Preliminary re­
sults. Proc 6th int. Coral Reef Symp. 2: 207-212.
Dahl, A. L., 1974. The structure and dynamics of algae in the
coral reef ecosystem. Proc. 2nd int. Coral Reef Symp. 1:
21-25.
Davies, P. J., 1988. Evolution of the Great Barrier Reef -
reductionist dream or expansionist vision. Proc. 6th int.
Coral Reef Symp. 1: 9-17.
Doherty, P. J. & D. McB. Williams, 1989. The replenishment
of coral reef fish populations. Oceanogr. mar. biol. ann.
Rev. 26: 487-551.
Dollar, S. J. & R. W. Grigg, 1981. Impact of akaolin clay spill
on a coral reef in Hawaii. Mar. Biol. 65: 269-276.
Done, T. J., 1985. Effects of two Acanthaster outbreaks on
coral community structure: the meaning of devastation.
Proc. 5th int. Coral Reef Symp. 5: 315-320.
Done, T. J., 1988a. Simulation of recovery of pre-disturbance
size structure in populations of Porites. Mar. Biol. 100:
51-61.
Done, T. J., 1988b. An assessment of coral reefs in Mauritius.
COM ARAF Ser. Doc. 2.
Done, T. J., K. Osborne & K. F. Navin, 1988. Recovery of
coral communities post-Acanthaster: Progress and pros­
pects. Proc. 6th int. Coral Reef Symp. 2: 137-142.
Done, T. J., P. K; Dayton, A. E. Dayton & R. Steger, 1991.
Regional and local variability in recovery of shallow coral
communities: Moorea, French Polynesia and central Great
Barrier Reef. Coral Reefs 9: 183-192.
Endean, R. & A. M. Cameron, 1985. Ecocatastrophe on the
G reat Barrier Reef. Proc. 5th int. Coral Reef Symp. 5: 309-
314.
Faure, G., 1989. Degradation of coral reefs at M oorea Island
(French Polynesia) by Acanthaster planci. J. Coast. Res. 5:
295-306.
Furnas, M. J., A. W. Mitchell, M. Gilmartin & N. Relevante,
1990. Phytoplankton biomass and primary production in
semi-enclosed reef lagoons of the Central G reat Barrier
Reef, Australia. Coral Reefs 9: 1-10.
Gabrie, C., M. Porcher & M. M asson, 1985. Dredging in
French Polynesian reefs: towards a general policy o f re­
source exploitation and site development. Proc. 5th int.
Coral Reef Symp. 4: 271-277.
Glynn, P. W., 1985. Corallivore population sizes and feeding
effects following el Niño (1982-1983) associated coral mor­
tality in Panama. Proc. 5th int. Coral Reef Symp. 1: 183—
188.
Hamner, W. M., M. S. Jones, J. H. Carleton, I. R. Hauri &
D. McB. Williams, 1988. Zooplankton, planktivorous fish,
and water currents on a windward reef face: G reat Barrier
Reef, Australia. Bull. mar. Sei. 42: 459-479.
Hansen, J. A., D. M. Alongi, D. J. W. Moriarty & P. C. Pol­
lard, 1987. The dynamics of benthic microbial communities
at Davies Reef, central G reat Barrier Reef. Coral Reefs 6:
63-70.
Hatcher, B. G., 1984. A maritime accident provides evidence
for alternate stable states in benthic communities on coral
reefs. Cora! Reefs 4: 199-204.
H atcher, B. G., R. E. Johannes & A. I. Robertson, 1989. Re­
view of research relevant to the conservation o f shallow
water tropical marine ecosystems. Oceanogr. mar. biol. ann.
Rev. 27: 337-414.
Hodgson, G. & J. A. Dixon, 1988. Logging versus fisheries

and tourism in Palawan. Occasional Papers of the East-
West Environment and Policy Institute, 7.
Hopley, D. & D. W. Kinsey, 1988. The effects of a rapid
short-term sea-level rise on the G reat Barrier Reef. In G. I.
Pearman (ed.), Greenhouse. Planning for climatic change.
E. J. BriH, Leiden: 189-201.
Hay, M., 1984. Patterns of fish and urchin grazing on Car­
ibbean coral reefs: are previous results typical? Ecology 65:
446-454.
Highsmith, R. C., 1980. Passive colonization and asexual col­
ony multiplication in the massive coral Porites lutea Ed­
wards. J. exp. mar. Biol. Ecol. 47: 55-67.
Hixon, M. A., 1983. Fish grazing and community structure of
reef corals and algae: a synthesis of recent studies. In M. L.
Reaka (ed.), The ecology of deep and shallow coral reefs.
NOAA, Rockville, Maryland. 1: 79-87.
Hughes, T. P., 1989. Community structure and diversity of
coral reefs: the role of history. Ecology 70: 275-279.
Hughes, T. P., D. C. Reed & M. J. Boyle, 1987. Herbivory on
coral reefs: community structure following mass mortalities
of sea urchins. J. exp. mar. Biol. Ecol. 113: 39-59.
Johannes, R. E., 1975. Pollution and degradation of coral reef
communities. In E. J. Ferguson-W ood & R. E. Johannes
(eds), Tropical marine pollution, Elsevier Scientific Publ.,
New York. Elsevier Oceanography Series 12: 15-51.
Johannes, R. E. & S. B. Betzer, 1975. Introduction: Marine
communities respond differently to pollution in the tropics
than at higher latitudes. In E. J. Ferguson-Wood & R. E.
Johannes (eds), Tropical marine pollution, Elsevier Scien­
tific Publ., New York. Elsevier Oceanography Series 12:
1- 12 .
Kaufman, L., 1977. The threespot damselfish: effects on
benthic biota of Caribbean coral reefs. Proc. 3rd int. Coral
Reef Symp. 1: 559-564.
Kinsey, D. W., 1985. Metabolism, calcification and carbon
production. I System level studies. Proc. 2nd int. Coral
Reef Symp. 4: 505-626.
Lapointe, B. E., 1989. Caribbean coral reefs: Are they becom­
ing algal reefs? Sea Frontiers 35: 82-91.
Larkum, A. W. D., 1983. The primary productivity o f plant
communities on coral reefs. In D. J. Barnes (ed.), Perspec­
tives on Coral Reefs. The Australian Institute of Marine
Science, Brian Clouston Publisher, Manuka: 221-239.
Lighty, R. G., 1982. Fleshy-algal domination of a modern
Bahamian barrier reef: example of an alternate climax reef
community. A bstract - Proc. 4th int. Coral Reef Symp. 1:
722.
Littler, M. M. & D. S. Littler, 1985. Factors controlling rela­
tive dominance of primary producers on biotic reefs. Proc.
5th int. Coral Reef Symp. 4: 35-39.
Lobel, P. S., 1980. Herbivory by damselfish and their role in
coral reef community ecology. Bull. mar. Sei. 30: 273-289.
Maragos, J. E., C. Evans & P. Holthus, 1985. Reef corals in
Kaneohe Bay six years before and after termination of sew­
erage discharges. Proc. 5th int. Coral Reef Symp. 4: 189—
194.
131
McClanahan, T. R. & N. A. Muthiga, 1988. Changes in Ken­
yan coral reef community structure and function due to
exploitation. Hydrobiologia 166: 269-276.
McClanahan, T. R. & N. A. Muthiga, 1989. Patterns of
predation on a sea urchin, Echinometra mathaei (de Blain­
ville), on Kenyan coral reefs. J. exp. mar. Biol. Ecol. 126:
77-94.
Moran, P. J., R. H. Bradbury & R. E. Reichelt, 1988. Distri­
bution of recent outbreaks of the crown o f thorns starfish
(Acanthaster planci) along the G reat Barrier Reef: 1985—
1986. Coral Reefs 7: 125-137.
Moriarty, D. J. W , P. C. Pollard, W. G. Hunt, C. M. Mori-
arty, T. Wassenberg, 1985. Productivity of bacteria and
microalgae and the effect o f grazing by holothurians on a
coral reef flat. Mar. Biol. 85: 293-300.
Munro, J. L. & D. McB. Williams, 1985. Assessment and
management of coral reef fisheries: biological, environmen­
tal and socioeconomic aspects. Proc. 5th int. Coral Reef
Symp. 4: 543-578.
Munro, J. L. & I. R. Smith, 1984. Management strategies in
multispecies complexes in artisanal fisheries. Proc. Gulf
Caribb. Fish. Inst. 36: 127-141.
Muthiga, N. A. & T. R. McClanahan, 1987. Population
changes of a sea urchin (Echinometra mathaei) on an ex­
ploited fringing reef. Afr. J. Ecol. 25: 1-8.
Oliver, J. K. & B. L. Willis, 1987. Coral-spawn slicks in the
Great Barrier Reef: preliminary observations. Mar. Biol.
94: 521-529.
Payri, P. E. & O. Naim, 1982. Variations entre 1971 et 1980
de la biomass et de la composition des populations de
macroalgues sur le récif corallien de Tiahura (Ile de Moorea,
Polynésie Française) Cryptogamie: Algologie 3: 229-240.
Pearson, R. G., 1981. Recovery and recolonisation of coral
reefs. Mar. Ecol. Prog. Ser. 4: 105-122.
Pichón, M. & J. Morrissey, 1985. Premieres mesures de bilan
métabolique dans l’ecosysteme recifal de Tulear (M adagas­
car). C.r. Acad. Sei. Paris 300: 99-101.
Potts, D. C., 1977. Suppression of coral populations by fila­
mentous algae within damselfish territories. J. exp. mar.
Biol. Ecol. 28: 207-216.
Ricard, M., 1982. Mean features of phytoplankton and pri­
mary productivity of M oorea Island (French Polynesia).
Abstract. Proc. 4th int. Coral Reef Symp. 1: 453.
Russ, G., 1984. The distribution and abundance of herbivo­
rous grazing fishes in the central G reat Barrier Reef. I.
Levels of variability across the entire continental shelf. Mar.
Ecol. Prog. Ser. 20: 23-34.
Russ, R. G. & A. C. Alcala, 1989. Effects of intense fishing
pressure on an assemblage of coral reef fishes. Mar. Ecol.
Prog. Ser. 65: 13-27.
Sale, P. F., 1988. W hat coral reefs can tell us about ecology.
Proc. 6th int. Coral Reef Symp. 1: 19-27.
Salvat, B. (ed.), 1987. Human Im pacts on Coral reefs: Facts
and Recommendations. Antenne Museum Ecole Pratique
des Hautes Etudes, French Polynesia, 253 pp.
Sammarco, P. W. & J. C. Andrews, 1988. Localised dispersal
132
and recruitment in G reat Barrier Reef corals: the Helix
Experiment. Science 239: 1422-1424.
Samoilys, M. A., 1988. Abundance and species richness of
coral reef fish on the Kenyan coast: the effects of protec­
tive management and fishing. Proc. 2nd int. Coral Reef
Symp. 2: 261-266.
Smith, S. V., W. J. Kimmener, E. A. Laws, R. E. Brock &
T. W. Walsh, 1981. Kaneohe Bay sewage diversion exper­
iment: perspectives on ecosystem responses to nutritional
perturbation. Pac. Sei. 35: 279-395.
Sournia, A., 1976. Oxygen metabolism of a fringing reef in
French Polynesia. Helgol. wiss. Meeres. 28: 401-410.
Steneck, R. S., 1988. Herbivory on coral reefs: a synthesis.
Proc. 6th int. Coral Reef Symp. 1: 37-49.
Walker, T. & G. O’Donnell, 1981. Observations on nitrate,
phosphate and silicate in Cleveland Bay, northern Queen­
sland. Aust. J. mar. Freshwat. Res. 32: 877-887.
Williams, D. McB. & A. I. Hatcher, 1983. Structure of fish
communities on outer slopes of inshore, mid-shelf and outer
shelf reefs of the G reat Barrier Reef. Mar. Ecol. Prog. Ser.
10: 239-250.
Williams, D. McB., E. Wolanski & J. C. Andrews, 1984.
Transport mechanisms and the potential movement of
planktonic larvae in the central region of the G reat Barrier
Reef. Coral Reefs 3: 229-236.
Woodley, J. D., 1979. The effects of trap-fishing on reef com­
munities in Jamaica. Proc. Is. Mar. Labs. Carib. 13: 27.
Woodley, J. D. & 19 others, 1981. Hurricane Allen’s impact
on Jam aican coral reefs. Science 214: 749-755.
Hydrobiologia 247: 133- 138, 1992 .
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 133
© 1992 Kluwer Academic Publishers. Printed in Belgium.

The incidence of hurricanes on the north coast of Jamaica since 1870:

are the classic reef descriptions atypical?

J. D. W oodley
Discovery B ay Marine Laboratory, P.O. B ox 35, Discovery Bay, Jamaica, West Indies

K ey words : Acropora, disturbance, hurricanes, Jam aica, coral reefs

Abstract

W hen the coral reefs o f north Jam aica were described and studied in the 1950’s, 60’s and 70’s, their
shallow zones were dom inated by dense thickets o f Acropora palmata and A . cervicornis. T hose stands
were dem olished by hurricanes in 1980 and 1988, but these seemed to be unusual events.
Published records o f storm tracks indicate that, in the last 120 years, about 35 hurricanes have passed
within 364 km (200 nautical miles) o f Discovery Bay, Jam aica. In the absence o f inform ation on their
intensity, an approxim ate m easure o f their im pact is the reciprocal o f their closest distance. R ecent
experience suggests th at passage within about 65 km, w hether to the north or the south, generates waves
capable o f destroying Acropora stands. There have been 13 years, out o f the last 120, in which hurri­
canes have passed th at close to J am aica, with a m edian interval o f 6.5 years. Large colonies o f A . palmata
(over 1 m high) might take about 12 years to develop. The total num ber o f years in which the reefs have
been free o f m ajor disturbance for longer than 12 years is only 33, and 24 o f them (1956-80) were in
the unusually long interval o f 36 years, from 1944 to 1980. The luxuriant Acropora stands o f the clas­
sic descriptions m ay therefore be atypical; one extrem e o f a variable condition.

Introduction H urricanes (and I use the C aribbean term )

Am ong tropical coastal ecosystem s, coral reefs
are the m ost exposed to wave action. The inten­
sity and direction o f prevailing wave regimes
partly determ ine their m orphology, com m unity
com position and zonation. M any reefs, however,
occasionally experience exceptionally large waves
generated by intense tropical revolving storm s
(cyclones, hurricanes, typhoons). These are brief
events, but the changes w rought by their violent
im pact m ay persist for m any years. W hat is their
p attern o f occurrence at a given locality?
Contribution Number 500 of the Discovery Bay Marine Lab­
oratory, University of the West Indies.
sweep narrow trails o f extreme im pact. There has
been an average o f only five per year in the W est­
ern A tlantic over the p ast century (N eum ann
et al., 1981). It is like a lottery where, in any year,
a given site will have a good chance o f escaping
hurricane influence. To a local observer, on that
short time-scale, it is a rare, unpredictable event.
O n larger scales o f time and space, however, h u r­
ricanes are com m on and alm ost ubiquitous; a
m ap showing a hundred years o f C aribbean tro p ­
ical storm s and hurricanes is black with their
tracks (N eum ann eta l., 1981).
N o reef in the area escapes their influence, but
the tem poral structure o f th at influence depends
on the tim e-scale o f the processes affected by
134
storm s. Thus, in relation to geological processes
o f reef grow th or sedim entation, on a tim e-scale
of hundreds or th o usands o f years, hurricanes
can be regarded as a continuous force. O ne could
recognize different intensities o f th at force due to
differences in hurricane frequency in space or over
tim e (e.g., H u b b ard, 1988). On shorter time
scales, the occurrence o f hurricanes is irregular.
Their influence on processes m easured on a scale
o f the sam e order as the interval betw een hurri­
canes, such as the grow th o f corals, m ay be bet­
ter understood in term s o f the time elapsed since
the previous storm , rather than their average fre­
quency (Kjerfve eta l., 1986).
C orals differ in their growth strategies; th at is
to say, in their growth rates, recruitm ent rates,
com petitive ability, fragility and regenerative
powers. As a result, they differ in their population
responses to storm im pacts. Montastraea
annularis (Ellis & Solander) is a m assive, slow-
growing fram e-builder on C aribbean reefs, resis­
tan t to storm s, but susceptible to overgrowth by
faster-grow ing branching corals such as Acropora
palmata (L am arck) and A . cervicornis (Lam arck).
These are strong enough to resist m oderate wave
energies and often dom inate shallow reef zones
th at are nam ed after them (G oreau, 1959). O n the
n o rth coast o f Jam aica, G oreau described the
breaker zone as ‘populated alm ost exclusively by
huge tree-like colonies o f Acropora palm ata', and
the upper p art o f the seaw ard slope as dom inated
by ‘im m ense beds o f staghorn coral., Acropora
cervicornis'. South coast reefs had been hit by
H urricane C harlie in 1951; the Acropora spp. were
less dom inant and there w as m ore rubble. G oreau
suggested, correctly but cautiously, th at ‘storm s
can play an im portant part in altering the basic
p attern o f species dom inance and zonal succes­
sion in the upper parts o f a climax re ef.
Lor another 20 years, the two Acropora species
rem ained conspicuous dom inants on the north
coast o f Jam aica. A . palmata form ed a continu­
ous thicket, up to 2 m high, from the surface to
- 6 m depth, while A cervicornis covered the shal­
low terrace to a depth of about - 2 2 m, punctu­
ated by M . annularis and other m assive corals. At
norm al wave base, in the M ixed or B uttress Z one
around - 5 to - 1 0 m, m assive corals were m uch
com m oner. In 1980, the huge w aves generated by
H urricane Allen sm ashed the Acropora thickets
and m ost o f their fragm ents died (K now lton et al.,
1981). The Acropora zones were transform ed into
rubble like th at which G oreau h ad seen on the
south coast. Allen w as the strongest C aribbean
storm of the century (at the time), and the first to
affect the n o rth coast for m any years (W oodley
et al., 1981). It was generally believed th at it was
only a m atter o f tim e before the n o rth coast reefs
returned to their form er luxuriance. H ow ever, not
only w as the recovery o f Acropora populations
unexpectedly slow (K now lton eta l., 1990), but
only eight years later they suffered another violent
disturbance, from H urricane G ilbert. The physi­
cal im pact o f this storm w as at least as great as
th at o f Allen, but the dam age to branching corals
w as m uch less spectacular because they h ad not
h ad tim e to regain their previous abundance
(W oodley, 1989).
T here are good records o f hurricanes in the
C aribbean, and it is possible to determ ine their
p attern o f occurrence with reasonable accuracy
for the last century or so. Im p o rtan t sources are
a set o f m aps showing tropical cyclones o f the
N o rth A tlantic, 1871-1980 (N eum ann e ta l.,
1981), and the very detailed Jam aica W eather
R eports w hich began in 1881. In this paper, I use
those sources to analyse the tracks o f hurricanes
th at m ight have dam aged reefs at Discovery Bay,
in the centre o f the north coast o f Jam aica, where
m any reef studies have been m ade.
Methods
Table 1 presents d ata on 39 storm s th at passed
closer than 364 km (200 nautical miles) to D is­
covery Bay since 1870. Their initial selection was
m ade from the m aps o f N eum ann et al. (1981).
O n the earliest m aps, tropical storm s (central
winds less than 3 3 m s _1 [7 4 m p h ]) and hurri­
canes are n o t distinguished. L rom 1886 to 1898,
storm s know n to have reached hurricane inten­
sity at some tim e are distinguished from the oth ­
ers, while from 1899 onw ards the various stages

o f storm developm ent are indicated along each
track. F o r the present w ork, the intensity cate­
gory o f storm s passing close to Jam aica from
1880 onw ards w as determ ined from the Jam aica
W eather R eports. These R eports were also used
for m ore accurate definition o f hurricane tracks
where they differed from the sm all-scale m aps of
N eum ann et al. The R eports on hurricanes are
often rem arkably detailed; especially those w rit­
ten by M axwell H aii, am ateur founder o f the
w eather service in 1881 and G overnm ent M ete­
orologist until at least 1919. H e collated the re­
ports o f other am ateur observers across the island
and interpreted them with great skill and enthu­
siasm.
The adjusted tracks o f the 35 rem aining storm s
th at were, or might have been, hurricanes when
they passed Jam aica, are illustrated for 12 suc­
cessive decades in Fig. 1. In the first decade, no
records were available for 1870, and three o f the
-20!
lo a k m
1 8 7 0 -7 9 75 W 1 9 3 0 -3 9 '
1 8 8 0 -8 9 1 9 4 0 -4 9
1 8 9 0 -9 9 1 9 5 0 -5 9
1 9 0 0 -0 9 1 9 6 0 -6 9
1 9 1 0 -1 9 1 9 7 0 -7 9
1 9 2 0 -2 9 1 9 8 0 -8 9
Fig. 1. Maps showing probable hurricane tracks near Jamaica
in 10 year periods, 1870-1989. Location o f Discovery Bay
shown by spot on the north coast. D ata from sources in
Table 1.
135
four storm s might no t have reached hurricane
intensity in this area. The sam e can be said for the
storm s o f 1889, 1894, 1895, 1909 and 1910 (T a­
ble 1). F urther evidence on the category o f these
storm s might be obtained from Jam aican (or
C uban) new spapers, but none is m entioned in
H all’s historical review (1916) o f hurricanes and
other phenom ena observed in Jam aica.
The height of the waves radiating from the
eye o f a hurricane decreases with distance
(B retschneider & Tam aye, 1976). A crude com ­
parative indicator o f their destructive pow er is the
reciprocal o f th at distance. If no other inform a­
tion is available on the intensity o f different hur­
ricanes, their im pacts can be com pared in this
way, as a first approxim ation. Table 1 lists the
track o f each storm , its direction (aspect) from
D iscovery Bay w hen closest, its distance at that
point (r), and the reciprocal o f th at distance (mul­
tiplied by 1000). T hose reciprocals are plotted in
Fig. 2, to an arbitrary m axim um of 33 (that co r­
responds to a hurricane centre passing 30 km
away, w hich is close enough to cause catastrophic
dam age, so there is little point in discrim inating
it from even closer encounters).
Results and discussion
Figure 1 gives a good im pression o f the irregular­
ity o f hurricane influence on a time scale o f de­
cades. But m any o f these storm s were sufficiently
distant th at their im pact at Discovery Bay m ust
have been relatively slight. Figure 2 com pares the
reciprocals o f their distance, as described above.
To calibrate them , we can consider those storm s
for which the im pact on reefs at Discovery Bay
is know n, and ask w hether or no t they destroyed
Acropora stands. T he best know n is H urricane
Allen (1980), for w hich the catastrophic im pact is
w ell-docum ented (W oodley et al., 1981). T hat de­
struction w as effected from a m inim um distance
of 45 km to the north. G ilbert (1988) was equally
violent, passing 40 km to the south. C arm en
(1974) w as 145 km south, and generated waves
th at destroyed experim ental cages in Discovery
Bay, but did no t destroy Acropora form ations.
136

Table 1. D ata on storms that might have reached hurricane intensity within 364 km (200 nautical miles) of Discovery Bay, Ja ­
maica since 1870. Direction (aspect) and distance (r) from Discovery Bay, and the reciprocal o f that distance (multiplied by 1,000)
are shown. Storm data were taken from Neumann etal. (1981) and modified from other sources as indicated. The Jam aica Weather
Reports (JWR) were published twice; first as monthly and occasional bulletins, then in collected volumes entitled Jam aica Me­
teorological Observations every 6 -8 years. Tropical storm (TS).

Year # /N a m e Date Track Aspect r(k m ) 1000/r Notes

1873 5 28 Sept WNW N 136 7 TS or Hurricane?

1874 7 2 Nov N E 22 45 H urricane (Hall, 1916)
1875 2 12 Sept NW NE 200 5 TS or Hurricane?
1879 6 12 Oct WNW S 90 11 TS or Hurricane?
1880 3 19 Aug NW E 20 50 Track adjusted (Haii, 1889)
1884 3 8 Oct N by E E N ot a hurricane (JW R 46)
1886 3 27 June WNW S 20 50 Track adjusted (JWR 67)
1886 6 20 Aug NW E 30 33 Track adjusted (JW R 69)
1886 7 18 Sept NW N N ot a hurricane (JW R 70)
1889 6 15 Sept W S 227 4 TS or Hurricane?
1894 3 23 Sept WNW NE 320 3 TS or Hurricane?
1895 2 25 Aug WNW S 180 6 TS or Hurricane?
1895 5 19 Oct NW sw N ot a hurricane (JW R 175)
1896 4 26 Sept W by N s N ot a hurricane (JW R 204)
1903 2 11 Aug W by N s 20 50 Track adjusted (Haii, 1911)
1905 4 5 Oct NE E 220 5
1909 5 24 Aug WNW N 90 11 N ot mentioned in JW R
1910 2 9 Sept WNW N 36 28 Further off? (Haii, 1911)
1912 6 18 Nov N W 13 77 Track adjusted (JWR 411)
1915 2 13 Aug WNW N 29 34 Track adjusted (JW R 446)
1916 4 15 Aug WNW S 58 17 Track confirmed (JW R 460)
1917 3 23 Sept WNW S 15 67 Track adjusted (JW R 474)
1928 2 11 Aug NW NE 320 3
1932 10 9 Nov NE NW 330 3
1933 19 30 Oct NE NW 50 20 Track confirmed (JW R 685)
1935 4 27 Sept N NW W 160 6
1935 5 22 Oct N NW NE 160 6
1938 2 11 Aug W by N S 180 6
1944 4 20 Aug W by N S 4 250 Track adjusted (JW R 828)
1950 King 16 Oct N by E W 190 5
1951 Charlie 17 Aug W by N s 90 11 Track adjusted (JW R (1951))
1954 Hazel 12 Oct NNE E 360 3
1958 Ella 1 Sept NW NE 255 4
1963 Flora 4 Oct NW NE 345 3
1964 Cleo 25 Aug WNW NE 135 7
1966 Inez 30 Sept WNW NE 320 3
1974 Carmen 31 Aug W S 145 7
1980 Allen 6 Aug WNW N 45 22 Taylor and Staff (1981)
1988 Gilbert 11 Sept W by N S 40 25 Jam aica Met Office, 1989

N o r did th at happen during any o f the distant
hurricanes from 1954-1966, w hen T. F. G oreau
w as working in J am aica, or he would surely have
docum ented it. H urricane Charlie (1951) passed
only 90 km south o f Discovery Bay. It caused
m uch dam age in K ingston and to the reefs o f the
P o rt Royal Cays, but little on the north coast (as
indicated by G oreau, 1959), w here m axim um
w inds were only 22 ms “ 1 (50 mph).
Assum ing th at all these hurricanes were o f
equal size and strength (which they were not), the
threshold distance for the destruction o f Acropora
 137

30

25

*. 20
i-
o
o
O 15

10

1880 1900 1920 1940 1960 1980

Fig. 2. Reciprocals ( x 1000) of the closest approach of hurricanes to Discovery Bay, 1870-1989, to a maximum of 33 (equiva­
lent to 30 km). Dotted line at 15 represents 65 km.

stands would be som ewhere between 45 and

90 km; say about 65 km ( 1000/r = 15, indicated
by the d o tted line across Fig. 2). There are 13
years in which hurricanes passed closer than that
to Discovery Bay, and it is instructive to look at
their distribution in time (Fig. 2, above the dotted
line). It is highly irregular; the intervals between
years o f potentially serious hurricane im pact
range from 1 to 36, with a m ean near 11, but a
m edian o f 6.5 (Fig. 3). T hat m eans wide varia­
tion, over the last 120 years, in the time available
for Acropora thickets to regenerate, before they
were knocked dow n again. F o r m uch o f the time,
Acropora rubble would have been as prevalent on
the north coast as it is today, or as it was at P ort
Royal in the 1950’s.
If we assum e th at Acropora palmata colonies
grow at 12 cm per year, it w ould take them 12
years to reach a height o f one m etre, growing up
at 45 ° . O f the 12 intervals between the 13 storm s,
only 3 exceed 12 years, and there have been only
33 years (since 1874) th at reefs at Discovery Bay
have been free o f m ajor disturbance for 12 years
or longer. 24 o f these (1956-1980) lie within the
single longest interval o f 36 years betw een 1944
and 1980. It w as during this unusual period, when
Acropora thickets flourished, th at m ost descrip-
12
10
8
6
4
2
10 20 30 40
Fig. 3. Histogram showing the number (vertical axis) of con­
tinuous periods o f years (horizontal axis) during which no
hurricane approached closer than 65 km to Discovery Bay, in
the period 1874-1988. Dotted line indicates median interval
of 6.5 years.
tions o f Jam aican north coast reefs were w ritten
(G oreau, 1959; K inzie, 1973; W oodley & R obin­
son, 1977). P erhaps they described an atypical
situation; one extreme o f a variable condition.
A nother perspective is provided by a com m ent
in Jam aica W eather R eport 685 (1934) on occa­
sions when Jam aica has been free o f hurricane
im pact for an unusually long time. In chronolog­
138
ical order these are 1655-1689 (34 years), 1786—
1812 (26 years), 1844-1874 (30 years) and 1917—
1933 (16 years). The next period w ould be 1951-
1980 (29 years) or, using the criterion of
‘traversing Jam aica’, 1944-1988 (44 years).
This analysis has left out any consideration of
differences in m agnitude between the hurricanes,
largely because such inform ation is not readily
available. It w ould be possible to gather d ata
about the im portant storm s, from the Jam aica
W eather R eports and from contem porary new s­
papers, to h indcast their im pacts on corals. Kjer-
fve et al. (1986) calculated th at turbulent flows of
2 -4 ms " 1 would bring dow n A . palmata colonies,
depending w hether or n o t they were bio-eroded.
Equivalent values for A . cervicornis were 0 .3 5 -
1.6 ms “ \ Only when th at analysis has been m ade
shall we know if the classic descriptions o f J a ­
m aican reefs are indeed atypical.
R e fere n ce s
Bretschneider, C. L. & E. Tamaye, 1976. Hurricane wind and
wave forecasting techniques. Look Lab/Hawaii 6: 1-29.
Goreau, T. F., 1959. The ecology of Jam aican coral reefs I.
Species composition and zonation. Ecology 40: 67-90.
Haii, M., 1889. Introduction. Jam aica Meteorological obser­
vations 1. Jam aica Government Printing Establishment,
Kingston.
Haii, M., 1911. Notes on the storms which passed over or
near Jamaica, between 1902 and 1910 inclusive. Jam aica
Meteorological Observations 4. Jam aica Government
Printing Office, Kingston.
Haii, M., 1916. Notes of hurricanes, earthquakes, and other
physical occurrences in Jamaica. Jam aica Weather Report
No. 465. In Jam aica Meteorological Observations 5. Ja ­
maica Government Printing Office, Kingston, 1918.
Hubbard, D. K., 1988. Controls of modern and fossil reef
development: common ground for biological and geological
research. Proc. Sixth Int. Coral Reef Symp. Townsville 1:
243-252.
Kinzie, R. A., 1973. The zonation of W est Indian gorgonians.
Bull mar. Sei. 23: 93-155.
Kjerfve, B., K. E. Magill, J. W. Porter & J. D. Woodley, 1986.
Hindcasting of hurricane characteristics and observed
storm damage on a fringing reef, Jamaica, West Indies. J.
mar. Res. 44: 119-148.
Knowlton, N., J. C. Lang & B. D. Keller, 1990. Case-study
of natural population collapse: post-hurricane predation on
Jam aican staghorn corals. Smithsonian Contributions to
the Marine Sciences 31, 25 pp. Smithsonian Institution
Press, Washington D.C.
Knowlton, N., J. C. Lang, M. C. Rooney & P. Clifford, 1981.
Evidence for delayed mortality in hurricane-damaged Ja ­
maican staghorn corals. Nature 294: 251-252.
Neumann, C. J., G. W. Cry, E. L. Caso & B. R. Jarvinen,
1981. Tropical Cyclones of the N orth Atlantic Ocean,
1871-1980. U.S. Departm ent of Commerce, NOAA.
Taylor, G. & Staff, 1981. Annual d ata and verification tabu­
lation Atlantic tropical cyclones 1980. National Hurricane
Center, NOAA tech. Mem. NW S N H C 15.
Woodley, J. D., 1989. The effects o f Hurricane Gilbert on
coral reefs at Discovery Bay. Appendix 9 in P. R. Bacon
(ed.), Assessment of the economic impacts of Hurricane
Gilbert on coastal and marine resources in Jamaica. U N E P
Regional Seas Reports and Studies 110.
Woodley, J. D., E. A. Chornesky, P. A. Clifford, J. B. C.
Jackson, L. S. Kaufman, N. Knowlton, J. C. Lang, M. P.
Pearson, J. W. Porter, M. C. Rooney, K. W. Rylaarsdam,
V. J. Tunnicliffe, C. M. Wahle, J. L. Wulff, A. S. G. Curtis,
M. D. Dallmeyer, B. P. Jupp, M. A. R. Koehl, J. E. Neigel
& E. M. Sides, 1981. Hurricane Allen’s impact on Jam ai­
can coral reefs. Science 214: 749-755.
Woodley, J. D. & E. Robinson, 1977. Field Guide to the
M odern and Ancient Reefs of J amaica. Atlantic Reef Com­
mittee, Miami.
 ■ÎSSÎ2
Hydrobiologia 247: 1 3 9 -1 4 0 , 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 139
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Short note

Importance o f the Lac-lagoon (Bonaire, Netherlands Antilles) for a

selected number of reef fish species

G . van der Velde,1 M. W. G o rissen ,1 C. den H arto g ,1 T. v a n ’t H o ff2 & G . J. M eijer1

1Laboratory o f Aquatic Ecology, Catholic University o f Nijmegen, Toernooiveld, 6525 ED Nijmegen, The
Netherlands; 2 Stinapa, Bonaire, Netherlands Antilles

M angroves and seagrass beds are know n to function as nursery for juveniles o f reef fishes. How ever,
the fish fauna o f these often linked ecosystem s is m ostly studied separately due to m ethodological
problem s. Therefore a selected num ber o f fish species w as studied both in Lac-lagoon (Bonaire, N eth ­
erlands Antilles) and the coral reef around Bonaire.
Lac-lagoon is the largest lagoon o f B onaire w ith an area o f about 8 km 2. It is a protected w etland
according to the R am sar convention. It consists o f a shallow basin with clear nutrient-poor w ater
bordered by m angroves and m udflats. T he bottom o f the lagoon is m ainly covered by seagrasses, es­
pecially Thalassia testudinum B anks ex König. T he investigations were focussed on the im portance of
the Lac-lagoon for the fishes on the reefs. T he reefs form the m ain attraction for the tourists visiting
Bonaire.

Table 1. Importance of various biotopes for juveniles of selected fish species and the importance of Lac-lagoon nursery for the
fish fauna of Lac-lagoon itself and the reef.

No. Species Juveniles Nursery

Mangroves Thalassia Shallow Deep Nursery Nursery

Reef Reef for Lae for reef

1 French grunt XX XXX X _ x x (x )

2 Bluestriped grunt - XXX - - XXX

3 Caesar grunt - - - - - -
4 Smallmouth grunt - - XX - - -
5 Black margate - - - - - -
6 Gray snapper XXX XX - - XXX (x )
7 Schoolmaster XXX XX - - XX X

8 Mahogany snapper - X XX - - X

9 Yellowtail snapper (x ) XXX - - - XXX

10 G reat barracuda XXX XX - - X XX

11 Sergeant major (x ) - XX - - X

12 Foureye butterfly XXX XX X X - xx(x)

13 Ocean surgeon - X XX - - X

14 Doctorfish (X ) XXX X - - X

15 Blue tang (x ) X X X - X

16 Stoplight parrotfish - XXX X X - XX

- not important
() doubtful
X minor importance
XX medium importance
XXX major importance
140

Sixteen species of reef fishes were selected based on easy identification and econom ic value (attrac­
tive for diving tourists, fishery) for a study o f abundance, distribution and size classes. D uring M a y -
N ovem ber 1981 abundance and size classes were estim ated by m eans o f a visual census technique;
estim ations in the seagrass beds were checked by catches w ith drop nets.
Six ‘biotopes’ have been studied, viz. seagrass beds (T . testudinum) and m angroves in the lagoon and
four depth zones on the reef. The m ajor results are sum m arized in Table 1.
M utual com parison indicates th at the seagrass beds form the m ost im portant biotope for the juve­
nile stages o f five fish species, viz. F rench grunt (Haemulon flavolineatum (D esm arest)), B luestriped grunt
(H. sciurus (Shaw )), Yellowtail snapper (Ocyurus chrysurus (Bloch)), Stoplight p arro t fish (Sparisoma
viride (B onnaterre)) and D o cto r fish (Acanthurus chirurgus (Bloch)), while m angroves are im portant for
four other species, viz. Schoolm aster (Lutjanus apodus (W albaum )), G ray snapper (L. griseus (L.)), G reat
b arracu d a (Sphyraena barracuda (B onnaterre)) and Foureye butterfly (Chaetodon capistratus (L.)).
F o r five species the lagoon seemed to be less im portant, viz. Sm allm outh grunt (Haemulon
chrysargyreum G ünther), M ahogany snapper (Lutjanus mahogoni (Cuvier)), Sergeant m ajor (.Abudefduf
saxatilis (L.)), O cean surgeon (Acanthurus bahianus C astelnau) and Blue tang (A. coeruleus Bloch &
Schneider): their juveniles mainly use the shallow reef. O f tw o species, C aesar grunt (Haemulon
carbonarium Poey) and Black m argrate (Anisotremus surinamensis (Bloch)), o f w hich adults were only
found on the reef, too few juveniles were found to draw conclusions.
 'IZShi.L
Hydrobiologia 247: 1 4 1 -1 6 1 , 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 141
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Hydrodynamics o f mangrove swamps and their coastal waters

Eric W olanski
Australian Institute o f Marine Science, P .M .B . No. 3, Townsville M .C ., Qld. 4810, Australia

Abstract

M angrove sw am ps help control the tidal hydrodynam ics o f m any tropical estuaries. They generate an
asym m etry o f the tidal currents in both the tidal creeks and the m angrove swam ps. This results in
self-scouring o f the tidal channels. M angrove land reclam ation results in siltation o f the channel. M an ­
grove sw am ps control the flushing rates o f the estuaries through the lateral trapping effect. Lateral
trapping leads to the aggregation o f m angrove litter along slick lines. E vapotranspiration plays a role in
the hot dry season by forming a salinity m axim um zone w hich isolates the estuary from the coastal w aters
for several m onths o f the year. In the absence o f runoff, évapotranspiration in the hot dry season gen­
erates an inverse estuarine circulation which can trap high salinity m angrove w ater, and m angrove de­
tritus, along the bottom o f a m angrove creek. This bottom layer can becom e anaerobic. G roundw ater
flow appears to play a key role in the nutrient budget o f m angrove creeks, exporting salt left behind by
évapotranspiration, and inhibiting runoff after rainfall. P articulates and dissolved nutrients outwelled
from m angrove sw am ps to coastal w aters are retained in a coastal boundary layer. This coastal boundary
layer w ater can be trap p ed along the shore for long periods if the coast is straight and mangrove-fringed
and the coastal w aters are shallow. H eadlands inhibit coastal trapping because they enhance mixing.
N utrient-rich coastal boundary layer w aters m ay be ejected offshore as tidal jets peeling off headlands
and locally enriching offshore waters.

Introduction 44 km long channel with 109 km 2 o f open w ater

A typical m angrove sw am p is C oral Creek
(Fig. 1). It is 5 km long, with fairly steep banks,
and m axim um depth decreasing with distance
from the coast. The cross-sections can be repre­
sented by polygons (Fig. lb). The fringing m an­
groves are typically 100 to 300 m wide on either
side o f the creek. T he surface o f the substrate
slopes gently upw ards from the creek, with ele­
vation near the creek o f about 0.3 m above m ean
sea level, and m axim um elevation rarely exceed­
ing 1.3 m.
W hen m any m angrove creeks exist one next to
the other, they form a large m angrove swamp.
H inchinbrook C hannel (Fig. 2) is an example of
such a vast m angrove swam p. The channel is
surrounded by 164 k m 2 of fringing m angrove
swam ps. Since the m angrove creeks draining the
swam p are oriented perpendicular to the C han­
nel, the sw am p is up to 5 km wide from the dry
land to the edge o f the Channel.
Tidal circulation in a creek-mangrove swamp
system
The barotropic tidal circulation in the C oral Creek
tidal creek-m angrove system w as first m odeled by
W olanski eta l. (1980). They noticed th at peak
tidal velocities often exceed 1 m s -1 in the tidal
creek bu t never exceeded 0.07 m s -1 in the
heavily vegetated m angrove sw am ps 50 m from

(b )
t' 1 9 m };t'
U p str e a m
Fig. 1. (a) M ap of Coral Creek, a mangrove creek in tropical Australia, with elevation in m. (b) shows two typical cross-sections
in the tidal creek.
the creek. They thus divided the system in its two
basic com ponent, firstly the tidal creek where ac­
celeration and inertia effects are im portant, and
secondly the m angrove swam p where they are
not. F o r m odeling purposes, the swam p w as di­
vided in a num ber of cells o f irregular shape and
size fitted to the topography o f the substrate
(Fig. 3a). The w ater circulation in the swam p was
calculated from the continuity equation, one such
equation for each cell i,
1 K ilo m e te r
100 m
H a lf w a y
where V¡ is the volum e o f w ater in cell i, QtJ is the
discharge (in m 3 s _1) betw een cells i and j , and
q¡ is the discharge from cell i direct to the tidal
creek. q¡ = 0 if cell i does no t touch the tidal creek.
Inertia and acceleration effects are assum ed to be
negligible in the m angrove swam p where the m o­
m entum equation reduces to a simple balance be­
tw een the frictional slope and the free surface
slope,
dq

ICARDW ELL
Point
Scraggy)
Point X cross-sectional area o f the creek, and S) the fric­
X X
a
km
LUCINDAi
Fig. 2. Map of Hinchinbrook Channel and its fringing giant
mangrove swamp, also in tropical Australia.
where t] is the elevation o f the w ater surface, x the
distance and Sf the frictional slope. Sf w as com ­
puted by M anning’s form ula and the M anning
coefficient w as set equal to 0.2 as characteristic of
flow through dense vegetation (Petryk & B osm a-
jan , 1975; W olanski e ta l., 1980; B urke &
S tolzenbach, 1983).
The tidal creek was divided in a num ber of
cross-sections show n in Fig. 3b. In the tidal creek,
acceleration and inertia effects are im portant and
the one-dim ensional barotropic equations o f m o­
tion for unsteady open channel flow are
continuity:
143
m om entum :
ÔQ 8 dri
~ T + T (2 2M ) + gA - Í + gA Sf + A = 0 (4)
at dx dx
where Q is the discharge, t is the time, A is the
tional slope calculated using a M anning’s rough­
ness coefficient n = 0.025 as characteristic o f open
channel flows. N ote th at in equation (4) is in­
cluded a term A. This term is the m om entum loss
due to w ater leaving the creek with its m om entum
at flood tide to enter the swam p. This m om entum
is rapidly dissipated in the first few m eters o f the
sw am p by friction in the heavily vegetated swamp.
A t falling tide, the m om entum o f the w ater leav­
ing the sw am p to flow in the tidal creek is so small
th at one can safely assum e A = 0.
The tw o m odels, one for the swam p and one
for the tidal creek, are linked by the condition that
the w ater levels are the sam e in the creek and in
a swam p cell where they touch each other. The
other condition th at the outflow from a m angrove
cell is the inflow into the tidal creek, and vice-
versa, is taken into account through the term q in
equations (1) and (3).
Since freshw ater runoff is negligible in Coral
Creek, the m odel needs only one open boundary
specification, nam ely a time series o f sea level at
the m outh.
The m odel w as verified against field observa­
tions o f currents at tw o points in the creek: the
m outh and a m id-channel point 2 km upstream .
Figure 4 shows a com parison between observed
and com puted tidal currents at the m outh during
a spring tide. N ote th at the peak tidal currents are
strong, being greater than 1 m s - 1 . Also there is
a large vertical shear o f the currents. Since the
w aters are vertically well-mixed in salinity and
tem perature, this shear can be attributed to b o t­
tom friction. The m odel cannot reproduce this
vertical shear since it is one-dim ensional (verti­
cally averaged). N evertheless the m odel appears
to be successful in reproducing the m ain features
 oL 1
JL j
2 km
Fig. 3. (a) shows the distribution of cells in the mangrove swamp. (b) shows the location of the cross-sections.
o f the depth-averaged currents, including the tidal
asym m etry, the peak ebb tidal currents being
m uch stronger th an peak flood currents. T he lat­
ter peak at 1 . 2 m s ‘ \ the form er at 1.6 m s “ 1.
This tidal asym m etry results in self-scouring o f
the tidal creek (W olanski et al., 1990). N um erical
experim ents suggest th at destruction o f the veg­
etation or reclam ation o f m angrove land, reduces
both the peak tidal currents and the asym m etry o f
the tidal currents, thereby resulting in siltation of
the channel.
Figure 5 show s the distribution o f the predicted
peak ebb and flood tidal currents in the m angrove
swam p. This figures illustrates another im portant
asym m etry of the tidal currents. The flood tidal
currents are small, typically o f order 0.03 m s - 1 ,
and are oriented at about 90 ° to the banks o f the
creek. On the other hand, the ebb tidal currents
are up to 3 times faster and are oriented at typ­
ically 20 to 30° from the creek. This asym m etry
in the tidal currents w as also confirmed in the
field through m easurem ents o f currents and cur­
rent directions at one point in the swam p. H ence,
over a tidal cycle there exists in the fringing m an­
groves a net circulation tow ards the m outh. The
im plications o f this circulation in the aggregation
o f m angrove litter are discussed later.
Groundwater flow
In the presence o f strong tidal currents such as at
C oral Creek, groundw ater flow probably does not
contribute m uch to the hydrodynam ics o f the sys­
tem. In other cases, groundw ater flow m ay be
im portant. Ovalle et al. (1990) have suggested that
the net outflow from the small Itacuruca m an­
grove sw am p in Brazil is due to groundw ater in­
flow in the swam p.
W olanski (unpublished data) and M az d a et al.
(1990a) have observed the initial stages o f the
flooding o f the sw am p at rising tide w hen a thin
sheet o f w ater advances in the sw am p from the
creek. This sheet is blocked by the high vegetation
density. The first few mm o f flooding w ater at a
given point are no t due to w ater from th at sheet
 145

200

O 150

100
1-

50 o.

•i
CD

ä O 0

-50

-100

-150
0 6 12 18 24
Tim e (h)

Fig. 4. Time series plot over one tidal cycle of the observed currents at different depths below the surface (from near the surface
to near the bottom) at the mouth of Coral Creek; the thick line shows the depth-averaged predicted velocity time series. The thin
line is the sea level time series.

spilling over obstructions on the ground, but are
observed to com e from the ground through the
num erous holes due to crabs and rotting vegeta­
tion.
W olanski & G ardiner (1981) have proposed
th at this groundw ater flow is an im portant flush­
ing m echanism preventing the accum ulation of
salt left behind after évapotranspiration.
M azd a et al. (1990) have carried out a study of
groundw ater flow in the 200 m long Bashita-
M inato m angrove sw am p in Japan. After storm s,
the tidal creek is occasionally ponded by a sill at
the m outh. They found th at when the creek is
ponded the w ater level in the creek decreases by
up to 10 cm d a y - \ about 15 times larger than the
evaporation rate. This observation suggests that
the creek w as draining out through groundw ater
flow. The flow appeared to be limited to the upper
90 cm o f the substrate, i.e. the m axim um depth of
crab burrow s, since below th at point the ground­
w ater was anaerobic.
The im portance o f crab burrow s in enhancing
groundw ater flow is apparent in C oral Creek and
in the Itacuruca swam p, where rainfall infiltrates
readily in the soil and produces no local runoff,
and in the K long N gao m angrove swam p in T hai­
land where crab burrow s are few and direct run­
off is observed (W attayakorn eta l., 1990).
Longitudinal diffusion in mangrove creeks
Salinity in m angrove creeks increases with in­
creasing distance from the m outh in the hot dry
season when runoff is negligible. The salinity in­
crease is due to évapotranspiration and is typi­
cally o f order 0.5 to 1 p.p.t. for 5 km long m an­
grove creeks such as C oral Creek and D ickson
146

20 cm s'
Velocity

03
CD

03
(a ) F lo o d ti d e
C
o
'</)
<o

0L i j2 km
(b) E b b ti d e

Fig. 5. Predicted peak tidal currents in the mangrove swamp of Coral Creek at flood and ebb tides.

