Giant Tortoise Distribution and Abundance in the Seychelles Islands: Past, Present, and Future
Abstract
The giant tortoises (Aldabrachelys or Dipsochelys) of the Indian Ocean Islands have been in decline since the first human settlement of the islands. They retain only a single natural population on Aldabra Atoll (and possibly Ile aux Cerfs, where tortoises are descendants of a mixture of indigenous and imported animals). Several additional wild populations are known, resulting from reintroductions to the historic range and introductions outside of that range. The historical distribution of tortoises in Seychelles is summarized, with reliable tortoise records from only 4 coralline islands and 23 granitic islands and the status of all the wild populations reviewed. This includes the first census of the Frégate Island population. In the granitic islands, only 9 islands support tortoises today and the wild population of these islands is estimated at 500–550 adults. In the coralline islands, tortoises are now present on 11 islands, with a total population of over 100,000 (almost all on Aldabra). Climate change impacts over the next 100 yrs are expected to be severe in low-lying areas of the Seychelles Islands attributable to sea-level rise and storm impacts on coastal erosion. These are projected to result in the loss of many populations and significant declines in the Aldabra population. As a result, the species should be regarded as Vulnerable by IUCN Red List Threat Criteria. Reintroduction to more of the high granitic islands could offset some of these projected declines, and it is recommended that such reintroductions be included in future conservation programs to restore ecosystem function.
Historically, giant tortoises occurred throughout the western Indian Ocean Islands, with 8–12 species (depending on the taxonomic system used) in two genera. These were recorded from Madagascar, at least five of the Mascarene Islands, and many of the Seychelles Islands. Shortly after human settlement, the giant tortoise populations declined catastrophically as a result of hunting and probably also of predation of hatchlings by introduced predators (dogs, rats, cats); by 1800 the only survivors were in the granitic Seychelles Islands and Aldabra Atoll (Fig. 1). Because giant tortoises were the largest herbivores of the islands, and reached high population densities at least in the Mascarenes, the extinction of most populations has resulted in dramatic changes in the ecosystems of the islands. The impact of these changes are not fully understood, although their significance in the Aldabra Atoll ecosystem has been demonstrated (Gibson and Hamilton 1984), and a significant role in plant evolution in Mauritius is suggested (Eskildsen et al. 2004).
Citation: Chelonian Conservation and Biology 12, 1; 10.2744/CCB-0902.1
The Aldabra giant tortoise has been considered to be under threat since 1874 (letter from Hooker et al., reproduced in Günther 1877; Gerlach 2004). In recent times, it has been included on the IUCN Red List of Threatened Species by IUCN as Rare (1982–1994) and more recently as Vulnerable on the basis of its restricted range (1996). Its status has not been updated in the past 14 yrs, and as such, this listing is indicated as “needs updating” by IUCN. The purpose of this paper is to review changes in the different populations of giant tortoises in Seychelles since 1996 to provide current information on their status. In addition, prospects for establishment of new populations and the potential impacts of projected sea-level rise attributable to climate change are considered.
Taxonomy and Nomenclature
The taxonomy and nomenclature of Indian Ocean giant tortoises has been highly contentious for over 100 yrs. The giant tortoises in the Seychelles Islands have been assigned to 7 species, many of which are considered invalid (Rothschild 1915; Gerlach 2004). Recent taxonomic publications recognize between one and three living taxa, variously considered species, subspecies, or morphotypes (Bour 1982; Gerlach and Canning 1998; Fritz and Havas 2007). These have been distinguished morphologically but not genetically (Palkovacs et al. 2003) although captive breeding demonstrates that the morphotypes breed true (Gerlach and Bour 2003; Gerlach 2011b). Recent lists of chelonian species published by the IUCN Tortoise and Freshwater Turtle Specialist Group lists three Seychelles taxa as subspecies (Turtle Taxonomy Working Group 2009, 2011). The group notes that currently genetic techniques are not satisfactory for identification of many chelonians. Even more contentious is the nomenclature. Recent taxonomic revisions are unanimous in recognizing the giant tortoises of Seychelles and Madagascar as a distinct genus (formerly included in Geochelone), for which the Aldabra tortoise is the type species (e.g., Le et al. 2006). However, three species names and two generic names have been applied to this species in recent years (Aldabrachelys gigantea, Dipsochelys elephantina, and Dipsochelys dussumieri), with no consensus on their validity. This issue is currently the subject of a petition to the International Commission on Zoological Nomenclature (see multiple author comments in Bulletin of Zoological Nomenclature 66). Pending the ruling of the ICZN, the scientific name of the Aldabra giant tortoise is not used here. Similarly, the validity of the Seychelles morphotypes lies outside the scope of this paper; to present all available information, the three morphotypes are referred to here as the “Aldabra,” “arnoldi,” and “hololissa” tortoises, without any interpretation of the taxonomic status of these morphotypes.
