Assessing the Fishery Resource Status of China’s Coastal Waters Using Surplus Production Models

ZHANG Qingqing, LIU Qun, and HAN Ya’nan

Assessing the Fishery Resource Status of China’s Coastal Waters Using Surplus Production Models

ZHANG Qingqing, LIU Qun*, and HAN Ya’nan

,,266003,

Surplus production models (SPMs) are among the simplest and most widely used fishery stock assessment models. The catch-effort data analysis (CEDA) and a surplus production model incorporating covariates (ASPIC) are softwares for analyzing fishery catch and fishing effort data using nonequilibrium SPMs. In China Fishery Statistical Yearbook, annual fishery production and fishing effort data of the Yellow Sea, Bohai Sea, East China Sea, and South China Sea have been publishedfrom 1979 till present. Using its catch and fishing effort data from 1980 to 2018, we apply the CEDA and ASPIC to evaluate fishery resources in Chinese coastal waters. The results show that the total maximum sustainable yield (MSY) estimate of the four China seas is 10.05–10.83 million tons, approximately equal to the marine fishery catch (10.44 million tons) reported in 2018. It can be concluded that China’s coastal fishery resources are currently fully exploited and must be protected with a precautionary approach. Both softwares produced similar results; however, the CEDA had a much higher2value (above 0.9) than ASPIC (about 0.2), indicating that CEDA can better fit the data and therefore is more suitable for analyzing the fishery resources in the coastal waters of China.

Chinese coastal waters; fishery resources; surplus production models (SPMs); catch-effort data analysis (CEDA); a surplus production model incorporating covariates (ASPIC); China Fishery Statistical Yearbook

1 Introduction

Surplus production models (SPMs), also known as bio- mass dynamic models (Hilborn and Walters, 1992), arewidely used in fish-stock assessment when analyzing catch and effort data, especially when age composition data is not available. SPMs treat the whole resource population as a single unit, without separately considering recruitment, growth, and mortality. Thus, they require less data than age- structured models (., cohort analysis) and produce results that can be easily understood (Haddon, 2011).

SPMs require time series fishery statistics data such as catch, fishing effort, or catch per unit effort (CPUE) data butdo not need biological life history parameters of the fish spe-cies concerned. SPMs output the maximum sustainable yield (MSY) estimate, which is commonly used as a manage- ment benchmark (Jacobson., 2002). Although they arevery simple, ‘SPMs may produce results just as useful andsometimes better for management than those produced us- ing an age-structured model’ (Haddon, 2011).

Since the late 1970s, owing to the reform and opening up policy, Chinese marine fisheries have entered a rapid development stage and provided high-quality proteins to Chinese people. However, the cost of rapid development is the over-exploitation of fishery resources; therefore, the assessments and management of the resources are becom- ing increasingly imperative. At present, most assessments focus on single species, because there are many difficulties for multispecies assessments. The SPMs can be used for both single species and multispecies assessments based on their own premises, and they are the simplest multispecies evaluation models (Zhan, 1995). In the absence of sufficient age-structure and environmental information, the results of SPMs may provide a valuable reference.

The China Fishery Statistical Yearbooks (1979–2019) are compiled by The Ministry of Agriculture and Rural Af- fairs (known as the Ministry of Agriculture before March 2018) (Su., 2019) and contain the annual fishery sta- tistical data in China (mainland unless otherwise specified) since 1979, which are comprehensive statistics on Chinese fisheries. Since the data in the China Fishery Statistical Yearbook were not derived from scientific surveys, their accuracy has been widely discussed (Zhao and Shen, 2016; Kang., 2018; Su., 2019). To assess the status of fishery resources in Chinese seas, this study used SPMs to analyze the marine capture fisheries catch and fishing ef- fort data recorded in the China Fishery Statistical Yearbooks.

