Home Journals Articles Updates About
CN
image
收藏切换
Population dynamics modelling with spatial heterogeneity for yellow croaker (Larimichthys polyactis) along the coast of China
收藏切换
PDF
Qiuyun Ma1, 2, 4, Yan Jiao2, Yiping Ren1, 3, Ying Xue1, *
Acta Oceanologica Sinica | 2020, 39(10) : 107 - 119
Less
收藏切换
Acta Oceanologica Sinica | 2020, 39(10): 107-119
• Marine Biology •
Population dynamics modelling with spatial heterogeneity for yellow croaker (Larimichthys polyactis) along the coast of China
Full
Qiuyun Ma1, 2, 4, Yan Jiao2, Yiping Ren1, 3, Ying Xue1, *
Affiliations
  • 1 Fisheries College, Ocean University of China, Qingdao 266003, China
  • 2 Department of Fish and Wildlife Conservation, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, USA
  • 3 Laboratory for Marine Fisheries Science and Food Production Processes, Pilot National Laboratory for Marine Science and Technology (Qingdao), Qingdao 266237, China
  • 4 College of Marine Sciences, Shanghai Ocean University, Shanghai 201306, China
Published: 2020-10-25 doi: 10.1007/s13131-020-1602-4
Outline
收藏切换

As one of the top four commercially important species in China, yellow croaker (Larimichthys polyactis) with two geographic subpopulations, has undergone profound changes during the last several decades. It is widely comprehended that understanding its population dynamics is critically important for sustainable management of this valuable fishery in China. The only two existing population dynamics models assessed the population of yellow croaker using short time-series data, without considering geographical variations. In this study, Bayesian models with and without hierarchical subpopulation structure were developed to explore the spatial heterogeneity of the population dynamics of yellow croaker from 1968 to 2015. Alternative hypotheses were constructed to test potential temporal patterns in yellow croaker’s population dynamics. Substantial variations in population dynamics characteristics among space and time were found through this study. The population growth rate was revealed to increase since the late 1980s, and the catchability increased more than twice from 1981 to 2015. The East China Sea’s subpopulation witnesses faster growth, but suffers from higher fishing pressure than that in the Bohai Sea and Yellow Sea. The global population and two subpopulations all have high risks of overfishing and being overfished according to the MSY-based reference points in recent years. More conservative management strategies with subpopulation considerations are imperative for the fishery management of yellow croaker in China. The methodology developed in this study could also be applied to the stock assessment and fishery management of other species, especially for those species with large spatial heterogeneity data.

yellow croaker  /  population dynamics  /  Bayesian hierarchical model  /  geographic variation
Qiuyun Ma, Yan Jiao, Yiping Ren, Ying Xue. Population dynamics modelling with spatial heterogeneity for yellow croaker (Larimichthys polyactis) along the coast of China[J]. Acta Oceanologica Sinica, 2020 , 39 (10) : 107 -119 . DOI: 10.1007/s13131-020-1602-4
Full Text
Less
1 Introduction
Less
Stock assessment could provide decision-makers with suggestions on the consequences of alternative management strategies (Jensen, 2005; Punt and Hilborn, 1997). These suggestions include predictions of the reactions between stocks and fishers to diverse levels of fishing pressure, and estimating the fishing effort level to obtain the maximum sustainable yield (MSY). One of the most associated models with MSY is the surplus production model, developed by Schaefer in 1954 (Schaefer, 1954).
Surplus production models (SPM), also known as biomass dynamics models or surplus yield models, rely on an assumption that a stock could produce an excess or surplus biomass (or abundance) that can be harvested. The SPM is assumed to combine the change of the stock’s biomass (except the fishing mortality), including the individual growth in weight, the recruitment of new individuals and the natural mortality (Quinn and Deriso, 1999). SPMs can provide inferences of the relationships between fishing effort to population biomass and catch, and estimate the biological reference points, such as MSY, fishing mortality and effort to achieve MSY (Jensen, 2005). The most important merit of this kind of model is their minimum requirement of data, i.e., they can be modelled with just the time series data of catch and abundance index. These models can be applied in fisheries without substantial size-structure, biological, selectivity or other survey data. These models are suitable in cases where other assessment methods with huge amount of data requirement are not feasible.
Yellow croaker (Larimichthys polyactis) is one of the top four most important fish species in China. Its annual catch ranges between 343 000–400 000 t in recent years and brings about 540 million of dollars to Chinese fisheries economy (Bureau of Fisheries and Fishery Administration of Ministry of Agriculture, 1969–2016). However, only time series data of fishery catch and effort are available for this species with a coarse spatial and temporal resolution (i.e., yearly data of each province) in the past 48 years (Bureau of Fisheries and Fishery Administration of Ministry of Agriculture, 1969–2016). As a result, surplus production models are likely to be the most appropriate stock assessment approach of yellow croaker populations along the coast of China.
Yellow croakers are widely distributed in the Northwest Pacific Ocean, including the coastal waters of China, Korea and Japan. In China, its natural habitat extends from the Bohai Sea, Yellow Sea to the East China Sea, across 22°–41°N. The existence of geographic subpopulations for yellow croaker in China has been confirmed by previous studies based on the differences in life history, body morphometries, migration trajectories, etc. (Lin et al., 1965; Liu, 1990; Xu and Chen, 2010; Zhang et al., 2014). The hypothesis of two subpopulations, the Bohai Sea and Yellow Sea subpopulation, and the East China Sea subpopulation, is widely accepted (Xu and Chen, 2010; Zhang et al., 2014). Their migration indicates little interactions between these two subpopulations, characterized with differences in spawning, feeding and wintering grounds (Xu and Chen, 2010). Therefore, it is essential to consider the spatial heterogeneity in the studies for yellow croaker.
No long term coastal wide population dynamics assessment has been explored so far for this important species. Two existing stock assessments for the East China Sea subpopulation of yellow croaker used surplus production models to evaluate fishery status and population status (Li et al., 2011, Liu et al., 2013). However, both of them used short-term time series data (1991–2003 and 1999–2008, respectively) and only considered observation error in their model construction. Potential variations on its population dynamics caused by long term environmental changes and fishing pressure have been reflected in its life history changes and fishing behavior, such as early maturity and increased catchability, but they have never been considered in the population dynamics and stock assessment (Jin, 1996; Jiao et al., 2006, 2008; Shan et al., 2011; Yan et al., 2014). A comprehensive study on the population dynamics of yellow croaker is critically needed to analyze the long-term subpopulation dynamics under multiple pressures.