Inlet in A ustralia (W olanski et al., 1980; R idd estuarine circulation exists on either side o f the
et al., 1990) and the Klong N gao m angrove creek salinity m axim um zone. In tropical A ustralia,
in T hailand (W attayakom et al., 1990). This sa­
linity gradient can introduce an inverse estuarine
circulation as is sketched in Fig. 6a. This circu­
lation becom es hydrodynam ic ally im portant and E s tu a ry
m easurable w hen both vertical and horizontal sa­
linity gradients are large, the existence o f a verti­ S alin ity m a x im u m z o n e

cal gradient is largely dependent on the strength

o f the tidal currents. In C oral Creek for instance
the tidal currents are particularly strong and
m aintain vertical hom ogeneity (Fig. 7), except
possibly at neap tides. In D ickson Inlet, vertical
gradients o f salinity are larger, particularly in the
upper reaches o f the tidal creek (Fig. 8a). The
resulting buoyancy effects inhibit vertical tu rb u ­
lent mixing.
In the presence o f a small runoff, a salinity
maxim um zone can develop, which, as sketched
in Fig. 6b, essentially isolates the upper reaches
o f the estuary from the coastal waters. An inverse
O cean
R iv e r e s tu a r y
S alin ity m a x im u m z o n e
Fig. 6. (a) sketches the inverse estuarine circulation gener­
ated in a mangrove creek by évapotranspiration in the fring­
ing swamp in the absence of runoff. (b) sketches the internal
circulation generated by the ‘p resence of a salinity maximum
zone in the presence of a vast evaporation area, such as a vast
mangrove swamp, and of a small freshwater runoff.
 147

(c ) S ite A D is ta n c e S ite B

C oral C reek
S ite B
m angrove
S ite A sw a m p

o 7 2

2 km

(b ) 1

E
Q_
CD
a
0
7

o
5
T im e (h o u r s )

Fig. 7. Longitudinal distribution of salinity in Coral Creek in the hot dry season.

S a li n i t y ( % » )
S ea Head

D is o lv e d O x y g e n (m g M )
a> u
a

1000 m

Fig. 8. (a) shows the salinity distribution in the Dickson Inlet mangrove creek in the hot dry season, at low tide, (b) shows the
corresponding distribution of dissolved oxygen, also at low tide.

such an unusual estuarine circulation can last In the presence o f an inverse estuarine circu-
several m onths (W olanski, 1986; R idd eta!., lation, high salinity m angrove w ater is then
1988). trapped in the creek at the bottom and is not
148
aerated. The decom position o f the m angrove d e­
tritus in th at layer and the biological oxygen de­
m and o f the m ud, can result in low values o f the
dissolved oxygen concentration (less than
1 mg 1” 1 in D ickson Inlet, Fig. 8b). In som e cases
anoxic conditions can result, such as occasionally
in C oral Creek (Boto, personal com m unication)
and in the B ashito-M inato m angrove swam p
(M azd a et al., 1990).
The value o f the longitudinal diffusion coeffi­
cient, B, hence also the flushing rates, m ay be
estim ated from the longitudinal salinity gradient.
Since the gradient reaches a steady state, the salt
accum ulated in the sw am p/creek system is ex­
ported by tidal diffusion, so that
dS
A ,E ,S = BA — (5)
dx
w here S is the salinity, E t the évapotranspiration
rate (0.2 to 0.6 c m d _1) , 4 ( the surface area o f the
swam p, and A is the creek cross-sectional area.
U sing observed salinity gradients in C oral Creek
and the Klong N gao estuary, equation (5) yields
for these tw o systems
B = 20 - 50 m 2 s - 1 (6)
This value o f B is surprisingly large for such small
tidal creeks. Indeed, if diffusion was due to ver­
tical turbulent mixing in the creek, then (Bow den
& H am ilton, 1975; Fischer e ta l., 1979) the tu r­
bulent eddy diffusivity (B, now called K to dif­
ferentiate observed from predicted values) would
be
K = CdhU (7)
where Cd is a drag coefficient, h is the depth and
U a characteristic tidal velocity. If vertical m o­
tions were dom inant, then C d = 0.003. It results
K = 0.03 m 2 s ” 1 for C oral Creek and the Klong
N gao m angrove creek, a result 1000 times too
small. This implies th at verticle turbulent m otions
are n o t the leading mixing m echanism in m an ­
grove creeks.
Alternatively, following Fisher etal. (1979)
K = d .\5 h u , (8)
where u* is the shear velocity.N okes (1986)
found that
u m = 0.05 U (9)
so th at equation (8) becom es
K = 0.0075 hU (10)
T he value o f K from equation (10) istwice that
from equation (7). This results in K = 0.06 m 2 s “ 1
for C oral Creek and the K long N gao m angrove
creek, a result still 500 times too small.
A nother mixing m echanism m ay be vertical
shear dispersion. This process w as first reported
by Taylor (1921) for flows in pipes, and has been
applied to open-channel flow by Elder (1959) and
F ischer et al. (1979) who proposed tw o alterna­
tive form ulae for diffusion in oscillatory currents,
K = 0.1 Uh
or
K = 5.93 hu, ~ 0 . 3 Uh (11)
which for C oral C reek and the K long N gao m an­
grove creek yields K = 0 .5 -1 .0 m 2 s " 1 a value still
at least 20 tim es too small. This suggests th at
longitudinal diffusion in m angrove-fringed tidal
creeks is controlled by another m echanism . This
m echanism m ay be lateral trapping due to lateral
shear. Fischer eta l. (1979) proposed th at in a
wide, shallow, tidal estuary lateral shear is m uch
m ore efficient than vertical shear at generating a
large longitudinal diffusion. H ow ever, their m odel
is no t readily applicable to m angrove creeks as it
requires a spatially hom ogeneous lateral shear.
M ore detailed studies o f this m echanism are dis­
cussed below.
Lateral trapping in mangrove swamps
The previous analytical m odels o f the longitudi­
nal diffusion coefficient in a m angrove creek ne-

gleet the presence o f the fringing m angrove
swam p. In fact this m angrove system is the rea­
son for the existence o f a large longitudinal eddy
diffusion coefficient in m angrove creeks. The pre­
vailing m echanism is lateral trapping, sketched in
Fig. 9. This m echanism w as first studied by
O kubo (1973) who m odeled dispersion in a tidal
estuary with a lateral em baym ent. A tagged w ater
m ass m oves upstream at flood tide, bu t a fraction
o f th at m ass ends trapped in the em baym ent. At
ebb tide, this m ass returns to the tidal creek and
mixes with untagged w ater. Mixing is thus greatly
enhanced and, following O kubo (1973),
A values o f the eddy diffusion coefficient are pre­
B = -------+ ------------------- ---------------- (12)
1 + s 2 k (1 + e)2 (1 + e + o/k)
where u0 is the peak tidal current, k~ 1 is the
characteristic exchange time betw een the em bay­
m ent and the estuary, eris the tidal frequency and
£ is the ratio o f volum e in the em baym ent to that
in the estuary.
W olanski & R idd (1986) m odified the O kubo
(1973) m odel to rem ove tw o assum ptions th at are
invalid in a m angrove creek, nam ely th at the
depths o f w ater in the estuary and in the swam p
are the sam e and are m uch larger than the tidal
range, and th at the exchange rate between w ater
in the swam p and in the estuary are constant in
time. They divided the m angrove creek in a rect­
angular m ain channel where tidal currents occur,
and a sw am p seen as the lateral em baym ent. The
elevation o f the swam p substrate w as taken to be
higher than m ean sea level so th at the swam p has
L a te ra l e m b a y m e n t
E s tu a ry
- ----------------------- —--------- .---------
Fig. 9. Sketch of the lateral trapping phenomenon in a tidal
estuary with an embayment, adapted from Okubo (1973).
149
no surface w ater for a fraction (1-a) of the tidal
cycle o f period 2T. They predicted
A su 2a 2 T
B = ---------+ --------------- (13)
(1 + £) 48 (1 + £)
E quation (13) yield B = 10-40 m 2 s - 1 for C oral
Creek, D ickson Inlet and the K long N gao estu­
ary, values th at agree qualitatively well with o b ­
servations. Equation (13) w as recently modified
by R idd eta l. (1990) to take into account the
decrease in peak tidal velocity from the m outh to
the upper reaches o f the m angrove creek. Smaller
dicted in the upper reaches o f the creek particu­
larly in the upperm ost 200 m or so.
Lateral trapping in m angrove sw am ps thus ap ­
pear to be the dom inant effect determ ining their
flushing rate. T he residence time in mangrove
creeks is o f order
T0 = L 2/B (14)
where L is the length o f the mangrove-fringed
creek. F or C oral Creek and the K long N gao m an­
grove creek, it results T0 = 6 days, suggesting that
the upper reaches o f m angrove creeks are flushed
slowly.
W olanski eta l. (1990) also applied equations
(13) and (14) to estim ate the values o f B and T0
in H inchinbrook C hannel. They found T0 = 50
days. This implies long-term trapping in this giant
m angrove swam p. In fact they found experim en­
tal evidence for this long-term trapping in the
flushing o f brackish w ater from the channel fol­
lowing cessation o f runoff. The salinity deficit fol­
lowed an exponential decay in time, with an e-fold
time scale o f 54 days, in surprisingly good agree­
m ent with the predictions.
Buoyancy effects in mangrove swamps
U ncles eta l. (1990) studied the dynam ics o f a
lärge, m angrove-fringed, estuary in M alaysia,
This estuary receives a large fairly steady fresh-
150

I«-tidal river -*(«.---------- mangrove swamp
2S
low salinity w ater
isohalines
(a ) H ig h tid e
.:M
'N>m ix in g ^ /S low salinity w ater
(b ) L o w tid e
Fig. 10. Cross-channel distribution of salinity in a mangrove-
fringed estuary at high and low tide following recovery of the
system from a major river flood.
w ater inflow. They found th at the salinity strati­
fication is strong during neap tides but th at the
system destratifies during spring tides. W ater
flooding the m angroves is thus fresher at neap
tides and saltier at spring tides. This m ay intro­
duce a cycle, with a period o f 2 weeks, in the
productivity o f the fringing m angrove forest.
W attayakorn etal. (1990) show ed th at the
small m angrove creek o f K long N gao in T hailand
stratifies and destratifies at a few hours period
closely following fluctuations in b oth the tidal cur­
rents and in the freshw ater inflow which varied
rapidly w ith local short-lived storm s in the small,
steep w atershed draining into th at swam p. This
unsteadiness precluded a reliable estim ate o f the
outwelling, if any, in the wet season.
In tropical A ustralia, river runoff can be very
intense bu t short-lived. Salt w ater can be totally
flushed out o f the estuary during floods. As soon
as the flood is over, salt w ater creeps back in the
estuary. If the estuary is m angrove-fringed, such
as the W enlock River in A ustralia (W olanski &
R idd, 1986), the intruding salt w ater intrudes
under the fresh w ater rem aining near the surface.
As the tide rises and salt w ater fills the estuary,
the freshw ater is pushed back laterally into the
fringing m angrove sw am ps (Fig. 10). This fresh­
w ater is then trapped there, with m inim al mixing

Fig. 11. Photograph showing an oceanographic front separating swamp water from creek waters, and made visible by a foam line.

with salt w ater in the estuary. As the tide falls, this
freshw ater returns to the estuary but stays trapped
as a narrow river plum e in shallow w aters along
the length o f the estuary banks (Fig. 10). In the
W enlock River estuary, this front m aintained its
integrity throughout the ebb tide. H ence there was
minim al mixing betw een fresh w ater in the m an­
grove swam p and sea w ater in the estuary.
The lateral trapping m echanism prevails also in
m angrove creeks and is occasionally m ade readily
visible by the presence o f an oceanographic front
forming at the start o f the ebb tide. This front,
shown in Fig. 11, separates swam p w ater from
creek w ater. In small m angrove creeks, this front
disappears in less than an hour but it indicates
th at the creek is n o t laterally well-mixed.
The front forming between creek and swam p
w ater is im portant as it aggregates floating detri­
tus, such as m angrove leaves. Visual observations
reveal th at at C oral Creek the flood tidal currents
are too w eak to break the organic film at the
surface o f the w ater in the m angrove swam p.
Floating m angrove leaves are often trap p ed in
this film and the flood tide does not carry the
H ead
32,
33.5
Q_
CD
Q
/'
High tid e
2
C'fiK.
Low tide
Fig. 12. Longitudinal distribution in Dickson Inlet at high and low tide about 10 days after cessation of runoff.
151
floating m angrove leaves further inland in the
swam p. A t ebb tide, the currents are stronger,
break the surface film and export floating leaves
to the creek. The leaves are then aggregated in
long straight lines (foam lines) m arking the small-
scale oceanographic fronts separating m angrove
swam p w ater from creek w ater (Fig. 11). These
leaves can be outwelled from the m angrove creek
to coastal w aters where they can som etimes be
found, still aligned in a straight line, several km
from the swam p in calm weather.
H ence, the fate o f m angrove detritus cannot be
estim ated using classical advection-diffusion
m odels since these m odels predict th at the detri­
tus would be dispersed in a cloud o f dim ension
increasing in time. The form ation o f small scale
barotropic and baroclinie fronts in m angrove
sw am ps has not been studied. These front are
im portant since they result for a depth-averaged
m odel in an apparent negative eddy diffusion as
the m angrove seeds and leaves are n o t dispersed
but are instead aggregated. This m ay have im por­
tan t biological consequences. F or instance one
often sees small fish aggregating under the leaves
D is ta n c e (km ) S ea
^ __ ^ 33
y? .?
1
Aä Æ t'
-m
152
floating in a sm all-scale front in a m angrove creek.
W olanski & H am ner (1988) docum ented several
sm all-scale m echanism s aggregating floating coral
eggs. T hese m echanism s are due to the three-
dim ensional circulation in topographically con­
trolled fronts. M ost o f the examples illustrated for
reefs can be applied to m angrove creeks. F o r in­
stance, leaves aggregate dow nstream o f a sharp
bend in the m angrove creek even in barotropic
conditions as a result o f the secondary circulation
in a zone o f strong flow curvature, in the same
m anner as coral eggs are aggregated in zones of
strong flow curvatures near coral reefs. Axial con­
vergence (Sim pson & Turrell, 1986), due to the
transverse tran sp o rt in estuaries, also appears to
be a process m aintaining the floating leaves and
m angrove seeds in a foamline, especially n ear the
m outh o f m angrove creeks.
A com bination o f lateral trapping and buoy­
ancy effects can result in very long residence time
o f brackish w ater in m angrove creeks. Figure 12
shows the longitudinal distribution o f salinity in
m angrove-fringed D ickson Inlet at both low and
high tide. B rackish w ater, a rem nant o f the ru n ­
off 10 days earlier, w as still present near the sur­
face in the creek at low tide. A t high tide, the bulk
o f th at w ater w as trapped in the swam p. Because
buoyancy effects inhibit vertical mixing, the
brackish w ater is basically simply moving back
and forth betw een the swam p at high tide and the
creek and low tide, but mixes very little with creek
w ater and is thus only very slowly exported from
the system.
C h w a k a B ay M a n g ro v e frin g e d tidal c re e k
o
30.5
.29.5'
Sill
34.5
2
S alin ity (% o )
4
Fig. 13. Longitudinal salinity distribution in the Chwaka bay
mangrove swamp in Zanzibar, in the dry season, July 1988.
The presence o f a sill or narrow opening in­
creases trapping. F o r instance, the m angrove
creek o f C hw aka Bay in Z an zib ar is nearly p o n ­
ded for a m easurable portion o f the tidal cycle by
a small coral lim estone sill at the m outh. Strong
tidal mixing over shallow w ater o f the sill results
in the form ation o f a front separating creek w a­
ters from oceanic w aters (Fig. 13). T he sill thus
inhibits mixing betw een creek w aters and oceanic
w aters, and ensures long-term trapping in the
m angrove-fringed tidal creek (W olanski, 1989).
Trapping in the coastal zone
Straight coastline
M aterial outwelled from m angrove sw am ps m ay
support offshore productivity but, if the coastal
w aters are shallow, this enrichm ent is restricted
to either a narrow coastal zone or specific points
near headlands.
W hen runoff is im portant, the outwelled m ate­
rial rem ains in the river plum es. In areas where
river runoff is always im portant, such as on the
m angrove-fringed coast o f T hailand facing the
A ndam an Sea near B urm a, longshore currents
ensure th at these river plum es usually rem ain
coastally trap p ed together with the outwelled
m angrove detritus and nutrients (e.g. W attaya-
korn eta l., 1990). Any nutrient enrichm ent from
the m angrove is then limited to coastal waters.
Sim ilar coastal trapping o f steady river plum es
for hundreds o f km has been reported in N icara­
gua (M urray & Young, 1985) and in South A m er­
ica (Bouyne & Roux, 1978).
In tropical A ustralia, river runoff is short-lived.
F reshw ater is injected into coastal w aters in a
pulse-like event. O nce runoff ceases, tidal cur­
rents m aintain vertical hom ogeneity in shallow
coastal w aters, and the brackish w aters behave
m ore or less like a passive tracer. T he fate o f this
tracer, i.e. the time it takes to be flushed out from
coastal w aters, provides a good indication o f the
dynam ics o f the coastal zone. Tw o examples are
illustrated below: the G ulf o f C arpentaria and the
coast o f the G reat B arrier Reef.

Figure 14 shows the distribution o f salinity in
the G u lf o f C arpentaria at 3 m onthly interval in
1976. The survey in A p ril-M a y corresponds to
the end o f the wet season. It shows th at the
brackish w ater w as coastally trapped and did not
mix readily with offshore waters. Five m onths
after, in A ugust-S eptem ber, runoff was negligible
but brackish w ater w as still found in the shallow
coastal w aters (depth less than 15 m or so). This
finding suggests th at brackish w aters is trapped
for several m onths in a coastal boundary layer,
and th at cross-shelf mixing between offshore w a­
ters and coastal boundary layer w ater is very slow
(W olanski & R idd, 1990). A satellite image in the
visible b ands taken in calm w eather (no wind
stirring o f bottom sedim ent) in the very dry sea­
son o f N ovem ber, 1986 (zero runoff hence no
river plum e) also shows a band o f w ater o f dif­
ferent colour in the dry season (negligible runoff)
along the co ast o f the G ulf of C arpentaria, with
a front running along the 15-20 m isobath
(Fig. 15). The presence o f a coastal boundary
layer m ay be explained by the effects o f friction in
shallow w ater. Figure 16 shows the predicted
m ean Eulerian circulation under the action o f both
the tides and wind in the tradew ind season. This
shows a narrow coastal jet extending to the 15 to
20 m isobath w ith a return flow in deeper water.
T he tides alone drive a negligible net circulation
even in shallow coastal w aters, while wind alone
in the absence o f tides drives a coastal jet about
x S u r f a c e s a lin ity ( 7oo)
G u lf o f C arpentaria
23 April -1 1 M ay 2 7 J u n e - 2 July
Fig. 14. Salinity distribution in the G ulf of Carpentaria at about three-monthly interval in 1976. These data were kindly provided
by P. Rothlisberg.
153
3 times as wide as th at shown in Fig. 16 (W olan­
ski & R idd, 1990). The non-linear interaction be­
tween tidal currents and wind-driven currents is
an essential factor in the dynam ics o f the coastal
boundary layer.
Rugged coastline
The east coast o f the G ulf o f C arpentaria is fairly
straight, with no m ajor headlands, and extensive
areas o f shallow w aters, the 20 m isobath being
located 20 km offshore. In contrast, the G reat
Barrier R eef continental shelf has num erous
headlands (Fig. 17) and the w aters are deeper, the
20 m contour line nearly touching these head­
lands. The salinity distribution in coastal w aters
w as m easured after the flooding o f the Burdekin
River and other smaller stream s in January 1981.
This distribution is shown in Fig. 17 for 2 6-27
January 1981, about 3 days past the peak o f the
flood and for 3 - 4 February 1981, 10 days after
the flood. N ote th at in the first survey the plume
m oved northw ard but w as continuous over
200 km and coastally trapped with the coast on
its left-hand-side, as dictated by Coriolis forces,
though offshore the current was southw ard
(W olanski & van Senden, 1983). In the second
survey, the plum e w as broken up in patches.
Brackish w ater rem ained preferentially in the wide
bays sheltered by headlands.
154
o f C a r p e n ta r ia ®
Fig. 15. NOAA satellite view in the visible bands of the east
and south coasts of the Gulf of Carpentaria on 23 November
1986, during the hot dry season, in calm weather.
M ixing and dispersion near headlands
These d ata imply th at coastal trapping is m uch
less efficient on the G reat B arrier R eef coast than
in the G u lf of C arpentaria. This difference m ay be
attributed to the headlands and reefs on the G reat
Barrier R eef continental shelf, these obstructions
to flow being absent in the G ulf o f C arpentaria.
Flow around headlands and islands generate ed­
dies (Pingree & M addock, 1979; W olanski et al.,
1984; F alconer eta l., 1986; P attiaratchi eta l.,
1987; G eyer & Signell, 1990). The tw o m ost in-
Fig. 16. Predicted mean circulation in the Gulf of Carpentaria
under the combined influence of all tides and a 15 m s -1
southeasterly trade-wind.
tensive field studies o f currents around a h ead­
land or an island in shallow w aters have been
those o f W olanski eta l. (1984) and G eyer &
Signell (1990). W olanski et al. (1984) docum ented
the presence o f energetic island w akes from the
d ata from 24 current m eters, float m easurem ents,
C TD studies and aerial photography. G eyer &
Signell (1989) m easured the circulation around a
headland using a ship-m ounted acoustic D oppler
current profiler. Figure 18 shows an example o f a
fully developed, energetic headland eddy as m ea­
sured with this latter technique.
W olanski eta l. (1984) and F alconer eta l.
(1986) have proposed respectively analytical and
num erical m odels th at attem pt to explain these
eddies in term s o f vorticity dynam ics at respec­
tively steady and unsteady cases. Their m odels
rely strongly on flow separation dynam ics at the
separation points at the island’s tips. Black &
G ay (1977) criticised the vorticity m odel and
claim ed th a t the eddy strength w as not prim arily
governed by flow separation, bu t is a ‘phase eddy’
form ed as a result o f the oscillatory pressure gra­
dient, which near the end o f the half cycle, re­
verses the frictionally controlled flow near the
 155

¡34
B arro n R iver
C a irn si

Jo h n sto n e
R iv e r i

Coral Sea

c u r r e n t m e te r
tr a n s e c t

'» » » •B rito m a rt R e e f
•“i -m* -
2 6 - 2 7 Jan u ary , <s?’ 3 - 4 F e b ru a ry ,

T ow nsville,

B urdekin
^Hiverw

Fig. 17. Surface salinity distribution on the Great Barrier Reef continental shelf between the mouth of the Burdekin River and
Cairns, following flooding of the Burdekin River and other smaller streams, in 1981.

— *• 5 0 c m s _1
S la c k h ig h w a te r

Gay Head

2 km

Fig. 18. Observations of the velocity field and evidence for a strong eddy behind G ay Head, M assachusetts, USA, adapted from
Geyer & Signell (1990).
156
coast before the freestream flow. How ever, Signell
(1989) carried out the m ost detailed, sm all-scale,
num erical studies so far o f headland w akes, and
show ed th at the vorticity which form s transient
eddies is produced near the headland tip and that
flow separation is indeed critical to tran sp o rt this
vorticity into the interior o f the eddy.
D eleersnijder eta l. (1989) have developed a
three-dim ensional m odel o f the tidal w akes and
show ed that, though the eddy dynam ics are es­
sentially tw o-dim ensional (independent o f depth),
there exist nevertheless vertical m otions, with up-
welling near the center o f the eddy and down-
welling along the island sloping boundaries, as
show n in Fig. 19.
Signell (1989) also show ed the im portance o f
adequate grid resolution near the headland tip.
H e show ed th at a lack o f sm all-scale resolution
in th at area leads to an underestim ate o f the wake
strength (Tee, 1976). This is not simply o f aca­
dem ic interest, but has profound im plications for
m odeling the fate o f w ater-borne m aterial. The
inclusion, or the neglect, o f flow separation effects
in num erical m odels lead to quite different results.
W hen flow separation is included, patches o f co n ­
tam inants in the vicinity o f the headlands are
strongly deform ed by straining associated with
the separated shear layer and transient vortices
(Signell, 1989). As is shown in Fig. 20, the dis­
tribution o f m aterial is extremely patchy and
streaky for those contam inants entrained in the
free shear layers, while those contam inants not
Fig. 20. Numerically predicted dispersion around a headland with reversing tidal currents in the absence of any net currents. Two
patches of particles are released at the beginning of the eastward flowing tide, one just off the tip of the headland and the other
some distance offshore (a). (b) shows the distribution after 6 tidal cycles, and (c) after 12 tidal cycles. Adapted from Signell &
Geyer (1990).
R atray
E
£ 0.5
È
1.0
,5 0 0 m i
Side view
Fig. 19. Numerical predicted three-dimensional circulation
around Rattray Island, Australia, after eddy spin-up. This
figure shows the velocity field in a vertical plane oriented along
the mean flow direction and intersecting the middle of the
island. Adapted from Deleersnijder etal. (1989).
entrained in the free shear layers are essentially
carried through with m inim al mixing. Figure 20
dem onstrates th at the large strain rate in the re­
gion o f flow separation tends to stretch particles
o f fluid into long filam ents, w hich are subsequently
rolled and distorted by the transient eddy field
(G eyer & Signell, 1990; Signell & Geyer, 1990).
The intense shear associated with the flow sep­
aration at the tip o f the headland, is the agent
responsible for the nearly explosive dispersion of
the patch th a t started close to the headland in
Fig. 20. Evidence for such patchiness and streak­
iness is found in field experim ents o f dispersion
around coral reefs (e.g. W olanski & H am ner,
1988; W olanski e ta l., 1988; W olanski & King,
1990). O n the other hand, m odels th at neglect
 157

x
H (x ) Hf
h (x)

) -- v e r tic a l
e n tra in m e n t

F lo o d ti d e E b b tid e

Fig. 21. Upwelling in a tidal jet at (a) flood tide and (b) ebb tide.

these flow separation dynam ics, such as the The inclusion, or no inclusion, o f separation
‘phase eddy’ m odel, are unreliable because they flow effects in num erical m odels, is not ju st a
result in sm ooth distributions from gradient-type question o f academ ic interest only. Indeed, the
diffusion since they do no t sim ulate straining and num erical m odels used at present to sim ulate the
folding o f vortex lines. dispersal of crow n-of-thorn starfish in the G reat
(a )
Surface Surface

1 m S'1 Flood tide

Bottom

Fig. 22. High-frequency echo-sounding transect at the mouth

of a reef passage between Ribbon Reefs Nos. 2 and 3, Great
Bariier Reef, at (a) flood tide and (b) ebb tide.
158
Barrier Reef, a problem o f great practical and
m anagem ent im portance, neglect these dynam ics
and predict sm ooth distributions and long reten­
tion times near reefs (e.g. Black, 1988). Such
m odel predictions are contrary to expectations if
straining and folding o f vortex lines were included
in the models.
A nother im portant mixing m echanism is th at
o f tidal jet-vortex pair system, the dynam ics o f
w hich is also intim ately linked with flow separa­
tion effects (Awaji eta l., 1980; Awaji, 1982; van
Senden, 1985; W olanski et al., 1988). Awaji et al.
(1980) and Awaji (1982) found th at the pattern of
tran sp o rt and dispersion varies greatly over short
length scales and, exactly as in the case o f island
wakes, does n o t display the sm earing tendency of
a typical diffusive process.
Usually m angrove-fringed coastal w aters are
sufficiently shallow that, through friction (Ozsoy,
1977), tidal jets do no t prevail at ebb tides at the
m outh o f tidal creeks (W olanski & R idd, 1990).
The frictional slope term balances the surface
slope term in the continuity equation, and thus
the flow resem bles potential flow. However, there
exist som e m angrove creeks, e.g. T udor Creek in
M om basa, Kenya, th at face deep water. Com plex
and biologically im portant flow patterns probably
exist at such places and m ay resem ble th at ob­
served in front o f reef passages through the R ib­
bon Reefs, G reat B arrier R eef o f A ustralia (T h­
om pson & Golding, 1981; W olanski et al., 1988).
A t flood tide, forced upwelling m ay occur as a
result o f the strong tidal currents at the m outh, as
is sketched in Fig. 21a. This upwelling is readily
visible in high-frequency echo-soundings, the
scatterer probably being plankton (Fig. 22a). At
ebb tide, a tidal jet m ay also develop, lifting off the
b ottom as a result o f buoyancy effects, and re­
sulting in upw ards entrainm ent o f deep w ater in
the je t (Fig. 21b). High-frequency echo-soundings
(Fig. 22b) also confirm the existence o f such a
m echanism .
Though a rugged coastline prevents the form a­
tion o f a long, stable coastal boundary layer, tra p ­
ping can occur in the shallow em baym ents shel­
tered by headlands such as Bowling G reen bay,
a wide, shallow em baym ent near Townsville,
A ustralia (Fig. 23). T he prim ary reason for this is
that, because o f friction, neither headland eddies
nor the longshore currents found further offshore
are generally able to propagate in these very shal­
low, sheltered, w aters (W olanski & R idd, 1990).
T rapping is reinforced w hen the em baym ents
are m angrove-fringed. Such is the case for in­
stance at the northern region o f H inchinbrook
C hannel where the w est coast is m angrove-fringed
and very shallow (Fig. 24a). W ater moves back
and forth with the tides betw een the m angrove
sw am p and the coastal boundary layer in shallow
Bowlini
G ro a n /
; M o d e l s h o re lin e
A c t u a l s h o re lin e
m ilii
mm
f t t i t l í
B ow ling G r e e n B ay ' tidal w e tla n d s;
ViVtWt't) ‘i ...............................i i
íí t IV i i
I
f t t i t ft ft ft t
tritt ft ft tit
ttttttY/yfff
IJ.t.tJ,/// / / / / ,
C a p e F e rg u so n
Fig. 23. N et circulation in Bowling Green Bay, a shallow
mangrove-fringed embayment near Townsville, Australia, in
calm weather. Tides and the southward flowing East A ustra­
lian Current generate strong currents offshore. The bathym­
etry is in metres. The bay has extensive tidal wetlands. To
make the very small nearshore currents visible on this graph
together with the strong offshore currents, the length of the
velocity arrows is proportional to the logarithm of the water
currents. As a result a factor of 10 difference in the water
velocity implies a difference by a factor o f 2 in the length of
the arrows.

CARDWELL
Oyster Point)
O y s te r
Scraggy
V Point
(b) L o w tid e
Fig. 24. (a) Map of the northern region of Hinchinbrook Channel with depths in m. (b) and (c) show the plume, at low and high
tide, of a tracer after 15 days of continuous discharge at point A in the upper reaches of a mangrove creek in calm weather.
w aters on the w est coast. The coastal boundary
layer w aters being very shallow, there is little
cross-shelf mixing between coastal boundary
layer w aters and w aters further offshore in the
deeper parts of the channel. W olanski et al. (1990)
have m odeled the w ater circulation in this area.
The m angrove sw am ps were considered to be
inter-tidal storage areas. The m odel results were
verified against field observations o f currents. A
tracer was released continuously for 15 days at
point A in the m odel, in the upper reaches o f a
m angrove creek, and followed through time. Fig­
ure 24 (b and c) shows the tracer’s plum e at both
low tide and high tide after 15 days in calm
w eather. T he co ntam inant is found in shallow
coastal w aters at low tide, and the bulk o f this
contam inant is found trapped in the m angrove
sw am p at high tide.
T he presence o f a coastal boundary layer in
shallow, m angrove-fringed, coastal w aters, pre­
sents problem s in interpreting outwelling studies
based on a budget o f m aterial exported over a
tidal cycle by a m angrove creek (e.g. B oto & Bunt,
1981) since w ater entering a m angrove creek in
159
(c) H ig h ti d e
such areas m ay simply have left another nearby
creek at the previous tidal cycle and is chemically
different from offshore water. Thus a m angrove
creek at the upstream side o f a coastal boundary
layer will have an apparent lower bulk return co­
efficient than one at the dow nstream end.
U nder the action o f w ind and tides, coastal
boundary layer w aters, enriched in nutrient by
outwelling from the tidal w etlands, can be ejected
offshore as tidal jets peeling off capes and h ead­
lands (B ow m an, 1988; W olanski & R idd, 1990).
This process m ay explain the presence o f baitfish
and billfish schools found a few km offshore from
C ape Bowling G reen (Williams, 1988).
References
Awaji, T. A., N. Imasato & H. Kunishi, 1980. Tidal exchange
through a strait: A numerical experiment using a simple
model basin. J. Phys. Oceanogr. 10: 1499-1508.
Awaji, T., 1982. Water mixing in a tidal current and the ef­
fect ot turbulence on tidal exchange through a strait. J.
Phys. Oceanogr. 12: 501-514.
Black, K. P., 1988. The relationship of reef hydrodynamics to
variation in number of planktonic larvae on and around
160
coral reefs. Proc. 6th. Int. Coral Reef Symposium, Towns­
ville, Vol. 2.
Black, K. P. & S. L. Gay, 1987. Eddy formation in unsteady
flows. J. geophys. Res. 92: 9514-9522.
Boto, K. & J. S. Bunt, 1981. Tidal export o f particulate or­
ganic matter from a northern Australian mangrove system.
Estuar. coast. Shelf Sei. 13: 247-255.
Bowden, K. F. & P. Hamilton, 1975. Some experiments with
a numerical model of circulation and mixing in a tidal es­
tuary. Estuar. coast. Mar. Sei. 3: 281-301.
Bowman, M. J., 1988. Estuarine fronts. In B. Kjerfve (ed.),
Hydrodynamics of estuaries, Vol. 1, CRC Press, Boca
Raton, Florida, 85-132.
Bouyne, P. & J. D. Roux, 1978. Remarques sur la circulation
d ’eaux turbides d’origine amazonienne le long de la cote des
Guyanes. C.r. Acad. Sei. Paris, 287, Serie D.: 203-205.
Burke, R. W. & K. H. Stolzenbach, 1983. Free surface flow
through salt m arsh grass. Massachusetts Institute o f Tech­
nology Sea G rant College Program, Publication No.
M ITSG 83-16, Cambridge, MA., 252 pp.
Deleersnijder, E., E. Wolanski & A. Norro, 1989. Numerical
simulation of the three-dimensional tidal circulation in an
island’s wake. In G. M. Carlomagno & C. A. Brebbia (eds),
Computer and experiments in fluid flow. Computational
Publications, Springer-Verlag, Berlin: 355-381.
Elder, J. W., 1959. The dispersion of marked fluid in turbu­
lent shear flow. J. Fluid Mech. 5: 544-560.
Falconer, R. A., E. Wolanski & L. M ardapitta-Hadjipandeli,
1986. Modelling tidal circulation in an island’s wake. J.
Waterway, Port, Coastal and Ocean Engineering, ASCE,
112: 234-254.
Fischer, H. B., E. J. List, R. C. Y. Koh, J. Imberger & N. H.
Brooks, 1979. Mixing in inland and coastal waters. Aca­
demic Press, New York, 484 pp.
Geyer, W. R. & R. P. Signell, 1990. A reassessment of the role
of tidal dispersion in estuaries and bays. Estuaries, in press.
Geyer, W. R. & R. P. Signell, 1990. Measurements of tidal
flow around a headland with a shipboard acoustic Doppler
current profiler. J. geophys. Res. 95: 3159-3197.
Mazda, Y., Y. Sato, S. Swamoto, H. Yokochi & E. Wolanski,
1990. Links between physical, chemical and biological pro­
cesses in Bashita-Minato, a mangrove swamp in Japan.
Estuar. coast. Shelf Sei., in press.
M azda, Y., H. Yokochi & Y. Sato, 1990. The behaviour of
groundwater in a mangrove area and the influence on the
properties of water and bottom mud. Estuar. coast. Shelf
Sei. 31: 621-638.
Murray, S. P. & M. Young, 1985. The nearshore current along
ahigh-rainfall, tradewind coast - Nicaragua. Estuar. Coast.
Shelf Sei. 21: 687-699.
Nokes, R. I., 1986. Problems in turbulent dispersion. Ph.D.
Thesis, Dept. Civil Eng., Univ. of Canterbury,
Christchurch, New Zealand, 229 pp.
Okubo, A., 1973. Effect of shoreline irregularities on stream-
wise dispersion in estuaries and other embayments. Neth.
J. Sea Res. 6: 213-224.
Ovalle, A. R. C., C. E. Rezende, L. D. Lacerda & C. A. R.
Silva, 1990. Factors affecting the hydrochemistry of a man­
grove tidal creek, Sepetiba Bay, Brazil. Estuar. coast. Shelf
Sei.: 639-650.
Ozsoy, E., 1977. Flow and mass transport in the vicinity of
tidal inlets. Technical Rep. U FL/CO EL/TR-036, Coastal
and Oceanographic Engineering Department, University of
Florida.
Pattiaratchi, C., A. Jam es & M. Collins, 1987. Island wakes
and headland eddies: a comparison between remotely
sensed data and laboratory experiments. J. geophys. Res.
92: 783-794.
Petryk, S. & G. Bosmajan, 1975. An analysis a flow through
vegetation. J. Hydraul. Div., Am. Soc. Civil Eng. 101: 871-
874.
Pingree, R. D. & L. M addock, 1979. The tidal physics of
headland flows and offshore tidal bank formation. Mar.
Geol. 32: 269-289.
Ridd, P. V., M. W. Sandstrom & E. Wolanski, 1988. Out­
welling from tropical tidal salt flats. Estuar. coast. Shelf Sei.
26: 243-253.
Ridd, P. V., E. Wolanski & Y. Mazda, 1990. Longitudinal
diffusion in mangrove-fringed tidal creeks. Estuar. coast.
Shelf Sei. 31: 541-554.
Robertson, A. I., D. M. Alongi, P. Daniel & K. G. Boto,
1988. How much mangrove detritus enters the G reat Bar­
rier reef lagoon? Proc. 6th. Inter. Symp. Coral Reefs,
Townsville, Aug. 1988.
Signell, R. P., 1989. Tidal dynamics and dispersion around
coastal headlands. Ph.D. thesis, MIT/W HOI-89-38.
Signell, R. P. & W. R. Geyer, 1990. Numerical simulation of
tidal dispersion around a coastal headland. In R. T. Cheng
(ed.), Residual currents and long-term transport in estuar­
ies and bays, Lecture notes on Coastal and Estuarine Stud­
ies, Springer-Verlag: 210-222.
Simpson, J. H. & I. D. James, 1986. Convergent fronts in the
circulation on tidal estuaries. In D. A. Wolfe (ed.), Estua­
rine variability. Academic Press, London.
Taylor, G. I., 1921. Diffusion by continuous movements. Proc.
London Math. Soc., Ser. A, 20: 196-211.
Tee, K. T., 1976. Tide-induced residual current, a 2-D non­
linear numerical tidal model. J. Mar. Res. 34: 603-628.
Thompson, R. O. R. Y. & T. J. Golding, 1981. Tidally in­
duced upwelling by the G reat Barrier Reef. J. geophys.
Res. 86: 6517-6521.
Uncles, R., J. E. Ong & W. K. Gong, 1990. Observations and
analysis of a stratification-destratification event in a tropi­
cal estuary. Estuar. coast. Shelf Sei. 31: 651-666.
van Senden, D., 1985. Structure o f tidal jets under variable
source conditions. Proc. 21st Congress I.A .H.R., Mel­
bourne, 3C, pp. A I/3-A I/8.
W attayakorn, G., E. W olanski & B. Kjerfve, 1990. Mixing,
trapping and outwelling in the Klong Ngao mangrove
swamp, Thailand. Estuar. coast. Shelf Sei., in press.
Williams, D. McB., 1988. Significance o f coastal resources to
sailfish and blackmarlin in N E Australia: an on-going re­

search program. Proc. International Billfish Symposium II,
Kailua-Kona, Hawaii, Aug. 1-5, 1988.
Wolanski, E., 1986. An evaporation-driven salinity maximum
zone in Australian tropical estuaries. Estuar. coast. Shelf
Sei. 22: 415-424.
Wolanski, E., 1989. Measurements and modeling of the water
circulation in mangrove swamps. UNESCO-CO M ARA F
Serie Documentaire No. 3, 43 pp.
Wolanski, E., 1989. Circulation anomalies in tropical Austra­
lian estuaries. In B. J. Kjerfve (ed.), Hydrodynamics of
estuaries. Vol. II, CRC Press, Boca Raton: 53-59.
Wolanski, E. & R. Gardiner, 1981. Flushing of salt from
mangroves. Aust. J. mar. Freshwat. Res. 32: 681-683.
Wolanski, E. & W. M. Hamner, 1988. Topographically con­
trolled fronts in the ocean and their biological influence.
Science 241: 177-181.
Wolanski, E. & B. King, 1990. Flushing of Bowden Reef la­
goon. Estuar. coast. Shelf Sei. 31: 789-804.
Wolanski, E. & P. Ridd, 1986. Tidal mixing and trapping in
mangrove swamps. Estuar. coast. Shelf Sei. 23: 759-771.
Wolanski, E. & P. Ridd, 1990. Mixing and trapping in Aus­
tralian tropical coastal waters. In R. T. Cheng (ed.), Re­
sidual currents and long-term transport in estuaries and
161
bays. Lecture notes on Coastal and Estuarine Studies,
Springer-Verlag: 165-183.
Wolanski, E. & D. van Senden, 1983. Mixing of Burdekin
River flood waters in the Great Barrier Reef. Austr. J. mar.
Freshwat. Res. 34: 49-63.
Wolanski, E., D. Burrage & B. King, 1989. Trapping and
dispersion o f coral eggs around Bowden Reef, G reat Bar­
rier Reef, following mass coral spawning. Contin. Shelf
Res. 9: 479-496.
Wolanski, E., J. Imberger & M. L. Heron, 1984. Island wakes
in shallow coastal waters. J. geophys. Res. 89: 10553—
10569.
Wolanski, E., M. Jones & J. S. Bunt, 1980. Hydrodynamics
of a tidal creek-mangrove swamp system. Aust. J. mar.
Freshwat. Res. 31: 431-450.
Wolanski, E., E. Drew, K. Abei & J. O’Brien, 1988. Tidal jets,
nutrient upwelling, and their influence on the productivity
of the alga Halimeda in the Ribbon Reefs, G reat Barrier
Reef. Estuar. coast. Shelf Sei. 26: 169-201.
Wolanski, E., Y. M azda, B. King & S. Gay, 1990. Dynam­
ics, flushing and trapping in Hinchinbrook Channel, a giant
mangrove swamp, Australia. Estuar. coast. Shelf Sei. 31:
555-580.
!
Hydrobiologia 247: 1 6 3 -1 7 1 , 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 163
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Intratidal fluctuations in stratification within a mangrove estuary

R. J. U ncles ,1 W .-K . G o n g 2 & J.-E . O n g 2

1Plymouth Marine Laboratory, Prospect Place, Plymouth, P L I 3DH, UK; 2 School o f Biological Sciences,
Universiti Sains Malaysia, Penang, Malaysia

K ey words: m angrove estuary, salt transport, estuarine stratification, dispersion, vertical shear
tran sp o rt

Abstract

The dependence o f salinity stratification on the vertical structure o f the longitudinal salinity gradient is
investigated by solving the equations o f longitudinal m om entum and salt conservation. The results are
used to interpret m easurem ents m ade throughout 31 tidal cycles in the M erbok Estuary, M alaysia, as
p art o f an ecological study o f a tropical m angrove estuary. The solutions show th at a substantial part
o f the salinity stratification was caused by vertical shear in the currents coupled with advection of the
longitudinal salinity gradient. This m echanism led to m axim um stratification around low w ater and
m inim um stratification around high w ater. O bserved intratidal variations in stratification at neap tides
greatly exceeded m odelled values when a depth-independent, longitudinal salinity gradient w as assum ed.
Sim ulations m ade using a longitudinal salinity gradient o f simple depth-dependent, power-law form
implied th at this difference arose m ainly because the near-surface, longitudinal salinity gradient in the
field w as significantly larger than th at near the bed during neap tides and at times o f high run-off. The
longitudinal, dispersive tran sp o rt o f salt, fresh w ater and solutes due to vertical shear is shown to be
im portant during periods o f stratification at neap tides and during high run-off. The relevance o f strat­
ification to m angrove estuaries is discussed.

Introduction rents, coupled with density stratification (Dyer,

M angroves are thought to be im portant in the
nutrient budgets of m any tropical coastal areas.
Boto and B unt (1981), B oto et al. (1984) and R idd
et al. (1988) have shown th at large quantities of
organic and inorganic nutrients are exported from
m angrove sw am ps to coastal waters. Such nutri­
ent outwelling m ay affect the productivity of
coastal fisheries. The longitudinal, tidally-induced
dispersive tran sp o rt of nutrients, fresh w ater and
detritus from m angrove to coastal areas has been
investigated by W olanski and R idd (1986), R idd
et al. (1990) and W olanski et al. (1990). D isper­
sive tran sp o rt due to vertical shear in the cur­
1974; R attray & D w orski, 1980) has received less
attention, but is likely to be im portant for these
systems during periods o f stratification at neap
tides and during times o f high freshw ater runoff.
U ncles et al. (1990) showed th at salinity strat­
ification and longitudinal transport due to verti­
cal shear dispersion o f fresh w ater (and presum ­
ably nutrients and other solutes) were large during
neap tides. The stratification and tran sp o rt were
considerably enhanced in the presence o f high
freshw ater runoff. A m odel o f the stratification, in
which stratification w as related to the longitudi­
nal advection o f the horizontal density gradient
and the effects o f vertical mixing, generated the
164

correct trends and tidal averages but failed to
reproduce sufficiently strong intratidal fluctua­
tions in stratification. In this paper we consider
intratidal processes in m ore detail and show th at
it is necessary to specify the vertical profile o f the
longitudinal density gradient if the intratidal fluc­
tuations are to be reproduced correctly.
This article briefly reviews observed d ata for
the M erbok Estuary, followed by an outline o f the
m athem atical m odel and its solutions. T he orig­
inal aspects are the treatm ent o f a schem atic form
for the vertical profile o f longitudinal density gra­
dient and the analysis o f intratidal fluctuations in
stratification. It is n o t our objective to derive ‘best
fits’ betw een m odel and data, but rather to em ­
phasize the im portance o f stratification and the
physical processes affecting it, as well as com ple­
m ent recent w ork on this topic (N unes et al., 1989;
S im pson et al., 1990; U ncles é ta l., 1990; Sim p­
son & Sharpies (in press)).
Observations
Study area and measurements
T he M erbok Estuary (Fig. 1) is situated in the
state o f K edah, M alaysia (5° 40' N , 1 0 0 °2 5 ' E).
The tidal influence extends for 35 km. The width
ranges from about 2 km at the m outh to about

#4 5 V

Tide
ga u g e

Sungai M erb ok U iC

Sungai
Petani

STRAITS

OF Effi Mangrove area

Xi R e c la im e d a rea
MALACCA

S. Muda

Fig. 1. The Merbok Estuary, northwest Malaysia, showing the deep-channel measurement site and the location of the tide-gauge.

