Mangrove forest composition and structure in Las Perlas Archipelago, Pacific Panama

14 Pages·2010·0.84 MB·English

Mangrove forest composition and structure in
Las Perlas Archipelago, Pacific Panama
Tom McGowan1, Sarah L. Cunningham2, Héctor M. Guzmán3, James M. Mair1, José M.
Guevara4 & Tanja Betts1
1. Centre for Marine Biodiversity and Biotechnology, School of Life Sciences, Heriot-Watt University, Edinburgh EH14
4AS, UK; [email protected]
2. Marine and Coastal Development Unit, Argyll & Bute Council, Lorn House, Oban, Argyll PA34 4LF, UK;
[email protected]
3. Smithsonian Tropical Research Institute, PO Box 0843-03092, Balboa, Ancon, Republic of Panama;
[email protected]
4. Environmental Resources Management Panama S.A., World Trade Center, 0832-0588 WTC, Panama, Republic of
Panama; [email protected]
Received 26-vII-2009. Corrected 11-III-2010. Accepted 12-Iv-2010.
Abstract: Mangrove forest is an important ecosystem that provides many services, but in Panama, as in other
countries, they are under threat due to a variety of human activities. Nowadays, large areas of mangroves con-
tinue to be lost without been described and lack of management strategies. This study focused on the mangrove
structure in the two largest islands, Isla del Rey and Isla San Jose, of Las Perlas Archipelago (LPA), Pacific
Panama. Assessment of Landsat satellite imagery revealed loss of mangroves in the LPA of 965ha in the period
1974-1986, and 248ha in the period 1986-2000. The majority of the loss (>77%) from the two study islands was
due to timber extraction and agricultural development. In May 2006, permanent plots following the CARICOMP
protocol were established at two sites on Isla del Rey (R1 and R2) and one site on Isla San Jose (SJ) where stan-
dardized metrics such as species, height and diameter at breast height of adult trees and seedlings were recorded.
Forest structure differed at the three sites, although R1 and R2 were most similar. At R1, Laguncularia racemosa
was the important species and R2 was dominated by Pelliciera rhizophorae. Examination of the forest structure
and classified imagery indicated that these sites are spatially dynamic and appear to be rejuvenating. The forest
structure would indicate that the sites have been growth-limited previously by human activities and possibly by
other factors. SJ was dominated by Rhizophora mangle and appears to have a mature forest with large adult trees
and few seedlings. It does not appear to have shown the same extent of spatial regrowth as the other two sites
between 1986 and 2000 and is relatively static. The establishment of permanent plots and monitoring will be
useful as part of the management plan, as the LPA shows a variety of mangrove structures and could be subject
to further coastal development. Rev. Biol. Trop. 58 (3): 857-869. Epub 2010 September 01.
Key words: mangrove structure, Rhizophora, Laguncularia, Pacific Panama, Las Perlas Archipelago.
Mangrove forests cover approximately quality, reduction in severity of storm, wave
20 million hectares worldwide and are the and flood damage, medicine, and as breeding
main vegetation type in protected intertidal and feeding areas for commercial and artisanal
areas along tropical and subtropical coastlines fishery species (Robertson & Phillips 1995,
(English et al. 1997, Cardona & Botero 1998, Yoshiro et al. 1997, Bandaranayake 1998,
Hogarth 1999, Krauss et al. 2008). Mangroves Baran & Hambrey 1998, Nagelkerken et al.
provide several important ecosystem services 2000, Wolfe et al. 2000, Kathiresan & Bing-
including the maintenance of coastal water ham 2001, Krauss et al. 2008).
Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010 857
The structure of a mangrove forest at any Rabinowitz 1978a,b,c, Smith et al. 1989, Duke
point in time is a function of its successional et al. 1997, Benfield et al. 2005, Defew et
stage, species composition, zonation, propagule al. 2005, Lovelock et al. 2005) no scientific
dispersal, growth and survival. These are all research to the authors’ knowledge has been
influenced by a number of biotic factors such carried out on the mangroves of LPA. This is
as propagule variables, propagule predation, largely due to the isolated nature of the islands,
herbivory, human interference and interspecific which lie some distance from the mainland
competition. Additionally, abiotic factors are coast of Panama and Panama City. Another
also influential including storm damage, rain- consequence of the LPA’s isolated nature is the
fall, tidal influence, freshwater input, tempera- lack of development on the islands leading to
ture, sedimentation rate, nutrient availability what can be regarded as a largely pristine envi-
and light (Krauss et al. 2008). To analyse all ronment. However, there is an increasing level
of the above components and relate them to of interest for further coastal tourist develop-
mangrove forest structure, and each other, is a ments in LPA such as on Isla del Rey (Fig.
