The production of global capture fisheries has remained stable over the last 10 years (FAO, 2020). The depletion of fish stocks has been increasingly reported in many marine ecosystems (Link and Watson, 2019). In addition to controlling catch and fishing effort, hatchery programs have been widely applied to fisheries aiming to improve the health of threatened or endangered natural resources (Taylor et al., 2017; Lorenzen et al., 2021). Globally, hatchery programs, such as pink salmon in Prince William Sound and European eel in River Ceira, have been practiced for several decades (Amoroso et al., 2017; Félix et al., 2020). The hatchery programs’ socio-economic benefits have been perceived such as seafood revenue and harvest opportunities for fishermen (Kitada, 2018). Many hatchery species, such as Pacific salmon, are recognized as keystone species in local ecosystems (Tatara and Berejikian, 2012). Hatchery-released individuals can mitigate the fishing pressure to the wild populations and help maintain the stability of local food webs (Anderson et al., 2020).

There are several risks associated with hatchery programs in a marine ecosystem. A large number of released hatchery individuals not only reduce the genetic diversity (Bell et al., 2005; Hagen et al., 2020), but also result in the competition of food and habitats with wild populations (Kitada, 2020), leading to calling for reducing the negative ecological impacts of hatchery programs (Osborne et al., 2020). A large number of released fish can affect the local trophic interactions. The combined hatchery and wild populations may temporarily exceed the local carrying capacity. This can deteriorate the status of the targeted stocks (Tatara and Berejikian, 2012). The cost and benefit of hatchery releases play an important role in the decision-making process. An increasing number of studies are now focusing on balancing economic profits and ecosystem impacts (Camp et al., 2017).

Many stocks of commercial importance are declining along the Chinese coast due to prolonged high fishing pressure (Liang and Pauly, 2020). Fenneropenaeus chinensis is a short-lived, fast-growing shrimp with a high commercial value in Chinese waters. Most of Fenneropenaeus chinensis individuals will die after spawning, and the span life is less than 2 years (Wang et al., 2006). High fishing pressure threatens the sustainable exploitation of F. chinensis stock. The hatchery program, which releases hatchery F. chinensis individuals into the ecosystem, has been developed and implemented over the past decades to rebuild the stock and maintain fishery productivity (Su et al., 2020). Hatchery programs cover the water areas from the East China Sea to the Bohai Sea. F. chinensis releases have been found to increase incomes and create employment opportunities for local residents. However, hatchery programs seemed to fail to prevent the decline of the wild population.

Hatchery release strategies need to consider release size, release habitat, release season, density-dependent effects, and predation processes (Bell et al., 2005). Single-species approaches have been developed to evaluate the profit benefits and risks (Lorenzen, 2005; Ye et al., 2005). Taylor and Suthers (2008) used the Predatory Impact Model to evaluate the optional release strategies associated with prey-predator relationships. In China, F. chinensis hatchery programs are now more likely to enhance the fishery production rather than to rebuild the wild population (Wang et al., 2006). F. chinensis fishery has become increasingly dependent on hatchery releases. For this reason, the number of released F. chinensis has increased every year. But the recapture rates are lower than 10% in many places (Wang et al., 2014). In addition to economic profits, there is a need to identify whether the released F. chinensis individuals influence the stability of local food webs. Studies applied the static food web model Ecopath to estimate the maximum ecological carrying capacity of F. chinensis releases (Lin et al., 2013, 2018; Wang et al., 2016). These studies are inadequate for advising to implement F. chinensis releases in a dynamic perspective. Moreover, the changes of trophic interactions caused by fishing may complicate the management of F. chinensis releases.