20 m tow ards the upper reaches o f the estuary,
w ith depths generally varying from 15 to 3 m. The
m ean annual freshw ater discharge is estim ated to
be 20 m 3 s " l .
M easurem ents were m ade throughout 31 tidal
cycles during June 1987 as p art o f an ecological
study o f a tropical m angrove estuary. F our sta­
tions were w orked near the m outh o f the M erbok
from 0730 h on 2 June until 1100 h on 18 June
1987. C urrent velocities were m easured every
lunar hour at five depths using calibrated current
vanes (Kjerfve & M edeiros, 1989). W ater sam ­
ples were collected and analysed for salinity at
three depths (surface, m id-depth and bed) every
alternate lunar hour and at one depth (m id-depth)
every other lunar hour. W e will present d ata from
the deep-channel (15 m) station in order to ex­
tend the analyses presented in U ncles et al. (1990)
and to provide additional insights into the phys­
ical processes at w ork in b oth m angrove and tem ­
perate estuaries.
Tides, salinity and runoff
The am plitudes o f the w ater-surface elevations
were small; root-m ean-square (R M S ) am plitudes
varied from about 0.3 m during the neap tide to
about 0.8 m during the following spring tide.
T hese am plitudes were less than about 5% o f the
(A): U ( - ) AND 10*R U N 0F F
20 0
co
CO lo o
S
c/i
100
0 10 0 200 300
LUNAR HOURS
Fig. 2. Currents, salinity and runoff data against time in lunar hours. (A): Estimated freshwater runoff (dashed curve, multiplied
by 10 for clarity, m 3 s “ 1) and depth-averaged tidal currents (cm s “ 1, ebb positive); (B): Surface and near-bed (dashed curve)
salinity (ppt).
165
m ean depth. The depth-averaged tidal currents
(ebb positive) are show n in Fig. 2A against time
in lunar hours. The R M S current speeds ranged
from about 0.2 m s -1 during the neap tide to
0.7 m s “ 1 during springs. F reshw ater runoff d ata
are also show n in Fig. 2A (plotted as the dashed
line at 10 times actual value for clarity). R unoff
reached a m axim um o f about 15 m 3 s 1 during
the neap tides.
The surface and bed salinity d ata are shown in
Fig. 2B. A considerable reduction in surface sa­
linity occurred during the neap tide, w hereas the
m id-depth salinity (not show n) and bed salinity
rem ained high. This reduction was due to weak
vertical mixing during the neap-tide period and
increased runoff (Fig. 2A). The stratification (the
difference in salinity between bottom and surface
w aters) becam e very small over the spring tide
due to increased mixing and show ed pronounced
intratidal variability over the period.
Longitudinal salinity gradient
The observations perm it only the depth-averaged,
longitudinal salinity gradient to be estim ated (U n ­
cles et al., 1990). A large peak occurred at 100 h
w hen salinity m inim ized and m axim um effects of
the runoff were experienced (Fig. 3A). However,
because the salinity o f w aters at the bed (and
(B ): SALINITY ( B E D , )
40
V
30
20
400 0 100 20 0 300 400
LUNAR HOURS
166
(A): SALINITY GRAD., d S / d X
S
M
>"
H 0 .5
O.
Pm
0 100 200 300
LUNAR HOURS
Fig. 3. Tidally averaged variables against time in lunar hours. (A): Estimated, depth-averaged, longitudinal salinity gradient (ppt
km~ '); (B): Magnitude of the up-estuary salt flux due to vertical shear dispersion (ppt cm s - 1 ) and the associated dispersion
coefficient, D (dashed line, plotted one-tenth actual value for clarity, m2 s " ').
m id-depth) were little affected by runoff and tidal
state over the period (Fig. 2B), m uch o f the strong
longitudinal salinity gradient at 100 h m ust have
been confined to the upper half o f the w ater col­
umn.
Vertical shear salt transport
The vertical shear dispersive flux o f salt is de­
noted by F and is defined as (U ncles et al., 1985;
M urray & Siripong, 1978; D ronkers & van de
Kreeke, 1986; H unkins, 1981):
F=jtrw)
with U' = U - U and S ' = S - S and where the
overbar is a depth average. An estim ate o f the
dispersion coefficient due to vertical shear is:
D = - <F}/dx (S >
w here < ) denotes a tidal average.
The tidally-averaged salt flux due to vertical
shear, < F > , is always directed up-estuary but is
small during spring tides (Fig. 3B). Dividing the
tidally averaged salt flux by the depth-averaged
longitudinal salinity gradient (Fig. 3A) gives an
estim ate o f the dispersion coefficient, D, due to
vertical shear (plotted as the dashed line at one
tenth actual value for clarity in Fig. 3B). The dis­
(B ): SALT F L U X (-) AND D / 1 0
5 0 --------------- 1--------------- 1--------------- 1------------
CO
s
u
Eh
CL,
CL,
400 0 100 200 300 400
LUNAR HOURS
persion coefficient is about 450 m 2 s ~ 1 during
strong stratification at neap tides and less than
about 50 m 2 s ” 1 during spring tides. Therefore,
during neap tides, the vertical shear flux is im ­
p o rtan t for flushing fresh w ater and solutes from
the M erbok E stuary to its coastal w aters and for
im porting salt water.
I n tr a tid a l s tra tific a tio n m o d el
Basic equations
A vertical mixing m odel has been presented in
U ncles et al. (1990). The m odel includes the com ­
peting influences o f tidal mixing and buoyancy.
T he longitudinal acceleration o f currents is the
result o f an im balance betw een vertical eddy vis­
cosity and horizontal pressure gradients. There is
no wind stress at the surface and a quadratic
friction law applies at 1 m above the bed. Longi­
tudinal advection is ignored, so th at the only n o n ­
linear aspects o f the m odel are the shear and
stratification dependence o f the eddy viscosity,
and the quadratic dependence o f frictional drag
on near-bed velocity.
The m om entum equation can be w ritten (with,
for example, dt = 8/dt):

d,U= - g p - 1 dx pdz - g [ p ( 0 / p ] d x t
» ation from the depth-averaged velocity:
z
+ dz (Ndz U)
where longitudinal current is denoted by U, den­
sity, p, is equal to its depth-averaged value p, ex­
cept when associated with buoyancy forcing.
E ddy viscosity, N (a function o f the R ichardson
num ber) is determ ined using a mixing length ap­
proach. The surface is at z = (, where £ is surface
elevation relative to datum , and the bed is at
z = - h. The surface elevation displacem ents
within a tidal cycle are ignored (£<ih).
W e will explore the effects o f depth variations
in the longitudinal density gradient (dx p) o f the
form:
Sx p = [1 + ß { \ - (n + \ ) a n}]dx p
where ß and n are constants, such th at « > 0 and
ß = 0 when n = 0 , and where a is defined by:
a = (z + /z)/(£ + h) = (z + h)/H
such th at a = 0 is the bed and a= 1 is the surface
with H = h+ £ used to denote total depth. If the
longitudinal density gradient is directed down-
estuary throughout the w ater column (that is, the
density always increases tow ards the sea) and is
never zero, then:
- \ < ß < l/n
Further, if the longitudinal density gradient
m onotonically increases in the fresher layers to­
w ards the surface, then:
- l< ß< 0
In hom ogeneous conditions, ß = 0 .
Averaging E quation (1) over depth and sub­
tracting the depth-averaged version from E qua­
tion ( 1) eliminates the barotropic pressure gradi­
ent and gives:
d,U' = GH[{ 1 - 2a) + ß (l - 2a) -
- 2 ß{(n + 2 )- 1 - (7"+ !}]
+ d , m U ' ) + C D - U 1\Ul \ / H
167
where a bottom drag coefficient of
CD = 2.5 X 10 ” 3 is used and where U' is a devi­
(1) U' = U - U
and
G = - gdx p/2p = - 3.9 X 10 _ 4gdx S (4)
(Bow den & H am ilton, 1975). B oundary condi­
tions for E quation (3) are: A t the surface (z = Ç):
Ndz U' = 0
A t the bed (z = - h):
Ndz U' = CD - U X\UX\
(2) where Ul is the current velocity at 1 m above the
bed. S tandard notation is used. Equation (3) is
solved by specifying the depth-averaged tidal
velocity from observations (U ncles etal., 1990)
together w ith the longitudinal density gradient.
This approach avoids the need to solve a conti­
nuity equation in order to specify the barotropic
pressure gradient (which varies in a com plicated
way with both intratidal and spring-neap tidal
state and freshw ater flow).
The salt-balance equation (in which lateral and
vertical flows, longitudinal variations in U, hori­
zontal mixing and tem poral variations in surface
elevation are ignored) m ay be written:
d,S = - Udx S + dz (Kdz S ) (5)
Subject to zero salt flux, Kdz S = 0, at the surface
and bed.
Averaging E quation (5) over the depth and
subtracting the depth-averaged version from
E quation (5) gives:
8t S ' = - [Udx S - Udx S ] + dz (Kdz S ' ) ( 6 )
In which the advection term can be com puted
having solved for U' and knowing U from the
observations.
The eddy coefficients, N and K are functions
(3) o f the local R ichardson N um ber, velocity shear
168

and mixing length and possess the classical
M unk-A nderson dam ping dependence on water-
colum n stability (U ncles et al., 1990).
Results of model
Longitudinal density-gradient profiles
H aving estim ated the depth-averaged, longitudi­
nal density (salinity) gradient (Fig. 3A), the strat­
ification, vertical profile o f longitudinal current
and the longitudinal salt flux due to vertical shear,
F, can be com puted using E quations 1-6. The
results depend upon the way in w hich the longi­
tudinal density gradient is assum ed to vary with
depth (that is, the values assigned to ß and n in
Equation 2).
Figures 4(A,B) show hypothetical, vertical p ro ­
files of the longitudinal density (or salinity) gra­
dient, relative to its depth-averaged values, for
ß= - 1/3 and ß = - 1 in Equation 2, respectively.
The case ß = - 1 corresponds to zero longitudinal
gradient at the bed; /?=0 (equivalent to n = 0 )
corresponds to a longitudinal gradient which is
everywhere equal to its depth-averaged value, and
ß= - 1/3 is an interm ediate and m ore realistic
case. Show n in Figs 4(A,B) are the vertical p ro ­
files corresponding to n = 5, 10 and 20, (E qua­
tion 2 ) together with vertical, dashed lines corre­
sponding to the depth averages (or n = 0 ).
Vertical profiles having values o f n greater than

(A): d S / d X PROFILE ( - 1 / 3 ) (B ): d S / d X PROFILE ( - 1 )

1

M K
\ 0 .5 \ 0 .5
N N

0
0 5 0 5 10
RELATIVE d S / d X RELATIVE d S / d X

(C ): STRATIFICATION (D ): MODEL STRATIFICATION

20 OU i------------------- 1------------------- 1-------------

15
20
H H
CU 10 CU
cu CU
10
5

0
0 100 200 400 0 100 20 0 300 400

LUNAR HOURS LUNAR HOURS

Fig. 4. The effects of a hypothetical, depth-dependent, longitudinal density (salinity) gradient. (A): Vertical profiles of dxS for
ß= - 1/3 in E quation2 and n = 5 ( - ) , 10 ( ---- ) and 20 ( - • - ) ; (B): The same except that ƒ!= - 1; (C): Observed stratification
(ppt) against time; (D): Modelled stratification (ppt) for ß= - 1/3 in Equation 2 and for n = 0 ( - • - ), 5 ( - ) and 10 ( ---- ) against
time. Note that Z = z + h in (A) and (B).

about 5 dem onstrate an alm ost uniform value of
the longitudinal density gradient in the bottom
half o f the w ater colum n (Figs 4(A,B)). This n ear­
uniform ity o f the hypothetical profiles is consis­
tent with the small intratidal variations in salin­
ity which were observed to occur during neap
tides in the bottom half o f the w ater colum n in the
M erbok (Fig. 2B).
Calculations o f intratidal variability
Results are presented o f a calculation in which
the depth-averaged, longitudinal density gradient
corresponds to the estim ated values show n in
Fig. 3A. By using the time-varying, longitudinal
density gradient we are also effectively taking into
account the time-varying freshw ater flow. A value
o f ß = - 1/3 is used in Equation 2 as an example
o f the effects o f a hypothetical, vertical profile in
the longitudinal density gradient and calculations
are m ade for values o f n = 0, 5, 10 and 20. The
case n = 0 corresponds to a constant value o f the
longitudinal density gradient throughout the w ater
colum n (equivalent to ß = 0). In nature, one would
expect b oth ß and n to vary with time and espe­
cially with the state o f the spring-neap cycle.
The observed stratification (bed m inus surface
salinity) is shown in Fig. 4C. The intratidal fluc­
tuations are very pronounced. M axim um strati­
fication tends to occur at low w ater and m inimum
stratification at high w ater. D uring ebb currents,
low salinity w ater is advected dow n-estuary in the
faster surface layers, which tends to minim ize sa­
linity at the surface around low w ater and m ax­
imize stratification. D uring the flood, vertical
shear in the tidal currents is reduced by the
density-driven circulation, so th at the tendency
for stratification to occur as a result of differen­
tial advection o f the longitudinal density gradient
is less. T he m odelled stratification is shown in
Fig. 4D for values o f n = 0 (dash-dot line), 5 (con­
tinuous line) and 10 (dashed line). There is very
little difference in the m odelled stratification be­
tween the cases for n= 10 and n = 20 , so th at only
the case for n = 10 is plotted in Fig. 4D.
It is clear from Figs 4(C ,D ) th at the intratidal
169
variations in stratification cannot be reproduced
with n = 0 (ƒ?= 0 ), which corresponds to a depth-
independent longitudinal density gradient, al­
though the m agnitude of the tidally-averaged
stratification is reproduced quite well. W ith n = 5
and n = 10 (Fig. 4D ) the intratidal fluctuations in
stratification are strong and are such th at maxi­
m um and m inim um stratification tend to occur at
low and high w ater, respectively, as observed in
the M erbok.
The m odelled stratification is greater with in­
creasing n. A lm ost-fresh w ater is predicted to
occur in the surface layers during m inim um neap
tides when n = 10. This result is a consequence of
our assum ption th at the longitudinal density gra­
dient is independent o f distance along the estuary
and is therefore no t affected by the com puted salt
flux. T he neglect o f density-gradient variations
along the estuary also precludes us from m odel­
ling the behaviour o f a surface, freshw ater lens. In
som e estuaries, such as the H aw kesbury, P arra­
m atta and P o rt H acking Estuaries in N ew South
W ales, A ustralia, a short-lived (few hours to a
day) freshw ater spate can flow to sea as a surface
lens and have little effect on the deeper w aters
(W olanski, personal com m unication).
T he m odelled, intratidal fluctuations in strati­
fication for n = 5 and 10 at neap tides are very
similar to those observed, w hereas at spring tides
they are m uch larger. This is a result o f our as­
sum ption th at the vertical profile o f longitudinal
density gradient does no t change with time. In
reality, enhanced vertical mixing at spring tides
will tend to reduce variations w ith depth and thus
decrease the absolute m agnitudes o f ß and n.
Relevance of stratification to mangrove systems
The presence o f m angrove sw am ps is m ost likely
to influence the w ater colum n stratification indi­
rectly, through an influence on m orphology and
thus tidal flow and mixing and through an ability
to trap fresher w ater and thus m odify longitudi­
nal salinity gradients. H ow ever, the longitudinal,
dispersive tran sp o rt o f salt and freshw ater due to
vertical shear in both density-driven and tidal
170
currents, coupled with stratification, will directly
influence the export o f nutrients, m angrove litter
and detritus from the m angrove system to the
coastal area and m ay thus affect the productivity
o f coastal fisheries.
W olanski & R idd (1986) have shown th at tidal
trapping in the sw am ps o f m angrove estuaries
can enhance longitudinal dispersion o f nutrients
and other m aterials by tem porarily storing w ater
in the sw am ps during the late flood and releasing
it to a different part o f the m ain flow during the
ebb. It is likely th at stratification will m odify this
effective longitudinal dispersion o f fresh w ater and
nutrients by supplying fresher and m ore nutrient-
rich surface w aters to the swam p. B ecause of
stratification, it is possible th at m angroves will
experience less stress and therefore spend less
energy dealing with salt in their effort to extract
fresh w ater for physiological and m etabolical p ro ­
cesses. It is also possible, as a result o f fresher
w ater flooding the m angrove m uds during the
neap stratification, th at m ore nutrient leaching
and export m ay occur.
Conclusions
(1) A substantial p art o f the salinity stratification
w as caused by vertical shear in the currents cou­
pled with advection o f the longitudinal salinity
gradient. This m echanism led to m axim um strat­
ification around low w ater and m inim um strati­
fication around high water.
(2) O bserved intratidal variations at neap tides
greatly exceeded m odelled values w hen a depth-
independent, longitudinal density gradient was
assum ed. Sim ulations m ade using a hypothetical,
longitudinal density gradient o f pow er-law form
with depth dem onstrated th at this difference arose
mainly because the near-surface, longitudinal
density gradient in the field w as significantly larger
than values lower in the w ater colum n during
neap tides and at times of high run-off.
(3) The tidally-averaged salt tran sp o rt due to
vertical shear was directed up-estuary and was
m uch smaller during spring tides. The greatest
longitudinal salt flux and dispersion coefficient
due to vertical shear occurred approxim ately d u r­
ing m axim um stratification, w hen vertical mixing
w as very small. This is consistent with theoreti­
cal analyses (Fischer etal., 1979) and with the
conclusion o f N unes et al. (1989) th at the spring-
neap cycle has an im portant influence on estua-
rine stratification and m ass transport. This tra n s­
po rt will influence the export o f nutrients,
m angrove litter and detritus from the m angrove
system to the coastal area.
Acknowledgements
This project (‘M angrove M alaysia’, C entre file
3-P-82-0192) w as funded by the International
D evelopm ent and R esearch C entre (IR D C ) of
C an ad a with infrastructure support from the Uni-
versiti Sains M alaysia. W e are particularly grate­
ful to D r B rian Davy, A ssociate D irector (F ish­
eries) and his staff at the IR D C A sia Regional
C entre in Singapore for their excellent support.
W e are also grateful to the B ritish C ouncil for
providing a travel grant th at allowed one o f us (R
J U ) to participate in this aspect o f the project.
O ur thanks go to D r W ong Chee H oong and D r
Z ubir D in as well as the 60 or so laboratory staff
and students o f U niversiti Sains M alaysia who
cheerfully assisted in the very arduous field and
laboratory work.
References
Boto, K. G. & J. S. Bunt, 1981. Tidal export of particulate
organic matter from a northern Australian mangrove sys­
tem. Estuar. coast. Shelf Sei. 13: 247-255.
Boto, K. G., J. S. Bunt & J. Wellington, 1984. Variations in
mangrove forest productivity in northern Australia and
Papua New Guinea. Estuar. coast. Shelf Sei. 19: 321-329.
Bowden, K. F. & P. Hamilton, 1975. Some experiments with
a numerical model of circulation and mixing in a tidal es­
tuary. Estuar. coast, mar. Sei. 3: 281-301.
Dronkers, J. & J. van de Kreeke, 1986. Experimental deter­
mination of salt intrusion mechanisms in the Volkerak Es­
tuary. Neth. J. Sea Res. 20: 1-19.
Dyer, K. R., 1974. The salt balance in stratified estuaries.
Estuar. coast, mar. Sei. 2: 273-281.
Fischer, H. B., E. J. List, R. C. Y. Koh, J. Imberger & N. H.

Brooks, 1979. Mixing in inland and coastal waters. Aca­
demic Press, New York, 483 pp.
Hunkins, K., 1981. Salt dispersion in the Hudson Estuary. J.
Phys. Oceanogr. 11: 729-738.
Kjerfve, B. & C. Medeiros, 1989. Current vanes for measur­
ing tidal currents in Estuaries. Estuar. coast. Shelf Sei. 28:
87-93.
Murray, S. P. & A. Siripong, 1978. Role o f lateral gradients
and longitudinal dispersion in the salt balance of a shallow
well mixed estuary. In: B. Kjerfve (ed.), Estuarine Trans­
port Processes. University of South Carolina Press: 113—
124.
Nunes Vaz, R. A., G. W. Lennon & J. R. de Silva Samaras-
inghe, 1989. The negative role of turbulence in estuarine
mass transport, Estuar. coast. Shelf Sei. 28: 361-377.
Rattray, M. & J. G. Dworski, 1980. Comparison o f methods
for analysis of the transverse and vertical circulation con­
tributions to the longitudinal advective salt flux in estuar­
ies. Estuar. Coast. Mar. Sei. 11: 515-536.
Ridd, P. V., M. W. Sandstrom & E. Wolanski, 1988. Out-
welling from tropical tidal salt flats. Estuar. coast. Shelf Sei.
26: 243-253.
Ridd, P., E. Wolanski & Yoshihiro Mazda, 1990. Longitudi­
nal diffusion in mangrove-fringed tidal creeks. Estuar. coast.
Shelf Sei. 31: 541-554.
171
Simpson, J. H., J. Brown, J. Mathews & G. Allen, 1990. Tidal
straining, density currents and stirring in the control of
estuarine stratification. Estuaries 13: 125-132.
Simpson, J. H. & J. Sharpies, In press. Dynamically-active
models in the prediction of estuarine stratification. In D.
Prandle (ed.), Physics of Estuaries and Coastal Seas.
Springer-Verlag, Berlin/Heideberg.
Uncles, R. J., R. C. A. Elliott & S. A. Weston, 1985. Disper­
sion of salt and suspended sediment in a partly mixed es­
tuary. Estuaries 8: 256-269.
Uncles, R. J., J. E. Ong & W. K. Gong, 1990. Observations
and analysis of a stratification-destratification event in a
tropical estuary. Estuar. coast. Shelf Sei. 31: 651-665.
Wolanski, E. & P. V. Ridd, 1986. Tidal mixing and trapping
in mangrove swamps. Estuar. coast. Shelf Sei. 23: 759-
771.
Wolanski, E., Yoshihiro M azda, B. King & S. Gay, 1990.
Dynamics, flushing and trapping in Hinchinbrook Channel,
a giant mangrove swamp, Australia. Estuar. coast. Shelf
Sei. 31: 555-579.
Wolanski, E. Personal Communication. Australian Institute
of Marine Science, PMB No. 3, Townsville M.C.. OLD
4810, Australia.
Hydrobiologia 247: 1 7 3 -1 7 9 , 1992.
V. Jaccarini & E. Martens (eds). The Ecology o f Mangrove and Related Ecosystems. 173
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Physical processes in the Rufiji delta and their possible implications on

the mangrove ecosystem

Julius Francis
Institute o f Marine Sciences, P.O. B ox 668, Zanzibar, Tanzania

K ey words: m angrove sw am ps, trapping, mixing, estuarine circulation, coastal currents, Rufiji delta

Abstract

The physical processes such as coastal currents, estuarine circulation and m onsoon w inds prevailing in
the Rufiji delta are discussed. The relationship between these processes and the occurrence o f long-term
trapping o f the river discharge and the outflow o f w aters from the m angrove sw am ps into the nearshore
zone has been observed. The trapped w aters in the nearshore zone significantly reduce the mixing be­
tw een the estuarine and offshore w aters, leading to the tw o w aters having distinctive properties. The
existence o f the trapped w aters in the nearshore zone is supported by evidence from a satellite image
and aerial photograph and m easurem ents o f salinity and suspended sedim ent concentration. The tra p ­
ping phenom enon is discussed in the light o f its possible im plications on the ecology o f m angrove ec­
osystem s. T rapping m ay explain the enhanced grow th o f the m angrove in the delta com pared to other
areas. This trapping effect may be providing m ore time for nutrient retention in the m angrove zone,
incorporation o f the decom posed leaf litter and fine sedim ents in the substrate, and settling o f fruits and
seedlings in the sw am ps, thereby enhancing the regeneration o f the m angrove.

Introduction sem i-diurnal tide, with an average tidal range of

The Rufiji delta, which is located in the southern
T anzania, is aligned approxim ately in a N - S di­
rection. The N - S extent o f the delta is about
65 km and extends for about 23 km inland. The
delta is traversed by a num ber o f channels am ong
which seven are m ajor (see Fig. 1). A part from
being the largest delta in E ast Africa, it is also
contains the largest estuarine m angrove forest in
the eastern p art o f the continent. The area cov­
ered by the m angrove forest is estim ated to be
about 1022 km 2 (M cN ae, 1968). The m angrove
sw am ps are dissected by tidal channels and n u ­
m erous tidal creeks. T he flow betw een the m an­
grove sw am ps and the ocean through these
channels and creeks is largely controlled by the
2 -2 .5 m.
T he Rufiji river has a m ean annual freshw ater
discharge o f about 30 x IO9 m 3. The m aximum
discharge occurs between M arch and M ay and
the m inim um in late O ctober and N ovem ber,
V HL, (1979) classified the Sim ba U ranga branch
as being a partially-m ixed estuary during the low
freshw ater discharge, and stratified during the
high discharge.
T he northern p art o f the delta (where m ost of
the observations were m ade), experiences strong
w inds during the N E m onsoon and therefore, the
sea is rough. D uring the SW m onsoon due to
relatively w eaker w inds prevailing, the sea is calm.
This is because the northern p art o f the delta is
m ore exposed to the N E m onsoon w inds and
174

\ KENYA

T A N Z A N IA
S im b a Uranga

D odom a
D a r -es-
S a la a m K iom boni

*rin9a /R ufiji"
' R.
0 200 km

M sala

>3A) M sa la

K iasi
K iasi

R. Rufiji 2A

M som eni

16km

.Mohoro

.Yaya

Fig. 1. M ap of Rufiji delta area; A -B Suspended sediment concentration transect line, J- Station at the mouth of tidal creek J,
and S- Station at the mouth o f Simba U ranga channel.

p rotected to the SW m onsoon winds. G enerally
however, T an zan ian coast experiences stronger
w inds during the SW m onsoon com pared to the
N E m onsoon (Bryceson, 1977).
The freshw ater discharge from the delta chan­
nels o f the Rufiji is trapped in the nearshore zone.
T he prevailing w inds and current conditions are
thought to be responsible for the existence o f this
trapping (Iversen et al., 1984). The existence and
the causes o f the trapping effect in the nearshore
zone and in the m angrove sw am ps have been
established in several areas such as C hw aka Bay
in Z an zib ar and in the G u lf o f C arpentaria and
in its adjacent m angrove sw am ps in A ustralia
(W olanski & R idd, 1986; W olanski & R idd, 1989;
W olanski, 1989).
Besides the report by Iversen etal. (1984),
which has discussed briefly the existence o f the
trapping effect in the nearshore zone off the Rufiji
delta, there are no other report th at have dis-

cussed the subject. K nowledge o f the existence of
the trapping effect and its im plications on the
m angrove ecosystem s is fundam ental to our u n ­
derstanding o f the processes th at might be affect­
ing the ecology o f the m angrove ecosystem s.
This p aper intends to establish the existence
and the causes o f the trapping effect in the m an­
grove sw am ps and in the nearshore zone off the
Rufiji delta, and to discuss the im plications of the
effect on the m angrove ecosystems.
Material and methods
The salinity m easurem ents were done along the
Sim ba U ran g a channel (see Fig. 1). The channel
w as selected in this study because it is the m ain
branch in the delta river system. The m easure­
m ents o f suspended sedim ent concentration were
done along the Sim ba U ranga and M sala chan­
nels however, in this paper only d ata from the
latter has been used.
A W PA cm 35 conductivity m eter was used to
obtain the conductivity o f the w ater. T he conduc­
tivity readings were later converted to salinity
readings. The conductivity d ata were taken from
tw o anchored stations. The first station at the
m iddle o f the m outh o f the J creek (see Fig. 1).
The creek is about 100 m wide at the m outh and
1 km long. This creek drains only the m angrove
sw am p around it. The second station w as at the
m iddle o f the m outh o f the Sim ba U ranga ch an ­
nel. The m easurem ents at the tw o stations were
taken for four tidal cycles in four consecutive days
i.e. from 21 July, 1990 to 24 July, 1990.
S uspended sedim ent concentration m easure­
m ents were done using the P artech Suspended
M atter M onitor. A t M sala channel the m easure­
m ents were m ade in a transect running from the
offshore to the m outh o f the channel (see Fig. 1).
This w as done purposely to determ ine the varia­
tion o f sedim ent concentration in the estuarine,
nearshore, and offshore w aters. The m easure­
m ents were done on the ebb tide as the sedim ent
concentration on the w ater is m axim um during
this period. At M sala channel, the d ata were col­
lected on 5 Septem ber, 1989. Inform ation for this
175
paper have also been obtained from the interpre­
tation o f satellite image and aerial photograph.
Results and discussion
A t the m outh o f J creek, the salinity values for the
surface w aters were found to fluctuate slightly at
tidal frequencies, being relatively high during flood
tide and low during the ebb tide period. The sa­
linity values ranged from 10.60%o to 14.26%0 (see
Fig. 2a). The d ata collected at a depth o f m ore
than 1 m below the surface showed th at during
the flood tide m ore saline w ater w as flowing into
the creek, as the values o f salinity ranged from
18.12%0 to 30.70%o. The small variation and small
range in salinity values at the J creek over tidal
cycle suggest there is little mixing between the
w ater from the creek and the w ater in the ch an ­
nel. Also the low salinity values obtained were
probably due to freshw ater discharged several
weeks earlier during the rainy season, as for the
period o f about one m onth preceding the d ata
collection there w as no rain in the areas around
the Sim ba U ranga Island and the floods o f the
Rufiji river h ad ceased about two m onths before.
All these facts are implying the existence o f the
trapping o f low salinity w ater in the mangrove
swam ps.
T he surface salinity values obtained at the
m outh o f the Sim ba U ranga channel showed a
relatively large variation between the peak flood
tide values and the corresponding ebb tide values.
The salinity values were low during the ebb tide
and high during the flood tide period (see Fig. 2b).
This large variation can be explained by the fact
th at the m ore saline nearshore w aters move in­
land along the channel during the flood tide and
back to the nearshore zone during the ebb tide.
M easurem ents o f suspended sedim ent concen­
tration taken during the ebb tide at the M sala
channel show ed th at there w as high concentra­
tion o f sedim ent in the nearshore w aters com ­
pared to either the estuarine or the offshore w a­
ters. The concentration ranged from less than
490 mgl “ 1 for the estuarine w aters to about
1000 mgl - 1 for the nearshore w aters. F or the off­
176

14-

O
o
o 10-

>
t—
z
—I
<
IO

HW LW

TIME (HOURS)

30-

25-
O
o

20 -
>
I-
z
_i 15-
<
io
10 -

HW

TIME (HOURS)
Fig. 2. Variation over a tidal cycle o f surface salinity; (a) at the mouth of tidal creek J - 21 July, 1990 and (b) at the mouth of
Simba Uranga channel - 24 July, 1990.

shore w aters zero values were dom inant (as
show n in Table 1). A parallel picture is also draw n
from the aerial photograph (Fig. 3), which indi­
cates th at nearshore w aters contain m ore sus­
pended sedim ent than either the estuarine or the
offshore w aters. It is clear th at the separation
between the nearshore and the offshore w aters is
well defined in m any areas, suggesting the ab­
sence o f intensive mixing between the tw o w ater
m asses. The suspended sedim ent concentration
 177

Table 1. Suspended sediment concentration along the m easurem ents and the interpretation o f the aerial
transect A -B .
photograph indicate a trapping effect in the n ear­
Time (hour) Depth (m) Concentration (mg 1 ') shore w aters off the delta.
The E ast African C oastal C urrent (EA CC )
1235 3.00 490 system m ay be responsible for causing this tra p ­
1235 1.00 480 ping effect. They are generally flowing northw ards
1245 1.00 500
throughout the year, being faster during the SW
1251 1.00 490
1304 0.50 720 m onsoon and slower during the N E m onsoon
1312 0.50 890 (Newell, 1959). The aerial photography shows the
1320 0.30 1000 existence o f northw ard flowing currents, which
1332 1.50 1000 could be the EA C C system. These currents are
1345 1.00 980
inhibiting the flow o f nearshore w aters into the
1256 1.00 1000
1405 1.00 50 deeper offshore w aters, consequently leading to
1416 1.00 100 the outflows from the delta channels being re­
1428 1.00 150 stricted in the nearshore zone.
1440 1.00 0 The Sim ba U ranga channel being stratified
1456 1.00 0
during the period o f high discharge and partially

Fig. 3. Satellite image and aerial photograph; (a) Landsat image of the Rufiji delta with dates: 23 June, 1981, 23 June, 1981, 24
June, 1981, and 18 March, 1985 and (b) Aerial photograph o f part of the delta taken on 10 June, 1965.
178
well mixed during the low freshw ater discharge
suggests the absence o f strong mixing processes.
It seems th at the m onsoon w inds an d /o r the tidal
currents are unable to cause intensive mixing b e­
tween either the estuarine and the nearshore w a­
ters or the nearshore and the offshore waters.
The trapping effect discussed above m ay have
several im plications on the ecology o f the m an­
grove ecosystem s as will be indicated in the fore­
going discussion. The trapping effect m ay be
am ong factors controlling the am ount o f sus­
pended particulate m atter incorporated in the
substrate in the m angrove sw am ps relative to the
am ount exported to the adjacent m arine ecosys­
tem s. O ther factors include the substrate charac­
teristics, the frequency o f tidal inundation, the
energy o f the tide, and the pattern o f production
and supply o f m angrove litter (W oodroffe, 1985).
This is due to the fact th at trapping provide m ore
tim e for the suspended m atter to settle out. Also
m ore tim e is provided for the fruits and seedlings
of m angrove to settle in the swam p and regener­
ate instead o f being w ashed away by the tidal
action as it is in the case where trapping does no t
exist or exist for few days. D ue to little mixing
w hich exists betw een the estuarine and the near­
shore w aters, the dissolved m atter is no t flushed
from the system by tidal action, and rem ains
within the sw am ps for a long time. This provides
m ore opportunity for absorption o f the nutrients
by the m angrove as practically the sam e w ater
w ould be passing through them over and over
again for prolonged period.
The m angroves develop best in areas which
experiences periodic deposition o f sedim ents
(C intron & Schaeffer-Novelli, 1983). In the Rufiji
delta, the m axim um sedim ent tran sp o rt occurs
during the period o f m axim um outflow (Eurocon-
sult, 1980). Som e o f those sedim ents are d epos­
ited in the m angrove sw am ps and the rest are
tran sp o rted to the nearshore zone where as dis­
cussed, they are trapped. H ow ever, during the
SW m onsoon, there is also considerable deposi­
tion o f sand in the m angrove sw am ps o f Kiom-
boni and Sim ba U ranga islands (pers. obser.).
The origin o f these sands is thought to be the
nearshore zone as they are well rounded and are
found together with broken shells. In view o f this
it seems th a t the m angrove are receiving sedi­
m ents from bo th terrestrial and m arine environ­
m ents at different times in a year.
Conclusion
Physical processes play a significant role in the
existence o f the long-term trapping effect in the
m angrove sw am ps and in the nearshore zone.
T he presence o f n o rthw ard flowing E A C C system
and the slow mixing processes are probably the
m ain factors responsible for the existence o f the
trapping effect in the sw am ps and in the n ear­
shore zone. The intensity o f trapping effect m ay
be varying w ith the m onsoon season, being m ore
pronounced during SW m onsoon than during the
N E m onsoon. This is due to the fact th at the N E
m onsoon are characterized by w eaker EA C C
system, lower freshw ater discharge, and relatively
strong m onsoon winds.
The enhanced grow th o f the m angrove in the
delta m ay be explained no t only by the fact th at
the m angrove sw am ps receive significant am ount
o f nutrients from freshw ater discharge, but also
from the fact th a t m ost o f nutrients supplied by
the river and from other sources are retained in
the sw am ps or in the nearshore zone for a p ro ­
longed period due trapping effect.
D ue to limited d ata available some o f the issues
raised in the discussion such as the ecological
im plications o f the trapping effect on the m an ­
grove and the possibility o f the seasonal varia­
tions in the trapping effect etc. are no t conclusive.
Acknowledgements
It is a pleasure to th an k F R O N T IE R Expeditions
for all the support they gave me during d a ta col­
lection. T hanks also should go to M essrs. J. P.
Shunula, A. N . N . M uzuka, Y. W. Shaghude,
and A. J. M m ochi all o f w hom helped m e in v ar­
ious ways, especially for their valuable construc­
tive com m ents and suggestions to the first draft
of the m anuscript.
 179

References MacNae, W., 1968. A general account of the fauna and flora
of mangrove swamps and forest in the Indo-W est Pacific
region. Adv. Mar. Biol. 6: 73-270.
Bryceson, I., 1977. An ecological study of the phytoplankton
Newell, D. S., 1959. The hydrography of the British East
of the coastal waters of Dar-es-Salaam. Ph.D. Thesis, U ni­
African Coastal Waters. Fishery Pubis. Colon. Off. No. 12.
versity of Dar-es-Salaam.
18 pp.
Cintron, G. & Y. Schaeffer-Novelli, 1983. Mangrove Forests:
VHL., 1979. Hydrological studies in the Lower Rufiji River.
Ecology and response to natural and man-induced stres­
Trondheim, Norway. 100 pp.
sors. In J. C. Ogden & E. H. Gladfelter (eds), Coral reefs,
Wolanski, E., 1989. Measurements and modelling of the water
seagrass beds, mangrove: Their interaction in the coastal
circulation in the mangroves swamps. Results of practical
zones of the Caribbean. U N ESC O report in marine science
work at the COM ARAF Regional Field Training W ork­
23: 87-113.
shop on Mangrove Ecosystems and Estuarine Processes.
Euroconsult., 1980. Identification study on the ecological im­
Chwaka Bay, Zanzibar. June 6-16, 1989. Regional Project
pacts of the Stiegler’s Gorge Power and Flood Control
RAF/87/038.
Project Part 2: Methodology, Definition of impact area and
Wolanski, E. & P. Ridd, 1986. Tidal mixing and trapping in
Resources inventory.
mangrove swamps. Estuar. coast, mar. Sei. 23: 759-771.
Iversen, S. A., S. Myklevoll, K. Lwiza & Y. Yonazi., 1984.
Wolanski, E. & P. Ridd, 1989. Coastal mixing and trapping
Tanzanian marine fish resources in the depth region 10-
in tropical Australia. In R. T. Cheng (ed.), Dynamics of
500 m investigated by R/V ‘Dr. Fridtjof N ansen’. In S. A.
Shallow Bays and Estuaries, (in press).
Iversen & S. Myklevoll (eds), The Proc. of the NORAD-
Woodroffe, C. D., 1985. Studies of a mangrove basin, Tuff
Tanzania Seminar to review the marine fish stocks and
Crater, New Zealand: 1. Mangrove biomass and produc­
fisheries in Tanzania. Mbegani, Tanzania. March 6-8,
tion of detritus. Estuar. coast, mar. Sei. 20: 265-280.
1984: 45-83.
. .
Hydrobiologia 247: 1 8 1 -1 8 6 , 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 181
© 1992 Kluwer Academie Publishers. Printed in Belgium.

Diurnal variation in physical-chem ical properties and primary

production in the interconnected marine, mangrove and freshwater
biotopes of Kakinada coast, Andhra Pradesh, India

V. Selvam, Jayapaul A zariah & H ilda A zariah*

Department o f Zoology, University o f Madras, Guindy Campus, M adras - 600 025, India; * C A S in
Botany, University o f Madras, Guindy Campus, Madras - 600 025, India

K ey words: D iurnal variation, m angrove w ater, dissolved inorganic nutrients, prim ary production

Abstract

D iurnal variation in hydrological variables and dissolved inorganic nutrients such as P O | -P, N
0 2~ -N, N 0 3 -N and N H 4+-N were studied in three interconnected biotopes including freshw ater, m a­
rine and m angrove brackish w ater o f the K ak in ad a coastal zone, A ndhra Pradesh. Sam ples were col­
lected at intervals o f 3 hours, for a period o f 24 hours. In the m arine environm ent salinity varied from
26%0 to 32%0 w hereas in the m angrove w aters it fluctuated from 12%„ to 20%o and in both biotopes salinity
show ed bim odal type o f oscillation. D issolved oxygen content w as high in the m angrove w aters during
day time but decreased rapidly during the night hours. In the m arine environm ent P O |_ -P concentra­
tion varied from 0.345 to 1.195 pg at I - \ N 0 3~ -N from 1.03 to 6.62 p,g at 1“ 1 and N 0 2“ -N from 0.086
to 0.506 pg at 1” The highest and the low est concentrations o f P O |_ -P, N O f - N , N 0 2_ -N recorded
in the m angrove w aters were 0.790 and 0.325 pg at 1~ 1, 7.10 and 1.60 pg at 1“ 1 and 0.278 and 0.060 jug
at I “ 1, respectively. The concentration o f PC>4 _ -P, N 0 3_ -N and N O y -N were high in the freshw ater
canal, the m axim um and m inim um values being 1.110 and 0.730 pg at 1” l, 26.40 and 9.98 pg at 1“ 1 and
0.520 and 0.252 ptg at 1” 1 respectively. The concentration o f am m onia was relatively high in the m an­
grove water. G ro ss and net prim ary production in the m angrove w ater w as 4 times higher than in the
m arine biotope. There w as no ‘export’ o f dissolved nutrients from the m angrove environm ent to the
adjacent m arine waters.

Introduction cles. M ost of the investigations carried out on the

U nderstanding the hydrography and nutrient cy­
cling in the m angrove w ater is com plicated by the
continual mixing o f w ater m asses with different
physical-chem ical properties (O dum et al., 1982).
S araya (1984) show ed th at the environm ental
conditions o f m angrove w aters are largely gov­
erned by tw o dom inant factors, namely, short
term changes resulting from tidal inundation and
seasonal changes induced by the m onsoonal cy­
hydrography and nutrients in m angrove w aters
are based on long term observations and are re­
stricted to the brackish w ater surroundings of the
m angroves. H ence, there is a lack o f u nderstand­
ing on the influence o f m arine and freshw ater
discharge on the diurnal and tidal variations in
the physico-chem ical properties o f the m angrove
w aters. In the present study diel and tidal changes
in the hydrographical variables and nutrients were
studied at three interconnected biotopes namely,
182

freshw ater, m arine and m angrove brackishw ater

in order to obtain: (i) the pattern o f the diurnal
and tidal variations in the physical-chem ical
£3 Mangrove
properties o f the m angrove w ater and (ii) to u n ­
d erstand the influence o f m arine and fresh w ater Kakinada
discharge on such variations, if any.

M aterials and methods

A m angrove swam p covering an area o f approx­
imately 96 hectare is present near the K akinada
tow n (16° 45 ' N and 82° 1 5 'E) in A ndhra
Pradesh. The vegetation o f the swam p consists of
Avicennia officinalis L., A . marina (F orsk.) Verh,
Excoecaria agallocha L., Acanthus ilicifolius L.
with A . officinalis as the dom inant species. The
swam p, on the southern side is bordered by the
C om m ercial C anal which receives surplus fresh­
w ater from irrigated land and is characterised by
the presence o f the inflow o f freshw ater through­
out the year. O n the w estern side o f the K akinada
Bay (average salinity 32%0) lies a m angrove
swam p. T he swam p, Com m ercial C anal and the
K ak in ad a Bay are interconnected as shown in
Fig. 1. W hen the freshw ater is minim al in the
C om m ercial C anal seaw ater from the K akinada
Bay enters it at high tide and mixes with the fresh­
w ater. T he resulting brackish w ater enters through
a channel branching off the left side o f the C om ­
m ercial C anal. D uring the m onsoon seasons,
heavy influx o f freshw ater in the Com m ercial
C anal prevents the entry o f seaw ater from the
K ak in ad a Bay and during such a period fresh­
w ater from the Com m ercial C anal directly reaches
the m angrove environm ent. In each biotope one
sam pling site w as fixed and designated as Statio-
n I (K ak in ad a Bay), Station II (M angrove
Sw am p) and Station III (Com m ercial Canal).
Single surface sam ples were collected from these
stations at 3 hour intervals from 0600 h, 20th A u­
gust, 1988 to 0600 h, 21st August, 1988. Salinity
was m easured by an optical salinom eter (Am er­
ican O pticals, H ong Kong, Cat. N o. 10423,
± 1%0). F o r dissolved oxygen analysis, sam ples
were fixed in the field using m anganous sulphate
and alkaline iodide solution. W inkler’s m ethod
KAKINADA
BAY
8 2 ° 1 5 'E
Fig. 1. M ap of the K akinada Bay showing the mangrove
swamp and sampling stations.
given by Strickland & P arsons (1968) w as fol­
lowed for the estim ation o f dissolved oxygen. F or
the analysis o f dissolved nutrients, w ater sam ples
collected from surface were brought to the labo­
ratory within 30 min o f collection and filtered
through a W hatm an G F /C filter p aper o f 0.05 ¡j.m
porosity. T he filtered sam ples were deep frozen
and the nutrients were analysed the next day.
A nalysis o f P O ^ - P , N O 3- N , N O f - N and N
H 4 -N w as carried out according to the m ethods
given by P arso n s et al. (1984). T he light and dark
bottle m ethod as described by Strickland & P a r­
sons (1968) w as followed for assessing the pri­
m ary production. W ater sam ples in duplicate
were incubated in situ for 6 hrs from 0900 hrs to
1500 hrs. D ue to logistical difficulty prim ary p ro ­
duction w as estim ated only at Stations I and II.
A tm ospheric tem perature w as noted at the time
o f sampling.
 183

Results Salinity

Tide Salinity show ed a sem idiurnal oscillation at Sts. I

The tide o f the study area is o f sem i-diurnal type
(Fig. 2a). Tidal level increased gradually from
0600 hrs and reached the first peak at 1200 hrs.
The second peak w as recorded at 2400 hrs. The
tidal am plitude w as 1.52 m.
Water temperature
and II. Salinity rem ained around 0%o at St. III
(Fig. 2c). A t St. I salinity varied from 26%0 to
32%0 with tw o peaks one occurring at 1200 hrs
and the other at 1800 hrs, indicating the sem idi­
urnal nature o f the tidal oscillation. A t St. II sa­
linity reached the peak values at the sam e tim e as
at St. I but the m agnitude was smaller, the m ax­
imum and m inim um values being 20%o and 12%0
respectively.

W ater tem perature at the three stations varied

from 26.9 to 30.8 °C (St. I), 26.9 to 31.2 °C
(St. II) and 27.4 to 30.4 °C (St. III). W ater tem ­
perature at all stations increased gradually from
0600 hrs and reached its peak at 1500 hrs
(Fig. 2b), closely following the atm ospheric tem ­
perature. W ater tem perature in the m angrove
w ater w as slightly higher than in the other two
biotopes during m ost o f the period o f observa­
tion.
Dissolved oxygen
Diel variation in dissolved oxygen concentration
at St. I and III w as smaller than at St. II. It w as
within a range o f 2.77 to 5.07 ml I - 1 at St. I and
2.65 to 4.83 ml 1" 1 at St. III. At St. II dissolved
oxygen concentration fluctuated from 2.41 to
7.24 ml I " 1 and at this station dissolved oxygen
content w as high during the day time, and de­
creased rapidly during the night hours (Fig. 2d).
H igh concentration o f dissolved oxygen in the

o— Mari ne (St. I)
»— S w a m p (St- II)
û— Ca n a i (Still)

40

— TO

>. 20
c
o 10
to
0
12 15 18 21 24 15 18 21 24
Time (hours) Ti me ( h o u r s )

Fig. 2. Diurnal variations in (a) tidal range (m), (b) water tem perature (°C ), (c) salinity (%0), and (d) dissolved oxygen (ml 1 ')■
184

m angrove w ater (St. II) might be due to a high ation in P O |~ -P concentration at all stations over
rate o f oxygen production by phytoplankton. the sam pling period.

