Fishing pressure and lifespan affect the estimation of growth parameters using ELEFAN
Introduction
Sustainable exploitation of marine fishery resources requires reasonable stock assessments as the foundation of fisheries management. Fish stock assessment builds an understanding of the retrospective status of harvested fish and forecasts population dynamics in the future, which provides guidance for rational management strategies (Hilborn and Walters, 1992). Traditional stock assessment models depend on substantial data sets, including time series of catch, absolute or relative abundance indices, fishing effort, age structure (Quinn and Deriso, 1999), some of which are typically unavailable for the majority of fisheries in developing countries and most small-scale fisheries. Consequently, approximately 90% of global fish stocks lack the quality and quantity of data for formal stock assessment and, thus, have been managed without sound grounds or been unmanaged (Costello et al., 2012). These fisheries are often categorized as data-limited, and there have been considerable research efforts to develop data-limited methods to fill the gap of traditional stock assessment methods (e.g., Cope and Punt, 2009; Costello et al., 2012; Carruthers et al., 2014; Carruthers and Hordyk, 2018; Dowling et al., 2015, 2019; Newman et al., 2015).
For data-limited fisheries, length-frequency (LFQ) data are often available due to the low time and financial cost of data collection. As a result, numerous length-based methodologies have been developed to understand the dynamics of fish populations in data-poor fisheries where the collection of age data is restricted (e.g., Peterson, 1891; Pauly and David, 1981; Wetherall, 1986; Fournier et al., 1998; Hordyk et al., 2015a, 2015b; Mildenberger et al., 2017; Froese et al., 2018; Rudd and Thorson, 2018). Among them, Electronic LEngth Frequency ANalysis (ELEFAN) is a system of fish assessment procedures that estimate growth and mortality parameters using LFQ data, originally described by Pauly and David (1981). ELEFAN has been widely applied in fisheries given the availability of LFQ data and the availability of software, including the FAO-ICLARM fish stock assessment tools FiSAT (Pauly and Sparre, 1991) and FiSAT II (Gayanilo et al., 2005), ELEFAN in R (Pauly and Greenberg, 2013), and TropfishR (Mildenberger et al., 2017). This method has contributed largely to the parameter estimation of the von-Bertalanffy Growth Function (VBGF, von Bertalanffy, 1938), which serves as the basis of life-history information, especially in data-limited fisheries (Mildenberger et al., 2017; Sun et al., 2017).
Despite its widespread uses, ELEFAN is vulnerable to several uncertainties. A series of studies have explored the performance and application of ELEFAN given gear selectivity (Isaac, 1990), sampling stochasticity (Schwamborn et al., 2019), individual growth variability (Isaac, 1990; Wang et al., 2020), selection of the moving average span (Taylor and Mildenberger, 2017) and bin size (Wolff, 1989; Wang et al., 2020), and the reliability of optimization algorithms (Schwamborn et al., 2019). Despite these sources of uncertainty, this study highlights that the representativeness of LFQ samples is fundamental to the application of ELEFAN and almost all of the length-based methods (Cope and Punt, 2009; Maunder et al., 2016; Punt et al., 2013). Specifically, the LFQ data obtained from catch or landings depend gear type, selectivity, fishing patterns, and stochasticity of sampling; therefore, they may not fully represent the actual length composition of the fish stock. These flaws in the data may lead to biased parameter estimates and increase the risk of error when determining stock status (Hordyk et al., 2015b).
This study is focused on the influence of fishing pressure on length-frequency data, given the impact of fishing on marine fish species (Bianchi et al., 2000; Kuparinen and Merilä, 2007; Law, 2000; Pauly et al., 1998). Typically, fishing gear is selective and inclined to harvest larger individuals in preference to smaller ones, which has led to significant declines in the average body size and trophic levels in global fisheries (Bianchi et al., 2000; Pauly et al., 1998; Tu et al., 2018). Large individuals may become rare in over-exploited fisheries (Barnett et al., 2017; Hiyama and Kitahara, 1993), which may affect the performance of length-based methods (Hufnagl et al., 2013; Schwamborn, 2018). For example, it has been demonstrated that the accuracy of the Powell-Wetherall (P–W) plot is influenced by the removal of large individuals by fisheries (Schwamborn, 2018). A lack of large individuals in a given LFQ sample will inevitably underestimate the asymptotic length by the common Lmax approach (estimating asymptotic length from the maximum length in the sample) and lead to bias in the estimation of mortality (Hufnagl et al., 2013; Schwamborn, 2018). However, there have been few studies considering the impact of data quality on ELEFAN outcomes, except given that Schwamborn et al. (2019) indicated that the robustness of ELEFAN is poor with low-quality LFQ data and a reasonable number of large individuals is necessary to ensure accuracy, but no previous studies have examined such effects explicitly. According to the report of the Food and Agriculture Organization of the United Nations (FAO, 2020), approximately 34.2% of fish stocks in the world have been over-fished, most of which are considered to be data-limited, which highlights the concern about whether length-based methods such as ELEFAN can adequately assess the status of the stocks and support sustainability of the associated fisheries. This is especially true for long-lived species, which are more vulnerable to size-selective harvesting due to their larger sizes, slower growth rate and lower resilience (Fromentin and Fonteneau, 2001; Jørgensen et al., 2007; Neubauer et al., 2013; Uusi-Heikkilä et al., 2008).
