Navigating Crisis: Insights into the Depletion and Recovery of Central Java’s Freshwater Eel (Anguilla spp.) Stocks

Abstract

The southern coast of Central Java Province is one of Indonesia’s largest eels (Anguilla spp.) fishing grounds. The annual eel catches in this region showed an increasing trend in 2011–2014 but then plummeted in 2019–2021. However, studies on stock status are lacking to provide an effective management plan to prevent the collapse of eel fisheries. Therefore, this study assessed the state of freshwater eels in Southern Central Java using a data-limited method, catch-based maximum sustainable yield (CMSY). The analysis revealed a swift stock depletion, shifting from a healthy state in 2011 to a concerning red zone in 2015–2018. Subsequently, the stock began a recovery phase in 2019, but the recent trajectory raised concerns as it approached the red zone again, signaling the urgency of implementing a rebuilding plan. The simulation of several rebuilding scenarios suggests that reducing fishing mortality to 80% of the F_MSY (maximum sustainable fishing mortality) can balance production and conservation objectives, achieving the quickest stock rebuilding with minimal catch loss. This study showed that eel stock could be rebuilt by reducing fishing pressure to preserve a surplus of individuals who can complete their migration and spawning cycles.

1. Introduction

The worldwide population of freshwater eels (Anguilla spp.) faces a serious threat due to extensive harvesting at different life stages. While adult eels, such as yellow- and silver-stage eels, have been consumed as foods, glass eels and elvers hold high aquacultural values [1]. In the last three to four decades, there has been a notable decline in the populations of European (A. anguilla), American (A. rostrata), and Japanese (A. japonica) eels, followed by other eel species, including Australian (A. australis) and New Zealand (A. dieffenbachia) [2]. To compensate for diminishing European and Japanese eel populations, countries such as China, Japan, Taiwan, and South Korea have turned to artificial cultivation to meet consumer demand [3]. Since artificially induced breeding techniques for eels have not been established, young wild juveniles (i.e., glass eels and elvers) captured in estuaries are used for cultivation [4]. Juvenile eels have become highly valuable commodities in aquaculture, exerting significant pressure on the natural environment owing to extensive fishing. Approximately 90% of the world’s eel supply is sourced from aquaculture, and the entire industry relies on wild juveniles [5]. This heavy dependence on wild catch for eel resources intensifies the threat to wild eel stocks, pushing their status from threatened toward further decline.

Southeast Asian tropical eels, specifically A. bicolor and A. marmorata, are targeted as alternatives to compensate for the diminishing European and Japanese eel resources in East Asia [3]. Anguillid eel capture fisheries are operational in Indonesia, the Philippines, and Vietnam, while these species are considered bycatch in Thailand [6]. In Indonesia, the focus is on capturing glass eels, elvers, and yellow eels of A. bicolor or A. marmorata during the rainy season spanning from September to December [7]. Meanwhile, the Luzon and Mindanao Islands in the Philippines primarily target A. marmorata for glass eels, elvers, and yellow eels, experiencing a notable decline in the combined annual catch from 12 tons in 2007 to approximately 0.3 tons in 2017 [8]. In Vietnam, the Ky Lo River in Phu Yen Province is a prominent site for capturing glass eels, predominantly A. marmorata (95%), and A. bicolor pacifica (5%). The peak of the fishing season occurs from November to May, recording an annual catch ranging between 0.60 and 0.75 tonnes (equivalent to 4,000,000–5,000,000 individuals) [9]. On the other hand, eel farming has also been expanding in this region. Indonesian, Filipino, and Burmese eel farmers rely on regionally sourced young eel stock for their eel farms within their respective countries, whereas Cambodian farmers import young eels from the Philippines and Indonesia [8]. Notably, large-scale farms cultivating tropical anguillid eels have been in operation, particularly on Java Island, Indonesia, since the late 2000s [10].

