ATOLL RESEARCH BULLETIN NO. 574 DEMISE OF MADAGASCAR’S ONCE GREAT BARRIER REEF - CHANGE IN CORAL REEF CONDITION OVER 40 YEARS BY ALASDAIR HARRIS, GEORGE MANAHIRA, ANNE SHEPPARD, CHARLOTTE GOUGH, AND CHARLES SHEPPARD ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. MARCH 2010 Figure 1 . Location of Toliara, with Landsat 7 image of the Grand Recif. Arrows mark the two survey sites. DEMISE OF MADAGASCAR’S ONCE GREAT BARRIER REEF - CHANGE IN CORAL REEF CONDITION OVER 40 YEARS BY ALASDAIR HARRIS, 1 2 GEORGE MANAHIRA, 1 ANNE SHEPPARD, 2 CHARLOTTE GOUGH, 1 AND CHARLES SHEPPARD 2 ABSTRACT In the 1960s and 1970s the biology and geology of the Grand Recif of Tulear, (now Toliara) in southwestern Madagascar, was thoroughly studied and reported. Toliara is the largest city in the south of the country, and the Grand Recif offshore provides both artisanal fisheries and coastal protection to the growing regional capital. Substantial research on the comparatively pristine reef was described in a volume of Atoll Research Bulletin in 1978. Since then, published scientific study of this reef has been largely lacking. The present study compares the condition of the Grand Recif of circa 40 years ago, with that seen in a brief resurvey undertaken in 2008, on transects corresponding to some of those documented previously. The trend has been of severe degradation; hard coral cover on the fore-reef slopes has declined substantially, and there has been a near total loss of the “architectural species” in particular. Coral has been replaced to great extent by fleshy algae. Observations also indicate severe decline on the broad reef flat, back reef and lagoon areas. Perhaps most seriously for the local fisheries and human communities, is that the fore reef is almost depleted in reef fish today. Comparisons are made of coral cover, coral morphological types and fish trophic structure with other reefs in southern Madagascar, which are not located near large human populations. Although a rise in mean sea surface temperature has occurred throughout the region of approximately 1°C over this 40 year period, which is probably a contributing cause of decline throughout, the Grand Recif is in much worse condition than most of the more remote reefs with which it is compared. It is suggested that the main reasons for the substantial decline in the Grand Recif over the past 40 years lies in the fact that the region’s population has grown substantially, there is a complete lack of any resources management, heavy overfishing, and no pollution control, resulting in massively increased discharges of sewage, sediments and other pollutants. Reef condition today is unrecognisable from that described in the 1970s. Unless far-reaching and effective management interventions are adopted to safeguard the Grand Recif the remaining ecosystem services upon which Toliara and its population depend wifi soon all but disappear. 1 Blue Ventures Conservation, 22-24 Highbury Grove, London, N5 2EA, UK 2 Department of Biological Sciences, University of Warwick, Coventry, CV4 7AL, UK Correspondence: Charles.Sheppard@warwick.ac.uk Manuscript received 17 June 2009; revised 9 October 2009. 2 INTRODUCTION Coral reefs run continuously for over 450 km along Madagascar’s southwestern coastline (Figure 1). The central geographical feature of this predominantly fringing reef system is the geomorphologically distinct Grand Recif which is 19 km in length, has a shallow reef area of approximately 33 km 2 , and which lies 1.5-12 km directly seaward of the coastal city. The city’s port lies protected within the reef’s lagoon, less than 2 km from the back-reef slope. The Grand Recif has historically been the most well-studied in Madagascar, and was the focus of intensive research efforts out of Toliara’s Marine Station from 1961-1970; during that period approximately 400 scientific reports were published on the region’s marine environment, culminating with respect to biological work in a 490 page volume of Atoll Research Bulletin (Pichon, 1978). These studies described the predominantly anthozoan communities of the Grand Recif and adjacent reef systems, and provide a valuable insight into the condition of Madagascar’s southwestern coral reefs prior to recent increased levels of direct anthropogenic and climatic disturbance. At that time, Toliara’s coral reef was amongst the world’s most well-studied. Its coral genus count of 62 was the highest recorded for the entire Indo-Pacihc, though Pichon (1978) acknowledges that this was an artifact caused by his intensity of sampling; Rosen’s (1971) compilation and the predictive biogeographic maps of