Forgot your password?
Document Actions
Re-Establishment of Epibiotic Communities in Reforested Mangroves of Gazi Bay, Kenya (Crona et al, 2006)
Wetlands Ecol Manage (2006) 14:527–538
DOI 10.1007/s11273-006-9005-7
ORIGINAL PAPER
Re-establishment of epibiotic communities in reforested
mangroves of Gazi Bay, Kenya
B. I. Crona Æ S. Holmgren Æ P. Ronnback
¨ ¨
Received: 6 December 2004 / Accepted: 17 February 2006 / Published online: 28 July 2006
Ó Springer Science+Business Media B.V. 2006
Abstract Recolonization of epibiotic flora and community composition between landward and sea-
fauna in two fringing Sonneratia alba reforestation ward zones were observed in all sites and trunk
plots was investigated and compared to a natural fouling fauna was distinctly different between sites.
mangrove stand and a denuded site in Gazi Bay, Reasons for the above patterns are discussed and it is
Kenya. The reforested sites differed with respect to suggested that zonation patterns affecting pneumato-
land history and planting density. Habitat availability phore surface and inundation time, in combination
in the form of pneumatophore surface differed among with proximity of sites to natural seeding areas, are
forested sites (P<0.001), and between landward and the most likely explanations for observed patterns of
seaward zones (P<0.05). Eighteen algal species were epibiotic community distribution in this study.
found in the natural area compared to 23 and 10 in
replanted sites. Only one species was encountered in Keywords Sonneratia alba Æ Algae Æ Epibiotic
the denuded area. SIMPER analysis distinguished communities Æ Recolonization Æ Replanted
Enteromorpha ramulosa, Polysiphonia sp., Hypnea mangroves Æ Sponges
sp. and Caloglossa leprieuri as the main algal species
responsible for differences between sites. Algal bio-
mass was positively correlated to pneumatophores
Introduction
area (P<0.001). Total algal biomass differed mark-
edly between forested sites: 1.4 (matrix replantation),
In response to the increasing decimation of man-
28.6 (natural stand) and 44.3 g mÀ2 (integrated
groves forests several re- and afforestation programs
replantation) in the seaward zones. The matrix
have been initiated world wide (Imbert et al. 2000;
replantation showed strong differences in algal com-
Kairo et al. 2001) and the need for mangrove reha-
munity assemblages compared to the other forested
bilitation has become recognized as a high priority in
sites, and this site also had significantly lower bio-
local coastal management plans for many developing
mass of sessile benthic fauna (P<0.001). Statistical
´
countries (Linden and Lundin 1996). Earlier man-
differences in algal (P<0.01) and sponge (P<0.05)
grove restoration goals have ranged from supply of
quality wood for logging to shore-line stabilization,
often overlooking the role of mangroves as habitats
B. I. Crona (&) Æ S. Holmgren Æ P. Ronnback
¨ ¨
for a diverse flora and fauna (Field 1996). Still today
Department of Systems Ecology, University of Stockholm,
the majority of rehabilitation programs focus mainly
S-106 91, Stockholm, Sweden
on structural aspects of reforestation (Ellison 2000),
e-mail: beatrice@system.ecology.su.se
123
528 Wetlands Ecol Manage (2006) 14:527–538
baseline information needed as the interest for man-
thus neglecting important issues such as the
grove rehabilitation increases world wide.
recolonization of associated flora and fauna, which is
essential if the ecological functions of a replanted
mangrove forest are to be restored.
The role of epibiotic communities of both flora and Methods
fauna in mangroves has received only limited scientific
attention to date. Ellison and Farnsworth (1990; 1992) Study area
and Ellison et al. (1996) demonstrated the role of root
fouling communities, both directly and indirectly, on Gazi Bay is located on the southern Kenyan coast at
the growth of mangrove roots. The fouling community 4°25¢ S and 39°50¢ E. The inner estuary is sheltered
effectively prevents isopod colonization which can from intense wave impact by shallow reefs at the
otherwise reduce root growth by more than 50% and mouth of the bay (Fig. 1). Seasonal monsoons dom-
facultative mutualism has been reported to occur be- inate the climate with two pronounced rainy seasons;
tween massive sponges and mangroves thus potentially a period of heavy rains from April–June (SE mon-
enhancing mangrove productivity (Ellison et al. soon) and a period of lighter rains from October–
1996). Proches et al. (2001) also showed that pneu- November (NE monsoon). Total annual rainfall
matophores and their associated epibiota provide a ranges from 1000 to 1600 mm and the salinity in the
unique structural feature to the physical environment study area ranges from 24 to 26.5 ppt during the SE
of mudflats resulting in increased microhabitat com- monsoon (Kitheka 1997). High flushing rates are
plexity and giving rise to arthropod assemblages dis- coupled with low water residence times (3 to 4 h) and
tinctly different from those of surrounding sediments. result in high rates of exchange between inshore and
Sponges and filamentous algae may constitute offshore waters (60–90% of the volume per tidal
food for fish foraging in the mangroves at high tide. cycle) (Kitheka 1997).
Several species of angelfish of the genus Pomacan-
thus, butterflyfish (Chaetodon) and some filefish Site description
(Cantherines) have a large sponge component in their
diet (Randall and Hartman 1968). Although adults of Four sites were investigated between March and June
these species are mainly reef-associated, some have 2002; one natural stand of Sonneratia alba (N) and
been documented to enter mangroves (de Troch et al. two replanted sites of the same species (MP and IP)
1998). Many juveniles of reef-bound species, such as of equal age (8 years) but differing planting density,
Tetradontidae, Ostraciidae and Chaetodontidae occur and one denuded site (D) included for comparison.
in mangroves and are known to feed on filamentous Site IP (Integrated Plantation) was replanted in a
algae and benthic invertebrates (Randall and Hartman degraded stand of S. alba with a planting density of
approximately 4330 trees haÀ1 (Kairo 1995). This
1968; Dunlap and Pawlik 1996; de Troch et al. 1998).
Substrate availability as a limiting resource is a site thus contained a certain amount of forest cover at
factor affecting most epibiotic communities (Osman the time of planting which distinguishes it from MP.
1977) in combination with variable complexity in Site MP (Matrix Plantation) was planted on denuded
habitat structure attributable to absolute abundance of ground which had been clear-felled under commer-
individual structural components such as pneumato- cial logging practices in the 1970’s (Kairo 1995) and
phores (Beck 2000). Since little information has been where no natural regeneration of mangroves had oc-
published on the return of ecosystem functions in curred. This site was planted in a 1.0·1.0 m matrix
rehabilitated mangroves the objective of this study with a planting density of approximately 10,250 trees
haÀ1 (Kairo 1995). Since their establishment pruning
was to investigate recolonization of the epibiotic flora
and fauna on pneumatophores and tree trunks in re- has been carried out twice but no thinning has been
planted mangroves, Sonneratia alba, in Gazi Bay, done in either of the plantations (Kairo personal
Kenya, and compare the findings with epibiotic communication). Site (D), adjacent to MP, was also
communities in adjacent natural mangroves and a logged in the 1970’s. Each site was subdivided into
denuded stand (currently a sandflat). This is valuable two zones, a landward zone (ZI) and a seaward zone
123
Wetlands Ecol Manage (2006) 14:527–538 529
Fig. 1 Map of the study
area, Gazi Bay, Kenya. The
area is located on the
southern Kenyan coast at
4°25¢ S and 39°50¢ E.
N=natural stand,
IP=integrated plantation,
MP=matrix plantation and
D=denuded site
(ZII) based on visual observations of a natural Relevant site characteristics are summarized in
epifaunal gradient and inundation patterns. Zones Table 1.
were approximately 25 meters deep and elevation
differences between them ranged between 0.61 to Sampling design
0.96 meters resulting in a slope around 2°, resulting
in differences in inundation, between zones, of Sampling was done using 0.5·0.5 m wood frames.
approximately 60 min per tide. All sites were lo- Ten replicates were taken in each zone and were
cated along a tidal channel (Fig. 1) and belong to selected by randomly paired coordinates, giving a
inundation class I as described by Watson (1928), total of 80 samples for all four sites. A sub-sample of
with inundation approx 12 h dayÀ1. The size of the roots in each frame were measured and counted. The
sites ranged from 1,715 m2 (MP) to 10,845 m2 (N). base diameter and height of each pneumatophore was
123
530 Wetlands Ecol Manage (2006) 14:527–538
Table 1 Summary of features (range or mean – SE) characterizing the Sonneratia alba sites studied in Gazi Bay, Kenya
Site Natural (N) Integrated plantation (IP) Matrix plantation (MP) Denuded (D)
Feature
Site area (m2) 10,845 7,931 1,715 3,741
Planting densitya (# haÀ1) – 4330 10,250 –
Current stand densityb (# haÀ1) 4300–1221 – 7640–600 –
Energy exposure High High Low Low
Relative site elevation (m) 0.14 0 0.37 0.37
Sediment typec ZI: mud ZI: mud ZI: muddy sand ZI: sand
ZII: muddy sand ZII: mud ZII: mud ZII: muddy sand
Sediment organic contentc 7.9–1.6 14.3–1.1 9.2–1.8 1.6–0.3
Pore water salinity* ZI: 34.6–0.3 ZI: 34.4–0.2 ZI: 34.0–0.8 ZI: 35.9–1.1
ZII: 34.5–0.3 ZII: 34.2–0.1 ZII: 34.6–0.2 ZII: 27.4–5.8
Water column salinity 36.5–38.0 36.5–38.0 36.0–37.5 36.0–37.5
Canopy cover 50–75% 50–75% 100% 0%
Pneumatophore density (# mÀ2)* ZI: 174–21 ZI: 424–24 ZI: 380–41 –
ZII: 280–37 ZII: 322–35 ZII: 400–26
Pneumatophore area (m2 per m2 forest)* ZI: 0.30–0.03 ZI: 0.41–0.06 ZI: 0.16–0.01 –
ZII: 0.55–0.07 ZII: 0.55–0.6 ZII: 0.24–0.03
Average trunk area (cm2 per m2 forest)* ZI: 1.2–0.2 ZI: 0.4–0.1 ZI: 1.3–0.4 –
ZII: 0.7–0.2 ZII: 0.8–0.1 ZII: 1.0–0.1
Trunk diameter range (cm)* Min: 4.3 Min: 2.5 Min: 1.2 –
Max: 93.5 Max: 48.0 Max: 8.6
a b c
¨ ¨
Data from Kairo (1995), Bosire et al. (2003), Crona and Ronnback (2005). ZI=landward zone, ZII=seaward zone, *n=10
determined and used to calculate root surface area for colonization, from the sediment surface and one
available for epifaunal colonization. For ease of meter up (Table 1). All sessile organisms found
surface computation pneumatophores were treated as growing on the lowest meter of each tree trunk were
perfect cones. Pore water samples were randomly counted and determined to genus, and species level
collected in each site by digging a hole in the sedi- wherever possible.