Both catch-effort data analysis (CEDA) (Hoggarth., 2006) and a surplus production model incorporating covariates (ASPIC) (Prager, 2005) software can use nonequilibrium surplus production models to analyze fishery catchand fishing effort data and estimate biological referencepoints, such as, stock biomass producing(B) and fishing mortality rate at(F). The abovementioned two softwares have been widely used in the existing literature (Wang and Liu, 2013; Kalhoro., 2015; Xu., 2015; Ji., 2019; Karim, 2019). In this study, both CEDA and ASPIC were used to assess thefishery resources in the Yellow and Bohai Seas (YBS), East China Sea (ECS), and South China Sea (SCS).

2 Materials and Methods

2.1 Data

The catch and fishing effort data used in this study wereacquired from the China Fishery Statistical Yearbooks. Catch data were directly derived from the yearbooks (Fishery Bu- reau of Agriculture Ministry of China, 1980–2019), while fishing effort data were obtained by adding the statistics of the fishing vessel powers of different provinces. Accord- ing to the division of China’s coastal waters (Fig.1) (Kang, 2018) and statistical fishing effort data in the Year- book,we combined the YBS data as one group. Eventu- ally, we evaluated the fishery resources in the three coas- tal areas,., the YBS, ECS, and SCS.

Fig.1 Map of China’s coastal waters with administration areas (a–n) and the four surrounding seas–Liaoning (a), Hebei (b), Tianjin (c), Shandong (d), Jiangsu (e), Shanghai (f), Zhejiang (g), Fujian (h), Guangdong (i), Hong Kong (j), Macau (k), Guangxi (l), Hainan (m), and Taiwan (n). BS, Bohai Sea; YS, Yellow Sea; ECS, East China Sea; SCS, South China Sea. Dashed black lines indicate the boundaries between the seas (Kang et al., 2018).

Fig.2 describes the total catch and CPUE of the fisheries in the three coastal waters in China from 1980 to 2018.

2.2 Surplus Production Models (SPMs)

The SPMs have three versions,., Schaefer, Fox, and Pella-Tomlinson. Schaefer model is built on a logistical po- pulation growth model:

Fig.2 Statistics of catch (t) and CPUE (t(kw)−1) for the fisheries in the Yellow and Bohai Seas (YBS), the East China Sea (ECS), and the South China Sea (SCS) from 1980 to 2018.

Next, Fox proposes an analysis based on the Gompertz population growth equation:

Pella and Tomlinson develop a generalized production equation:

whereis fish-stock biomass;is time (year);∞is carrying capacity;is the intrinsic rate of population increase, andis a shape parameter. The Schaefer and Fox models are special cases of the Pella-Tomlinson model.

2.3 Catch-Effort Data Analysis (CEDA)

The CEDA (version 3.0.0.1) software contains three non- equilibrium production models (Schaefer, Fox, and Pella- Tomlinson models) and three error assumptions (normal, lognormal, and gamma distributions). These models are able to calculate the population parameters and biological reference points, including intrinsic rate of increase (), carry- ing capacity (), catchability coefficient (),, replace- ment yield (yield) and coefficient of determination (2) (Hoggarth., 2006). It needs to set an initial starting bio-mass over carrying capacity (1/) value and allow the boot- strap process (Efron and Tibshirani, 1993; Haddon, 2011) to calculate the confidence intervals for the parameters.

2.4 A Surplus Production Model Incorporating Covariates (ASPIC)

ASPIC (version 5.0) is another software developed to estimate parameters of nonequilibrium SPMs. It includes two types of SPMs: Fox and Logistic. The output parame- ters include,,,2,B,F, andf(optimum fishing effort at). It also allows the bootstrap estima- tion of variability and is extremely flexible in handling dif- ferent patterns of fishing (Prager, 2005).

Both CEDA and ASPIC need an initial proportion (IP) of1/. This value is prior information on stock status at the start of a dataset. When IP is set near 1, it indicates that the data started from a virgin population, and when set near zero, it hints that the data started from a seriously exploited population.