Cilacap, on the southern coast of Central Java Province, serves as one of the largest fishing grounds for yellow eels in Indonesia, supplying a significant portion of elvers and yellow eels to cultivation farms [9]. In this regency, elvers and yellow eels are primarily captured using scoop nets or PVC traps during the period from October to November, whereas eels intended for consumption are harvested through lift nets from April to September [7]. The annual catch of eel in this region was provided by the Marine and Fishery Department of Central Java Province (Figure 1), which was mainly recorded from three regencies: Cilacap, Kebumen, and Purworejo [11,12]. The eel fishery in this region is predominantly characterized as artisanal, with eel fishermen engaging in the sale of their catches to intermediary eel collectors. Subsequently, these collectors facilitate the distribution of the eels to various eel farms [9,13].

Despite the increase in tropical eel catches, there is a notable absence of studies on tropical eels compared to their European, American, Japanese, Australian, and New Zealand counterparts. Studies on tropical eels have been conducted on species composition, early life history, recruitment behavior of the glass eel, and the relationship between eel biology and habitat [14,15,16,17,18]. Studies on the stock’s condition, particularly regarding its exploitation level and spawning biomass status, are lacking in formulating a comprehensive management strategy. This strategy should include diverse management methods such as regulating effort, establishing a minimum mesh size for nets, and setting maximum catch limits [19,20]. The 2017 International Eel Symposium highlighted a great concern regarding the stock status of eel fisheries, particularly for tropical species that remain poorly understood [21].

Unfortunately, the available fishery data, in which only time series of landing data and some biological information exist, may not facilitate utilizing conventional methods to assess the stock status. Therefore, there is a critical need to shift from traditional stock assessment methods to more adaptable approaches that can function effectively, even with limited data [21]. Various data-limited methods have been developed for species using fish catch data [22,23,24,25,26,27], fish length data [28,29,30], and fish abundance data [31].

This study assesses the state of freshwater eels in southern Central Java, where glass and adult eels are exploited. Due to the scarcity of data, the assessment of freshwater eel stock status and exploitation rate relied on a data-limited model constructed from commercial landing data of adult eels. The evaluation of this stock serves as fundamental information for prospective strategies aiming to avert the decline of eel fisheries. Additionally, based on the knowledge of the biology and fisheries of this eel, potential strategies to manage the utilization of these resources were explored to improve the eels’ sustainability.

2. Materials and Methods

2.1. Study Area

Central Java Province was selected as the unit of analysis in this research, where eel fishing is practiced on the southern coast of Central Java, with only one species identified, A. bicolor bicolor [32]. The south coast of Central Java has many sloping estuaries where glass eels prefer to enter rivers. The Indian Ocean on southern Java Island also has many deep trenches preferred by these eels for mating and spawning [33]. Therefore, a massive exploitation of glass and adult eels has occurred in this region. Moreover, several eel aquacultures managed by fishermen have expanded rapidly to meet export demands. The eel fishing grounds in this region are spread across three watersheds: Citandui, Serayu, and Ijo [34] (Figure 2). In this region, the average annual rainfall reaches 3.444 mm/year, with monthly rainfall ranging from 734 mm during the dry season to 852 mm during the wet season, accompanied by an average monthly temperature of 26.7 °C [13]. Meanwhile, the pH conditions in this area are neutral (pH = 7), with salinity ranging between 0 and 4 ppt and dissolved oxygen (DO) levels approximately between 4 and 5.6 ppm. These water conditions are deemed ideal for the habitat of eels and fall within the normal range for glass eel migration [33].

2.2. CMSY Analysis

Stock assessment in a data-limited condition requires an approach to stock assessment and fishery management different from classical age-based models, which incorporate novel techniques for calculating fishery reference points when biological data are limited and catch/biomass time series lack [23]. Catch-only methods (COMs) are data-limited stock assessment methods that mainly use the time series data for catch or landing to estimate stock biomass status (e.g., B/B_MSY or depletion) and other common fishery reference points and quantities [35]. One of the COMs that has been widely applied to assess stock status is the catch-based maximum sustainable yield (CMSY) method, which has been used in the assessment of global [36], regional [37,38,39,40,41,42], single species [43,44,45,46,47], and multispecies fisheries [42,48,49]. CMSY uses the Monte Carlo technique to predict fishery reference points (MSY, F_MSY, and B_MSY), relative stock size (B/B_MSY), and exploitation (F/F_MSY). This estimation relies on catch data and general assumptions regarding resilience or productivity (r) and stock status (B/k) at the start and conclusion of a temporal data set [23]. The CMSY has utility for short-term estimates, except for relatively flat harvest dynamic scenarios, probably because it uses more informative prior distributions and thus has more information for a short time series [25]. Hence, the CMSY may be more appropriate for fisheries in developing countries, where data collection programs have recently been initiated so that a short-term series of catches is only available.