ment 10–15 cm deep, and poor water and water
column salinity was measured using an optical Statistical analysis
refractometer (Atago brand). All epibiota on both
roots and sediment within the frame was removed and Species abundance, biomass and frequency data were
dried at 60°C to constant weight. Sessile fauna double square root transformed and subjected to non-
growing on the roots was determined to the taxo- metric multi-dimensional scaling ordination (nMDS)
nomic level of phyla. Although a detailed account of using the Bray–Curtis similarity coefficient (Field
the diversity of the sponges and ascidians would have et al. 1982). The nMDS is a method based on a non-
been preferable, the poorly documented taxonomy of parametric regression of distance on dissimilarity of
these organisms in East African waters made this samples. The goodness-of-fit of the regression line is
difficult. As a result they were grouped into respec- evaluated by calculating a stress value which is a
tive phyla both representing the functional group of measure of how well the MDS succeeded in fitting the
sessile filter feeders on pneumatophores. The poten- multidimensional data onto a 2-dimesional plane. An
tial benefits of such grouping when studying stability alternative assessment of the dimensionality of the
and persistence of marine benthic communities has dataset is presented by non-parametrically correlating
been reviewed by Steneck and Dethier (1994). Algae the original sample similarity matrix with that repre-
were determined to genus and species wherever senting distance between samples in ordination space.
possible (Jaasund 1976; Richmond 1997). For each Biomass data for both flora and fauna was tested for
site ten random trees were also selected in each zone significant differences using analysis of similarity
(a total of 60 trees) and the circumference was randomization tests (ANOSIM, Clarke and Green
measured and used to calculate trunk area available 1988; Clarke and Warwick 2001). Algal species
123
Wetlands Ecol Manage (2006) 14:527–538 531
Algal biomass and total epibiotic biomass on
responsible for differences in sites observed in nMDS
roots and sediment
plots were identified with a dissimilarity percentage
program (SIMPER, Warwick et al. 1990). Data on
For all sites, algal biomass and combined epibiotic
root complexity and trunk area were analysed using
biomass (algae, sponges and ascidians) were mea-
nested, one-way ANOVA. When conditions for use of
sured. Biomass was consistantly higher in the seaward
parametric statistics were not met Kruskal–Wallis
zone (ZII) for all sites and the natural (N) and
and Mann–Whitney U tests as well as Spearman rank
replanted site (IP) had the highest total biomass of
correlation were applied. For multiple comparisons
algae and sessile fauna (Table 2). The replanted site
significance levels were adjusted using the Bonferroni
(MP) had lower algal biomass in the seaward zone
method (Rice 1989). All statistical analysis
(ZII) than the adjacent denuded plot which was a re-
were performed using Primer 5 (version 5.2.1) or
sult of the high presence of Enteromorpha ramulosa in
Statistica 6.0.
the cleared area, a green algae favoured by intense
sunlight and growing directly on the sediment. nMDS
Results ordination revealed a high similarity between sites IP
and N for both algal and total epibiotic biomass
Surface available for colonization (Fig. 2a, b) expressed by the small distance between
sites in the plot. Sites D and MP differed markedly
Results from a hierarchically nested ANOVA (zones from all sites except for combined epibiotic biomass
nested in sites) performed on number of pneumato- where sites MP and D showed slightly higher resem-
phores and total pneumatophore area (Table 1) reveal blance (Fig. 2a, b). This similarity was due, in part, to
significant differences between sites (P<0.001) and a common lack of poriferans and ascidians among MP
zones (P<0.05), which supports the sub-division of and D samples. The low stress of the nMDS plot
each site into zones in the analysis of biological data (<0.05) suggests a very good fit of the multidimen-
that follows. sional data onto the 2-dimensional plot (Clarke 1993).
Average trunk area mÀ2 available for colonization The dimensionality of the datasets for both algae and
by sessile fauna was compared for sites N, IP and MP total biomass was further assessed through non-para-
(Table 1). Results of a one way hierarchically nested metric Spearman correlation between distance in the
ANOVA (zones nested in sites) show a highly sig- ordination space and distance in the original
nificant difference in available trunk area between p-dimensional space (ralgae=1.0, rtot bio=1.0) for both
sites (P<0.001). nMDS plots presented in Fig. 2. To test the relation-
ship between algal biomass and pneumatophore area a
Species richness of epibiotic communities Spearman rank order correlation was conducted,
resulting in a positive relationship (ralgae=0.55,
The algal species encountered in each zone and site P<0.001). A similar correlation was also run to test
are shown in Table 2. Site D had only one species of differences between zones in terms of biomass of
algae, Enteromorpha ramulosa. This algae was found algae and sponges (ralgae=0.39, P<0.01; rsponge=0.22,
growing both on hard substrate and freely on the P=0.046).
sediment surface. Site MP had a total of ten algal Analysis of similarities (ANOSIM) between sites
species, all of which were found in zone II, while for algal biomass and combined epibiotic biomass
only four species were found in zone I. In sites IP and (R=0.422 and R=0.407 respectively, P<0.001) indi-
N a total of 23 and 18 species were found respec- cates dissimilarity between sites. Furthermore, the
tively. For all forested sites the algae found belong to results reveal a high degree of dissimilarity between
classes Rhodophyta and Chlorophyta. Several dif- all pair wise between-site comparisons (P<0.001)
ferent taxa of both ascidians and poriferans were except for sites IP and N which were similar for algal
found growing on the pneumatophores of Sonneratia biomass but not for average combined epibiotic bio-
alba stands in Gazi Bay. Tedania digitata vulcanis, mass (P=0.015). To explore why, Mann–Whitney U
which could be positively identified to species, tests were used to test differences in average sponge
dominated the poriferan community. and algal biomass between and within the natural
123
Table 2 List of organisms encountered at each studied site of Sonneratia alba in Gazi Bay, Kenya
532
Zone I Zone II
123
N IP MP D N IP MP D
Biomass of organisms on pneumatophores
Rhodophyta (Algae)
Catenella nipae 1.8 (1.8) 6.2 (5.4) – – 329 (217) 19.9 (19.9) 46.8 (23.3) –
Ceramium brevizonatum var. caraibicum 41.0 (23.4) 68.0 (20.7) – – 10.4 (8.3) 29.2 (24.9) – –
Ceramium sp. – 9.8 (9.8) – – – – – –
Coelothrix irregularis 0.8 (0.8) – – – – 2.1 (2.1) – –
Gelidiopsis intricata – 0.3 (0.3) – – – 2.7 (2.7) – –
Gracilaria crassa – – – – – 6.4 (6.4) – –
Gracilaria salicornia 321 (300) 352. (266) – – 1503 (821) 3294 (1537) – –
Heterosiphonia sp. – – – – 0.9 (0.9) – – –
Hypnea sp. 5.6 (3.5) 107 (70.1) – – 325 (190) 160 (50.2) – –
Laurencia perforate – 5.0 (5.0) – – 12.8 (12.1) – – –
Polysiphonia sp. 180 (71.9) 177 (173) – – 451 (284) 675 (432) 80.6 (35.2) –
Unidentified red algae. T1 – 30.5 (26.9) – – 36.6 (32.6) – – –
Unidentified red algae. T2 – 1.2 (1.2) – – – – – –
Phaeophyta (Algae)
Bostrychia binderi 0.6 (0.6) – – – – – – –
Bostrychia radicans – 20.1 (11.2) – – 38.6 (21.5) 5.9 (4.2) 2.6 (2.6) –
Caloglossa leprieuri 90.1 (48.1) 139 (35.7) 1.7 (1.7) – 140 (47.1) 140 (44.3) 7.5 (7.4) –
Endosiphonia clavigera – – – – – 12.4 (12.4) – –
Chlorophyta (Algae)
Chaetomorpha crassa – 5.8 (4.3) – – – 2.0 (1.3) – –
Cladophora patentiranea 5.2 (4.2) 9.6 (6.2) 3.5 (2.4) – 10.9 (8.4) 2.8 (2.8) 1.6 (1.6) –
Enteromorpha kylinii – 1.3 (0.9) – – 0.7 (0.7) – – –
Enteromorpha ramulosa 0.4 (0.4) – – – – – – 688 (407)
Ulva pertusa – – – – – 73.9 (49.9) – –
Unidentified green algae – 0.7 (0.7) – – – – –
Cyanophyta unidentified 12.4 (8.4) 1.5 (1.5) – – – – 27.2 (15.1) –
Diatoms – 12.7 (12.7) – – 64.9 (63.5) 16.3 (14.7) – –
Hydrozoans 21.6 (21.6) 14.3 (12.4) – – 39.2 (23.4) 77.0 (53.1) 0.8 (0.8) –
Total algal biomass 681.0 961.9 5.2 0 2963 4519 167.0 688.8
Chordata
Ascidians 0.26 (0.2) 0.4 (0.2) – – – 1.5 (0.7) – –
Porifera
Poriferans 20.7 (8.3) 20.0 (9.1) – – 43.8 (9.1) 75.9 (14.7) – –
Total combined biomass 21.65 21.32 0.005 0 46.73 81.88 0.169 0.689
Wetlands Ecol Manage (2006) 14:527–538
Wetlands Ecol Manage (2006) 14:527–538 533
Mean (SE) dry weight (mg m ) for algae, (g m ) for ascidians and poriferans on pneumatophores (n=10). Mean (SE) abundance (# m trunk area) of trunk epifauna (n=10).
stand and integrated plantation. Tests between sites
proved non-significant (psponge=0.516; palgae=0.250)
while Zone I and II of the integrated plantation dif-
D
–
–
–
–
–
–
–
–
–
fered significantly (psponge=0.008; palgae=0.041) and
Zone I and II of the natural stand showed somewhat
389 (162)
547 (198)
weaker trends (psponge=0.059; palgae=0.049).
179 (50)
62 (19)
Analysis of similarity percentages (SIMPER) of
17 (7)
2 (2)
2 (2)
MP
algal species distributions was done to reveal the
–
–
main species responsible for observed dissimilarities
among sites (Table 3). Enteromorpha ramulosa was
the sole species responsible for characterizing site D
À2
Total combined biomass (algae, ascidians and poriferans) expressed in g mÀ2Natural stand (N), replanted (IP and MP) and denuded (D)
40 (12)
as it was the only species of algae found in this
2 (2)
2 (2)
4 (3)
4 (3)
2 (2)
habitat. Consequently it also played a dominant role
IP
–
–
–
in distinguishing this site from all other areas sam-
pled. Two species of algae contributed to over 80%
of total algal biomass in site MP as compared to four
436 (247)
574 (215)
109 (76)
Zone II
53 (45)
45 (19)
and three species in the natural stand and the inte-
4 (3)
2 (2)
grated plantation, respectively (Table 2). Polysipho-
N
–
–
nia sp. was responsible for over 50% of the biomass
in the matrix plantation, but was also present as a
D
–
–
–
–
–
–
–
–
–
potential characterizing species in both the natural
stand and integrated plantation. Three out of four
species characterizing sites N and IP were the same
3.3 (1.1)a
125 (44)
115 (70)
11 (11)
02 (2)
despite differences in their percentage contribution.