METHODS
Compiling Existing Information
All available published information on distribution and numbers or giant tortoises in Seychelles (current and past) was compiled. In addition, verbal inquiries regarding personal observations on past and current distribution and numbers on outer islands were made through discussion with island managers and island workers.
Capture–Recapture Survey on Frégate
Study sites on Frégate Island were stratified into zones according to vegetation type and sampled randomly thereafter. A total of 37 plots, each covering 50m2, were searched for tortoises. An average of 12% of each of the island's 12 different vegetation types (as defined by Hill 2002) was sampled. Two separate surveys were conducted over a two-week period. Temporary paint markings were applied to all tortoises encountered in the initial survey to obtain the recapture figures on the second survey. The total population on Frégate was estimated using the unbiased Lincoln–Petersen index for the full sample, with the survey being conducted over a short time period and in a closed population; therefore, mortality can be discounted, and complex population modeling is unnecessary.
Modeling of Carrying Capacity
Tortoise populations were modeled based on the published model of the Aldabran tortoise population (Gerlach 2001), which produces a model population close to the recorded census data. This model is based on the formula:
where K = carrying capacity (individuals per hectare), taken as the maximum number of tortoises recorded in each major habitat type; i = yr group; r = ln(λ + 1); λ = annual rate of production of female offspring per female; N = number of individuals; s = survival from i – 1; t = yr.
Based on published relationships between reproductive parameters, habitat, rainfall, and population density (listed in Gerlach 2001), the components are specified as: λ = ln(R(8.69 – 2.60 N/A + 0.05[N/A]2); R = rainfall (mm per year); s1–5 = 0.2284 – (0.0062 N/A); A = area (hectares); s5–10 = 0.75; s10–100 = 0.997; K is based on productivity (modified from Gerlach 2001) of the five most abundant food plants in up to five main habitat types:
where E = evapotranspiration (mm per year); Vp = percentage cover of each plant species; and Yp = productivity of each plant species in g/m/yr.
Sea-Level Rise Predictions
The above model was used to predict population size estimates of giant tortoises following a reduction in plateau and low-lying areas, as a result of projected sea-level rise attributable to climatic change. However, projected reduction in rainfall that may occur as a result of climatic change was not built into the model to ascertain its impact on giant tortoises.
RESULTS
Historical Distribution and Observed Changes
Historical records of giant tortoises in the Seychelles Islands are divided into three time periods for convenience (Tables 1–3). The 1609–1850 period covers the time between the first descriptions of the islands to the approximate date of extinction of wild tortoises on the granitic islands (reported to still exist in the wild on Mahé and Silhouette Islands in 1848). Giant tortoises were recorded by the first visitors to the Seychelles Islands. However, in many cases, descriptions are highly ambiguous and in some cases may refer to marine turtles rather than to tortoises. This issue was discussed by Stoddart and Peake (1979), and their conclusions are followed here, with some modifications (see Table 1). It is difficult to ascertain the original distribution of different forms of tortoises in the islands, because early accounts give only superficial descriptions, and collected specimens were rarely labelled with any accuracy. Subfossil remains are scarce and highly fragmentary. Elongate remains attributable to the arnoldi morphotype have been found on Mahé, and thin shelled bones possibly referable to this form have been collected on North Island (Ile du Nord), whereas honeycombed remains characteristic of the hololissa morphotypes have been found on Silhouette. Pleistocene remains from Aldabra all appear to be very similar to the form currently present on the atoll.
The years from 1850–1990 cover a period when tortoises were largely extinct in their original range but were widely translocated (Table 2). Unfortunately, there is very little information on these movements until the late 1970s. Thus, the origins of most captive and free-ranging tortoises are unknown. The years from1990–2012 cover data from the recent past and are summarized in Table 3.