Responsible approach to marine stock enhancement requires a set of principles to ensure success and avoid mistakes (Blankenship and Leber, 1995; Lorenzen et al., 2010). Management strategy evaluation (MSE) can facilitate fisheries management achieving by assessing candidate strategies (Punt et al., 2016). Ecosystem models are recognized as a necessary complement to support managing fisheries at an ecosystem level (Hilborn, 2011). In this study, we use a published ecosystem model OSMOSE-JZB (Object-oriented Simulator of Marine ecOSystEms) to evaluate the impacts of F. chinensis release in Jiaozhou Bay, China in a dynamic perspective. OSMOSE-JZB can simulate the seasonal variations in Jiaozhou Bay ecosystem (Xing et al., 2017) and inform fisheries management using simulation (Xing et al., 2020a,2021a,2021b). The objectives of this study are to (1) analyze the economic profit with increasing F. chinensis releases; (2) investigate ecological impacts of F. chinensis release; and (3) evaluate fishing impacts on the F. chinensis release and ecosystem. The study can help us better understand economic profits and ecological impacts of hatchery programs.

Seasonal variations appeared in biomass, abundance, mean trophic level, predation mortality, and size structure for F. chinensis (Figure 3). F. chinensis biomass increased slowly after the second half of July and peaked in the second half of August (Figure 3A). F. chinensis abundance showed the largest decline during the 1st month after the release (Figure 3B). The trophic level of F. chinensis increased rapidly before the first half of July and then remained stable (Figure 3C). F. chinensis experienced high predation mortality (higher than 0.4) during the first half month of release (Figure 3D). In July and August, the dominant size group of F. chinensis was 10–16 cm, which accounted for over 60% of the total biomass (Figure 3E).

The scenario S1 simulated 1-year F. chinensis hatchery program under different released amounts. In this scenario, the estimated biomass and estimated commercial value (ECV) of F. chinensis was highest when 234 million F. chinensis individuals (Multiplier = 2.6) were released (Figures 4A,C). The abundance increased with the increase of released amount (Figure 4B). The economic profit (EP) increased rapidly when the annual released amount increased from 18 million (Multiplier = 0.2) to 90 million (Multiplier = 1.0). The EP increased slightly when the released amount was between 90 million and 144 million, and then declined when the released amount was higher than 198 million (Multiplier = 2.2) (Figure 4D). The ratio of ECV and Cost (REC) was highest with the annual released amount of 18 million (Multiplier = 0.2), and then decreased rapidly along with increasing released amount (Figure 4E). The proportion of individuals larger than 10 cm exceeded 85% when the released amount was between 18 million to 108 million (Multiplier = 1.2). There was a downward trend for the proportion of individuals larger than 10 cm with increasing released F. chinensis individuals (Figure 4F).

The scenario S2 simulated a long-term hatchery program that F. chinensis releases implemented for 10 years and then stopped during following 10 years. In the scenario S2, the sum of economic profit in the release stage (i.e., total EP) was highest under P1 and lowest under P2 especially when the 10-year released amount was more than 900 million (Rx1). The sum of economic profit under P2 to P4 declined when the 10-year released amount increased from 900 million to 1980 million (Rx2.2). The annual economic profit (i.e., EP) did not exceed 2.8 million RMB despite increasing the released F. chinensis amount (Figure 5A). The relative changes of two large fish groups (SP6 and SP11) biomass were bigger than other wild species. The biomass of two large fishes at the end of recovery stage (Bio₁₀₀) were lower than those before F. chinensis released (Bio₈₀). It suggested that they did not recover to the previous levels. The large amount of F. chinensis (Rx2.2) had negative impacts on small shrimps (SP1). In contrast to the two large fishes, the biomass of small shrimps could rose back to the status before releasing. It suggested large fishes recovered slowly compared to small shrimps. Additionally, F. chinensis releases seemed not to influence other wild species (Figure 5B).

The scenario S3 simulated a long-term hatchery program under different fishing pressure. For this scenario, the sum of economic profit had small changes after fishing changed. The sum of economic profit increased when released amount increased (Figure 6A). Fishing had more negative impacts on the two large fishes (SP6 and SP11) than other spices. The Bio₉₀/Bio₈₀ values for Sebastes schlegelii (SP6) were more than 1.2. It suggested that reducing fishing pressure could offset the negative impact of F. chinensis release. The Bio₉₀/Bio₈₀ of Johnius belangerii (SP11) under the medium (Rx1) and high (Rx2.2) levels of F. chinensis releases were less than 1. It suggested the impacts of F. chinensis release were bigger than fishing when the released amount was large (Figure 6B).

