Introduction

Sulawesi Island, the largest landmass in the Wallacea region, is renowned for high levels of endemicity, including aquatic taxa (^{Hadiaty, 2018}; ^{von Rintelen & Cai, 2009}; ^{von Rintelen, Stelbrink, Marwoto, & Glaubrecht, 2014}). The “ancient lakes” of Sulawesi, including the Malili Lake complex in South Sulawesi and Poso Lake in Central Sulawesi, are especially rich in endemic fishes and invertebrates (^{Meisner, 2001}; ^{von Rintelen & Glaubrecht, 2006}; ^{Meixner et al., 2007}; ^{Schubart & Ng, 2008}; ^{Walter, Hogan, Haffner, & Heath, 2011}; ^{von Rintelen & Cai, 2009}; ^{Mokodongan & Yamahira, 2015}; ^{Vaillant, Bock, Haffner, & Cristescu, 2013}; ^{von Rintelen et al., 2014})

The ricefishes (Adrianichthyidae) are a family with 33 recognised (valid) species in 2013 (^{Kottelat, 2013}), a number which has since increased to at least 37 (^{Mandagi, Mokodongan, Tanaka, & Yamahira, 2018}), including the description of two new species from Sulawesi: Oryzias soerotoi (^{Mokodongan, Tanaka, & Yamahira, 2014}); Oryzias dopingdopingensis (^{Mandagi et al., 2018}). Of these 37 species, 21 ricefishes are native to Sulawesi, a recognised “hotspot” of biodiversity for this fish family (^{Parenti, 2008}; ^{Mokodongan et al., 2014}; ^{Hadiaty, 2018}; ^{Mandagi et al., 2018}). While Oryzias javanicus is widely distributed across Southeast Asia (^{Mokodongan, 2019}a), and O. celebensis (previously considered endemic to Sulawesi alone has also been reported from East Timor (^{Parenti, 2008}; ^{Hadiaty, 2018}; ^{Lumbantobing, 2019}a), the majority (19 species or 90.5 %) of the Sulawesi ricefishes are endemic to one or more of the ancient lakes and/or certain riverine systems in Sulawesi (^{Parenti, 2008}; ^{Hadiaty, 2018}; ^{Mandagi et al., 2018}; ^{Lumbantobing, 2019}b; ^{Lumbantobing, 2019}c; ^{Lumbantobing, 2019}d; ^{Mokodongan, 2019}b; ^{Mokodongan, 2019}c; ^{Mokodongan, 2019}d; ^{Mokodongan, 2019}e; ^{Mokodongan, 2019}f; ^{Mokodongan, 2019}g; ^{Mokodongan, 2019}h; ^{Mokodongan, 2019}i; ^{Mokodongan, 2019}j; ^{Mokodongan, 2019}k; ^{Mokodongan, 2019}l; ^{Mokodongan, 2019}m; ^{Mokodongan, 2019}n; ^{Mokodongan, 2019}o; ^{Mokodongan, 2019}p; ^{Mokodongan, 2019}q).

Species with such restricted natural ranges are intrinsically vulnerable to extinction (^{Walter et al., 2011}). The 2019 version of The IUCN Red List of Endangered Species™ (hereafter referred to as the IUCN Red List) provides assessments of the conservation status according to the criteria of the International Union for the Conservation of Nature (IUCN) for 36 ricefishes (Family Adrianichthyidae), including all but one (O. dopingdopingensis) of the 21 ricefishes reported from Sulawesi. One Sulawesi endemic (O. wolasi) is classed as Data Deficient (DD), meaning the data available are insufficient to assign a conservation status.

Globally, half of the assessed ricefishes are classed within one of the IUCN Red List categories indicating a significant risk of extinction; it is remarkable that 89 % of these are ricefishes endemic to Sulawesi. A third of all assessed ricefishes are classed as Least Concern (LC) including the Poso Lake endemic Xenopoecilus oophorus (synonyms Adrianichthys oophorus and Oryzias oophorus) (^{Mokodongan, 2019}m) and O. celebensis (^{Lumbantobing, 2019}a). However, concerns have been expressed regarding the threats to and urgent need for conservation of all Sulawesi endemic freshwater fishes (^{Parenti, 2011}), including X. oophorus (^{Gundo, 2010}).