In the granitic islands, tortoises were recorded on 23 islands during the first time period (1609–1850), but were subsequently extirpated from all but Ile aux Cerfs. During the second period (1850–1990), wild tortoises were reintroduced to 8 islands (Curieuse, Frégate, Ile du Nord, Grande Soeur, Moyenne, Cousin, Cousine, and Silhouette). However, 15 granitic islands (Mahé, Praslin, La Digue, Félicité, Marianne, Petite Sœur, Aride, Conception, Thérèse, Ronde, Longue, Anonyme, Ile du Sud-Est [now linked to Mahé], Ile aux Récifs, and Mamelles), which used to have wild giant tortoises, no longer have them.
In the coralline islands, tortoises were only known to be definitely present on four islands (Aldabra, Astove, Bird, and Denis) prior to human colonization. Early extinction followed on Astove, Bird, and Denis; after 1850, the Aldabra population was dramatically reduced. During the 20th century, tortoises were transferred to six coralline islands from which tortoises were definitely not recorded in the historical past and that should be regarded as introductions (D'Arros, Desroches, Rémire, Alphonse, Farquhar, and Providence) and to two islands where they had been present but possibly only in the Pleistocene (Assomption and Cosmolédo). Additional information on the main populations is given below.
Ile aux Cerfs
This island was proposed as a location for a depot of tortoises collected from the granitic islands in 1787 (Malavois in Fauvel 1909). It appears that tortoises have been present on the island continuously, and at times, the island has probably supported a mixture of tortoises originally found on the island, others imported from other granitic islands, and recent imports from Aldabra. The tortoises currently on the island are all owned by island residents, but many are free-ranging. A survey of the free-ranging tortoises in 1996 recorded 7 tortoises, of which 4 males were ascribed to the hololissa morphotype and the others to the Aldabran type. Of the hololissa morphotypes, one juvenile was moved to captive-breeding enclosures on Silhouette Island in 1997, and two adults were exported to Cousine in 2001 (one via Ste Anne and then Curieuse in 1999) in 1999–2000. The last female recorded breeding was exported to the United States in 1986, and the last wild hatching was recorded on 25 March 1986 (W. Gardner, pers. comm.). There are no reports of subsequent breeding.
Silhouette
In 1990, 5 tortoises were maintained in captivity, 4 by the Islands Development Company (IDC), probably imported from Aldabra in the late 1980s, and one privately owned, originating from North Island. In 1998, these were moved to a new enclosure and breeding started in 2008. In 2009, all the IDC tortoises were transferred to Desroches Island. In 1997, 6 arnoldi morphotypes and 6 hololissa morphotypes were brought from Mahé, Ile aux Cerfs, and Praslin for captive breeding in the Giant Tortoise Conservation Project of the Nature Protection Trust of Seychelles. This project aimed to breed these tortoises and reintroduce them into the wild on Silhouette and received the approval of the Ministry of Environment and the support of IDC in 1997. By 2010, this had produced 200 juveniles. Five adult arnoldi were released in the Grande Barbe area in 2006 with the support of IDC and the lessee of the area and appear to be still present in 2012. In 2010, plans were developed for the release of juveniles at three locations on the island, but these were rejected by IDC and the Department of Environment. In 2011, NPTS was given notice to vacate its premises on Silhouette and, as a consequence, had to stop its conservation projects on this island (see Gerlach 2011a). The tortoises were removed from Silhouette and transferred to North, Cousine, and Frégate Islands.
North Island (Ile du Nord)
Tortoises were present in the late 20th century. The date of introduction and their origins are unknown; all observed animals were typical Aldabran forms. At the purchase of the island by Wilderness Safaris in 1997, 3 to 5 animals were apparently found already present on the island, including the 2 largest males still present on the island; the identities of the other 1–3 animals are unknown. Wilderness Safaris subsequently invited Seychellois to donate animals to the island in 2003 and received donations from Mahé. These animals (presumably about 12–15) were shipped in and released in the wild. Reproduction may have been limited because of predation by black rats and feral cats, although it is unlikely that these long-term captive tortoises in poor condition would have bred for at least 5–10 yrs after release. In August 2005, during rat eradication efforts by aerial raticide drops, 14 tortoises were rounded up and placed in a pen. At least 4 others were later located in the hills. Further releases of captive tortoises were made in 2005 from Mahé (1 juvenile) and in 2007 from Anonyme (5 adults and 2 immatures). During a census held at the end of 2008, a total of 32 tortoises were found (all of the 22 described above and a further 10 animals). The majority of the animals were seen on the plateau and in the forest, but 6 remained in the hills permanently or over extended periods. In December 2005 to March 2006, 11 baby tortoises estimated to be about a month old were found on the eastern plateau. These were reared in captivity and released on the eastern side of the island from October 2008 to September 2009. Subsequently, some nests were found to contain infertile eggs, and new juveniles were discovered in November 2009 (two very small animals, one in the forest and the other at West plateau, the latter indicating that the Anonyme tortoises probably successfully bred, because only a few females of the original population were seen traversing the island from east to west). Further juveniles were found in December 2009 (a very small blind animal found on west plateau, which subsequently died), August 2010 (very small animal high up on hills), and November 2010 (one approximately 1 yr old and one new hatchling). At the beginning of 2011, the population stood at 33 adults, 3 subadults, and some 15–20 juveniles. In February and March 2011, 38 juveniles of the arnoldi morphotype were released on North Island; these had been captive bred by NPTS on Silhouette Island.