3 Results

3.1 CEDA

Results in Table 1 show that CEDA is sensitive to different IP values. Since the results of the Pella-Tomlinson model are similar (or identical) to those from the Schaefer model (Table 1), and the former model has a production curve shape parameter that is difficult to estimate, the re- sults from the Pella-Tomlinson model are not shown in this study. Because of the minimization failures under the assumption of a gamma-error structure, the SPMs of Fox and Schaefer (namely, Logistic in the ASPIC) with two error assumptions of normal and lognormal distributions were used in CEDA.

Table 1 Maximum sustainable yield (MSY) estimates for the fishery in three coastal waters of China via catch-effort data analysis (CEDA) with the initial proportions (IP) ranging from 0.1 to 0.9

()

()

AreaIPFoxSchaeferPella-Tomlinson NormalLognormalGammaNormalLognormalGammaNormalLognormalGamma 0.532565852.42E+07312405729494603012710 –29494603012710 – (0.2517)(2.7827)(0.1954)(0.1480)(0.0901) –(0.1678)(0.0886) – 0.634625433732515 –302091531097663020112302091531097663020112 (0.2856)(0.1022) –(0.1817)(0.0850)(0.2249)(0.2312)(0.0933)(0.2131) SCS0.736101064439103 –30616093756127 –30616093756127 – (0.2942)(0.2405) –(0.2774)(0.3100) –(0.2646)(0.3108) – 0.834654715540587 –29208674544310 –29208674544310 – (0.4364)(1.2623) –(0.3540)(1.1442) –(0.3702)(0.7808) – 0.9949280.87630878265135612865366802346 –12865366802346 – (3.0769)(4.8029)(1.2343)(1.5915)(4.0606) –(1.6615)(4.0278) –

Notes: CV, coefficient of variation are shown beneath the eachvalue in brackets. Cells with ‘–’ indicate minimization failure.

Table 2 shows the CEDA outputs in the three coastal waters with the IP values 0.2, 0.3, and 0.2 in YBS, ECS, and SCS, respectively. These IP values were chosen because the starting catch was roughly 20%, 30%, and 20% of the maximum catch in each area and the results were relatively stable in the sensitivity analyses. In YBS, theestimates with coefficient of variation () from the Fox model with normal and lognormal distribution assumptionsare 3751369t (=0.0296,2=0.948) and 3750400t (=0.0355,2=0.957), respectively, while from the Schaefer’s model with these two error assumptions the corresponding estimates are 4288870t (=0.0462,2=0.953) and 4697484t (=0.0219,2=0.958), respectively. In ECS, theestimates withfrom the Fox model with two error assumptions are 5002601t (=8244.4,2=0.916) and 3292739t (=0.6055,2=0.901), respectively, while from the Schaefer model are 5502317t (=5352.9,2=0.916) and 4310149t (=0.4846,2=0.901), respective- ly. In SCS, the estimateds withfrom the Fox mo- del and two error assumptions are 3006917t (=0.0177,2=0.972) and 2942940t(=0.0285,2=0.987), respectively, while from the Schaefer model are 3812418t (=0.0163,2=0.971) and 3506701t (=0.0174,2=0.971), respectively. The2values are greater than 0.9, indicating a good fit of the models to the data.

Table 2 Surplus production model outputs from CEDA

Notes:, intrinsic population growth rate;, carrying capacity;, catchability coefficient;, maximum sustainable yield;, coef- ficient of variation for;yield, replacement yield;2, coefficient of determination.

The observed and estimated catches in the three coastal waters from 1980 (ECS from 1982) to 2018 are shown in Fig.3, using the four models in CEDA. The models fitted the data well, and the estimated catches from the four modelsare similar. The catch estimates show a trend consistent with the observed catches in YBS and SCS, but there are some discrepancies in ECS, indicating that the results for ECS may not be as reliable as those for YBS and SCS. We be- lieved that the results obtained from SCS are optimum.

3.2 ASPIC

Table 3 shows the results for fisheries in the two coas- tal waters with IP values ranging from 0.1 to 0.9 in ASPIC, which illustrates that ASPIC is sensitive to the IP values. In ECS, no viable results were obtained because the2values were negative (Section 4.4).