2 (2)
MP
–
–
–
Epibiotic communities of trunks
A total of nine animal species from three phyla were
17 (17)
recorded (Table 2). No sponges, ascidians or algae
9 (8)
2 (2)
2 (2)
IP
were found on the trunks and the majority of speci-
–
–
–
–
–
mens recorded were found around the upper limits of
the 1 m range of trunk area investigated. There is a
404 (139)
clear separation between sites although some inter-
132 (50)
28 (28)
Zone I
spersion of samples from site N and IP is found
(2)
(3)
(3)
(3)
4 (3)
N
(Fig. 3). No correlation was found between the
2
4
4
6
–
À2
number of species of sessile fauna present and
available trunk area. Correlation between trunk area
and abundance of the three most common fouling
Abundance of organisms on tree trunks
species was done revealing no correlation for Chi-
rona tenuis, a positive correlation for Chthalamus
À2
dentatus (r=0.29, P<0.05) and a negative correlation
Clypeomorus bifasciatus
for Balanus amphitrite (r=0.35, P<0.01).
Brachidontes variabilis
Patellid (unidentified)
Chthalamus dentatus
Cerithidea decollata
Terebralia palustris
Cirripedia (Crustacea)
Balanus amphitrite
Discussion
Table 2 Continued
Chirona tenuis
Patterns of biomass distribution
Nerita sp.
Gastropoda
Bivalvia
When examining community composition of the
investigated sites in terms of algal biomass two
123
534 Wetlands Ecol Manage (2006) 14:527–538
Fig. 2 (a) nMDS plots of total algal biomass per mangrove site tion, D=denuded site, I=landward zone, II=seaward zone. For
and zone in Gazi Bay, Kenya. Each symbol represents the site D, ZI contained no epibiotic biomass and consequently was
average value of that zone (n=10). Stress=0 (b) nMDS plots of not included in the MDS analysis. For both nMDS plots
combined epibiotic biomass per site and zone. Stress=0 distances between points are represented as equal distances in all
N=natural stand, IP=integrated plantation, MP=matrix planta- dimensions of the plot
Table 3 Algal species responsible for similarities within and dissimilarities among sites of Sonneratia alba mangroves in Gazi Bay,
Kenya
di/SD(di)
Site Species Contrib % Cum contrib % Avg sim
Species responsible for observed similarity within sites
1.64a
IP Caloglossa leprieuri 28.60 28.60 40.61
1.22a
Hypnea sp. 24.80 53.40
0.85a
N Polysiphonia sp. 45.36 45.36 36.13
0.84a
Caloglossa leprieuri 26.85 72.21
MP Polysiphonia sp. 0.47 51.45 51.45 26.03
0.36a
Catenella nipae 30.83 82.28
D Enteromorpha ramulosa ##### 100.00 100.00 100.00
Sites Species Dissim/SD Contrib % Cum contrib % Avg dissim
Species responsible for observed dissimilarity between sites
6.87b
D, MP Enteromorpha ramulosa 45.00 45.00 100.00
3.87b
D, IP Enteromorpha ramulosa 26.51 26.51 100.00
1.92b
Caloglossa leprieuri 12.38 38.88
1.17b
Hypnea sp. 12.27 51.16
1.22b
MP, IP Polysiphonia sp. 13.66 13.66 84.77
1.19b
Hypnea sp. 13.53 27.18
1.59b
Caloglossa leprieuri 13.20 40.38
3.25b
D, N Enteromorpha ramulosa 31.63 31.63 99.37
1.21b
Polysiphonia sp. 19.05 50.69
1.06b
Caloglossa leprieuri 13.20 63.89
1.07b
MP, N Polysiphonia sp. 19.53 19.53 75.50
1.07b
Caloglossa leprieuri 16.47 36.01
1.17b
IP, N Polysiphonia sp. 14.19 14.19 63.58
1.01b
Gracilaria salicornia 13.59 27.77
1.07b
Hypnea sp. 11.86 39.63
1.10b
Caloglossa leprieuri 10.88 50.51
1.03b
Ceramium brevizonatum var caraibicum 9.62 60.13
a
indicates that the species potentially characterizes the species assemblage of a site and b indicates a possible discriminating species
between sites. Analysis is based on algal biomass mÀ2
Natural (N), replanted (IP and MP) and denuded (D)
123
Wetlands Ecol Manage (2006) 14:527–538 535
species present in the former was E. ramulosa, a
green algae thriving in sun exposed environments
(Oliveira et al. 2005) and consequently lacking in the
adjacent shaded habitat.
If biomass of sponges and ascidians is included in
the nMDS analysis the pattern and spatial relationship
between sites change (Fig. 2b). While sites IP and N
remain tightly clustered site MP is now found posi-
tioned far from N, IP and D. The reason for this is the
lack of sponges and ascidians in all but one sample
from site MP. Since both of these organism groups
require hard substrate for their growth and survival
none were present in site D. Studies of sponge
assemblages suggest a number of factors responsible
for species distributions including water flow rate
(Maldonado and Young 1996), sedimentation and
nutrient levels (Bell and Barnes 2003), depth (Alv-
arez et al. 1990), light (Cheshire and Wilkinson
Fig. 3 nMDS of epibiotic trunk fauna based on species 1991), and habitat availability (Barthel and Tendal
abundance per standardized available mangrove trunk area 1993). Sedimentation and nutrient levels were not
(m2) in Gazi Bay, Kenya. N=natural stand (squares), IP=inte-
investigated in this study but are assumed to be
grated plantation (circles), MP=matrix plantation (triangles).
similar due to the close proximity and fringing po-
Stress=0.16
sition of all sites in the bay as well as the thorough
mixing of the waters entering the bay (Kitheka 1997;
distinct patterns emerge (Fig. 2a). For algal biomass
Mwashote and Jumba 2002). Lower root surface
there is a clear gradient with algal biomass of the
availability was significant in site MP, but the lack of
denuded site being markedly lower than replanted
pneumatophore surface may not provide a sole sat-
site IP and the natural site (N), while replanted site
isfactory explanation. Several investigations on fac-
MP has intermediate biomass values. This is also
tors affecting distribution and composition of
supported by a positive correlation between algal
mangrove epibiotic communities have concluded that
biomass and total pneumatophore area. Pneumato-
larval behaviour and longevity may affect recruit-
phore area in the matrix plantation (MP) is half of
ment patterns (e.g. Sutherland 1980; Bingham and
that in both the replanted site IP and the natural stand
Young 1995). New colonists seem to recruit primar-
(N) (Table 1). Although no causal relation can be
ily from local sources and sponge colonization has
determined with certainty the lower root area is
been suggested to occur from root to root (Farnsworth
surely important in determining the colonization
and Ellison 1996) or through fragmentation of adult
efficiency and recruitment possibilities of algae. If
colonies (Bingham and Young 1995). Ascidian larvae
algal recruitment follows the lottery model proposed
are known to swim short distances (2–15 m) (Davis
for epibiota (Sutherland 1980), lower pneumatophore
and Butler 1989) and many have developed swim-
area available would certainly reduce chance
ming behaviours promoting settlement in close
recruitment of algal spores from the water column. A
proximity to parental colonies (Bingham and Young
higher level of canopy cover observed in the matrix
1991). The location of site MP up-stream along the
plantation as opposed to the natural and integrated
creek from sites N and IP may thus be another reason
stands, and consequently lower solar radiation in the
why larval recruitment or entrapment of colonial
former plot, could also affect the abundance, biomass
fragments has not occurred.
and composition of algal species present. This is
Maldonado and Young (1996) showed that distri-
likely the main reason for observed higher algal
bution of four tropical demosponges, including
biomass values in the seaward zone of the cleared
Tedania ignis, a close relative to the dominant sponge
compared to the replanted area (MP), since the only
123
536 Wetlands Ecol Manage (2006) 14:527–538
in extreme ecological conditions (Oliveira et al.
in this study (Tedania digitata vulcani), was likely
2005). Average dissimilarity values between sites
affected by post-settlement mortality of larvae due to
reinforce the pattern and provide further evidence
suboptimal habitat conditions. This resulted in adult
that site MP has a distinctly different community
populations strongly associated with both high and
composition in terms of epiphytic algae as compared
low irradiance, but always with a high level of water
to the other forested sites (Table 3). Davey and
movement. Since sites IP and N have a higher wave
Woelkerling (1985) observed fluctuating frequency
energy exposure and presumable higher levels of
of colonizing red algae over time and related this to
water movement this may explain their abundant
grazing, competition for substratum and sloughing
sponge communities compared to site MP. Differ-
off of pneumatophore bark. De Troch et al. (1998)
ences in canopy cover may thus play a subordinate
reported that between 6 and 7% of the fish commu-
role in determining the presence of sponges.
nity sampled in Gazi Bay consist of herbivores, but
Observed differences in combined epibiotic bio-
effects of grazing on newly established algae in the
mass at the zone level for IP and N was a result of
area is unknown. Hence it is difficult to speculate on
much higher total biomass values for sponges and
the importance of herbivory to the lack of algal
algae found in zone II of IP as compared to zone II of
coverage in site MP. It is possible that the less diverse
N (Table 2). Correlation values also support this and
algal community in site MP is stable and may not
suggest that the main difference in terms of both
reach the higher species diversity of N and IP. Sim-
sponge and algal biomass is at the zone level, linked
ilarly altered community assemblages were observed
to inundation, rather than differences in vegetation
after disturbance by Davey and Woelkerling (1985)
structure and habitat availability between the natural
and in colonization experiments by Eston et al.
and replanted site.
(1992). Another explanation is that since replanting
in plot IP was integrated with sparse but existing
Algal species assemblages
forest cover the existence of epibiotic communities
on remaining root structures at the time of planting
A few studies have described mangrove associated
may have facilitated the ease and speed of recolon-
algal communities (Burkholder and Almodovar 1974;
ization. Closer proximity of site IP to the natural site
Beanland and Woelkerling 1983; Davey and Woel-
(N) compared with site MP is also plausible, although
kerling 1985; Rodriguez and Stoner 1990) with focus
a difference in proximity of approximately 500 me-
on algal communities in Rhizophora and Avicennia
ters seems unlikely to account entirely for the low
stands. To our knowledge none have dealt with epi-
diversity and recruitment after eight years.
biotic assemblages associated with Sonneratia sp.
