The Spotted Softshell Turtle (Pelodiscus variegatus) has been recognised since 2019 from Vietnam and Hainan Island, China, but little information about its population status and distribution range is currently available. The species has been provisionally listed as Critically Endangered by the Turtle and Tortoise Working Group, although the status has not been officially accepted by the IUCN, due to the threats the species is facing, including habitat loss and degradation, overexploitation for food, competition with other non-native softshell turtles and pollution. To identify conservation priority sites for P. variegatus in mainland Indochina, this study combines molecular analyses and species distribution modelling. Our results show that, in Vietnam, Phong Nha-Ke Bang National Park has the largest suitable area and high probability of species occurrence, followed by Vu Quang National Park and Song Thanh and Ke Go Nature Reserves. In addition, the central provinces, from Thanh Hoa to Thua Thien Hue in Vietnam, constitute a key part of the species distribution and should be prioritised for conservation actions. According to the study’s findings, although P. variegatus is possibly found in Laos, the probability decreases sharply at the border between both countries and there is also a gap in the occurrence of wetlands, arguing for strong natural barriers. Unfortunately, to date, only part of the species potential distribution is protected, while no records are known from protected areas, highlighting the need for extended or even new reserves. To recover natural populations of the species and following the IUCN’s One Plan Approach to Conservation, breeding programmes have been established in Vietnam with a potential to expand to other facilities in the country and abroad. Once suitable sites are identified, offspring can be released into the protected areas to improve the current conservation status of this highly-threatened softshell turtle.

Cytochrome b, Laos, ND4, species distribution modelling, Vietnam

The Chinese Softshell Turtle, Pelodiscus sinensis, was described nearly 200 years ago and was believed to represent a morphologically variable, geographically widespread taxon (from the Russian Far East through the Korean Peninsula, eastern and central China to Vietnam). The Northern Chinese Softshell Turtle, P. maackii (Brandt, 1857), was described 23 years later, but was then thought to be a synonym of P. sinensis. Only 35 years ago, the populations from the northernmost part of the P. sinensis distribution range were shown to represent a distinct species, based on osteological differences (Chkhikvadze 1987).

Two additional species of this complex from central China were described in the 90s, based on morphological differences: the Hunan Softshell Turtle (P. axenaria) and the Lesser Chinese Softshell Turtle (P. parviformis) (Zhou et al. 1991; Tang 1997). Recent molecular studies confirmed that the genus Pelodiscus constitutes a species complex (Fritz et al. 2010, see also Stuckas and Fritz (2011); Yang et al. (2011); Gong et al. (2018)). Based on genetic and morphological analyses, three new taxa were described: the Spotted Softshell Turtle, P. variegatus from northern Vietnam and Hainan (China), the Horse-hoof Softshell Turtle, P. huangshanensis from southern Anhui Province of China and the Chinese Stone Slap Softshell Turtle, P. shipian from Jiangxi and Hunan Provinces of China (Farkas et al. 2019; Gong et al. 2021; Gong et al. 2022).

Thus, the Pelodiscus sinensis complex at the moment comprises seven species, with six species being distributed in China and four endemic to the country (Gong et al. 2021). The revisions have consequences on the species conservation, as taxonomic splitting implies that the range size and number of individuals decrease for each species. However, these new research results have yet to be reflected in the conservation assessment of the genus. According to the IUCN (2023), only P. sinensis is listed as Vulnerable, but the data are greatly outdated with the last update in 2000 (Asian Turtle Trade Working Group 2000). The Turtle Taxonomy Working Group (TTWG) recently evaluated conservation status of six species in the genus and provisionally categorised three species as Data Deficient (P. axenaria, P. huangshanensis and P. maackii), one as Vulnerable (P. sinensis) and two as Critically Endangered (P. parviformis and P. variegatus; TTWG 2021). The last species, P. shipian, has not been assessed by any previous study.

The exact range of the recently-described P. variegatus is still largely unknown, but historical records suggest that the species occupies lowland areas of central and northern Vietnam and parts of southern China, viz. Hainan Province (Farkas et al. 2019; TTWG 2021). Recently, Ziegler et al. (2020) investigated whether natural populations of P. variegatus still exist, since the description of this species was mostly based on historical museum specimens. To find potential members of P. variegatus, this study focused on surveys of central lowland freshwater habitats, as well as local markets, restaurants and farms in Vietnam. Individuals with the species-specific dark blotched plastron pattern were identified and subsequently genetically screened to confirm their identity. By using this approach, Ziegler et al. (2020) demonstrated that P. variegatus is still extant in the wild in Vietnam, both based on evidence from the trade and on surveys in the natural habitat. The study also showed that P. variegatus occurs in the central provinces of Thanh Hoa, Nghe An, Ha Tinh and Quang Binh, primarily in lakes with flat shores consisting of mud and soft soil, rivers in agricultural landscape and medium-sized streams in secondary forests.