Nutrients
Dissolved inorganic phosphate
D issolved inorganic phosphate at the three sta­
tions varied from 0.345 to 1.195 pg at 1“ 1 (St. I),
0.325 to 0.790 pg at I “ 1 (St. II) and 0.730 to
1.110 jug at I “ 1 (St. III) (Fig. 3a). At Stations I
and II P O |_ -P concentration, during the day
time, w as low and showed no relationship with
tides w hereas during the night hours PO¿¡ -P
concentration increased at the time o f high tide
and decreased during the low tide regime. F rom
Fig. 3a it is evident th at there w as significant vari­
Nitrite-nitrogen
At Stations I and II, N 0 2“ -N concentration v ar­
ied within the range o f 0.086 to 0.506 pg at 1“ 1
and 0.060 to 0.278 pg at 1“ 1 respectively which is
com paratively m uch low er than the levels re­
corded at St. III. A t Sts. I and II, N 0 2“ -N was
lower during the day tim e bu t rose at night. At
St. I, N 0 2 -N concentration during the night
hours increased the flood tide w hereas at St. II
nitrite showed significant variation between day
and night time but this did no t correlate with the
tide. A t St. III, N 0 2“ -N fluctuated from 0.252 to

• — Marine (St- I)
* — S w am p(S t. II )
a — Canal (St. III)

20
300
Ö
o>
=3. 200
Z
I
o
CO o
z 100

S tI St.II
24

04

z
0-2
CM
O
z

1-3

z
IV'
I
z

6 12 18 24 3 6 6 12 18 24 3 6
T im e (h o u rs) Time (h o u rs)
Fig. 3. Diurnal variations in PC>4 - -P, N O f-N , N 0 3~-N, N H 4+-N and gross and net primary production.

0.520 jug at 1 1 and showed only a m inor varia­
tions from 1500 hrs to 2100 hrs.
Nitrate-nitrogen
In the K akinada Bay (St. I) and m angrove w aters
(St. II) N 0 3“ -N show ed a m ore or less similar
range o f variation, the value being 1.03 to 6.62 ¡ig
at 1” 1 and 1.60 to 7.10 jug at 1“ 1 respectively. At
these two stations N 0 3~ -N concentration de­
creased gradually from 0600 hrs and reached the
lowest value at 1500 hrs and thereafter an
increasing trend w as noticed till 0300 hrs. Like P
0 | “ -P and N 0 2" -N , N 0 3_ -N concentration at
these two stations, w as independent o f the oscil­
lations in tides during day time. However, during
night hours N 0 3- -N concentration increased dur­
ing the high tide period and decreased during the
low tide regime. A t St. III, N 0 3- -N concentration
during day time fluctuated erratically with the fol­
lowing values: 0600 hrs: 14.86 / r g a t l - ^0 9 0 0 hrs:
26.40 ^rg at I - 1 ; 1200 hrs: 13.88 ¿rg at I ” 1;
1500 hrs: 25.08 ¿ig at 1—1; 1800 hrs: 9.98 /rg at I - 1
(Fig. 3c). At St. III there was intensive hum an
activity, with the m ovem ent o f m echanised fish­
ing boats. These bo ats leave around 0500 hrs to
0900 hrs and return around 1500 hrs. It might be
thought th at boat activity after 0500 hrs and
1500 hrs explains the peaks in N 0 3_ -N in the
freshw ater canal. H owever, this is an unlikely ex­
planation because one would expect similar peaks
in the concentrations o f the other nutrients which
were not observed.
A mmonia-nitrogen
A m m onia concentration w as low at St. I, the
highest and lowest values being 1.120 and
7.128 yug at I -1 respectively. At St. II, N H 4+-N
concentration w as high, the m axim um and m in­
imum values being 23.65 and 8.85 /tg at 1“ \ A t
St. I, N H 4+ -N concentration decreased gradually
from 0600 hrs to 1800 hrs w hereas at St. II, N
H ^ - N concentration showed a steady decrease
from 0900 hrs to 2100 hrs. At St. III, N H 4+-N
185
show ed no definite pattern in diel changes. At
St. I and II, unlike P O |- - P , N O f - N and N
0 3“ -N , N H / -N did no t show any relationship
with tides even during night hours which may
indicate th at the regeneration and utilization o f N
H 4+ -N is independent in each biotope.
Primary production
G ross production in the m arine (St. I) and m an­
grove w ater (St. II) w as 82.4 m g C m ~ 3 h ” 1 and
262.3 mg C m -3 h - 1 respectively. N et produc­
tion at Sts. I and II w as 3 1 .2 m g C m - 3 h -1 and
136.5 mg C m _ 3 h~ 1 respectively. High prim ary
production at St. II m ay attributed to the opti­
m um salinity condition ( 12%0 to 20 %o) prevailing
in the m angrove waters.
Discussion
M angrove ecosystem s are tidally affected to such
an extent th at tides are the m ain ecological fac­
tor serving as m ajor energy input (O dum , 1971).
It is generally believed th at coastal w etlands in­
cluding m arshes and m angroves export large
quantities o f dissolved inorganic nutrients to the
adjacent estuarine or coastal w aters through tidal
m ovem ent (W oodwell & W hitney, 1977; Bacon,
1981). In contrast Boto & W ellington (1988)
show ed th at only dissolved organic m aterials
are exported from the m angroves. They esti­
m ated th at nearly 24% o f the phosphate re­
quirem ent of the m angrove environm ent is sup­
plied from coastal w aters. G uerrero et al. (1988)
also confirm ed such a phenom enon and reported
th at no t only dissolved inorganic phosphate but
also dissolved nitrate w as im ported to the m an­
grove environm ent from coastal waters. The
present study show s th at concentration o f P
0 |~ - P , N 0 2” -N , and N 0 3_ -N were m ore or less
similar in the m angrove and bay w ater at first low
tide.
A t second low tide concentrations o f all these
nutrients were slightly higher in the bay w ater
than th at o f the m angrove. This indicates that
186
though there is a little leaching o f dissolved nu­
trients into the bay from m angrove at low tide this
phenom enon seems to be uncom m on at all low
tides. Secondly, the increased concentration o f P
0 | _ -P, N 0 3_ -N and N O ¿"-N during the high tide
period in the m angrove w ater indicates th at som e
am ount o f nutrients are im ported to this m an­
grove environm ent through the tidal m ovem ent.
How ever, this phenom enon is only apparent d u r­
ing the night hours and it is no identifiable d u r­
ing the day time. This might be due to continuous
utilization o f dissolved inorganic nutrients by
phytoplankton w hich is indicated by high prim ary
production values. In addition, the increased con­
centration o fP O !~ -P , N 0 2“ -N, N 0 3" -N at night
in the bay w ater suggests th at the freshw ater
flow from C om m ercial C anal acts as a source o f
these nutrients. The low concentration o f these
n utrients in the m angrove w ater clearly indicates
th at m angrove acts as a sink for land-driven n u ­
trients.
H eald (1971) and Boto & W ellington (1988)
show ed th at m angrove w aters are high in am m o­
nia concentration resulting from continuous de­
com position o f m angrove litter. H eald (1971)
show ed th at m ineralization o f m angrove organic
m atter is rapid in low to m oderate salinity. In the
p resent study high concentration o f am m onia was
observed in the m angrove w ater when the salin­
ity w as m oderate. Such an increase in am m onia
concentration and the prevalence o f optim um sa­
linity, 12%0 to 20%„ (Q asim & Devassy, 1971)
might be the reasons for the observed higher pri­
m ary production in the m angrove w aters than in
the K ak in ad a Bay.
Acknowledgement
The first author (VS) thanks the C ouncil o f Sci­
entific and Industrial R esearch, G ovt, o f India,
N ew Delhi for the financial support.
References
Bacon, P. R., 1981. Nutrient export from a Carribean la-
goonal ecosystem. Proc. Intl. Symp. Coastal Lagoons, Bra­
zil: 17-21.
Boto, K. G. & J. T. Wellington, 1988. Seasonal variations in
concentrations and fluxes of dissolved organic and inor­
ganic materials in a tropical, tidaly dominated mangrove
waterway. Mar. Ecol. Prog. Series. 50: 151-160.
Guerrero, G. R., D. R. Cervantes & I. A. Jimenez, 1988. N u­
trient variation during a tidal cycle at the mouth o f a coastal
lagoon in the Northwest of Mexico. Indian J. mar. Sei. 17:
235-237.
Heald, E., 1971. The production of organic detritus in a South
Florida estuary. Univ. Miami Sea G rant Tech. Bull. 6: 110.
Odum, E. P., 1971. Fundamental of Ecology. W.B. Saunders
Co, Philadelphia, 574 pp.
Odum, E. P., C. C. M clvor & T. J. Smith, 1982. The ecology
of the mangroves o f South Florida, a community profile.
U .S. Fish and Wild Life Service, Washington D.C., 144 pp.
Parsons, T. R., Y. M aita & C. M. Lalli, 1984. A manual of
chemical and biological methods of seawater analysis. Per­
gamon Press, Oxford, 173 pp.
Qasim, S. Z. & V. P. Devassy, 1971. The influence o f salin­
ity on the rate of photosynthesis and abundance of some
tropical phytoplankton. Mar. Biol. 12: 200-206.
Saraya, A., 1984. The physico-chemical properties o f a mussel
farm at Samaekho, Thailand. Proc. Asi. Symp. Mangr.
Envir. - Research Manag. 405-428.
Strickland, J. D. H. & T. R. Parsons, 1968. A practical hand
book o f seawater analysis. Fish. Res. Bd Canada, Ottawa,
Bull. 167, 311 pp.
Woodwell, G. M. & D. E. Whitney, 1977. Flux pond ecosys­
tem study: exchange of phosphorus between a slat marsh
and the coastal waters of Long Island Sound. Mar. Biol.
41: 1-6.
 ' 13,581
Hydrobiologia 247 : 187- 193, 1992 .
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 187
© 1992 Kluwer Academic Publishers. Printed in Belgium.

M ethanogenic bacteria in mangrove sediments

R. M o h a n ra ju 1 & R. N a ta ra ja n 2
1Centre o f Advanced Study in Marine Biology, Annamalai University, Parangipettai-608 502, India;
2Institute fo r Ocean Management, Centre fo r Water Resources, Anna University, Madras-600 025, India

K ey words: m ethanogenic bacteria, m angrove sedim ents

Abstract

The occurrence o f m ethanogenic bacteria in the K odiakkarai (10° 18' N; 79° 52' E) m angrove sedi­
m ents, where Avicennia spp are predom inant, w as studied. Trim ethylam ine under N 2:C 0 2 (80:20% v/v)
w as used as the substrate. M ost Probable N um ber (M P N ) o f m ethanogenic bacteria w as determ ined
for a period o f one year from July 1987 to June 1988 with m onthly sampling. The m ethanogenic b ac­
terial populations were found to be at the m axim um o f 1.1 x IO5 M P N g m “ 1 of wet sedim ent during
August 1987 and from February to June 1988. The bacterial num bers were found to decrease during
O ctober to D ecem ber 1987 with a m inim al value o f 3.6 x IO2 M P N g m “ 1 during D ecem ber 1987. E n­
vironm ental factors were correlated with the m ethanogenic bacterial population.

Introduction (Balch e ta l., 1979; D aniels e t a ! , 1984). The

M angroves are characteristic o f tropical coasts
and are the counterparts o f the tidal salt m arshes
o f the tem perate regions (W alsh, 1974; C hapm an,
1976). They are bordered by shallow w aters and
protected from wave action. This environm ent is
typically dom inated by large am ounts o f organic
m atter brought in during the m onsoon season and
form s a com plex ecosystem containing plant, an ­
imal and m icrobial life. The conversion o f this
organic m atter buried in sedim ents to natural gas
by bacterial decay under anaerobic conditions
yields ‘biogenic m ethane’. This gas is form ed by
the activities o f m ethanogenic bacteria.
M ethanogenic bacteria are of m uch interest due
to their role in the term inal dissim ilation o f or­
ganic carbon and also in nitrogen m etabolism by
utilization o f m ethylam ines (M athrani & Boone,
1985) and produce m ethane as their end product
m ethanogenic bacteria com m only occur in anaer­
obic environm ents, where organic m atter under­
goes decom position. M ah eta l. (1977) observed
th a t environm ents with an E h lower than
-2 0 0 mV have the largest num ber o f m ethano-
gens. The m ethanogens are characterized by their
ability to produce m ethane from H 2 + C 0 2, for­
m ate, acetate, m ethanol, m ethylam ines, dimethyl
sulphide, etc. In m arine sedim ents trim ethylam ine
is the m ajor substrate for m ethane production
and m etabolism o f m ethylated am ines is o f spe­
cial significance because high concentrations are
produced by m arine organism s (Yancey eta l.,
1982) and as a result o f bacterial decom position
o f choline and glycine betaine (King, 1988). The
aim o f the present study was to study the m eth­
anogenic bacterial population in these mangrove
sedim ents and the possible influence o f environ­
m ental param eters on them.
188
Description o f study area
The K odiakkarai m angrove is located along the
south east coast o f India (10° 18' N ; 79° 52' E)
(Fig. 1). It is bordered by the Bay o f Bengal on the
north east and the P alk strait on the southw est
with the presence o f some m angrove o f Avicennia
spp. W hen sam pled, the sedim ent w as black in
colour, saline and smelled strongly o f hydrogen
sulphide. The m angrove w as flooded twice daily.
Material and methods
M onthly sam pling o f sedim ent and w ater was
carried out for a period o f one year (July 1987 to
Ju n e 1988) during low tide. Sedim ent sam ples
were collected using a polycarbonate corer from
a depth o f 70 cm following the m ethod o f P eder­
sen & Sayler (1981) and transferred to a bottle
filling it completely so as to avoid any pockets of
air. Surface (overlying) w ater sam ples were col­
lected in sterile plastic containers. Both the sam ­
ples were returned to the laboratory within two
h ours under iced conditions (tem perature
< 5 °C).
Sedim ent and w ater tem peratures were m ea­
sured in situ with a m ercury therm om eter (accu­
S am p lin g site Swam p
79 52 E
Fig. 1. Location of sampling site.
racy o f + 0 .1 °C ). The p H and Eh o f the sedi­
m ent were m easured using a portable p H /E h
m eter (Industrial Electronics C orporation, India).
Rainfall d ata were obtained from the m eterolog-
ical unit o f the A vifauna P roject (B N H S ) at K o ­
diakkarai. Salinity w as determ ined by the m ethod
of Strickland & P arsons (1972). N itrite and ni­
trate were also determ ined by the m ethod of
Strickland & P arsons (1972). F o r total ph o sp h o ­
rus, sedim ent sam ples were ground and digested
by the m ethod o f R ochford (1951) and the co n ­
centration determ ined by the single solution
m ethod o f M urphy & Riley (1962). Organic car­
bon w as estim ated by the m ethod o f el W akeel &
Riley (1956).
O ne gram wet sedim ent w as transferred into
anaerobic serum bottles containing a m ixture of
99 ml o f sterile sea w ater and distilled w ater (in
the ratio o f 1:1). All m anipulations were done
under a co n stan t flow o f 0 2-free N 2. T he bottles
were then sealed with butyl rubber stoppers and
crim ped with alum inium seals.
The culture m edium used by M athrani & B oone
(1985) w as used. The anaerobic technique of
H ungate (1950) for m edia preparation as m odi­
fied by B ryant (1972) & Balch et al. (1979) using
serum bottles (M iller & W olin, 1974) w as fol­
lowed. T he m edium contained (in gram s per litre
IN DIA
ÍKodiakarai
(Light h o u s e
Mangrove
 189

o f 50% estuarine water): N H 4C1, 1.0; K 2H P 0 4, 250-

0.4; M gCl2, 0.1; yeast extract, 0.5; diatom aceous _ 200-
earth, 10.0; N a 2S, 0.2; N a H C 0 3, 2.0; Trim ethy- I 1 20'
lam ine hydrochloride, 1.91 (50 m M ), trace m in­ “ 100-

eral solution 10m l (W olin eta!., 1963), vitam in Ö 25-

solution 10 ml (W olin et al., 1963) and resazurin, i 50-
£ 25-
0.0001. The final p H w as adjusted to 7.2. The gas
phase w as N 2:C 0 2 (80:20 v/v) 15 ml o f m edium J y A S O N D J F M A M J
w as dispensed into 65 ml capacity serum bottles 1987 - 1988
w hich were pregassed with N 2:C 0 2. T he gassing Fig. 2. Monthly variation in rainfall.
w as continued for 3 to 5 m inutes to avoid oxygen
contam ination. The bottles were then stoppered
with black butyl stoppers, crim ped with alum in­ X— X w a ter (t°c)
ium seals and sterilized. Syringes flushed with N 2 30- X— X w ater (F^*)
sedim ent (fc)
were used for all transfers (M acy et al., 1972).
sedim ent (PH)
Filter-sterilized penicillin solution w as added to a 29-
7.9
final concentration o f 3.3 f i g m l” ’, as antibiotic
o
offers a selective advantage for m ethanogens o

(B ock & K andier, 1985). cu

The M o st P robable N um ber (M P N ) technique 3
O
w as used to enum erate m ethanogens (Heukeliki- a>
an & H einem ann, 1939; Sieberi & H attingh, 1967; £ 26-
a*
Belyaev & Finkelshtein, 1973; Zeikus & Winfrey,
1976; Jones & Paynter, 1980; Phelps & Zeikus, 25-

1985; F ranklin eta l., 1988). A three dilutions

three tube M P N w as prepared by transferring
serial dilution (IO - 2 ) o f 10, 1.0 and 0.1 ml into
the serum bottles using sterile syringes. The b o t­
tles were then incubated at 37 ° C for 20 days.
The M P N bottles were com puted using M cC rady
tables (1918). M ethane was m easured in a
Chem ito G as C hrom atograph (M odel 3800) with
a P o rap ak Q colum n equipped with a flame ion­
ization detector.
Results
Rainfall along the southeast coast o f India is
largely influenced by the northeast m onsoon with
a heavy rainfall at K odiakkarai (Fig. 2). The an ­
nual rainfall am ounted to 530.5 mm with a
m onthly average o f 44.21 mm. The m inim um
rainfall o f 1.8 mm w as observed during June 1988
and a m axim um o f 201 m m in O ctober 1987.
M onthly variations in tem perature are shown
in Fig. 3. The minim um surface w ater tem pera-
J y A S O N D J F M A M J
1987 - 1988
Fig. 3. Monthly variation in temperature and pH.
ture observed w as 25 ° C during O c to b e r-
D ecem ber 1987 and January 1988. The sedim ent
tem perature showed a m inim um value o f 26 ° C in
D ecem ber 1987, January and February 1988. The
m axim um value o f 30 ° C in July 1987, M arch and
M ay 1988.
T he p H values o f the surface w ater ranged from
a m inim um o f 7.2 in N ovem ber, 1987 to a m ax­
imum o f 8.1 in M ay, 1988; w hereas in sedim ents
it ranged from a m inim um o f 7.1 in Novem ber,
D ecem ber 1987 and January 1988 and a maxi­
m um value o f 7.6 in April and M ay, 1988. (Fig. 3).
R edox potential (Eh) values showed a m ini­
m um value o f -7 5 mV in N ovem ber 1987 and a
m axim um o f -1 2 9 mV in June 1988 (Fig. 4).
Salinity w as found to be lowest (5.5%0) during
N ovem ber, 1987 and highest (61.7%0) in Septem ­
ber, 1987 (Fig. 4).
190

The organic carbon content o f the sedim ent

Eh sedim ent -60
-130-
• — • Salinity
ranged from a m inim um value o f 1.8 mg 1” 1 in
12 0 -
-
July 1987 to a m axim um o f 4.2 mg g -1 in Sep­
-50 tem ber and N ovem ber 1987 (Fig. 5).
-115-
The M ost P robable N um ber for trim ethylam ine
-«0
utilizing m ethanogens are shown in Table 1.
>
£ The m ethanogenic bacterial population showed
-30
x :
Ui
a m axim um value o f > 1.1 x IO5 M P N g~ 1 o f wet
-95-
sedim ent during August, 1987. A decreasing trend
-20
w as observed from then on to reach the minimum
-8 5 -
value o f 3.6 x IO2 M P N g -1 during Decem ber,
-10
1987. Thereafter an increase in the population
-7 5 -
was noticed, with the m axim um value during
J y A S O N D J F M A M J February to June, 1988. The average m ethano­
1987 - 1988 genic bacterial population w as found to be
Fig. 4. Monthly variation in salinity and Eh. 5.9 x IO4 M P N g m “ 1.
The correlation co-efficient (r) values calculated
between the m ethanogenic bacteria (M B ) and
The nitrite concentration o f the w ater showed various environm ental param eters show ed posi­
a m inim um level o f 0.5 pg 1" 1 in June 1988 and tive correlation with tem perature, pH , Eh and
a m axim um o f 6.3 yug 1“ 1 in N ovem ber 1987. The
n itrate concentration of w ater show ed a m ini­
Table I. Methanogenic bacterial population.
m um level o f 0.2 ,ug 1~ 1 during M ay and June
1988, and the maxim um concentration of Month Bacterial numbers
5.5 yug 1_ 1 in O ctober and N ovem ber, 1987. The (M PN per gm. wet sediment)
total phosphorus concentration in sedim ent
July 1987 4.6 x IO4
show ed a m inim um o f 0.8 yug 1~ 1 during M ay and
August > 1.1 x IO5
June 1988, while the m axim um value o f 9.2 yug 1_ 1 September 4.3 x IO3
was recorded in Septem ber and O ctober 1987 October 3.9 x IO3
(Fig. 5). November 9.1 x IO2
December 3.6 x IO2
January 1988 2.7 x IO3
X X NO2
February > 1.1 x IO5
March > 1.1 x 10s
• — • N0 3
6. 0 - 10-0 April > 1.1 x IO5
X-— X P0 4
5-0- • — • Org. C 9-0 May > 1.1 x IO5
o> June > 1.1 x IO5
2 3.0- 7.0

5-0 Table 2. Correlation co-■efficient between methanogenic bac-

z terial population and environmental parameters.
Om
Z 3-0
«5 Temperature pH Eh Salinity
Sz '
0 .2 - 0.6746** 0.8842**** 0.8554**** 0.6343**

J y A S O N D J F M A M J Nitrite Nitrate Organic Carbon Total phosphorus

1987 - 1988
-0.6530** - 0.6489** -0.7813*** -0.5120NS
Fig. 5. Monthly variation in Nitrite, Nitrate, Total Phospho­
rus and Organic Carbon. P value: **<0.02; ***<0.01 and ****<0.001; N s>0.05.

salinity; while the nutrients nitrite, nitrate, total
p hosphorus and organic carbon showed negative
correlation (Table 2 ).
Discussion
Generally, m axim um tem perature w as recorded
during sum m er and m inim um during the m on­
soon as observed earlier by A nbhazhagan (1988),
K arthikeyan (1988) and S am path (1989) for the
K odiakkarai region. N o distinct seasonal pattern
w as observed for pH . The values ranged from 7.0
to 7.5 which are com parable to previous reports
by A nbhazhagan (1988) & K arthikeyan (1988).
W illiams (1981) attributed the variation in pH to
photosynthesis by subm erged vegetation in the
area.
D uring sum m er and early prem onsoon m axi­
m um salinity values were observed. The reasons
for high salinity values are due to the addition of
harvested salt heaped from salt p ans close by and
also due to the brine from the salt pan being
drained in the vicinity of the study area. The fresh
w ater discharge w as also low during this period.
W ith the advent o f prem onsoon showers the sa­
linity declined slightly because o f freshw ater run-
olf from sources like irrigation channels. The
n ortheast m onsoon starting in O ctober and con­
tinuing up to D ecem ber decreased the salinity
values to the minimum. The sam e trend has been
previously observed by A nbhazhagan (1988),
K arthikeyan (1988) & S am path (1989) in the K o ­
diakkarai m angroves and by V asantha (1989) in
the P ichavaram m angroves.
T he m axim um nitrite and nitrate concentra­
tions during the m onsoon were apparently due to
the influence o f seasonal freshw ater discharges
(Carlucci e ta l., 1970). Higher concentrations of
nitrite th an nitrate could be attributed to bacterial
reduction o f nitrate and oxidation o f am m onia by
nitrifying bacteria. D uring sum m er, the total
phosphorus concentration increased gradually
and reached a m axim um during the m onsoon
owing to enrichm ent by m onsoonal floods as
pointed out by C handran & R am am oorthi (1984)
for the P ichavaram m angroves. A nother reason
191
could be the regeneration o f phosphorus in the
b ottom m uds and the subsequent release o f the
sam e into the w ater column due to turbulence and
mixing caused by tides as well as the heavy winds
prevailing during this season.
N o definite trend w as observed in the sedim ent
organic carbon. H igher concentrations were ob­
served during m onsoon and postm onsoon sea­
sons probably due to the high rate o f freshw ater
inflow enriched with organic debris from land
run-off.
T he m ethanogenic bacteria are a m ajor com ­
ponent o f the m icrobial population in the m an­
grove sedim ent (R am am urthy et al., 1990). They
are restricted with respect to the n ature o f the
substrate utilized for m ethane form ation. F ran k ­
lin eta l. (1988) observed a tw o to three times
higher num ber of m ethanogens in a short Spartina
zone than in a tali Spartina zone, and they attrib­
uted this to the abundance o f m ore com pact Spar­
tina roots in the short Spartina zone. B artlett et al.
(1985) on m easuring m ethane flux from a coastal
salt m arsh found th at tem perature increase was
accom panied by an increase in the m ethane flux.
Jones & P aynter (1980) observed a decrease in
the m ethanogenic population which utilized hy­
drogen as a substrate on the m ud flats away from
the Spartina plants, showing a population level of
3.7 x IO4 to 1.2 x IO6 cells gm~ 1 o f dry sediment.
King (1988) found th at m ethane production in
sedim ents o f a hypersaline pond w as associated
with algal m ats which were apparently sources of
m ethylated amines.
O bservations by Phelps & Zeikus (1985) in the
sedim ents o f Lake M endota show ed a bacterial
population range o f IO3—IO5 cells g m -1 o f dry
sedim ent. Higher population levels were observed
during sum m er and lower ones during winter. The
change in population level w as due to the fall in
tem perature and acetate availability. The present
study o f m ethanogenic bacterial population from
the m angrove sedim ents show ed higher bacterial
num bers during the sum m er and prem onsoon and
lower values during the m onsoonal m onths. The
difference in population levels is m ainly due to the
decline of salinity due to freshw ater influx, tem ­
perature changes and changes in pH and Eh.
192
Trim ethylam ine w as an im portant precursor
(3 1 .5 % -6 1 .1 % ) in the estuarine intertidal sedi­
m ents o f Lowes Cove, M aine (King et al., 1983).
The im portance o f this com pound as a m ethane
precursor w as due to the abundance o f decom ­
posed plant m aterial in the sedim ent system.
T M A has been confirm ed as a non-com petitive
substrate for m ethane production (King, 1984) by
sulphate reducers. In sedim ents choline and be-
taine undergo initial attack by ferm entative b ac­
teria rather th an sulphate reducers. This results in
the production o f trim ethylam ine and acetate
which are then respectively channelled to m eth-
anogenesis and sulphate reducers.
In conclusion the presence o f m ethanogenic
b acteria in m angrove sedim ent appears to be in­
fluenced by environm ental param eters such as
tem perature, p H , Eh, salinity and other nutrients.
Acknowledgement
W e wish to th an k the D epartm ent o f N on-
conventional Energy Sources, G ovt, o f India, new
D elhi for funding this study.
References
Anbhazhagan, P., 1988. Hydrobiology and benthic ecology of
Kodiakkarai coastal sanctuary (Southeast coast of India).
Ph.D. Thesis, Annamalai Univ., India, pp. 208.
Balch, W. E., C. E. Fox, L. J. Magrum, C. R. Woese & R. S.
Wolfe, 1979. Methanogens: Re-evaluation of a unique bi­
ological group. Microbiol. Rev. 43: 260-296.
Bartlett, K. B., R. C. Hariss & D. I. Sebacher, 1985. M eth­
ane flux from coastal salt marshes. J. Geo-phy. Res. 90:
5710-5720.
Belyaev, S. S. & Z. I. Finkelshtein, 1973. M ethod o f count­
ing methane producing bacteria in media containing organic
substrates. Mikrobiologiya 42: 1102-1107.
Bock, A. & O. Kandier, 1985. Antiobiotic sensitivity of ar-
chaebacteria. In: C. R. Woese & R. S. Wolfe (eds), The
Bacteria, Vol. 8. Academic Press, New York: 525-544.
Bryant, M. P., 1972. Commentary on the Hungate technique
for culture of anaerobic bacteria. Am. J. Clin. Nutr. 25:
1324-1328.
Carlucci, A. F., E. O. Hertwig & P. M. Bowes, 1970. Biolog­
ical production of nitrite in seawater. Mar. Biol. 7: 161—
166.
Chapman, V. J., 1976. Mangrove vegetation. Cramer Vaduz.
Chandran, R. & K. Ramamoorthi, 1984. Hydrobiological
studies in the gradient zone of the Vellar estuary. I. Physico­
chemical parameters. Mahasagar-Bull. N at. Inst. Ocean-
ogr. 17: 69-77.
Daniels, 1., R. Sparling & G. D. Sprott, 1984. The bioener-
getic of methanogenesis. Biochim. Biophys. Acta 768: 113—
163.
el Wakeel, S. K. & J. P. Riley, 1956. The determination of
organic carhon in marine muds. J. du. Coun. 22: 180-183.
Franklin, M. J., W. J. Wiebe & W. B. Whiman, 1988. Popu­
lations of methanogenic bacteria in a Georgia salt marsh.
Appl. Envir. Microbiol. 54: 1151-1157.
Heukelikian, H. & B. Heinemann, 1939. Studies on the meth­
ane bacteria. I. Development of a method for enumeration.
Sew. Wks. J. 11: 426-435.
Hungate, R. E., 1950. The anaerobic mesophilic celluloytic
bacteria. Bact. Rev. 14: 1-49.
Jones, W. J. & M. J. B. Paynter, 1980. Populations o f meth­
ane producing bacteria and in vitro methanogenesis in salt
m arsh and estuarine sediments. Appl. Envir. Microbiol. 39:
864-871.
Karthikeyan, E., 1988. Organochlorine pesticides in the trop­
ical estuary (The Vellar) mangrove (Pichavaram) and
swamp (Kodiakkarai) in the east coast of India. Ph.D.
Thesis, Annamalai Univ., 151 pp.
King, G. M., 1984. On the metabolism o f trimethylamine,
choline and glycinebetaine by sulfate reducing and metha­
nogenic bacteria in marine sediments. Appl. Envir. M icro­
biol. 48: 719-725.
King, G. M., 1988. Methanogenesis from methylated amines
in a hypersaline algal mat. Appl. Envir. Microbiol. 54: 130—
136.
King, G. M., M. J. Klug & D. R. Lovely, 1983. Metabolism
of acetate, methanol and methylated amines in intertidal
sediments o f Lowes Cove, Maine. Appl. Envir. Microbiol.
45: 1848-1853.
Macy, J. M., J. E. Schnellen & R. E. Hungate, 1972. Use of
syringe method for anaerobiosis. J. Clin. Nutr. 25: 1318—
1323.
M ah, R. A., D. M. Ward, L. Baresi & T. L. Glass, 1977.
Biogenesis of methane. Ann. Rev. Microbiol. 31: 309-341.
Mathrani, I. M. & D. R. Boone, 1985. Isolation and charac­
terization of a moderately halophilic methanogen from a
solar saltern. Appl. Envir. Microbiol. 50: 140-143.
McCrady, A. H., 1918. Tables for rapid interpretation of fer­
mentative tube results. Can. J. Publ. Health. 9: 1153-1158.
Miller, T. L. & M. J. Wolin, 1974. A serum bottle modifica­
tion of the Hungate technique for cultivating obligate anaer­
obes. Appl. Microbiol. 27: 985-987.
Murphy, J. & J. P. Riley, 1962. A modified single solution
method for the determination o f phosphate in natural wa­
ters. Analyt. Chem. Acta. 12: 162-176.
Pedersen, D. & G. S. Sayler, 1981. Methanogenesis in fresh­
water sediments, inherent variability and effects of environ­
mental contaminants. Can. J. Microbiol. 27: 198-205.

Phelps, J. & J. G. Zeikus, 1985. Effect of fall turnover an
terminal carbon metabolism in Lake M endota sediments.
Appl. Envir. Microbiol. 50: 1285-1291.
Ramamurthy, T., R. Mohanraju & R. N atarajan, 1990. Dis­
tribution and ecology of methanogenic bacteria in man­
grove sediments of Pichavaram, east coast of India. Ind. J.
Mar. Sei. 19: 269-273.
Rochford, D. J., 1951. Studies in Australian estuarine hydrog­
raphy I. Introductory and comparative features. Aust. J.
mar. Freshwat. Res. 2: 1-116.
Sampath, K., 1989. Studies on the ecology of shore birds
(Aves: Charadriformes). Ph.D. Thesis, Annamalai Univ.,
202 pp.
Siebert, M. L. & W. H. J. Hattingh, 1967. Estimation of
methane producing bacterial numbers by the most proba­
ble number (M PN ) technique. Wat. Res. 1: 13-19.
Strickland, J. D. H. & T. R. Parsons, 1972. A practical hand­
book of sea water analysis. Bull. Fish. Res. Bd Can. 167:
310.
193
Vasantha, K., 1989. Studies on hydrobiology and decompo­
sition of macrophytes in Portonovo marine environments
(Southeast coast of India). Ph.D. Thesis. Annamalai Univ.,
132 pp.
Walsh, G. E., 1974. Mangroves: a review In: R. Reinhold &
W. Queen (eds), Ecology of Halophytes. Academic Press,
New York, 174 pp.
William, W. D., 1981. The limnology of saline lakes in W est­
ern Victoria. A Review of some recent studies. Hydrobio-
logia 82: 233-259.
Wolin, E. A., M. J. Wolin & R. S. Wolfe, 1963. Formation of
methane by bacterial extracts. J. Biol. Chem. 238: 2882—
2886.
Yancey, P. H., M. E. Clark, S. C. H and, B. D. Bowlus &
G. N. Somero, 1982. Living with water stress: evolution of
osmolyte system. Sei. 217: 1214-1222.
Zeikus, J. G. & M. Winfrey, 1976. Temperature limitation of
methanogenesis in aquatic sediments. Appl. Envir. Micro­
biol. 31: 99-107.
Hydrobiotogia 247 : 195- 207 , 1992.
V. Jaccarini & E. Martens (eds), The Ecology of Mangrove and Related Ecosystems. 195
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Mangroves as a habitat for fish and prawns

A. Sasekum ar, V. C. Chong, M. U. Leh & R. D ’Cruz

Department o f Zoology, University o f Malaya, 59100 Kuala Lumpur, Malaysia

Abstract

M angrove inlets and creeks in Selangor, M alaysia are the habitat for 119 species o f fish and 9 species
o f praw ns. The m ajority o f fish and all praw ns sam pled in the inlets were juveniles. The com m on fish
species in the inlets in term s of weight were Arius sagor, Ambassis gymnocephalus, Liza subviridis, Toxotes
jaculator, Sphyraena barracuda and Lates calcarifer. P raw ns were represented by juvenile Penaeus
penicillatus, P. merguiensis, P. indicus, Metapenaeus brevicornis and M . affinis. Sam ples from enclosure
trap s set on m udflats during ebbing w ater captured 37 species o f fish and 11 species o f praw ns. The role
o f m angroves as nursery and feeding grounds for fish and praw ns is reviewed in the light of recent work
in Selangor. It is apparent th at m angroves support fisheries by providing habitat and food.

Introduction did not carry the m angrove carbon signal, it was

T he im portance o f m angroves as habitat for fish
and praw ns continues to be a topic o f debate
am ong m arine ecologists (M aenae, 1974; R obert­
son & D uke, 1987; H atcher eta l., 1989). The
prevailing view o f m angrove-offshore interactions
is th at m angroves export large quantities o f de­
tritus to estuarine and near shore w aters (Boto &
Bunt, 1981; R obertson, 1986). This detritus is
either refractory and sinks to the bottom , presum ­
ably never enters the food chain, or is dispersed
over a large area w ithout any local im pact (e.g.
magnification o f benthic secondary production).
However, it should be noted th at m arine ecolo­
gists generally agree th at inshore areas constitute
a crucial part o f the life-support systems o f off­
shore populations by providing nursery and feed­
ing areas (Jan sso n e ta l., 1988).
The correlation between offshore praw n
catches and areal extent o f m angroves (M aenae,
1974; T urner, 1977) has been suggested to be
indicative o f the dependence o f juvenile praw ns
on m angroves rather than o f a food chain link
(H atcher e ta l., 1989). Though praw ns offshore
shown th at several species o f juveniles praw ns
collected from a m angrove inlet consum ed
m angrove-derived carbon (Rodelli e ta l., 1984).
C om parative analysis o f the tissue carbon o f the
com m on consum ers captured in the inshore and
offshore w aters in Selangor indicated th at praw ns
in m angrove inlets derived an average o f 65% of
their carbon from m angrove trees (Rodelli eta l.,
1984). Eleven species o f fish were also found to
have derived an average of 60% o f their carbon
from m angroves.
The coast o f central Selangor experiences
strong sem i-diurnal tides, with ranges o f 5 m and
above during extreme spring tides. The coast may
be classified as a high m esotidal environm ent.
The w aters are well m ixed vertically, and salini­
ties range from 20 to 32%0 (Chong, 1980). M ajor
dilution occurs from the discharge o f freshw ater
by the Klang and Langat Rivers.
The objective o f this paper is to review the role
o f m angroves as a habitat for fish and praw ns
based on our previous studies on the mangrove-
fringed Selangor coast, and on new evidence from
the present study.
196

'fR7n^
■ KUALA SELANGOR STUDY
AREA

BAGAN SUNGA1 BULOH

\* fe---
V. Í * KAMPONG SUNGAI JANGGUT
V- .

KLANS
RIVER

KETAM IS'.

KLANG IS:
■ENGAH,
IS.'i^

SELAT KERING ISä ¡CHE MAT/

• : ; zi n I S / . LUMUT
IS. 4
: ;
PINTU GEDONGJS ?

CAREY ISLAND

LE G E N D
S a n d b anks exposed at
e x t r e m e lo w t i d e

EM Mangroves
• D e p t h in f a t h o m s
Bund

Sampling sites

Fig. 1. The study area indicating sampling sites in Angsa Bank, Klang Strait and mangrove inlets on the Selangor coast. The
enclosure trap sampling sites are situated north and south of the Buloh River.

M aterials and methods
The study site consisted o f several habitats on the
coast o f central Selangor (Fig. 1). Sites sam pled
included (i) m angrove creeks and inlets, (ii) m ud­
flats, and (iii) the adjacent shallow w aters, which
consists o f the near inshore Klang Straits and the
far inshore A ngsa Bank. The A ngsa B ank and the
K lang Straits are less than 10 m etres in depth.
Eight m angrove inlets were sam pled during
daylight, tw o on m ainland Selangor and six on
the m angrove islands (Fig. 1). All inlets are on the
average 2 to 5 km long, 50 to 150 m wide at the
m outh and about 5 m deep in the lower reaches
at low tide. T he creeks are narrow (5 to 10 m
wide), shallow (3 m deep at high tide) and are
com pletely drained o f w ater during low spring
tides.
D ue to site differences, different sam pling tech­
niques were em ployed. Gili nets o f several mesh
sizes (1.3 cm, 2.5 cm, 5.1 cm, 7.6 cm and 15.2 cm)
and a 2 m wide bean traw l were used in sampling
inlets while creeks were sam ples with a bag net.
The fish com m unity o f the m udflats at tw o sites
north and south o f the m outh o f the Buloh River
were sam pled with a commercially operated en­
closure trap. T he A ngsa B ank (far inshore) and
K lang Strait (near inshore) w aters were sam pled
with praw n and fish otter traw ls. D etails of the
sam pling techniques and analysis o f catches are
given in C hong eta l. (1990) m acroscopic exami­
nation o f the gonads (K esteven, 1960). F or
praw ns, the ovaries or p etasm a and sperm ato-
phores were exam ined (Chong, 1980).
Results
Creeks and inlets
A total o f 119 species o f fish and 9 species o f
praw ns were found in the creeks and inlets o f
these m angrove sw am ps (Table 1). T he creeks
were dom inated by schooling fish species belong­
ing to the families A m bassidae, Engraulidae, C lu­
peidae, Leiognathidae, Eleotridae and M ugilidae
(Table 2). The praw ns were mainly penaeids,
197
nam ely Penaeus penicillatus, P. merguiensis, P. in­
dicus as well as Metapenaeus brevicornis and
M . affinis.
The inlets were also dom inated by the same
schooling fish and 70% o f the total catches in gili
nets consisted o f Ambassis gymnocephalus,
Sardinella melanura, Thryssa kammalensis,
T. hamiltoni and Stolephorus indicus (Table 3).
Flowever, in term s o f biom ass, the fish which co n ­
stituted as m uch as 62.5% o f the catch wer e Arius
sagor (15.5% ), Ambassis gmnocephalus (14% ),
Liza subviridis ( 8 % ), Toxotes jaculator (5% ) and
6 others, including Lates calcarifer which m ade up
3% o f the total catch. The beam traw l w as effec­
tive in capturing the praw ns as well some o f the
dem ersal fish (Table 4).
Mudflats
Thirty-seven species o f fish and 11 species of
praw ns (Table 5) were captured in the intertidal
m udflats. A pproxim ately 80% o f the total num ­
ber o f these fish species were also caught in the
inlets and creeks. F o r praw ns the species overlap
w as 55% .
The com position o f catches by weight and
num bers w as different on each sam pling occasion
(Table 6 ). T he m agnitude o f the catch appeared
to be dependent on the tidal am plitude. A higher
catch biom ass w as obtained during extrem e high
tides ( > 5 m). The percentage com position o f fish
and praw ns were also m arkedly different on each
sam pling occasion. C arnivores (catfish and cat-
fish-eel) and detritivores (mullets) dom inated the
catch on one occasion while the praw n Palaemon
styliferus w as dom inant on another occasion (T a­
ble 6 ).
Angsa bank and Klang straits
A total o f 102 species o f fish and 15 species of
praw ns w ere found in the area (Table 1). The fish
com m unity w as dom inated by m em bers o f the
families Sciaenidae, Cynoglossidae, Engraulidae,
C lupeidae and Leiognathidae. The praw n com-
198

Table I . Fish and prawn species found in the inshore waters of the Selangor coast, Malaysia, c = creeks/inlets; m = mudflats; n = near inshore waters;
f = far inshore waters.

Fish species

Abalistes stellaris f Gobiidae species A cmnf Protonibea diacanthus m

Ablennes hians c Gobiidae species B mf Psettodes erumei n
Acentrogobius caninus c Gobiidae species C mn f Pseudorhombus arsius c
Acentrogobius sp. A c Harpadon nehereus nf Pseudorhombus javanicus cnf
Alepes djeddaba f Ilisha ßigera cnf Rastreflliger kanagurta f
Allaneta forskali c Ilisha megaloptera cmnf Sardinella gibbosa n
Ambassis commersoni c Ilisha melastoma cf Sardinella melanura cmf
Ambassis gymnocephalus cmnf Johnius aneus c Sardinella fimbriata n
Ambassis kopsii cf Johnius belangerii cmnf Saurida tumbil cmnf
Amphotistius imbricatus c Johnius carutta cnf Saurida undosquamis f
Anguilla nebulosa c Johnius coitor cmnf Scatophagus argus cn
Anodontostoma chacunda cmf Johnius dussumieri cn Scomberoides commersonianus cf
Apogon hyalosoma c Johnius soldado cmn Scyllium marmoratum mn
Apogon quadrifasciatus mf Johnius sp. mnf Secutor insidiator cmnf
Arius caelatus cm Laputa cingalensis f Selaroides leptolepis f
Arius maculatus cn m Lates calcarifer c Secutor ruconius f
Arius sagor cm Leiognathus bindus cf Setipinna taty cmnf
Arius tonggol cm Leiognathus brevirostris cmnf Sideria thyrsoidea cm
Arius sp. A cf Leiognathus daura c Siganus canaliculatus cf
Arius venosus cmf Leiognathus fasciatus c Siganus javus cf
Arothron leopardus c Leiognathus lioneolatus c Siganus vermiculatus c
Atropus atropus f Leiognathus splendens c Sillago sihama cmnf
Bagridae sp. I c Liza argentea m Sphyraena barracuda c
Bagridae sp. II c Liza malinoptera cmf Sphyraena jello c
Batrachus grunniens cnf Liza subviridis cmnf Stigmatogobius sadanundio cm nf
Batrachocephalus mino c Lutjanus johni cf Stolephorus indicus cf
Boleopthalmus boddaerti c Lutjanus sanguineus cmf Stolephorus macroleptus m
Butis butis c Macrones gulio cmf Stolephorus macrops cnf
Callionymus schaapi nf Monodactylus argenteus c Stolephorus tri cmnf
Carangoides malabaricus cmnf Muraenesox cinereus mnf Synaptura commersoniana f
Caranx melampygus c Ophichthys rhytidodermatoides mf Tenualosa sinensis cmnf
Caranx sansun c Ophiocara porocephala c Therapon jarbua cmnf
Carangidae sp. c Opisthopterus tardoore f Thryssa mystax c
Chelonodon fluviatilis cm Osteogeneiosus militaris cm Thryssa dussumieri nf
Chiloscyllium indicum cnf Oesteogeneiosus sthenocephalus n Thryssa hamiltonii cmnf
Chirocentrus dorab cf Otolithus argenteus nf Thryssa kammalensis cmnf
Coilia dussumieri f Otolithes ruber cmnf Thryssa setirostris c
Cynoglossus macrolepidotusc mnf Otolithoides brunneus f Thunnidae f
Cynoglossus lingua cf Pampus argenteus mnf Thysanophrys indicus c
Dasyatis uarnak cf Pampus chinensis mnf Tilapia mossambica c
Dasyatis zugei cmnf Pennahia argentata m Toxotes jaculator c
Dendrophyssa russelli cmn Periophthalmodon scholsseri c Trichiurus savala cmnf
Dendroscorpaena sp. f Platax teira f Trypauchen sp. nf
Drepane longimana cmf Platycephalus asper f Trypauchen vagina cmnf
Drepane punctata cmf Platycephalus indicus c Tylosurus annulatus cm
Eleutheronema tetradactyluc mnf Platycephalus scaber cmnf Tylosurus leiurus c
Elops machnata c Platycephalus tuberculatus cmf Tylosurus strongylurus cm
Ephinephelus fario c Plotosus canius cm Upeneus sulphureus mnf
Ephippus orbis f Polynemus plebeius cn Uranoscopus sp. f
Formio niger nf Polynemus sextarius cmnf Uropterygius concolor c
Gastrophysus lunaris cmnf Pomadasys hastga cmf Zebrias quaga mf
Gerres abbreviatus cmnf Pomadasys maculatus cmf Zenarchopterus buffoni c
Gerres filamentosus c Potamolosa sp. A c Zenachopterus candovittatus c
Glossogobius giuris cm Prinobutis koilomatodon c
 199

Table 1. (Continued)

Prawn species

Macrobrachium equidens cm Parapenaeopsis coromandelica mf Penaeus indicus cmf

Metapenaeus affinis cmnf Parapenaeopsis gracillima mf Penaeus merguiensis cmnf
Metapenaeus brevicornis cmnf Parapenaeopsis hardwickii mnf Penaeus monodon c
Metapenaeus ensis c Parapenaeopsis hungerfordi mnf Penaeus penicillatus cmf
Metapenaeus lysianassa m Parapenaeopsis maxillipedo mnf Solenocera subnuda mnf
Metapenaeus stridulans f Parapenaeopsis sculptilis cmnf Trachypenaeus fulvus mf
Palaemon styliferus m

H abitat Num ber of species

group
Creeks/inlets Mudflat N ear in shore F ar in shore

Fish 119 37 58 92
Prawns 9 11 8 15

Table 2. Percentage abundance by number of major fish species caught by the bag net in two mangrove creeks in S. Sementa Kecil
and S. Perepat Kecil from November 1987 to May 1988, ( + ) means present in small numbers.