The dominating algal species in this study were
Fouling fauna of mangrove tree trunks
Polysiphonia sp., Catenella nipae, Hypnea sp. and
Caloglossa leprieuri. Apart from Hypnea sp., these
species have all been found as dominant components The fouling community of mangrove tree trunks
among the studies mentioned above. Their respective differs entirely from that of pneumatophores. The
dominance within the different sites of this study primary reason behind the distinct difference in
varies however. Relatively low values of average communities is likely a higher tolerance to wave
similarity within all sites are likely due to the marked exposure and dessication by the barnacles dominating
zonation pattern described earlier. Algal diversity of the trunk fouling community. Patterns of species
forested sites was in the upper range of values re- abundance of sessile fauna on tree trunks reveal a
ported previously (Beanland and Woelkerling 1983; clear separation between sites (Fig. 3). Despite the
Davey and Woelkerling 1985; Rodriguez and Stoner lack of correlation between species richness and
1990). However, while replanted site IP and the available trunk area, species assemblages between the
natural stand (N) had a high diversity, replanted site sites apparently differ. Negative correlation for Bal-
MP counted approximately half the number of spe- anus amphitrite appears because of the highly local-
cies and the denuded site grew only one species, ized presence of this species in the matrix plantation
growing freely on the sediment and known to thrive (MP) where trunk diameters are smaller than the
123
Wetlands Ecol Manage (2006) 14:527–538 537
location and pre-planting history of sites may affect
other forested sites (Table 1). Barnacle larvae are
the return of associated flora and fauna such as those
known to exhibit patchy distributions in the water
studied in Gazi Bay.
column (Grosberg 1982) and availability of suitable
substrata, competition from other species and pres-
Acknowledgements This study was funded by the Swedish
ence of sexually mature conspecifics have been sug-
International Development Agency (Sida). The authors thank
gested to influence their settlement and consequent the Kenya Marine and Fisheries Research Institute (KMFRI) for
distribution (Bayliss 1993; Coates and McKillup logistical support. A special thank you goes to J. Bosire and
1995; Satumanatpan and Keough 2001). Bayliss J. Kairo for assistance in Kenya, to Abdul for tireless field
assistance, to Nils and Lena Kautsky for valuable comments and
(1993) observed highly aggregated distributions of
help with taxonomic identification, and to two anonymous
Balanus amphitrite on pneumatophores of Avicennia reviewers for helpful comments on how to improve the manuscript.
marina and suggested that mechanisms responsible
may be settling behaviour of cyprid larvae, post-set- References
tlement mortality, selective predation among others.
Furthermore, Satumanatpan et al. (1999) have shown Alvarez B, Diaz MC, Laughlin RA (1990) The sponge fauna on
a fringing coral reef in Venezuela, I. Composition, dis-
that settlement determines large scale distributions
¨
tribution and abundance. In: Rutzler K (eds) New per-
such as along shore variations in abundance observed
spectives in sponge biology. Smithsonian Institution
in this study while post-settlement mortality, in Press, Washington DC, pp 358–366
combination with settlement cues mentioned above, Barthel D, Tendal OS (1993) The sponge association of the
abyssal Norwegian Greenland Sea: species composition,
is responsible for small scale vertical distribution.
substrate relationships and distributions. Sarsia 78:83–96
This may explain the distinct pattern in barnacle
Bayliss DE (1993) Spatial distribution of Balanus amphitrite
species distribution and composition between studied and Elminius adelaide on mangrove pneumatophores. Mar
sites in Gazi Bay. Biol 116:251–256
Beanland WR, Woelkerling WmJ (1983) Avicennia canopy
effects on mangrove algal communities in Spencer Gulf,
South Australia. Aquat Bot 17:309–313
Beck MW (2000) Separating the elements of habitat structure:
Conclusion
independent effects of habitat complexity and structural
components on rocky intertidal gastropods. J Exp Mar
Results presented indicate that after eight years, Biol Ecol 249:29–49
marked differences in epibiotic biomass exist be- Bell JJ, Barnes KA (2003) Effect of disturbance on assem-
blages:an example using porifera. Biol Bull 205:144–159
tween replanted stands of Sonneratia alba and con-
Bingham BL, Young CM (1991) Larval behaviour of the
sequently also between the replanted and natural
ascidian Ecteinascidia turbinata Herdman; an in situ
stands studied. Differences among replanted sites as experimental study of the effects of swimming on dis-
well as between replanted and natural sites also exist persal. J Exp Mar Biol Ecol 145:189–204
Bingham BL, Young CM (1995) Stochastic events and
with regard to algal species assemblages. As both
dynamics of a mangrove root epifaunal community. Mar
reforestation plots were initiated in the same year the
Ecol 16(2):145–163
effect of time cannot solely account for the discrep- Burkholder PR, Almodovar LR (1974) Studies on mangrove
ancies in epibiotic communities between these two algal communities in Puerto Rico. Florida Sci 36:66–73
Cheshire AC, Wilkinson CR (1991) Modeling the photosyn-
sites nor are the physical parameters investigated
thetic production by sponges on Davis Reef, Great Barrier
significantly different between sites. Instead, clear
Reef. Mar Biol 109:13–18
zonation patterns of pneumatophore surface and Clarke KR (1993) Non-parametric multivariate analyses of
inundation time, in combination with proximity of changes in community structure. Aust J Ecol 18:117–143
Clarke KR, Green RH (1988) Statistical design and analysis for
sites to natural seeding areas, seem more plausible
a ‘biological effects’ study. Mar Ecol Progress Ser
explanations for the observed patterns of epibiotic
46:213–226
community distribution in this study. Due to lack of Clarke KR, Warwick RM (2001) Change in marine commu-
replication at the level of planting strategy and his- nities: An approach to statistical analysis and interpreta-
tion, 2nd edn. Primer-E, Plymouth
tory of sites prior to planting it is risky to draw too
Coates M, McKillup SC (1995) Role of recruitment and growth
strong conclusions from this study with respect to
in determining the upper limit of distribution of the
these factors. It would, nonetheless, be interesting to intertidal barnacle Hexaminus popeiana. Marine and
investigate how such aspects as planting density, Freshwater Research 46:1065–1070
123
538 Wetlands Ecol Manage (2006) 14:527–538
¨ ¨
Crona BI, Ronnback P (2005) Utilization of replanted man- Kairo JG, Dahdouh-Guebas F, Bosire J, Koedam N (2001)
groves as nursery ground for shrimp communitites in Gazi Restoration and management of mangrove systems – a
Bay, Kenya. Estuarine Coastal Shelf Sci 65:535–544 lesson for and from the East African region. South Afr J
Davey A, Woelkerling WmJ (1985) Studies on Australian Bot 67(3):383–389
mangrove algae. III. Victorian communities: structure and Kitheka JU (1997) Coastal tidally-driven circulation and the
recolonization in western Port Bay. J Exp Mar Biol Ecol role of water exchange in the linkage between tropical
85:177–190 coastal ecosystems. Estuarine Coastal Shelf Sci 45:177–
Davis AR, Butler AJ (1989) Direct observations of larval 187
´
dispersal in the colonial ascidian Podoclavella molucc- Linden O, Lundin CG (1996) Integrated coastal zone man-
ensis Sluiter: evidence for closed populations. Exp Mar agement in Tanzania. World Bank, Environmental
Biol 127:189–203 Department and Sida, Department of Research Coopera-
de Troch M, Mees J, Wakwabi E (1998) Diets of abundant tion/SAREC, Sweden, pp 166
fishes from beach seine catches in seagrass beds of a Maldonado M, Young C (1996) Effects of physical factors on
tropical bay (Gazi Bay, Kenya). Belgian J Zool larval behavior, settlement and recruitment of four tropi-
128(2):135–154 cal demosponges. Mar Ecol Progress Ser 138:169–180
Dunlap M, Pawlik JR (1996) Video-monitored predation by Mwashote BM, Jumba IO (2002) Quantitative aspects of
Caribbean reef fishes on an array of mangrove and reef nutrient fluxes in the Gazi Bay (Kenya): implications for
sponges. Mar Biol 126:117–123 coastal ecosystems. Mar Pollut Bull 44(11):1194–1205
¨
Ellison AM, Farnsworth EJ (1990) The ecology of Belizean Oliveira EC, Osterlund K, Mtolera MSP (2005) Marine plants
¨
mangrove-root fouling communities: I. Epibenthic fauna of Tanzania. In: Sporrong N, Bjork M (eds) A field guide
are barriers to isopod attack of red mangrove roots. J Exp to the seaweeds and seagrasses. Botany Dept., Stockholm
Mar Biol Ecol 142:91–104 University, Sweden, pp 267
Ellison AM, Farnsworth EJ (1992) The ecology of Belizean Osman RW (1977) The establishment and development of a
mangrove-root fouling communities: patterns of distribu- marine epifaunal community. Ecol Monogr 47(1):37–63
tion and abundance and effects on root growth. Hydro- Proches S Marshall K, Ugrasen K, Ramcharan A (2001)
¸,
biologia 247:87–98 Mangrove pneumatophore arthropod assemblages and
Ellison AM, Farnsworth EJ, Twilley RR (1996) Facultative temporal patterns. J Mar Biol Assoc UK 81:545–552
mutualism between red mangrove and root-fouling spon- Randall JE, Hartman WD (1968) Sponge-feeding fishes of the
ges in Belizean mangal. Ecology 77(8):2431–2444 West Indies. Mar Biol 1:216–225
Ellison AM (2000) Mangrove Restoration: Do we know en- Rice WR (1989) Analyzing tables of statistical sets. Evolution
ough? Restor Ecol 8(3):219–229 43:223–225
Eston VR, Braga MR, Cordeiro-Marino M, Fujii MT, Yokoya Richmond MD (ed) (1997) A guide to the seashores of East
NS (1992) Macroalgal colonization patterns on artificial Africa and the Western Indian Ocean Islands. Sida-
substrates inside southeastern Brazilian mangroves. Aquat SAREC Stockholm, pp 448
Bot 42(4):315–325 Rodrigues C, Stoner AW (1990) The epiphyte community of
Farnsworth EJ, Ellison AM (1996) Scale-dependent and tem- mangrove roots in a tropical estuary: distribution and
poral variability in biogeography of mangrove root epi- biomass. Aquat Bot 36:117–126
biont communities. Ecol Monogr 66(1):45–66 Satumanatpan S, Keough M, Watson GF (1999) Role of set-
Field CD (eds) (1996) Restoration of mangrove ecosystems. tlement in determining the distribution and abundance of
International Society for Mangrove Ecosystems, barnacles in a temperate mangrove forest. J Exp Mar Biol
Okinawa, Japan, pp 250 Ecol 241:45–66
Field JG, Clarke KR, Warwick RM (1982) A practical strategy Satumanatpan S, Keough M (2001) Roles of larval supply and
for analyzing multispecies distribution patterns. Mar Ecol behavior in determining settlement of barnacles in a
Progress Ser 8:37–52 temperate mangrove forest. J Exp Mar Biol Ecol 260:133–
Grosberg RK (1982) Intertidal zonation of barnacles: the 153
influence of planktonic zonation of larvae on vertical Steneck RS, Dethier LE (1994) A functional group approach to
distribution of adults. Ecology 63:894–899 the structure of algal-dominated communities. Oikos
Imbert D, Rousteau A, Scherrer P (2000) Ecology of mangrove 69:476–498
growth and recovery in Lesser Antilles: State of knowl- Sutherland JP (1980) Dynamics of the epibenthic community
´
edge and basis for restoration projects. Restor Ecol on roots of the mangrove Rhizophora mangle, at Bahıa de
8(3):230–236 Buche, Venezuela. Mar Biol 58:75–84
Jaasund E (1976) A field guide to the intertidal seaweeds in Warwick RM, Clarke KR, Suharsono (1990) A statistical
¨
Tanzania, 1st ed. University of Tromso, Norway analysis of coral community responses to the 1982–3
˜
Kairo JG (1995) Community participatory forestry for El Nino in the Thousand Islands, Indonesia. Coral Reefs
the rehabilitation of deforested mangrove areas of Gazi 8:171–179
Bay Kenya. A first approach. Final technical report. Watson JG (1928) Mangrove forests of the Malay Peninsula,
BSP/WWF-(USA) and University of Nairobi (Botany vol. 6. Malay Forest Records, 275 pp
Department)
123
DOI 10.1007/s11273-006-9005-7
ORIGINAL PAPER
Re-establishment of epibiotic communities in reforested
mangroves of Gazi Bay, Kenya
B. I. Crona Æ S. Holmgren Æ P. Ronnback
¨ ¨
Received: 6 December 2004 / Accepted: 17 February 2006 / Published online: 28 July 2006
Ó Springer Science+Business Media B.V. 2006
Abstract Recolonization of epibiotic flora and community composition between landward and sea-
fauna in two fringing Sonneratia alba reforestation ward zones were observed in all sites and trunk
plots was investigated and compared to a natural fouling fauna was distinctly different between sites.
