Understanding the larval dispersal patterns of exploited marine species is important to fisheries managers as it informs the spatial scales at which these harvested populations should be managed (e.g. Sale et al., 2005; Pelc et al., 2010; Jones, 2015). From an ecological perspective, when mean larval dispersal distances are short (i.e. tens of kilometers, e.g. Almany et al., 2007) and a large proportion of larvae are locally retained, local marine reserves can be self-sustaining, and nearby fisheries can receive larval subsidy. From the perspective of economic and social theory, local larval dispersal makes small-scale management more viable (Almany et al., 2007; Hamilton et al., 2021), and creates the conditions for self-sustaining territorial user-rights fisheries – in particular, it means that the community who manage their fisheries are the primary beneficiary of sustainable behavior (Costello et al., 2015). Such management becomes much more challenging where mean larvae dispersal distances are large (i.e. in the hundreds of kilometres, e.g. Nanninga et al., 2015; Herrera et al., 2016; Williamson et al., 2016). Where larval dispersal scales substantially exceed management scales, connectivity creates economic and social externalities.

Our understanding of the scales and patterns of population connectivity in marine species has increased markedly in recent decades. Empirical field studies based on new methodologies such as genetic parentage analysis have allowed direct observations of larval dispersal events (Jones et al., 2005). These studies have assisted in the validation and calibration of increasingly sophisticated biophysical dispersal simulations (Bode et al., 2019). However, most experimental studies on larvae dispersal have focused on small fish species that are not harvested (e.g., Planes et al., 2009; Pusack et al., 2014; Shulzitski et al., 2016). Literature pertaining to large commercial species (i.e., reef fish) is gradually increasingly (e.g., Harrison et al., 2012; Almany et al., 2013; Hamilton et al., 2021). Yet despite the global interest in managing sea cucumber fisheries, we know of only one published study that has attempted to quantify sea cucumber larval dispersal (Brown et al., 2022). The study by Brown et al. (2022) involved conducting a high-resolution genomic audit of sandfish (Holothuria scabra) populations at six sites in Fiji that were separated by a maximum of 300 km. Brown et al. (2022) could not infer larval dispersal patterns of sandfish in Fiji, as they did not identify any parent-offspring relationships, however they did detect some full sibling and half sibling relationships, and concluded that sandfish in Fiji are made up of three genetically distinct populations.

Sea cucumbers, or bêche-de-mer, as they are known in their dried and processed form, have been a luxury sea food in Asia for centuries (Toral-Granda et al., 2008). Many commercially valuable sea cucumbers occur in shallow tropical waters, where they are typically exploited by small-scale fisheries that operate in low income and low governance contexts (Barclay et al., 2019). Consequently, many sea cucumber fisheries are severely overfished (Toral-Granda et al., 2008; Friedman et al., 2011), with patterns of boom and bust inherent in sea cucumber fisheries globally. In the Pacific Islands region both bottom-up and top-down management efforts have been utilized to manage these high value fisheries. Coastal communities in Fiji, Papua New Guinea and Solomon Islands have established locally managed marine areas (LMMAs) on their customarily owned reefs to allow populations of sea cucumbers to recover (e.g. PNG National Fisheries Authority, 2007; Hamilton and Lokani, 2011; Hamilton et al., 2015; Lalavanua et al., 2017). This type of community-based fisheries management is made possible by the existence of customary marine tenure systems that are common throughout the Pacific, often involving communities closing an area to some or all types of fishing for a certain period (Ruddle, 1996; Johannes, 2002; Almany et al., 2015). National fishery agencies also routinely place moratoriums on these fisheries once they are considered over-exploited at a regional scale (e.g. Christensen, 2011; Hair et al., 2019; Brown et al., 2022).

For the past two decades, the Titan communities, located along the south coast of Manus Island in Papua New Guinea, have established LMMAs within their traditional fishing grounds as a fisheries management tool for a variety of harvested species (e.g. Hamilton et al., 2005). To evaluate the effectiveness of these LMMAs, Titan fishers and staff from James Cook University and The Nature Conservancy investigated the dispersal of squaretail grouper (Plectropomus areolatus) larvae that were spawned at a fish spawning aggregation that is protected by one of these LMMAs (Almany et al., 2010, Almany et al., 2013, Almany et al., 2015). This work showed that P. areolatus had a mean larval dispersal distance of 14 km, with 20% of larvae retained within the fishing grounds of where they spawned (Almany et al., 2013).