Results from the nonequilibrium models of Fox and Logistic applying ASPIC are shown in Table 4, and the IP values were the same as in the CEDA. The2values of the two models in three coastal waters are low or even negative, which indicates poor fitting. In YBS, theestimate withfrom the Fox model is 3517000t (=0.0403,2=0.196) and from Logistic model is 4823000t (=0.0410,2=0.227). In ECS, theestimate from the Fox model is 4050000t (2=−0.149) and from the Logistic model is 4072000t (2=−0.129). In SCS, theestimate withfrom the Fox model is 2870000t (=0.0434,2=0.208) and from the Logistic model is 3498000t (=0.0343,2=0.246).

Fig.3 Observed (dot) and expected (line) catch (tons) from four CEDA models in three coastal waters. Fox-normal, Fox model with normal error distribution; Fox-lognorm, Fox model with lognormal error distribution; Schaefer-norm, Schaefer model with normal error distribution; Schaefer-lognorm, and Schaefer model with lognormal error distribution. YBS, Yellow and Bohai Seas (1980–2018, IP=0.2); ECS, East China Sea (1982–2018, IP=0.3); SCS, South China Sea (1980– 2018, IP=0.2).

Table 3 A Surplus production model incorporating covariates (ASPIC) results for fishery in YBS and SCS with IP ranging from 0.1 to 0.9

()

()

AreaModelIPMSY (t)CVKqBMSYFMSYfMSYR2 Fox0.52.695E+060.0716 3.154E+085.925E−091.160E+082.323E−023.921E+060.197 0.63.306E+060.0960 2.261E+086.755E−098.319E+073.975E−025.884E+060.167 0.85.189E+060.1141 2.366E+085.024E−098.703E+075.963E−021.187E+070.109 0.96.565E+060.2833 4.241E+082.549E−091.560E+084.208E−021.651E+070.011 0.16.187E+060.0330 1.707E+095.663E−098.533E+087.250E−031.280E+060.248 0.23.498E+060.0343 8.574E+085.640E−094.287E+088.159E−031.447E+060.246 SCS0.32.680E+060.0379 5.761E+085.602E−092.880E+089.305E−031.661E+060.243 0.42.360E+060.0430 4.380E+085.535E−092.190E+081.078E−021.947E+060.240 Logistic0.52.279E+060.0470 3.592E+085.413E−091.796E+081.269E−022.344E+060.235 0.62.373E+060.0641 3.149E+085.172E−091.575E+081.507E−022.914E+060.228 0.72.643E+060.0801 3.011E+084.684E−091.505E+081.756E−023.749E+060.216 0.83.061E+060.0307 3.358E+083.740E−091.679E+081.823E−024.875E+060.198 0.92.732E+060.1487 5.514E+082.078E−092.757E+089.910E−034.770E+060.177

Notes:, maximum sustainable yield;, coefficient of variation for;, carrying capacity;, catchability coefficient;B, stock biomass giving;F, fishing mortality rate at MSY;f, optimum fishing effort at; and2, coefficient of determination.

Table 4 Surplus production model outputs from ASPIC

Notes: Same as those in Table 3.

Fig.4 Observed (dot) and estimated (line) CPUE (t(kw)−1) from two ASPIC models (Fox model and Logistic model) in three coastal waters. YBS, Yellow and Bohai Seas (1980–2018, IP=0.2); ECS, East China Sea (1982–2018, IP=0.3); and SCS, South China Sea (1980–2018, IP=0.2).

Fig.4 shows the observations and estimates of CPUEs in the three coastal waters using the Fox and Logistic mo- dels in ASPIC. It can be observed that CPUE estimates from these two models are close but inconsistent with the trend of the CPUE observations, indicating poor fittings of the models. We believe that the results obtained from the SCS are optimum.