mangrove stand and a denuded site in Gazi Bay, Reasons for the above patterns are discussed and it is
Kenya. The reforested sites differed with respect to suggested that zonation patterns affecting pneumato-
land history and planting density. Habitat availability phore surface and inundation time, in combination
in the form of pneumatophore surface differed among with proximity of sites to natural seeding areas, are
forested sites (P<0.001), and between landward and the most likely explanations for observed patterns of
seaward zones (P<0.05). Eighteen algal species were epibiotic community distribution in this study.
found in the natural area compared to 23 and 10 in
replanted sites. Only one species was encountered in Keywords Sonneratia alba Æ Algae Æ Epibiotic
the denuded area. SIMPER analysis distinguished communities Æ Recolonization Æ Replanted
Enteromorpha ramulosa, Polysiphonia sp., Hypnea mangroves Æ Sponges
sp. and Caloglossa leprieuri as the main algal species
responsible for differences between sites. Algal bio-
mass was positively correlated to pneumatophores
Introduction
area (P<0.001). Total algal biomass differed mark-
edly between forested sites: 1.4 (matrix replantation),
In response to the increasing decimation of man-
28.6 (natural stand) and 44.3 g mÀ2 (integrated
groves forests several re- and afforestation programs
replantation) in the seaward zones. The matrix
have been initiated world wide (Imbert et al. 2000;
replantation showed strong differences in algal com-
Kairo et al. 2001) and the need for mangrove reha-
munity assemblages compared to the other forested
bilitation has become recognized as a high priority in
sites, and this site also had significantly lower bio-
local coastal management plans for many developing
mass of sessile benthic fauna (P<0.001). Statistical
´
countries (Linden and Lundin 1996). Earlier man-
differences in algal (P<0.01) and sponge (P<0.05)
grove restoration goals have ranged from supply of
quality wood for logging to shore-line stabilization,
often overlooking the role of mangroves as habitats
B. I. Crona (&) Æ S. Holmgren Æ P. Ronnback
¨ ¨
for a diverse flora and fauna (Field 1996). Still today
Department of Systems Ecology, University of Stockholm,
the majority of rehabilitation programs focus mainly
S-106 91, Stockholm, Sweden
on structural aspects of reforestation (Ellison 2000),
e-mail: beatrice@system.ecology.su.se
123
528 Wetlands Ecol Manage (2006) 14:527–538
baseline information needed as the interest for man-
thus neglecting important issues such as the
grove rehabilitation increases world wide.
recolonization of associated flora and fauna, which is
essential if the ecological functions of a replanted
mangrove forest are to be restored.
The role of epibiotic communities of both flora and Methods
fauna in mangroves has received only limited scientific
attention to date. Ellison and Farnsworth (1990; 1992) Study area
and Ellison et al. (1996) demonstrated the role of root
fouling communities, both directly and indirectly, on Gazi Bay is located on the southern Kenyan coast at
the growth of mangrove roots. The fouling community 4°25¢ S and 39°50¢ E. The inner estuary is sheltered
effectively prevents isopod colonization which can from intense wave impact by shallow reefs at the
otherwise reduce root growth by more than 50% and mouth of the bay (Fig. 1). Seasonal monsoons dom-
facultative mutualism has been reported to occur be- inate the climate with two pronounced rainy seasons;
tween massive sponges and mangroves thus potentially a period of heavy rains from April–June (SE mon-
enhancing mangrove productivity (Ellison et al. soon) and a period of lighter rains from October–
1996). Proches et al. (2001) also showed that pneu- November (NE monsoon). Total annual rainfall
matophores and their associated epibiota provide a ranges from 1000 to 1600 mm and the salinity in the
unique structural feature to the physical environment study area ranges from 24 to 26.5 ppt during the SE
of mudflats resulting in increased microhabitat com- monsoon (Kitheka 1997). High flushing rates are
plexity and giving rise to arthropod assemblages dis- coupled with low water residence times (3 to 4 h) and
tinctly different from those of surrounding sediments. result in high rates of exchange between inshore and
Sponges and filamentous algae may constitute offshore waters (60–90% of the volume per tidal
food for fish foraging in the mangroves at high tide. cycle) (Kitheka 1997).
Several species of angelfish of the genus Pomacan-
thus, butterflyfish (Chaetodon) and some filefish Site description
(Cantherines) have a large sponge component in their
diet (Randall and Hartman 1968). Although adults of Four sites were investigated between March and June
these species are mainly reef-associated, some have 2002; one natural stand of Sonneratia alba (N) and
been documented to enter mangroves (de Troch et al. two replanted sites of the same species (MP and IP)
1998). Many juveniles of reef-bound species, such as of equal age (8 years) but differing planting density,
Tetradontidae, Ostraciidae and Chaetodontidae occur and one denuded site (D) included for comparison.
in mangroves and are known to feed on filamentous Site IP (Integrated Plantation) was replanted in a
algae and benthic invertebrates (Randall and Hartman degraded stand of S. alba with a planting density of
approximately 4330 trees haÀ1 (Kairo 1995). This
1968; Dunlap and Pawlik 1996; de Troch et al. 1998).
Substrate availability as a limiting resource is a site thus contained a certain amount of forest cover at
factor affecting most epibiotic communities (Osman the time of planting which distinguishes it from MP.
1977) in combination with variable complexity in Site MP (Matrix Plantation) was planted on denuded
habitat structure attributable to absolute abundance of ground which had been clear-felled under commer-
individual structural components such as pneumato- cial logging practices in the 1970’s (Kairo 1995) and
phores (Beck 2000). Since little information has been where no natural regeneration of mangroves had oc-
published on the return of ecosystem functions in curred. This site was planted in a 1.0·1.0 m matrix
rehabilitated mangroves the objective of this study with a planting density of approximately 10,250 trees
haÀ1 (Kairo 1995). Since their establishment pruning
was to investigate recolonization of the epibiotic flora
and fauna on pneumatophores and tree trunks in re- has been carried out twice but no thinning has been
planted mangroves, Sonneratia alba, in Gazi Bay, done in either of the plantations (Kairo personal
Kenya, and compare the findings with epibiotic communication). Site (D), adjacent to MP, was also
communities in adjacent natural mangroves and a logged in the 1970’s. Each site was subdivided into
denuded stand (currently a sandflat). This is valuable two zones, a landward zone (ZI) and a seaward zone
123
Wetlands Ecol Manage (2006) 14:527–538 529
Fig. 1 Map of the study
area, Gazi Bay, Kenya. The
area is located on the
southern Kenyan coast at
4°25¢ S and 39°50¢ E.
N=natural stand,
IP=integrated plantation,
MP=matrix plantation and
D=denuded site
(ZII) based on visual observations of a natural Relevant site characteristics are summarized in
epifaunal gradient and inundation patterns. Zones Table 1.
were approximately 25 meters deep and elevation
differences between them ranged between 0.61 to Sampling design
0.96 meters resulting in a slope around 2°, resulting
in differences in inundation, between zones, of Sampling was done using 0.5·0.5 m wood frames.
approximately 60 min per tide. All sites were lo- Ten replicates were taken in each zone and were
cated along a tidal channel (Fig. 1) and belong to selected by randomly paired coordinates, giving a
inundation class I as described by Watson (1928), total of 80 samples for all four sites. A sub-sample of
with inundation approx 12 h dayÀ1. The size of the roots in each frame were measured and counted. The
sites ranged from 1,715 m2 (MP) to 10,845 m2 (N). base diameter and height of each pneumatophore was
123
530 Wetlands Ecol Manage (2006) 14:527–538
Table 1 Summary of features (range or mean – SE) characterizing the Sonneratia alba sites studied in Gazi Bay, Kenya
Site Natural (N) Integrated plantation (IP) Matrix plantation (MP) Denuded (D)
Feature
Site area (m2) 10,845 7,931 1,715 3,741
Planting densitya (# haÀ1) – 4330 10,250 –
Current stand densityb (# haÀ1) 4300–1221 – 7640–600 –
Energy exposure High High Low Low
Relative site elevation (m) 0.14 0 0.37 0.37
Sediment typec ZI: mud ZI: mud ZI: muddy sand ZI: sand
ZII: muddy sand ZII: mud ZII: mud ZII: muddy sand
Sediment organic contentc 7.9–1.6 14.3–1.1 9.2–1.8 1.6–0.3
Pore water salinity* ZI: 34.6–0.3 ZI: 34.4–0.2 ZI: 34.0–0.8 ZI: 35.9–1.1
ZII: 34.5–0.3 ZII: 34.2–0.1 ZII: 34.6–0.2 ZII: 27.4–5.8
Water column salinity 36.5–38.0 36.5–38.0 36.0–37.5 36.0–37.5
Canopy cover 50–75% 50–75% 100% 0%
Pneumatophore density (# mÀ2)* ZI: 174–21 ZI: 424–24 ZI: 380–41 –
ZII: 280–37 ZII: 322–35 ZII: 400–26
Pneumatophore area (m2 per m2 forest)* ZI: 0.30–0.03 ZI: 0.41–0.06 ZI: 0.16–0.01 –
ZII: 0.55–0.07 ZII: 0.55–0.6 ZII: 0.24–0.03
Average trunk area (cm2 per m2 forest)* ZI: 1.2–0.2 ZI: 0.4–0.1 ZI: 1.3–0.4 –
ZII: 0.7–0.2 ZII: 0.8–0.1 ZII: 1.0–0.1
Trunk diameter range (cm)* Min: 4.3 Min: 2.5 Min: 1.2 –
Max: 93.5 Max: 48.0 Max: 8.6
a b c
¨ ¨
Data from Kairo (1995), Bosire et al. (2003), Crona and Ronnback (2005). ZI=landward zone, ZII=seaward zone, *n=10
determined and used to calculate root surface area for colonization, from the sediment surface and one
available for epifaunal colonization. For ease of meter up (Table 1). All sessile organisms found
surface computation pneumatophores were treated as growing on the lowest meter of each tree trunk were
perfect cones. Pore water samples were randomly counted and determined to genus, and species level
collected in each site by digging a hole in the sedi- wherever possible.