Stehli and Wells (1971) already indicated, for example, that highest diversity resided in the Southeast Asian region, but that a secondary high diversity locus probably existed in the Western Indian Ocean. Following this, very little research from Toliara’s coral reefs was published over the intervening decades, when the city’s population increased markedly. A search of the words ‘coral’ and ‘Madagascar’ in the online bibliographic reference archive Scopus yields a total of 54 items published since 1971, of which just 6 refer to ecological and fisheries issues relating to Toliara’s reefs, none of which provide quantitative information on the ecological condition of coral reefs in light of the growing anthropogenic and climatic stresses. A small number of non-peer reviewed studies shed some light on the status of, and threats to, nearby reef systems but, notably, there are no records of occurrences of coral bleaching in Toliara between 1998 and 2002, a period which saw several mass coral bleaching and mortality episodes elsewhere in the Indian Ocean. It is likely that Toliara’s reefs suffered widespread bleaching on at least one occasion during this time (Ahamada et al., 2008). Many people living in the region have described this to the present authors, but it has been undocumented. Toliara’s population increased by 53% between 1993 and 2008, and is forecast to increase by a further 49% over the next 14 years (INSTAT, 2007). Population growth comes inevitably with a concomitant increase in artisanal fishing pressure in adjacent reefs. Mangrove stands once lining the shoreline between the city and the Grand Recif ’s lagoon have been almost entirely removed in recent decades. There is no centralised organic waste disposal system for the region, and human waste is commonly disposed of on the deforested mudflats for removal by each outgoing tide. Tagoonal water is highly turbid, and there are no management controls in place to regulate artisanal fisheries, 3 pollution or waste disposal. Those environmental regulations that do exist are rarely implemented due to a lack of enforcement capacity. A series of sedimentalogical and geomorphological surveys carried out between 1990 and 1995 on the shallow reef flat of the Grand Recif provides valuable biological information of the very shallow areas from that time (Thomassin et al., 1998; Vasseur et al., 2000). Notable changes in the condition of the reef flat include a reduction in the height and width of the outer boulder tracts; a decrease in coral cover within the “inner moat” of the reef flat, with greatly reduced living coral cover; widespread mortality of most branching and digitate corals ( Acropora humilis , A. muricata , A. arbuscula ); a considerable increase in sea urchin abundance (. Diadema , Echinothrix , Echinometra ); substantially increased algal growth on all shallow hard substrates, predominantly brown algae ( Sargassum ) in summer and green algae ( Enteromorpha , Ulvaria) in winter; and increased cover (up to 90% benthos in some cases) of zooanthids ( Palythoa ) and soft corals on the outer reef flat in habitats previously dominated by Acropora. On soft substrates, there was almost complete disappearance of seagrass beds within the lagoon, where the once dominant Syringodium , Thalassia and Thalassodendron , were by then replaced by a sparse coverage of Halodule uninervis and Cymodocea rotundata only. Regarding wave energy, a reduction in the sheltering effect of the barrier reef against heavy swells was noted, with waves breaking on the reef flat landwards of the boulder tract, whereas in the 1960s it was noted that waves did not break landwards of the boulder tract. These studies also inferred an increase in current velocity across the reef surface, as well as an overall reduction in the roughness of the reef surface over time as a result of decreased coral cover and complexity (Thomassin et al., 1998; Vasseur et al., 2000). The changes in the condition of the reef flat were considered to be caused by over-fishing, particularly of herbivorous species, pollution from Toliara, and heavy sedimentation from deposition of silts and clays from the nearby Onilahy and Fiherenana rivers (Vasseur, 2008 and personal communication). The present study presents results from a brief but quantitative assessment of coral cover and fish composition of the reef slope of Toliara’s Grand Recif. It compares today’s fore-reef slope condition with that of circa 40 years ago. The comparison has the limitation that much of the older work was heavily descriptive rather than quantitative, but sufficient re-analyses and inferences can be