When the findings on P. areolatus connectivity were shared with Titan fishers in 2011, communities realized that while they would obtain some direct larval subsidies from protecting fish spawning aggregations located within their fishing grounds, the scale of larval connectivity meant that this fishery would be best managed at a larger spatial scale. This information provided Titan leaders with the ecological rationale for establishing the Mwanus Endras Asi Resource Development Network (MEARDN) tribal network. MEARDN encompasses a sea area of approximately 24,000 km² and 10,000 people; the Titan communities that make up MEARDN share a common language, religion, and ethnic identity. Most MEARDN inhabitants live in communities that are located along a continuous 65 km stretch of the south coast of Manus Island, with a smaller number of people residing on outer islands that lie up to 100 km south or east of Manus Island (Almany et al., 2015). MEARDN was established in 2013, with the goal of improving the livelihoods of the Titan people, by maintaining healthy marine ecosystems and strengthening the governance and financial capacity of the Titan people.

One of the most valuable marine resources that is found within MEARDN is H. scabra. Often called “sandfish”, this is the most valuable species in the Papua New Guinea sea cucumber fishery (Purcell et al., 2014; Hair et al., 2019). Sandfish have been commercially exported from Papua New Guinea since the 1870s (few local people traditionally consumed sea cucumber), but the trade accelerated dramatically in the 1980s, by which time Papua New Guinea had become the third largest supplier of bêche-de-mer to Hong Kong (Kinch et al., 2008). Historically sandfish were very abundant in nearshore seagrass habitats in Papua New Guinea, with reported densities in the 1980s-1990s often exceeding 1000 per ha (i.e. Lokani, 2001; Hamilton and Lokani, 2011). Because of its shallow distribution children, women, men, and the elderly participate in the sandfish fishery (Barclay et al., 2019). Its shallow distribution and high value have resulted in sandfish populations being depleted throughout Papua New Guinea (e.g. Dalzell, 1990; Hamilton and Lokani, 2011; Hair et al., 2019) and much of its range (Hamel et al., 2022). Given this vulnerability to overexploitation, H. scabra was listed as Endangered on the IUCN Red List in 2010 (Hamel et al., 2013). At the time of this study, 2014, a Papua New Guinea wide moratorium on sea cucumber fisheries had been in place since 2009, following low stock assessments for all commercially valuable sea cucumber species (Hair et al., 2016).

Sexes are separate in H. scabra and this species is a broadcast spawner. Their larvae are planktonic, settling in shallow seagrass habitats, approximately 14 days post-spawning (see review in Hamel et al., 2022). Adults inhabit seagrass areas and deeper sand and silt habitats. The species has a narrow habitat range, and is only found in shallow, low-energy environments that have muddy or sand substrates (see review in Hamel et al., 2022). Uthicke and Benzie (2001) found restricted gene flow between H. scabra populations along the north-east coast of Australia, suggesting a low dispersal ability in this species. Because of their broadcast spawning behavior many closely related congeneric species of sea cucumbers can hybridize, particularly in areas where both populations are abundant (e.g. Gkafas et al., 2023). In Papua New Guinea the golden sandfish (H. lessoni) occurs in similar habits to H. scabra, although in Manus this species occurs in very low abundances (authors, personal observations). The two species are readily distinguishable based on markedly different color patterns and presence/absence of wrinkles in the body wall (Uthicke et al., 2005). H. scabra and H. lesoni have been shown to hybridize, although the genetic integrity of either holothurian species remains intact through an unknown postzygotic mechanism that may be hybrid sterility (Uthicke et al., 2005).