ment 10–15 cm deep, and poor water and water
column salinity was measured using an optical Statistical analysis
refractometer (Atago brand). All epibiota on both
roots and sediment within the frame was removed and Species abundance, biomass and frequency data were
dried at 60°C to constant weight. Sessile fauna double square root transformed and subjected to non-
growing on the roots was determined to the taxo- metric multi-dimensional scaling ordination (nMDS)
nomic level of phyla. Although a detailed account of using the Bray–Curtis similarity coefficient (Field
the diversity of the sponges and ascidians would have et al. 1982). The nMDS is a method based on a non-
been preferable, the poorly documented taxonomy of parametric regression of distance on dissimilarity of
these organisms in East African waters made this samples. The goodness-of-fit of the regression line is
difficult. As a result they were grouped into respec- evaluated by calculating a stress value which is a
tive phyla both representing the functional group of measure of how well the MDS succeeded in fitting the
sessile filter feeders on pneumatophores. The poten- multidimensional data onto a 2-dimesional plane. An
tial benefits of such grouping when studying stability alternative assessment of the dimensionality of the
and persistence of marine benthic communities has dataset is presented by non-parametrically correlating
been reviewed by Steneck and Dethier (1994). Algae the original sample similarity matrix with that repre-
were determined to genus and species wherever senting distance between samples in ordination space.
possible (Jaasund 1976; Richmond 1997). For each Biomass data for both flora and fauna was tested for
site ten random trees were also selected in each zone significant differences using analysis of similarity
(a total of 60 trees) and the circumference was randomization tests (ANOSIM, Clarke and Green
measured and used to calculate trunk area available 1988; Clarke and Warwick 2001). Algal species
123
Wetlands Ecol Manage (2006) 14:527–538 531
Algal biomass and total epibiotic biomass on
responsible for differences in sites observed in nMDS
roots and sediment
plots were identified with a dissimilarity percentage
program (SIMPER, Warwick et al. 1990). Data on
For all sites, algal biomass and combined epibiotic
root complexity and trunk area were analysed using
biomass (algae, sponges and ascidians) were mea-
nested, one-way ANOVA. When conditions for use of
sured. Biomass was consistantly higher in the seaward
parametric statistics were not met Kruskal–Wallis
zone (ZII) for all sites and the natural (N) and
and Mann–Whitney U tests as well as Spearman rank
replanted site (IP) had the highest total biomass of
correlation were applied. For multiple comparisons
algae and sessile fauna (Table 2). The replanted site
significance levels were adjusted using the Bonferroni
(MP) had lower algal biomass in the seaward zone
method (Rice 1989). All statistical analysis
(ZII) than the adjacent denuded plot which was a re-
were performed using Primer 5 (version 5.2.1) or
sult of the high presence of Enteromorpha ramulosa in
Statistica 6.0.
the cleared area, a green algae favoured by intense
sunlight and growing directly on the sediment. nMDS
Results ordination revealed a high similarity between sites IP
and N for both algal and total epibiotic biomass
Surface available for colonization (Fig. 2a, b) expressed by the small distance between
sites in the plot. Sites D and MP differed markedly
Results from a hierarchically nested ANOVA (zones from all sites except for combined epibiotic biomass
nested in sites) performed on number of pneumato- where sites MP and D showed slightly higher resem-
phores and total pneumatophore area (Table 1) reveal blance (Fig. 2a, b). This similarity was due, in part, to
significant differences between sites (P<0.001) and a common lack of poriferans and ascidians among MP
zones (P<0.05), which supports the sub-division of and D samples. The low stress of the nMDS plot
each site into zones in the analysis of biological data (<0.05) suggests a very good fit of the multidimen-
that follows. sional data onto the 2-dimensional plot (Clarke 1993).
Average trunk area mÀ2 available for colonization The dimensionality of the datasets for both algae and
by sessile fauna was compared for sites N, IP and MP total biomass was further assessed through non-para-
(Table 1). Results of a one way hierarchically nested metric Spearman correlation between distance in the
ANOVA (zones nested in sites) show a highly sig- ordination space and distance in the original
nificant difference in available trunk area between p-dimensional space (ralgae=1.0, rtot bio=1.0) for both
sites (P<0.001). nMDS plots presented in Fig. 2. To test the relation-
ship between algal biomass and pneumatophore area a
Species richness of epibiotic communities Spearman rank order correlation was conducted,
resulting in a positive relationship (ralgae=0.55,
The algal species encountered in each zone and site P<0.001). A similar correlation was also run to test
are shown in Table 2. Site D had only one species of differences between zones in terms of biomass of
algae, Enteromorpha ramulosa. This algae was found algae and sponges (ralgae=0.39, P<0.01; rsponge=0.22,
growing both on hard substrate and freely on the P=0.046).
sediment surface. Site MP had a total of ten algal Analysis of similarities (ANOSIM) between sites
species, all of which were found in zone II, while for algal biomass and combined epibiotic biomass
only four species were found in zone I. In sites IP and (R=0.422 and R=0.407 respectively, P<0.001) indi-
N a total of 23 and 18 species were found respec- cates dissimilarity between sites. Furthermore, the
tively. For all forested sites the algae found belong to results reveal a high degree of dissimilarity between
classes Rhodophyta and Chlorophyta. Several dif- all pair wise between-site comparisons (P<0.001)
ferent taxa of both ascidians and poriferans were except for sites IP and N which were similar for algal
found growing on the pneumatophores of Sonneratia biomass but not for average combined epibiotic bio-
alba stands in Gazi Bay. Tedania digitata vulcanis, mass (P=0.015). To explore why, Mann–Whitney U
which could be positively identified to species, tests were used to test differences in average sponge
dominated the poriferan community. and algal biomass between and within the natural
123
Table 2 List of organisms encountered at each studied site of Sonneratia alba in Gazi Bay, Kenya
532
Zone I Zone II
123
N IP MP D N IP MP D
Biomass of organisms on pneumatophores
Rhodophyta (Algae)
Catenella nipae 1.8 (1.8) 6.2 (5.4) – – 329 (217) 19.9 (19.9) 46.8 (23.3) –
Ceramium brevizonatum var. caraibicum 41.0 (23.4) 68.0 (20.7) – – 10.4 (8.3) 29.2 (24.9) – –
Ceramium sp. – 9.8 (9.8) – – – – – –
Coelothrix irregularis 0.8 (0.8) – – – – 2.1 (2.1) – –
Gelidiopsis intricata – 0.3 (0.3) – – – 2.7 (2.7) – –
Gracilaria crassa – – – – – 6.4 (6.4) – –
Gracilaria salicornia 321 (300) 352. (266) – – 1503 (821) 3294 (1537) – –
Heterosiphonia sp. – – – – 0.9 (0.9) – – –
Hypnea sp. 5.6 (3.5) 107 (70.1) – – 325 (190) 160 (50.2) – –
Laurencia perforate – 5.0 (5.0) – – 12.8 (12.1) – – –
Polysiphonia sp. 180 (71.9) 177 (173) – – 451 (284) 675 (432) 80.6 (35.2) –
Unidentified red algae. T1 – 30.5 (26.9) – – 36.6 (32.6) – – –
Unidentified red algae. T2 – 1.2 (1.2) – – – – – –
Phaeophyta (Algae)
Bostrychia binderi 0.6 (0.6) – – – – – – –
Bostrychia radicans – 20.1 (11.2) – – 38.6 (21.5) 5.9 (4.2) 2.6 (2.6) –
Caloglossa leprieuri 90.1 (48.1) 139 (35.7) 1.7 (1.7) – 140 (47.1) 140 (44.3) 7.5 (7.4) –
Endosiphonia clavigera – – – – – 12.4 (12.4) – –
Chlorophyta (Algae)
Chaetomorpha crassa – 5.8 (4.3) – – – 2.0 (1.3) – –
Cladophora patentiranea 5.2 (4.2) 9.6 (6.2) 3.5 (2.4) – 10.9 (8.4) 2.8 (2.8) 1.6 (1.6) –
Enteromorpha kylinii – 1.3 (0.9) – – 0.7 (0.7) – – –
Enteromorpha ramulosa 0.4 (0.4) – – – – – – 688 (407)
Ulva pertusa – – – – – 73.9 (49.9) – –
Unidentified green algae – 0.7 (0.7) – – – – –
Cyanophyta unidentified 12.4 (8.4) 1.5 (1.5) – – – – 27.2 (15.1) –
Diatoms – 12.7 (12.7) – – 64.9 (63.5) 16.3 (14.7) – –
Hydrozoans 21.6 (21.6) 14.3 (12.4) – – 39.2 (23.4) 77.0 (53.1) 0.8 (0.8) –
Total algal biomass 681.0 961.9 5.2 0 2963 4519 167.0 688.8
Chordata
Ascidians 0.26 (0.2) 0.4 (0.2) – – – 1.5 (0.7) – –
Porifera
Poriferans 20.7 (8.3) 20.0 (9.1) – – 43.8 (9.1) 75.9 (14.7) – –
Total combined biomass 21.65 21.32 0.005 0 46.73 81.88 0.169 0.689
Wetlands Ecol Manage (2006) 14:527–538
Wetlands Ecol Manage (2006) 14:527–538 533
Mean (SE) dry weight (mg m ) for algae, (g m ) for ascidians and poriferans on pneumatophores (n=10). Mean (SE) abundance (# m trunk area) of trunk epifauna (n=10).
stand and integrated plantation. Tests between sites
proved non-significant (psponge=0.516; palgae=0.250)
while Zone I and II of the integrated plantation dif-
D
–
–
–
–
–
–
–
–
–
fered significantly (psponge=0.008; palgae=0.041) and
Zone I and II of the natural stand showed somewhat
389 (162)
547 (198)
weaker trends (psponge=0.059; palgae=0.049).
179 (50)
62 (19)
Analysis of similarity percentages (SIMPER) of
17 (7)
2 (2)
2 (2)
MP
algal species distributions was done to reveal the
–
–
main species responsible for observed dissimilarities
among sites (Table 3). Enteromorpha ramulosa was
the sole species responsible for characterizing site D
À2
Total combined biomass (algae, ascidians and poriferans) expressed in g mÀ2Natural stand (N), replanted (IP and MP) and denuded (D)
40 (12)
as it was the only species of algae found in this
2 (2)
2 (2)
4 (3)
4 (3)
2 (2)
habitat. Consequently it also played a dominant role
IP
–
–
–
in distinguishing this site from all other areas sam-
pled. Two species of algae contributed to over 80%
of total algal biomass in site MP as compared to four
436 (247)
574 (215)
109 (76)
Zone II
53 (45)
45 (19)
and three species in the natural stand and the inte-
4 (3)
2 (2)
grated plantation, respectively (Table 2). Polysipho-
N
–
–
nia sp. was responsible for over 50% of the biomass
in the matrix plantation, but was also present as a
D
–
–
–
–
–
–
–
–
–
potential characterizing species in both the natural
stand and integrated plantation. Three out of four
species characterizing sites N and IP were the same
3.3 (1.1)a
125 (44)
115 (70)
11 (11)
02 (2)
despite differences in their percentage contribution.