complicated exercise in ecological modelling 1) and sand extraction operations in marine
but attempts have been made (Twilley et al. and coastal areas have been licensed for some
1998, Schaeffer-Novelli et al. 2005, Twilley & locations. The marine environments of the
Rivera-Monroy 2005). LPA were designed in 2007 as a protected area
Mangroves are, however, under threat from under the status of a Marine Special Manage-
a variety of anthropogenic activities, and it is ment Zone. It was therefore considered desir-
estimated that one-third of the world’s man- able and timely to collect data on mangrove
grove forests have already been lost (Alongi forests within the LPA before they potentially
2002). Large areas of mangrove forests are become affected by future development.
cleared for shrimp aquaculture in developing The aims of the study were to assess the
countries (Primavera 1993, Adeel & Pomeroy structure of mangrove forest on the two larg-
2002). Other impacts include coastal urbaniza- est islands, Isla del Rey and Isla San José, and
tion, wood extraction, agriculture conversion, to establish permanent plots at these sites to
salt production and tin mining (Ong 1995, allow future monitoring of mangrove forest
Macintosh 1996, Alongi 2002, Benfield et al. structure.
2005, Walters 2005). Oil spills may also cause
large-scale damage to mangrove ecosystems
MATERIAL AND METHODS
(Duke et al. 1997). As well as direct impacts,
mangroves, due to their coastal nature, could Study Site: The LPA lies between
be under threat from global warming if these 08°40’19”-08°11’46” N, 79°03’ 49”-78°46 ’31”
results in sea level rise (Field 1995). W within the Gulf of Panama (Fig. 1A). It is
Within Pacific Panama, most of the man- composed of 250 mostly uninhabited basaltic
groves in the Gulf of Panama are classified as rock islands and islets that lie within the 50m
either critical or endangered, on the conserva- isobath and falls within the Tropical Eastern
tion status scale, due to losses associated with Pacific (TEP) biogeographic zone. The LPA
human development (D’Croz 1993). Areas of has a tropical moist climate with annual rainfall
mangroves under threat from human interfer- between 250cm and 300cm. The mean tidal
ence and not previously studied should be range for the archipelago is 3.8m with maxi-
priority targets for research. With this in mind mum of ca. 6m (Glynn & Maté 1996).
the Las Perlas Archipelago (LPA) in Pacific The LPA was designated as a marine pro-
Panama was selected as a survey location as the tected area, locally known as a Marine Special
forests found on the islands meet these criteria. 688km2 and encompasses all of the islands and
Whilst there have been several mangrove stud- Bajo Trollope, a rocky shallow bank at the
ies conducted in Panama (Rabinowitz 1975, South-East corner of the archipelago (Fig. 1A).
858 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010
Fig. 1. A. The location of Las Perlas Archipelago in Pacific Panama and the survey sites for this study as indicated; (B.-D.)
the distribution and decline of mangrove forest areas in the archipelago since 1974 mapped using Landsat satellite images.
Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010 859
The MSMZ was created to implement an inte- were easily accessible. The distribution of man-
grated coastal management approach including groves within the archipelago and the location
the production side of fisheries and tourism of the three sites chosen are shown in Fig. 1A.
and long-term landscape conservation planning The two sites on Isla del Rey are subsequently
(sensu Margules & Pressey 2000). The law referred to as R1 and R2, and one site on Isla
protects coral reefs, reef fish, mangrove forests, San José (SJ). R1, R2 and SJ were located at
except where there is an overriding need for 08°24’ N, 078°51’ W; 08°16’ N, 078°55’ W;
tourism development, and prohibits the use of and 08°14’ N, 079°06’ W respectively.
certain fishing methods, allowing only artisa- During the month of May 2006, within
nal line fishing. Whilst land use is unregulated each of these three sites, five 10 m by 10 m
under Law 18, the relatively pristine state of the plots were randomly chosen and marked out
islands is a conservation priority. Over 98.2km2 using four PvC poles inserted into the sedi-
of forest, including some mangrove forest and
ment (sensu English et al. 1997, Benfield et
the largest watersheds on Isla del Rey were
al. 2005). Their locations were recorded using
designated a Hydrological Reserve in Octo-
GPS and photographic records were taken to
ber 2006, while the entire island of San José,
allow re-location. At each plot measuring tapes
including its mangrove forests, is protected as
were laid between the PvC poles to represent
a Private Nature Reserve.