Among the ancient lakes of Sulawesi, the ichthyofauna of Lake Poso is particularly rich in ricefishes with seven endemic species (^{Hadiaty, 2018}). The taxonomy of ricefishes at the genus and species levels is a matter of ongoing research (^{Mokodongan et al., 2018}; ^{Parenti, 2008}), with several synonyms in current use for some species (e.g. X. oophorus). Using the taxonomy according to ^{Kottelat (2013}), the ricefishes recorded in Poso Lake are: Oryzias nigrimas (^{Parenti, 2008}; ^{Serdiati, Arfiati, Widodo, Lelono, & Gosari, 2019}a; ^{Serdiati et al., 2020}), O. nebulosus (^{Parenti & Soeroto, 2004}; ^{Serdiati, Arfiati, Widodo, Lelono, & Toha, 2019}b), O. orthognathus (^{Parenti & Soeroto, 2004}), Xenopoecilus oophorus and X. poptae (^{Parenti, 2008}), Adrianichthys kruyti (^{Parenti, 2008}), and A. roseni^{(Parenti & Soeroto, 2004}).

Anthropogenic activities pose many direct and indirect threats to the environment in and around Poso Lake (^{Nursangaji et al., 2014}; ^{Mamondol, 2018}), Despite their small size, several ricefishes are exploited as food fish (^{Gundo, 2010}; ^{Parenti & Soeroto, 2004}), while the remainder may well be caught as by-catch in targeted (ricefish) and other fisheries. Sulawesi ricefishes are also being proposed as potential model organisms for research in fields as diverse as the study of evolution, behavioural, biological and medical research (^{Mokodongan & Yamahira, 2015}; ^{Mokodongan et al., 2018}; ^{Sari, Andriani, & Yaqin, 2018}; ^{Hilgers & Schwarzer, 2019}; ^{Sutra et al., 2019}). Seven of the eight ricefish species assessed as Near Threatened (NT) in the 2019 IUCN Red List are endemic to Sulawesi, and three to Lake Poso: O. nebulosus (^{Mokodongan, 2019}g), O. orthognathus (^{Mokodongan, 2019}h), and O. nigrimas (^{Mokodongan, 2019}n).

The purpose of this study was to evaluate the Poso endemic ricefish O. nigrimas from a holistic conservation management perspective. Based on primary survey data combined with secondary data and information, the study aimed to provide guidance on potential management measures to sustain this species.

Materials and methods

Field study site: Poso lake (Fig. 1) is situated in Poso District, Central Sulawesi Province, Indonesia (1º41’18.42”-2º18’3.41” S & 120º21’27.10”-120º51’9.28” E) at an altitude of approximately 657 m above sea level (^{Nontji, 2016}). Based on measurements made in 2007 (^{Lukman & Ridwansyah, 2009}), Poso Lake covers an area of 368.9 km² (36 890 ha), with a shoreline of 127 km, length and width of approximately 35.9 km and 15.3 km. The narrow shallow shelf around the shore reaches depths of around 5-7 m then plunges steeply to a maximum depth of ≈ 385 m, with a mean depth of 194.7 m and nearly 2m fluctuation in water level between rainy and dry seasons.

Fig. 1 Map of Poso Lake showing Oryzias nigrimas observation stations and the surrounding sub-districts (adapted from ^{Serdiati et al., 2020}). 

Data collection: Primary data were collected primarily at three survey stations around Poso Lake (Fig. 1) on a monthly basis from May 2017 to April 2018. These sites were selected based on information from local fishermen and data from previous surveys. Quantitative in situ measurements visibility (m, Secchi disc), dissolved oxygen (DO) (mg/l, Lutron DO-5510) temperature (ºC) and pH (ATC Tri-meter) were supplemented by qualitative observations on habitat condition. As no fisheries statistics were available for O. nigrimas and other endemic fishes in Poso Lake, qualitative data on O. nigrimas exploitation and trends in abundance were obtained through direct observation and key informant interviews (government officials, fishermen, local residents). Secondary data were sourced from the scientific literature and databases, unpublished data (e.g. previous survey reports by academics), and other documents available online or from government agencies.

Data analysis: Data were analysed descriptively. A synopsis of bioecological parameters was compiled for of O. nigrimas and related species, focused on aspects relevant for conservation. Current and historical data on fish populations, habitat and exploitation levels were evaluated to identify actual or potential threats and changes relevant to the conservation of Poso endemic species, especially O. nigrimas. Management measures were formulated to reduce the risk of extinction and maintain or promote recovery of O. nigrimas populations along with other endemic species in Poso Lake.

Results

Poso Lake ecosystem and O. nigrimas habitat: A synopsis of data (primary and secondary) on the habitat of O. nigrimas in Poso Lake is shown in Table 1. Table 2 presents an overview of major threats to the Poso Lake ecosystem in general and the habitat of O. nigrimas in particular.