The study provided the insights of F. chinensis population dynamics. Mortality of fish populations is strongly density-dependent in the early stages of the life cycle (Lorenzen and Camp, 2019). High M_natural–P1 value suggested most of the released F. chinensis juveniles died within the first 2 weeks. One reason is that high density has the negative impacts on the postrelease survival (Schloesser et al., 2021). Another possible reason is that some individuals died during the transportation process. Changes in environmental factors influenced F. chinensis especially for the species’ locomotive ability (Li et al., 2019). The mean trophic level of F. chinensis was lower than 2.95 at their initial stage of inhabiting the Jiaozhou Bay. They could not feed on large prey as they were too small (Figure 3C). The mean trophic level (TL) was 3.35 after June similar to the results of stable carbon and nitrogen isotopes (TL = 3.27) in the Laizhou Bay, which is located in the north of the Shandong Peninsula (Li et al., 2021). Similar to Jiaozhou Bay, Laoshan Bay is a small bay located in the south of the Shandong Peninsula. Yuan et al. (2015) analyzed the survey data of F. chinensis release in the Laoshan bay during 2012. They reported that the size range of F. chinensis was 10.3∼12.7 cm on July 13, 11.4∼15.7 cm on August 1, and 13.1∼17.2 cm on August 17. Our predicted F. chinensis size was similar to Yuan et al. (2015) in July but smaller in August (Figure 3E). Intense shellfish culture reduced the ecosystem productivity in Jiaozhou Bay (Han et al., 2017). F. chinensis individuals needed more food to support the growth in August compared to July. Therefore, F. chinensis growth was slower in August (Figure 3E). The F. chinensis predation mortality was higher within the 1st month in Jiaozhou Bay than in other periods (Figure 3D) due to small F. chinensis individuals were more susceptible to predation.

Releasing hatchery individuals has become the preferred approach to enhance fishery productivity. High population density may limit individual growth as a result of competition of food and habitat (Kitada, 2018). The economic profit (i.e., EP) peaked when the annual released F. chinensis were 198 million (Figure 4B). However, the proportion of individuals which were larger than 10 cm (approximately 72%) at the end of August was much lower than scenarios whose annual released amount was below 90 million (Figure 4D). This may reduce the prospective benefit of F. chinensis release to the fishery production. The results highlighted the population density should consider in a long-term F. chinensis hatchery program. For example, the sum of economic profit in the release stage was maximized under the uniform release (P1), especially when the total released amount was large (Figure 5A). We found annual economic profit could not exceed 2.8 million RMB regardless of the increasing released amount of F. chinensis (Figure 5A). Therefore, we recommend that the annual released amount should be lower than 90 million individuals per year since the economic profit is high and individual size is large. The pattern P1 seemed to be a preferable release strategy that can avoid excessive release in a long-term hatchery program compared to other release patterns.

Historically, the F. chinensis metapopulation of wild individuals and artificially released individuals was up to 242 t in August (Wu et al., 1997). It exceed over four times of our optimal F. chinensis biomass. Intensive shellfish aquaculture especially for Ruditapes philippinarum culture is an apparent reason for Jiaozhou Bay failing to provide abundant food for F. chinensis growth as before. The shrimp biomass decreased from 4.70 t/km² in 1982 to 0.38 t/km² in 2015, while cultured R. philippinarum biomass increased from 25 t/km² to 100 t/km² (Ma et al., 2018). R. philippinarum culture had marked negative impacts on shrimp as they consumed a large amount of phytoplankton (Han et al., 2018). Shellfish aquaculture is important for local residents that creates billions of income about 230 million dollars per year (Han et al., 2017). It is difficult to remove shellfish aquaculture from the Jiaozhou Bay. Future F. chinensis release strategy should consider the impact of shellfish aquaculture such as food competition.