x and y-axes from a corner acting as 0,0. This
produced a grid system to obtain a position of
Distribution of mangrove forest: Man-
individual trees. The pole acting as 0,0 was
grove forest was mapped in a previous study
marked and the orientation of each plot in rela-
by Guevara (2005) from three Landsat satellite
tion to 0,0 was common to all plots (0,0 was
images acquired on 23rd February 1974 (Land-
located on the most seaward side of the plot on
sat MSS), 1 February 1986 (Landsat TM) and
the left when back was to the sea).
23 November 2000 (Landsat ETM+). Seven
land cover types, including mangrove forest,
Data collection and processing: All live
were predefined and ground truth data were
trees greater than 1m in height above the sedi-
collected in May 2004. A maximum likelihood
ment surface were tagged with marking tape so
supervised classification was performed of the
they could be identified. This height restriction
year 2000 image using ground truth data and a
ensured the inclusion of saplings, defined as
combination of bands 1,2,3,4,5 and 7 in Erdas
trees greater or equal to 1m in height with a
Imagine 8.7. Unsupervised classifications were
girth of less than 4cm (English et al. 1997).
conducted on the 1974 and 1986 datasets using
For each individual, the species was noted,
the same number of categories. Areas of man-
height above ground sediment measured and
grove forest were calculated for each island
the x and y coordinates taken to allow future
greater than 20ha in size. Deforestation rates of
mangroves were calculated from 1974 to 1986 monitoring. Tree diameter at breast height
and 1986 to 2000. (DBH) was measured at 130cm from the
ground (Brokaw & Thompson 2000) using a
Selection of survey sites: Three study sites pre-calibrated tape measure when DBH was
chosen in the LPA were selected from the land greater or equal to 2.5cm. In circumstances
cover mapping using Landsat ETM+ (Guevara when the tree branched below this height stan-
2005) and previous visual surveys of the area dardised rules were followed (sensu English et
by boat. These sites were chosen as they were al. 1997, Benfield et al. 2005). Trees measured
located on islands with some of the largest for DBH were marked at the location where
areas of mangroves in the LPA, were felt to be this variable was recorded and numbered using
representative of the general forest structure, stainless steel tags to assist future relocation
had experienced decline in forest area, and and recording.
860 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010
Within each plot, data on seedlings were silt-clay fraction to very-coarse sand (Went-
also taken. Many plots had very low numbers worth 1929). The percentages of carbonate and
of seedlings (less than 50) allowing the direct organic carbon contents of the sediments were
measurement of every seedling. For those plots derived respectively by treatment with hydro-
that contained greater than 50 seedlings, five chloric acid and combustion (sensu Walkley &
1m by 1m subplots were randomly chosen Black 1934, Barnes 1959).
and marked using four PvC poles so that they
could be monitored (sensu CARICOMP 2001).
RESULTS
One meter rules were used to form the x and y
axis. The species, height and coordinates of the Mangrove forest distribution and change:
seedlings in subplots (or within the whole plot A summary of mangrove area covered and
where there were less than 50 seedlings) were area loss is provided in Table 1 and the area
recorded. Seedlings were also numbered with a loss over the period studied is presented geo-
stainless steel tag. graphically in Fig. 1(B-D). The greatest loss
The importance value of each species in of mangrove forest in the LPA during the
each plot for the adult/sapling population was period studied occurred between 1974 and
calculated using the equations given by Cin- 1986. An area of 965ha was lost according to
trón & Schaeffer-Novelli (1984). This was to the analysis performed on the satellite data, of
ascertain the contribution of each species to the which the majority of the loss (90.3%) was on
mangrove forest and the importance value was Isla del Rey and Isla San José. Between 1986
used as an indicator of the degree of mono- and 2000 mangrove forest in the LPA also saw
specificity (Cintrón & Schaeffer Novelli 1984, a decrease in cover, although not as much as
English et al. 1997). To examine differences in in the previously examined period, of which
height and DBH between sites Kruskal-Wallis Isla del Rey and Isla San José combined again
tests and Dunn’s multiple comparison proce- represented the majority of the loss (77.1%).