This study tends to investigate geographical variations in the population dynamics of two subpopulations of yellow croaker, i.e., the Bohai Sea and Yellow Sea subpopulation (BYSS) and the East China Sea subpopulation (ECSS), using fishery data back to late the 1960s. Bayesian hierarchical state-space surplus-production models have been implemented to describe the population dynamics in most previous researches (Jiao et al., 2009a, 2011; Li and Jiao, 2015). The hierarchically structured models with spatial heterogeneity could be used to address the variability of population parameters among subpopulations that have been reflected in life history variation and fishery characteristics changes (Li and Jiao, 2015). These models could test hypotheses on its population dynamics characteristics and potential catchability variations (Jiao et al., 2006, 2008). Based on these results, conservative management strategies with spatial considerations would be discussed and suggested for the management of yellow croaker fishery in China.
2 Materials and methods
Less
2.1 Data sources
Time series catch and effort data of yellow croaker were extracted from China Fishery Statistical Yearbooks from 1968 to 2015 (Bureau of Fisheries and Fishery Administration of Ministry of Agriculture, 1969–2016). Catch record given in 105 t, and the power of fishing vessels in 107 kW were used as proxies for fishing effort (Fig. 1). Data included both national and province-specific data for catch and effort. The yellow croaker fishery in Shandong Province and its northern provinces targeted BYSS, while that in Jiangsu Province and its southern provinces targeted ECSS.
Generally, China’s catches for yellow croaker kept at low levels (0.21×105–0.55×105 t) from 1968 to 1989, then appeared to increase reaching the highest value (4.07×105 t) in 2010, and slightly decreased in the subsequent five years (Fig. 1). Temporal trends in respective catches for the two subpopulations (BYSS and ECSS) were much similar, which were kept at low level during the first 23 years and increased after 1990. The fishing effort increased in these 48 years for both the whole population and the two subpopulations, except in 1991 when the effort decreased abruptly (Fig. 1). The geographical difference in the population parameters and overall dynamics among the two subpopulations of yellow croaker were evaluated in this study (Fig. 2).
2.2 Population dynamics model
The basic model structure was the state-space surplus production model:
${B_{t + 1}} = ({B_t} + f\left( {{B_t}} \right) - {C_{\rm t}}) \times {{\rm e}^{{\sigma _{1,t}}}},$
$\frac{{{C_{\rm t}}}}{{{E_t}}} = CPU{E_t} = q \times {B_t} \times {{\rm e}^{{\sigma _{2,t}}}},$
where Bt is the stock biomass in year t; Ct is the total catch; Et is the effort in year t; CPUEt is the catch per unit effort in year t (10–2 t/kW); and q is the catchability coefficient. The observation-process-error estimator used in this study, considered both the observation error in the function of CPUE and the process error in the function of stock biomass (Prager, 1994; de Valpine and Hastings, 2002).
The likelihood function for B is
$\begin{split}L\left( {B|{\theta _1}} \right) =& \mathop \prod \limits_{1968}^{2014} \frac{1}{{{B_t}{\sigma _1}\sqrt {2{\text π} } }}\times\\& \exp\left\{ {\frac{{{{\left[ {\ln {{B_{t + 1}}}- {\rm{ln}}({B_t} + f\left( {{B_t}} \right) - {C_{\rm t}}}) \right]}^2}}}{{2\sigma _1^2}}} \right\}.\end{split}$
The likelihood function for CPUE is
$\begin{split}L\left( {CPUE|{\theta _2}} \right) =& \mathop \prod \limits_{1968}^{2015} \frac{1}{{CPU{E_t}{\sigma _2}\sqrt {2{\text π} } }}\times\\& \exp\left\{ {\frac{{{{\left[ {\ln {CPU{E_t}} - {\rm{ln}}\left( {q{B_t}} \right)} \right]}^2}}}{{2\sigma _2^2}}} \right\}.\end{split}$
In the above two likelihood functions, θ1 and θ2 represent all the parameters in Eqs (1) and (2), respectively.
Schaefer model was used for the production function f (Bt), which is the most widely used and accepted surplus production model (Schaefer, 1954):
$f\left( {{B_t}} \right) = r{B_t}\left( {1 - \frac{{{B_t}}}{K}} \right),$
in which, r is the growth rate parameter (derived from the intrinsic rate of natural increase), and K represents the carrying capacity.
In the basic global model, yellow croaker was assumed to be one population along the coast of China without subpopulation-specific considerations.
$K \sim {\rm{ln}}N\left( {{\rm{ln}} {\bar K},{v_K}} \right)T\left( {{K_1},{K_2}} \right);\bar K \sim U\left( {4,10} \right);{v_K} \sim U\left( {0,1} \right),$
$r \sim {\rm{ln}}N\left( {{\rm{ln}}{\bar r},{v_r}} \right);\bar r \sim U\left( {0.6,0.8} \right);{v_r} \sim U\left( {0,1} \right),$
$q \sim U\left( {{q_{\_L}},{q_{\_U}}} \right).$
The carrying capacity K and growth rate r were assumed to follow lognormal distribution (lnN), with hyperparameters mean and variance. The values of K were limited to be higher than the maximum catch (K1=4×105 t) and lower than five times of the maximum catch (K2=5×4×105 t), given the reality that it is a heavily fished species. The hyperparameter $ \bar K$ was set to follow uniform distribution between 4×105 t and 10×105 t, since K for yellow croaker subpopulation in the East China Sea was estimated to be about 3.9 ×105 t (Li et al., 2011). The hyperparameter $ \bar r$ was set to follow uniform distribution between 0.6 and 0.8, which should be at least larger than the natural mortality (0.49 from Lin et al. (2006)) to keep the population survive. As the biomass was between the maximum of catch and the carrying capacity, and K was between max(catch) and 5×max(catch), the catchability q (CPUE/biomass) should be between the lower level q_L=CPUE/(5×max(catch) and the upper level q_U=CPUE/min(catch). The expectation of biomass in 1966 was assumed to follow uniform distribution between 70% and 100% of K, i.e., $ {\bar B_{1966}} \sim U\left( {0.7K,K} \right)$, based on the Bayesian model setting and the condition of this population (Fisheries Bureau and Yellow Sea Fisheries Headquarters in the Ministry of Agriculture, 1990; McAllister and Kirkwood, 1998). The variance of biomass vb=σ12 and the variance of CPUE vi=σ22 were all assumed to follow uniform distribution between 0 and 0.5, while other variances were set to follow uniform distribution between 0 and 1.