During a MEARDN tribal meeting in 2013, Titan leaders expressed interest in establishing a business arm within their tribal network, as a means of enhancing the financial security of their people. Titan leaders were particularly interested in sustainably harvesting, processing, and exporting H. scabra directly to Hong Kong once the national moratorium on sea cucumber harvesting was lifted. By running all aspects of the bêche-de-mer business themselves, MEARDN and its fishers hoped to capture a greater share of the profits in this fishery. To ensure the sustainability of their H. scabra fisheries, Titan leaders also discussed the merits of establishing a network of nearshore LMMAs within MEARDN, to protect local spawning stocks of H. scabra once the fishery was opened. Since nothing was known about the larval dispersal patterns of H. scabra, TNC staff who were present at the 2013 tribal meeting were asked if they would conduct research on their larval dispersal patterns within MEARDN. This request formed the premise for the research that is reported on here.

In this study, we combined genetic parentage analysis and modelling to provide the first description of larval dispersal patterns in a commercially valuable sea cucumber. We identified 15 juvenile H. scabra in the Mbuke, Mbunai, Pere, Tawi and Timonai fishing grounds that were the offspring of parents sampled from the Pere fishing grounds. A fitted dispersal kernel based on the sampling location of adult and juvenile H. scabra predicted that the mean distance of larval dispersal is 15 km. These findings imply high levels of self-recruitment within the Pere fishing grounds, as well as larval connectivity between the five MEARDN fishing grounds investigated in this study – particularly between adjacent communities ( Figure 2 ).

We found that 50% of larvae settle within 6.7 km of their parents and 95% of larvae settle within 59 km of their parents, indicating that the studied H. scabra population is reasonably demographically closed (Sale et al., 2005). This implies that the spatial scale of MEARDN, which extends across 65 km of the Manus coastline, is likely adequate for sustainably managing the stock. Establishing a network of nearshore LMMAs that protect adult spawning stock of H. scabra would be a culturally appropriate strategy for MEARDN that has merit from both an ecological and fisheries standpoint. Although demographically closed populations can be managed at restricted spatial scales that match the size of tribal networks such as MEARDN, they are more vulnerable to recruitment overfishing (Begg et al., 1999; Walters and Martell, 2004) and localized extinction at that scale. Once depleted, demographically closed populations are unlikely to be rescued by the arrival of larvae from distant intact populations (van der Meer et al., 2015; Bonin et al., 2016). Indeed, some sea cucumber fisheries that have been overfished have failed to recover, despite many decades of complete closure (Friedman et al., 2011). The limited dispersal observed in this study confers with findings from Brown et al. (2022), who found three genetically distinct populations of H. scabra across sites that were separated by a maximum of 300 km in Fiji.

Aside from limited larval dispersal, there are several additional reasons why H. scabra stocks are well-suited to being managed through the establishment of a network of LMMAs in southern Manus. Firstly, adult H. scabra habitat is common near MEARDN communities. Establishing LMMAs for sandfish that were in line-of-sight of communities makes social enforcement more feasible (McClanahan et al., 2006), and was a topic of discussion and the 2013 MEARDN meeting. Secondly, in Papua New Guinea and across the Pacific, most LMMAs are usually< 0.5 km² (Mills et al., 2010; Weeks et al., 2010). Because sea cucumbers are relatively sedentary, small LMMAs are likely to be sufficient for protecting their spawning stocks. For example, some large-bodied species such as the elephant trunkfish (Holothuria fuscopunctata) move less than 10 m on average over annual timescales (Purcell et al., 2023). Movement studies on H. scabra indicate that H. scabra will cover a total distance of 2–8 m per day and have home ranges of<0.05 km² (Lee et al., 2018).

For LMMA networks to serve their fisheries management objectives individual reserves must be large enough to protect adult spawning stocks that export larval subsidies to fished areas and nearby LMMAs (Green et al., 2015). For sea cucumbers, the configuration (size, spacing and location) of LMMAs within a network should be informed by the home ranges of benthic life stages and the dispersal patterns of pelagic larvae (Green et al., 2015). We suggest that a LMMA network for H. scabra should be made up of LMMAs ≥ 40 ha (Purcell and Kirby, 2006) and separated by no more than 15 km (the approximate distance that the majority of larvae travel from their natal sites). It is noteworthy that a study by Hamel et al. (2019) challenges the notion of that H. scabra adults have restricted home ranges. Hamel et al. (2019) describes how juveniles and adult H. scabra in aquaculture enclosures in Madagascar were able to modify their buoyancy, leading them to tumble or float at speeds much faster than they can obtain through benthic crawling. Hamel et al. (2019) termed this behavior active buoyancy adjustment, and found that in experimental trials this behavior was triggered by high conspecific density, increasing water turbidity, and decreasing salinity, with active buoyancy adjustment only observed in enclosures that were located near land. If active buoyancy adjustment is common in wild populations of H. scabra, then LMMAs would need to be large to be effective once protection leads to high densities.