4 Discussion

4.1 CEDA

Table 1 shows that the three SPMs in the CEDA are sen- sitive to the choice of initial population relative to carry- ing capacity. Thevalues were estimated using the boot- strapping confidence-limit method. In YBS and SCS, theofestimates are small when the IP is set small, and are large when the IP is set large. When IP is between 0.1 and 0.9, theestimates fluctuate greatly; in contrast, they are more stable with normal error distributions than with lognormal distributions. Thevalues esti- mated by the Fox model are lower than those by the Scha- efer model, which is consistent with the traditional theory that the Fox model is more conservative (Panhwar, 2012). In ECS, theestimates are different and irre- gular when IP changes from 0.1 to 0.9; thus, it may be dif-ficult to compare the results of the abovementioned mo- dels. From Table 2, the2values from the CEDA are higher than 0.9, showing a good fit to the data, especially in SCS. Because of the inherent properties of SPMs, which are the same as for any other fish-stock assessment model, the mo- del output should be treated with caution. Therefore, we believe that theestimates from the Fox model with normal error distribution were more stable and reasonable. Therefore, we suggest thefor the fisheries in YBSand SCS are 3751369 and 3006917t, respectively. Because of the uncertainty about the results of ECS, thevalue may not be well determined, but we may still use theestimates from 3292739 to 5502317t in ECS as reference points.

4.2 ASPIC

The SPMs of the Fox and Logistic models were used in ASPIC. Table 3 shows that the two SPMs in ASPIC are sensitive to the choice of initial proportion, which is contrary to the common belief that ASPIC is not sensitive to the IP values (Wang and Liu, 2013; Memon., 2015; Xu., 2015). Kalhoro(2015) also indicated that ASPIC was sensitive to the IP values, though he calculated an appropriateestimate (with a2value above 0.8) for the Greater LizardfishFishery in Pakistan. The reason is unknown and further work is required to investigate this. Theestimates may vary when IP changes from 0.1 to 0.9, and the2estimates from these two models for the three coastal waters are less than 0.25, even negative, as observed in ECS. These resultsin- dicate that ASPIC did not perform well with the Chinese data. Finally, we recommend the results obtained from the same IP values in CEDA as references,., thebe- tween 3517000–4823000t in YBS, 4050000–4072000t in ECS, and 2870000–3498000t in SCS.

4.3 Status of Fishery Resources in China’s Coastal Waters

The results show that CEDA and ASPIC can get similar results (Tables 2, 4), but both are sensitive to the setting of IP. The2values in the CEDA fitting are greater than 0.9, much higher than those in the ASPIC assessments. There- fore, CEDA might be more suitable for the assessment of China’s coastal capture fishery. According to the results of CEDA and ASPIC, theestimates of the fisheries in the YBS, ECS, and SCS might be 3751369, 3292739–4072000, and 3006917t, respectively, and the totalin all the China seas lies between 10.05 and 10.83 million tons. In 2018, the total catch of marine fishery in the YBS was 3.18 million tons, lower than the estimated. In contrast, the total catch from the ECS was 4.17 million tons, higher than the estimated. In SCS there was a total catch of 3.10 million tons, also higher than the estimated. Domestic marine fisheries in Chinese coastal waters in 2018 produced 10.45 million tons, approximately equal to the estimated. Therefore, the fisheries in the coastal waters along China has been fully exploited, and the fisheries must be managed in a precautionary manner. Zhai and Pauly (2019) evaluated 21 economically important, trawl-caught species in China’s coastal seas and found that all species suffered from overfishing. Liang and Pauly (2020) confirmed that both catches and biomass in the Bohai Sea exhibited reductions in the mean tropic levels, implying that substantial changes have occurred in the underlying structure of the Bohai Sea ecosystem.