When the subpopulation-specific data were used to assess population dynamics, instead of modelling each subpopulation separately, a Bayesian hierarchical model was developed to assess the two subpopulations simultaneously (Parent and Rivot, 2012). In the hierarchical models, all three parameters (K, r and q) were assumed to be hierarchically structured, following lognormal distributions (Table 1).
${K_i} \sim {\rm{ln}}N\left( {{\rm{ln}} {\bar K} ,{v_K}} \right)T\left( {{K_{1i}},{K_{2i}}} \right),$
${r_i} \sim {\rm{ln}}N\left( {{\rm{ln}}{\bar r} ,{v_r}} \right),$
${q_i} \sim {\rm{ln}}N\left( {{\rm{ln}} {\bar q} ,{v_q}} \right)T\left( {{q_{\_Li}},{q_{\_Ui}}} \right),$
in which i is the subscript for subpopulations, with i=1 for BYSS and i=2 for ECSS.
The population size, structure and growth rate of yellow croaker, have undergone great changes due to the environmental variations and increased fishing pressure (Yan et al., 2014). Thus, the growth rate parameter r was set to be time-varying with random walk in some alternative models (Table 1), which meant rt following lognormal distribution with mean rt–1 and variance vr:
${r_{t = 1}} \sim {\rm{ln}}N\left( {{\rm{ln}} {\bar r} ,{v_r}} \right);{r_t} \sim {\rm{ln}}N\left( {{\rm{ln}}{r_{t - 1}}},{v_r}\right),{\rm{when}}\;t \geqslant 2.$
Due to the fast-economic development in China since 1980 and the improvement of the fishing technique, the catchability q was assumed to increase after 1980, with the ∆q following uniform distribution between 1 and 1.05:
$ \begin{split}{q_t} =& {q_{t - 1}} \times \Delta {q_{t - 1}};\Delta {q_{t - 1}} \sim U\left( {1,1.05} \right)\;\\& {\rm{between}}\;{\rm{1981}}\;{\rm{and}}\;{\rm{2015,}}\;{\rm{i}}{\rm{.e}}{\rm{.,}}\;t = {\rm{14}}\;{\rm{to}}\;{\rm{48}}{\rm{.}}\end{split}$
Therefore, in both the global models and subpopulation models, there were alternative models in which r or/and q were set to be time-varying. These four alternative models had parameters with constant r and q, time-varying q, time-varying r, as well as time-varying q and r (Table 1).
2.3 Bayesian methods
Just Another Gibbs Sampler (JAGS) is an effective program to fit Bayesian models using Markov Chain Monte Carlo (MCMC) (Plummer, 2003). In this study, coda, rjags and runjags packages were used to fit models with JAGS in R version 3.4.3 (Plummer et al., 2006; Denwood, 2016; Plummer, 2016). Except 50 000 burn-in iterations, 100 000 iterations and 3 chains were used to estimate the posterior distributions of parameters with a thinning interval of 100. Different initial values were generated for each chain during the model fitting, and Gelman-Rubin convergence statistic was calculated to diagnose model convergence (Gelman and Rubin, 1992).
In the global model, p(θ|CPUE), the posterior density for all parameters (θ) given the observations CPUE, can be written as
$p(\theta |CPUE) \propto L(CPUE|B,q,{{{\sigma }}_2}){\text π} (B|r,K,{\sigma _1}){\text π} \left( {q,{{{\sigma }}_2}} \right){\text π} \left( {r,K,{\sigma _1}} \right).$
In the subpopulation hierarchical models, the posterior probability can be expressed as
$\begin{split}p\left( {\theta '{\rm{|}}CPU{E_{i = 1,\,2}}} \right) \propto & \mathop \prod \limits_i \{ L\left( {CPU{E_i}{\rm{|}}{B_i},{q_i},{{{\sigma }}_{2,i}}} \right){\text π} \left( {{B_i}{\rm{|}}{r_i},{K_i},{\sigma _{1,i}}} \right)\times \\& {\text π} \left( {{K_i}{\rm{|}}\bar K,{v_K}} \right){\text π} \left( {{r_i}{\rm{|}}\bar r,{v_r}} \right){\text π} \left( {{q_i}{\rm{|}}\bar q,{v_q}} \right)\times \\& {\text π} \left( {\bar K} \right){\text π} \left( {\bar r} \right){\text π} \left( {\bar q} \right){\text π} \left( {{v_K}} \right){\text π} \left( {{v_r}} \right){\text π} \left( {{v_q}} \right)\} .\end{split}$
In the above equations, L(CPUE|B, q, σ2) is the likelihood function of CPUE. The expression π(B|r, K, σ1) is the joint distribution for biomass from year 1968 to 2015. π(vK) is the probability density function of parameter vK. i ranges from the first to the second subpopulation.
The performance of Bayesian models was measured based on the Deviance Information Criterion (DIC) (Spiegelhalter et al., 2002), which has been widely used and demonstrated by many previous studies (Jiao et al., 2009b, 2011; Chang et al., 2015; Wang et al., 2016). The DIC can be calculated by
${\rm {DIC}} = 2\bar D - \hat D,$
$\bar D\left( \theta \right) = {E_{\theta |y}}\left( { - 2\ln L(y|\theta )} \right),$
$\hat D\left( \theta \right) = - 2\ln L(y|\bar \theta ),$
where D is deviance used to measure the prediction of goodness, $ \bar D$ is the posterior mean of the deviance and $ \hat D$ is the deviance of the posterior mean. As a hierarchically modelling generalization of the Akaike Information Criterion and Bayesian Information Criterion, lower DIC values indicate better model fitting.