As softshell turtles are common and prized as food where they occur, natural populations are threatened by local hunting, with further threats of habitat loss and competition with introduced softshell turtles (Shi et al. 2008; Le Duc et al. 2020). At the moment, there is little evidence of any viable population of the P. variegatus existing in the wild. It is, therefore, imperative to implement conservation measures in priority areas, where natural populations still likely survive. To establish conservation priorities for the species in Vietnam, the present study performed climatic niche modelling, based on comprehensive distribution data derived from detailed molecular analyses of existing and new samples collected from the country, the Emys system (emys.geo.orst.edu), the Turtle Taxonomy Working Group (TTWG 2021) and the Global Biodiversity Information Facility (GBIF).

The molecular matrix contained 1921 aligned characters. Both BI and ML analyses showed that new samples belong to Pelodiscus sinensis and P. variegatus. The former species was only moderately supported (PP = 89%, UBF = 90%), while the latter received strong support from both BI and ML (PP = 100%, UBF = 99%). In total, 61 newly-collected and four GenBank samples were identified as P. variegatus (see Suppl. material 1: fig. S1, for the full tree). The localities of the samples were used for the species distribution modelling.

The niche modelling trained with only climatic variables had a good overall performance (AUC_test = 0.79; AUC_training = 0.84, Suppl. material 1: fig. S2). The predictor with the highest explanative power was precipitation of the driest month (bio_14: 77%), followed by the maximum surface temperature of the warmest month (ED15078_bio6: 11.8%), precipitation of the warmest quarter (bio_18: 5.3%), surface temperature annual range (ED15078_bio7: 3.9%), annual mean surface temperature (ED15078_bio1: 1.4%) and annual mean precipitation (bio_12: 0.7%). Based on climatic conditions, the potential distribution covers major parts of central Vietnam, wherein the probability of occurrence sharply decreases towards Laos (Fig. 2A).

The niche modelling, trained with microhabitat variables, had a good overall performance (AUCtest = 0.73; AUCtraining = 0.80, Suppl. material 1: fig. S3). The predictor with the highest explanative power was the categorical map of wetlands (60.3%) (open water, mangroves, swamps, flood-out swamp, fens, riverine, floodplain, marshes, marshes—dryland/wetland and marshes—wet meadows), followed by annual mean NDVI (ED1514_bio1: 6.6%), annual mean EVI (ED1515_bio1: 3.6%), annual range of EVI (ED1515_bio7: 1.9%) and annual range of NDVI (ED1514_bio7: 1.7%). Density of suitable microhabitats is highest near the coast, with some suitable wetlands/water bodies in higher elevations towards Laos (Fig. 2B).

When integrating the probabilities of occurrence derived from climatic and land-cover variables, the most suitable habitats for P. variegatus are near the Vietnam coastline, where extensive freshwater wetlands exist (Fig. 2C). Rivers and other water bodies in mountainous areas are partly suitable, but they cover much less area.

Only part of the potential distribution for P. variegatus in Vietnam is protected and no known occurrence is directly located within reserves, although they are close, such as nearby Ke Go Nature Reserve and Vu Quang National Park (Fig. 2D). Table 3 has information on the relative ranking of protected areas, based on their IUCN status, the total area suitable for P. variegatus, the sum of potentially suitable sites and the average potential for the species. The Table is sorted with descending total suitable areas per Reserve.

Farkas et al. (2019) stated that, in Vietnam, most records of P. variegatus fall within the “Northeast Lowlands Subregion” of Bain and Hurley (2011). The zoogeographical affinities of Hainan are closely related to this area as well as mainland south-western China, while the southern portion of the purported range forms part of the “Central– South Vietnam Lowlands Subregion” as defined by Bain and Hurley (2011). However, our molecular data show that most distribution localities of the species occur in the north-central region of the country, except for records from Dong Mo Lake (Fig. 1). It is likely that a majority of extant populations of the species is restricted to this region in Vietnam, although it is possibly found in Laos, based on climatic niche modelling results. However, the probability based on climate decreases sharply at the border between both countries and there is also a gap in the occurrence of wetlands. Thus, the border area may represent a strong natural barrier.