We evaluated the performance of ELEFAN under different exploitation levels for three species with different life-history types. We aimed to clarify the role of fishing pressure in the utility of length-based methods. We used a Length-based Integrated Mixed Effects (LIME) model with a Monte Carlo approach to simulate the population dynamics of short-, medium-, and long-lived fish species under different fishing intensities. ELEFAN was implemented using the simulated LFQ data to evaluate its performance in terms of growth parameter estimation. This study may contribute to the understanding of the reliability of ELEFAN in exploited fisheries and may strengthen the confidence in implementing ELEFAN for data-limited fisheries.
Section snippets
Overview
We conducted the simulation study using an age-structured population model as the operating model, which generated the population dynamics and length-frequency (LFQ) data. The study considered three target species with different life-history characteristics in terms of growth rates and maximum age, namely (i) a short-lived species, small yellow croaker (Larimichthys polyactis), (ii) a medium-lived species, yellowtail flathead (Platycephalus endrachtensis), and (iii) a long-lived species,
Length-frequency distribution
Based on the OM, we generated the length-frequency data for each species under different scenarios. The LFQ distribution of the three tested species was substantially modified by fishing. Fig. 1 shows the length-frequency distributions for one replicate (the first three months in year 19). The maximum body lengths in the LFQ samples for all species reached their corresponding asymptotic length in the under-exploited (E = 0.2) fishing scenario. Only the short-lived species reached its asymptotic
Discussion
This study is dedicated to analyzing the performance of ELEFAN given different levels of fishing pressure. We simulated the growth variation and population dynamics of three fish species and the sampling process of trawl surveys, using a simulation-estimation analysis framework. We concluded that the outputs of ELEFAN are reliable in most cases, whereas in some species the outputs may be biased by heavy fishing. Specifically, the relative bias of the VBGF parameter estimates is insensitive to
CRediT authorship contribution statement
Kun Wang: Conceptualization, Methodology, Software, Validation, Formal analysis, Visualization, Writing - original draft, Writing - review & editing. Chongliang Zhang: Conceptualization, Writing - review & editing, Funding acquisition, Validation. Ming Sun: Writing - review & editing, Funding acquisition, Validation. Binduo Xu: Methodology, Software. Yupeng Ji: Software, Validation. Ying Xue: Validation, Data curation. Yiping Ren: Supervision, Project administration, Funding acquisition.
Declaration of competing interest
The authors report no declarations of interest.
Acknowledgments
We want to thank two anonymous reviewers for their constructive comments. André E. Punt deserves substantial thanks for his patience and careful help during the editorial process. We also thank Nathan Willse (University of Maine) for proofreading this paper. This study was supported by the by the Marine S & T Fund of Shandong Province for Pilot National Laboratory for Marine Science and Technology (Qingdao) (No. 2018SDKJ0501-2).
References (60)
- et al.
Evaluating methods for setting catch limits in data-limited fisheries
Fish. Res.
(2014) - et al.
Empirical harvest strategies for data-poor fisheries: a review of the literature
Fish. Res.
(2015) - et al.
Fishing effects and life history traits: a case study comparing tropical versus temperate tunas
Fish. Res.
(2001) - et al.
Detecting and managing fisheries-induced evolution
Trends Ecol. Evol. (Amst.)
(2007) - et al.
Growth: theory, estimation, and application in fishery stock assessment models
Fish. Res.
(2016) - et al.
Current methods for setting catch limits for data-limited fish stocks in the United States
Fish. Res.
(2015) - et al.
Assessing sources of uncertainty in length-based estimates of body growth in populations of fishes and macroinvertebrates with bootstrapped ELEFAN
Ecol. Modell.
(2019) - et al.
Reproductive cycle in reared male common Japanese conger, Conger myriaster
Aquaculture
(2004) - et al.
A behavioral perspective on fishing-induced evolution
Trends Ecol. Evol. (Amst.)
(2008) - et al.
Selecting optimal bin size to account for growth variability in Electronic LEngth Frequency ANalysis (ELEFAN)
Fish. Res.
(2020)