made for comparative purposes. METHODS Surveys were undertaken on the outer reef slope of the Grand Recif in July 2008. Survey localities were as close as possible to those examined and described in detail by Pichon (1978). Firstly, the reef slope in four areas was surveyed to 30 m depth at intervals of 5 m. Using approximately 20 replicate visual quadrats of approximately 5x5 m at each depth, percent cover by coral, soft coral, algae, sponge, coralline algae, bare rock and sand was recorded. Secondly, line intercept transects were used at 10 m depth to examine that region in greater detail, where the composition of benthic and substrate types were recorded in two replicate 10 m line intercept transects (TIT) (English et al., 4 1997). Stony coral community composition was recorded along each transect line by measuring the longest diameter of every coral lying within a 10 m 2 belt running 50 cm either side of each LIT. Corals with >50% of their surface area within the 10 m 2 belt were recorded. These corals, identified to genus, provide size frequency data on the major components. Wherever possible, these data are presented with data extracted from the largely descriptive work of Pichon (1978). In the latter work, the cover of different categories of reef occupancy was derived from Figure 3; in his “kite diagram,’' the width of the black blocks represents the relative cover of each group, and these were used to infer the percentage cover of each category. For each depth, the widths for each category appear to sum to the same value (taken to be 100%). One complication which cannot be properly resolved in this case, however, is the apparent absence or non-recording of sand in the 1970s data. In 2008, sand was a very minor component (1-2%) down to depths of 15 m, and, being an exposed reef, this is likely to have been fairly similar in the 1970s, but from 15 m depth and deeper, sand occupies up to 18% of the substrate. Simultaneously, fish diversity and biomass were measured at a depth of 10 m using an underwater visual census (UVC) technique. At each site a 50 m line was laid along the reef benthos and all fish seen within 2.5 m either side of the line were identified and recorded, with length assigned to 10 cm size classes from 0-80 cm, or >80 cm. Each transect was passed four times, with 1-3 fish families sampled during each pass along the transect. Finally, HadlSSTl monthly sea surface temperature (SST) (Rayner et al., 2003) data were obtained for Toliara, with a spatial resolution of 1 degree latitude and longitude. RESULTS Figure 2 shows cover profiles of corals and algae with depth, in 2008. Hard coral declines with increasing depth while algae broadly increase in cover with depth. Sand provided significant substrate cover, while a fourth category, labelled ‘bare rock’ is also shown, which is substrate devoid of macro algae and animal life but which is covered mainly with filamentous or turf algae. Soft corals accounted for 11% of cover in the shallowest region, but became negligible below 15 m, while sponges, which were very conspicuous 40 years ago between 20-30 m depth, were negligible throughout in this study. In Pichon (1978) there is no similar quantitative data on cover, but a kite diagram of relative percentages (Figure 3) showed a decline in the proportion of coral cover with depth and a simultaneous increase in fleshy algae. Bare substrate and sand appeared to be unmeasured in Pichon (1978) but are clearly high in 2008. More detailed LIT results at 10 m depth, and extraction of approximate, equivalent data from Pichon (1978), further show deterioration at this formerly high diversity region. There is now greatly reduced coral and soft coral cover, reduced coralline algae, and a massive increase in macro and turf algae (Figure 4). 5 Figure 2. Hard coral (dark grey), fleshy algae (light grey), bare substrate (patterned) and sand (open) cover of fore-reef slope of the Grand Recif to 30 m depth. Figure 3. Relative cover of the main biotic groups from Pichon (1978). 