Species Site

S. Sementa Kecil S. Perapat Pooled

mean
Nov Dec Jan Feb Jan M ar May

Ambassis gymnocephalus 28.78 + 13.27 3.9 5.61 11.87

Butis butis + 1.29 1.69 12.99 0.46 1.9 2.18
Chelonodon flaviatilis 0.51 1.13 14.56 + 4.67 4.67 3.18 7.56
Eleutheronema tetradactylum + + + 0.92 0.93 4.44 0.41
Ilisha megaloptera 14.29 3.72 + 3.24
Ilisha melastoma 20.56 1.68
Leiognathus brevirostris 23.37 + 4.09
Liza malinoptera 1.33 1.13 8.44 13.99 27.1 12.06 3.63
Liza subviridis 33.49 18.69 53.65 5.93
Macrones gulio + 15.6 + 2.22 1.4
Periophthalmodon schlosseri 0.93 + 0.04
Plotosus canius + 0.92 + 0.19
Pomadasys hasta + 5.17 1.95 1.38 4.67 2.22 3.05
Scatophagus argus 0.71 + 3.74 1.52 1.59 2.21
Scomberoides commersonianus + + 4.47 12.99 3.29
Secutor insidiator 10.41 2.26 16.1 9.83
Stigmatogobius sadanundio + + + 0.46 + 0.19
Stolephorus indicus 4.18 6.3 + 0.87 1.98
Thryssa hamiltonii 5.82 73.02 2.86 11.26 0.93 + 11.35
Toxotes jaculator 0.81 + + 0.44
Tylosurus strongylurus + + 11.47 23.36 + 1.34
Zenarchopterus buffoni 10 2.91 3.05 22.08 + 12.15 13.33 6.32
Zenarchopterus candovittatus 2.1 28.27 20.87 15.5

Others 0.6 5.33 6.82 3.44 0.44 0.96 3.19 2.08

200

m unity w as dom inated by tw o penaeid species,
Parapenaeopsis sculptilis and P. hungerfordi.
S exual maturity
Ninety-seven percent o f the fish sam pled in the
inlets (including creeks) were im m ature. This was
in co n trast to 80% over the m udflats and 60% in
the nearshore w aters. O f the total fish population
in the far inshore area, only 29% were im m ature
(Table 7). All the praw ns sam pled in the inlets
were im m ature. Their carapace lengths ranged
from 7 to 28 m m CL. The percentage o f im m ature
praw ns in the shallow coastal w aters o f the Klang
Straits and A ngsa B ank were 8 and 38% respec­
tively.
D is c u s s io n
T he com m unity o f fish found in m angrove inlets
in Selangor resem bles th a t o f similar habitats

Table 3. Percentage abundance by numbers o f the major fish species caught by gili nets in the mangrove inlet S. Sementa Kecil
from November 1987 to May 1988, ( + ) means present in small numbers.

Species Month Pooled

mean
Nov Dec Jan May Aug Sep Oct

Ambassis commersoni + 0.65 0.2

Ambassis gymnocephalus 74.44 17.32 14.25 64.42 24.87 4.11 21.82 53.36
Anodontostoma chacunda + + + 2.07 + 3.64 0.43
Arius sagor + + + 4.11 0.22
Arius tonggol + 5.03 0.28 0.31
Arothron leopardus 0.19 0.04
Carangoides malabaricus 1.55 0.06
Chelonodon fluviatilis 0.59 0.23
Dasyatis uarnak 3.64 0.04
Dasyatis zugei + 1.82 0.06
Dendrophyssa russelli + + 5.45 0.37
Eleutheronema tetradactylum 1.04 + + 0.16
Ilisha megaloptera + + 10.96 0.27
Gerres abbreviatus 1.55 0.06
Johnius aneus 1.01 0.29
Johnius coiter + + 6.85 5.45 0.22
Johnius soldado + 5.03 0.19 0.25
Leiognathus brevirostris + 0.94 29.53 5.48 7.27 1.62
Leiognathus fasciatus 0.29 + 0.65 0.33
Liza malinoptera + 1.89 + 0.59
Liza subviridis 1.08 3.35 1.02 2.56 4.11 + 1.58
Osteogeniosus militaris 2.07 3.64 0.12
Pomadasys hasta 0.73 2.79 + + + 0.49
Sardinella melanura 2.64 17.88 80.76 15.30 28.50 24.28
Secutor insidiator + 1.95 0.59
Setipinna taty + 1.15 10.96 0.51
Stolephorus indicus 3.67 5.03 + + 1.58
Stolephorus macrops 4.47 0.16
Stolephorus tri + 8.22 38.18 0.55
Thryssa hamiltonii 2.69 20.67 0.19 2.16 1.55 9.59 3.64 2.7
Thryssa kammalensis 10.90 5.03 3.64 4.15 19.18 5.99
Toxotes jaculator 1.12 + 0.37 0.7

Others 1.85 13.4 1.64 4.98 3.12 16.43 5.45 1.66

 201

Table 4. Percentage abundance by numbers of the major and prawn species caught by beam trawl in the mangrove inlet S . Sementa
Kecil from Januray 1988 to August 1989. ( + ) present in small numbers.

Species Month Pooled

mean
Jan Apr May Jun Jul Aug

Fish
Ambassis gymnocephalus 18.46 4.02 89.90 56.18 8.33 3.31 34.94
Thryssa hamiltoni 22.81 1.21 16.38 12.14 1.24 14.53
Ilisha megaloptera 0.43 0.22 5.17 39.52 3.72 2.48
Thryssa kammalensis 43.32 1.31 3.66 + 23.01
Leiognathus brevirostris 1.25 3.66 6.19 26.86 1.90
Glossogobius giuris 10.07 0.76 0.10 0.93 5.24 13.22 1.28
Pomadasys hasta 17.79 0.31 0.06 0.93 3.57 5.37 1.01
Scatophagus argus 9.73 0.52 2.55 4.24 4.76 2.07 1.98
Chelonodon fluviatilis 1.68 3.68 0.38 2.95 2.62 10.74 2.73
Leiognathus daura 18.31 2.45 0.14 10.08

Others 42.27 4.59 1.82 5.76 17.63 33.47 6.06

Prawn
Penaeus penicillatus 21.21 23.89 20.74 36.15 43.14 22.81 28.65
Metapenaeus brevicornis 20.61 46.02 61.17 17.79 1.96 15.79 30.01
Penaeus merguiensis 9.70 17.70 6.91 31.40 31.37 43.86 21.93
Penaeus indicus 48.48 2.65 2.66 8.97 11.76 1.75 13.53
Metapenaeus mutatus 7.08 5.32 1.85 11.76 10.53 3.88
Macrobrachium equidens 2.65 2.66 1.58 3.51 1.68
Penaeus monodon 0.26 1.75 0.21
Parapenaeopsis sculptilis 0.53 0.10

elsewhere (Little e ta l., 1988; Pinto, 1988; R ob­
ertson & D uke, 1987). Six species of small fish of
low and m edium econom ic value com prised 70%
o f a total o f 21670 specim ens (119 species) ex­
am ined. The m ost abundant, Ambassis gymno­
cephalus, is n o t eaten by m an because o f its small
size. H ow ever, others like the Sardinella sp.,
Thryssa spp. and Stolephorus sp. are consum ed.
Although the catfish-eel, Plotosus canius, was
rarely captured in the gili nets or the beam trawl,
it constitutes an im portant fishery and is norm ally
landed by eel trap or hook and line (Leh, u npub­
lished).
Fish occurring in creeks feed on benthic food
resources including crabs, penaeids, insects,
m angrove detritus, diatom s, m acroalgae, and on
pelagic fauna like Acetes and copepods (Leh &
S asekum ar, 1991). The mullets, Liza malinoptera
and L. subviridis consum e significant quantities of
m angrove detritus (Chong, 1977; Sasekum ar
eta l., 1984). P redation o f benthic m angrove in­
vertebrates including sesarm id crabs (Chiroman­
thes spp.), M etaplax spp., bivalves, gastropods
and sipunculids by catfish eel and catfish is com ­
m on (S asekum ar et al., 1984). In addition, m an­
grove invertebrates liberate larvae into coastal
w aters th at becom e food for planktivores. M a­
cintosh (1979) estim ated th at two species o f fid­
dler crabs ( Uca spp.) in K uala Selangor produced
200000 m egalopae m - 2 y r - 1 . C rab m egalopae
are im portant food for several species o f ancho­
vies (Engraulidae) and Ambassis spp. (pers. obs.).
The praw ns in the m angrove inlets are repre­
sented largely by juvenile penaeids as has been
observed in Q ueensland (R obertson & Duke,
1987). Several species o f praw ns which include
the predom inant species o f Penaeus merguiensis,
P. penicillatus and P. indicus have been shown to
use these inlets as nursery areas (C hong eta l.,
1990). Metapenaeus brevicornis, M . affinis and
202

Table 5. Percentage abundance by numbers of the major fish and prawns species caught by the enclosure net in intertidal mudflats,
north and south of the river the River Buloh, Selangor, Peninsular Malaysia, from January 1988 to August 1989.

Species M onth Total Pooled

mean
Jan Aug Oct

Fish
Macrones gulio 16.82 46.40 63.22 26.50
Cynoglossus macrolepidotus 10.28 14.70 15.24 40.22 13.65
Arius tonggol 20.56 8.21 28.81 9.75
Liza subviridis 14.02 2.45 7.22 23.69 6.99
Pomadasys hasta 3.27 6.34 11.23 20.84 6.72
Arius caelatus 20.56 20.56 5.92
Ambassis gymnocephalus 1.73 15.24 16.97 4.64
Johnius coiter 1.64 11.38 3.74 16.76 6.72
Arius sagor 14.71 14.71 3.70
Thryssa kammalensis 0.29 9.89 10.18 2.62

Others 12.85 8.50 22.73 44.09 12.78

Prawn
Palaemon styliferus 45.18 75.82 121 70.34
Metapenaeus brevicornis 41.03 3.07 3.62 47.72 4.36
Metapenaeus mutatus 7.69 16.67 12.86 37.22 13.23
Penaeus indicus 25.64 7.02 32.66 1.43
Parapenaeopsis coromandelica 12.82 17.54 0.65 31.01 3.03
Penaeus merguiensis 7.69 4.39 3.68 15.76 3.86
Macrobrachium equidens 5.13 2.19 2.91 10.23 2.87
Penaeus penicillatus 3.51 3.51 0.44
Parapenaeopsis gracillima 0.44 0.44 0.06
Metapenaeus lysianassa 0.26 0.26 0.22
Parapenaeopsis hardwickii 0.19 0.19 0.17

Penaeus merguiensis occurred together both in the
inlets and the adjacent m udflats. However, there
is evidence o f spatial distribution o f species within
the inlet. The three Penaeus species are restricted
mainly to the upper and m iddle reaches o f the
inlet, w hereas the tw o Metapenaeus species are
found in the lower reaches and m udflats (Figs 2a
& 2b). Spatial partitioning o f juvenile praw ns in
m angrove estuaries has also been observed in
E ast Africa (de Freitas, 1986).
The presence o f juvenile praw ns in m angrove
inlets m ay be attributed to three m ain factors:
shelter from predators, high food abundance and
turbidity. Although the catfish eel and ariid ca t­
fish are com m on in the inlets, praw ns could take
refuge from these predators am ongst the prop
roots at high tide. How ever, praw ns are vulner­
able at low tide as the w aterw ays do no t have sea
grass cover or other refugia. Therefore, high food
abundance is m ore likely the m ain benefit to ju ­
venile praw ns. Turbidity (including food avail­
ability and tem perature) is know n to influence the
m ovem ent and distribution o f juvenile fish in es­
tuarine w aters (B laber & Blaber, 1980; C yrus &
Blaber, 1987). Similar studies have n o t been done
with praw ns though high turbidity could be im­
p o rtan t in reducing predation.
Penaeid praw ns are opportunistic generalists
in their food habits, being able to consum e a v a­
riety o f anim al and plant m atter and bacteria
(C hong & Sasekum ar, 1981; Leh & Sasekum ar,
1984; M oriarty, 1976).
The abundance o f food m aterial on the surface
o f m ud or sand-m ud substrates associated with
 203

Table 7. Percentage composition of the maturity stages of fish

o q 0X
o <N ox q and prawns sampled in the study area (Selangor, Peninsular
Ü Ö O h r--.
tí- w Malaysia).

-tí H abitat Total Immature Maturing M ature

<tí co numbers
\¿: On
S-t
<D
xi io q
S ■tí X Inlets fish 15894 97.4 1.4 1.2
o 'ÑT s^—✓
prawn 935 100 0 0

Mudflats fish 2909 55.7 18.2 26.1

q
Os < tí prawn 2319 49.2 13.4 37.4
O
*( t í 3
Near inshore fish 1619 60.7 15.6 23.7
X prawn 253 8.3 68.9 22.8

Far inshore fish 8760 29.1 41.5 29.4

of the total catch).

oo oo
■tí" \¿ prawn 3886 37.8 31.4 30.8

o
Qh xi <N oX
a
o
'S
3 OO OO Os OO
the m angrove environm ent provides a conducive
U environm ent for the growth o f praw ns. O n aver­
age there are IO6 m eiofauna m -2 (1 -2 g m -2 dry
Table 6. Biomass and numbers of fish and prauwns captured in the enclosure trap subsamples (10%

m ass) on the surface layers o f subtidal m ud sub­

X
a■x strates (Coull & Bell, 1979). The m ud substrate
£ in m angroves contains 0 .5 -1 .Ox IO6 m eiofauna
m ~2 (Sasekum ar, 1981). M icrofauna such as cil­
iates and flagellates are also abundant in m an­
a. grove m ud in densities o f 1.2 ( + 2 .3 )x 106 m -2
S
CÖ and 1.5 ( + 2.6) x IO6 m -2 respectively (Alongi,
oo
1986). M ean bacterial abundance and produc­
tivity o f m angrove sedim ents is estim ated at
<u
3 1. 1 x 10" cells g -1 dry weight and
£
Xi Q 1 . 6 g C m “ 2 d _I respectively (Alongi, 1988). In
bß H
x O C/5 contrast, the surfaces o f sandy substrates are an
»> -Ö
£ im poverished environm ent, where the m eiofauna
O Ü
^
tc 'S a are distributed within a broad depth o f 10 to 15 cm
below the surface (Coull & Bell, 1979).
-c -tí -tí
H atcher eta l. (1989) have suggested th at fac­
JD JD jD
’S 3 3
CQ 03 03 tors related to freshw ater outflow m ay be im por­
bß
00 bß
CO ta n t in the correlation betw een offshore catches of
0-H
o O O penaeid praw ns and m angrove sw am p area. Our
-tí - tí -tí
tío S
o
S-H
O studies in Selangor indicate th at the occurrence of
tí tí juvenile praw ns cannot be related to freshw ater
c
_o tí 3
tío qtí
Ttí3 ftí outflow as they also occur in abundance in inlets
•o
tí ’O
tí
o located on m angrove islands (off Port Klang)
J S S s
w here there are no rivers or land runoff (except
following heavy rainfall).
bß
G iven the present knowledge o f the distribu­
tí
tí < tion o f praw ns, it appears th at som e species of
Q praw ns are m ore linked to m angroves th an oth-
204

40
BSSSSSSSS8
35 LOWER

30

S 25

i 20

15

IO
i
9
1
IO
ISSÛ
11
12
13
14
15 17
18
19
20
21
22 23 25 27

C A R A P A C E LENGTH (mm)

Penaeus penicillatus

60

LOWER
50 - C" ' I
M/FLATS
40 mmmM
MODLE
ffiHWtfWWa
30 UPPER

20

io -

IO 12 14 16 18 20 22 24 26 28
11 13 15 17 19 21 23 25 27
CARAPACE LENGTH (mm)

Penaeus merguiensis

70

60

50

40

30

20

IO

IO 12 14 16 18 20 22 24 26 28
11 13 15 17 19 21 23 25 27
CARAPACE le n g th <mm) Penaeus indicus
Fig. 2a & 2b. Size frequency distribution of six prawn species by location in the upper, middle and lower reaches of the man­
grove inlet (Sungai Sementa Kecil) and the adjacent mudflat.
 205

140
gggggaggft
LOWER
120
c u : .1
M/FLATS
FREOflCVfê)
100
MIDDLE
80
tMMWBB
UPPER
60

40

20

13 15 17
a io 12 16
C A R A P A C E LENGTH ( m m )

Metapenaeus brevicornis

70

LOWER
60 -
M/FLAT5
50
MODLE
FREQUENCY'(?^

40 IM B H T O
LFFER
30

20

IO

i 11 13 15 17 10 21 23
10 12 14 16 IS 20 22
C A R A PA C E LENGTH ( m m )

Metapenaeus mutatus

LOWER
M
M ODLE
löUUÖUPPPB
UPPER

Macrobrachium spp.

C A R A PA C E LENGTH ( m m )

Fig. 2b.
206
ers. Penaeus merguiensis, P. penicillatus, P. indi­
cus, M etapenaeus brevicornis and M . affinis m ay be
m ore dependent on m angrove estuaries than other
species. T hus it is possible th at the species com ­
position o f offshore catches o f praw ns in M alay­
sia m ay change with the reclam ation or destruc­
tion o f large tracts o f m angroves.
Fish and praw ns generally occur in greater
num bers and biom ass in m angrove habitats than
in adjacent near shore habitats (Thayer eta l.,
1987; R obertson & D uke, 1987; Blaber eta l.,
1989; Chong eta l., 1990). The abundance o f a
variety o f food item s and the turbid w aters appear
to be the m ain contributing factors. Although ju ­
venile praw ns m ay not derive nutrition directly
from m angrove detritus (S toner & Zim m erm an,
1988), it is ap p aren t th at significant quantities of
m angrove carbon are being assim ilated through
secondary sources (Rodelli eta l., 1984). W e
therefore conclude th at m angroves provide sig­
nificant support for fish and praw n com m unities.
Acknowledgements
W e th an k Jalal M o khtar and Jam ri T ohid for
field assistance, and C heong Nge, G . H . Wee,
P. K. G eeth a and Siti N uraini for assistance in
the laboratory. W e also thank D r. S usan Lim for
her critical com m ents on the m anuscript. O ur
studies were supported by research grants from
A ID A B (A ustralia) and R & D funds from the
U niversity o f M alaya.
References
Alongi, D. M., 1986. Quantitative estimates of benthic
protozoa in tropical marine systems using silica gel: a
comparison of methods. Estuar. coast. Shelf Sei. 23: 443-
450.
Alongi, D. M., 1988. Bacterial productivity and microbial bio­
mass in tropical mangrove sediments. Microb. Ecol. 15:
59-79.
Blaber, S. J. M. & T. J. Blaber, 1980. Factors affecting the
distribution of juvenile estuarine and inshore fish. J. Fish.
Biol. 17: 143-162.
Blaber, S. J. M., D. T. Brewer & J. P. Salini, 1989. Species
composition and biomass of fishes in different habitats of
a tropical Northern Australian estuary: their occurrence in
the adjoining sea and estuarine dependence. Estuar. coast.
Shelf Sei. 29: 509-531.
Boto, K. G. & J. S. Bunt, 1981. Tidal export of particulate
organic matter from a N orthern Australian mangrove sys­
tem. Estuar. coast. Shelf Sei. 13: 247-255.
Chong, V. C., 1977. Studies on the small grey mullet Liza
malinoptera. J. Fish. Biol. 11: 293-308.
Chong, V. C., 1980. The biology of the white prawn Penaeus
merguiensis de M an (Crustacea: Penaeidae) in the Pulau
Angsa Klang Strait waters (Straits of Malacca). M. Sc.
thesis. D epartm ent of Zoology, University of Malaya,
K uala Lumpur. 141 pp.
Chong, V. C. & A. Sasekumar, 1981. Food and feeding habits
of the white prawn Penaeus merguiensis. Mar. Ecol. Prog.
Series 5: 185-191.
Chong, V. C., A. Sasekumar, M. U. Leh & R. D ’Cruz, 1990.
The fish and prawn communities of a Malaysian coastal
mangrove system, with comparisons to adjacent mudflats
and inshore waters. Estuar. coast. Shelf Sei. 31: 703-722.
Coull, B. C. & S. S. Bell, 1979. Perspectives in marine meio-
faunal ecology. In R. J. Livingston (ed.), Ecological Pro­
cesses in Coastal and Marine Systems. Plenum Press, New
York: 189-216.
Cyrus, D. P. & S. J. M. Blaber, 1987. The influence o f tur­
bidity on juvenile marine fishes in estuaries. Part I. Field
studies at Lake St. Lucia on the southeastern coast of
Africa. J. exp. mar. Biol. Ecol. 109: 53-70.
de Freitas, A. J., 1986. Selection of nursery areas by six
Southeast African Penaeidae. Estuar. coast. Shelf Sei. 23:
901-908.
Hatcher, B. G., R. E. Johannes & A. I. Robertson, 1989. Re­
view o f research relevant to the conservation of shallow
tropical marine ecosystems. Oceanogr. Mar. Biol. annu.
rev. 27: 337-414.
Jansson, B. O., A. D. McIntyre, S. W. Nixon, M. M. Pam-
atmat, B. Zeitzschel & J. J. Zijlstra, 1988. Coastal-offshore
interactions - an evaluation of presented evidence. In B. O.
Jansson (ed.), Coastal-Offshore Ecosystem Interactions.
Springer-Verlag: 357-363.
Kesteven, G. L., 1960. M anual of field methods in fisheries
biology. FAO M anuals in Fisheries Science. No. 1, FAO,
Rome. 152 pp.
Leh, M. U. & A. Sasekumar, 1984. Feeding ecology of prawns
in shallow waters adjoining mangrove shores. In E. Soe-
padmo et al. (eds), Proc. Asian Symp. Mangrove Environ­
ment - Res. & Management: 331-353.
Leh, M. U. & A. Sasekumar, 1991. Ingression of fish into
mangrove creeks. In Alcala (ed.), Proc. Asean-Australian
‘Living Resources in Coastal Areas’ Project. 30 Ja n -1 Feb.
1989. Manila: 495-501.
Little, C., P. J. Reay & S. J. Grove, 1988. The fish commu­
nity of an East African mangrove creek. J. Fish. Biol. 32:
729-747.
Macintosh, D. J., 1979. Predation of fiddler crabs ( Uca spp.)
in estuarine mangroves. In Srivastava, P. B. etal. (eds),

Proc. Symp. mangrove and estuarine vegetation in South­
east Asia. University of Agriculture, Malaysia: 101-110.
Maenae, W., 1974. Mangrove forests and fisheries. FAO,
Rome. 35 pp.
Moriarity, D. J. M., 1976. Quantitative studies on bacteria
and algae in the food of the mullet Mugil cephalus L. and the
prawn Metapenaeus bennettae (Racek & Dali). J. exp. mar.
Biol. Ecol. 22: 131-143.
Pinto, L., 1988. Population dynamics and community struc­
ture of fish in the mangroves of Pagbilao, Phillipines. J.
Fish. Biol. 33 (Supplement A): 35-43.
Robertson, A. I., 1986. Leaf-burying crabs: their influence on
energy flow and export from mixed mangrove forests
{Rhizophora spp.) in northeastern Australia. J. exp. mar.
Biol. Ecol. 102: 237-248.
Robertson, A. I. & N. C. Duke, 1987. Mangroves as nursery
sites: comparisons of the abundance and species com­
position of fish and crustaceans in mangroves and other
nearshore habitats in tropical Australia. Mar. Biol. 96:
193-205.
207
Rodelli, M. R., J. N. Gearing, P. J. Gearing, N. Marshall &
A. Sasekumar, 1984. Stable isotope ratio as a tracer of
mangrove carbon in Malaysian ecosystems. Oecologia 61:
326-333.
Sasekumar, A., 1981. The ecology of meiofauna on a Malayan
mangrove shore. Ph. D. thesis. Departm ent of Zoology,
University of Malaya, K uala Lumpur. 183 pp.
Sasekumar, A., T. L. Ong & K. L. Thong, 1984. Predation of
mangrove fauna by marine fishes. In E. Soepadmo et al.
(eds), Proc. Asian Symp. Mangrove Environment - Res.
and Management: 378-384.
Stoner, A. L. & R. J. Zimmerman, 1988. Food pathways as­
sociated with penaeid shrimps in a mangrove-fringed estu­
ary. Fish. Bull. 86: 543-551.
Thayer, G. W., D. R. Colby & W. F. Hettler, Jr., 1987. Uti­
lization of red mangrove prop habitat by fishes in South
Florida. Mar. Ecol. Prog. Series 35: 25-38.
Turner, R. E., 1977. Intertidal vegetation and commercial
yields of penaeid shrimp. Trans, am. Fish. Soc. 106: 411-
416.
■ -
Hydrobiologia 247 : 209 - 214 , 1992.
V. Jaccarini & E. Martens (eds), The Ecology of Mangrove and Related Ecosystems. 209
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Mariculture development in Kenya: alternatives to siting ponds in the

mangrove ecosystem

Joseph R asow o
M oi University, Zoology Department, B ox 3900, Eldoret, Kenya

K ey words: Artemia, salinas, m angrove, pond, shrim p, praw n

Abstract

P ond culture o f brackishw ater praw ns and shrim ps has recently generated m uch interest in Kenya. The
m angrove areas are the target zones for the construction o f these ponds. W ith the increasing aw areness
o f the unique ecological role played by the m angroves, there is an urgent need to stop the conversion
o f m angrove sw am ps into aquaculture ponds. To develop pond aquaculture w ithout destroying the
m angroves, a shift from tide-fed to pum p-fed pond system s is recom m ended in order to divert the farming
from the m angroves to higher grounds. ‘M angrove-friendly’ m ariculture practices like pen, cage and raft
culture are discussed. M ethods o f efficiently utilising the already destroyed m angrove areas are consid­
ered.

Introduction surge o f interest in the pond culture o f brackish­

M angroves and other coastal w etlands are highly
com plex and productive ecosystem s which serve
a wide variety of useful functions including pre­
vention o f coastal erosion, encouraging soil dep­
osition, providing food, shelter and sanctuary for
birds and m am m als, providing spawning, nursery
and forage ground for a wide variety o f aquatic
organism s, and an organic food factory through
its litter production (Bailey, 1988; Singh, 1987;
Saenger eta l., 1983; M acN ae, 1974; Turner,
1977). M angroves thus support im portant coastal
and nearshore fisheries and indeed good correla­
tions have been found between praw n and fish
yields and the area o f m angroves adjacent to the
fishing ground (Brusher, 1974; M acN ae, 1974).
Despite these unique physical, biological and eco­
nom ical functions, m angroves and other coastal
w etlands are being destroyed with little regard for
the consequences. This has been due to an up­
w ater praw ns and shrim ps as a result o f the high
dem and o f these products in the m arkets o f W est­
ern Europe, Jap a n and U S A . The export poten­
tial has increased so m uch th at m ariculture has
becom e established as a significant source o f for­
eign exchange earnings in a num ber o f develop­
ing countries. The success story o f countries like
E cuador has attracted w orld wide attention and
encouraged m any developing tropical countries
with large stands o f m angrove forests to invest in
shrim p m ariculture developm ent.
K enya has a total o f about 59980 h a of m an­
grove forests on its 450 km long semi arid coast­
line (D oute eta l., 1981). These m angroves have
been cleared m ainly for salt farm ing and for w ood
(fuel and building purposes). Although aquacul­
ture w as introduced in K enya soon after the first
w orld w ar, well over 250 h a o f the total pond area
in the country has been devoted for the culture of
freshw ater finfish especially Tilapia (Alíela, 1986).
2 1 0

The m ariculture subsector has been lagging be­
hind and it was in 1978 th at a pilot project w as
set up to test the technical and econom ic viability
o f brackishw ater aquaculture o f shrim ps on the
K enya coast. The production levels o f an average
o f 525 kg h a - 1 y r _ 1 has created considerable in­
terest am ong private sector entrepreneurs and n a ­
tional policym akers resulting in plans being draw n
up for large scale com m ercial operations.
This p aper proposes a cautious approach to
the envisaged m ariculture developm ent which
w ould involve the integration o f aquaculture
with the m anagem ent o f m angroves and the effi­
cient utilisation o f already reclaim ed m angrove
areas.
Efficient utilisation of already reclaimed man­
grove lands
orated in controlled areas to produce brine for
crystallising salt. A typical saltw orks in K enya
consists o f a series o f ponds w here the brine is
transferred as it gets progressively m ore saline.
As m any as six transfers are involved from the
first reservoir through a series o f evaporation
ponds to a serving pond and finally to the crys-
tallisers (Fig. 1). The first reservoir ponds which
generally occupy 7 -1 5 % o f the saltw orks total
area are indeed ready-m ade extensive ponds suit­
able for fish and shrim p farming. The sea w ater
in the reservoir is always freshly draw n and has
a salinity which does not differ m uch from th at of
the ocean since the pum ps are constantly in o p ­
eration. U sing the estim ates o f Sam m y & Tyler
(1986) th a t some 65 tonnes o f sea w ater are
needed to produce one tonne o f salt, a solar salt
producer could culture plankton feeding anim als

P ond aquaculture is currently recognized w orld­

wide as the m ajor culprit in the large scale clear­
ing o f m angrove forests and other coastal w et­
lands (Bailey, 1988; Singh, 1987). So far in Kenya,
it is salt farm ing which has been mainly respon­
sible for the destruction o f large tracts o f m an ­
groves with 50% o f the total salt farm acreage in
the coast being located within form er m angrove
FA C ILITIE S EVAPORATION P O N D -2 /p um p
areas (Jira, pers. com m .). The N gom eni aquac­
EVAPORATION
ulture pilot project executed under the F A O / P O N D -1

U N D P funding is responsible for the destruction

o f about 60 h a o f m angrove forests in the N go­
meni creek. Utilising the reclaim ed m angrove land
m ore efficiently rather than further clearing of o /EVAPORATION
S / POND- 3
m ore m angroves is an avenue which should be
considered. O ne viable alternative is for the solar
salt producing com panies to diversify into aquac­
ulture. A nother alternative is the intensification
o f pond aquaculture instead o f the extensive
m ethod which was used in the F A O /U N D P pilot
project.

PUMP MAYBE INSTALLED AS

Integrated solar salt - aquaculture production NECESSARY

Irrespective o f locality, the solar salt m aking pro- p¡g I Generalized layout of a saltwork in Kenya (Source:
cess is fairly uniform , with sea w ater being evap- Yap & Landoy, 1986).

like oysters, m ussels and milkfish on the plankton
content o f this volum e o f pum ped sea water.
T hese reservoir ponds could also be used as ex­
tensive shrim p ponds in which the shrim ps could
subsist and grow on the naturally occurring food
in the ponds. Considering th at the physical facil­
ities, civil w orks, and equipm ent required is com ­
m on with th at for the salt industry and th at there
is an existing pool o f m anpow er with actual ex­
perience in all aspects o f pond construction and
m aintenance, the establishm ent o f an aquaculture
facility in the first reservoir m akes good econom ic
sense. T he necessity and cost o f pum ping sea
w ater are inherent in the operation of a salt field
hence no extra costs w ould be incurred in obtain­
ing the plankton biom ass.
S e a w a te r in ta k e
o voviviparous
50 g- s a Its / 1 Artemia
■(Continuous re c ru itm e n t)
dyin g A r t e m i a
sal t precipitation 250 g /l
Fig. 2. Schematic diagram of solar salt operation with natural occurrence of Artemia (Source: Sorgeloos, 1983).
2 1 1
Integrated solar salt - A rtem ia culture
As the sea w ater gets progressively m ore saline
w ith passage through a salt field there is a n atu ­
ral m ortality o f organism s because o f the increas­
ing salinity. T hose organism s that survive tend to
proliferate through lack o f com petition. O ne of
these organism s is the brine shrim p Artemia which
thrives in the evaporator ponds where the brine
attains a concentration o f 60-3 0 0 % (Fig. 2). In
the low salinity ponds (1 0 0 -1 5 0 g salt I - 1 ) the
Artemia reproduces by ovoviviparity (free-swim­
ming larvae) while at higher salinities
( > 1 5 0 g I - 1 ) the anim al responds to the sub-
optim al conditions by producing cysts (ovipari-
ty). Jones et al. (1981) have shown th at Artemia
w in d d ir e c tio n
100 g/l.
ovoviv i p arou s
Artemia
ovipar ous Artemia
(no m o r e r e c r u i t m e n t )
g r a v i t y f l o w of w a te r
+ Ar t emi a
cyst accum ulation
212
plays a fundam ental role in the salt m aking p ro ­
cess. The continuously non-selective filter feeding
Artemia keeps under control the algal bloom s
which usually prevent early precipitation o f gyp­
sum and in extreme situations even ham per salt
crystallisation. T h q Artemia m etabolites or the de­
caying Artem ia also provide a suitable substrate
for the halophilic m icrobes o f the genus Halobac-
terium, the presence o f which assures a red col­
oration o f the cyrstallisers. D ark red crystallisers
assures m uch better heat absorption and thus
faster salt crystallisation than in a situation where
the white salt at the bottom o f the pond reflects
m ost o f the light (Sorgeloos, 1983; Jones e ta l.,
1981). In crustacean hatcheries Artemia is used as
a practical and suitable larval food source either
in the form o f nauplii th at can easily be hatched
from the dry cysts or as adults (biom ass) th at can
be harvested from evaporator ponds. Artemia
biom ass w ould alleviate the feed supply problem
in the ponds while the Artemia cysts which can
com m and as high as $50 per kilogram in the world
m arket could be a source o f foreign exchange
earnings. Artem ia culture in the coastal saltfarm s
has already been investigated and found viable
(R asow o & R adui, 1986) and only needs m an­
agement to enhance cyst and biom ass p roduc­
tion.
Intensification o f existing pond aquaculture
M axim ising the output in the existing ponds
w ould negate the need for expanding the pond
area. T hus intensification in place o f extensifica-
tion o f the pond culture seems to be a m eans of
saving the m angroves from further destruction.
Since intensive farm ing involves great production
costs because o f the use o f sophisticated technol­
ogy, K aptesky (1987) points out th at it can only
be a viable m anagem ent solution if appropriate
local technology and m anpow er trained to utilise
it is available. K enya lacks this technology and
hence sem i-intensive m ethods which also include
feeding, pond fertilization, predator control and
aeration should be adopted.
Aquaculture in the mangroves
Som e types o f aquaculture activities can be p ra c­
ticed in m angrove areas w ithout adverse effects
on the m angrove ecosystem . Pen, cage, raft, long
line, and tray culture are good alternatives to pond
culture in the m angroves although the m ethods
require biological, technological and infrastruc­
tural inputs. O yster culture is currently being
practiced in one creek in K enya w here a simple
and inexpensive culture system using a com bina­
tion o f rack and raft m ethod for grow th and roof­
ing tiles for spat collection has given very im pres­
sive results w ith the local species Crassostrea
cucullata. K enya possesses 22 extensive m angrove
creeks covering 52 000 h a in surface area, thus the
potential for growing oysters is vast provided that
a m arket can be established. Som e species of
Tilapia, e.g., T. spilurus and T. mossambica, show
rem arkable tolerance to salinity. T he feasibility of
tilapia cage culture within the m any sheltered tidal
creeks should be studied. T he culture o f brack-
ishw ater fish like rabbit fish (Siganus spp.), m ul­
lets, and milkfish in floating cages or fish pens
should be encouraged. B w athondi (1986) reports
successful attem pts at cage culture o f Siganus
canaliculatus in neighbouring T anzania, though
the identity o f the species in question is probably
S. sutor (F ischer & W hitehead, 1974).
Pump-fed ponds
A lm ost all tropical countries with m angrove for­
ests have enacted legislation regulating the utili­
sation o f the m angrove sw am ps. B ut as pointed
out by C savas (1988), conservation m easures
however strictly enforced cannot com pletely stop
further destruction of m angroves. H ence, alter­
native solutions have to be found. Diverting
coastal aquaculture developm ent out o f the m an­
grove and onto adjacent higher grounds could be
an answer. L ocating p o nds on higher grounds
requires pum ping to provide w ater circulation. It
has been show n th at the pum ping costs are m ore
than counterbalanced by the savings on construc­
tion costs as there is a m arked reduction in the

volume o f earthw ork and structures. H igher p ro ­
duction is also achieved with the im proved cir­
culation independent o f tidal rise and fall (G ed-
ney e ta l., 1984). Since the ponds are easier to
drain a m ore efficient and com plete harvesting
facilitating the drying and treatm ent o f the pond
b ottom is possible. C onstruction out o f the m an­
grove areas m eans a substantially decreased d a n ­
ger of hitting acid sulphate soils, especially as the
ponds are also only shallowly excavated.
Conclusion
M any tropical countries with extensive areas of
inter-tidal land suitable for shrim p production
have began to develop the resource for shrim p
production with a view to foreign exchange earn­
ings, agricultural diversification and em ploym ent
generation. As a result, large tracts of m angrove
forests are being converted into aquaculture
ponds. This large-scale conversion m ay eventu­
ally be detrim ental to the aquaculture industry
itself because the viability o f shrimp m ariculture
is closely linked to the status o f m angroves in the
local ecosystem. Indeed, in the m ajority o f these
countries the hatchery systems are not well de­
veloped hence m uch o f the shrim p culture indus­
try is dependent on the m angrove areas to supply
wild fry for stocking the ponds. A num ber of
conservation m easures m ay be em ployed to m in­
imize the loss o f the m angrove habitats: ponds
should be sited on the landw ard side o f the m an­
grove area so as to have the m ore productive
seaw ard area undisturbed. W here cleared inter­
tidal land is available this should be used in pref­
erence to intact m angroves. M angroves should be
retained and planted on the banks o f ponds. There
is a clear indication th at aquaculture activity can
exist in harm ony with the m angrove ecosystem
when ‘m angrove-friendly’ aquaculture m ethods
like pen, cage, raft and long-line are employed.
Ultim ately em phasis should be laid on a m ore
efficient use o f existing pond areas by encourag­
ing intensification as well as integrated aquacul­
ture practices.
213
Acknowledgement
This paper w as prepared with the support o f a
grant from M oi University.
References
Alíela, S. O., 1986. Development and research of aquaculture
in Kenya. In E. A. Huisman (ed.), Aquaculture Research in
the Africa Region. Pudoc, Wageningen, 274 pp.
Bailey, C., 1988. The social consequences o f tropical shrimp
mariculture development. Ocean & Shoreline Management
11: 31-44.
Brusher, H. A., 1974. The magnitude, distribution and avail­
ability of prawn (Penaeidae) resources in coastal and estu­
aries of Kenya. J. mar. Biol. Ass. India 16: 335-348.
Bwathondi, P. O., 1986. A review of aquaculture practices in
Tanzania. In E. A. Huisman (ed.), Aquaculture Research
in the Africa Region. Pudoc, Wageningen, 274 pp.
Csavas, I., 1988. Shrimp farming developments in Asia. In-
fofish International 2: 11-16.
Doute, R., N. O chanda & H. Epp, 1981. A forest inventory
of Kenya using remote sensing techniques. K REM U Tech­
nical Report Series No. 30. Kenya Ministry of Environ­
ment and N atural Resources.
Fischer, W. & P. J. P. Whitehead (eds), 1974. FAO Species
Identification Sheets for Fishery Purposes Vol. 4. Rome,
FAO, Pag. Var.
Gedney, R. H., Y. C. Shang & H. L. Cook, 1984. Compara­
tive study of tidal and pumped water supply for brackish-
water aquaculture ponds in Malaysia. Coastal Aquaculture
Demonstration and Training Project, FI:DP/M AL/77/008.
Field Document No. 2: 119-60.
Jones, A. G., C. M. Ewing & M. V. Melvin, 1981. Biotech­
nology of solar salt fields. Hydrobiologia 81: 391-406.
Kapetsky, J. M., 1987. Conversion of mangroves for pond
aquaculture: some short-term and long-term remedies.
W orkshop on the Conversion of Mangrove Areas to
Aquaculture. U N D P /U N E S C O Regional Project RAS/79/
002 .
MacNae, W., 1974. Mangrove forests and fisheries. Indian
Ocean Programme, Indian Ocean Fishery Commission,
Roma, IOFC/DEV/74/34.
Rasowo, J. & J. Radui, 1986. Inoculation of brine shrimp,
Anemia salina, in Kenya; expected impact on aquaculture
development. In E. A. Huisman (ed.), Aquaculture Re­
search in the Africa Region. Pudoc, Wageningen 274 pp.
Saenger, P., E. J. Hegerl & J. D. S. Davie, 1983. Global sta­
tus of mangrove ecosystems. Commission on Ecology Pa­
pers No. 3, International Union for Conservation of Nature
and Natural Resources, 88 pp.
Sammy, N. & J. Tyler, 1986. Opportunity knocks for solar
salt field aquaculture. Australian Fisheries, November 1986,
2-4.
214
Singh, T., 1987. The use of mangrove-managed areas for
aquaculture. Workshop on the conversion of Mangrove
areas to Aquaculture. U N D P/U N E SC O Regional Project
RAS/79/002.
Sorgeloos, P., 1983. Potential of the mass production of brine
shrimp Artemia. Journal of the Society for Underwater
Technology. Spring 1983, 27-29.
Turner, R. E., 1977. Intertidal vegetation and commercial
yields of penaeid shrimp. Trans, am. Fish. Soc., 106:411-6.
Yap, G. W. & R. J. Landoy, 1986. Development of Coastal
Aquaculture. Report on a survey of the coastal areas of
Kenya for shrimp farm development. KEN/80/018, Fish­
eries Department, FAO, Rome.
 A
Hydrobiologia 247: 215-222, 1992.
V. Jaccarini & E. Martens (edsj, The Ecology o f Mangrove and Related Ecosystems. 215
© 1992 Kluwer Academic Publishers. Printed in Belgium.

The effect of oocytic atresia on fecundity estim ates o f the rabbit fish
S i g a n u s s u t o r (Pisces: Siganidae) of Kenyan marine inshore waters

M. J. N tiba & V. Jaccarini

Department o f Zoology, University o f Nairobi, P.O. B ox 30197, Nairobi, Kenya

K ey words: oocytic atresia, fecundity, Siganus sutor, K enya coast

Abstract

In the strongly group-synchronized oocyte developm ent o f Siganus sutor (Valenciennes, 1835) the group
o f oocytes to be released in the following spawning, is identified. The sm allest size o f oocyte belonging
to this group w as identified by the presence o f cytoplasm ic vacuoles in oocytes in histological sections.
These vacuolated oocytes corresponded to oocytes o f 150 p m diam eter obtained by treatm ent with
G ilson’s fixative. T he m ean num ber o f such oocytes in stage 4 (late developing) ovaries w as found to
be 638000. T he p roportion o f these oocytes rem oved by atresia before spawning w as determ ined on
histological sections to be 5 %. T he corrected estim ate o f m ean fecundity w as thus 606 000 oocytes per
spawning.

Introduction which m ay be spaw ned together with the m ore

A m ajor problem in fecundity studies o f fish is
how to identify those oocytes which are poten­
tially capable o f release in the com ing spawning
season (M acer, 1974). Some w orkers count the
actual num ber o f oocytes which are released in a
single spaw ning season. The difficulty with this
m ethod is th at it is often no t easy to determ ine the
num ber o f spaw ning seasons for a female in a
year. So m ost other w orkers use the presence o f
yolk in the oocytes as the criterion for determ in­
ing the oocytes to be spaw ned. H ow ever, M acer
(op. cit.), working with the carangid Trachurus
trachurus, argued th at the presence o f vacuoles in
the cytoplasm is an earlier indication o f active
m aturation o f the oocytes than the presence of
yolk. H e pointed out th at using the acquisition o f
yolk as the sole criterion o f oocyte developm ent
could seriously underestim ate fecundity by om it­
ting a later batch o f oocytes which are actively
developing but have no t yet laid any yolk and
advanced yolked oocytes.
A nother com plicating factor in the estim ation
o f fecundity is the phenom enon of atresia (Y am a­
m oto, 1956; M acer, 1974; G uraya eta l., 1975;
M onaco e ta l., 1978; W allace & Selm an, 1981;
Cyrus & Blaber, 1984). This has m ostly been dis­
regarded in the estim ation o f fecundity, in spite of
the fact th at it m ay affect a considerable pro p o r­
tion o f the oocytes at all stages o f developm ent
(M acer, 1974; W allace & Selman, 1981; Cyrus &
Blaber, 1984). M ost studies o f atresia in the te-
leost ovary have failed to investigate its causes,
and no w ork has so far quantified its role in reg­
ulating the num bers o f oocytes as they m ature
in tropical fishes. Y et it is clear th at atresia can
have an im portant effect on recruitm ent into the
fishery.
The object o f this study is to determ ine fecun­
dity in S. sutor, which is by far the m ost im por­
tan t siganid o f the K enyan inshore w ater (N tiba,
1986), taking into account the effect o f atresia.
216
The siganids contribute 50% o f the total artisanal
catch o f the reef fisheries o f K enya (N zioka,
1984). The m angrove-lined creeks provide n urs­
ery grounds for these economically im portant fish
(G rove eta l., 1986).
Materials and methods
Fish were caught using traditional bottom traps
in shallow w ater ( < 14 m depth) in the vicinity of
M om basa, Kenya. Sam pling was done fortnightly
around neap tides throughout 1985, with an o c­
casional additional sam ple taken on spring tides.
The total num ber o f fish caught was 904. All fish
were sexed and m aturity stages assigned accord­
ing to a m odified N ikolsky (1963) scheme, which
h as been described elsewhere (N tiba & Jaccarini,
1990). In this scheme six m aturity stages are dis­
tinguished, viz. ( 1) virgin, (2 a) developing virgin,
(2b) resting and recovering, (3) early developing,
(4) late developing, (5) ripe, som etim es running,
( 6 ) spent.
Histological techniques
F o r histology, portions o f ovaries were either fixed
for 24 h in Sm ith’s form al dichrom ate, w ashed in
running tap w ater for another 24 h, and stored in
10% form alin, or fixed and stored in B ouin’s fluid.
The stored m aterial w as dehydrated in graded
alcohols, cleared in xylene and em bedded either
in paraffin wax or in ester wax (Steedm an, 1960).
Sections were cut at 4 -1 0 fim and stained in iron
haem atoxylin and eosin.
Gilson’s treatment and counting o f oocytes
Pieces o f ovary were cut longitudinally and turned
inside out and treated with G ilson’s fluid in 50 ml
specim en bottles for three m onths. The bottles
were vigorously shaken from time to time to aid
in the release o f oocytes from the ovarian walls.
Before counting, the contents o f the bottles were
poured into a petri dish and those oocytes which
h ad n o t been liberated from the ovarian tissue
were released by teasing. T he debris o f the ova­
rian walls w as then rem oved and the rem aining
fluid containing the oocytes poured into 11 b ea­
kers. T he oocytes were w ashed in tap w ater which
w as changed several times by decanting after al­
lowing the oocytes to settle. E ach tim e the w ater
w as decanted, the supernatant w as checked under
a dissecting m icroscope for any small oocytes.
F o r counting, the clean oocytes were poured
into a 11 beaker and w ater m ade up to a know n
volume. To ensure the even distribution o f o o ­
cytes in the w ater colum n, a plastic ruler w as used
to stir the oocytes vigorously with a to and fro
m otion. A subsam ple w as taken after 10 strokes
o f the ruler by a Labsystem finelet 1 ml pipette.
O ne aliquot usually gave sufficient num bers o f the
large and small oocytes to yield satisfactory
counts and diam eter distributions.
T he oocytes were pipetted into a solid w atch
glass and their diam eter m easured along a hori­
zontal axis, regardless o f their shape, under a
stan d ard m icroscope using a calibrated eye-piece
graticule using a total m agnification o f x 40. The
accuracy o f the subsam pling m ethod w as tested
by taking seven replicates and calculating the co­
efficient o f variation, w hich w as found to be 6.2 %.
T he total num ber (N ) o f oocytes in any size class
in the ovary w as calculated as follows:-
V W
N =— nx— , where
yi wl
n = num ber o f oocytes o f a given size class
in the subsam ple;
V = volum e o f sam ple;
VI = volum e o f the sub s ample;
W = weight o f w hole ovary;
W i = weight o f portion o f ovary preserved.
Distribution o f oocytes within the ovary
To establish w hether there were any differences in
the distribution o f oocytes o f different stages
within the ovary, counts were m ade on portions
 217

o f the m aterial treated with G ilson’s taken from

the anterior, mid- and posterior regions o f the
ovary and from the right and left lobes. An anal­
ysis o f variance showed th at there were no sig­
nificant differences in oocyte diam eter distribu­
tion am ong the different regions o f the ovary
f/ j > 0.05). Therefore portions for oocyte counts
could be cut from any region o f the ovary.
Throughout this w ork, however, portions were
cut from the mid-region o f left or right lobes.