Biology, ecology, conservation status and exploitation of O. nigrimas and other Poso ricefishes: A synopsis of data on key life history parameters for the black buntingi O. nigrimas (including data obtained during this study) and the other six Poso Lake endemic ricefishes (Table 3) reveals major gaps in current knowledge. Similar gaps are reflected in a synopsis of data on the ecology, conservation status and exploitation of Poso ricefishes (Table 4).

The fisheries targeting endemic fishes in Poso Lake were not regulated or recorded by any government agency. The field survey revealed that O. nigrimas and other Poso endemic ricefishes were among the species targeted by night-time fisheries using lights to attract phototaxis positive fish. The main light fishing gears were a variety of lift nets and gillnets; scoop nets were also used and beach seines also caught ricefishes. Lift nets generally operated 15-20 days per month, avoiding the full moon period. The O. nigrimas fishing grounds were in shallow waters close to the lake shore, and included all three survey sites (Fig. 1) where the rocky substrate is particularly suitable as O. nigrimas habitat. Fishermen and other locals consistently reported a decline in ricefish abundance with a decreasing trend in fisheries catch despite an increase in effort, indicating overfishing of O. nigrimas populations.

The mesh size of all liftnets and some gillnets observed was very small, making them intrinsically non-selective (size and species). The ricefish O. nebulosus was visually identified in the catch of fishers targeting O. nigrimas. This was confirmed by molecular (DNA) analysis (^{Serdiati et al., 2019}b). Fishing activities targeting O. nigrimas were intensive and ricefishes in all age-size classes were caught. Very few captured ricefishes were in a fit condition to survive if certain categories (e.g. juvenile fish, egg-bearing females or specific species) were released.

Invasive alien species in Poso Lake: A growing number of non-native (alien invasive) species have been reported in Poso Lake. The list of introduced species reported in Poso Lake in Table 5 is likely incomplete.

The field survey revealed that local communities do not like eating these introduced fishes; with little fishing pressure, they have become established and spread. While the origins of some alien species (Table 5) are known (e.g. releases under government projects), others are speculative. The bony-lipped bard (Osteochilus vittatus) and Java barb (Barbonymus gonionotus), called nilem and tawes locally, are thought to have been released by migrants under the government “transmigrasi” program to reduce over-population in Java (and Bali). Nilem were the most abundant fish visible during a survey on Poso anguillid eels (Ndobe and Moore, unpublished data, 2008), in an area where endemic (native) species were abundant during surveys a decade earlier (Ndobe, unpublished data, 1996-2001). Several species seem to have been introduced through escape from aquaculture facilities or through well-meaning but misguided release to the wild of ornamental fishes (^{Parenti & Soeroto, 2004}; Samliok Ndobe & Abigail Moore, unpublished data, 2008). One of the more recent and potentially most dangerous introductions for endemic fish species, the janitor fish (ikan sapu-sapu), Family Locariidae, was documented in Poso Lake in 2018 (Gundo, pers. com., 2018).

Management options for O. nigrimas. Based on the data presented above, proposed priorities, interventions and anticipated benefits are outlined in Table 6. One potential intervention is the development of captive breeding followed by restocking. The national guidelines for restocking threatened aquatic (marine and freshwater) species (^{Sadili, Haryono, Kamal, Sarmintohadi, & Ramli, 2015}) and their relevance to O. nigrimas are summarised in Table 7.

Discussion

Need for a holistic approach: The direct and indirect threats to Oryzias nigrimas (e.g. habitat degradation and loss, introduced alien species, and intensive exploitation as a locally important food fish) are complex and interrelated with each other as well as with the status of other species (native and introduced) in Poso Lake. Ensuring the survival of this species in its native (endemic) habitat therefore requires a holistic approach, crossing socio-political boundaries (e.g. between government sectors, scientific disciplines, administrative jurisdictions and community segments/actors), as well as linking aquatic and terrestrial ecosystems. Action to mitigate these threats should benefit other endemic fish and invertebrate species in Poso Lake. Proposed approaches discussed below under four thematic groupings would each benefit greatly from and should be synergised with the other approaches.