6 Figure 4. Relative percentage composition of living benthos categories at 10 m depth from LITs in 2008 (white bars with s.e.) compared with data inferred from Pichon (1978) (figure 3) (grey bars) from the same depth. Comparing results from the two methods used in 2008, estimates of hard coral cover obtained by LITs showed an average cover of 13% ± 5.0 SE, which is slightly lower (but with overlapping error bars) than the mean result of 20% obtained using the visual quadrat estimates taken at the same time. For erect fleshy and turf algae both sets of values were similar. With both 2008 methods, the deterioration over 40 years is marked. Of the corals present, only 1.6% ± 1.3 SE of the benthic composition in 2008 were of structurally complex hard coral growth forms (branching, digitate or tabular colonies). The present paucity of architecturally complex coral genera is reflected in the coral community composition data. Only 15% and 7% of colonies recorded belonged to the families Acroporidae and Pocilloporidae. However, although sparser, corals remained diverse, with 30 genera encountered in 2008 with a median colony size of 20cm maximum diameter. Colony density averaged 539 colonies 100 nr 2 , the community being dominated by massive and encrusting Poritidae and Faviidae. The two most dominant genera were Echinopora and Porites. Large expanses of Acropora of several 7 kinds recorded by Pichon (1978) were missing in 2008, and no evidence existed of any dead or eroded tabular coral framework, suggesting that the widespread loss of Acropora had taken place several years previously. On the reef flats, there was, in 2008, very little live coral. This contrasts markedly with photographs of reef flats shown in Pichon (1978) (Figure 5). Figure 5. Photographs from Pichon (1978): Top: Pichon’s Figure page 38, ‘Plateforme superieur des eperons. Peuplement a Acropora cf pen guis et Acropora humilis .’ Bottom: Pichon’s Figure 63: ‘Champ de Scleractiniaires branchus (Recif sud-Ifaty).’ 8 Fish species richness was particularly low in 2008, with a total of just 14 reef fish species observed during the survey. Mean reef fish biomass was estimated to be 134.6 kg ha 1 (± 33.9 SE) in 2008. Seventy five percent of the fish species were herbivorous, a trophic composition which differs markedly with that seen in 1979 (Figure 6) (Harmelin- Vivien, 1979). 100% 90% 80% 70% 60% 50% 40% 30% 20% 10 % 0% 1979 Harmelin-Vivien This study 2008 Figure 6. Reef fish trophic-guild composition at the Grand Recif in 2008 (right column), compared with data from Harmelin-Vivien (1979) (left column). Sea surface temperature for this region has risen over the time span considered (Figure 7). The linear rise of mean SST is over 1°C, and warm summer peaks are seen in both 2005 and 2007, with the latter at 29. 12 being the highest in the data series. No quantitative observations of widespread coral bleaching have been recorded in the Toliara region in recent decades, but regular monitoring has only been carried out in southwestern Madagascar since 2003, 5 years after the Indian Ocean- wide mass bleaching and mortality episode that may have heavily impacted the region’s reefs. □ Other □ Herbivore ■ Carnivore 9 Figure 7. Sea surface temperature (HadlSSTl data) for the Toliara region. Trend line is the 12 month moving average. Comparison with Other Sites in Madagascar The benthic composition of the Grand Recif today is similar to that on seaward slopes of the barrier islands of Nosy Fasy and Nosy Hao, which are both similar reef structures located some 200 km north of the Grand Recif (Figure 8) (unpublished data and Nadon et al. 2007). In these sites also, coral cover is very low at 8-14%, with similarly high coverage of fleshy macroalgae of around 60%. Colony density is similar too, at between 200 and 600 colonies 100 nr 2 (unpublished data). These sites, including the Grand Recif, contrast greatly with two protected reef sites in the southwestern region that lie within the Velondriake marine protected area (established in 2008, but previously benefiting from de facto managed status from low fishing effort); in the latter, coral cover approaches 80% and colony densities are three times higher with around 1,400 corals 100 nr 2 . Moreover, these protected reefs show a far higher relative abundance of structurally complex genera, with approximately 15-40% of colonies belonging to the architecturally complex Acroporidae and Pocilloporidae respectively. In the protected area, human populations are very much lower, the region being populated by remote fishing villages and settlements of nomadic fishers, with no major towns. 10 Fish diversity and biomass on the Grand Recif may also be contrasted with other reef surveys conducted in the region (Figure 9). Both parameters on the Grand Recif are dramatically lower than in the more remote and less populated areas. CM E o a> n £ D Fasy Hao Figure 8. Comparison of hard coral colony density of major scleractinian families of Grand Recif Tulear (GRT) with four reefs located approximately 200 km north. These are two barrier reefs which are fished and (right pair) two patch reefs within a marine protected area with comparatively low human population density. 11 Fish Biomass