The oocytes to be spawned

In order to m ake accurate fecundity estim ates, it
is necessary to identify the oocytes th at will be
spaw ned in the season following capture. This
was done in the first place by examining the o o ­
cyte frequency distribution o f G ilson’s treated
m aterial for all the m aturity stages o f S. sutor
(Fig. 1). It is seen th at stage 4 and 5 ovaries have
an oocyte diam eter distribution th at is strongly
bim odal, with small and large oocytes. In some
ovaries interm ediate-sized oocytes are completely
missing. In contrast, stage 1, 2, 3, and 6 ovaries
have only one m ode, th at o f small oocytes.
W ith such a strongly bim odal distribution of
oocytes in stage 4 and 5 ovaries, it is reasonable
to assum e th at the m ode o f larger oocytes will be
spaw ned in the season following capture. The
question now arises: w hat is the sm allest size to
include in the m ode of large oocytes? Following
M acer (op. cit.), it was decided th at the m ost re a­
sonable answ er is to include those oocytes which
have started to form cytoplasm ic vacuoles.
The size o f such oocytes w as determ ined on
histological sections. Eight stage 4 ovaries were
exam ined at x 40, and the size frequency distri­
bution o f all oocytes in random ly selected sec­
tions w as determ ined. The sections were m oved
along the y-axis by m eans o f a m echanical stage
and all the oocytes cut through the nucleus and
falling within the graticule were m easured. Care
w as taken to avoid counting the sam e oocyte twice
by calculating how m any sections h ad to be
skipped to avoid revisiting even the largest oo­
cytes.
cn
Oocyte diam eter (um )
Fig. 1. Oocyte frequency distribution in Gilson’s treated ova­
ries o f Siganus sutor in various maturity stages, indicated by
number in each polygon. Total counts of oocytes from 6 ova­
ries were made from each maturity stage.
Results
Figure 2 shows the percentage o f oocytes con­
taining vacuoles in each oocyte size class. Since
the sm allest size o f oocyte to show vacuoles was,
of necessity, determined on sectioned m aterial,
account had to be taken of the possible effect of
differential shrinkage between m aterial fixed in
Bouin’s for sectioning and G ilson’s treated m a­
terial used for the isolation of oocytes. To m ea­
sure this effect the m ean diam eters of oocytes of
the sam e stage o f developm ent were determ ined
using six stage 4 ovaries, portions o f each o f which
were fixed in B ouin’s and G ilson’s, respectively.
It w as found th at B ouin’s fixative shrunk m aterial
9.3% m ore than G ilson’s fluid. All m easurem ents
218

3 .0 -

2 .9 -

1 . 2-

1.1-

1. 0 -

to
3 °-9 "

c/>
•2 0 .8 -
o
o
cn
c
0 .7 -
‘o.
o
§Ü 0 . 6-
-a
'S 0 .5 -

0.4-

0.3-

/90 150 210 270 330 390 450 510 560

0 . 2- ■

Oocyte diameter Cum) 0 .1-

Fig. 2. Percentage of oocytes containing cytoplasmic vacu­

oles in histological sections of eight stage 4 ovaries of Siganus 0 5 10 15 20
sutor. The cu b e o f b o d y le n g th
n~3
( S . L ) ( c m ) x 10 '

Fig. 3. Number of developing oocytes against the cube of

done on histological m aterial were thus increased
by this percentage to obtain G ilson’s equivalents,
which are used for plotting Fig. 2. This plot shows
th at vacuoles start appearing in the cytoplasm of
the oocytes on their reaching a diam eter o f 90 ¿um,
and that, at 1 5 0 /rm, about 87% o f the oocytes
have a vacuolated cytoplasm . F rom about 200 p m
the cytoplasm fills with yolk granules and the vac­
uoles are obliterated, so th at at about 290 /um no
vacuolated oocytes rem ain. Therefore, for fecun­
dity estim ates, all oocytes isolated by G ilson’s
treatm ent which were > 1 5 0 ¿um were counted.
standard body length in cm in fifteen stage 4 ovaries of Siganus
sutor.
total num ber o f such oocytes is obtained and
hence the potential fecundity o f each female d e­
term ined. Fifteen stage 4 ovaries were used and
the total oocytes to be released at the next spaw n­
ing plotted against the cube o f body length in cm
(Fig. 3). A correlation w as found betw een fecun­
dity and cube o f body length (r= 0.546, d.f. 13,
P < 0 .0 5 ). A m ean o f 638000 oocytes to be re­
leased in the next spaw ning w as determ ined.

Counting the oocytes to be spawned Effect o f egg atresia on fecundity estimates

Applying the equation N = (V /V ^ n x (W /W f) on Atretic oocytes have a chorion w hich initially b e­

the num bers o f oocytes > 1 5 0 p m in diam eter com es less distinct, then disintegrates and sinks
obtained by G ilson’s treatm ent, an estim ate o f the into the cytoplasm o f the oocyte. A t the sam e time

the yolk is resorbed and the oocytes collapse.
A ssociated with such oocytes is a m ultitude of
squam ous follicular cells and m any blood vessels
w hich invade the atretic oocytes. These observa­
tions are in close agreem ent with those o f Cyrus
& B laber (1984) working with Gerres species.
Figure 4 gives the percentage frequency of
atretic oocytes for each m aturity stage o f the
ovary. It is apparent from this figure th at the de­
gree o f oocyte atresia is low est in stage 1 ( < 20 % ),
rises to a m inor peak o f 45% in stage 3, and a
m ajor peak o f 58% in stage 6 . Stages 4 and 5
have a rate o f atresia o f 27 and 32% , respectively.
While some stages 4 and 5 ovaries treated with
G ilson’s fluid h ad the interm ediate-sized oocytes
m issing altogether, this was never the case with
histological sections o f stage 4 ovaries, which al­
ways show ed developing oocytes o f all sizes in­
cluding an appreciable proportion o f oocytes o f
interm ediate size. This apparent discrepancy can
be explained by the observation th at m any o f the
interm ediate-sized oocytes (as well as some
smaller and larger ones) were atretic and h ad a
disintegrated chorion and were already invaded
by a num ber o f squam ous follicular cells. One
would expect these oocytes to disintegrate under
G ilson’s treatm ent. A com parison of oocyte di-
3 40
cr
Maturity s ta g e
Fig. 4. The mean ( + sem) percentage of atretic oocytes in six
ovaries of Siganus sutor in each maturity stage.
219
am eter frequency distributions from portions of
the sam e ovaries fixed in B ouin’s and in G ilson’s
fluid separately supports this interpretation. Fig­
ure 5 shows the proportion o f atretic oocytes in
different size classes in 10 ovaries in histological
sections o f m aterial fixed in Bouin’s and the size
frequency distribution for the sam e 10 ovaries
fixed in G ilson’s fluid. It is clear th at the batch of
the interm ediate-sized oocytes (210-330 fim),
present in the histological sections of stage 4 ova­
ries, is alm ost com pletely m issing from the oocyte
population derived from G ilson’s treated m ate­
rial. As seen from histological sections, atresia
peaks precisely in the oocyte-size class missing
from G ilson’s treated m aterial.
It follows th at the estim ate o f 638000 as the
m ean fecundity o f S. sutor in K enyan w aters de­
rived from stage 4 ovaries m ust be corrected by
the proportion o f oocytes lost through atresia by
20-
cn
IQ-
30 90 150 210 270 330 390 450 510 570
Oocyte diam eter ( jum )
Fig. 5. Cross-hatched histogram gives the percentage fre­
quency distribution of the various oocyte classes (non-atretic)
from 10 ovaries of Siganus sutor isolated with Gilson’s fluid.
Open histogram gives the percentage frequency distribution of
the atretic oocytes in histological sections o f the same ovaries.
One lobe of each ovary was treated with Gilson’s fluid, while
the other lobe was fixed in Smith’s fluid.
220
the time they reach stage 5, i.e., before they are
spaw ned. F rom Fig. 4, it is seen th at this am ounts
to 5% over the stage 4 level. Applying th at cor­
rection factor the estim ated resultant fecundity
becom es 606000 oocytes per spawning.
Though there are tw o spawning peaks per year
in the population o f S. sutor (N tiba & Jaccarini,
1990), it is n o t clear from the evidence available
w hether individual fish spaw n once or twice a
year. O n the latter assum ption, the m ean cor­
rected fecundity estim ate w ould be about 1.2 mil­
lion.
Discussion
Fecundity
T he strongly bim odal size frequency distribution
o f oocytes in stage 4 and 5 ovaries o f S. sutor
suggests th at oocyte developm ent in this species
is group-synchronized. In this type o f develop­
m ent tw o populations o f oocytes are distin­
guished: a m ore advanced, fairly synchronous,
population o f large oocytes, and a less advanced
heterogeneous population o f small oocytes from
w hich the larger ones are recruited (W allace &
Selm an, 1981). Taking the fecundity for the im ­
m ediately following spawning season to be the
total num ber of all developing oocytes in the ovary
at stage 4 m inus w hat will be lost to atresia by
stage 5, the estim ated m ean for S. sutor in K en­
yan coastal w aters is 606000. D e S ouza (1988)
obtained a total fecundity o f 700000 eggs for
S. sutor at the K enya coast by counting only
yolked eggs from stage 4 and 5 ovaries treated
w ith G ilson’s fluid. H is counts therefore involved
only non-atretic oocytes. D e S ouza’s estim ate
agrees num erically fairly well with ours, but this
is h ard to explain since we counted both yolked
oocytes and the earlier vacuolated ones to obtain
our estim ate. M onacop (1937, cited in Lam , 1974)
working with S. canaliculatus, a species closely
related to S. sutor, reports 300000-500000 o o ­
cytes in its ovary. In the ovaries o f artificially
spaw ned S. canaliculatus individual variation
ranging from 50000 to over 2 million eggs were
found (Bryan eta l., 1975), while H asse etal.
(1977) reported a m ean per female o f 300 000 eggs
w ith a m ean diam eter o f 0.5 mm. This wide range
agrees with our findings in w hich the fecundity
ranges from 200000 in a fish o f 18.0 cm to well
over 3 million in one o f 25.5 cm SL. A similar
situation w as found for Trachurus trachurus o f the
N orth Sea (M acer, op. cit.). M acer found a cor­
relation betw een fecundity and the cube o f body-
length (P < 0.001). T hough we also see such a
correlation (P < 0.05) in S. sutor inspection o f re­
siduals show s th at the linear regression m odel
w as no t a good fit. B ut F isher’s z transform shows
no significant difference (z = 1.31) betw een M ac-
er’s and our correlation coefficients. So m ore d ata
are required to test the apparent discrepancy with
the m odel. A possible explanation o f our findings
is th at S. sutor is short-lived, e.g. the largest fish
in a sam ple o f over 900 w as less than tw o years
old (N tiba & Jaccarini, 1988). O ne perhaps might
not expect a well-defined trend in fecundity to be
established in the short tim e available.
Som e w orkers use already running fish ob­
tained from com m ercial fishermen (H asse e ta l.,
1977; D e Souza, 1988). B ut the uncertainties in­
troduced by this procedure are obvious, since
such fish m ay have already shed some eggs.
It is certain th at S. sutor populations spaw n
twice a year in E ast A frica w aters (N tiba & Ja c ­
carini, 1990 and this Sym posium ). This is indi­
cated by the tem poral variation in the weights o f
the gonads, the seasonal appearance o f juveniles,
the tem poral changes in the relative condition fac­
tor, K n, and especially by the seasonal occurrence
of spent males and females in the catches. This is
in agreem ent with results o f other w orkers study­
ing different species o f siganids in Palau, Sin­
gapore, and the Philippines (Lam , 1974; K am i &
Ikehara, 1976; H asse eta l., 1977). H ow ever
w hether the individual S. sutor spaw n once or
twice a year it is im possible to decide on the
present evidence.
Functional significance o f atresia
In group-synchronous ovaries, a variety o f oocyte
recruitm ent strategies exist (W allace & Selman,
 221

340um

Fig. 6. Sections from stages 4 (a) and 5 (b) ovaries of Siganus sutor. Iron haematoxylin and eosin. AO: atretic oocyte; RO: resting
oocyte; LPP: late protoplasmic phase; EVP: early vitellogenic phase; LVP: late vitellogenic phase; CV: cytoplasmic vacuoles; EM:
egg membrane; YG: yolk granules.

1981). R ecruitm ent can be (a) directly from oo-
gonia, at the end o f the gonadotropin-independent
stage, (b) from pre-vitellogenic to vitellogenic
stages, or (c) from oocytes which have term inated
the vitellogenic stage. In S. sutor recruitm ent
seems to be at the end o f the gonadotropin-
independent stage, since oocytes in this stage are
always present in the ovary. This type o f oocyte
recruitm ent w as show n in the herring, Clupea
harengus (Bowers & Holliday, 1961). C onsidering
S. sutor (Fig. 1), it is clear th at at stage 4 there is
a considerable num ber o f the interm ediate-sized
oocytes, while at stage 5 there are two distinct
m odes separated by a gap. These results suggest
th at during developm ent from stages 4 to 5, the
group o f largest oocytes is augm ented in num bers
by the later developm ent of the smaller ones. This
is further supported by the histological appear­
ance of stages 4 and 5 ovaries (Figs 6 a and b) in
which the form er present a wide variety o f oocyte
stages, from early resting ones, through vacu­
olated oocytes o f a wide range o f sizes, to ad­
vanced vitellogenic oocytes; while stage 5 ovaries
present an alm ost uniform distribution o f ad ­
222
vanced vitellogenic oocytes, with som e scattered
groups o f m uch smaller resting oocytes situated in
the interstices.
A tresia o f oocytes is prevalent in both pre- and
post-spaw ning stages o f S. sutor. While in the lat­
ter atresia aids in the rem oval o f unw anted m a­
terial from the ovary and helps to re-utilize valu­
able m aterial, the functional significance o f p re­
spaw ning atresia is not so clear. W allace &
Selm an (1981) suggest th at poor nutrition and
horm onal im balances, especially o f the gonadot­
ropins, m ay cause involution o f developing oo­
cytes in the pre-spaw ning phase. The ovaries m ay
function as storage organs. The occurrence o f
atresia at all oocyte m aturation stages m ay sug­
gest th at this storage function is a finely-tuned
process and th at this fish m ay be responding to
one or m ore physiological and environm ental cues
(M om m sen & W alsh, 1988).
Acknowledgements
W e th ank M r. Alíela, D irector, K enya M arine
and Fisheries R esearch Institute, M om basa, for
use o f facilities. This w ork w as supported by a
G erm an D A A D research grant and another grant
from the N ational C ouncil for Science and Tech­
nology, K enya. W e th an k M r. K igondu and M r.
Oswago, D epartm ent o f Zoology, U niversity of
N airobi, for help with the illustrations. W e also
th an k the B otany D epartm ent for help with the
m icrographs.
References
Bowers, A. B. & F. G. T. Holliday, 1961. Histological changes
in the gonad associated with the reproductive cycle of the
herring (Clupea harengus L.). Mar. Res. 5: 1-16.
Bryan, P. G., B. B. M adraisau & J. P. McVey, 1975.
Hormone-induced and natural spawning of captive Siganus
canaliculatus (Pisces: Siganidae) year round. Micronesica
11: 199-204.
Cyrus, D. P. & S. J. M. Blaber, 1984. The reproductive biol­
ogy o f Gerres in Natal estuaries. J. Fish Biol. 24: 491-504.
De Souza, T. F., 1988. Reproduction, length-weight relation­
ship and condition factor in Siganus sutor (Valenciennes,
1835) (Pisces: Siganidae) from the Kenyan waters of the
western Indian Ocean. Kenya J. Sei. Ser. B. 9: 89-101.
Grove, S. J., M. C. Little & P. J. Reay, 1986. Tudor Creek
Mombasa. The early life-history stages o f fish and prawns
1985. Overseas Development Administration, Research
Project R 3888, U K, 133 pp.
Guraya, S. S., R. K aur & P. K. Saxena, 1975. Morphology of
the ovarian changes during the reproductive cycle of the
fish, Mystus tengara (Ham.). A cta anat. 91: 222-260.
Hasse, J. J., B. B. M adraisau & J. P. McVey, 1977. Some
aspects of the life history of Siganus canaliculatus (Park)
(Pisces: Siganidae) in Palau. Micronesica 13: 297-312.
Kami, H. T. & 1.1. Ikehara, 1976. Notes on the annual juve­
nile siganid harvest in Guam. Micronesica 12: 323-325.
Lam, T. J., 1974. Siganids: their biology and maricultural po­
tential. Aquaculture 3: 325-354.
Macer, C. T., 1974. The reproductive biology of the horse
mackerel, Trachurus trachurus (L.) in the N orth Sea and
English Channel. J. Fish Biol. 6: 415-438.
Mommsen, T. P. & P. J. Walsh, 1988. Vitellogenesis and oo­
cyte assembly. In W. S. H oar & D. J. Randall (eds), Fish
Physiology 11A. Academic Press, New York: 374-406.
M onaco, P. J., E. M. Rasch & J. S. Balsano, 1978. Cytolog-
ical evidence of temporal differences during asynchronous
ovarian maturation of bisexual and unisexual fishes of the
genus Poecilia. J. Fish Biol. 13: 33-44.
M onacop, P., 1937. The artificial fertilization of dangit,
Amphacanthus oramin (Bloch & Schneider) (Probably = Si­
ganus canaliculatus). Philip. J. Sei. 62: 229-237.
Nikolsky, J. R., 1963. Ekologiya Ryb. Moscow: Vysshaya
Schkola Press, 368 pp.
Ntiba, M. J., 1986. The biology of the Kenya reef fish of the
genus Siganus. M.Sc. Thesis, University o f Nairobi,
xviii + 143 pp.
Ntiba, M. J. & V. Jaccarini, 1988. Age and growth param e­
ters of Siganus sutor o f Kenya marine inshore waters de­
rived from numbers of otolith microbands and fish lengths.
J. Fish Biol. 33: 465-470.
Ntiba, M. J. & V. Jaccarini, 1990. G onad maturation and
spawning times of Siganus sutor off the Kenya coast: evi­
dence for definite spawning seasons in a tropical fish. J.
Fish Biol. 37: 315-325.
Nzioka, R. M., 1984. The evaluation of marine fisheries re­
sources in Kenya. In S .A . Iversen & S. Myklevoll (eds),
Proceedings of the NORAD -K enya seminar to review the
marine fish stocks and fisheries in Kenya. M ombasa,
Kenya, 13-15 M arch 1984. Institute of Marine Research,
Bergen, Norway: 27-32.
Steedman, H. F., 1960. Section cutting in microscopy. Ox­
ford: Blackwell, 172 pp.
Wallace, R. A. & K. Selman, 1981. Cellular and dynamic
aspects of oocyte growth in teleosts. Am. Zool. 21: 325-
342.
Yamamoto, K., 1956. Studies on the formation of fish eggs.
1. Annual cycle in the development of the ovarian eggs in
the flounder, Liopsetta obscura. J. Fac. Sei. Hokkaido U ni­
versity Ser. 6, Zoology 12: 352-373.
Hydrobiologia 247 : 2 2 3 -2 2 7 , 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 223
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Differences in biomass production and carrageenan yields among four

strains of farmed carrageenophytes in Northern Bohol, Philippines

G avino C. T rono, Jr. & A rturo O. Lluism a

Marine Science Institute, College o f Science, University o f the Philippines, Diliman, Quezon City,
Philippines

K ey words: carrageenan, carrageenophytes, biom ass production, farming, Philippines

Abstract

C om parative studies on the biom ass and carrageenan production o f tw o strains of Eucheuma denticulatum
and tw o strains o f Kappaphycus alvarezii were m ade to assess the seasonality in their production ca ­
pacities.
The high and similar refined carrageenan (R C ) yields (43 -5 3 % o f dry w t.) o f the four strains in the
first cropping season (Ju n e -O cto b er) coincided w ith their high biom ass production with plants averaging
from 1.1 to 1.8 kg each at harvest. The poor R C yields (2 1 -3 3 % ) recorded in the second cropping
(O c to b er-F eb ru ary ) coincided with their season o f low biom ass (av. wt: 0.34 to 1.0 kg). The four strains,
however, recorded contrasting perform ance in the third cropping season (F ebruary-July) with the two
E. denticulatum strains recording high R C yields (43 and 42.5% ) together with high biom ass (av. wt: 1.5
and 1.6 kg) in co n trast to the low R C yields (30 and 39% ) and low biom ass (av. wt. 0.21 and 0.28 kg)
o f the tw o K. alvarezii strains. R ecords for semi-refined carrageenan (S R C ) yields in the second and third
cropping seasons were quite consistent and similar for the four strains (4 2 -5 5 % ), except in the second
cropping where the tw o K. alvarezii strains recorded low SRC. These differences in production poten­
tials highlight the need for cropping m anagem ent o f the four strains to im prove their cropping perform ­
ance.

Introduction als such as other seaweeds, sand, and plastic m a­

The great dem and for carrageenans on the w orld
m arket has encouraged the com m ercial p roduc­
tion o f carrageenophytes through m ariculture
which is presently one o f the few im portant live­
lihoods am ong the coastal populations in the
Philippines. The industry, however, is faced with
problem s like the apparent decrease in vigor o f
farm ed stocks resulting in low farm production
outputs, poor quality o f the produce caused by
poor p o st harvest handling practices, e.g., high
m oisture content, presence o f extraneous m ateri­
terials, and low yields o f refined carrageenans
from farm grow n crops.
The different strains o f carrageenophytes pres­
ently farm ed in D anajon Reef, N orthern Bohol
show distinct differences in their grow th rates and
production (biom ass) during the different seasons
of the year (T rono & O hno, 1989). The current
cropping practice o f farm ers is to m ix-plant the
different strains and select seedstocks for the next
cropping season from the present harvest. Be­
cause the different strains differ in their p roduc­
tion capacities during the different parts of the
224
year, the seedstocks selected for the succeeding
cropping season are not at all acclim ated to this
season resulting in low production yields o f pure
carrageenan o f crops.
The green (G S P 1 ) and the brow n (B SP1)
strains o f Eucheuma denticulatum (B urm an) C ol­
lins et H ervey and the green (G S T 1) and the
brow n (BST1) strains o f Kappaphycus alvarezii
(D oty) D oty are com m only farm ed throughout
the year. E. denticulatum produces iota carrag­
eenan and is called the ‘spinosum ’ o f com m erce
while K. alvarezii produces kappa carrageenan
and is the ‘cottonii’ o f com m erce. W e report here
the seasonal differences in biom ass production
and carrageenan yields (semi-refined and refined)
of the four strains from three different cropping
seasons with the view to illucidating the season­
ality in their production potentials as basis for
their cropping m anagem ent.
M aterials and methods
This research w as conducted at the A ntonio-A gro
Seaw eed F arm on D anajon Reef, N orthern
Bohol, Philippines (10° 1 6 ' N. L at; 124° 3 ' E.
Long.) from M arch 1983 to Septem ber 1984.
T he fixed bottom m onoline m ethod (T rono &
Fortes, 1988) w as used in the culture o f the four
strains. The experim ental plots for each strain
consisted o f 20 m onolines, each m onoline planted
to 20 plants/cuttings, the individual cuttings
weighing 100 g.
F o ur croppings were m ade, the first was from
13 June to 5 O ctober 1983; the second from 12
O ctober 1983 to February 1984; the third from 7
April to 20 July 1984. The fourth cropping from
5 August to Septem ber 1984 was destroyed by a
typhoon. The growing period lasted from 12 to 18
weeks.
The sam pling for biom ass w as done one week
after planting and weekly thereafter. Sam pling
w as done by random ly harvesting the whole
m onoline from experim ental plots o f each strain.
The average biom ass (fresh) o f the plants in the
m onoline w as recorded. The harvests (biom ass)
were sun dried. The dried m aterials starting with
the fourth week harvest were analyzed for semi­
refined (S R C ) and refined carrageenan (R C )
yields. T he extraction o f S R C w as done by M a ­
rine Colloids Philippines, Inc. in C ebu City. RC
w as processed by the alcohol precipitation
m ethod (Towle, 1973).
The d ata on standing crop and carrageenan
yields o f the four strains from the different cro p ­
ping periods were analyzed and com pared.
Results
N o significant differences in the am ount o f RC
betw een strains were observed in the first cro p ­
ping (Ju n e -O cto b er) (Table la). T he R C yields of
the four strains were within the narrow range of
4 3 -5 3 % o f dry weight and were the highest
am ong values recorded in the three cropping sea­
sons. T he biom ass o f the plants w as also the
highest recorded as all strains show ed good
growth.
The second cropping (O c t-F e b ) w as a period
o f poor grow th for all strains as the strains suf­
fered the ‘ice-ice’ disease, with the exception of
the brow n strain o f K. alvarezii (B ST1) which re­
corded relatively good grow th (attaining m ore
than 1000 g m axim um average biom ass). The
m axim um average biom ass attained by the other
strains w as barely 500 g.
Differences betw een strains in am ount o f car­
rageenan yields in the second cropping were o b ­
served. T he E. denticulatum strains (BSP1 and
G S P 1 ) yielded significantly higher am ount of
S R C than did the K. alvarezii strains (Table lb).
T he am ount o f R C appeared to be significantly
different am ong the four strains although the green
strains (G SP1 and G S T 1) seemed to yield m ore
R C than did the brow n strains (BSP1 and B ST1)
(Table la). The yields o f R C relative to SR C from
the four strains during this cropping season dif­
fered, i.e., the R C yields o f the K. alvarezii strains
were alm ost as m uch as the S R C in co n trast to
those o f E. denticulatum strains w hich were con­
siderably less th an th at o f the SRC.
In the third cropping (F e b -Ju l), the R C yields
o f the E. denticulatum strains were significantly
 225

Table 1. Comparison of the RC (Table la ) and SRC (Table lb) yields (± std. deviation), in per cent of dry weight, o f the two
strains of Eucheuma (BSP1 and G SP1) and Kappaphycus (BST1 and GST1). Results of the ANOVA-Duncan’s Multiple Range
Test are shown on the left (for comparisons between cropping seasons within a strain) and on the right (for comparisons between
strains within a cropping season) of the values; means with the same letter are not significantly different from each other (p < 0.05).
For Table lb, the t-Test was used to compare the yields of a strain from the two cropping seasons; s = significant, ns = not sig­
nificant (p< 0.05).

Table la

Strain Cropping season

Ju n -O ct O ct-F eb Feb-Jul

GST1 a 51.30 (± 4 .6 5 ) a b 33.18 (± 5 .2 7 ) a b 37.05 (± 3 .8 8 ) a

BST1 a 49.74 ( ± 2.22) a b 26.51 (± 7 .5 5 ) b b 29.98 ( ± 2.82) b
GSP1 a 49.68 (± 3 .6 6 ) a b 30.82 (± 6 .1 6 ) ab c 42.47 (± 3 .8 7 ) c
BSP1 a 49.88 (± 3 .6 5 ) a b 20.86 (± 6 .1 8 ) c c 43.00 (± 5 .4 6 ) c

Table I b

Strain Cropping season

O ct-F eb F eb-Jul

GST1 s 33.68 (± 6 .0 4 ) a 54.86 (± 4 .7 1 ) a

BST1 s 31.40 (± 7 .5 5 ) a 41.48 (± 5 .0 0 ) b
GSP1 ns 52.84 (± 3 .3 5 ) b 49.16 (± 6 .4 0 ) c
BSP1 ns 52.28 (± 7 .2 5 ) b 47.63 (± 9 .4 1 ) c

higher th an those o f the K. alvarezii strains
(Table la). All strains, however, yielded fairly the
sam e am ount o f SRC but the am ount o f R C from
the K . alvarezii strains w as significantly less than
th at o f the SRC while the am ount o f R C from the
E. denticulatum strains w as alm ost as m uch as
their SRC.
The four strains appeared to follow a com m on
pattern of intercropping differences in the yields
o f RC, i.e., highest in the first cropping, lowest in
the second, and interm ediate in the third
(Table la). These differences appeared significant
am ong the E. denticulatum strains while they did
n o t appear substantial in the K . alvarezii strains
in the third cropping.
Differences in biom ass production o f the dif­
ferent strains during the three cropping seasons
were also observed. The E. denticulatum strains
showed a tren d similar to th at o f their R C yields:
high in the first and third croppings but low in the
second cropping although a longer period (16
weeks) w as required to attain m axim um biom ass
in the third cropping com pared to the first crop­
ping (7 weeks). In contrast, high biom ass pro­
duction am ong the K. alvarezii strains w as ob­
served only in the first cropping periods. Their
R C yields appeared to follow the trend in their
biom ass production, e.g., high during the first and
low during the second and the third croppings.
Discussion
The low quality o f farm produce (low R C yields)
has always been attributed to poor post-harvest
processes. D espite efforts to induce farm ers to
im prove the quality o f their produce such as
through graded pricing based on the physical
quality, processors still receive batches o f sea­
weeds with low R C yields. The results o f this
study show th at the state o f the crop at the time
o f harvest, i.e., w hether healthy or not, appeared
226
to be an im portant factor affecting the SR C and
R C yields. The am ount o f yields appeared to vary
with the grow th perform ance o f the strains, as
indicated by biom ass production. High R C yields
o f the four strains were recorded in the J u n e -
O ctober cropping period when grow th w as good
(high standing crop) and significantly low yields
in the O cto b er-F eb ru ary cropping period when
grow th w as poor. Thus, the observed intercrop­
ping differences in the yields o f R C appear to take
exception to the general observation th at rapid
grow th o f carrageenophytes, as a result o f ele­
vated nutrient levels is accom panied by less ca r­
rageenan production (D aw es et al., 1974a, b; M c-
C andless & Craigie, 1979; C hopin e ta l., 1990).
The lack o f agreem ent between the results o f the
present studies and those o f the other w orkers
might have been caused by differences in am bient
environm ental conditions. D aw es et al. (1974a, b)
reported th at m axim um grow th and m inim um
carrageenan yields occurred in period o f low tem ­
perature. T rono & O hno (1989) reported high
growth and biom ass during periods o f relatively
high tem perature regimes.
Differences in the RC yields and biom ass p ro ­
duction observed am ong the four strains appeared
to be species-specific. T he two E. denticulatum
strains showed similarly seasonality in biom ass
and R C yields as did the two strains o f K. alva­
rezii. T rono & O hno (loc. cit.) reported the same
species-specificity in the seasonality o f biom ass
production of the different strains. But differences
in R C yields between the brow n and green strains
o f the tw o species were also noted during the
second cropping periods with the green strains of
the two species recording higher R C yields than
their brow n counterparts.
The distinctly higher ratio o f SR C to R C yields
o f crops during cropping seasons o f low biom ass
production and lower ratio during cropping sea­
sons of high biom ass production are very ap p a r­
ent. The incidence o f w hat is know n as the ‘ice-
ice’ disease is preceded by the paling o f the color,
the loss o f the glossy surface and thinning o f the
branches, and occurrence o f epiphytes and other
weeds. The SRC yields during this period rem ain
approxim ately similar to th at recorded in crops
from good cropping period but the R C yields were
significantly lower. A large portion o f the carra­
geenan in diseased or sickly crops apparently is
degraded to other products and cannot be recov­
ered by alcohol precipitation during processing of
R C. But during S R C processing, this portion is
apparently retained. The presence o f these de­
graded products m ay reduce the gel strength and
viscosity o f the carrageenan. This im portant
problem needs clarification.
The results o f this study show th at the produc­
tion potentials o f the different farm ed strains v ar­
ied with species and with cropping seasons. The
production in the first cropping period w as no t at
all affected by choice o f species or strains as all
show ed good perform ance (high biom ass and
carrageenan yields). In the second cropping pe­
riod, however, the green strains o f both species
appeared to record better crop perform ance than
did the brow n strains. In the last cropping, both
strains o f E. denticulatum recorded good grow th
com pared to those o f K . alvarezii. This inform a­
tion can be used as basis for cropping m anage­
m ent o f different strains to im prove the p ro d u c­
tivity o f the farm s.
Acknowledgements
This study w as supported by the International
D evelopm ent and R esearch C enter (ID R C ) -
C an ad a and the Philippine Council for A quatic
and M arine R esearch and D evelopm ent (PC A M -
R D ).
References
Chopin, T., M. D. Hanisak, F. E. Koehn, J. Mollion & S.
Moreau, 1990. Studies on carrageenans and effects of sea­
water phosphorus concentration on carrageenan content
and growth of Agardhiella subulata (C. Agardh) Kraft and
Wynne (Rhodophyceae, Solieriaceae). J. appl. Phycol. 2:
3-16.
Dawes, C. J., A. C. Mathieson & D. P. Cheney, 1974a. Eco­
logical studies of Floridean Eucheuma (Rhodophyta, G i­
gartinales) I. Seasonal growth and reproduction. Bull. mar.
Sei. 24: 235-273.

Dawes, C. J., J. M. Lawrence, D. P. Cheney & A. C. Mathie-
son, 1974b. Ecological studies of Floridean Eucheuma
(Rhodophyta, Gigartinales) III. Seasonal variation of car­
rageenan, total carbohydrate, protein and lipid. Bull. mar.
Sei. 24: 286-299.
M cCandless, E. L. & J. S. Craigie, 1979. Sulfated polysac­
charides in red and brown algae. Ann. Rev. PI. Physiol. 30:
41-53.
Towle, Gordon, A., 1973. Carrageenan. In R. L. Whistler &
J. N. BeMiller (eds), Industrial Gums: Polysaccharides and
227
their Derivatives. Academic Press, New York/London: 8 3 -
113.
Trono, G. C. Jr. & E. G. Fortes, 1988. Philippine Seaweeds.
National Bookstore, Metro Manila. 330 pp.
Trono, G. C. Jr. & M. Ohno, 1989. Seasonality in the biomass
production of the Eucheuma strains in Northern Bohol,
Philippines. In I. Umezaki (ed.), Scientific Survey of M a­
rine Algae and their Resources in the Philippine Islands.
Monbushio International Scientific Research Program,
Japan.
■- ü i
H

Aiyj w p i M i^iii^li
T, iw fc a y in w w w ^ 'i ,it o j
 35 ??
Hydrobiologia 247: 2 2 9 -2 3 4 , 1992.
V. Jaccarini & E. Martens (eds). The Ecology o f Mangrove and Related Ecosystems. 229
© 1992 Kluwer Academie Publishers. Printed in Belgium.

Status and exploitation of the mangrove and associated fishery resources

in Zanzibar

M . A. K. Ngoile & J. P. Shunula

Institute o f Marine Sciences, P. O. B ox 668, Zanzibar

K ey words: exploitation, m angroves, fishery resources, Z anzibar

Abstract

The distribution and growth form s o f the m angroves on Z anzibar island (U nguja and Pem ba) were
investigated. P em ba island, with approxim ately 12000 h a o f m angroves, h ad nine species; and U nguja
island, with 6000 ha, had eight. Average cropping intensities in P em ba island varied between 350-1937
cut plants per h a while in U nguja these were between 777-3567 cut plants per ha. In Pem ba, m angrove
forests were relatively less heavily exploited com pared to those in U nguja, where clear felling was
com m on. C harcoal, lime and salt production are the m ajor activities which involve heavy utilization o f
m angrove wood.
The fishery resources (shellfish and finfish) associated with the m angrove vegetation were also inves­
tigated by examining the com position o f fish catches at landing stations in m angrove as well as n o n ­
m angrove zones of fishing districts. A survey on the fauna associated with m angrove forests was also
conducted. A num ber o f the local com m ercial fish species were found to be associated with the m an­
grove vegetation. The fishes o f the genus Lethrinus and the rabbit fish Siganus were found to constitute
the bulk o f the fish caught from w aters adjacent to m angroves using m ovable traps. The m ollusc Pyrazus
sp., which is com m only used by the local people for food and bait, was found only in the m angroves.

Introduction extracted from com pleted and ongoing research

Studies on the m angroves o f Z anzibar islands
(4° 5 0 '- 6 ° 30' S and 38° 1 0 '- 3 9 ° 5 5 ' E) are few.
Griffith (1949) enum erated the m angrove species
found in several stands in both U nguja and
Pem ba islands, and attem pted a visual estim ation
o f the percentage abundance o f the various spe­
cies in addition to estim ating the areas under
m angrove cover using a planim eter. Since his
time, there were virtually no other studies on the
m angroves o f Z anzibar. M ost studies on the
m angroves of T an zania have concentrated on the
m ainland coastal stands.
The inform ation on m angroves and fisheries of
Z an zib ar islands reported in this paper have been
w ork by the authors. These studies were initiated
in view o f recent developm ents tow ards environ­
m ental conservation, m arine coastal resources
exploitation and m anagem ent, hence the need for
inventory and assessm ent.
M aterials and methods
The sites in w hich the studies were m ade included
U nguja-U kuu K isakasaka, Pete, C hw aka and
M uw anda in U nguja island (See Fig. 1), while in
Pem ba islands the sites w ere Micheweni,
C hakechake and W ete (See Fig. 2).
A m inim um of two visits were m ade to each
230

39* 15' E 39* 30’ E 35' 40' 39* 4 5 'E 50’ 55’
las Kigomasha
Ras Kiuyu
Mangrove areas
surveyed.
WSifà Mangrove areas not
surveyed during the
study. Njao Is, ¡heweni
5* OCfS-

Bumbwini 05‘-

5*00' S- PEM BA
IS LA N D
10 -

UNGUJA ISLAND
£> Chake Chake Ba)
15 -
Chwaka
I Bay
Zanzibar Town

20-

Kisakaka« Mkoanl
8*15-5-
(Unguja Pete>
Kiwani
S Bay iMuungoni

30- Mangrove areas surveyed

Mangrove areas not surve-

yed during the study______

Fig. 2. M ap of Pemba island showing the distribution of man­

39* 30’ E groves.

Fig. 1. M ap of Unguja island showing the distribution of

mangroves.
Results and discussion

study site. The inform ation collected at each site
included a list o f m angrove species, an estim ate
o f cut m angrove trees in random sam ple plots of
5 m X 5 m, percent coverage by species, diam eters
and average lengths o f random ly sam pled m an­
grove poles, and the fauna associated with the
m angroves. Inform ation regarding the local uses
of the various m angrove species w as gathered
from local residents through interviews. The spe­
cies com position o f the fish caught by artisanal
fishermen w as obtained by identification o f fish
species at landing sites and calculation o f the per­
centage com position o f the species for each fish­
ery type e.g. dema trap, towe trap, hand lines etc.
Mangroves
The distribution o f the species of m angroves to ­
gether with estim ates o f their percentage abun­
dance and cutting intensities at the study sites is
given in Table 1.
Prior to these studies, eight species o f m an ­
groves were recorded for T an zan ia ( S emesi, 1986)
and these were found both on the m ainland and
the islands. Xylocarpus moluccensis (Lam ) Roem
w as identified during these studies and has so far
been recorded only in P em ba islands (J.P .S .) and
T anga area on the m ainland (Sem esi, 1988). The
other species include Bruguiera gymnorrhiza (L)
Lam , Rhizophora mucronata Lam ., Ceriops tagal
 231

Table 1. In fo rm a tio n o n m a n g ro v e s o n U n g u ja a n d P e m b a , Z a n z ib a r.

Item U sea Unguja sites b Pemba sites b

Unguja- Kisaka­ Pete Chwaka M uwanda Miche­ Chake­ Wete

Ukuu saka weni chake

Total mangrove areas (ha)c _ 344 460 320 3240 520 1288 1324 1128
Cutting (N r/ha)d - 2317 1463 777 n.a. 3567 1937 350 335

Coverage by species (in %):

A. marina F, C 15 15 10 10 25 10 5 20
B. gymnorrhiza F, C, P 30 20 25 30 4 30 0 2
C. tagal P, F 10 15 15 15 20 10 3 3
H. littoralis F, M 0 2 0 2 3 1 3 2
L. racemosa F, C, P 5 5 5 0 5 15 2 3
R. mucronata P, C, D, L, T 20 30 35 25 7 25 15 8
S. alba P, B, F 20 8 5 15 35 6 70 60
X. granatum F, M 0 5 5 3 1 2 2 2
X. moluccensis M(?) 0 0 0 0 0 1 0 0

Size distribution of poles

Average diameters (cm) - 5.7 n.a n.a 6.7 n.a 11.5 6.7 n.a
Average length (m) - 3.50 n.a n.a 4.00 n.a 3.45 3.57 n.a
Sample size - 418 326 374 345

a F = firewood, C = charcoal, P = poles, B = boat building, D = dye, L = leather tanning, T = fish traps, M = medicine.
b Underlined % value refer to species m ost affected by cutting at the location indicated.
c From Griffith (1949).
d Average of three plots in each area.

(P err.) C.B. R obinson, Avicennia marina (F orsk)
Vierh, Lumnitzera racemosa Willd, Xylocarpus
granatum K oen, Sonneratia alba J. Smith, and
Heritiera littoralis D ryand in Ait. The m ost com ­
m on species found on the sites are R. mucronata,
B. gymnorrhiza, and S. alba.
A bout tw o thirds o f the m angrove forest stands
on Z an zib ar are located in Pem ba island while
one third is found in U nguja island. The areas
with m angroves in U nguja (See Fig. 1) are lo­
cated m ostly to the w est and include M enai bay
(U nguja U kuu, Pete and M uungoni, Kiwani bay
(K isak asak a) and M akoba bay (M uw anda and
Bumbwini). On the eastern side o f U nguja island
there is only one stand o f m angroves, found
around C hw aka bay. T here are also small stands
at M aruhubi and C hukw ani to the north and
south o f Z an zib ar tow n respectively. The U nguja
m angrove stands cover about 6000 ha. However,
m ore accurate d ata about the area and cover of
m angroves would be available after the interpre­
tation o f m ore recent aerial photographs.
O n P em ba island (See Fig. 2) the m angrove
forests are similarly found on the w estern side.
They stretch southw ard from N jao island north
o f W ete tow n through C hakechake to M koani.
M ore th an three quarters o f the m angroves lie in
this region. O ther com paratively less extensive
but im portant stands are found at A dam son bay
on the east coast, and in the bay betw een R as
K iuyu and R as K igom asha on the northern part
o f the island. The total area covered by m angrove
vegetation here has been estim ated (Griffith, 1949)
at 12000 ha. This figure will also be revised after
interpretation o f m ore recent data.
The various local use o f the m angrove species
are enum erated in Table 1. The m ost destructive
o f these are firewood and charcoal production
because the trees are cut indiscrim inately result­
ing in wide gaps in the area. Several affected areas
were seen at U nguja U kuu, K isakasaka, M u­
w anda, around C hw aka bay (U nguja) and to
some extent at M icheweni (Pem ba) where cutting
w as done for lime production. Salt production by
232
boiling seaw ater, although less practised now has
caused dam age to som e areas, e.g. the M uw anda
area along M ak o b a bay (U nguja). This area w as
know n for salt production and collection o f ho-
lothurians (bêche de mer) for export in the early
seventies, both activities involve boiling trea t­
m ent, thus requiring firewood.
The use o f m angrove tim ber is also im portant
especially in U nguja U kuu, C hw aka, K isakasaka,
and to some extent at M ichew eni and
C hakechake (as show n in Table 1). Pole p roduc­
tion w as qualitatively estim ated to be m ore in­
tense in U nguja th an in Pem ba, as judged by the
am ount of poles seen staked at various points on
the islands. This is probably due to a higher de­
m and for construction m aterials because o f the
relatively higher population density on U nguja in
co n trast to Pem ba. T he larger area o f m angroves
in P em ba and the low cutting pressure from the
smaller hum an population probably explains the
superior grow th form in term s o f tree stem thick­
ness, observable in m ost stands (as show n in T a ­
ble 1).
L eather tanning is another use o f the m angrove.
The b ark o f R . mucronata is stripped off the trees,
pounded and m ixed with lime, and soaked in
m etal or concrete troughs into w hich anim al skins
are im m ersed and left to soften for one to tw o
weeks. T he skins are then ready for use in the
m anufacture o f shoes (makubadhi), in small local
shoe factories. L eather tanning w as com m on at
U zi island in U nguja; and at M icheweni, W ete
and C hakechake in Pem ba. D ebarked trees die
on drying. The problem is serious because the
trees affected are the m ature seed producing
plants.
T he m aking o f fixed stake trap s utilizing the
roots o f R . mucronata is another destructive prac­
tice. It rem oves the supporting roots, som etim es
o f an entire plant leading to the collapse o f the
trees. In the m aking o f the trap the stilt roots of
the m angrove are cut and split into thin trips
which are then woven together in a long line.
T hese long lines are then stuck erect in the m ud
or sand in m id-shore at low tide to trap fish m ov­
ing out with the w ater on the next low tide. Al­
though some other m angrove species are also used
e.g. L . racemosa, C. tagal and the m angrove as­
sociate, Pemphis acidua, the roots o f R . mucronata
are the m ost com m only used.
F or m edicinal purpose, only X . granatum and
H. littoralis are used. A pow der obtained from the
seed coat from both species is dissolved and orally
adm inistered for the treatm ent o f abdom inal ail­
m ents, however, overdose o f this concoction can
be dangerous.
The utilization o f m angroves in Z an zib ar is
thus carried out at as shown in various intensities
(Table 1). Som e o f the stands are under m ore
serious threat than others. F rom interviews with
villagers it w as apparent th at in som e places, in­
discrim inate m angrove cutting had already re­
sulted in serious exposure o f the coastline to ero­
sion. T he M uw anda village, for example, was
reported to have been shifting landw ard away
from the approaching ocean. Ruins o f the older
settlem ents th at stand surrounded by w ater could
be seen as evidence o f loss o f land to the sea. This
phenom enon has been partly due to the cleared
m angrove vegetation.
Associated fishery resource
D uring the survey o f the fauna associated with
m angroves, the m ost com m on fish encountered
were Lethrinus harak (F orskal) p arro t fishes,
Chanos chanos (F orskal) and fishes o f the genus
Siganus. C rustaceans included Scylla serrata
(Forskal), herm it crabs, and praw ns; and m ol­
luscs included Pyrazus sp., Pinctada sp., Modiolus
sp., Crassostrea sp.
The association o f fish with m angrove vegeta­
tion has been pointed elsewhere (A hm ad, 1966),
and the potential o f several tropical fisheries have
been know n to correspond to the respective area
o f m angrove vegetation (M acN ae, 1974). It was
evident also from the interviews conducted d u r­
ing the study th at a considerable am ount o f co n ­
ventional as well as non-conventional fishing
takes place either within or in the vicinity o f the
m angrove stands. N otable am ong the non-
conventional fishery is the collection o f m olluscs.
Evidence for the m agnitude o f edible m ollusc col­
lected w as provided by the large heaps o f em pty
 233

Table 2. Species composition of fish landings at three stations on Unguja island, Zanzibar (Tarbit, 1976).