After comparing the four global Bayesian models (M1–M4) and four Bayesian hierarchical models (M5–M8), five scenarios were used to test the sensitivity of the model outcomes respectively, to the specified priors for the best model scenarios (Table 1). In Scenarios S1–S3, the variance (vi) of the CPUE (observation-error) was assumed to follow uniform distribution U(0, 0.5), while the prior distribution of the variance (vb) of biomass (process-error) were U(0, 0.4), U(0, 0.5) and U(0, 0.7), respectively. In Scenarios S4 and S5, the prior distribution of variance of CPUE were U(0, 0.4) and U(0, 0.7) respectively, while U(0, 0.5) was for the variance of biomass.
2.4 Biological reference points
The fishery biological reference points, including MSY (maximum sustainable yield), EMSY, FMSY and BMSY (the fishing effort, fishing mortality and the biomass to achieve MSY, respectively) were computed as follows (Haddon, 2011):
${\rm{MSY}} = \frac{{rK}}{4};{F_{\rm {MSY}}} = \frac{r}{2};{B_{\rm {MSY}}} = \frac{K}{2};{E_{\rm {MSY}}} = \frac{r}{{2 \times q}}.$
The risk of overfishing and overfished were defined by the MSY-based reference points, which are the probability of F being greater than FMSY, i.e., P(F>FMSY), and the probability of the biomass being smaller than BMSY, i.e., P(B<BMSY), respectively. The probability of E being greater than EMSY, P(E>EMSY), was also estimated, since the fishing effort was the main control target of the fishery management in China (Cao et al., 2017). The probabilities were estimated as the percentage of iterations where F>FMSY, E>EMSY or B<BMSY in the total number of iterations in the Bayesian approach. The depletion was also calculated by the biomass of the most recent year (2015) relative to the carrying capacity K (Caddy, 1998).
3 Results
Less
Among the four global models, Model M4 with time-varying r and q outperformed, which yielded much smaller DIC values than other models (Table 1). The subpopulation Model M8 with time-varying r and q, performed much better than other subpopulation models. Both the global Model M4 and the hierarchical Model M8 were relatively robust to the priors of the observation and process errors. The results were not sensitive to the changes in the upper bound of the noninformative priors of the CPUE and biomass variance under different sensitivity scenarios (Figs 3 and 4).
In Model M4, the posterior population growth rate r values were time-varying, with low level from 1965 to 1985 (median estimates lower than 0.50), and increased sharply to about 2.42–2.66 from 1985 to 2010 in different scenarios, with a little decline in recent five years (Fig. 5). The posterior r values for each subpopulation from Model M8 showed similar trend (Fig. 6), however the r values of ECSS after 1985 was higher than that of BYSS with overlapped 95% credible intervals (95% CI).
The catchability q for the global population and the two subpopulations, kept at low level before 1980, but increased by more than two times from 1981 to 2015. The q of the subpopulation ECSS was much lower than that of BYSS, with non-overlapped 95% CI (Fig. 6). The median of the posterior carrying capacity K for the whole population along the coast of China was 6.74×105–7.48×105 t in different scenarios. The posterior K for the two subpopulations were much similar, with the median values being 3.78–4.05 and 3.97×105–4.05 ×105 t in different scenarios for ECSS and BYSS, respectively.
The Bayesian state-space surplus production models, both M4 and M8, fitted the CPUE data well (Figs 7 and 8). When yellow croaker was assumed as a whole population along the coast of China in Model M4, its posterior estimated biomass decreased from 1966 to 1990, recovered in 1990 and slightly declined during the last decade. Similarly, the biomass for the two subpopulations in Model M8, varied with the same trend. Before 1980, the estimated biomass of the subpopulation BYSS was lower than that of ECSS, while it recovered to better conditions with higher biomass estimates than that of ECSS.
The whole population of yellow croaker likely became overfished ever since the late 1970s and overfishing occurred since from the mid-1970s, with depletion rate varying from 0.26–0.28 in 2015 in different scenarios (Fig. 9 and Table 2). The P(F>FMSY) and P(E>EMSY) of the whole population showed similar patterns. The subpopulation BYSS became overfished since early 1970s and overfishing occurred ever since the mid-1980s, and the depletion rate observed from 0.36–0.37 (Fig. 10 and Table 2). The subpopulation ECSS became overfished since 1980 and overfishing occurred periodically during the 1980s and absolutely since 1990, and recorded depletion rate in 2015 was 0.19–0.20 (Fig. 10 and Table 2). Estimates of P(F>FMSY), P(E>EMSY) and P(B<BMSY) for the ECSS subpopulation were much higher than those of BYSS in the last 15 years (Fig. 10).
According to the results obtained via Bayesian models, the biological reference points including MSY, EMSY, FMSY and BMSY were estimated for both the whole population and the two subpopulations of yellow croaker (Table 2). The MSY for the whole population was 4.19×105–4.47 ×105 t according to the global Model M4, while the MSY for the subpopulations BYSS and ECSS were 1.97×105–2.00×05 and 2.37×105–4.43×105 t based on the subpopulation Model M8, respectively. The EMSY for the subpopulation ECSS was 0.65×1070.67×107 kW in 2015, while the observed fishing effort in 2015 was 1.26×107 kW. For the subpopulation BYSS, the observed effort in 2015 was 0.45×107 kW, which was much higher than the EMSY 0.28×107–0.29×107 kW in the same year.
4 Discussion
Less
In this study, the whole yellow croaker population in China seas during the past 48 years was covered with and without considering subpopulation dynamics, which could greatly improve our understanding of the status and dynamics of this species. Bayesian hierarchical models were applied in this study to conduct stock assessment for yellow croaker, considering the spatial and temporal variations of population growth rate and catchability, and provided the biological reference points for the fishery management of yellow croaker. Based on the results of this study, more conservative management strategies with spatial considerations and specific reference points for MSY, biomass and effort, were suggested for the management of yellow croaker fishery in China.