dures were implemented in SigmaStat. visual examination of land cover maps for
the same period also derived from the Landsat
Sediment samples: Within each plot one images of changes in land cover type revealed
sediment-surface sample of approximately 1kg that the mangroves were being replaced dur-
(wet weight) was taken for analysis of sediment ing this period by regenerating forest (forest
characteristics (organic carbon content, car- regenerating after logging and burning), brush
bonate content and particle size). Samples for wood-stubble and farmland, grasslands and
sediment analysis were frozen until required. In paddocks (Guevara 2005). This indicates that
the laboratory these samples were assessed for the mangroves lost were due to human extrac-
particle size distribution by passing dried sub- tion and agricultural development.
samples through a stack of differently sized Adult and sapling data: A summary of
mesh sieves (63, 125, 250, 500, 1.0, 2.0µm) adult, sapling and seedling parameters derived
and classified using the Wentworth scale from from the field data is given in Table 2 and
TABLE 1
Areas of mangrove and area lost in Las Perlas Archipelago (LPA) marine protected area and on Isla del Rey and
Isla San Jose as calculated from three classified Landsat satellite images
1974 1986 2000 Area change (ha/y) Area change (ha/y)
Area(ha) Area(ha) Area(ha) 1974-1986 1986-2000
LPA 2867.7 1901.5 1653.4 -48.3 -17.7
Isla del Rey 2288.2 1563.9 1356.3 -60.4 -14.8
Isla San Jose 105.9 85.1 68.7 -1.7 -1.2
Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010 861
d Fig. 2. All sites contained a full vertical struc-
2om the five 1m subplots where required Adult mean Seedling mean Importance % species % species rank of adult compositioncompositionspecies at siteLRPLRP L>R>P2670434660 P>R>L3238303970 R>L>P496001000 dlings <1m in height (English et al. 1997). Mean stan tRlSbabh(t(twiuiohnTHduoinJ2hrg edau tt=ge h ib hhibldchl sc1ate ewelh,i aae2dw ssst tbdi rt. stsitsea6 2reu n hihisr smg6e)tgsb eaio asn5ima u apem twdi,m hwtnal fguiiid iianldonurllcaaatf tea ngcas r= sttarD hnssh,r 2ovtn ne k tseeBi,feluheuean r yrHpmsD iDwpm y ng c=lg Bd bBeuhoib0secndiememeHHif.nogr mr0rfesmb pd s e0ao tr ioeairhao2latpfserfra na lft)a t inarlro .hnsdary idtsofeel b iF aus lnihnd fua ipnetl rga. shdtteRl oisid.tTirSudot omn 1e ea2lhaJln get sin.B es nRa sl d. seteSnw g n o1i asaiHdJsstul ascne le. tomes auhhsdRheR rwn,s a deb f21tgdsadoo lere e ili l tauvaarnlreehtRt tnnnehgfodoeehrt2dddessf-fr,,.
r e
gs, was extrapolated f Mean seedling height (m±SE) 0.45±0.004 0.63±0.003 0.740.003 >2.5cm but <4cm; se RatvaRin av12sdle iu mlbameynu i)ed(tl. a2 d irR7itR a cm2h2nm a eDbd)hdo BaicatadhoH n mm scachupl loacaasshwsrlsesi eeg ldadvsh r att(goal2lyeu csR rem i hm1D( i 4agBi(nlch2aHmde2r r c3) dr mcaitmmson)t. gar RrxieSbe,i1J smu wpthauieinoatmcdhdn-
TABLE 2which, for seedlin Mean stand basal area 2 /ha±SE)(m 8.9±2.3 17.1±3.7 16.0± 8.1 n height and D130 aamfsot naelm lnaiondanwcx tboieomamdtsa upablm lly sea t tae6rSne 6Jradc a wmndbgeha, essifcp aaohlir tf e a hl rcoaehlwdaaa s vewsairen ashgfs ri (eghaFqh iigmu ghme.u en2ceschBati ne i)lsn .o t TowaRtnhe2aderl,