4.4 Limitations and Concerns

There have been many studies using CEDA and ASPIC to estimate. Wang and Liu (2013) evaluated the hair- tail () fishery in the ECS with these two methds and both obtained good results. Xu. (2015)assessed the Albacore fishery in the southern Atlantic Ocean and believed that ASPIC is more suitable for the fishery. The studies of Memon. (2015) and Kalhoro. (2015) for different fisheries in Pakistani waters show- ed that both softwares could achieve reliable results, al- though ASPIC had better2values. Karim. (2019) studied the hilsa shad () fishery in Bangladesh and got betterestimates form CEDA. Among these studies, only Xu(2015) standardizes the CPUE data. In this study, thevalues calculated using different models are very different and the CPUE data are not standardized, which may lead to some degree of uncertain- ty; therefore, the use of these two softwares may help mi- nimize uncertainty in the obtained results.

For the ECS, results from both softwares do not seem desirable, and we guess that this may be caused by the data itself. The fishing effort data we used were estimated by summing up fishing boat powers in each province aroundeach marine area. Fishing activities do not follow the boun- dary division between marine areas, and bias in effort data is unavoidable. We combined the data from the YBS to reduce data deviation, and this had a positive effect. How- ever, we can not reduce data uncertainty in the ECS, whichcan actually lead to the failure of sensitivity analysis. There- fore, the results of ECS were not presented in Table 3. The negative2value is owing to the formula used to calculate adjusted2, which becomes negative when the model fit is very poor (Table 4). Both softwares had the best eva- luation results for the SCS, wherein fishery activities are relatively separate from the other Chinese marine areas.

4.5 SPMs

Ecosystem-based resource assessment is a current research hotspot. However, as Zhan (1995) has mentioned, in the process of multispecies assessment, there are three difficulties that need to be resolved: the interactions among fishing activities, the interactions among fish species, and the difficulty in obtaining data. Study of multispecies fish- ery resources requires the introduction of ecology theories and methods, which leads to the complexity of modeling and difficulty in obtaining data. SPMs can be used for bothsingle species and multispecies resource assessment, and they are the simplest multispecies evaluation models (Zhan, 1995). When the catch is the total catch of mixed fish spe- cies and the effort is the total fishing effort exerted on mix- ed populations, theof the mixed population can then be calculated.

SPMs can only give a rough idea about the fish stock because the biological and environmental data are not con- siderd in this program (Quinn and Deriso, 1999). How- ever, even if other more complex and realistic models are implemented when more information is available,SPMs still can be employed as a comparison (Haddon, 2011). In the absence of enough information, SPMs may be a good choice with reliable results.

Acknowledgements

This study is supported by the project from the Food and Agriculture Organization of the United Nations (FAO) (No. GF.FIRFD.RA20403020400). We thank Dr. Yimin Ye of FAO for comments.

Efron, B., and Tibshirani, R. J., 1993.. Chapman & Hall, London, UK, 436pp.

Fishery Bureau of Agriculture Ministry of China, 1980–2019.. China Agriculture Press, Beijing, 17-85 (in Chinese).

Fox, W. W., 1970. An exponential surplus-yield model for optimizing exploited fish populations., 99 (1): 80-88, DOI: 10.1577/1548-8659 (1970)99<80:AESMFO>2.0.CO;2.

Haddon, M., 2011.2nd edition. Chapman & Hall/CRC Press, New York, 285- 333.

Hilborn, R., and Walters, C. J., 1992.. Chapman and Hall, New York, 570pp.

Hoggarth, D. D., Abeyasekera, S., Arthur, R. I., Beddington, J. R., Burn, R. W., Halls, A. S.,., 2006.(). Food and Agriculture Organization of the United Nations, Rome, 261pp.

Jacobson, L. D., Cadrin, S. X., and Weinberg, J. R., 2002. Tools for estimating surplus production and FMSYin any stock assessment model., 22: 326-338, https://doi.org/10.1577/1548-8675(2002) 022<0326:TFESPA>2.0.CO;2.

Ji, Y. P., Liu, Q., Liao, B. C., Zhang, Q. Q., and Han, Y. N., 2019. Estimating biological reference points for largehead hairtail () fishery in the Yellow Sea and Bohai Sea., 38 (10): 20-26, DOI: 10.1007/s13 131-019-1343-4.