A sequence of yields produced from the available biomass at a given productivity level can be used as time-series catch data. When two of the three variables—yield, biomass, or productivity—are available, the third variable can be estimated. Typical production models [50] use a catch-and-abundance time series to estimate productivity. In contrast, the CMSY method uses catch and productivity to estimate biomass, exploitation rate, MSY, and related fishery reference points. A Monte Carlo approach can search for probable ranges for the maximum intrinsic rate of population increase (r) and unexploited population size or carrying capacity (k) for identifying ‘feasible’ r–k pairs.

(i)

Initial r–k range determination

The resilience of A. bicolor bicolor in the FishBase is ‘low’ and converted as ranges for initial values of r at 0.05–0.5 (

Table 1. Initial ranges for parameter r.

Resilience	Initial r Range
High	0.6–1.5
Medium	0.2–0.8
Low	0.05–0.5
Very low	0.015–0.1

). Alternatively, the preliminary estimates of r can be calculated using the empirical relationships below [23]:

$$\mathrm{r}\approx 2\mathrm{M}\approx 2{\mathrm{F}}_{\mathrm{M}\mathrm{S}\mathrm{Y}}\approx 3\mathrm{K}\approx \frac{3.3}{{\mathrm{t}}_{\mathrm{g}\mathrm{e}\mathrm{n}}}\approx \frac{9}{{\mathrm{t}}_{\mathrm{m}\mathrm{a}\mathrm{x}}}$$

(1)

where r is the intrinsic growth of population; M is the natural mortality; F_MSY is the fishing mortality that produces MSY; K is the rate of somatic growth derived from the von Bertalanffy growth equation; t_gen is the generation time; and t_max is the maximum age.

Next, the initial range for ‘k’ can be estimated using the following Equations (2) and (3) [23]:

$$k_{low}=\frac{\mathit{max}\left(C\right)}{r_{high}}; k_{high}=\frac{4\mathit{max}\left(C\right)}{r_{low}}$$

(2)

$$k_{low}=\frac{2\mathit{max}\left(C\right)}{r_{high}}; k_{high}=\frac{12\mathit{max}\left(C\right)}{r_{low}}$$

(3)

where k_low and k_high are the minimum and maximum limits of the preceding k-range; max(C) is the maximum recorded catch within the temporal data sequence; and r_low and r_high are the minimum and maximum limits of the r values, respectively, determined by CMSY. Equation (2) is utilized when the initial biomass of the stock is depleted at the end of the temporal data sequence, whereas Equation (3) is used when the biomass is elevated.

(ii)

Estimating the initial ranges of relative biomass

The initial range of the relative biomass at the start and end of the temporal data sequence was estimated based on the fishery status classification (

Table 2. Initial ranges of relative biomass as inputs [23].

Strongly Overfished	Overfished	Optimum Fished	Lightly Fished
0.01–0.2	0.01–0.4	0.2–0.6	0.4–0.8

). Based on the annual catch data (Figure 2), it was assumed that the initial biomass range relative to the unexploited biomass (B/k) at the start of the catch temporal data sequence was lightly fished, whereas it was overfished at the end.