A responsible approach to stock enhancement requires to assess and manage ecological impacts (Lorenzen et al., 2010). 1-year F. chinensis release had small impacts on wild species (Figures 5, 6). The impact on other species is mainly attributed to (1) F. chinensis is a low-trophic level shrimp feeding on small size prey and influenced wild species mainly via food competition; (2) F. chinensis stayed in the simulated area less than 4 months; and (3) Ecopath model studies evaluated the ecological carrying capacity of F. chinensis biomass were 0.85 t/km² in Haizhou Bay (Wang et al., 2016), 2.95 t/km² in Laizhou Bay (Lin et al., 2013), and 0.70 t/km² in Bohai Sea (Lin et al., 2018). The F. chinensis density in our study was much lower than those in the Ecopath model studies. The simulations highlighted that the negative impacts of a long-term F. chinensis release on two large fishes (S. schlegelii and J. belangerii) should be considered (Figure 5B). The prey biomass to S. schlegelii and J. belangerii were abundant in the Jiaozhou Bay (Xing et al., 2017). The increase of F. chinensis biomass made a very small contribution to the growth of S. schlegelii and J. belangerii. S. schlegelii and J. belangerii spawned in the late spring and summer. F. chinensis biomass reached the peck in August (Figure 3). Large biomass of F. chinensis resulted in a strong food competition with S. schlegelii and J. belangerii juveniles (Xing et al., 2021a). Ecological processes can influence stock enhancements and fisheries (Johnston et al., 2018). The effectiveness of F. chinensis release seemed not to change when fishing pressure increased or declined by 20% (Figure 6A). Predator pressure to F. chinensis was low since June due to their large body size (Figures 3D,E). We found that managing fishing effort could offset the negative impacts of F. chinensis release (Figure 6B). It was noteworthy that F. chinensis resulted in a bigger decline of large fishes when fishing pressure was high.

Additional aspects should be considered using our approach to advise management. The release size can influence the effectiveness of stock enhancement and wild populations (Garlock et al., 2019). The sensitivity analysis highlighted the accuracy of estimation for short-term post-release mortality (M_natural–P1, the mortality of F. chinensis at the initial stage of release). Moreover, reducing M_natural–P1 could increase F. chinensis biomass (Supplementary Figure 3). Studies have demonstrated large release size can improve the effectiveness of stock enhancement (Hamasaki and Kitada, 2006; Ye et al., 2005). The cost of hatchery program will increase as the individuals at a large size have a higher price (Li, 2011). Future study will consider releasing the F. chinensis individuals at other body sizes. In addition to the release size, short-term post-release mortality is dependent on the season and release method (Leber et al., 1997; Hervas et al., 2010). Given the food web of Jiaozhou Bay has seasonal variations, the timing of F. chinensis release should be considered in the management. The economy indicators need to further improve in co-operating into fluctuations in market price and transport charge. Climate change can influence the population processes such as growth and reproduction (Szuwalski and Hollowed, 2016; Martino et al., 2019). Understanding the impact of environmental changes and shellfish culture is important for the decision-making process. Ecosystem models are usually “data-hungry,” and imprecise parameters will bias model outputs (Xing et al., 2020b). Although we compared our results to other studies, inadequate data certainly increase model uncertainty. It is difficult to involve all species in the modelled food web. This is a potential source of bias to our model predictions.

We evaluated the F. chinensis release in terms of economic profits and ecological impacts. Our findings highlight that the effectiveness of F. chinensis release cannot be improved by simply increasing the released amount. F. chinensis release gradually influences non-released species instead of instantaneously resulting in a big change of fish community. The potential negative impact of a long-term F. chinensis release on the ecosystem should be considered in the decision-making process. We encourage implementing the hatchery program in a dynamic perspective and ecosystem consideration instead of simply emphasizing socio-economic benefits. Our modeling approach can be applied to other ecosystems and promote implementing hatchery programs in an ecosystem context.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

LX contributed to the conceptualization, carried out the data curation, formal analysis, and software, performed the methodology, and wrote the original draft. YC contribute to the conceptualization, performed the methodology, supervised the data, and wrote, reviewed, and edited the manuscript. KT performed the methodology and wrote, reviewed, and edited the manuscript. NB performed the methodology and carried out the software. YR carried out the data curation and funding acquisition and supervised the data. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.