Curieuse
In 1978, 95 adult tortoises were transferred from Aldabra to Curieuse in a conservation release. The aim of this was to establish a wild breeding population in the granitic islands, to provide a tourist attraction, and a source of animals for trade. Further releases were made in 1980 (78 animals) and 1982 (74 animals), giving a total of 248. A further 3 tortoises of unknown origin were transferred from Aride to Curieuse in 1983 and in 2000; 22 tortoises from Ste Anne Island were released on the island (all long-term captives of unknown origin from other islands, and one hololissa morphotype originally from Ile aux Cerfs [subsequently moved to Cousine in 2002]). During the 1980s, high levels of loss were apparent attributable to poaching; in 1990, only 111 adults were located (44% of the originally released animals) and 94 in 1992 (Hambler 1994). The 1990 census recorded 73 males, 38 females, and 6 Curieuse-bred juveniles. Censuses of the population were made by searching all areas and recording the presence of individual tortoises, all of which were permanently tagged when released on the island. Censuses are not fully comparable because they used different methods and varied in effort. However, the results show a consistent pattern of decline through the 1980s (as summarized in Hambler 1994). Subsequently, poaching of adults appears to have ceased, although theft of juveniles has continued.
Reproduction has been recorded in this population since 1980; however, only one Curieuse-bred juvenile survived in the wild from this time. It is assumed that the majority of these lost juveniles were stolen from the island. Head-starting was initiated in 1992 with hatchlings being collected and reared until at least 5 yrs of age. Problems with enclosure security led to frequent theft from the head-starting facility. In 1992, 150 adults and 12 juveniles were reported to be present (Schramm 1992, cited in Ebersbach 2001).
Cousine
Tortoises were introduced to Cousine Island in 1993 with long-term captives from Mahé and Praslin. In 1993–2010, 32 tortoises had been introduced, comprising 29 Aldabran morphotypes (13 female, 15 male, 1 juvenile), one female arnoldi morphotype, and two males of the hololissa morphotype. These last two were introduced from Ile aux Cerfs (one via Curieuse). Several nests have been recorded, but there has been no successful hatching to date. The juvenile was found dead in 2001 (aged 7 yrs). In April 2011, 42 juveniles and 6 adults of the hololissa morphotype were released on Cousine from the NPTS captive breeding project on Silhouette Island.
Frégate
In 1954, a total of 31 tortoises were present on the island (Honegger 1966). They were reputedly introduced from Aldabra sometime during the 20th century as a conservation measure for the species (Hambler 1994). Prior to this date, there is no documented evidence that they occurred on the island. Reproduction has been observed in this population since 1990. Hatchlings have been collected and reared since 2001 with the initiation of the Frégate Island Tortoise Conservation Project. A fund was set up in early 2001, which assists in supporting the protection of the island's juvenile tortoise population. An adoption program exists in conjunction with the project, which brings in additional support for the fund. Juveniles are reared in captivity and released when they reach 3 yrs of age. The first 15 tortoises were released in 2002, and a total of 240 juveniles were collected in the same year. By the end of 2005, there were a total of 362 juveniles in captivity. Releases have taken place on average twice a year. The total release number up to now has not been recorded. Historical census and monitoring data prior to 2010 is unreliable and is based purely on nonmethodical observations or estimations. In 2004, an attempted individual marking system was initiated over a 6-mo period; 402 captive juveniles were notch-marked prior to their release the same year.