(iii)

Estimating likely reference points using feasible r–k pairs

CMSY randomly chooses an r–k pair from within the initial ranges of r and k to identify a feasible r–k pair. Following this, the initial biomass was selected from the preceding biomass range for the first year, and then the biomass of subsequent years was projected using Equations (4) and (5).

$$B_{t+1}=B_t+r{B}_t(1-\frac{B_t}{k})+C_t$$

(4)

where B_t+1 is the harvested biomass in year t + 1; k is the carrying capacity; r is the resilience; and B_t and C_t are the biomass and catches in year t, respectively. In cases of extensive depletion, when the biomass dropped below 1/4k, the prediction formula also included a linear reduction in surplus production, outlined as follows:

$$B_{t+1}=B_t+4\frac{B_t}{k}r\left(1-\frac{B_t}{k}\right)B_t-C_t \mathrm{when} {\mathrm{B}}_{\mathrm{t}}/\mathrm{k} < 0.25$$

(5)

where 4B_t/k emulates a gradual reduction in recruitment that falls under half the biomass that could produce MSY.

The r–k combinations and projected biomass paths were deemed feasible if the estimated biomass is not less than 0.01 k and remains within the prior biomass range for the middle and concluding years [23]. CMSY derived the most likely values for ‘r’ and ‘k’ from the ranges of the feasible r–k pairs, employing the default rules of the method. The fishery reference points were calculated as MSY = rk/4, fishing mortality aligned with MSY (F_MSY) = 0.5r, and biomass-producing MSY (B_MSY) = 0.5k [51]. The R code for the CMSY analysis was obtained from http://oceanrep.geomar.de/33076/ (accessed on 3 February 2022).

2.3. Stock Status and Recovery Scenarios

The CMSY analysis provides reference points that can be used to evaluate the stock status (

Table 3. Stock status based on fishery reference points [46].

B/BMSY	F/FMSY	Stock Status
≥1	<1	Healthy stock
0.5–1.0	<1	Recovering stock
<0.5	<1	Stock beyond safe biological limits
0.5–1.0	>1	Overfished stock
0.2–0.5	>1	Stock beyond safe biological limits
<0.2	>1	Severely depleted stock

). The output of the CMSY analysis can also be used to assess future stock status under several exploitation scenarios implemented in a modified Schaefer model, as follows:

$$\frac{B_{t+1}}{B_{MSY}}=\frac{B_t}{B_{MSY}}+2 F_{MSY}\frac{B_t}{B_{MSY}}\left(1-\frac{B_t}{{2B}_{MSY}}\right)-\frac{B_t}{B_{MSY}}{F}_t when \frac{B_t}{B_{MSY}} \ge 0.5$$

(6)

$$\frac{B_{t+1}}{B_{MSY}}=\frac{B_t}{B_{MSY}}+4 F_{MSY}{\left(\frac{B_t}{B_{MSY}}\right)}^2\left(1-\frac{B_t}{{2B}_{MSY}}\right)-\frac{B_t}{B_{MSY}}{F}_t when \frac{B_t}{B_{MSY}}<0.5$$

(7)

Four exploitation scenarios were implemented as follows: (1) at 0.5 F_MSY, implying no fishing activity when the biomass falls below 0.5 B_MSY (B < 0.5); otherwise, maintaining a fishing mortality rate (F) equal to 0.5 F_MSY (F = 0.5 F_MSY). (2) Under 0.6 F_MSY, with F set at 0.6 F_MSY when B ≥ 0.5 B_MSY; however, F diminishes linearly to zero as the biomass declines below 0.5 B_MSY. The reduced fishing mortality (F_reduced) was computed using Equation (8) [38]. (3) The 0.8 FMSY exploitation scenario, where F equals 0.8 F_MSY for B ≥ 0.5 B_MSY; otherwise, F similarly diminishes linearly. (4) The 0.95 F_MSY exploitation scenario, where F remains fixed at 0.95 F_MSY.

$$F_{reduced}=2F_t \frac{B_t}{B_{MSY}}$$

(8)

Changes in stock biomass and fishing catches were forecasted and compared across four fishing pressure scenarios. The catch in 2022 marked the initial point for the forecasted path, and the fishing pressure in 2022 determined the biomass and catch estimation for 2023–2024. Resource recovery was estimated within these four exploitation scenarios. All analyses were performed in R version 4.2.2 using the CMSY forecasting methodology [38], modified for a single-species fishery. The code was downloaded from https://github.com/SISTA16/cmsy_multispecies_forecast/ (accessed on 30 October 2023).