plings and seedlings per plot (10mx10m) Mean adult tree Mean stand gsheight (m±SE)diameter (cm±SE) 3.5±0.415.3±0.03 4.9±0.026.4±0.04 10.7±0.4113.2±0.7 m in height and D>4cm; saplings>1m i130ulated only for stems with D>2.5cm.130a mangle; P=Pelliciera rhizophorae. ehdpvbafRnrsrei<eeue2eftirim0fq ) gnyae.d.uhg bT0nroReset e0dh mninsr21 meot ktt i)o h fenhic.bm fo lwalae atFsaaertdtreeeseiwda gd sdeniaui e.n s n l 2be t mht 2o. nRywa5etA a rd im1aaretgx uhel seshi lwll ems ahtattf . ethotouh olirewReuvmeef f ne si1sttsg il d r hhtyheatie oehn tensts irados dhgdt(t i e Hoh ici3bsnsRrta=t.et t t0 r1eoS i5ai nr fsbnJ3a agi utd.n1gwnr2 t de2nidtt6aoeh hmi 03fssnaeRi , . tc , g5(w md2amrmbtf nhehe=oaoat ieatlr2alsthynyee-tt,
Mean number of adults, sa No.Sample No.No.adult Sitesaplingsseedlinstems R1756662 R2687876 SJ17617 Adults classified as stems which were >1basal area and mean stand diameter calcL=Laguncularia racemosa; R=Rhizophor tatmci6auehht lnnmnsaweaadt rsm nx,,e i sm iRadpTaeeomnernsa2ihsetdr udxt,rh so em irfeweemfaeirnra t en i sShctutoqg eshheJmfrupde i amegte o.w nhincahfngaee ittt ten e r hhsseoRs yiiee wftz 1ai eioegt4an lsh fshw0s tc rooe tmlea mn rnaecmsso s lak (IstsaFse1ne islwdc0tigwasegoi.re aas5e.dons s tdme2vt we i amlA en. tanhR souu)sdeS .eo ci tJdwgyThuht ia,h seehef llhf tr xlae eleelar chier rhgkvleif aepbeeittnn srir ritgbe,stgRe eeehuisa 1ndass-ttt
862 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010
70
60
50
y
c
n 40
e
u
q
e
Fr 30 R1 R2 SJ
20
10
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
Height (m)
60
50
40
y
c
n
ue 30
q
e
Fr R1 R2 SJ
20
10
0
2 5 8 11 14 17 20 23 26 29 32 35 38 41 44 47 50 53 56 59 62 65
DBH (cm)
12
10
5
R1 R2 SJ
y
c
n
e
u 6
q
e
Fr
4
2
0
0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Height (m)
Fig. 2. A. The distribution of adult height classes; (B.) tree diameter at breast height (DBH); and (C.) seedling height at the
three survey sites.
Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010 863
terms of their most important species (Table 2). found to be significantly different between SJ
At R1 Laguncularia racemosa was the most and R1 (Q=4.027, df=1, p<0.05) and SJ and R2
important, at R2 Pelliciera rhizophorae and at (Q=2.629, df=1, p<0.005) but not between R1
SJ Rhizophora mangle. Both sites on Isla del and R2. Fig. 2C describes the height distribu-
Rey had all three species but only L. racemosa tion of seedlings at the three sites, and shows
and R. mangle were present at SJ. The percent- that R1 and R2 had a similarly wide range in
age composition confirms the dominance of R. seedling height distribution. R1 had an almost
mangle at all three sites. However, although R. normal distribution apart from high frequen-
mangle comprised a greater percentage of the cies at 10 and 15cm. The height distribution of
adults surveyed, at R1, L. racemosa was found R2 was roughly similar with no seedling less
to be the most important species (Table 2). than 15cm. The height distribution of SJ had
Analysis of species composition shows a much a much narrower range than R1 and R2, with
more even spread of species at R2 compared to a lowest value class of 40cm and a median of
the other sites. 95cm (Fig. 2C). At all three sites R. mangle was
Seedling data: Table 2 shows that R2 the dominant species amongst seedlings, with
had the highest mean number of seedlings, L. racemosa being the second most abundant
although there was much variation between at R1 and R2, although it was not found at SJ
plots surveyed. SJ had the lowest number of (Table 2). Seedlings of P. rhizophorae were not
mean seedlings. The mean height of seedlings recorded at any of the sites.