A total of 29 of the captive juveniles were released into the free-roaming population in April 2010. Additions and removals from the free-roaming population into captivity and vice versa are now strictly monitored and managed. Poaching has not been a problem on the island since the initiation of the project in 2001. The primary objective of the Frégate Island Tortoise Project is conservation and safe-guarding of the species. A secondary objective is to provide a tourist attraction for guests, which ultimately also leads to an opportunity to educate and create awareness for the species.
A census was initiated in July 2010 over a two week period using mark–recapture methods. This estimated the population at 191 ± 10.8 (Table 4). Population size could only be evaluated reliably using the entire dataset; estimating population densities in different habitats was unreliable because of uneven patterns of movement between habitats (Table 5). The habitats with the greatest levels of migration were native woodland and exotic scrub. Individuals were categorized into age classes from A to E (A = <1 yr, B = 1–5 yrs, C = >5–20 yrs, D = >0–100 yrs, and E = >100 yrs), estimated by size. Results (Fig. 2) indicated that 1% of the free roaming population is <1 yr old; 10% is 1–5 yrs old; 38% is >5–20 yrs; 38% is >20–100 yrs old; and 13% is >100 yrs old. The very low proportion of juveniles is probably inaccurate because of sampling bias toward older, more conspicuous animals. There is a discrepancy in the proportion of males to females in the older age classes. In age class C, the sex ratio was 66∶34 (females to males); in age class D, it was 40∶60; and in age class E, 100% of the individuals were male. However, the proportions of males and females in these age categories should be treated cautiously because the smaller size of females makes direct comparison difficult and may generate a male bias for older categories and a female bias for younger ones. The combined sex ratio of tortoises (>5 yrs old), marked in the initial survey, was 45∶55 (females to males). At the time of the survey, there were a total of 177 juveniles in captivity between the ages of 1 wk and 3 yrs old. In May 2011, 92 juveniles of the arnoldi morphotype were released into the head-starting enclosure on Frégate; these had been captive bred by NPTS on Silhouette Island.
Citation: Chelonian Conservation and Biology 12, 1; 10.2744/CCB-0902.1
Aldabra
Tortoises were reported to be highly abundant in 1744 (Dalrymple, quoted in Günther 1877) but “very scarce” in 1879 (Wharton 1879), “more than 1000” in 1892 (Riseley-Griffith 1892), or considerably fewer (Abbott 1893). Other accounts from the time appear to support the view that they were very scarce, except in the southeast of the atoll (Roberts 1905; Nicoll 1908; Fryer 1911). Later, Dupont (1929) reported “a great many.” The first attempt at a census of the Aldabra population was conducted in 1969 when Grubb (1971) estimated that there were 100,000 tortoises on Grande Terre.
The most reliable census was conducted in 1973–1974 using semipermanent marking of 7882 adult or semi-adult tortoises and 292 1-ha census quadrats (Bourn et al. 1999). This survey was repeated in 1997 using a subset of 157 quadrats. These detailed surveys estimated 129,415 tortoises in 1973–1974 and 100,473 in 1997. Significant declines were recorded in the areas with the highest population densities in 1973–1974 but no significant change, or even increases, in other areas. These findings were in accordance with expectations that tortoise populations in parts of the atoll had reached a carrying capacity and were in decline (up to 30–40% in 24 yrs), whereas others were still in a phase of increase, through recruitment or migration. In addition to the quadrat surveys, transects have been monitored intermittently from 1981; results of these are partially confounded by observer differences and seasonal effects. However, they also show a similar pattern of decline in high-population density areas from 1981 to 1997 (Bourn et al. 1999). An evaluation of transect data in 1997–2000 noted nonsignificant increases on Malabar, a decline in one transect and an increase in another on Picard, and nonsignificant declines on Grande Terre except in the southeast where the declines continued to be highly significant (Bourn 2001). Habitat-based comparisons (data in Bourn and Goodridge 2004) have suggested that the declines on Grande Terre have leveled out, with no significant declines apparent in most habitats since 2000. On Malabar, populations continue to increase, although this is only strong in coastal rock-based habitat and open-mixed scrub. Since 1999, all habitats on Picard have shown increases. Changes in tortoise distribution are also apparent (Fig. 3).
Citation: Chelonian Conservation and Biology 12, 1; 10.2744/CCB-0902.1
Citation: Chelonian Conservation and Biology 12, 1; 10.2744/CCB-0902.1
Other Southern Islands
No tortoises appear to have ever been reported from Saint-Pierre and Marie-Louise (A. Constance and G. Savy, pers. comm.), nor from Poivre, although there may have been some before the 1980s (A. Constance, pers. comm.).