The best Bayesian hierarchical model revealed that the growth rate r and the catchability q of the yellow croaker’s population increased over a long time series and showed a large geographical difference. The hierarchical Bayesian models with multilevel priors and time-varying r and q consistently proved to fit the data much better. So, it was suggested to be a mathematically viable and valuable approach in modelling the spatial complex fishery data (Jiao et al., 2008, 2009a). The use of multilevel priors in Bayesian models was found to be a better choice than the commonly used priors because multilevel priors could lead to the stability of model results (Roberts and Rosenthal, 2001; Jiao et al., 2011). The methodology developed and demonstrated in this study also applies to the stock assessment and fishery management of many other species, especially for those species with large spatial heterogeneity and limited fishery data, as is the case with most fishery in China.
The population growth rate was not assumed to be constant, which allowed the investigation of its variation over time to likely be related to regime shifts and changes in productivity regimes (Beamish et al., 1999; Jiao, 2009; Clark, 2003). Both data fitting and the improved fishing techniques confirmed the increasing pattern of the catchability q. Thus, the growth rate r and catchability q of yellow croaker might exhibit some variations caused by the intensive fishing activities and catch technique improvements, and should not be assumed consistent in the model.
The population of yellow croaker in China has depleted under high fishing pressure. Under this circumstance, the growth rate of this species might increase to survive and recover its population. Previous studies have shown that individuals of yellow croaker grow faster and mature in their early years. Yellow croaker usually spawns during April and May (Ren et al., 2001), and the younger ones feed and grow faster in summer during the closed fishing season. In recent years, the juvenile yellow croaker of 4 or 5 months old grew to the mean length of 12.75 cm (Cheng et al., 2004), which reached the minimal mesh size of fishing nets. The percentage of the 4- or 5-month-old individuals in the total catch increased from 43% in 1992–1994 to 75% in 2000 (Cheng et al., 2004). Therefore, the population growth rate of yellow croaker has been increasing to high values (even larger than 2). Additionally, the predation mortality on yellow croaker decreased with larger predator fish species being removed by fishing in this region (Matsuda and Abrams, 2006). So the existence of trophic cascade might also have contributed to the increasing growth rate of yellow croaker.
For both the whole population and the subpopulations, the biomass of yellow croaker in 1968 was higher or close to the carrying capacity, which led to the rapid decline of the population size under exploitation. The increased fishing effort led to the continuous depletion of its population. Faced with this situation, the fishery management agency of China began to apply more restrictive strategies, such as Spawning Ground protection program since 1981, introducing Fisheries Law in 1986, and Seasonal Moratorium since mid-1990s (Cao et al., 2017). In this way, the population of yellow croaker began to recover since 1990, and kept relatively stable at a level lower than BMSY in recent years. In the future, further control on fishing effort or overall quota will be needed to recover this population to BMSY level.
The results of the best Bayesian hierarchical models indicated considerable subpopulation variations of yellow croaker population along the coastal areas of China. Before 1990, the biomass of the subpopulation ECSS was much higher than that of the subpopulation BYSS. The conservation management effort has been enhanced to protect fishery resources in China since 1990. Then in 2015, the biomass of the two subpopulations had recovered to two-thirds and one-third of the biomass in 1968 for BYSS and ECSS, respectively. This changes in population size along latitudes might be as a result from global warming. With the increased water temperatures under global warming, the preferred habitat of yellow croaker changed northward, meaning that the northern region becomes its favored habitat, which was also observed for many other aquatic species (Tanaka et al., 2019; Torre et al., 2019). Furthermore, the fishing efforts in the East China Sea was 2.5 times higher than that in the Bohai Sea and Yellow Sea, which might be the other factor that caused geographical variations of yellow croaker subpopulations.
The q estimate for ECSS was much lower than the BYSS. One possible reason is that the fishery targeting yellow croaker of ECSS are from all the south coast provinces, from Jiangsu to Guangdong and Hainan (Bureau of Fisheries and Fishery Administration of Ministry of Agriculture, 1969–2016). For provinces situated far away from the East China Sea, such as Guangdong and Hainan Provinces, would record much lower catchability q. A part of the ECSS is distributed relatively far away from the coast, while BYSS is distributed near the coast, which might lead to the lower q of ECSS.
Apart from the heterogeneity in large spatial scale among two subpopulations of yellow croaker, multiple previous studies have proven the existence of this species local spatial pattern (Lin et al., 2011; Shan et al., 2016; Xiong et al., 2016). The spatial pattern in smaller scale would also influence the population structure and population dynamics of this stock to a degree. Intensive surveys over time with local spatial population distributions are needed for further smaller scale population dynamics analysis (Jiao et al., 2016).
In this study, the values of P(F>FMSY), P(E>EMSY) and P(B<BMSY) were high, almost close to 100%, which indicated unfavorable condition of yellow croaker population in recent years. Additionally, previous studies revealed that smaller sizes at age and earlier maturity might indicate overfishing of yellow croaker over a longer time period (Shan et al., 2016), and the fishing effort has increased 40 times since the 1950s, which contribute to the life history traits shift of yellow croaker (Shan et al., 2013). The status for the two subpopulations was dramatically different, and the risk of overfishing and being overfished for the subpopulation ECSS was higher and increased since 2010, which was consistent with the results of previous studies (Lin, 2004; Li et al., 2011; Liu et al., 2013). Therefore, conservative management strategy was imperative for yellow croaker, especially for the subpopulation ECSS. The biological reference points, estimated from Bayesian models, can be referred to make management strategy for yellow croaker in the future. For instance, the observed fishing effort (0.45×107 and 1.26×107 kW for BYSS and ECSS, respectively) were much higher than the EMSY (about 0.28×107 and 0.67×107 kW respectively) for both subpopulations. Therefore, the fishing effort should be controlled more strictly in the fishery management process.