was greatest in SJ and those in R1 were short- Sediment Data: Table 3 shows the average
est, the same pattern as observed for the adults sediment composition of the three sites in terms
(Table 2), and the mean height of trees was of percentage organic carbon, carbonate and
TABLE 3
Organic carbon, carbonate and particle size analysis (Wentworth classification) in terms of percentage total
for all plots at the three sites surveyed. Very coarse sand (VC Sand), coarse sand (C Sand), medium sand (M Sand),
fine sand (F sand), very fine sand (VF Sand)
Sample % % Particle Size Analysis (%)
Site Organic C Carbonate Granule vC Sand C Sand M Sand F Sand vF Sand Silt-Clay
R1 0.5 6.0 3.4 2.3 4.8 23.3 44.6 19.5 2.1
R2 2.0 29.1 0.8 2.0 4.9 29.2 50.1 5.9 7.1
SJ 12.0 13.3 7.8 4.1 5.2 8.2 11.5 15.7 47.5
particle size. R1 had lower organic carbon and region, with R. mangle and L. racemosa being
carbonate content than the other two sites and particularly common but P. rhizophorae less
was composed of medium to very fine sand. so. The presence of only three species is not
R2 had the highest carbonate content of the uncommon as Central American mangroves are
three sites and was also dominated by medium generally species poor (Jimenez 1992, Murray
to fine sand. SJ showed the highest levels of et al. 2003). L. racemosa is usually regarded
organic carbon and a large proportion of its as a pioneer species that is out-competed as
composition was from the silt-clay fraction. the forest develops and therefore moves back
to higher elevations. The importance of L. rac-
emosa at R1 either indicates that this area is an
DISCUSSION
early stage community, or alternatively it could
Species composition: The species found be explained by mangroves at this site being
at the three sites studied are typical for this distributed across a relatively narrow band
864 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010
between the shore and the terrestrial forest, was low when compared to other values for the
resulting in larger specimens of L. racemosa in Pacific Central America region, which range
a very narrow band behind the more numerous from 6 to 96.4m2/ha (Pool et al. 1977, Jime-
R. mangle. The importance of P. rhizophorae nez 1992). Additionally, the density of adult
at R2 is noteworthy given P. rhizophorae has trees was also low when compared to other
a very limited distribution and Jimenez (1984) areas (Ong et al. 1995, Cox & Allen 1999).
reported that significant stands of this species The heights recorded fell within the dwarf and
are found only in neighbouring Costa Rica. P. medium range outlined by Murray et al. (2003)
rhizophorae was found to have a high basal and were substantially less than the values of
area, implying it must be relatively well devel- 16m and 22m reported in Costa Rica (Pool et
oped in this area. This species requires high al. 1977) and in other Panamanian mangroves
rainfall or terrestrial run-off (Jimenez 1984) (Mayo 1965) respectively. However, Tomlinson
and its distribution at the site corresponds to a (1999) notes that mature trees of P. rhizophorae
creek flowing from the land. The species com- can be expected to grow between 5 and 10m,
position of SJ was the most monospecific of the and the heights recorded in R2 for this species
three sites, being dominated by R. mangle. This fell within this range. The mean heights of the
is not particularly unusual as it is common for seedlings were well spread but the density of
mangroves to form monospecific stands, and seedlings at the sites was low. It appears that
there are examples of R. mangle out-competing the seedlings were being prevented from estab-
L. racemosa (Ball 1980, Sherman et al. 2000). lishing by a possible combination of limiting
Additionally, R. mangle is known to be the factors such as human intervention, salinity,
most common species in the area (Lacerda et light, nutrients and predation (Lopez-Hoffman
al. 2002). et al. 2007, Krauss et al. 2008).
The seedling species composition of R1 Examination of the classified satellite data
and R2 was found to differ from the adult (Fig. 1B-D) reveals that mangrove forest was
population; for R1 there were no individuals of lost in both R1 and R2 over the period 1974-
P. rhizophorae present, although the proportion 1986 and 1986-2000. However, whilst in this
of L. racemosa and R. mangle were roughly latter period there was a net loss, visual inspec-
representative of the adult species composi- tion of the classified satellite images showed
tion. At R2 there were also no seedlings of P. that there was some regrowth between 1986-
rhizophorae despite it being the most important 2000 indicating that the forest is spatially
species in the adult phase. Tomlinson (1999) dynamic (Dahdouh-Guebas & Koedam 2002).