Distribution Outside of Seychelles
Aldabra giant tortoises have been established in the wild on non-Seychelles Islands in the western Indian Ocean since the mid-1900s when tortoises were released on Flat Island, Mauritius. This population bred successfully but was reduced by poaching. By the 1950s, none remained on the island (Cheke and Hume 2008). Tortoises were released on Changuu Island, Zanzibar, in 1919 from Seychelles. By 1993, the population had fallen from some 200 animals to only 50, with a further fall to 9 in 1996. Attempts to increase the population in 1996 failed until better protection was put in place. By 2000, there were 17 adults and 140 juveniles present (Lever 2003).
In Mauritius, the replacement of extinct giant tortoises (Cylindraspis spp.) with Aldabran tortoises as “ecological replacements” has been proposed (Griffiths et al. 2010), and an ambitious project to restore tortoises to the ecosystem of part of Rodrigues Island has released two species of tortoises in 2007 at the Reserve François Leguat, at Anse Quitor. Aldabra giant tortoises are being used as replacements for Cylindraspis vosmerai and radiated tortoises (Astrochelys radiata) from Madagascar for Cylindraspis peltastes. This project currently has over 400 giant tortoises roaming over 19 ha (A. André, pers. comm.).
Experimental releases have been made on Isle aux Aigrettes, with the release of 18 giant tortoises in 2000 with some breeding recorded (Hansen et al. 2010). A similar release took place on Round Island in 2007 (Hansen et al. 2010). Hundreds of additional tortoises have been established at Réserve de la Vanille, and more releases are planned in two forested areas of Mauritius totaling approximately 20 ha, where vegetation is currently being restored through control of invasive plants and replanting with native and rare endemic trees (O. Griffiths, pers. comm.). A similar restoration is now planned in Madagascar (Griffiths, in press).
Future Prospects
Changes in populations are probable in the future with declines attributable to habitat changes and the possibility of establishment of new wild populations. Negative impacts are expected mainly from erosion and submersion of low-lying plateau areas, which represent high-quality habitat that can support high densities of tortoises and where most favorable soils for nesting are found. These changes are considered for all wild populations and for islands that have the potential to support tortoises in the future. For each island, the impacts of probable sea-level rise are considered by using the population model with reductions in land area (Table 6), using a sea-level rise estimate of approximately 100 cm by 2100 (IPCC 2007 estimated a rise in the range of 18–59 cm, but more recently this was concluded to be unrealistically low and probably half the true value; Allinson et al. 2009).
Moyenne
This is a small but high island with virtually no plateau that will be little affected by sea-level rise. However, the island has very limited nesting habitat. It is estimated that only 50–100 adult tortoises could be supported in the long term, with little breeding likely.
Ile aux Cerfs
Most tortoise habitat on Ile aux Cerfs is at sea level and may be affected by sea-level rise. The island could support more tortoises than it does at present, but current high rates of development on the island make this unlikely.
Silhouette
Reintroduction of the arnoldi and hololissa morphotypes was proposed for 3 locations around the island that could support at least 580 adult tortoises, most (about 300) at Grand Barbe. The first 5 adults released at Grande Barbe in 2006 have adapted well to the wild environment (Gerlach and Gerlach 2011); additional releases of tortoises should now be envisaged as an effort to restore ecosystem function in certain areas of the National Park, especially with regards to plant seed dispersal. Sea-level rise is likely to increase coastal erosion in the main areas that could support tortoises, although the island is steep, and most of the land area lies above the coastal plateau.
North Island (Ile du Nord)
Breeding has started and the island should be able to support a large tortoise population. It is secure from most threats, but sea-level rise is likely to have a severe impact, especially through coastal erosion. Currently the island should be able to support 1000–2000 adult tortoises.
Curieuse
The tortoise population on Curieuse has been severely affected by poaching in the past, and this may remain a problem for juveniles. Sea-level rise is likely to reduce the extent and quality of the main area used by the tortoises but will not affect most of the island. Currently the island should be able to support 2000–2500 adult tortoises, although with projected sea-level rise this is likely to be reduced significantly.
Cousin
This island does not support a regularly breeding tortoise population at present. Sea-level rise may reduce the area of available habitat, although the plateau is raised and the main effects are likely to be increases in coastal erosion. The island is estimated to be able to support 160–240 adults.