The data used in this study were derived from China Fishery Statistical Yearbooks from 1968 to 2015. Fishery data, especially about the fishery dynamics, are essential to conduct stock assessment, and to evaluate and guide sustainable fishery production. China Fishery Statistical Yearbooks is the unique fishery-dependent survey data source with high temporal-spatial coverage, which covers all the Chinese coast during these decades. Although there are some controversies about the reliability of this official data (Blomeyer et al., 2012; Campbell and Pauly, 2013; Pauly et al., 2014), there actually has been much researches conducted based on this dataset and yielded valuable knowledge and inferences (Li et al., 2017; Szuwalski et al., 2017; Kang et al., 2018). The official fishery statistics may be with uncertainty but it captured the general trend of the catch and effort, which is much useful to estimate the population dynamics. Additionally, the effort used in this study, power of fishing vessels is considered to be relatively reliable, since during the strict supervision for fishing vessel permit, the power is the important criteria to calculate fuel subsidies that a vessel receives from Chinese government (Kang et al., 2018). Therefore, this study tried the data from this unique and valuable dataset with uncertainties to explore population dynamics by Bayesian model considering different level of errors.
The uncertainty resulted from data collection could be considered to explore its influence for yellow croaker stock assessment. For example, the existence of unregistered vessels could lead to underestimated catch and effort data, which can be tested by complicated sensitivity analysis with different underreporting level. The implementation of zero-growth strategy for total catch of marine fishery in 1999 and negative-growth strategy in 2000, could influence the fishery dynamics and might increase the bias of the estimated population dynamics. Additionally, the fishery dependent data are not available from Korean fishery, in which BYSS of yellow croaker has been one target species. The catchability, fishing season and ground are different among Chinese and Korean fishery, leading to some stock assessment uncertainty. Therefore, much more efforts are required to improve the fishery monitoring program and to consider the uncertainty of data for stock assessment in future studies.
5 Conclusions
Less
The population dynamics modelling in this study revealed that the population growth rate of yellow croaker and its fishery catchability along the coast of China exhibit increasing pattern, which are likely due to the evolutionary response of the population facing intensive fishing and the improvement in catching technique separately. The biomass of yellow croaker depleted and recovered to a relatively stable level with spatial heterogeneity, but this important fishery is still overfished and subject to overfishing in recent years. Consequently, more conservative management strategy based on the biological reference points from the population dynamics models was imperative for yellow croaker along the coast of China, considering the heterogeneity of two subpopulations (BYSS and ECSS).
Acknowledgements
Less
We gratefully thank Xiaoxiao Liu for helping in data synthesis and equally thank Can Zhou during the modelling process. Writing of this manuscript was improved by the comments from Kindong Richard. We thank the Department of Fish and Wildlife Conservation of the Virginia Polytechnic Institute and State University for the opportunity given to Qiuyun Ma to work on this project under the supervision of Yan Jiao. We also thank the China Scholarship Council who provided the funding for Qiuyun Ma to work at Virginia Tech.
Funding
Less
  • The National Key R&D Program of China under contract No. 2017YFE0104400; the National Natural Science Foundation of China under contract No. 31772852; the Fundamental Research Funds for the Central Universities under contract Nos 201512002 and 201562030.
References
Less
Beamish R J, Noakes D J, McFarlane G A, et al. 1999. The regime concept and natural trends in the production of Pacific salmon. Canadian Journal of Fisheries and Aquatic Sciences, 56(3): 516–526, doi: 10.1139/f98-200
Blomeyer R, Goulding I, Pauly D, et al. 2012. The Role of China in World Fisheries. Brussels: European Parliament, Directorate General for Internal Policies, Policy Department B: Structural and Cohesion Policies-Fisheries
Bureau of Fisheries and Fishery Administration of Ministry of Agriculture. 1969–2016. China Fishery Statistical Yearbook (in Chinese). Beijing: China Agriculture Press
Caddy J. 1998. A short review of precautionary reference points and some proposals for their use in data-poor situations. Rome: Food & Agriculture Organization
Campbell B, Pauly D. 2013. Mariculture: a global analysis of production trends since 1950. Marine Policy, 39: 94–100, doi: 10.1016/j.marpol.2012.10.009
Cao Ling, Chen Yong, Dong Shuanglin, et al. 2017. Opportunity for marine fisheries reform in China. Proceedings of the National Academy of Sciences of the United States of America, 114(3): 435–442, doi: 10.1073/pnas.1616583114
Chang Y J, Brodziak J, O’Malley J, et al. 2015. Model selection and multi-model inference for Bayesian surplus production models: A case study for Pacific blue and striped marlin. Fisheries Research, 166: 129–139, doi: 10.1016/j.fishres.2014.08.023
Cheng Jiahua, Lin Longshan, Ling Jianzhong, et al. 2004. Effects of summer close season and rational utilization on redlip croaker (Larimichthys polyactis Bleeker) resource in the East China Sea Region. Journal of Fishery Sciences of China (in Chinese), 11(6): 554–560
Clark J S. 2003. Uncertainty and variability in demography and population growth: A hierarchical approach. Ecology, 84(6): 1370–1381, doi: 10.1890/0012-9658(2003)084[1370:UAVIDA]2.0.CO;2