Habitat conservation and rehabilitation: Like other ancient lakes, Poso Lake was until quite recently classified as oligotrophic (^{Giesen, 1994}). Various human activities around the lake and throughout the watershed have increased both nutrient levels and sedimentation rates in the lake (^{Nursangaji et al., 2014}; ^{Nontji, 2016}; ^{Mamondol, 2018}). Increased nutrients are thought to arise from urbanisation around the Lake as well as the increased use of fertiliser on arable and plantation crops which have increasingly replaced the natural forest vegetation and waste from cage culture of introduced fishes. The loss of around 50-60 m in maximum depth over four decades (^{Lukman & Ridwansyah, 2009}) is most likely related to increasing rates of erosion in the Poso Lake watershed (^{Lukman & Ridwansyah, 2009}; ^{Nursangaji et al., 2014}; ^{Mamondol, 2018}). Erosion rates of 15-60 tonnes/ha/year have been reported in the watershed around Poso Lake, with the highest rates in areas with predominantly steep slopes and high rates of deforestation (^{Mamondol, 2018}). The data in Table 1 and Table 2 indicate an ongoing degradation of the Poso Lake ecosystem in general and O. nigrimas habitat in particular.

On the positive side, there are initiatives with potential to mitigate or even to some extent reverse environmental degradation, as highlighted in Table 6. In particular, as Poso Lake is one of 15 lakes in the national lake rehabilitation program (^{Nursangaji et al., 2014}), there are many opportunities for synergy with this program. A key recommendation is that this and other Poso Lake watershed programs should include endemic species as a priority factor. Specific measures for endemic fish conservation should be synergised with this program, including direct protection of key habitat and habitat restoration through improved watershed management and restoration.

Sustainable ricefish fisheries in Lake Poso: Data collection for fisheries purposes (Poso Lake and surrounding watershed) has concentrated on anguillid eels and introduced food fish species (e.g. ^{Makmur et al., 2011}). The 2019 IUCN Red List assessments referring to surveys in 2012 and 2017 (^{Mokodongan, 2019}c, ^{Mokodongan, 2019}g, ^{Mokodongan, 2019}h, ^{Mokodongan, 2019}m, ^{Mokodongan, 2019}n, ^{Mokodongan, 2019}p, ^{Mokodongan, 2019}q) and research on O. nigrimas and Xenopoecilus oophorus (^{Gundo, 2010}; ^{Gundo, 2015}) confirm the field survey findings that Poso endemic ricefishes in general are target and/or bycatch species in light fisheries which are of importance for local food security and livelihoods.

The biology and ecology of most Sulawesi freshwater endemic species is still poorly understood, including the Poso Lake ricefishes in general and O. oryzias in particular (Table 3, Table 4). Knowledge gaps include ecological roles and interactions in the environments in which they have evolved as well as parameters used in both traditional and ecosystem-based approaches to fisheries management. Despite these gaps, we recommend an assessment of the Poso ricefish fisheries using the Ecosystem Approach to Fisheries Management (EAFM) indicators developed primarily for marine fisheries in Indonesia (^{NWG-EAFM, 2014}; ^{Pomeroy et al., 2015}) and currently being adapted to freshwater fisheries.

As reported previously for the X. oophorus fishery (^{Gundo, 2010}; ^{Gundo, 2015}), the intensive and currently unregulated light-assisted liftnets arguably pose the most serious direct fishing-related threat to Poso Lake endemic fishes, including O. nigrimas. By their very nature, these liftnets are unselective in terms of species and size. Therefore, attempts to reduce or manage the impact of these gears on endemic fish populations need to focus on temporal and spatial aspects. Specific priorities include spatial (key feeding, spawning and nursery grounds) and temporal (peak spawning periods) protection for O. nigrimas populations. The current de facto temporal fishing limitation to darker nights, during which light fishing is most effective, could be adapted to conserve fisheries resources, in particular the phototaxis positive endemic ricefish populations in Poso Lake. November and February appear to be key spawning seasons for both O. nigrimas and X. oophorus (Table 3). Extending the non-fishing period to the spawning peaks of the ricefishes present at each fishing site should increase the number of fish reproducing before capture, and thus help to maintain stocks. Spatial restrictions could include fisheries reserves (suaka perikanan in Indonesia), to support the conservation (including sustainable use) of O. nigrimas and other endemic fishes in Poso Lake.

While data poor approaches could be adopted now, we recommend multidisciplinary studies to fill the many knowledge gaps, in particular to identify spawning periods, investigate trophic ecology and map key habitats of Poso Lake endemic species, as well as to collect data on preferred fishing grounds and monitor catch compositions. Such data would enable the identification of times and places where temporal and/or spatial restrictions could have a maximum positive impact on endemic fish resources while minimising negative impacts on subsistence and commercial fisheries.