Fish % Species composition

species/group --------------------------------------------------------------------------------------------------------------------------------------
Malindi (non-mangrove) Chwaka (mangrove) M kokotoni (mangrove)

Towe Dema To we Dema Towe Dema

Lutjanidae 0.4 0.4 - 1.0 4.2 9.6

Lethrinidae - 10.5 35.5 6.7 39.0 31.7
Siganus 13.7 20.1 - 45.7 26.3 37.5
Scardae 26.2 13.0 46.1 24.1 10.1 5.8
Scolopsia 36.1 39.3 - 0.6 3.0 -

Gaterin 0.7 0.3 0.3 3.0 0.8 0.2

Acanthurus - - - 4.7 0.4 -
Epinephelus 0.2 1.2 1.2 6.0 3.4 8.0
Nemipterus - - - 0.1 0.3 0.5
Rajidae - - - 0.1 2.8 -
Caesionidae - - 15.1 0.4 - -
Parupeneus 22.6 10.2 - 7.1 7.6 4.2
Others 0.1 5.2 1.8 0.5 2.1 2.5

shells left at various points on the beach after the
rem oval o f the m eat. Villagers reported th at m ost
o f the molluscs are used to supplem ent protein
especially for those families th at cannot afford
finfish. A small proportion is used as bait in hook
and line fishing. M olluscs are also collected for
their opercula especially those from the genus
M urex. T hese opercula cost as m uch as U S $ 40
per kilogram. A t C hakechake for instance the
trade in this article w as reported to have been
quite attractive and th at the products were col­
lected and exported to M om basa, Kenya.
Three fishery types (based on the type o f fish­
ing gear) were found to be associated with the
m angroves and adjacent habitats. These included
fixed stake trap s (uzio), m ovable traps (madema
and towe) and hand line fisheries (H oekstra et a i,
1990). A com parison o f trap fishery landings at
three stations (M alindi, C hw aka and M kokotoni)
indicates th at m ore fish are caught from m an­
grove-lined areas than elsewhere (Tarbit, 1976).
Im p o rtan t fish species belonging to the families
Lethrinidae, Siganidae and Scaridae are caught
in greater num bers in m angrove-lined area (as
show n in Table 2). How ever, fish landed at M a ­
lindi (Z anzibar Town), which are caught m ostly
from non-m angrove lined areas, were found to
contain less o f the m angrove associated groups
m entioned above, instead, the coralline fishes o f
the genera Scolopsis and Pampeneus dom inated
the catches.
C intron & Schaeffer-Novelli (1983) observed
th at due to the presence o f juvenile organism s and
their m igration in m angroves as well as seagrass
beds, these regions m ake excellent fishing grounds
even for larger predatory fishes such as red snap­
pers, yellow fin tuna, sharks and other highly ca r­
nivorous species th at forage these regions. On
Z an zib ar islands, 62% and 82% o f all the fixed
stake trap fishery units operate on the w est coast
o f U nguja and Pem ba, respectively. O f the m ov­
able traps, 77% and 85% operate on the respec­
tive w est coasts o f U nguja and Pem ba. Likewise,
m ore troll line fishery units (89.5 % in U nguja and
76.6% in Pem ba), a gear which is used to catch
large pelagic fish, operate on the w est coasts where
extensive m angrove stands are found (H oekstra
eta l., 1990).
Conclusion
M angrove sw am ps support several types o f fish­
eries on Z an zib ar islands and indeed worldwide.
M angroves, also provide direct resources to the
coastal com m unities and play an im portant role
234
in protecting the coastline from erosion. H ow ­
ever, it is n o t realized th at over-exploitation o f the
m angrove forests leads to a decline in the fisheries
resources and severe environm ental dam age.
There is therefore an urgent need for the intro­
duction o f effective m anagem ent m easures that
w ould guarantee sustained utilization o f the m an­
groves.
Acknowledgement
W e wish to express our deep appreciation to the
C om m ission o f L ands and Environm ent (C oLE),
Z an zib ar, for suggesting and providing funding to
support the m angrove survey on U nguja and
P em ba island through F IN N ID A ’S financial as­
sistance. W e would also like to thank all those
who have contributed in one way or another in
the preparation o f this paper.
References
Ahmad, N., 1966. Fish and fisheries of the Sundarbans. In
Scientific problems of humid tropical zone deltas and their
implications. Proceedings of the D accar Symposium
February/M arch 1964, U N ESC O Paris: 271-276.
Cintron, G. & Y. Schaeffer-Novelli, 1983. Mangrove forests:
Ecology and response to natural and man induced stres­
sors. U N ESC O Report in Marine Science # 23 ROST-
LAC. 133 pp.
Griffith, L. A., 1949. Reconnaissance report on the forest
problems o f Zanzibar protectorate. Govt. Printer, Zanzi­
bar. 21 pp.
Hoekstra, M., M. A. K. Ngoile, N. S. Jiddawi & C. Rotteglia,
1990. Census of the marine fishing units of Zanzibar in 1989
F A O /U N D P: RAF/87/008/DR/60/99/E. 119 pp.
MacNae, W., 1974. Mangrove forest and fisheries. Indian
Ocean Program Publication Number 34, Rome, Indian
Ocean Fishery Commission. 35 pp.
Semesi, A. K., 1988. Status and utilization of mangrove along
the coast of Tanga, Tanzania. In Mainoya, J. R. (ed.), Pro­
ceedings of workshop on Ecology and Bioproductivity of
marine coastal waters of Eastern African, D ar es Salaam,
Tanzania, 18-20 January 1988. Faculty o f Science, U ni­
versity of D ar es Salaam: 174-181.
Semesi, A. K., 1986. Zonation and vegetation structure of
mangrove communities in Tanzania. In Mainoya, J. R. and
P. R. Siegel (eds), Proceedings of W orkshop on ‘Save the
mangrove ecosystems in Tanzania’. D ar es Salaam, Tan­
zania, 21-22 February 1986. Faculty of Science, University
of D ar es Salaam: 15-35.
Tarbit, J., 1976. Demersal Fisheries Research Terminal Re­
port 1974-76. EAM FRO Annual Report 1976, EAM FRO,
Zanzibar.
Hydrobiologia 247: 2 3 5 -2 4 1 , 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 235
© 1992 Kluwer Academic Publishers. Printed in Belgium.

The management of Insular Caribbean mangroves in relation to site

location and community type

Peter R. B acon & G erard P. Alleng

Zoology Department, University o f the West Indies, Mona, Kingston 7, Jamaica

K ey words: assessm ent, Caribbean, ecosystem interactions, m anagem ent, m angroves, site variability

Abstract

M angrove ecosystem s occupy different locations on C aribbean island coasts, ranging from open bays
(fringe m angals) to totally enclosed salt ponds and salinas. On géom orphologie ally active coastlines, such
as south Jam aica, systems are at varying degrees o f m aturity and productivity. Furtherm ore, because
o f system variability, the interactions betw een m angroves and associated m arine systems, such as coral
reefs and seagrass beds, are developed to different degrees.
Com m unity structure and productivity of a range o f m angals on different islands o f the G reater and
Lesser Antilles are discussed. Forcing functions and levels o f interaction with the m arine environm ent
are identified.
The rational choice o f m anagem ent options m ust be based on the range o f goods and services p ro ­
vided by the different system s; and a good understanding o f their ecology is essential w hen choosing sites
for protection, w aste disposal, landfill, m arina developm ent and fisheries enhancem ent. Exam ples are
given from current studies in Jam aica, St. Lucia, the B ritish Virgin Islands and Trinidad, o f a flexible
m anagem ent response to m angrove ecosystem diversity.

Introduction risonii Leechm an, R . racemosa G . F. W. M eyer

The im portance o f w etland conservation is widely
recognised and C aribbean island governm ents are
working to produce inventories and m anagem ent
plans. M angrove dom inated w etlands, or m an­
gals, are the m ost com m on type in the Insular
C aribbean, particularly in the Lesser Antilles.
M angroves are threatened and im pacted by de­
velopm ent activity, particularly because o f their
coastal location and the im portance o f coastal
econom ies in the C aribbean (H udson, 1983;
B acon, 1987). In m ost coastal areas the m angals
contain only three species o f m angrove trees,
which in order o f im portance are Rhizophora
mangle L., Avicennia germinans (L.) S team and
Laguncularia racemosa (L.) G aertn.; with R . har­
and A . schaueriana S tapf & Leechm an at scat­
tered locations in the Southeastern C aribbean.
T he buttonw ood, Conocarpus erectus L., is com ­
m on throughout the region on the drier w etland
m argins. The invertebrate, fish and waterfowl
faunas show a high degree of com m onality and
C aribbean m angals carry out the classic roles of
coastal protection, organic m atter production,
provision o f fish nursery area and wildlife habitat.
This suggests th at a uniform m angrove m anage­
m ent strategy can be applied and th at this might
follow the globally applicable principles laid down
by H am ilton & Snedaker (1984).
How ever, as Lugo & Snedaker (1974) dem on­
strated, a range o f m angrove types occurs in the
N eotropics; the riverine, fringe, basin and scrub
236

types, w hich have different ecological attributes.
F urtherm ore, B acon (1986)pointed out th at these
m angals were associated with a variety o f coastal
ecosystem s w hose geographic characteristics d e­
term ined their ecology and productivity. This
p aper examines the influence o f site location on
the ecological and econom ic attributes o f m angals
in selected C aribbean islands and discusses how
these features can be incorporated into the m an­
agem ent decision-m aking process.
Description of sites studied
M angrove systems were surveyed on a 50 km
stretch o f coastline near K ingston, south Jam aica;
these ranged from open bay fringe m angal to en­
closed salina with scrub m angroves (see Fig. 1).
O n this geom orphologically active coast, barrier
and lagoon developm ent has influenced m angrove
distribution and successional processes; so that
there are pioneering, m ature and senile m angal
com m unities (B acon, 1986). Site aspect also in­
fluences the degree o f hurricane im pact, which is
a further factor controlling m angrove forest struc­
ture and productivity (Alleng, 1990).
M angrove areas studied in St. Lucia show ed a
similar site variability on both east and w est
coasts o f this geom orphologically different island
(see Fig. 2). T he generally small m angrove stands
occupied open bay fringes or were isolated from
the sea behind beach barriers which breach only
in tim es o f heavy flooding. Site characteristics
influence plant distribution, tidal exchanges, sa­
linity range, groundw ater and riverine inputs,
sedim ent grain size and com position and nutrient
status.

Hunts Bay
Riverine
M a n g ro v e s Kingston

Port Royal
Lagoon fringe
Gt . S a l t Pond
Ephemeral hypersaiine pond

C ab a r ita
Complex fringe/
basin wetland M anatee Bay Lousy Bay Pond
Open bay fringe Closed salina

y West Harbour
Lagoon fringe

Jam aica

Fig. 1. Selected mangrove sites in south Jamaica.

 237

E s p e r a ne e
H b r.

dH Sand bar
Mangr ove s

Bo i s
D'Or a n g e
Bay

Choc
Bay

Ma n
Kote'

Fig. 2. Selected mangrove sites in St. Lucia. Scale for a -c , & f= 1:10000; Scale for d & e = 1:25000.
238

Methodology for mangrove area assessment Results and discussion

The ecological and socioeconom ic values o f these
tw o groups o f m angals were assessed using a
m ethodology adapted from A dam us (1983). The
m ethod assigns a rating to attributes and func­
tions o f these w etlands in order to obtain a ra n k ­
ing th a t can be used for m anagem ent decision­
m aking (see Table 1). T he rating system used w as
simple, scoring High (3), M edium (2), Low (1)
and N one (0), because o f the relative paucity o f
d ata for some categories. Field observation and
m apping by the senior author and reports on the
Jam aican sites (Chow, 1989; M cC ain & Bacon,
1989; Alleng, 1990) and St. Lucian sites (Porte-
cop & B enito-Espinal, 1985) provided detailed
infirm ation on areal extent, plant distribution and
use o f the habitats by wildlife, but assignm ents
u nder the socioeconom ic categories (such as, im ­
portance to fisheries, education value) were based
largely on inform ed opinion.
The danger o f generalizing about m angrove m an­
agem ent is apparent when this assessm ent m eth­
odology perm its the ranking o f system s according
to their respective values. All the sites were p o s­
itive for some values, but the ranking identified
areas o f priority for conservation while singling
out sites with lower value th at might be available
for use as m arinas, dum p sites or landfill. All
isolated and sem i-isolated sites scored low values
for wildlife habitat and fisheries, and the potential
for m ariculture, educational and recreational use
(see the W H , EC & R E categories in Table 1).
Em baym ent and Lagoon-type sites also received
the highest rankings when this evaluation system
w as used in the B ritish Virgin Islands (Blok-
Meeuwig, 1990); in T ortola, for example, top pri­
ority sites for conservation were P araquita (la­
goon fringe), W itches Brew (lagoon fringe) and
H odges Creek (em baym ent fringe).

Table 1. Assessment of ecological & socioeconomic values of selected mangrove sites in Jam aica & St. Lucia.

Site SZ CD VD OW SD SL WH EC RE Value index

A. Jam aica
Kingston Hbr. 3 3 2 3 3 3 2 2 2 2.3
West Hbr. 3 3 2 3 3 3 2 2 2 2.3
G SP/Flashes 3 3 2 2 3 3 2 2 2 2.2
Cabarita 3 2 2 3 2 2 2 2 2 2.0
M anatee Bay 3 1 2 1 3 2 1 1 1 1.5
Long Pond 2 2 2 1 3 1 1 0 1 1.3
Lousy Bay 1 1 1 1 1 1 1 0 0 0.7

B. St. Lucia
Savannes Bay 3 1 2 1 3 2 2 2 2 1.8
Esperance Hbr. 3 2 2 1 3 2 2 1 1 1.7
Marigot Hbr. 2 1 2 1 2 2 2 1 1 1.4
M an Kote 3 2 3 1 0 1 1 1 1 1.3
Bois d’Orange 2 2 2 1 0 1 1 0 1 1.0
Choc 1 1 1 1 3 1 1 0 0 0.9

Key: SZ = size of mangal CD = community diversity

VD = vegetation dispersion OW = open water dispersion
SD = species dominance SL = sytem linkages
WH = wildlife habitat EC = economic values
RE = recreational/educational value

Score: High = 3, Medium = 2, Low = 1, None = 0.


O pen m angals perm it m ovem ent o f m arine and
estuarine organism s in and out of the system and
export o f organic m atter. Ogden & G ladfelter
(1983) noted th at interaction betw een m angals,
seagrass beds and coral reefs was a com m on fea­
ture o f C aribbean coasts, but did n o t com m ent on
the degree o f variability which exists in such link­
ages. The nature o f the association between three
o f the south coast Jam aica m angals and neigh­
bouring m arine ecosystem s is show n in Fig. 3, to
illustrate this point. The com m unities at these lo­
cations are fringe m angroves dom inated by
Rhizophora. In P o rt Royal they occupy approxi­
m ately 6 ha, w ith m ean canopy height 4.6 m (Al­
leng, 1990), while in the G reat Salt P ond/F lashes
area they occupy approxim ately 25 ha, with m ean
canopy height 5.4 m. Fringe m angroves in W est
H arb o u r occupy approxim ately 260 ha, with
m ean canopy height at 6.0 m. A part from varia­
tions in areal extent, other structural param eters
are within the sam e order o f m agnitude and ac­
cord well with the description o f fringe m angroves
given by Lugo & Snedaker (1974).
How ever, the geographic setting and hydrog­
raphy influence the fate o f the organic m atter and
nutrients generated within the three systems, with
im plications for the nature o f their linkages with
contiguous ecosystem s. Site geography indicates
th at the im pacts o f exported m aterials are prob­
ably in inverse relation to the area o f m angroves
at each site. P o rt Royal m angroves are function­
ally part o f the K ingston H arb o u r system, which
has strong fluviatile influences, w ind-driven and
tidal circulation patterns. K ingston H arb o u r is a
m ajor exporter o f nutrients and organic m atter to
the inshore environm ent and its effluent strongly
influences the neighbouring reefs and seagrass
beds (H ead & H endry, 1986; B acon et al., 1989).
Furtherm ore, this system is an open, continuously
exporting system. In contrast, the G re at Salt
P o n d /F lash es m angroves developed around an
hypersaline p o n d closed by an ephem eral sand
bar. A lagoon form ed briefly during the wet sea­
son when flood w ater breached the barrier per­
m itting entry o f sea water. N ow th at there is a
perm anent artificial channel stable, m ore m arine
conditions have been established in the pond.
239
H owever, it still acts as a source o f nitrite and
phosphorus to the inshore w aters w hen there is
positive net discharge only during the wet season
and particularly after storm s (M cC ain & Bacon,
1989). This situation is similar to the N ariva m an­
grove lagoon, Trinidad, where discharge to the
Atlantic O cean occurs only during tw o to three
m onths o f the wet season, when the effluent
stream can carry up to 100 t org.C d a y ” 1 from
the 90 km 2 w etland (B acon, 1981). Such pulses of
nutrient-rich, brackish w ater have a seasonal, or
interm ittent im pact on neighbouring m arine h ab ­
itats.
As W est H arb o u r lacks riverine inputs, has
prevailing onshore w inds, a w eak circulation p a t­
tern, and experiences a high evaporation rate be­
cause it is located in one o f the driest areas of
Jam aica, a net outflow o f w ater containing dis­
solved and particulate organic m atter has no t been
detected. This m angal probably has little direct
influence on the ecology o f the adjacent P ortland
Bight, bu t functions as a net accum ulator o f o r­
ganic m atter. Such lagoonal system s which act as
sinks for organic m atter occur in several p arts of
the Jam aican coast, but are poorly described.
M angrove production appears to be available to
consum ers only if they enter the system, hence
W est H a rb o u r’s im portance as fish habitat and
nursery environm ent and its potential for mari-
culture.
T he relative roles o f the three fringe mangals
show n in Fig. 3 have been assessed as indicated
in Table 1 using differences in com m unity struc­
ture and ecological forcing functions (Lugo &
S nedaker, 1974) and degree o f site variability
(B acon, 1986). This allows these ecologically and
economically im portant systems to be rated for
m anagem ent purposes.
W hat has no t been done is to include in the
assessm ent m ethodology criteria which express
the value o f the m angals to associated m arine
habitats. Although the value o f the m angal itself
has been assessed, the relationship between p ar­
ticular m angrove sites and the continued health
and productivity o f an island’s fisheries, seagrass
beds or coral reefs is a critical factor in coastal
resources m anagem ent. If this dim ension can be
240

& K i n g s t o n H a r b o u r

<¿r Port R o y a l
iling

S a lt Pond P o rt R o y a l
& Cays
FI a s h e s
M a n g r o v e s
5 km

M C o ra I reefs

il S e a g ra s s
beds

P o r tla n d
P r e V a i I i n9
Bight
w i nd

West
Harbour

3 km

Fig. 3. Associations of three Jam aican mangals with neighbouring coastal ecosystems.

Table 2. Role of mangals in interactions with neighbouring
coastal and marine ecosystems.
1. Physical interactions
1.1. Filtering terrestrially derived sediment
Reduction in sediment load reaching the sea
1.2. Buffering salinity changes
Reduction in volume of freshwater reaching the sea
2. Nutrient interactions
2.1. Export of dissolved & particulate organic matter
2.2. Sinks for particulate organic matter
3. Biotic interactions
3.1. Export of meroplanktonic larvae to marine food chains
3.2. provision of feeding habitat
- marine invertebrates & fin fish
(diurnal & seasonal migrations)
- seabirds, shorebirds, marine mammals
3.3. Provision of juvenile (nursery) habitat
- crustacea (shrimp, lobster) & molluscs (conch)
- fin fish (reef fish, seagrass bed fish, pelagios)
- seabirds & shorebirds nesting (also roosting)
- marine mammals (dolphins)
added, a m ore flexible system o f m angrove site
assessm ent w ould be produced which could re­
spond to conservation requirem ents in other sec­
tors o f the economy. Table 2 lists some o f the
identified interactions with associated ecosystem s
which influence the value th at should be attached
to m angrove sites.
T he n ature o f interactions between these sys­
tem s is u n d erstood (Ogden & Gladfelter, 1983),
but insufficient d ata is available on which to base
evaluations. F u rther research is needed to quan­
tify (a) the interactive physical forcing functions,
(b) the m agnitude o f nutrient and biotic exchanges
and (c) the degree o f nursery dependence between
key C aribbean ecosystem s, in order to im prove
the m angrove site assessm ent m ethodologies cur­
rently in use.
A c k n o w le d g e m e n ts
T h anks are due to the G overnm ent, B ritish Vir­
gin Islands, and the C aribbean Environm ental
H ealth Institute, St. Lucia, for opportunities to
test the assessm ent m ethodology in the field.
241
R e fere n ce s
Adamus, P. R., 1983. A method for wetland functional as­
sessment. U.S. D epartm ent of Transport, Washington. 1:
176 pp.
Alleng, G. P., 1990. Elistorical development, present status
and management guidelines for Port Royal mangal, King­
ston, Jamaica. M. Phil. Thesis, Zoology Department, U ni­
versity of the West Indies, Mona: 171 pp.
Bacon, P. R., 1981. Nutrient export from a Caribbean la-
goonal ecosystem. Int. Symp. Coastal Lagoons, U NESCO,
Bordeaux, Sept. 1981. Abstracts p. 17.
Bacon, P. R., 1986. Mangrove ecosystem responses to gradi­
ents in factors limiting primary production in the Carib­
bean. U N ESC O Rep. Mar. Sei. 41: 19-24.
Bacon, P. R., 1987. U se of wetlands for tourism in the Insular
Caribbean. Ann. Tourism Res. 14: 104-117.
Bacon, P., I. Goodbody, A. Greenaway, M. Hendry, P. Devi-
Prasad & J. Woodley, 1989. The Caribbean Coastal M an­
agement Study: Investigations on the Hellshire coastline.
Proc. 1st. N at. Conf. Sei. & Tech., Kingston, April 1987.
Scientific Research Council, Jamaica: 276-295.
Blok-Meeuwig, J., 1990. Mangrove systems of the British
Virgin Islands: Resource mapping and assignment to pro­
tection categories. Conservation Office, Government of the
British Virgin Island, Tech. Rep. 5: 45 pp.
Chow, B., 1989. Evaluation of the importance of mangroves
to the ecology of W est H arbour, Jamaica. M. Phil. Thesis,
Zoology Department, University of the West Indies, Mona,
Jamaica: 261 pp.
Hamilton, L. S. & S. C. Snedaker, 1984. Handbook for man­
grove area management. U N E P & East W est Centre:
123 pp.
Head, S. M. & M. D. Hendry, 1986. Development of the
Hellshire Hills, Jamaica: implications for management
strategy of reefs and related ecosystems. Proc. 5th Coral
Reef Congress, Tahiti, 1985. 2: 519-524.
H udson, B. J., 1983. Wetland reclamation in Jamaica. Carib.
Geog. 1: 75-88.
Lugo, A. E., S. C. Snedaker, 1974. The ecology of mangroves.
Ann. Rev. Ecol. Syst. 5: 39-64.
McCain, C. & P. Bacon, 1989. The G reat Salt Pond. Carib­
bean Coastal Management Study: The Hellshire coast, St.
Catherine Jamaica. Marine Science U nit Res. Rep. 1: 100—
128.
Ogden, J. C. & E. H. Gladfelter, 1983. Coral reefs, seagrass
beds and mangroves: their interactions in the coastal zones
of the Caribbean. U N E SC O Rep. Mar. Sei. 23: 133 pp.
Portecop, J. & E. Benito-Espinal, 1985. The mangroves of
St. Lucia: A preliminary survey. ECN A M P, St. Lucia:
56 pp.
 A 36oV
Hydrobiologia 247: 2 4 3 -2 5 1 , 1992.
V. Jaccarini & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 243
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Regeneration strategy of mangroves along the Kenya coast:

a first approach

D. Van Speybroeck
University o f Nairobi, Botany Department, P.O. B ox 30197, Nairobi, Kenya

K ey words: m angroves, zonation, regeneration, seedlings

A b s tra c t

In June 1990 a research project sponsored by D A N ID A and A W F w as carried out by botany students
o f N airobi U niversity to investigate the regeneration strategy o f m angroves at G azi bay and M ida creek.
Statistical analysis o f 449 quadrats (5 x 5 m ) sam pled along 35 line transects in 4 m angrove forests
showed th at m angrove seedlings follow the sam e distribution pattern in the intertidal zone as their parent
trees. In other w ords, m angrove seedlings mainly develop within a well-defined species specific zone.
T hese distribution zones for the various m angrove trees and their seedlings are defined in term s of el­
evation above the m ean low w ater level o f spring tides.
T hrough assigning m angrove seedlings in the intertidal zone to the categories (i) ‘fixed’ or ‘not-fixed’,
and (ii) ‘covered’ or ‘not-covered’, evidence w as found th at propagule dispersal followed both the
‘self-planting’ theory and the ‘stranding’ theory. The ‘self-planting’ theory appeared to be the m ajor
m echanism o f propagule dispersal in ‘undisturbed’ m angrove forests, whilst the ‘stranding theory’ proved
to be predom inant in colonizing ‘over-exploited’ and cleared m angrove forests.
It is concluded th at re-afforestation o f m angrove seedlings in the intertidal zone will be m ost successful
w hen the seedlings are planted in their specific distribution zones under fixed conditions.

In tro d u c tio n Lam ., Ceriops tagal (Perr.) C. B. R obinson,

M ore than 50 years ago the zonation pattern of
the E ast A frican m angroves w as described by
G rah am (1929) and by W alter & Steiner (1936),
and it w as only in recent years th at m ore studies
concentrated on the zonation pattern of m an ­
groves along the K enya coast; at G azi bay (G al-
lin, 1988; C oppejans & Gallin, 1989; G allin et al.,
1989; Beeckm an et al., 1990; Ruw a, 1990a,
1990b; Van Speybroeck, 1990), and at M ida creek
(G ang & Agatsiva, 1990; Van Speybroeck, 1990);
revealing a topography-determ ined zonation p a t­
tern, with Sonneratia alba Sm. growing closest to
the low w ater line, followed by Rhizophora
mucronata Lam ., Bruguiera gymnorrhiza (L.)
Avicennia marina (L orrsk.) Vierh., Lumnitzera
racemosa Willd. and Xylocarpus granatum Koen.
The distribution o f m angroves along the K enya
coast also had been studied by Isaac, W. & Isaac,
F. (1968), M oorjani (1977), K okw aro (1980,
1985) and R uw a & Polk (1986).
So far, however, the regeneration strategy of
m angroves along the K enya coast has not been
studied in detail. Only G raham (1929) briefly dis­
cusses the life cycle o f K enyan m angrove trees.
A n understanding o f regeneration strategies, nev­
ertheless, form s the basis for foresters, ecologists
and policy m akers to develop a m anagem ent plan,
which em phasizes the sustainable use o f m an­
grove resources along the K enya coast. This paper
244
is a first contribution to the understanding o f
m angrove regeneration strategies in Kenya.
D o m angrove propagules follow any species
specific distribution pattern for anchorage in the
intertidal slope and further developm ent? D o
m angrove seedlings mainly develop under the p a ­
rental trees or are they rather w ashed away by the
tidal action and dispersed by the w ater to anchor­
age places far away from their parental trees?
W hen discussing the dispersal o f m angrove
propagules, R abinow itz (1978a, 1978b), D aw es
(1981) and T om linson (1986) adhered to the
‘stranding’ theory, stressing th at all m angrove
propagules are w ater dispersed before establish­
m ent in the substratum takes place. Rabinow itz
(1978b: 123) even dem onstrated th at m angrove
seedlings in tran sp lant experim ents ‘generally
show better survival in other m angroves’ habitats
than below their own adults’.
M uch less attention, however, has been paid to
the ‘self-planting’ theory (D aw es, 1981: 518;
Tom linson, 1986: 159), highlighting th at m an­
grove propagules are pointed to prom ote self­
planting as they fall from the tree.
The present study concentrates on m angrove
forests under different anthropogenic pressures.
In total 449 vegetation quadrats were sam pled in
undisturbed, used and over-exploited m angrove
forest along 37 line transects perpendicular to the
coastline. F rom each quadrat the m angrove spe­
cies com position (adult trees and seedlings) was
recorded as well as their respective positions
above M LW S. In the vegetation quadrats the
growing conditions o f the m angrove seedlings
were examined and sim ultaneously assigned to
the categories (i) fixed or not-fixed, and (ii) cov­
ered or not-covered (explained below).
Statistical analysis o f these d ata allow us (i) to
describe the zonation pattern o f m angrove trees
and m angrove seedlings, and (ii) to estim ate the
relative im portance o f the ‘stranding’ and the
‘self-planting’ strategy in the dispersal o f m an ­
grove propagules in the intertidal zone. Differ­
ences in regeneration strategy o f undisturbed and
over-exploited m angrove forests will be discussed
throughout the text.
Definition of terms used in the text:
- covered/not-covered: w hether the m angrove
seedling is covered by the p arent tree or not.
- fixed/not-fixed: w hether the seedling uses a
fixed object, such as a pneum atophore or a
rock in the substratum for anchorage and
protection or not).
- undisturbed m angrove forest: canopy cover
> 8 0 % , obviously no logging activities.
- used m angrove forest: canopy cover ± 50 %,
logging activities apparent.
- over-exploited m angrove forest: canopy cover
± 1 5 % , heavily logged by m an.
- M LW S: m ean low w ater level o f spring tides
(0 cm = datum level).
D e s c rip tio n o f s ite s s tu d ie d
Field w ork w as carried out in June 1990 at G azi
bay (Kenya: 4° 25' S ,3 9 ° 5 0 ' E ,F ig. 1) and M ida
creek (Kenya: 3° 20' S, 40° 00' E, Fig. 1). A t
G azi bay 79 vegetation quadrats were sam pled
along 9 line transects in an undisturbed m angrove
forest (G azi 1; N o f the fisherm en’s hut) whilst
147 vegetation quadrats w ere sam pled along 9
line transects in an over-exploited m angrove for­
est (G azi 2; S o f the fisherm en’s hut). A t M ida
creek 223 vegetation quadrats were sam pled along
17 line transects in used m angrove forests (M ida 1
and 2; at the E and N N W o f the creek).
M a te r ia ls a n d m e th o d s
Since the elevation above the m ean low w ater
level is considered to be the predom inant envi­
ronm ental factor effecting the distribution of
m angrove trees (W alter & Steiner, 1936; Clarke
& H annon, 1970), line transects following this
altitude gradient (perpendicular to the coastline)
were used to sam ple the m angrove vegetation.
Line transects ranged betw een 100 m and 360 m
in length, covering the w idth o f the m angrove
forests at the various sites. At least ten 5 x 5 m
quadrats were pegged out at regular intervals

S o m a lia
Kenya
L am u
S a b a k i river
M ida c re e k
Indian O cean
4°S
M om basa
G azi bay
km
T a n za n ia
Fig. 1. Map of Kenya situating the study sites at Gazi bay
(226 vegetation quadrats) and Mida creek (223 vegetation
quadrats).
along each line transect. In each quadrat the m an­
grove vegetation w as described according to the
principles o f the floristic-sociological approach
(W esthoff & van der M aarel, 1978). Species cover
and abundance were estim ated by the ordinal
scale o f van der M aarel (1979). All the m angrove
seedlings ( = young trees having at least 4 leaves,
and less th an 1 m high) present in the quadrats
were counted and assigned to the categories fixed
or not-fixed and covered or not-covered.
The elevation in cm above M LW S w as m ea­
sured for each quadrat, according to the line level
m ethod (D aw es, 1981). Heights above M LW S
were com puted using the tide table (June 1990)
for Kilindini harbour (M om basa).
G raphical representation o f d ata distributions
in Figs 1 and 2 follows the box plot display
(C ham bers eta l., 1983). A box plot sum m arizes
the distribution o f a param eter by m eans o f 5
specific values from the d ata set: the m edian
( = the 50 percentile), the upper and lower quartile
(the 75 and 25 percentiles) and the upper and
245
lower extreme (the 90 and 10 percentiles). Outlier
values in the box plot fall above and below the
upper and lower extremes, and are represented by
isolated points.
F acto r analysis (Hotelling, 1933; Orlóci &
K enkel, 1985) is used to visualize the relationship
betw een the m angrove forests studied, in term s of
their respective seedling regeneration strategies
(Table 1 and Fig. 5).
R e s u lts
Zonation o f mangrove trees and mangrove seedlings
Figure 2 (G azi bay) and Fig. 3 (M ida creek) sum ­
m arize the distribution in the intertidal zone o f 7
species o f m angrove trees and their respective
seedlings in term s o f height above M LW S.
100 Inspection o f Fig. 2 shows th at m angrove trees
and m angrove seedlings clearly grow in species
specific distribution zones. M edian elevation val­
ues (in cm ) for adult trees (tr.) and seedlings (sd.)
for the respective m angrove species are:
Sonneratia alba (tr. = 160, sd. = 148), Rhizophora
mucronata (tr. = 212, sd. = 198), Bruguiera
gymnorrhiza (tr. = 212, sd. = 209), Ceriops tagal
(tr. = 245, sd. = 240), Avicennia marina (tr. = 258,
sd. = 256), Lumnitzera racemosa (tr. = 271,
sd. = 299) and Xylocarpus granatum (tr. = 282,
sd. = 270). Figure 2 suggests the existence of
m onospecific m angrove belts along the intertidal
slope, starting with Sonneratia alba at the sea­
w ard side, bu t it nevertheless allows mixed m an­
grove stands as well, since there is a m ajor over­
lap in distribution pattern betw een Rhizophora
mucronata and Bruguiera gymnorrhiza, between
Ceriops tagal and Avicennia marina and between
Lumnitzera racemosa and Xylocarpus granatum.
Figure 3 show s the box plots for the distribu­
tion o f m angrove trees and m angrove seedlings at
M ida creek. The sam pled m angrove forests at
M ida creek are described as used m angrove for­
ests. Lumnitzera racemosa and Xylocarpus
granatum, although present at M ida creek, did
not figure in our transects. Only 6 Sonneratia alba
trees were present in the line transects with a
246

Table 1. Percentage of seedlings growing under ‘fixed’ and ‘covered’ conditions as observed at G azi bay and M ida creek. G azi 1
refers to an ‘undisturbed’ mangrove forest, M ida 1 and M ida 2 to ‘used’ mangrove forests and G azi 2 to an ‘over-exploited’
mangrove forest. The percentage of seedlings growing under ‘not-fixed’ and ‘not-covered’ conditions are not listed since they add
up to 100% with the ‘fixed’ and ‘covered’ seedlings respectively. The table summarizes growing conditions o f 9479 seedlings in
the four line transects.

Used or undisturbed Over-exploited

G azi 1 Mida 1 M ida 2 G azi 2

Sonneratia alba % fixed _ _ _ 91

% covered - - - 17
Rhizophora mucronata % fixed 88 90 67 76
% covered 98 93 87 57
Bruguiera gymnorrhiza % fixed 60 99 92 -
% covered 100 32 23 -

Ceriops tagal % fixed 58 36 73 79

% covered 67 93 32 77
Avicennia marina % fixed 12 37 21 87
% covered 75 68 90 19

400.
Gazi bay H i mangrove tree EU mangrove seedling
E 350
o
03
o
300
o
O o
XX
E o
3 250- o
CO
-O
03
> 200
o 1
-
? ±
n o
CO
£ 150-1
03
0)
s: loo.!
TV s i o o

T 1--------1--------1-------- 1--------1--------1-------- 1-------- 1--------1-------- 1-------- 1-------- 1-------- r

Sonneratia Rhizophora Bruguiera Ceriops Avicennia Lumnitzera Xylocarpus
alba mucronata gymnorrhiza tagal marina racemosa granatum

Fig. 2. Box plots showing the position above MLWS of mangrove trees and mangrove seedlings at G azi bay. The figure includes
data from 79 quadrats from an undisturbed mangrove forest (G azi 1) and 147 quadrats from an over-exploited mangrove forest
(Gazi 2).

m edian elevation above M LW S o f 101 cm , but
no Sonneratia alba seedlings were found. F rom
Fig. 3 it is seen as well th at m angrove trees and
their seedlings tend to grow in species specific
distribution zones.
A lthough no direct evidence is given in the
present paper, no difference w as observed in the
zonation pattern o f m angrove trees and seedlings
in undisturbed, used and over-exploited m an­
grove forests.
 247

400-
Mida creek |ü mangrove tree □ mangrove seedling
E 350-
o

(D
> 300- o
o
E
3 o
250- o
co
X3 i
<D
> 200 -
O
_Q
CO
— 150-
D)

X
CD
100- TT T TTT
50- — i--------1-------- 1--------1-------- 1-------- r
Sonneratia Rhizophora Bruguiera Ceriops Avicennia Lumnitzera Xylocarpus
alba mucronata gymnorrhiza tagal marina racemosa granatum

Fig. 3. Box plots showing the position above MLWS of mangrove trees and mangrove seedlings at M ida creek. The figure in­
cludes data from respectively 103 quadrats (M ida 1) and 120 quadrats (M ida 2) from used mangrove forests.

100 h eig h t a b o v e d atu m (cm ) 400 forests occupy the intertidal zone between 100
and 370 cm above M LW S. Although there is a
Sonneratia alba
substantial overlap betw een the distribution p a t­
terns in the intertidal zone for the various m an­
Rhizophora mucronata grove species, a clear zonation pattern is appar­
ent and easily recognized in the field. If present,
Bruguiera gymnorrhiza
Sonneratia alba form s the outerm ost forest m ar­
Ceriops tagal gin tow ards the open w ater, followed by pure
stands o f Rhizophora mucronata or mixed stands
Avicennia marina of R. mucronata and Bruguiera gymnorrhiza, then
followed by pure or mixed stands o f Ceriops tagal
Lumnitzera racemosa
and Avicennia marina, and with at the m ost land­
Xylocarpus granatum w ard m argin o f the forest stands o f Lumnitzera
racemosa and a very small margin, if still present,
Fig. 4. Profile diagram showing the overall zonation pattern of Xylocarpus granatum. Seedling developm ent is
of mangrove trees (full lines) and mangrove seedlings (dashed
restricted to the species specific distribution zones
lines) at Gazi bay and M ida creek. Thick lines are defined by
the upper and lower quartiles, whilst thin lines are defined by along the intertidal slope.
the upper and lower extremes. Albeit no t the case in our transects, it should
be noticed th at Avicennia marina m ay replace
Figure 4 shows the profile diagram o f m angrove Sonneratia alba at the seaw ard side o f m angrove
trees and m angrove seedlings along the intertidal stands in small channels and creeks (G raham ,
slope at G azi bay and M ida creek. M angrove 1929; M acN ae & K alk, 1962).
248
G azi 2,
-a '
o .6
VW
c/5 .4
A
-.4-
■1 .8 .6 -.4 .2 0 .2 .4
Fig. 5. Factor analysis of the four mangrove forests studied,
using the figures of table 1 on seedling growing conditions as
input. The first axis explains 60% = 2.5) of the original
variance, whilst the second axis accounts for 20% of the
original variance (A2 = 1-0).
Establishment o f mangrove seedlings
Table 1 gives the percentages o f fixed and covered
seedlings for the 5 m ost dom inant m angrove spe­
cies at G azi 1 (undisturbed), G azi 2 (over-
exploited), and M ida 1 and 2 (used).
F acto r analysis is used to facilitate interpreta­
tion o f the m ultivariate d ata represented in T a ­
ble 1 (Jongm an et al., 1987). The relationships be­
tween G azi 1, G azi 2, M ida 1 and M ida 2 (in
term s o f fixed and covered seedlings) are sum ­
m arized in Fig. 5, showing the position o f the four
sites in a tw o-dim ensional space determ ined by
the tw o principal axes extracted by factor analy­
sis. It is clear th at G azi 1 (undisturbed) and
M ida 1 and M ida 2 (used) are very similar in
term s o f percentage fixed and covered seedlings,
and th at G azi 2 (over-exploited) is quite different
from the first three sites. Indeed, according to
Table 1, m ost o f the seedlings at G azi 2 (over-
exploited) are fixed and not-covered, whilst seed­
lings in the form er three sites are predom inantly
not-fixed and covered.
D is c u s s io n
Zonation o f mangrove trees and mangrove seedlings
The zonation pattern o f m angrove trees described
in this study confirms earlier observations m ade
in the E ast African region (G raham , 1929; W alter
& Steiner, 1936; M acN ae & K alk, 1962; Beeck-
m an et al., 1990). Innovative for the K enya coast
Mida 1
how ever is the observation th at m angrove seed­
M ida 2 '
lings do not develop at random in the intertidal
Gazi 1
zone, but are restricted to sites where their p a ­
rental trees are found. Even w hen the parental
trees have been felled, such as in the over-
exploited m angrove forest G azi 2, m angrove
seedling developm ent only takes place in the spe­
first axis
cies specific distribution zones.
.6 .8 1
This observation supports the idea already p ro ­
m oted by R abinow itz (1978a) who dem onstrated
th at requirem ents o f propagules for natural es­
tablishm ent correlate with position in the zo n a­
tion. B ut we cannot support her idea o f tidal sort­
ing being the only m echanism for the control of
zonation (R abinow itz, 1978b). If it were true th at
propagule establishm ent is controlled by
propagule size (and its m ovem ent in shallow
w ater), then we would expect the zonation p a t­
tern o f a specific m angrove species to be deter­
m ined by the m axim al am plitude o f the tidal
m ovem ents. K now ing th at the tidal am plitude at
the K enya coast varies betw een less than 0.5 m to
m ore than 4 m, tidal sorting only can no t explain
m angrove distribution zones o f less than 4 m.
F rom Figs 2, 3 and 4 it is clear th at m angrove
seedlings do not develop below 1 m above
M L W S, and th at the distribution o f any given
m angrove species in the intertidal zone hardly
exceeds a range o f 2 m elevation in the intertidal.
Establishment o f mangrove seedlings
W hen considering the ratio o f fixed/not-fixed and
covered/not-covered seedlings in the studied
m angrove forests (Table 1, Figs 6 and 7), argu­
m ents for both the ‘self-planting’ theory and the
‘stranding’ theory can be form ulated.
It is assum ed th at dispersal according to the
‘self-planting’ theory will lead to a m ajority of
covered seedlings growing either in a fixed or not-
fixed condition, whilst dispersal according to the
‘stranding’ theory rather will lead to a m ajority o f
fixed seedlings growing either in a covered or not-

100
CD
- - ■© " o v e r -e x p lo ite d "
T3
" u n d is tu r b e d "
Sonn, fíhiz. Brug. Ceri. Avie.
Fig. 6. Percentage seedlings growing under fixed conditions
in an over-exploited (Gazi 2; dashed line) and an undisturbed
mangrove forest (Gazi 1, M ida 1 and M ida 2; full line). M an­
grove species are listed according to their position on the
intertidal slope, with Sonneratia alba (Sonn.) closest to the low
water line, followed by Rhizophora mucronata (Rhiz.),
Bruguiera gymnorrhiza (Brug.), Ceriops tagal (Ceri.) and
Avicennia marina (Avic.).
covered situation. W hen we also take into co n ­
sideration the uprooting effect o f the wave action,
we m ay assum e th at there is a positive correlation
betw een the intensity o f the wave action and the
num ber o f seedlings growing under fixed condi­
tions. Figure 6 shows th at tow ards the seaw ard
side o f the m angrove forests (where exposure to
wave action is m axim al) nearly 90 % o f the m an­
grove seedlings are fixed. In undisturbed m an­
grove forests this percentage decreases with
higher positions in the intertidal gradient, coin­
ciding with a gradient o f decreasing wave action.
Figure 6 shows as well th at at the very landw ard
site o f the undisturbed m angrove forest (where
exposure to wave action is m inim al) only 25% of
the seedlings grow under fixed conditions. In the
over-exploited m angrove forests on the other
han d the percentage o f fixed seedlings does not
decrease with higher positions in the intertidal
gradient and even never drops beyond 75% . This
can be explained (i) by an inadequate protection
o f the intertidal slope from wave action (more
than 75 % o f the m angrove trees have been felled)
and (ii) by the suprem acy o f the ‘stranding’ the­
ory over the ‘self-planting’ theory in an over-
exploited m angrove forest.
Inspection o f the percentage covered seedlings
in undisturbed and over-exploited m angrove for-
249
a> " u n d is tu r b e d "
" o v e r -e x p lo ite d "
Sonn. Rhiz. Brug. Ceri. Avic.
Fig. 7. Percentage seedlings growing under covered condi­
tions in an over-exploited (Gazi 2; dashed line) and an un­
disturbed mangrove forest (Gazi 1, M ida 1 and M ida 2; full
line). Mangrove species are listed according to their position
on the intertidal slope, with Sonneratia alba (Sonn.) closest to
the low water line, followed by Rhizophora mucronata (Rhiz.),
Bruguiera gymnorrhiza (Brug.), Ceriops tagal (Ceri.) and
Avicennia manna (Avic.). The Ceriops zone in the over-,
exploited mangrove forest behaves as in an undisturbed man­
grove forest (it still has an overall tree cover of 70 % ), resulting
in an untypical high percentage covered seedlings for an over-
exploited mangrove forest.
ests (Table 1, Fig. 7), further supports the idea of
coexistence o f the ‘stranding’ and the ‘self­
planting’ theory. Seedlings in undisturbed m an­
grove forests how ever are principally covered, in­
dicating a preponderance o f the ‘self-planting’
theory over the ‘stranding’ theory, whilst seed­
lings in over-exploited m angrove forests are prin­
cipally not-covered, supporting here the idea of a
preponderance o f the ‘stranding’ theory.
C o n c lu s io n
Analysis o f 449 vegetation quadrats along 36 line
transects in m angrove forests at G azi bay and
M ida creek show ed th at establishm ent o f m an­
grove propagules, once dispersed, takes place in
species specific distribution zones in the intertidal
zone, which we defined in term s o f height above
M LW S. The resulting zonation o f m angrove
seedlings starts with Sonneratia alba closest to
the low w ater level, respectively followed by
Rhizophora mucronata, Bruguiera gymnorrhiza,
Ceriops tagal, Avicennia marina, Lumnitzera
250
racemosa and Xylocarpus granatum. This zonation
pattern perfectly corresponds with the zonation
pattern displayed by their parent trees. This clear
m angrove zonation pattern nevertheless allows
the existence o f mixed stands o f m angrove trees
in the intertidal zone.
Analysis o f growing conditions o f seedlings in
undisturbed and over-exploited m angrove forests
supported the idea th at propagule dispersal fol­
lows b oth the ‘self-planting’ and the ‘stranding’
theory. The ‘self-planting’ theory being the m ajor
m echanism o f propagule dispersal in undisturbed
m angrove forests, where only the seaw ard m argin
is exposed to strong wave action and where the
m angrove forest as such provides sufficient p ro ­
tection to prevent uprooting o f dom inantly ‘self­
p lan ted ’ seedlings in the intertidal zone. The
‘stranding’ theory, on the other hand, becom es
m ore im portant th an the ‘self-planting’ theory in
places w ithout a direct input o f ‘self-planting’
propagules such as in over-exploited or cleared
m angrove forests.
M angrove seedlings are m ore likely to be found
under fixed conditions with an increasing p oten­
tial o f the wave action to uproot not-fixed seed­
lings. This is why fixed seedlings, as a rule, are
found at the seaw ard m argin o f the m angrove
forests, and at sites exposed to strong wave ac­
tion, such as gaps in m angrove forests or over-
exploited and totally cleared m angrove forests.
Propagules dispersed by m echanism s belonging
to the ‘stranding’ theory, as a result, rather grow
under fixed conditions than under not-fixed co n ­
ditions.
W hen considering m angrove re-afforestation
projects along the K enya coast, it is clear from the
present study th at seedlings should be planted
u n d er fixed conditions and in their species spe­
cific distribution zones.
A c k n o w le d g e m e n ts
This w ork w as undertaken as p art o f the N airobi
University (B otany D epartm ent) undergraduate
course SBT 203 - plant ecology - in June 1990.
W e are grateful to the university authorities for
their overall support, and to D r. M . J. N tiba, M r.
A. O. O luoch, M r. C. A. A ketch, M r. J. M ur aya,
M r. S. M athenge and 50 second year students for
their help with the field w ork. W e th an k the D a n ­
ish International D evelopm ent Agency (D A N I­
D A ) at N airobi, K enya, for financing this study,
as well as the A frican Wildlife F oundation of
W ashington D. C., A frican operations, for their
contribution.
R e fe re n c e s
Beeckman, H., E. Gallin & E. Coppejans, 1990. Indirect gra­
dient analysis of the mangal formation of Gazi Bay (Kenya).
Silva G andavensis 54: 57-72.
Chambers, J. M., W. S. Cleveland, B. Kleiner & P. A. Tukey,
1983. Graphical methods for data analysis. The W adsworth
statistics/probability series. W adsworth International
Group, Belmont, California. 395 pp.
Clarke, L. D. & N. J. Hannon, 1970. The mangrove swamp
and salt marsh communities of the Sydney district. 3. Plant
growth in relation to salinity and waterlogging. J. Ecol. 58:
351-369.
Coppejans, E. & E. Gallin, 1989. Macroalgae associated with
the mangrove vegetation of G azi Bay (Kenya). Bull. Soc.
r. Bot. Belg. 122: 47-60.
Dawes, C. J., 1980. Marine Botany. John Wiley & Sons, New
York. 628 pp.
Gallin, E., 1988. Vegetatie-ecologische studie van de m an­
grove langs de W estkust van G azi Bay (Kenia). State U ni­
versity Ghent. M. Sc. Thesis, 157 pp.
Gallin, E., E. Coppejans & H. Beeckman, 1989. The m an­
grove vegetation of G azi Bay. Bull. Soc. r. Bot. Belg. 122:
197-207.
Gang, P. O. & J. L. Agatsiva, 1990. The current status of
mangroves along the Kenyan coast: a case study of the
M ida Creek mangroves based on remote sensing. Paper
presented at the International Symposium on Mangroves
and related Ecosystems, 1990, M ombasa, Kenya.
Graham, R. M., 1929. Notes on the mangrove swamps of
Kenya. J. E. Afr. U. Nat. Hist. Soc. 36: 157-164.
Hotelling, H., 1933. Analysis o f a complex of statistical var­
iables into principal components. J. Ed. Psych. 24: 417—
441, 498-520.
Isaac, W. M. E. & F. M. N. Isaac, 1968. Marine botany of
the Kenya coast. 3. General account of the environment,
flora and vegetation. J. E. Afr. N at. Hist. Soc. 27: 1-28.
Jongman, R. H. G., C. J. F. ter Braak & O. F. R. van Ton­
geren, 1987. D ata analysis in community and landscape
ecology. Pudoc, Wageningen. 299 pp.
Kokwaro, J. O., 1980. Economic importance and local use of
the Kenyan mangroves. Proceedings of the Kenya National
Seminar on Agroforestry, ICRAF, Nairobi: 377-386.