states that P. rhizophorae experiences compe- During the first period, agricultural activity
tition from R. mangle, whose seedlings were was known to be higher compared to the later
almost completely dominant and this may offer period, when agriculture decreased and com-
an explanation for this pattern. The seedling munities started to rely more on fishing. This
cohort at SJ is entirely of R. mangle, which is was due to improved transport links to the LPA
unsurprising given the adult species composi- allowing more provisions to reach the island
tion. However, it was the only site not to have communities by air and sea, thus reducing the
seedlings of L. racemosa, a common species need to grow their own crops (Guevara 2005).
in the surrounding area. It seems unlikely that The reduction in agricultural pressure appears
propagules of L. racemosa were not reaching to have allowed some recovery and growth of
the site but there may be factors preventing mangroves in these areas, although the grow-
their establishment and/or they are outcom- ing older population may return to the area and
peted by R. mangle. expand agricultural activities by the end of the
decade (pers. observ.)
Forest structure and recovery: For both A combination of the satellite data and the
R1 and R2, the mean basal area for the site proportion of adults to saplings to seedlings,
Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010 865
seems to indicate that R1 and R2 are probably coupled with the proportion of adults, saplings
areas of rejuvenating forest with elements of a and seedlings, and its developed structure sug-
static nature, where seedlings are originating gests that SJ is probably a mature forest for the
from young trees/saplings (sensu Dahdouh- area that is relatively static in nature.
Guebas & Koedam 2002). The growth at these This study has found that a variety of for-
sites could have previously been restricted est types and structures, with different domi-
though felling and agriculture development nant species exist within the LPA. The sites
and/or by other abiotic or biotic factors. R1 studied showed a variety of structures and
is a fringing forest and it may be that it is the stages of development, and each had a different
least productive of the three community types dominant species in terms of importance rank.
owing to a lack of nutrients (sensu Lacerda et The area of SJ showed the most developed
al. 2002). It was noted that little or no leaf litter structure of the three sites, which is of interest
was found on the sediment surface at R1 and as mature forests are generally an exception
the sediment analysis contained relatively little and forests with such a maturity were observed
organic carbon. It could be inferred from this only once in early surveys of the LPA by boat.
that the sediment may also have had a lower Additionally, unpublished work on the genetics
nutritive quality, compounding the slow rate of R. mangle in the LPA using microsatellite
of recovery or that the tidal regime removes analysis has demonstrated a mixed gene flow
the litter. between islands of the LPA and the mainland
SJ had a low mean stand basal area, a very but that the genetic diversity of the LPA is
low density of adult stems and trees had a wide higher than many other areas of Pacific Central
range of heights and DBH, with several trees America. Previous analysis of satellite imagery
over 20m and having a DBH over 25cm. The of LPA forest cover has provided some indica-
site was a basin forest community behind a tions of forest dynamics from snapshots since
large sand bank and Lacerda et al. (2002) note 1974 and the detailed study of sample sites in
that basin communities retain nutrients and 2006 in three mangrove areas of Las Perlas
therefore can have higher productivity. This is Archipelago has added more understanding of
corroborated by the sediment analysis which actual structure and potential trends in man-
showed a higher organic carbon content rela- grove dynamics. It appears from these that the
tive to the other two sites. Whilst there were mangrove forests of the LPA offer an interest-
seedlings at this site, there was none under ing array of attributes for further research and
40cm and there were relatively less seedlings it is hoped that the establishment of permanent
compared to the other two locations. As previ- plots in this study will aid further research and
ously mentioned, for sites R1 and R2 it appears monitoring, as recommended for the manage-
that the seedlings are being prevented from ment plan of the marine protected area. It
establishing by a possible combination of lim- would be highly desirable to gather further
iting factors. However, possible explanations data on these poorly studied mangroves and
may be reduced irradiance due to the increased to monitor change brought about by the future
canopy cover of mature trees or predation planned development on the islands which in
by grapsid crabs which have been noted to our opinion is the only potential threat this
increase in abundance when larger tree cano- habitat.
pies develop (Smith 1992, Dahdouh-Guebas et Historical changes in Las Perlas archi-
al. 1998). pelago have been exclusively associated to
The classified satellite data (Fig. 1B-D) the expansion of the agriculture frontier and
reveals that mangrove forest was lost over both only affected the largest island of Del Rey. The
periods examined in the area of the study site mangrove areas are not adequate and extensive
and has not shown recovery, indicating that it for aquaculture. The creation of the marine
is spatially static in nature. This information, protected area regulates and prohibits fishing
866 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 857-869, September 2010