2 (2)
MP
–
–
–
Epibiotic communities of trunks
A total of nine animal species from three phyla were
17 (17)
recorded (Table 2). No sponges, ascidians or algae
9 (8)
2 (2)
2 (2)
IP
were found on the trunks and the majority of speci-
–
–
–
–
–
mens recorded were found around the upper limits of
the 1 m range of trunk area investigated. There is a
404 (139)
clear separation between sites although some inter-
132 (50)
28 (28)
Zone I
spersion of samples from site N and IP is found
(2)
(3)
(3)
(3)
4 (3)
N
(Fig. 3). No correlation was found between the
2
4
4
6
–
À2
number of species of sessile fauna present and
available trunk area. Correlation between trunk area
and abundance of the three most common fouling
Abundance of organisms on tree trunks
species was done revealing no correlation for Chi-
rona tenuis, a positive correlation for Chthalamus
À2
dentatus (r=0.29, P<0.05) and a negative correlation
Clypeomorus bifasciatus
for Balanus amphitrite (r=0.35, P<0.01).
Brachidontes variabilis
Patellid (unidentified)
Chthalamus dentatus
Cerithidea decollata
Terebralia palustris
Cirripedia (Crustacea)
Balanus amphitrite
Discussion
Table 2 Continued
Chirona tenuis
Patterns of biomass distribution
Nerita sp.
Gastropoda
Bivalvia
When examining community composition of the
investigated sites in terms of algal biomass two
123
534 Wetlands Ecol Manage (2006) 14:527–538
Fig. 2 (a) nMDS plots of total algal biomass per mangrove site tion, D=denuded site, I=landward zone, II=seaward zone. For
and zone in Gazi Bay, Kenya. Each symbol represents the site D, ZI contained no epibiotic biomass and consequently was
average value of that zone (n=10). Stress=0 (b) nMDS plots of not included in the MDS analysis. For both nMDS plots
combined epibiotic biomass per site and zone. Stress=0 distances between points are represented as equal distances in all
N=natural stand, IP=integrated plantation, MP=matrix planta- dimensions of the plot
Table 3 Algal species responsible for similarities within and dissimilarities among sites of Sonneratia alba mangroves in Gazi Bay,
Kenya
di/SD(di)
Site Species Contrib % Cum contrib % Avg sim
Species responsible for observed similarity within sites
1.64a
IP Caloglossa leprieuri 28.60 28.60 40.61
1.22a
Hypnea sp. 24.80 53.40
0.85a
N Polysiphonia sp. 45.36 45.36 36.13
0.84a
Caloglossa leprieuri 26.85 72.21
MP Polysiphonia sp. 0.47 51.45 51.45 26.03
0.36a
Catenella nipae 30.83 82.28
D Enteromorpha ramulosa ##### 100.00 100.00 100.00
Sites Species Dissim/SD Contrib % Cum contrib % Avg dissim
Species responsible for observed dissimilarity between sites
6.87b
D, MP Enteromorpha ramulosa 45.00 45.00 100.00
3.87b
D, IP Enteromorpha ramulosa 26.51 26.51 100.00
1.92b
Caloglossa leprieuri 12.38 38.88
1.17b
Hypnea sp. 12.27 51.16
1.22b
MP, IP Polysiphonia sp. 13.66 13.66 84.77
1.19b
Hypnea sp. 13.53 27.18
1.59b
Caloglossa leprieuri 13.20 40.38
3.25b
D, N Enteromorpha ramulosa 31.63 31.63 99.37
1.21b
Polysiphonia sp. 19.05 50.69
1.06b
Caloglossa leprieuri 13.20 63.89
1.07b
MP, N Polysiphonia sp. 19.53 19.53 75.50
1.07b
Caloglossa leprieuri 16.47 36.01
1.17b
IP, N Polysiphonia sp. 14.19 14.19 63.58
1.01b
Gracilaria salicornia 13.59 27.77
1.07b
Hypnea sp. 11.86 39.63
1.10b
Caloglossa leprieuri 10.88 50.51
1.03b
Ceramium brevizonatum var caraibicum 9.62 60.13
a
indicates that the species potentially characterizes the species assemblage of a site and b indicates a possible discriminating species
between sites. Analysis is based on algal biomass mÀ2
Natural (N), replanted (IP and MP) and denuded (D)
123
Wetlands Ecol Manage (2006) 14:527–538 535
species present in the former was E. ramulosa, a
green algae thriving in sun exposed environments
(Oliveira et al. 2005) and consequently lacking in the
adjacent shaded habitat.
If biomass of sponges and ascidians is included in
the nMDS analysis the pattern and spatial relationship
between sites change (Fig. 2b). While sites IP and N
remain tightly clustered site MP is now found posi-
tioned far from N, IP and D. The reason for this is the
lack of sponges and ascidians in all but one sample
from site MP. Since both of these organism groups
require hard substrate for their growth and survival
none were present in site D. Studies of sponge
assemblages suggest a number of factors responsible
for species distributions including water flow rate
(Maldonado and Young 1996), sedimentation and
nutrient levels (Bell and Barnes 2003), depth (Alv-
arez et al. 1990), light (Cheshire and Wilkinson
Fig. 3 nMDS of epibiotic trunk fauna based on species 1991), and habitat availability (Barthel and Tendal
abundance per standardized available mangrove trunk area 1993). Sedimentation and nutrient levels were not
(m2) in Gazi Bay, Kenya. N=natural stand (squares), IP=inte-
investigated in this study but are assumed to be
grated plantation (circles), MP=matrix plantation (triangles).
similar due to the close proximity and fringing po-
Stress=0.16
sition of all sites in the bay as well as the thorough
mixing of the waters entering the bay (Kitheka 1997;
distinct patterns emerge (Fig. 2a). For algal biomass
Mwashote and Jumba 2002). Lower root surface
there is a clear gradient with algal biomass of the
availability was significant in site MP, but the lack of
denuded site being markedly lower than replanted
pneumatophore surface may not provide a sole sat-
site IP and the natural site (N), while replanted site
isfactory explanation. Several investigations on fac-
MP has intermediate biomass values. This is also
tors affecting distribution and composition of
supported by a positive correlation between algal
mangrove epibiotic communities have concluded that
biomass and total pneumatophore area. Pneumato-
larval behaviour and longevity may affect recruit-
phore area in the matrix plantation (MP) is half of
ment patterns (e.g. Sutherland 1980; Bingham and
that in both the replanted site IP and the natural stand
Young 1995). New colonists seem to recruit primar-
(N) (Table 1). Although no causal relation can be
ily from local sources and sponge colonization has
determined with certainty the lower root area is
been suggested to occur from root to root (Farnsworth
surely important in determining the colonization
and Ellison 1996) or through fragmentation of adult
efficiency and recruitment possibilities of algae. If
colonies (Bingham and Young 1995). Ascidian larvae
algal recruitment follows the lottery model proposed
are known to swim short distances (2–15 m) (Davis
for epibiota (Sutherland 1980), lower pneumatophore
and Butler 1989) and many have developed swim-
area available would certainly reduce chance
ming behaviours promoting settlement in close
recruitment of algal spores from the water column. A
proximity to parental colonies (Bingham and Young
higher level of canopy cover observed in the matrix
1991). The location of site MP up-stream along the
plantation as opposed to the natural and integrated
creek from sites N and IP may thus be another reason
stands, and consequently lower solar radiation in the
why larval recruitment or entrapment of colonial
former plot, could also affect the abundance, biomass
fragments has not occurred.
and composition of algal species present. This is
Maldonado and Young (1996) showed that distri-
likely the main reason for observed higher algal
bution of four tropical demosponges, including
biomass values in the seaward zone of the cleared
Tedania ignis, a close relative to the dominant sponge
compared to the replanted area (MP), since the only
123
536 Wetlands Ecol Manage (2006) 14:527–538
in extreme ecological conditions (Oliveira et al.
in this study (Tedania digitata vulcani), was likely
2005). Average dissimilarity values between sites
affected by post-settlement mortality of larvae due to
reinforce the pattern and provide further evidence
suboptimal habitat conditions. This resulted in adult
that site MP has a distinctly different community
populations strongly associated with both high and
composition in terms of epiphytic algae as compared
low irradiance, but always with a high level of water
to the other forested sites (Table 3). Davey and
movement. Since sites IP and N have a higher wave
Woelkerling (1985) observed fluctuating frequency
energy exposure and presumable higher levels of
of colonizing red algae over time and related this to
water movement this may explain their abundant
grazing, competition for substratum and sloughing
sponge communities compared to site MP. Differ-
off of pneumatophore bark. De Troch et al. (1998)
ences in canopy cover may thus play a subordinate
reported that between 6 and 7% of the fish commu-
role in determining the presence of sponges.
nity sampled in Gazi Bay consist of herbivores, but
Observed differences in combined epibiotic bio-
effects of grazing on newly established algae in the
mass at the zone level for IP and N was a result of
area is unknown. Hence it is difficult to speculate on
much higher total biomass values for sponges and
the importance of herbivory to the lack of algal
algae found in zone II of IP as compared to zone II of
coverage in site MP. It is possible that the less diverse
N (Table 2). Correlation values also support this and
algal community in site MP is stable and may not
suggest that the main difference in terms of both
reach the higher species diversity of N and IP. Sim-
sponge and algal biomass is at the zone level, linked
ilarly altered community assemblages were observed
to inundation, rather than differences in vegetation
after disturbance by Davey and Woelkerling (1985)
structure and habitat availability between the natural
and in colonization experiments by Eston et al.
and replanted site.
(1992). Another explanation is that since replanting
in plot IP was integrated with sparse but existing
Algal species assemblages
forest cover the existence of epibiotic communities
on remaining root structures at the time of planting
A few studies have described mangrove associated
may have facilitated the ease and speed of recolon-
algal communities (Burkholder and Almodovar 1974;
ization. Closer proximity of site IP to the natural site
Beanland and Woelkerling 1983; Davey and Woel-
(N) compared with site MP is also plausible, although
kerling 1985; Rodriguez and Stoner 1990) with focus
a difference in proximity of approximately 500 me-
on algal communities in Rhizophora and Avicennia
ters seems unlikely to account entirely for the low
stands. To our knowledge none have dealt with epi-
diversity and recruitment after eight years.
biotic assemblages associated with Sonneratia sp.