A routine monitoring program for O. nigrimas and other Poso Lake endemic species is needed to build time series data sets as a basis for science-based adaptive and holistic management. Sustainable levels and patterns of exploitation could support local livelihoods including fisheries and other economic activities (e.g. eco-tourism) and maintaining local culture, including culinary specialties based on O. nigrimas and other endemic fishes.

The conundrum of alien species: Among the threats to the endemic Ichthyofauna of Poso Lake, introduced (alien) species are a particular concern. Government projects, misleadingly called “restocking” programs, have (often repeatedly) introduced non-native species. Several of these are known to be invasive and have negatively impacted native ichthyofauna in Indonesia and other countries, including two species of Tilapia (Oreochromis niloticus and O. mossambicus); African catfish Clarias sp.; and aggressive and omnivorous South American piranha relatives of the genera Colossoma and Piaractus (^{Cagauan, 2007}; ^{Sadili et al., 2015}; ^{Ndobe et al., 2019}).

The status of two species, the striped snakehead Channa striata (local name ikan gabus) and climbing perch Anabas testudineus (local name ikan betok), is unclear. Some regard these species as native to (at least parts) of Sulawesi (^{Kottelat, Whitten, Kartikasari, & Wirjoatmodjo, 1996}) while others consider them introduced (^{Herder et al., 2012}). Both are considered at risk from introduced species in some areas of Central Sulawesi (^{Ndobe et al., 2019}). If introduced by humans, their place in local culture and cuisine indicates the introductions must have occurred a long time ago, centuries or even millennia. Considering that all known Poso endemic fishes (including those considered critically endangered or possibly even extinct) have been found in recent decades, it is unlikely that A. testudineus or C. striata posed a significant threat to Poso ricefishes in the past. However, being predatory fishes, they could potentially accelerate the decline or impede recovery of ricefish populations severely depleted from other causes.

The national guidelines for restocking threatened aquatic (marine and freshwater) species (^{Sadili et al., 2015}, summarised in Table 7) advocate precautionary approaches. It is unfortunate that similar precautions have rarely (if ever) been followed for the introduction of non-native fishes (by government and non-government parties) across Indonesia, most of which have occurred without any scientific evaluation of the risks to native biodiversity (including locally important fisheries species) and ecosystems. This paradigm is reflected in the list of introduced (alien invasive) species in Poso Lake (Table 5). It is vital to avoid further introductions (of existing or new alien species), and serious efforts should be made to mitigate the impacts of past alien species introductions.

No reports could be found of successful eradication after breeding populations were established of the invasive species already present in Poso Lake from any other natural lakes. It seems logical that fisheries for alien species should be encouraged to control (and hopefully reduce) their abundance and environmental impact. However, it is crucial that such fisheries are operated (and regulated) in ways which minimise negative impacts on O. nigrimas and other endemic species, both direct (e.g. as by-catch or through behavioural disturbance) and indirect (e.g. through impacts on habitat, such as lacustrine vegetation and water quality).

While eradication of introduced species is unlikely to be feasible on practical grounds (and in some cases livelihood or food security considerations), the prevention of further introductions should be a priority. Government should lead the way, ceasing so-called “restocking” programs which are in fact alien species introductions, and implementing mitigation measures. Education and public awareness strategies need to be accompanied by effective surveillance and enforcement, with meaningful penalties strictly applied for any violations such as the release of alien species.

Potential contribution of aquaculture and ex-situ measures: In addition to in-situ conservation measures, ex-situ research and conservation-oriented breeding could be considered for O. nigrimas and other endemic ricefishes. The medaka O. latipes is widely-used as a model organism, and ricefishes are considered relatively easy to maintain and breed in captivity (^{Parenti & Soeroto, 2004}; ^{Hilgers & Schwarzer, 2019}). Several Sulawesi ricefishes have been successfully bred in the laboratory (^{Parenti & Soeroto, 2004}). Recent research on the food habits of O. nigrimas (^{Serdiati et al., 2019}a) provides data for the development of captive breeding for this species.

Restocking is not considered necessary at present for O. nigrimas, based on the condition of the endemic population. Furthermore, the scientific/technical basis has not yet been developed and the necessary knowledge is not available regarding historical range of O. nigrimas in Poso Lake. However, research to develop the knowledge and tools to support an eventual need for responsible restocking is highly recommended, both as “insurance” and as a means of learning and building the scientific basis and human capacity to support in-situ management. True restocking (sensu^{Sadili et al., 2015}) of native/endemic fish species should follow the guidelines in Table 7. The development of ex-situ breeding could also potentially support aquaculture as an economic activity, and reduce fisheries pressure.