Denwood M J. 2016. Runjags: An R package providing interface utilities, model templates, parallel computing methods and additional distributions for MCMC models in JAGS. Journal of Statistical Software, 71(1): 1–25
de Valpine P, Hastings A. 2002. Fitting population models incorporating process noise and observation error. Ecological Monographs, 72(1): 57–76, doi: 10.1890/0012-9615(2002)072[0057:FPMIPN]2.0.CO;2
Fisheries Bureau and Yellow Sea Fisheries Headquarters in the Ministry of Agriculture. 1990. Investigation and Regionalization of Fishery Resources in the Yellow Sea and Bohai Sea (in Chinese). Beijing: China Ocean Press
Gelman A, Rubin D B. 1992. Inference from iterative simulation using multiple sequences. Statistical Science, 7(4): 457–472, doi: 10.1214/ss/1177011136
Haddon M. 2011. Modelling and Quantitative Methods in Fisheries. 2nd ed. New York: CRC Press
Jensen A L. 2005. Harvest in a fluctuating environment and conservative harvest for the Fox surplus production model. Ecological Modelling, 182(1): 1–9, doi: 10.1016/j.ecolmodel.2003.08.004
Jiao Yan. 2009. Regime shift in marine ecosystems and implications for fisheries management, a review. Reviews in Fish Biology and Fisheries, 19(2): 177–191, doi: 10.1007/s11160-008-9096-8
Jiao Yan, Cortés E, Andrews K, et al. 2011. Poor-data and data-poor species stock assessment using a Bayesian hierarchical approach. Ecological Applications, 21(7): 2691–2708, doi: 10.1890/10-0526.1
Jiao Yan, Hayes C, Cortés E. 2009a. Hierarchical Bayesian approach for population dynamics modelling of fish complexes without species-specific data. ICES Journal of Marine Science, 66(2): 367–377, doi: 10.1093/icesjms/fsn162
Jiao Yan, Lapointe N W R, Angermeier P L, et al. 2009b. Hierarchical demographic approaches for assessing invasion dynamics of non-indigenous species: An example using northern snakehead (Channa argus). Ecological Modelling, 220(13–14): 1681–1689, doi: 10.1016/j.ecolmodel.2009.04.008
Jiao Yan, Neves R, Jones J. 2008. Models and model selection uncertainty in estimating growth rates of endangered freshwater mussel populations. Canadian Journal of Fisheries and Aquatic Sciences, 65(11): 2389–2398, doi: 10.1139/F08-141
Jiao Yan, O’Reilly R, Smith E, et al. 2016. Integrating spatial synchrony/asynchrony of population distribution into stock assessment models: a spatial hierarchical Bayesian statistical catch-at-age approach. ICES Journal of Marine Science, 73(7): 1725–1738, doi: 10.1093/icesjms/fsw036
Jiao Yan, Reid K, Nudds T. 2006. Variation in the catchability of yellow perch (Perca flavescens) in the fisheries of Lake Erie using a Bayesian error-in-variable approach. ICES Journal of Marine Science, 63(9): 1695–1704, doi: 10.1016/j.icesjms.2006.07.002
Jin Xianshi. 1996. Ecology and population dynamics of small yellow croaker (Pseudosciaena polyactis Bleeker) in the Yellow Sea. Journal of Fishery Sciences of China (in Chinese), 3(1): 32–46
Kang Bin, Liu Min, Huang Xiaoxia, et al. 2018. Fisheries in Chinese seas: What can we learn from controversial official fisheries statistics?. Reviews in Fish Biology and Fisheries, 28(3): 503–519, doi: 10.1007/s11160-018-9518-1
Li Jilong, Cao Kun, Ding Fang, et al. 2017. Changes in trophic-level structure of the main fish species caught by China and their relationship with fishing method. Journal of Fishiery Sciences of China (in Chinese), 24(1): 109–119, doi: 10.3724/SP.J.1118.2017.16164
Li Yan, Jiao Yan. 2015. Evaluation of stocking strategies for endangered white abalone using a hierarchical demographic model. Ecological Modelling, 299: 14–22, doi: 10.1016/j.ecolmodel.2014.11.031
Li Jiuqi, Ye Changchen, Wang Wenbo, et al. 2011. A stock assessment of small yellow croaker by Bayes-based Pella-Tomlinson model in the East China Sea. Journal of Shanghai Ocean Univeristy (in Chinese), 20(6): 873–882
Lin Longshan. 2004. Analysis on extant abundance of small yellow croaker Pseudosciaena polyactis in the East China Sea. Marine Fisheries (in Chinese), 26(1): 18–23
Lin Xinzhuo, Deng Siming, Huang Zhengyi. 1965. Study of population on biometrics of small yellow croaker (Pseudosciaena polyactis Bleeker). In: Zhu Yuanding, Zhu Shuping, eds. Collections of Marine Fishery Resource (in Chinese). Beijing: China Agricultural Press, 84–108
Lin Longshan, Liu Zunlei, Jiang Yazhou, et al. 2011. Current status of small yellow croaker resources in the southern Yellow Sea and the East China Sea. Chinese Journal of Oceanology and Limnology, 29(3): 547–555, doi: 10.1007/s00343-011-0182-8
Lin Longshan, Zheng Yuanjia, Cheng Jiahua, et al. 2006. A preliminary study on fishery biology of main commercial fishes surveyed from the bottom trawl fisheries in the East China Sea. Marine Sciences (in Chinese), 30(2): 21–25, 42
Liu Xiaoshun. 1990. Investigation and Division of the Yellow Sea and Bohai Sea Fishery Resources (in Chinese). Beijing: China Ocean Press
Liu Zunlei, Yan Liping, Yuan Xingwei, et al. 2013. Stock assessment of small yellow croaker in the East China Sea based on multi-source data. Journal of Fishery Sciences of China (in Chinese), 20(5): 1039–1049, doi: 10.3724/SP.J.1118.2013.01039
Matsuda H, Abrams P A. 2006. Maximal yields from multispecies fisheries systems: Rules for systems with multiple trophic levels. Ecological Applications, 16(1): 225–237, doi: 10.1890/05-0346
McAllister M K, Kirkwood G P. 1998. Bayesian stock assessment: a review and example application using the logistic model. ICES Journal of Marine Science, 55(6): 1031–1060, doi: 10.1006/jmsc.1998.0425
Parent E, Rivot E. 2012. Introduction to Hierarchical Bayesian Modeling for Ecological Data. Boca Raton: Chapman and Hall
Pauly D, Belhabib D, Blomeyer R, et al. 2014. China’s distant-water fisheries in the 21st century. Fish and Fisheries, 15(3): 474–488, doi: 10.1111/faf.12032
Plummer M. 2003. JAGS: A program for analysis of Bayesian graphical models using Gibbs sampling. In: Hornik K, Leisch F, Zeileis A, eds. Proceedings of the 3rd International Workshop on Distributed Statistical Computing (DSC 2003). Vienna, Austria: Technische Universität Wien, 1–8