Kalhoro, M. A., Liu, Q., Memon, K. H., Waryani, B., and Soo- mro, S. H., 2015. Maximum sustainable yield of Greater li- zardfishfishery in Pakistan using the CEDA and ASPIC packages., 34 (2): 68- 73, DOI: 10.1007/s13131-014-0463-0.

Kang, B., Liu, M., Huang, X., Li, J., Yan, Y. R., Han, C. C., and Chen, S. B., 2018. Fisheries in Chinese seas: What can we learn from controversial official fisheries statistics?, 28 (3): 503-519, DOI: 10.1007/s11160- 018-9518-1.

Karim, E., Liu, Q., Sun, M., Barman, P. P., Hasan, S. J., and Hoq, M. E., 2019. Assessing recent gradual upsurge of marine cap- tured Hilsa stock () in Bangladesh., 4 (4): 156-165, DOI: 10.1016/j.aaf.2019.02. 005.

Liang, C., and Pauly, D., 2020. Masking and unmasking fishing down effects: The Bohai Sea (China) as a case study., 184: 105033, DOI: 10.1016/j.ocecoa man.2019.105033.

Memon, A. M., Liu, Q., Memon, K. H., Baloch, W. L., Memon, A., and Baset, A., 2015. Evaluation of the fishery status for King Soldier Breamin Pakistan using the software CEDA and ASPIC., 33 (4): 966-973, http://dx.doi.org/10.1007/ s00343-015-4330-4.

Panhwar, S. K., Liu, Q., Khan, F., and Pirzada, J. A., 2012. Ma- ximum sustainable yield estimates of Ladypees,(Forssk), fishery in Pakistan, using the ASPIC and CEDA packages., 11 (1): 93- 98, https://doi.org/10.1007/s11802-012-1880-3.

Pella, J. J., and Tomlinson, P. K., 1969. A generalized stock pro- duction model., 13(3): 416-497.

Prager, M. H., 2005. A stock-production model incorporating co- variates (version. 5) and auxiliary programs. CCFHR (NOAA) Miami laboratory document MIA-92/93-55, Beaufort Laboratory Document bl-2004-01.

Quinn, T. J., and Deriso, R. B., 1999.. Oxford University Press, New York, 542pp.

Schaefer, M. B., 1954. Some aspects of the dynamics of populations important to the management of the commercial marine fisheries., 1 (2): 23- 56.

Su, S., Tang, Y., Chang, B. W., Zhu, W. B., and Chen, Y., 2019. Evolution of marine fisheries management in China from 1949 to 2019: How did China get here and where does China go next?, 21 (2): 435-452, DOI: 10.1111/faf. 12439.

Wang, Y., and Liu, Q., 2013. Application of CEDA and ASPIC computer packages to the hairtail () fishery in the East China Sea., 31 (1): 92-96, DOI: 10.1007/s00343-013-2073-7.

Xu, Y. W., Zhang, K., and Chen, Z. Z., 2015. Stock assessment for the albacore fishery in the southern Atlantic Ocean base on CEDA and ASPIC programs., 3: 45-54, DOI: 10.13984/j.cnki.cn37-1141.2015. 03.006 (in Chinese with English abstract).

Zhai, L., and Pauly, D., 2019. Yield-per-recruit, utility-per-re- cruit, and relative biomass of 21 exploited fish species in Chi- na’s coastal seas., 6: 724, DOI: 10. 3389/fmars.2019.00724.

Zhan, B. Y., 1995.. China Agriculture Press, Beijing, 224-228 (in Chinese).

Zhao, W. W., and Shen, H. H., 2016. A statistical analysis of Chi- na’s fisheries in the 12th five-year period., 1: 41-49, DOI: 10.1016/j.aaf.2016.11.001.

. E-mail: qunliu@ouc.edu.cn

August 24, 2020;

October 9, 2020;

January 13, 2021

© Ocean University of China, Science Press and Springer-Verlag GmbH Germany 2021

(Edited by Qiu Yantao)