3. Results

3.1. CMSY Analysis

The outputs of the CMSY analysis on the eel fishery in Central Java are listed in

Table 4. Fishery parameters with their 95% confidence intervals.

Parameters	Value	95% CI
r (1/year)	0.282	0.163–0.487
k (103 MT)	1.112	0.483–2.554
MSY (103 MT)	0.078	0.045–0.137
FMSY (1/year)	0.141	0.082–0.244
BMSY (103 MT)	0.556	0.242–1.277

.

Based on the above estimates, the CMSY offered management insights for eel fisheries, as outlined in

Table 5. Management information for the eel fishery.

Parameters	Value	95% CI
Biomass in the last year	0.275	0.021–0.438
B/BMSY in the last year	0.5	0.038–0.787
Fishing mortality in the last year	0.13	0.082–1.683
Exploitation F/FMSY	0.93	0.584–12.056

. Predicted MSY trends for eels suggested that sustainable harvesting rates were surpassed by 2012 and reached their highest point in 2014, which was more than 100% higher than expected (Figure 3a).

The results showed that fishing mortality in 2022 (0.13) was close to the figure that produced the MSY (F_MSY = 0.141). For F/F_MSY, a rapid upward movement was noted, crossing the sustainable threshold in 2012 and peaking in 2014 (Figure 3c). However, the exploitation (F/F_MSY) dropped to 0.93 in the final year of the dataset, which was deduced to be the effect of the global pandemic in 2020–2021.

The biomass pattern affected by fishing mortality and recruitment peaked in 2011, steadily declined until 2014, and dropped below the B_MSY threshold in 2019. This signifies a transitional phase for the stock, potentially leading to depletion if the present fishing mortality does not return to the F_MSY. The B/B_MSY path in Figure 3b indicates a steep decrease from 2014 that reaches a minimum in 2020, which is under the target reference point (B/B_MSY = 0.495).

The Kobe plot depicts the concurrent growth of B/B_MSY and F/F_MSY (Figure 3d). The plot illustrates a swift decline in stock from the green zone in 2011 to the dangerous red zone from 2015 to 2018. The stock entered the recovery phase from 2019 until the end of the dataset because of reduced fishing pressure during the global pandemic. However, the stock trajectory was worrisome, returning to the red zone, which indicates a collapsing fishery.

3.2. Stock Status and Recovery Scenarios

From the fishery reference points produced by the CMSY analysis, the eel fishery stock status was in the recovery phase (B/B_MSY = 0.5 and F/F_MSY < 1). Under the four future predictive exploitation scenarios, the eel fishery showed biomass recovery by 2031, with catches remaining below the MSY until the end of the simulation year (Figure 4). The scenario with 0.5 F_MSY exhibited the most rapid biomass recovery rate, whereas under the 0.95 F_MSY scenario, the biomass would never recover to a sustainable level. Both the 0.6 F_MSY and 0.8 F_MSY scenarios produced a moderate biomass recovery rate, and these scenarios could recover the biomass in 2032 and 2036, respectively. However, both the forecasted biomass and catches exhibited significant uncertainty.

4. Discussion

Indonesia has led Southeast Asia in capturing anguillid eel fisheries and farming activities [8]. Notably, Japanese companies import these eels from Indonesia because of the decline in temperate eel production, elevating the significance of tropical anguillid eels in the global market [7]. The annual export value of these eels to Japan averages approximately USD 2.3 million [52]. Cilacap, located in Central Java Province, is a pivotal hub for yellow eel fishing in Indonesia and supplies the majority of elvers and yellow eels to the country’s eel farms [9]. To ensure the sustainable management and utilization of freshwater eel resources, comprehensive information regarding the status and trends of eel resources is of paramount importance.