The dominating algal species in this study were
Fouling fauna of mangrove tree trunks
Polysiphonia sp., Catenella nipae, Hypnea sp. and
Caloglossa leprieuri. Apart from Hypnea sp., these
species have all been found as dominant components The fouling community of mangrove tree trunks
among the studies mentioned above. Their respective differs entirely from that of pneumatophores. The
dominance within the different sites of this study primary reason behind the distinct difference in
varies however. Relatively low values of average communities is likely a higher tolerance to wave
similarity within all sites are likely due to the marked exposure and dessication by the barnacles dominating
zonation pattern described earlier. Algal diversity of the trunk fouling community. Patterns of species
forested sites was in the upper range of values re- abundance of sessile fauna on tree trunks reveal a
ported previously (Beanland and Woelkerling 1983; clear separation between sites (Fig. 3). Despite the
Davey and Woelkerling 1985; Rodriguez and Stoner lack of correlation between species richness and
1990). However, while replanted site IP and the available trunk area, species assemblages between the
natural stand (N) had a high diversity, replanted site sites apparently differ. Negative correlation for Bal-
MP counted approximately half the number of spe- anus amphitrite appears because of the highly local-
cies and the denuded site grew only one species, ized presence of this species in the matrix plantation
growing freely on the sediment and known to thrive (MP) where trunk diameters are smaller than the
123
Wetlands Ecol Manage (2006) 14:527–538 537
location and pre-planting history of sites may affect
other forested sites (Table 1). Barnacle larvae are
the return of associated flora and fauna such as those
known to exhibit patchy distributions in the water
studied in Gazi Bay.
column (Grosberg 1982) and availability of suitable
substrata, competition from other species and pres-
Acknowledgements This study was funded by the Swedish
ence of sexually mature conspecifics have been sug-
International Development Agency (Sida). The authors thank
gested to influence their settlement and consequent the Kenya Marine and Fisheries Research Institute (KMFRI) for
distribution (Bayliss 1993; Coates and McKillup logistical support. A special thank you goes to J. Bosire and
1995; Satumanatpan and Keough 2001). Bayliss J. Kairo for assistance in Kenya, to Abdul for tireless field
assistance, to Nils and Lena Kautsky for valuable comments and
(1993) observed highly aggregated distributions of
help with taxonomic identification, and to two anonymous
Balanus amphitrite on pneumatophores of Avicennia reviewers for helpful comments on how to improve the manuscript.
marina and suggested that mechanisms responsible
may be settling behaviour of cyprid larvae, post-set- References
tlement mortality, selective predation among others.
Furthermore, Satumanatpan et al. (1999) have shown Alvarez B, Diaz MC, Laughlin RA (1990) The sponge fauna on
a fringing coral reef in Venezuela, I. Composition, dis-
that settlement determines large scale distributions
¨
tribution and abundance. In: Rutzler K (eds) New per-
such as along shore variations in abundance observed
spectives in sponge biology. Smithsonian Institution
in this study while post-settlement mortality, in Press, Washington DC, pp 358–366
combination with settlement cues mentioned above, Barthel D, Tendal OS (1993) The sponge association of the
abyssal Norwegian Greenland Sea: species composition,
is responsible for small scale vertical distribution.
substrate relationships and distributions. Sarsia 78:83–96
This may explain the distinct pattern in barnacle
Bayliss DE (1993) Spatial distribution of Balanus amphitrite
species distribution and composition between studied and Elminius adelaide on mangrove pneumatophores. Mar
sites in Gazi Bay. Biol 116:251–256
Beanland WR, Woelkerling WmJ (1983) Avicennia canopy
effects on mangrove algal communities in Spencer Gulf,
South Australia. Aquat Bot 17:309–313
Beck MW (2000) Separating the elements of habitat structure:
Conclusion
independent effects of habitat complexity and structural
components on rocky intertidal gastropods. J Exp Mar
Results presented indicate that after eight years, Biol Ecol 249:29–49
marked differences in epibiotic biomass exist be- Bell JJ, Barnes KA (2003) Effect of disturbance on assem-
blages:an example using porifera. Biol Bull 205:144–159
tween replanted stands of Sonneratia alba and con-
Bingham BL, Young CM (1991) Larval behaviour of the
sequently also between the replanted and natural
ascidian Ecteinascidia turbinata Herdman; an in situ
stands studied. Differences among replanted sites as experimental study of the effects of swimming on dis-
well as between replanted and natural sites also exist persal. J Exp Mar Biol Ecol 145:189–204
Bingham BL, Young CM (1995) Stochastic events and
with regard to algal species assemblages. As both
dynamics of a mangrove root epifaunal community. Mar
reforestation plots were initiated in the same year the
Ecol 16(2):145–163
effect of time cannot solely account for the discrep- Burkholder PR, Almodovar LR (1974) Studies on mangrove
ancies in epibiotic communities between these two algal communities in Puerto Rico. Florida Sci 36:66–73
Cheshire AC, Wilkinson CR (1991) Modeling the photosyn-
sites nor are the physical parameters investigated
thetic production by sponges on Davis Reef, Great Barrier
significantly different between sites. Instead, clear
Reef. Mar Biol 109:13–18
zonation patterns of pneumatophore surface and Clarke KR (1993) Non-parametric multivariate analyses of
inundation time, in combination with proximity of changes in community structure. Aust J Ecol 18:117–143
Clarke KR, Green RH (1988) Statistical design and analysis for
sites to natural seeding areas, seem more plausible
a ‘biological effects’ study. Mar Ecol Progress Ser
explanations for the observed patterns of epibiotic
46:213–226
community distribution in this study. Due to lack of Clarke KR, Warwick RM (2001) Change in marine commu-
replication at the level of planting strategy and his- nities: An approach to statistical analysis and interpreta-
tion, 2nd edn. Primer-E, Plymouth
tory of sites prior to planting it is risky to draw too
Coates M, McKillup SC (1995) Role of recruitment and growth
strong conclusions from this study with respect to
in determining the upper limit of distribution of the
these factors. It would, nonetheless, be interesting to intertidal barnacle Hexaminus popeiana. Marine and
investigate how such aspects as planting density, Freshwater Research 46:1065–1070
123
538 Wetlands Ecol Manage (2006) 14:527–538
¨ ¨
Crona BI, Ronnback P (2005) Utilization of replanted man- Kairo JG, Dahdouh-Guebas F, Bosire J, Koedam N (2001)
groves as nursery ground for shrimp communitites in Gazi Restoration and management of mangrove systems – a
Bay, Kenya. Estuarine Coastal Shelf Sci 65:535–544 lesson for and from the East African region. South Afr J
Davey A, Woelkerling WmJ (1985) Studies on Australian Bot 67(3):383–389
mangrove algae. III. Victorian communities: structure and Kitheka JU (1997) Coastal tidally-driven circulation and the
recolonization in western Port Bay. J Exp Mar Biol Ecol role of water exchange in the linkage between tropical
85:177–190 coastal ecosystems. Estuarine Coastal Shelf Sci 45:177–
Davis AR, Butler AJ (1989) Direct observations of larval 187
´
dispersal in the colonial ascidian Podoclavella molucc- Linden O, Lundin CG (1996) Integrated coastal zone man-
ensis Sluiter: evidence for closed populations. Exp Mar agement in Tanzania. World Bank, Environmental
Biol 127:189–203 Department and Sida, Department of Research Coopera-
de Troch M, Mees J, Wakwabi E (1998) Diets of abundant tion/SAREC, Sweden, pp 166
fishes from beach seine catches in seagrass beds of a Maldonado M, Young C (1996) Effects of physical factors on
tropical bay (Gazi Bay, Kenya). Belgian J Zool larval behavior, settlement and recruitment of four tropi-
128(2):135–154 cal demosponges. Mar Ecol Progress Ser 138:169–180
Dunlap M, Pawlik JR (1996) Video-monitored predation by Mwashote BM, Jumba IO (2002) Quantitative aspects of
Caribbean reef fishes on an array of mangrove and reef nutrient fluxes in the Gazi Bay (Kenya): implications for
sponges. Mar Biol 126:117–123 coastal ecosystems. Mar Pollut Bull 44(11):1194–1205
¨
Ellison AM, Farnsworth EJ (1990) The ecology of Belizean Oliveira EC, Osterlund K, Mtolera MSP (2005) Marine plants
¨
mangrove-root fouling communities: I. Epibenthic fauna of Tanzania. In: Sporrong N, Bjork M (eds) A field guide
are barriers to isopod attack of red mangrove roots. J Exp to the seaweeds and seagrasses. Botany Dept., Stockholm
Mar Biol Ecol 142:91–104 University, Sweden, pp 267
Ellison AM, Farnsworth EJ (1992) The ecology of Belizean Osman RW (1977) The establishment and development of a
mangrove-root fouling communities: patterns of distribu- marine epifaunal community. Ecol Monogr 47(1):37–63
tion and abundance and effects on root growth. Hydro- Proches S Marshall K, Ugrasen K, Ramcharan A (2001)
¸,
biologia 247:87–98 Mangrove pneumatophore arthropod assemblages and
Ellison AM, Farnsworth EJ, Twilley RR (1996) Facultative temporal patterns. J Mar Biol Assoc UK 81:545–552
mutualism between red mangrove and root-fouling spon- Randall JE, Hartman WD (1968) Sponge-feeding fishes of the
ges in Belizean mangal. Ecology 77(8):2431–2444 West Indies. Mar Biol 1:216–225
Ellison AM (2000) Mangrove Restoration: Do we know en- Rice WR (1989) Analyzing tables of statistical sets. Evolution
ough? Restor Ecol 8(3):219–229 43:223–225
Eston VR, Braga MR, Cordeiro-Marino M, Fujii MT, Yokoya Richmond MD (ed) (1997) A guide to the seashores of East
NS (1992) Macroalgal colonization patterns on artificial Africa and the Western Indian Ocean Islands. Sida-
substrates inside southeastern Brazilian mangroves. Aquat SAREC Stockholm, pp 448
Bot 42(4):315–325 Rodrigues C, Stoner AW (1990) The epiphyte community of
Farnsworth EJ, Ellison AM (1996) Scale-dependent and tem- mangrove roots in a tropical estuary: distribution and
poral variability in biogeography of mangrove root epi- biomass. Aquat Bot 36:117–126
biont communities. Ecol Monogr 66(1):45–66 Satumanatpan S, Keough M, Watson GF (1999) Role of set-
Field CD (eds) (1996) Restoration of mangrove ecosystems. tlement in determining the distribution and abundance of
International Society for Mangrove Ecosystems, barnacles in a temperate mangrove forest. J Exp Mar Biol
Okinawa, Japan, pp 250 Ecol 241:45–66
Field JG, Clarke KR, Warwick RM (1982) A practical strategy Satumanatpan S, Keough M (2001) Roles of larval supply and
for analyzing multispecies distribution patterns. Mar Ecol behavior in determining settlement of barnacles in a
Progress Ser 8:37–52 temperate mangrove forest. J Exp Mar Biol Ecol 260:133–
Grosberg RK (1982) Intertidal zonation of barnacles: the 153
influence of planktonic zonation of larvae on vertical Steneck RS, Dethier LE (1994) A functional group approach to
distribution of adults. Ecology 63:894–899 the structure of algal-dominated communities. Oikos
Imbert D, Rousteau A, Scherrer P (2000) Ecology of mangrove 69:476–498
growth and recovery in Lesser Antilles: State of knowl- Sutherland JP (1980) Dynamics of the epibenthic community
´
edge and basis for restoration projects. Restor Ecol on roots of the mangrove Rhizophora mangle, at Bahıa de
8(3):230–236 Buche, Venezuela. Mar Biol 58:75–84
Jaasund E (1976) A field guide to the intertidal seaweeds in Warwick RM, Clarke KR, Suharsono (1990) A statistical
¨
Tanzania, 1st ed. University of Tromso, Norway analysis of coral community responses to the 1982–3
˜
Kairo JG (1995) Community participatory forestry for El Nino in the Thousand Islands, Indonesia. Coral Reefs
the rehabilitation of deforested mangrove areas of Gazi 8:171–179
Bay Kenya. A first approach. Final technical report. Watson JG (1928) Mangrove forests of the Malay Peninsula,
BSP/WWF-(USA) and University of Nairobi (Botany vol. 6. Malay Forest Records, 275 pp
Department)
123