Plummer M. 2016. Rjags: Bayesian Graphical Models using MCMC. https://cran.r-project.org/package=rjags [2016-02-19/2017-04-04]
Plummer M, Best N, Cowles K, et al. 2006. CODA: Convergence diagnosis and output analysis for MCMC. R News, 6(1): 7–11
Prager M H. 1994. A suite of extensions to a nonequilibrium surplus-production model. Fishery Bulletin, 92(2): 374–389
Punt A E, Hilborn R. 1997. Fisheries stock assessment and decision analysis: the Bayesian approach. Reviews in Fish Biology and Fisheries, 7(1): 35–63, doi: 10.1023/A:1018419207494
Quinn T J II, Deriso R B. 1999. Quantitative Fish Dynamics. New York: Oxford University Press
Ren Yiping, Gao Tianxiang, Liu Qun, et al. 2001. Study on the population structure and reproduction of Pseudosciaena plyactis in southern Yellow Sea. Transactions of Oceanology and Limnology (in Chinese), (1): 41–46
Roberts G O, Rosenthal J S. 2001. Infinite hierarchies and prior distributions. Bernoulli, 7(3): 453–471, doi: 10.2307/3318496
Schaefer M B. 1954. Some aspects of the dynamics of populations important to the management of commercial marine fisheries. Bulletin, Inter-American Tropical Tuna Commission, 1(2): 26–56
Shan Xiujuan, Li Zhonglu, Dai Fangqun, et al. 2011. Seasonal and annual variations in biological characteristics of small yellow croaker Larimichthys polyactis in the central and southern Yellow Sea. Progress in Fishery Sciences (in Chinese), 32(6): 7–16
Shan Xiujuan, Jin Xianshi, Dai Fangqun, et al. 2016. Population dynamics of fish species in a marine ecosystem: A case study in the Bohai Sea, China. Marine and Coastal Fisheries, 8(1): 100–117, doi: 10.1080/19425120.2015.1114543
Shan Xiujuan, Sun Pengfei, Jin Xianshi, et al. 2013. Long-term changes in fish assemblage structure in the Yellow River estuary ecosystem, China. Marine and Coastal Fisheries, 5(1): 65–78, doi: 10.1080/19425120.2013.768571
Spiegelhalter D J, Best N G, Carlin B P, et al. 2002. Bayesian measures of model complexity and fit. Journal of the Royal Statistical Society. Series B: Statistical Methodology, 64(4): 583–639, doi: 10.1111/1467-9868.00353
Szuwalski C S, Burgess M G, Costello C, et al. 2017. High fishery catches through trophic cascades in China. Proceedings of the National Academy of Sciences of the United States of America, 114(4): 717–721, doi: 10.1073/pnas.1612722114
Tanaka K R, Cao Jie, Shank B, et al. 2019. A model-based approach to incorporate environmental variability into assessment of a climatically-influenced commercial fishery: A case study with the American lobster fishery in the Gulf of Maine and Georges Bank. ICES Journal of Marine Science, (fsz024): 1–13
Torre M P, Tanaka K R, Chen Y. 2019. Development of a climate-niche model to evaluate spatiotemporal trends in Placopecten magellanicus distribution in the Gulf of Maine, USA. Journal of Northwest Atlantic Fishery Science, 50: 37–50, doi: 10.2960/J.v50.m721
Wang Jintao, Yu Wwei, Chen Xinjun, et al. 2016. Stock assessment for the western winter-spring cohort of neon flying squid (Ommastrephes bartramii) using environmentally dependent surplus production models. Scientia Marina, 80(1): 69–78
Xiong Ying, Zhong Xiaming, Tang Jianhua, et al. 2016. Migration and population structure characteristics of the small yellow croaker Larimichthys polyactis in the southern Yellow Sea. Acta Oceanologica Sinica, 35(6): 34–41, doi: 10.1007/s13131-016-0844-7
Xu Zhaoli, Chen Jiajie. 2010. Population division of Larimichthys polyactis in China Sea. Chinese Journal of Applied Ecology (in Chinese), 21(11): 2856–2864
Yan Liping, Liu Zunlei, Zhang Hui, et al. 2014. On the evolution of biological characteristics and resources of small yellow croaker. Marine Fisheries (in Chinese), 36(6): 481–488
Zhang Chi, Ye Zhenjiang, Wan Rong, et al. 2014. Investigating the population structure of small yellow croaker (Larimichthys polyactis) using internal and external features of otoliths. Fisheries Research, 153: 41–47, doi: 10.1016/j.fishres.2013.12.012
Appendix
Less
Year 2020 volume 39 Issue 10
PDF
28
15
Cite this Article
BibTeX
Article Info
doi: 10.1007/s13131-020-1602-4
  • Receive Date：2019-09-06
  • Online Date：2026-03-31
  • Published：2020-10-25
Article Data
Affiliations
History
  • Received：2019-09-06
  • Accepted：2019-11-01
Funding
The National Key R&D Program of China under contract No. 2017YFE0104400; the National Natural Science Foundation of China under contract No. 31772852; the Fundamental Research Funds for the Central Universities under contract Nos 201512002 and 201562030.
Affiliations
    1 Fisheries College, Ocean University of China, Qingdao 266003, China
    2 Department of Fish and Wildlife Conservation, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, USA
    3 Laboratory for Marine Fisheries Science and Food Production Processes, Pilot National Laboratory for Marine Science and Technology (Qingdao), Qingdao 266237, China
    4 College of Marine Sciences, Shanghai Ocean University, Shanghai 201306, China

Corresponding:

References
Share
https://castjournals.cast.org.cn/joweb/aos/EN/10.1007/s13131-020-1602-4
Share to
QR

Scan QR to access full text

Cite this article
BibTeX
Citations
表12种不同金属材料的力学参数
科
Family		 属数
Number of
genus		 种数
Number of
species		 占总种数比例
Percentage of
total species (%)		 属
Genus		 种数
Number of
species		 占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏
  • BibTeX
  • EndNote
  • RefWorks
  • TxT
Links
Articles
Latest Articles
Most Read
Collections
Updates
Events
News
Multimedia
About
About Us
Contact
No. 86 Xueyuan South Road, Haidian District, Beijing
100081
010-62199257
qkjq@cast.org.cn

Copyright © 2025 China Association for Science and Technology. All rights reserved. For all open access content, the relevant licensing terms apply.
Sponsored by the Office of the Leading Group for Cybersecurity and Informatization of CAST, and supported by Science and Technology Review Publishing House