In this study, a data-limited assessment model (CMSY) was employed to evaluate the stock status of A. bicolor along the southern coast of Central Java. The findings indicate a recovery phase for this stock after years of overexploitation until 2019, notably influenced by a significant decrease in fishing efforts during the COVID-19 pandemic (2020–2021). Nevertheless, the trajectory of the stock is cause for concern, as it has reverted to the red zone. Thus, a recovery plan should be implemented to prevent eel fisheries from collapsing. Simulation outcomes suggest that reducing fishing efforts by 50–80% of the F_MSY could facilitate the rebuilding of eel stock to a sustainable level. However, scenarios 0.5 and 0.6 exhibit rapid stock rebuilding, surpassing biomass that produces the maximum sustainable yield (B_MSY) but yields lower catch levels. Despite modeling uncertainties, the scenario of lowering fishing mortality to 80% of the F_MSY strikes a balance between production and conservation objectives, achieving the quickest stock rebuilding with minimal catch loss over 13 years. However, managing fishing efforts remains one of the most challenging aspects of fishery management [53], understandably causing hesitation regarding this policy. Therefore, a potential shift for managers could be from effort control to establishing a total allowable catch (TAC) guided by a catch projection graph (Figure 4b).

Applying TAC (fishing quota) should be combined with size limitations specifically to protect spawning adults because the current study underscores that an overfished stock might experience a slow recovery, particularly evident in A. bicolor bicolor, which is unable to reach a sustainable level under the 0.95 F_MSY scenario. Contrary to the common belief that F = F_MSY offers optimal fishing pressure to rebuild fisheries [37], the results of this study challenge this notion, particularly for A. bicolor bicolor, which has a low intrinsic growth rate (r = 0.282). Moreover, freshwater eels display semelparous behavior, indicating that they undergo a single exhaustive spawning event at the end of their life cycle, making their reproductive potential highly sensitive to fishing pressure [4]. Thus, combining fishing quotas and size limitations could preserve a surplus of individuals that can complete their migration and spawning cycles [1]. In addition to fishing pressure, eel populations are also sensitive to climate change, which can significantly impact eel migration patterns, alter their access to suitable habitats, and affect their population dynamics [21]. Therefore, implementing additional management strategies beyond fishing effort reduction and fishing quotas is crucial, such as adaptable seasonal closures to align with the altered timing of eel migration due to changing environmental conditions. Habitat protection can also mitigate the impact of climate change on eel fisheries by providing stable environments for eels to reproduce and grow [54]. Integrating these approaches into fishery management is imperative to ensure the sustained recovery of eel stocks.

However, it is crucial to acknowledge that the approach and models utilized in this study rely heavily on assumptions regarding the life history traits and exploitation status of stocks. Notably, the CMSY model’s limitation in accounting for size and age structure might lead to overestimating sustainable productivity, especially in populations with disrupted structures due to excessive fishing pressure [38]. Therefore, future research could be conducted to accommodate these factors for better estimation results. Moreover, the forecasting algorithm assumes the static nature of evolving fishing strategies. In future research, this issue could be further explored using system modeling, such as agent-based and system dynamics models, to depict the complex and nonlinear dynamics inherent in fishery systems, providing a more nuanced representation of the interplay between various factors influencing stock dynamics, fishing practices, environmental variables, and policy impacts. For instance, an agent-based model can create simulated individual agents representing fishermen with diverse behaviors, decision-making processes, and interactions with the environment to explore in more detail the impact of different fishing strategies on stock dynamics [55,56]. Meanwhile, a system dynamics model can explore feedback loops, delays, and nonlinear relationships between factors such as fishing effort, environmental variables, market demand, and fishing policies to understand their potential impacts on stock and fish population dynamics [57,58].

5. Conclusions

This assessment of the status of eel fisheries along the southern coast of Central Java can serve as a guide for fishery management. The CMSY model revealed a decline in biomass despite low fishing efforts, indicating a recovery phase for this stock. Immediate protective actions are necessary to restore this stock; otherwise, its decline will be exacerbated, making recovery more challenging. Decreasing fishing pressure is likely to facilitate the restoration of stocks to a sustainable level, and greater reductions in fishing pressure would accelerate their recovery. Despite the potential years of reduced catch during the rebuilding phase, the regenerated stock should generate heightened yields and financial advantages.