Kokwaro, J. O., 1985. The distribution and economic impor­
tance of the mangrove forests in Kenya. J. E. Afr. Nat.
Hist. Soc. 75: 1-12.
MacNae, W. & M. Kalk, 1962. The ecology of the mangrove
swamps at Inhaca island, Moçambique. J. Ecol. 50: 19-34.
Moorjani, S. A., 1977. The ecology of marine algae of the
Kenya coast. University of Nairobi, Ph.D. Thesis. 285 pp.
Orlóci, L. & N. C. Kenkel, 1985. Introduction to data anal­
ysis. Statistical Ecology monographs 1. International Co­
operative Publishing House, Maryland, USA. 339 pp.
Rabinowitz, D., 1978a. Dispersal properties of mangrove
propagules. Biotropica 10: 47-57.
Rabinowitz, D., 1978b. Early growth of mangrove seedlings
in Panama, and an hypothesis concerning the relationship
of dispersal and zonation. J. Biogeogr. 5: 113-133.
Ruwa, R. K. & P. Polk, 1986. Additional information on
mangrove distribution in Kenya: some observations and
remarks. Kenya Journal of Science and Technology Series
B 7 (2): 41-45.
Ruwa, R. K., 1990a. The effects of habitat complexities cre­
ated by mangroves on macrofaunal composition in brack-
251
ishwater intertidal zones at the Kenyan coast. Discovery
and Innovation 2: 49-55.
Ruwa, R. K., 1990b. Impact on mangroves and crabs in creek
and fringe mangrove vegetation due to disturbances at the
Kenyan coast. Paper presented at the International Sym­
posium on Mangroves and related Ecosystems, 1990, M o­
mbasa, Kenya.
Tomlinson, P. B., 1986. The botany of mangroves. Cambridge
university Press Tropical Biology Series. 413 pp.
Van Speybroeck, D., 1990. Field course to the Kenya coast.
June 1990. Report on the activities. University of Nairobi,
Botany Department, internal report. 65 pp.
van der Maarel, E., 1979. Transformations of cover-
abundance values in phytosociology and its effects on com­
munity similarity. Vegetatio 39: 97-114.
Walter, H. & M. Steiner, 1936. The ecology of East African
mangroves. Zeitschrift für Botanik 30.
Westhoff, V. & E. van der Maarel, 1978. The Braun-Blanquet
approach. In: Whittaker, R. H. (ed.), Classification of plant
communities. D r W. Junk Publishers, The Hague: 287-
399.
| t i ^ J * # ){ É | ¡ '» y -» "W ;* i 3
|Ífj^fl^ij^(p^lÍ'Í^ÍB
ntíjoAlk è A
Hydrobiologia 247: 2 5 3 -2 5 9 , 1992.
F. Jaccaríni & E. Martens (eds), The Ecology o f Mangrove and Related Ecosystems. 253
© 1992 Kluwer Academic Publishers. Printed in Belgium.

Impact of past management practices on the present status of the

Muthupet mangrove ecosystem

Jayapaul A zariah, V. Selvam & S. G unasekaran

Department o f Zoology, University o f Madras, Guindy Campus, M adras - 600 025, India

K ey words: M angrove ecosystem , species com position, zonation, com m unity structure, past
m anagem ent practices

A b s tra c t

M uthupet m angrove ecosystem (10° 25' N and 79° 30' E) is located at the southern end o f the Cauvery
delta on the east coast o f India. The swam p is characterized by the presence o f only 4 species o f ex­
clusive m angroves nam ely Avicennia marina, Excoecaria agallocha, Aegiceras corniculatum and Acanthus
ilicifolius.
C om m unity structure analysis carried out at four random ly selected quadrats show ed th at the rela­
tive density, relative dom inance, relative frequency and im portance value o f A . marina w as relatively
higher than for the other three species. H ence, M uthupet m angrove ecosystem is characterized by the
mono-specific dom inance of A . marina. Earlier palynological studies indicate th at species like Rhizophora
were present in this swam p some 200 years ago. It is believed th at the p ast m anagem ent practices to ­
gether with the present poor scientific policy o f the decision m akers and harsh environm ent conditions
m ay be considered as a few factors responsible for the disappearance o f Rhizophora from this mangrove
swamp.

In tro d u c tio n system can be m aintained in as natural or close

T he present status o f the m angrove plants with
respect to their ecological distribution and com ­
m unity structure, is poorly know n in m ost of the
A sian countries. M anagem ent policies form ulated
on the basis o f this inadequate knowledge o f the
resource m ay result in unanticipated and irrevo­
cable loss o f valuable m angrove resources (Bur-
bridge, 1984). The m angroves are essentially a
forest form ation and, like other forests, m an­
groves are also a renew able natural resource. It is
thus possible to m anage the m angrove ecosystem
on a sustainable basis although, unfortunately,
this is often not done. The concept o f sustainable
use involves either sustainable harvest or sustain­
able econom ic returns while at the same tim e the
to its original status as possible (Saenger eta l.,
1983). India is one o f the few countries which
realized the im portance o f the study and m anage­
m ent o f m angroves. As early as 1903 the flora of
Sunderban including m angroves and the m an­
grove vegetation o f Bom bay Presidency were
studied (Prain, 1905). In the second stage o f m an­
grove m anagem ent, two sym posia were con­
ducted during the years 1956-1957. They were
the 9th Silvicultural C onference (1956) and the
M angrove Sym posium (1957) which was organ­
ised by the M inistry o f F ood and Agriculture,
G overnm ent o f India. L ater on significant con­
tributions were m ade by the study o f Sidhu (1963),
V enkatesan (1966), Blasco (1975) and U ntaw ale
(1984). Along with scientific contributions, the
254

G overnm ent departm ents also contributed the

w orking plan during the year 1918 and 1935 and
a revised scheme o f F orest m anagem ent was pub­
lished during the year 1935-1948 (V enkatesan,
1966). In India, at present, only the m angroves of
G anges delta and o f A ndam an and N icobar is­
lands are under the M anagem ent Plan (U ntaw ale,
1987). In the present study, the status o f the
M uthupet m angrove ecosystem, with respect to
its species com position, zonation and com m unity
lu d f la t
structure are dealt with in detail and the im pact
o f the p ast m anagem ent practices and the present
status of the m angrove plant com m unity are dis­
cussed.

D e s c rip tio n o f th e stu d y a r e a

1. Pamini river r s r * Prosopis zone
2. Korayar river S u a e d a zone
M uthupet m angrove forest is located at the
southern end o f the C auvery delta on the east 3. Kanthaperichan ^ ^ M a n g r o v e zon e
can al
coast o f India. The Cauvery river is the m ain
riverine system o f the T hanjavur district. M any 4. Marakakorayar river 5. Kilaithangai river

drainage arteries o f the Cauvery basin namely,

Pam iniyar, K orayar, Kilaithangi, M arakakorayar
flow through the M uthupet and adjacent villages
and create a lagoon before they em pty into the
Palk Strait (Fig. 1). The northern as well as the
w estern border o f the lagoon are occupied by a
dense m angrove com m unity. The southern p o r­
tion o f the lagoon is form ed by a sand spit which
is devoid of m angroves.
Fig. 1. M ap showing the location of the Muthupet mangrove
swamp and the zonation of the flora. 1. Pamini river 2. Ko­
rayar river 3. Kanthaperichan canal 4. M arakakorayar
5. Kilaithangi river, 't 4 'f Prosopis zone; ° ° ° ° Suaeda zone;
HH mangrove zone.
Zonation

The zonation pattern o f the flora w as studied by

M a te r ia ls a n d m e th o d s
Species composition
In order to prepare a com prehensive list o f plant
species available in the study area, a num ber of
surveys were m ade. Twigs and inflorescences
were collected and an herbarium w as m ade and
the species were identified referring to appropri­
ate taxonom ic keys (G am ble & Fischer, 1957).
The herbarium m aterial w as then counterchecked
with the International H erbarium , Presidency
College, M adras, India. In the final account, all
the identified species were listed and placed in the
relevant family.
visual observation. A zone w as dem arcated
clearly on the basis o f a resident species which
w as predom inant in it and the zone w as nam ed
after th at dom inant species. F lo ra in each zone
w as collected and identified.
Community structure
Com m unity structure o f the flora w as studied fol­
lowing the ran d o m q uadrat m ethod described by
C intron & Novelli (1984). A total num ber o f 5
quadrats, each covering an area o f 1 hectare, was
laid at five different locations. O ut o f 5 quadrats,

four were established in the pure thickets form ed
by exclusive m angroves and the rem aining one
w as laid in the non-m angrove zone. F rom each
q u ad rat the following d ata w ere collected: i) n-
um ber o f species o f plants ii) num ber o f individ­
uals o f each species iii) circum ference o f the trees
at b reast height. The circum ference readings were
divided by n to obtain the diam eter at breast which
were used for basal area com putation, using the
following formula.
g(m2) = 0.00007854 dbh2, where ‘g’ is the basal
area in m 2, dbh is the diam eter at breast height.
The relative frequency (R F), relative dom i­
nance (R D o), relative density (R D e) and im por­
tance value (IV) were calculated using the stan ­
dardized m ethodology (C intron & Novelli, 1984).
The im portance value o f a given species is the
sum o f R F, R D o and R D e o f th at species.
Past management practices
Inform ation on the p ast m anagem ent practices
used to exploit the m angrove resources o f M uth­
upet w as collected from the working plans p re­
p ared by C hengappa (1918) and A hm ed (1937)
and from the G overnm ent G azette released in
1937.
R e s u lts
Species composition
T he qualitative studies revealed th at the M uthu­
pet m angrove forest contains only four species of
exclusive m angroves nam ely Avicennia marina
(F orsk.) Vierh, Excoecaria agallocha L., Acanthus
ilicifolius L. and Aegiceras corniculatum (L.)
Blanco. The above species were recorded only on
the border area o f the lagoon.
Zonation
O n the basis o f distribution o f flora, M uthupet
m angrove ecosystem can be divided into 4 zones
255
namely, P rosopis zone, S uaeda zone, m angrove
zone and m ud flat zone (Fig. 1). The P rosopis
zone is present on the landw ard side o f the lagoon
and its length and breadth were approxim ately 1
to 1.2 km and 30 to 50 m respectively. This zone
w as com pletely occupied by the non-m angrove
species Prosopis spicigera L. which is a thorny
w oody plant. A few individuals o f E. agallocha
were also present on the banks o f the small tidal
creeks of the P rosopis zone. N ext to the P rosopis
zone, a well m arked Suaeda zone w as present.
This zone w as occupied by the halophyte, Suaeda
maritima is an undershrub. T he m angrove zone
w as located next to the S uaeda zone and was
restricted to the edges o f the brackish w ater la­
goon. The length o f the m angrove zone varied
from 0.2 km to 1.0 km and the breadth w as ap­
proxim ately 0.05 km to 1.5 km. All the four ex­
clusive species o f the m angroves were present in
this zone.
Community structure
The density o f the different plant species in the
random ly selected quadrats and the diam eter of
the each individual tree are given in Table 1. In
the P rosopis zone, a total num ber o f 513 individ­
uals w ere present per hectare (Q uadrat I), out of
which 466 individuals belonged to the species,
P. spicigera', 32 w ere identified as E. agallocha and
14 individuals w ere o f the species Opuntia
monocantha H aw . N early 75% o f the P. spicigera
h ad a diam eter between 10.1 and 20 cm. In all
other q uadrats the density o f Av. marina w as high,
the num ber o f individuals per hectare being 302
(Q u ad rat II) 168 (Q uadrat III) 274 (Q uadrat IV)
and 213 (Q u ad rat V) which represents 76.26% ,
66.67% , 81.55% and 89.87% , o f the total indi­
viduals recorded in the respective quadrats. In
Q uadrats II and III, S. maritima w as the codom ­
inant species and its density was 94 per hectare
in Q uadrat II and 64 per hectare in Q uadrat III.
In Q u ad rat IV, E. agallocha w as the codom inant
tree which constituted 18.45% o f the total indi­
viduals present in th at quadrat. Ae. corniculatum
w as the second dom inant species in Q uadrat V
256

Table 1. P la n t d e n s ity (n o h : ) a n d th e ir d ia m e te r c la ss.

Location Species Number of Diameter class (cm)

individuals
per Less than 3.1-10.0 10.1-20.0 20.1-30.0 G reater than
hectare 3.0 30.0

Q uadrat I Prosopis spicigera 466 62 - 354 36 14

(91.02) (13.40) (75.97) (7.73) (3.00)

Excoecaria agallocha 32 - 8 24 - -
(6.25) (25) (75)
Opuntia monocantha 14 14 - - - -
(2.73) (100)

Total 512

Quadrat II Avicennia marina 302 - - 26 272 4

(76.26) (8.61) (90.07) (1.32)

Suaeda maritima 94 9 — — — —
(23.74) (100)

Total 396

Q uadrat III Avicennia marina 168 - - 14 126 28

(66.67) (8.33) (75.00) (16.67)
Suaeda maritima 84 — — — — —
(33.33)

Total 252

Q uadrat IV Avicennia marina 274 - - 11 251 12

(81.55) (4.01) (91.61) (4.38)

Excoecaria agallocha 62 - 18 37 7 -
(18.45) (29.03) (59.68) (11.29)

Total 336

Q uadrat V Avicennia marina 213 - 6 21 172 14

(89.87) (2.81) (9.86) (80.75) (6.57)

Aegiceras corniculatum 24 20 4
(10.13) (83.33) (16.67)

237

Value in parenthesis indicates percentage.

- Indicates absence of individuals o f species in that particular diameter class.

which constituted 10.13% o f the total individu­
als.
The com m unity structure indices such as rela­
tive dom inance, relative frequency, relative den­
sity and im portance value o f each species o f the
five quadrats are given in T able 2. In Q uadrat I,
P. spicigera was the dom inant species w hose rela­
tive dom inance, relative density, relative fre­
quency and im portance values were 93.54% ,
91.01% , 33.34% and 217.89 respectively. In all
 257

Table 2. Basal area (m2), relative domainance (% ) relative density (% ) relative frequency (% ) and importance value of flora of
Muthupet mangrove swamp.

Species Basal Relative Relative Relative Importance

area dominance density frequency value
(m2) (% ) (% ) (% )

Q uadrat I
Prosopis spicigera 0.7321 93.54 91.01 33.34 217.89
Excoecaria agallocha 0.0506 6.36 6.25 33.33 46.94
Opuntia monocantha - - 2.74 33.33 -

Total 0.7827 100 100 100

Q uadrat II
Avciennia marina 1.8324 100 75.26 50 226.26
Suaeda martima - - 23.74 40 -

1.8324 100 100 100

Quadrat III
A. marina 1.2346 100 72.41 50 222.41
S. maritima - - 27.59 50

1.2346 100 100 100

Quadrat IV
A. marina 1.9080 87.35 81.55 50 218.90
E. agallocha 0.2766 12.65 18.45 50 81.01

2.1846 100 100 100

Q uadrat V
A. marina 1.5388 93.70 89.87 50 233.57
Aegiceras
corniculatum 0.1035 6.30 10.13 50 66.43

1.6423 100 100 100

- indicates value could not be calculated as the stem diameter was less than 3 cm.

other q u adrats A . marina was the dom inant spe­
cies. The relative dom inance value o f A . marina in
Q uadrats II and III w as 100% w hereas in Q ua­
d rat IV and V its value was 87.35% and 93.70%
respectively. The relative density o f A . marina in
Q u ad rat II to V w as 75.26% , 72.40% , 81.55%
and 89.87% respectively. The im portance value
o f A . marina was also high in all the last four
q u adrats the values being 226.20 (Q .II) 222.41
(Q .III) 218.90 (Q.IV) and 233.57 (Q.V). Thus the
com m unity structure analysis indicates th a t the
landw ard portion o f the swam p is dom inated by
the non-m angrove species P. spicigera w hereas
A. marina show ed monospecific dom inance in the
m angrove zone.
Past management practices
A ccording to C hengappa (1918), the M uthupet
m angrove forest w as under the control o f the
C hatram D epartm ent from 1853 to 1912. The
C hatram D epartm ent, a branch o f the then D is­
trict B oard o f Tanjore, looked after the num erous
C hatram s (free boarding and lodging houses)
258
which were constructed by the R aja o f T anjore
for the convenience o f pilgrims and travellers. To
m aintain these C hatram s, the C hatram D ep art­
m ent collected revenue from M uthupet m angrove
forest. They harvested the m angrove w ood and
sold it in the m arket as tim ber and firewood. The
revenue thus realized was spent to m aintain the
C hatram s. In 1912, the Revenue D epartm ent took
charge and during this year the First W orking
Plan for T anjore m angroves w as proposed and
advocated. A ccording to this plan the coppice
system o f clear felling with 12 years rotation was
introduced. The revenue collected from selling the
m angrove w as given to the C hatram D epartm ent.
In 1923, once again the C hatram D epartm ent
took over the responsibility o f m anaging the m an­
grove forest and followed a 20 years rotation with
clearfelling as advised by the F orest D epartm ent
(V enkatesan, 1966). The C hatram D epartm ent
exploited the m angrove resources o f M uthupet till
1936. F rom 1923 to 1936, half o f the revenue
obtained from selling m angrove products was
paid to the Revenue D epartm ent and the rem ain­
ing half was spent to m aintain the C hatram s. In
February 1937, M uthupet m angrove forest was
declared as reserve forest by the G overnm ent and
accordingly the m angrove forest w as handed over
to the F o rest D epartm ent o f the M ad ras Presi­
dency (G azette, 1937). After 1937, the F orest D e­
partm ent m anaged the M uthupet m angroves fol­
lowing the 20 years rotation system. B ased on the
frequency and dom inance o f Avicennia which was
found to be a coppicing species, 50 trees per acre
were left as seedbearers. A ccording to the study
o f V enkatesan (1966) it w as found th at the above
ro tation pattern w as followed w ith the added p re­
scription o f retaining all advanced grow th below
1.8 m (6 feet) in height.
N o detailed inform ation is available on the yield
and revenue obtained by the above m anagem ent
bodies from M uthupet m angrove Forest. H ow ­
ever, V enkatesan (1966) m entioned th at from
1935 to 1948, the average yield w as 16 tonnes/
hectare in T anjore district. In a neighbouring
m angrove forest area, C uddalore, South Division
o f Tam il N adu, a cycling pattern o f 40 years w as
followed leaving 38 m other trees/acre as seed
bearers. The average yield w as 25 tonnes/
hectares.
D is c u s s io n
The species com position o f M uthupet m angrove
forest is very poor w hen com pared to the num ber
o f exclusive m angrove species present in the Pit-
chavaram m angrove forest which is located ju st
140 km north o f the M uthupet m angrove. U n ­
controlled and unregulated exploitation by the
authorities o f the C hatram D epartm ent prior to
1912 and the unscientific m anagem ent policies
followed by the Revenue as well as C hatram D e­
partm ents after 1912 and some environm ental
factors like poor inundation and dam age due to
cyclonic storm seem to be the reason for the poor
species diversity o f the M uthupet m angrove for­
est. The available records show, th at prior to 1912,
the w oody plant species o f M uthupet m angrove
were exploited throughout the year for sales as
tim ber as well as for fuelwood by the C hatram
D epartm ent (C hengappa, 1918). After 1912, a
coppice m ethod with 12 years rotation period w as
followed for the harvest o f m angrove wood. A c­
cording to this m ethod an area o f m angrove forest
w as allotted for clear felling by the R evenue D e­
partm ent which w as then brought to the notice of
the C hatram D epartm ent. Since authorities of
both the R evenue and C hatram D epartm ents
were not aw are o f the im portance o f leaving seed
bearers in the clear felling area they rem oved all
the trees which resulted in the poor regeneration
of the m angrove plants in th at particular area
(Ahm ed, 1937). To correct the above m istake,
A hm ed (1937) introduced another system accord­
ing to which all the trees, except 50 seed bearers/
hectare were harvested. H e also changed the 12
years clear felling rotation to 20 years. A hm ed
(1937) also noticed the poor regeneration o f the
m angrove plants and prescribed artificial regen­
eration o f A . marina by direct sowing. But this
practice w as w ithdraw n after a few trials due to
unsatisfactory results. H e also suggested the dig­
ging up o f channels in areas where the ground
level was raised due to fresh accretion o f sand

and where Avicennia w as showing signs o f death
due to insufficient inundation. This suggestion
w as discontinued after a few attem pts due to high
cost involved in it. The coppice system with 20
years rotation period w as followed by the F orest
D epartm ent till 1970. The m anagem ent policy
suggested by C hengappa (1918) w as unscientific
because it was based on the wrong assum ption
th at all m angrove plant species could regenerate
from the ro o t stock. Blasco (1975) showed th at
only the species such as Avicennia and Excoecaria
have the capacity to regenerate from ro o t stocks.
The palynological studies carried out by T issot
(1987) indicate th at the species o f Rhizophora and
Sonneratia were present in the M uthupet m an­
grove ecosystem some 200 years ago but they
have disappeared com pletely now. The uncon­
trolled and unregulated exploitation carried out
by the authorities o f the C hatram and Revenue
D epartm ents, some unscientific policies intro­
duced in 1920 and thereafter as well as the poor
attention paid tow ards the replantation of the
m angrove plants seem to be some o f the causes
responsible for the present poor species diversity
o f the M uthupet m angrove forest.
R e fere n ce s
Ahmed, S., 1937. Working plan for the Tanjore Forest Divi­
sion Revenue Department, District Board of Tanjore,
82 pp.
259
Blasco, F., 1975. Mangroves of India. Inst. Fr. Pondicherry,
Trav. Sect. Sei. Tech., Vol. 14. 180 pp.
Burbridge, P. R., 1984. Management and Planning of M an­
grove Resources of India. Proc. As. Symp. Mangr. Env. -
Res. Manag., 27-42.
Chengappa, D., 1918. Working Plan for the Tanjore Forest
Division. Revenue Department, District Board of Tanjore,
78 pp.
Cintron, G. & Y. S. Novelli, 1984. Methods for studying man­
grove structure. In: Snedaker, S. C. & Snedaker J. G. (eds),
The Mangrove Ecosystem: Research Methods. M ono­
graphs on Oceanographic Methodology. U N ESCO , Paris,
91-113.
Gazette, 1937. Government of the Presidency of M adras,
M adras, 318 pp.
Gamble, G. S. & C. E. S. Fischer, 1957. Flora of the Presi­
dency of M adras Vol. I—III, Government of India, Cal­
cutta, 1388 pp.
Prain, D., 1905. Flora of the Sunderbans. Ree. Bot. Sur. India
2: 237-370.
Saenger, P., E. J. Hegerl & J. D. S. Davis, 1983. Global sta­
tus of mangrove ecosystems. Commission of Ecology paper
No. 3 IU C N , Gland, 38 pp.
Sidhu, S .S ., 1963. Studies on the mangroves of India. I-East
G odavari region. Indian Forester 89: 337-351.
Tissot, C., 1987. Recent evolution of mangrove vegetation in
the Cauvery Delta: A palynological study. J. mar. biol. Ass.
India 29: 16-22.
Untawale, A. G., 1984. Present status of mangrove along the
west coast of India. Proc. As. Symp. Mangr. Env. - Res.
Manag: 57-74.
Untawale, A. G., 1987. Conservation of Indian mangroves -
A national perspective. In: Rao, T. S. S. (ed.), Contribution
in Marine Sciences. National Institute of Oceanography
Goa: 85-104.
Venkatesan, K. R., 1966. The mangroves of M adras state.
Indian Forester 92: 27-34.
 m m & k

w w e w iii

L.a1,^uL j-i
WWNPWFWR
. ¿ 4 J XjjJ'.fiï'ïi'lî^â,
lifîillW friiA«1<iái ip i

^ swmfwmwytf-""iwm

■
Hydrobiologia 2 4 7 : 2 6 1 - 2 6 6 , 1992
V. Jaccarini & E. M artens (eds.), The Ecology o f Mangrove and Related Ecosystems. 261

Index

A budefduf saxatilis 140 Bacillariophyceae 79, 80

Acanthaster planci 121, 126, 127 Bacteria 84, 187,190, 203
Acanthophora spicifera 64 Bag net 199
Acanthurus bahianus 140 Bagamoyo district 7
A. coerulus 140 Balanus 94
Acanthus ilicifolius 11, 14, 15, 21, 182, 253 Barnacle 87, 94
Acropora 127, 133, 135, 136, 137 Baroclinie 151
A. cerviconis 133, 134, 138 Barotropic 141, 143, 151, 152, 167
A . palm ata 1 3 3 ,1 3 4 ,1 3 7 ,1 3 8 Beam trawl 201
Aegiceras corniculatum 11, 14, 15,16, 21, 38, 253 Belizean mangrove 87
Aeluropus lagopoides 16 Benthic communities 126
Aerial photography 1, 3, 173, 175, 176, 111 Biodiversity 8
Aeva lanata 34 Boodlea siamensis 126
Afzelia quanzensis 35 Boodleopsis pusilla 59, 63, 65, 69
Agriculture 1, 30 Bostrychia spp 90, 94
Aiptasia pallida 90, 95 Brachystegia speciformis 35
Algae invasions 121,122 Brachyura 120
Ambassidae 197 Bruguiera 5
Ambassis gymnocephalus 195, 198, 199, 200, 201 B. cylindrica 21
Amphimedon viridis 90 B. gymnorhiza 21, 29, 32, 49, 50, 54, 55, 56, 99,110,
Amphipoda 1 01 ,1 0 5 ,1 2 0 230,243
Amphiroa fragilissim a 64 Bryopsis hypnoides 63
A. rigida 64
Andhra Pradesh (India) 11, 181 Caesionidae 233
Anemone 87, 95 Calenoidea 101,105
Anisotes parviflora 34 Canaralia cartharica 34
Anisotremus surinamensis 140 Canonical Correspondence Analysis 62, 67, 69
Annelida 90 Caribbean m angroves 235
Appendicularia 120 Caridae 120
Aquaculture 6, 209, 210 C arrageenan 223, 224
Arabuko Sokoke forest 35 Carrageenophytes 223-226
Aristogeitonia monophylla 34 Caspian Tern 6
A rius sagor 195 Caulerpa sertularioides 63, 90
Artemia 209 C. verticillata 90, 94
culture 211, 212 Cauvery delta 254
cyst 212 Ceasalpinia nuga 14
Ascidian 87, 95 Ceranium strictum 64
Asparagus africanus 34 Ceriops tagal 3, 20, 21, 22, 29, 32, 34, 38, 49, 54, 55, 56,
Atmosperic pollutants 13 99 1 1 0 ,2 3 0 ,2 4 3
Avicennia 187 Cerithidea 5
A. germinans 88, 235 Chaetodon capistratus 140
A. marina 3 ,1 1 , 15, 17, 20, 22, 23, 25, 29, 32-34, 38, C haetognatha 120
40-45, 50, 54, 77, 99,110, 182, 231, 243, 253 Chaetomorpha crassa 63, 64
A . nitida 44 Chanos chanos 232
A . officinalis 11, 14, 15, 16,182 Chenolea diffusa 55
A. schaueriana 235 Chlorophyta 63, 90, 126
Avicenniaceae 3 Chthamalus 94
Avrainvillea obscura 59, 63 Cirripeda 91, 120
Azadrachta indicata 34 Cladocera 101,105
Azim a tetracantha 34 Cladophora patentiramea 63
 262

Clerodendrum inerme 14 Digital mapping technique 36

Cliona inconstans 126 Digitaria milanjiana 34
Clupeidae 197 Dinoflagellate 77
Cnidaria 90, 91 Dinophyceae 79, 81
Coastal currents 141, 170, 177 D iplosoma glandulosum 91
developm ent 126 Discovery Bay (Jamaica) 125
waters 141, 159, 163 Dispersion 154,158, 170
Coastline configuration 49, 50 coefficient 166, 170
Collembola 101,105 Dotilla 5
Combretaceae 3 D. fenestrata 5
Condylactis gigantea 90 Dredging 126
Conocarpus erectus 88, 235 Dutch bay (Sri Lanka) 17,19, 24, 25, 37, 3 8 ,4 6
Continuity equation 142,143
Copepoda 1 0 1 ,1 0 5 ,1 2 0 Echinometra mathaei 125,127
Coral 2, 87, 133, 134, 139, 152, 158 Ecological models 123
coverage 128 Ecomorphology 109-117
mining 128 Eddy coefficients 167
mortality 121 diffusity 1 4 8 ,1 4 9 ,1 5 1 , 156, 166
platform s 59 Eleaodendron schweinfurthianum 34
recolonisation 126, 128 Eleotridae 197
recovery 125, 126, 134, 137 Enclosure net 202
reef 121, 133 Engraulidae 197
reef management 121, 129 Enhalus acoroides 59, 63, 66
reef communities 1 2 1 ,1 2 3 ,1 2 4 ,1 3 4 Enteromorpha clathara 63
reef degradation 121, 122, 126, 134, 136 E. flexuosa 63, 65, 69
Coral Creek mangrove 141 E. kylinii 64
Corallinaceae 65 E. ramulosa 64, 65, 69
Coringa mangrove 11,16 Enteropogon macrostachysus 34
Crab burrows 145 Environmental parameters 64, 78, 82, 83, 109, 112, 181,
Crassostrea cucullata 5, 99, 109-117, 212 183, 190
Cremaspora trifloa 34 Epibiont 87-98
Cropping season 224 abundance 88, 89
Croton p seu dopulchella 35 distribution 88, 92
Crustacea 91 predation 93
Cumacea 120 species 87, 8 9 ,9 5
Curlew sandpiper 6 succession 95
Currents 153,159, 163, 165,1 6 8 ,1 6 9 , 1 7 0 ,1 7 4 ,1 7 7 transplantation 92
Cyanobacteria 90, 95 zonation 89
Cyanophyceae 79, 82 Epiphyte 59, 73
Cyclones 134 Erosion 6, 121
Cymodocea rotundata 59, 63, 65, 66 Estuarine circulation 173
C. serrulata 59, 63 stratification 163
Cynoglossidae 197 Estuary 1 4 1 ,1 5 0 ,1 5 2 ,1 6 3 ,1 7 6 ,1 7 8
Cynometra iripa 21 Eucheuma denticulatum 223
C. webberi 35 Euclea schimperi 34
Cyperus rotundus 14 Euphausiacea 120
C. tuberosus 34 Eutrophication 125, 127
Evaporation pans/ponds 6, 210
Dactyloctineum g em in a tu m 34 Evapotranspiration 141, 145
Dalbergio sp 16 Excoecaria agallocha 11, 14, 15, 16, 21, 25, 182, 253
Damselfish 126, 127
Density gradient 1 6 4 ,1 6 7 ,1 6 9 ,1 7 0 Fecundity 215, 220
D erris sp 16, 34 Fim bristylis cymosa 16
D. trifoliata 14 F. triflora 34
D iadem a antillarum 125 Fish 1, 195, 198, 229
D. setosum 127 abundance 199, 200-202
Dictyosphaeria cavernosa 125 biomass 203
Dictyota 125 eggs 120, 216
Didemnum conchyliatum 91, 95 larvae 120
Diffusion coefficient 148 maturity 200, 203, 216
 263

production 122 H yphaene compressa 34

species 197,198, 200-201, 233 H ypnea cornuta 64
Fishery 30, 121, 163, 170, 229, 232 H. hamulosa 64
Flacourtia indica 34 H. nidifica 64
Flow separation 156,158 H. valentiae 64
Flushing 145, 149, 150, 178
Fouling 87 ‘Ice-ice’ disease 226
Fourier analysis 109 Indicator species 65
Freshwater 143, 150, 175,178, 181, 191, 203 Inhaca Island 49, 50
Frictional slope 142 Insecta larvae 101, 105
Fringing mangroves 17, 37,141, 239 Intertidal zone 72, 247
reefs 127 Intratidal stratification 166
Fronts 151, 152 variability 169
Fuelwood 6, 29, 3 0 ,3 5 Ipomoea tuba 14,15
Fuirena ciliaris 34 I. pescaprae 34
Isipod 8 7 ,9 1 ,9 5
Gastropod 87, 93, 101, 105 Isopoda 1 0 1 ,1 0 5 ,1 2 0
Gastrotricha 101,105
Gazi Bay/Creek (Kenya) 59, 99, 109, 119, 243 Jamaican mangroves 236, 240
Gili net 200 reef 133
Godavari delta 11
Gracilaria corticata 59, 64 Kaneohe Bay (Hawaii) 125
G. crassa 126 Kappaphycus alvarezii 223
G. salicornia 59, 64 Kenya 29, 5 9 ,9 9 ,2 0 9 , 215
G. verucosa 64 Kinorhyncha 101, 105
Great Barrier R eef 127 Kodiakkarai mangrove (India) 187
Grewwia vaughanii 34 Kosi estuary (South A frica) 56
Ground reconnaissance 1, 3, 29, 31, 51
Groundwater flow 144, 145 La Saline R eef (Reunion) 126
Gypsum 212 Lac-lagoon (Bonaire) 139
Laguncularia racemosa 88, 235
Haemulon carbonarium 140 Landsat M ultispectral Scanner imagery 30
H. chrysargyreum 140 Lateral trappings 141, 148,149, 151-154,159
H. flavolineatum 140 Lates calcarifer 195,198
H. sciurus 140 Laurencia papillosa 64, 65
Halcyon senegaloiides 5 L eather tanning 232
Halecium spp 90 Leiognathidae 197,198
Haliclona sp 89, 95 Lesser Black-backed Guii 6
Halimeda 65, 90, 125 Lethrinidae 233
H. macroloba 59, 64 Lethrinus harak 232
H. opuntia 59, 64 Litterfall 37, 38, 40, 41, 4 2 ,4 3 , 4 4 ,4 6
Halobacterium 212 Little Stint 6
H alocordyle disticha 90 Littorina scabra 5
H alodule uninervis 59, 63 Liza subviridis 195, 198-200
H. wrightii 59, 63, 65, 66 Longitudinal advection 163
H alophila m inor 63 Lum nitzera racemosa 3, 15, 20, 21, 22, 29, 32, 38, 50, 110,
H. ovalis 59, 63, 65, 66 2 31,243
H. stipulacea 59, 63, 65, 66 Lutjanidae 198, 233
H alophilic microbes 212 Lutjanus apodus 140
Halophytes 3 L. griseus 140
Heliotrophium curassaricum 16 L. mahogoni 140
H elobiales 63
H eritiera 5 M acroalgae 59, 6 3 ,6 9 , 78, 94, 121,124, 125,127
H. littoralis 231 Macrobrachium equidens 199
Hermit crab 232 M afia island 2
Hinchinbrook Channel 141 Mangals 17, 25, 37, 38, 43, 45, 46, 78, 238
Hurricane 125, 133, 134, 135, 136,137 M angrove 1-10, 17, 29, 30, 37, 94, 99, 139,140, 169,181,
Hydrodynamic 121, 141 187, 195, 209, 235
Hydrography 181 assessment 238
Hydrozoan 87 biomass 21, 23, 37, 39
 264

M angrove (continued) M ontipora 127

cays 88 M oorea (French Polynesia) 126
community structure 11, 17, 24, 25, 29, 235, 253-255 M or modica trifoliata 34
conversion 1,209 Mozambique current 55
creeks 109, 119, 145, 149, 150,151, 152, 158, 160, 197 M twara district 7
degradation 1, 144, 210 M ud content 105
density 256, 257 M ud skipper 5
detritus 141, 148, 151, 163, 170, 201 M udflats 197, 255
distribution 1, 11, 15,17, 29, 49, 52, 230, 243, 246, 247 Mugilidae 197
ecosystems 6, 17, 3 7 ,1 7 5 ,1 7 8 , 185, 253 M urex 233
estuary 141, 150, 163,1 6 5 ,1 7 6 M uthupet mangrove 253
exploitation 1, 6, 30,22 9 , 244, 258 M yriostachya wightiana 11, 14, 15
fauna 3, 5, 6, 88, 195, 229, 232
forest 3, 5 ,1 5 0 Nauplii 101,105, 120
inlets 195, 197 Nearshore zone 173, 174, 177, 178
interactions 241 N ematoda 9 9 ,1 0 1 , 105
litter 37, 40, 46, 141, 144,170, 178,186 Nereodictyon 64
management 1 ,7 , 8, 36, 213, 229, 235, 253, 257 N. imitans 64
meiofauna 99-106 Nerita 5
mortality 49, 54, 57 Netherlands Antilles 139
poles 6, 29, 34 Nutrient 2, 78, 83, 127, 163, 170,173, 178, 181,184, 191,
production 3 7 ,4 5 ,4 6 , 239 239
propagules 56, 244 budget 141
regeneration 4 9 ,5 1 , 54 ,1 7 3 , 243,259 cycling 181
root 87, 96 stripping 59
seedling 51, 57, 173, 243,248
species 14,17, 20, 21, 24, 32, 33, 35, 50, 244, 253 Ocean currents 2
swamp 141, 142, 143, 144, 148, 149, 151, 159, 169, Ocyurus chrysurus 140
173,174, 175, 178, 182 Oecophyllia 5
timber 29, 35, 231, 232 Offshore reefs 127
uses 6, 29, 30, 34, 229, 231, 232, 258 Oligochaeta 99, 101,105
vegetation 1, 3, 19, 34, 54, 99, 100, 229, 254 Oocytes 216
zonation 20, 25, 78, 243,245, 247,253 developm ent 221
M anning coefficient 143 distribution 216
M ariculture 209, 223 recruitem ent 220
M aturity stages 203, 216 Oocytic atresia 215, 219
M aytenus undata 35 O puntia monocantha 255
Medusae 120 O rganic m atter 45, 239
Meiobenthos 99 Oscillatory currents 148
Meiofauna 99, 203 Ostracoda 9 9 ,1 0 1 , 105, 120
composition 106 Outwelling 152, 159, 160, 163
densities 9 9 ,1 0 1 -1 0 6 Overfishing 1 2 5 ,1 2 6 ,1 2 9
distribution 99, 101, 105 Oxygen 82, 181, 183
M eliaceae 3 Oyster 99
Melongena 96 shell 109, 112, 115
Melongena melongena 87, 93
M erbok estuary (Malaysia) 164 Padina 125
Metapenaeus affinis 195,199 P. boryana 64
M. brevicornis 195, 199 Palythoa caribaea 90
M ethane production 187 Panicum infestum 34
M ethanogenic bacteria 187, 191 P. piniplium 34
M icrobiota 77 Parapeneus 199, 233
Microfauna 203 Particle size 59, 9 9 ,1 0 5
Microphytoplankton 77 Particular m atter 13, 14
Mida Creek 29, 30, 33, 35, 243 Particulate organic carbon 77, 84
Millepora alicornis 90 Pellina 94
M ollusc 120, 229, 232 P. carbonaria 90
Momentum equation 142,143 Pemba island 2, 229
Monastraea annularis 134 Penaeidae 120, 197, 201
Monsoon 2, 119, 173, 174, 178, 198 Penaeus indicus 195, 199, 201
 265

Penaeus merguiensis 195, 199, 201 Rufiji delta 6, 173

P. penicillatus 195, 199, 201 R unoff 145,146, 152, 153,163, 165, 166
Perophora form osana 91, 95
Phaeophyta 64, 90,12 6 Sabellidae 90
Phase-shift 121,124 Salicornia sp 34
Philippines 223 Salinas 209, 235
Phoenix paludosa 34 Salinity 2, 3 8 ,5 6 ,7 8 , 83, 8 7 ,9 2 , 9 3 ,1 0 1 ,1 4 1 ,1 5 2 ,1 5 3 ,
P. reclinata 34 154,165, 1 6 9 ,1 7 3 ,1 7 5 , 181, 182, 211
Phragnites communis 34 gradient 146, 148, 163, 165,166, 168,169, 170
Phycolumnoria clarkae 91 stratification 150,163
Physical processes 164 Salt pans/ponds 1,191, 253
Physiological tolerance 3 production 6, 210, 231
Phytoplankton biomass 83 transport 163,170
comm unity 77 Salvadora persica 34
density 83, 85 Sargassum sp 126, 127
distribution 7 Satellite image 29, 30, 31, 153, 173,175
species 77, 80-82 Scaridae 233
Pitchavaram mangals 77 Sciaenidae 197
Pocillopora spp 127 Scolopsis 233
Pollution 7 ,1 3 ,1 2 1 Scylla serrata 4, 232
Polychaeta 99, 101, 105, 120 Scyphiphora hydrophyllaceae 21
Polygons 141 Sea urchin 125, 126
Pond 1,210 Seagrass 59, 63, 139, 140
Porifera 90 succession 59
Porites 127 uses 60
Portugal bay 19 zonation 59, 69
Portuguese Island 49, 50, 54, 55 Seaweed 63
Potamides decollatus 5 Sediment 2, 5 9 ,9 9 ,1 0 1 ,1 0 5 ,1 7 5 ,1 7 7 ,1 7 8 ,1 8 8
Praw n 1, 195, 198, 202, 209, 232 Seepage 64
biomass 203 Selangor (M alaysia) 195
food 202 Sergestidae 120
species 197,199 Serpulidae 90
Predation 201 Sertularia turbinata 90
Premna spp 34 Sesarma 3, 107
Pressure gradient 156 S. guttata 50
Prim ary production 3 7 ,1 8 1 ,1 8 4 S. meinerti 50
Propagule 56, 243, 244 Sesuvium portulacastrum 3, 34, 55
Prosopis spicigera 255 Sewage pollution 121, 125,129
Psilotrichum sericeum 34 Shear 148
Pum p-fed ponds 212 Sideroxylon inerme 34
Puttalam lagoon 17,19, 24, 25 Siganus 215, 232
S. canaliculatus 220
Rabbit fish 215 5. sutor 215
Rajidae 233 Siltation 7 ,1 2 5 ,1 2 9 , 141,144
Redox potential 189 Siphonophora 120
R eef fishes 139, 140, 216 Sonneratia 60
Remote sensing 29, 31 S. alba 3, 21, 29, 32, 33, 34, 99, 110, 231, 243
Rhizom es 60 S. apatela 11
Rhizophora apiculata 15, 21, 77 Sonneratiaceae 3
R. harrisonii 235 Sparisoma viride 140
R. mangle 87, 88,95 , 235 Spawning 215
R. m ucronata 3, 17, 20, 21, 22, 25, 29, 32, 33, 34, 38, Sphacelaria rigidula 64
40, 41—46, 50, 5 4 ,9 9 ,1 1 0 , 238, 243 S. tribuloides 64
R. racemosa 235 Sphaeroma terebrans 91, 95
Rhizophytic algae 73 Sphyraena barracuda 140,195
Rhodophyta 64, 90 Sponge 87, 95
Rice farming 7 Sporobolus virginicus 55
Roseate Tern 6 Sri Lanka 17, 37, 38
Rosenvingea orientalis 64 St Lucia mangroves 237
Rotifera 99, 101, 105 Stegastes nigricans 127
266

Sterculiaceae 3 Trap 169, 170, 173, 174, 177, 178

Stomatopoda 120 Tunicata 91
Storms 133, 134, 145 Turbellaria 99, 101, 105
Stratification 164, 166, 168, 169, 170 Turbinaria ornata 126
Suaeda maritima 1 1 ,1 4 ,1 5 ,1 6 , 255
Subtidal zone 72 Uca 3, 107, 201
Sueda 34 U. annulipes 50
S. monoica 3 U. gaimardi 50
Syringodium isoetifolium 59, 63, 65, 66 U. inversa 50
U. urvillei 50
Tanga district 5 Udotea orientalis 64
Tannins 29 Ulva lactuca 64
Tanzania 1, 173 U. reticulata 64
Tardigrada 101,105 Unguja island 229
Tedania 95 Up welling 158
T. ignis 90
Terebralia 5 Vertical mixing 163, 165, 166, 169, 170
Thalassia 59, 69 shear transport 1 6 3 ,1 6 6 ,1 6 8
T. hemprichii 59, 63, 65, 66 Vortex lines 158
T. testudinum 139 Vorticity dynamics 156
Thalassodendron ciliatum 59, 63, 65, 66
Thespesia danis 34 W ater mixing 146,152, 1 5 4 ,1 5 7 ,1 5 9 ,1 7 3 ,1 7 6 , 178
Tidal amplitude 248 tem perature 87, 92, 93, 183, 188
assy me try 144
circulation 1 4 1 ,1 4 3 ,1 4 6 , 156 Xylocarpus 5, 32, 33
creeks 141, 1 4 2 ,1 4 3 ,1 4 6 ,1 7 3 X. granatum 21, 29, 32, 34, 110, 231,243
currents 141,144,146,149,153,158,165 X. moluccensis 230
diffusion 148
je t 158, 159, 160 Zanzibar 2, 173, 229
recharge 54 Zoning 9
velocities 141, 148, 149, 167 Zooplankton composition 120
Tilapia culture 212 distribution 119
Tourism 8, 121 diversity 119
Toxic wastes 129 Zostera 60
Toxotes jaculator 195 Zostera capensis 63