Cousine
The tortoises on this island do not form a breeding population at present. Sea-level rise is affecting coastal erosion but may have little impact on the tortoise population because of the raised plateau. It is estimated that the island could support 160–220 adults.
Aride
This island does not support tortoises at present. Sea-level rise may reduce the area of available habitat, principally increased coastal erosion. The island could potentially support a reintroduced population of 300–400 adults.
Grande Soeur
The small number of tortoises present on Grande Soeur would be expected to increase to 300–400 adults. However, the impact of sea-level rise would be notable, because a large part of its small plateau (approximately 4 ha) is a low-lying grassy area that would be completely submerged.
Frégate
This island's growing tortoise population is secure from most threats. Sea-level rise is likely to affect only a small part of the island and would not significantly affect the area occupied by tortoises. It is estimated that the island could ultimately support a population of approximately 3000 adult tortoises.
Bird (Ile aux Vaches)
The small number of free-ranging tortoises on Bird Island does not breed at present, but the island has sufficient suitable habitat to support a small wild population (450–500 adults). However, the entire island lies less than 2 m above sea level, and projected sea-level rise makes it questionable whether the island would support tortoises in the long term.
D'Arros
Currently a small number of free-ranging tortoises are present on D'Arros. These do breed, and the island has sufficient suitable habitat to support a small wild population (250–500 adults). However, the entire island lies less than 2 m above sea level, and projected sea-level rise makes it questionable whether the island would support tortoises in the long term.
Desroches
There are both captive and some free-ranging tortoises on Desroches presently. The island has sufficient suitable habitat to support a small wild population (1500–3000 adults). However, as with D'Arros, the island may not be able to support a viable tortoise population when sea-level rise starts creating significant impacts.
Alphonse
Currently a small number of free-ranging tortoises are present on Alphonse, in addition to a small number of captive juveniles. The island has sufficient suitable habitat to support a wild population of 184–303 adults. As with the previous two islands, the future of tortoises on the island following projected sea-level rise is highly doubtful.
Farquhar
At present, there is only a small population on Farquhar, but there is probably sufficient suitable habitat for a substantial population to develop (2500–4000 adults). Only a very small proportion of the atoll is >1 m above sea level, and the impacts of sea-level rise are likely to be considerable. Accordingly, it is unlikely that Farquhar will be suitable for tortoises in the future.
Cosmolédo
Currently, there does not appear to be a viable population on Cosmolédo, although there is sufficient suitable habitat on Menai and Grande Ile to support a population of 2000–4000 adults. Only a proportion of the atoll is >1 m above sea level, and the impacts of sea-level rise are likely to be considerable. Accordingly, it is unlikely that Cosmolédo will be suitable for tortoises in the future except possibly for a small area of Menai Island.
Astove
Presently, there does not appear to be a viable population on Astove, although there is a large area of suitable habitat that could support 2500–4000 adults. Most of the atoll is 2 m above sea level; projected sea-level rise is likely to reduce the land area by approximately 30%.
Assomption
At present, there does not appear to be a viable population on Assomption, although there is a large area of suitable habitat that could support 6000–8000 adults. The main area of the island is 2–3 m above sea level; projected sea-level rise is likely to reduce the land area by approximately 10% and carrying capacity by a similar proportion.
Aldabra
The Aldabran population can be considered secure from all threats except for sea-level rise. Because most of Aldabra is less than 2 m above sea level, the impact of sea-level rise is likely to be considerable. Sea-level rise of 1 m is likely to reduce the area of suitable habitat by 30% and the population to an estimated 60,000 (25% reduction). A rise of 2 m may lead to habitat reduction by 60% and population reduction to 35,000 (56% reduction). These changes would not be uniform because of the topography of the different islands and the current distribution of tortoises; Picard and Malabar would be least affected, and the Dune d'Messe area of Grande Terre would be expected to suffer the greatest declines.
DISCUSSION
Human colonization of the Seychelles led to the exploitation of giant tortoises for meat and for captive pets, resulting in their extirpation from much of their former range. This has resulted in the extinction of several endemic forms or morphotypes and has paradoxically extended the giant tortoise distribution in the outer islands. The history of tortoises in the granitic and coralline islands has been very different. In the granitic islands, strong declines from 1772 until the mid-1800s resulted in 22 islands losing tortoises completely, with only Ile aux Cerfs retaining tortoises and these being a mixture of tortoises moved from different islands at different times. Since then, tortoises have been reintroduced to eight islands, with a total wild population in the granitics of only 500–550 adults.