Figure 1.

Trimmed phylogram, based on the Bayesian analysis. Number above and below branches of major nodes are Bayesian posterior probabilities and ML ultrafast bootstrap values, respectively. Sample highlighted in bold and orange is the paratype of Pelodiscus variegatus.

Our habitat suitability analysis predicts that two most important protected areas for P. variegatus include Phong Nha-Ke Bang and Vu Quang National Parks in Vietnam. Other protected sites with the largest suitable sizes consist of Song Thanh and Ke Go Nature Reserves and Ben En National Park (Table 3, Fig. 2). In terms of Average Probability, Hue Saola Nature Reserve receives the highest value (0.422), followed by Phong Dien Nature Reserve, Bach Ma National Park and Ke Go Nature Reserve (Table 3). The findings show that the central provinces from Thanh Hoa to Thua Thien Hue form an important part of the species distribution. Although one record in our study suggests that the species occurs in Laos (Fig. 2), its approximate field coordinates could not be used to absolutely confirm the species’ presence in the country. It is, therefore, essential to conduct field surveys at suitable sites in Laos to verify its presence or absence.

Figure 2.

Potential distribution of Pelodiscus variegatus in Vietnam based on A climate B land cover C climate and land cover and D coverage with protected areas as number of suitable grid cells.

Prior to its discovery, P. variegatus was considered part of P. parviformis. The latter species was already assessed as threatened and included on appendix II of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES). As its southern population has been split into a different species, P. variegatus, the overall population of each species becomes even smaller than previously thought. Rhodin et al. (2018) regarded P. parviformis as “Critically Endangered (CR)”. Consequently, TTWG (2021) provisionally categorised P. variegatus as CR, although the assessment has not been officially accepted by the IUCN Red List. The species was recently proposed to list as Vulnerable (VU) in the Vietnam Red Data Book, based on estimation of population reduction approximately of over 30% in the past of 30 years (Nguyen, per. comm. 2023).

Ziegler et al. (2020) found during their market and trade surveys in Vietnam that human overexploitation of softshell turtles appears massive. In addition, interbreeding events were observed to occur amongst softshell turtle species in farms. Most of the inhabited freshwater bodies and their surroundings showed signs of human encroachment, such as fishing, vegetation transformation, conversion and pollution. Thus, both in situ and ex situ conservation measures seem essential for protecting P. variegatus from extinction. Some of the investigated freshwater habitats are already located inside protected areas, such as Ben En, Phong Nha-Ke Bang and Vu Quang National Parks and Ke Go Nature Reserve in central Vietnam. Some of the protected areas are well known internationally as either special bird areas (Ke Go Nature Reserve) or with spectacular mammals, including Saola (Pseudoryx nghetinhensis) and the Large-antlered Muntjac (Muntiacus vuquangensis) (Vu Quang National Park). Improving conservation in and around the protected areas will benefit a suite of critically-endangered and endemic species, including the P. variegatus.

As a first measure, based on the genetically-identified individuals, a conservation breeding programme has been established. This is following IUCN’s One Plan Approach to Conservation, developed by the Conservation Planning Specialist Group (CPSG), which combines in situ with ex situ conservation measures for the optimum protection of a given species (Byers et al. 2013). For the build-up of the conservation breeding programme, the individuals identified as P. variegatus were transferred to the Melinh Station for Biodiversity of the Institute of Ecology and Biological Resources (IEBR), Hanoi. In the Station, located in Vinh Phuc Province, northern Vietnam, besides existing outdoor tank facilities, an exclusive softshell turtle breeding facility was constructed recently.

To maximise positive outcomes and for security reasons, viz., to extend the conservation breeding network, another group of the genetically-identified P. variegatus was provided to another softshell breeding facility in northern Vietnam. Successful breeding has already been observed in the colony and offspring are ready for release to the original habitat sites. To extend the conservation breeding programme and, thus, contribute to the build-up of a stable assurance colony and conservation breeding network, a plan has been developed to transfer a limited number of surplus offspring to other facilities in Vietnam and overseas. In addition to these ex situ conservation measures already being in place, focus should now be directed to improving in situ conservation of this beautiful, but threatened softshell turtle species.

We thank Jonathan Fong, Luca Luiselli and two anonymous reviewers for their insightful comments on the earlier version of the paper and John B. Iverson and Anders G. J. Rhodin for providing valuable distribution data.