The findings of this study demonstrate that H. scabra, and possibly other species of the genus Holothuria, are well suited to parentage analysis. In this study, 19 H. scabra microsatellites were successfully genotyped. Each of these microsatellites had high levels of polymorphism, which provided strong exclusion power in the parentage analysis and high confidence in assigned parent-offspring pairs. The H. scabra dispersal kernel presented here is based on a low number of assignments, 15 out of out of 827 total juvenile samples. This 2% assignment rate is relatively low in dispersal analyses based on parentage assignments (Bode et al., 2018), most of which assign between 1% and 15% of juveniles. This low juvenile assignment is a result of only sampling a small proportion of a single source population, and is reflected in the large uncertainty around the best-fit larval dispersal kernel. UVC surveys indicate that there were 24,653 adult sandfish in the Pere fishing grounds when this fieldwork was conducted ( Supplementary Table S2 ), of which only 3.1% (n = 765) were analysed for the purpose of this study.

In this study, the shallow distribution of abundant and sedentary H. scabra, combined with the assistance of knowledge local fishers, made it possible to sample 6,465 sandfish in a month. However, the human resources required to amplify all the samples were unavailable. – In particular, the costs of the technical specialists’ time to extract and genotype the samples, the cost of the necessary extraction kits, and the time to process the output from the genotyping. In detail, to extract batches of samples (e.g., 2 x 96 samples) and prepare them for fragment analysis would take 1–3 days, while processing of genotyped samples would take 3–4 days. As a consequence, we were only able to process 36.5% (n = 765) of the adult and 18.9% (n = 827) of the juvenile samples. Advances in genetic technologies may allow genotyping of a very large number of tissue samples in a small number of sequencing runs, saving months of data processing and making large-scale parentage studies on sea cucumbers more feasible.

The limited larval dispersal of H. scabra that we observed in this study provides compelling evidence that coastal communities who establish LMMAs to protect spawning stocks of sea cucumbers can directly benefit from their efforts. These LMMAs can replenish stocks within the fishing grounds they were established in, complementing national management measures such as closed seasons and size limits. Our study has also shown that although H. scabra appear to have limited larval dispersal, the animals nevertheless move across spatial scales that are larger than most communities’ traditional fishing grounds. Networks of LMMAs that span multiple fishing grounds – at the scale of MEARDN, for example – are needed to generate the greatest benefits for these fisheries. LMMAs that are protect prime adult habitat of H. scabra, are in line-of-sight of communities, are ≥ 40 ha, and are separated by other LMMAs in a network by a maximum of 15 km are likely to provide the greatest fisheries management benefits for sandfish.

The code for fitting dispersal kernels to parentage data are available at https://github.com/MikeBode/Parentage_kernel_fitting/. The cleaned datasets used for the Manus example, including the code for creating Figure 2, can be found in this repository in the archive “Manus_Holothuria_scabra.zip”. The GenBank accession number for the sequences reported in this paper are PP548095-PP548120. Additional markers are available upon request to the authors.

Ethical approval was not required for the study involving animals in accordance with the local legislation and institutional requirements because Research clearance, which included ethics clearance, was provided by the Mwanus Endras Asi Resource Development Network and the Papua New Guinea National Fisheries Authority.

PW: Project administration, Data curation, Formal analysis, Software, Validation, Writing – review & editing. DF: Data curation, Writing – review & editing, Project administration, Investigation, Methodology. MB: Formal analysis, Methodology, Software, Writing – original draft. MM: Conceptualization, Data curation, Investigation, Project administration, Writing – review & editing. HH: Data curation, Writing – review & editing, Formal analysis, Methodology. MLB: Writing – review & editing, Project administration, Resources. CM: Writing – review & editing, Data curation, Investigation. MK: Data curation, Writing – review & editing, Investigation. RH: Investigation, Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Writing – original draft.