In the coralline islands, giant tortoises became extinct on the three islands where they had been recorded prior to human colonization, apart from Aldabra, where the population was dramatically reduced. Recovery on Aldabra was followed by widespread introduction in both granitic and coralline islands, using a mixture of Aldabran and captive stock, resulting in Aldabra tortoises being moved to a total of 9 outer islands and atolls. The wild population in the coralline islands remains over 100,000, largely attributable to the population on Aldabra.
Many of the introductions from Aldabra in the 1970s and 1980s have had very limited success, probably attributable to poaching. Where poaching has been controlled (e.g., Frégate in recent years, North Island, and D'Arros), there is evidence of rapid population growth. This indicates that giant tortoises should be able to repopulate islands with suitable habitat if steps are taken to prevent poaching or if introductions are made to localities with difficult access.
In the next 100 or 200 yrs, severe negative impacts from sea-level rise are expected from climate change in all coralline islands, where a significant part of the suitable habitat is likely to disappear. The projected reductions obtained through our model should be taken as orders of magnitude; however, these may well be conservative because the model does not take into account the predicted reduction in rainfall linked to climate change or possible indirect or cumulative effects. Most of the low-lying islands, including Bird, Denis, and most of the Amirantes, are unlikely to be able to sustain viable tortoise populations in the long term. The southern islands, including Providence, Farquhar, and Cosmolédo, would also experience a drastic reduction of their potential carrying capacity. However, this would not be as extreme on the raised islands and atolls of Assomption, Astove, and Aldabra. In the granitic islands, despite a certain reduction in carrying capacity, no dramatic declines would be expected.
The projected overall decline in population estimates of 40–65% over the next 100 yrs affects the conservation status of this species. It is currently considered to be Vulnerable on the IUCN Red List on the basis of restricted range (criterion D2). With a generation time of approximately 50 yrs (using Red List definition of a generation length equal to the average age of breeding animals, estimated here at 50 yrs), declines should be considered over a 100-yr period. This would qualify the species for Vulnerable status on criteria A3a; a decline of 50% would make the species Endangered on the same criteria. Restoration of parts of the natural range and the habitats currently occupied could allow these declines to be partially offset, reducing the decline to 20% with a sea-level rise of 1 m (50% with 2 m).
Declines attributable to sea-level rise in the granitic islands could be compensated by the establishment of new wild populations in small islands where rehabilitation activities are currently being conducted or are planned (e.g., Grande Soeur, Conception, Félicité, Marianne, Thérèse, etc.). Despite the significant numbers that still exist in captivity on the islands of Mahé, Praslin, and La Digue, free-ranging tortoises in the granitic islands are now too scarce for these megaherbivores to fulfill their former role in the islands' ecosystems (herbivory, seed dispersal, manuring, attracting decomposing, affecting soil fertility through nutrient composition, etc.). On Mauritius, the passing of seeds through the digestive system of Aldabra giant tortoises was found to significantly improve germination for the Mauritian ebony Diospyros egrettarum, a rare endemic tree (Moolna 2007). Aldabra giant tortoises are currently being used as taxon substitutes in several restoration programs in Mauritius and Rodrigues, as an attempt to replace extinct tortoises and restore ecosystem function, especially with regard to plant seed dispersal (Hansen et al. 2008, 2010; Griffith et al. 2010). The initial results from these releases indicate that these tortoises are selectively grazing problem invasive plants species and may provide a positive contribution to invasive plant management in the future (Hansen et al. 2010). The Aldabra giant tortoise appears to be an ideal species for the “rewilding” of island ecosystems, because it can be fairly rapidly bred, easily translocated, and is highly adaptable. Island conservation initiatives in Seychelles should include reintroduction of this species as an important contribution to restoring ecosystem function and conservation of one of the most iconic species of the region.
Seychelles Islands mentioned in the text. (a) All islands, (b) granitic islands. Islands marked in black have currently wild tortoise populations; islands marked in grey have historical records only.
Approximate age structure of tortoises on Frégate.
Recorded tortoise distribution on Aldabra in 1997 and 2000. Dark grey—1997 distribution; black—new areas occupied by 2000 (J.G., pers. obs.); light gray—not occupied.
Modeled and real population data from Aldabra.
