Eutrophication is a worldwide problem in coastal ecosystems that severely threatens environmental equilibrium (Nichols et al., 2019; Liu et al., 2020). Nutrient loads are very important drivers of eutrophication (Meier et al., 2019). Excessive nutrient enrichment causes alternated algal flourish and collapse cycles (Farmaki et al., 2014), which can influence food web and ecosystem.

Mariculture is one of the fastest-growing food-producing industries in the world with an annual production of over 63×10⁶ t in 2018 (FAO, 2020). The potential effects of the mariculture industry on its surrounding environment have attracted a large interest from both marine scientists and policymakers (e.g., Bambaranda et al., 2019; Joesting et al., 2016; Bannister et al., 2016; Carballeira et al., 2018). Mariculture is dominated by extractive culture of aquatic plants, filter-feeding bivalves, and fed culture of marine finfish and crustacean. Most studies have illustrated that mariculture has remarkable effects on the environment (Holmer, 2010) mainly due to waste load, which promotes eutrophication of coastal waters (Boyd, 2003; Farmaki et al., 2014; Filgueira et al., 2017). Two generally accepted views are that fed culture (cages and pond) is the source of nitrogen (N) and phosphorus (P) (Wang et al., 2012; Bannister et al., 2016; Carballeira et al., 2018), and photosynthetic seaweed is the sink of nutrient (Xiao et al., 2017). Due to different estimation methods, the calculations of source-sink effect of bivalve are different. The major eco-physiological pathways in which bivalves interact with coastal nutrient cycling are filtration of sestons from water column, nutrient storage in tissue and shell (growth), excretion of inorganic metabolic waste, and production and mineralization of biodeposits (Newell, 2004). The calculation methods on N and P loading of bivalve mariculture include, taking account of biodeposition+excretion (Gallardi, 2014), growth–excretion–biodeposition (Bouwman et al., 2013), growth+biodeposition–excretion (Christensen et al., 2003) and growth+biodeposition–excretion–remineralization (Richard et al., 2007), which lead to significantly different results of bivalve being source or sink of nutrient.

Most of the previous studies on nutrient loads mainly focused on one specific species or on the region, such as a relatively short distance (i.e., hundreds meters) from the farms. However, the discharge of N and P could transport and distribute in large-scale area, and the spatial-temporal distribution and impact on the ecosystem depend on the ocean characteristics (current, water depth, temperature, etc.) and mariculture practices, such as culture density, species, etc. (Sarà et al., 2011; Filgueira et al., 2017; Oh et al., 2015; Wu, 1995). It is important to understand the interactions between mariculture and environment on ecosystem scale and provide management strategies in the ecosystem approach to aquaculture (Aguilar-Manjarrez et al., 2017). There were a few studies on national or global scale about nutrient loads of fed culture type of cage finfish culture (Bouwman et al., 2013; Cai and Sun, 2007; Islam, 2005), extractive culture types of shellfish and seaweed farming (Bouwman et al., 2013), and on both culture types (Wang et al., 2020; Bouwman et al., 2013).

China is one of the fastest-growing mariculture countries in the world. To protect and restore the marine ecological environment and to pursue a sustainable development of mariculture, China currently proposes a nationwide mariculture management strategy. Therefore, an improved understanding of the influence of mariculture on both province and national scale is crucial for appropriate regulation and management of mariculture operations at the ecosystem level. The potential environmental impact of mariculture and how to keep mariculture sustainable development are the issues of concern (Cao et al., 2007; Yang et al., 2017; Shen et al., 2018). In this paper, we aimed to (1) evaluate N and P released from 4 different typical culture systems in China; (2) quantify the amount of N and P loading from mariculture on province scale in China; (3) analyze the association between N and P loading and culture structures and provide strategic advice on mariculture.

From this research, we found that the N and P loads were negative from the national perspective in China, which indicates that through mariculture, a large amounts of N and P were removed from the coastal environments every year. One of the reasons was Chinese mariculture characteristics. In China, the maricultural species were mainly extractive low-trophic macroalgae and filter-feeding bivalves, and this culture structure was relatively stable over the past 30 a (Fig. 1). Compared with other countries in the world, China’s mariculture had the characteristics of high production, large scale, rich diversity, low nutritional level, and high ecological efficiency (Tang et al., 2016). Fed culture types (e.g., cages and pond) brings bigger nutrient loads than the extractive culture types comparatively (Wang, et al., 2012; Bannister et al., 2016; Carballeira et al., 2018).

Ours results showed N and P loads from mariculture of 2019 were 5.08×10⁴ t and 2.9×10³ t respectively, which were remarkably lower compared with the 0.4×10⁶ t/a (according to N) and 0.06×10⁶ t /a (according to P) in 2017 and the 0.184×10⁶ t/a (according to N) (Wang et al., 2020) and 0.022×10⁶ t/a (according to P) in 2010 (Zhang et al., 2015), probably due to differences in the approaches used. The bivalves reared using raft and tidal flat systems (the yield in 2019 was 11×10⁶ t, occupy nearly 90% of total mariculture yields of bivalves) were ignored in Zhang et al. (2015), only fish, shrimps, crabs, and mollusks in pond system were estimated, which would overestimate the nutrient loads. In addition, the fish production and waste coefficient were different comparing with Wang et al. (2020). In this paper, the yield of fish from re-circulating aquaculture system were ignored because the waste coefficient was low. For the fish cultured in pond, not all the pollution materials were released to coast; therefore, we calculated the nutrient loads based on PDC, which was lower than PPC. The above reasons led to the underestimation of the nutrient loads in this paper. Another main problem was the nutrient loads of bivalves. In this paper, we consider bivalve mariculture as a negative nutrient loads and subtract the N and P removed by bivalves from the total nutrient loads, while the calculation method was just opposite in Wang et al. (2020). As filter feeders, bivalves acquired their food (mostly, phytoplankton) directly from the water column, and part of them were released into water, which includes biodeposits (feces and pseudo-feces) and dissolved waste (such as ammonia); and part of ingestion food was utilized as growth, which was eventually harvested from the coast (Tang et al., 2011). The farming of bivalves begins with the seeding of spat and ends with harvesting. Therefore, nutrient extraction by bivalves represented a cumulative effect from bivalve feeding activity over the full growth cycle and could be readily quantified from harvest data (Petersen et al., 2016). Phytoplankton as the main food of bivalve has 2 sources, from primary production of local phytoplankton, and from transportation of outside phytoplankton (Sun et al., 2021). In area with relatively slow flow rate, phytoplankton mainly comes from local photosynthetic production. Bivalve cultivation sites are situated in nutrient-rich coastal areas mostly, thereby taking advantage of high primary production rates to achieve rapid growth. Under excessive nutrient availability, filtration of phytoplankton may help to prevent or overcome eutrophication problems (Brigolin et al., 2009; Ferreira et al., 2014). While the area with higher flow rate, usually in relatively open area, phytoplankton mainly comes from transportation outside the farming area. In such harsh areas, most of biodeposits from bivalve can be transported to outside and the amount that settled in the bottom bed of farming area is limited. According to current research reports, after more than 30 a of longline bivalves and seaweed culture in the Sanggou Bay, the organic pollution of the sedimentary environment is very light (Zhang et al., 2020b). Recently, Bivalve farming was widely considered to be an efficient means of nutrient removal in coastal areas affected by eutrophication as a result of the consumption of anthropogenic sources of N and P contained within phytoplankton and the sequestration of these excess nutrient into harvestable bivalve biomass (Petersen et al., 2016).The major eco-physiological pathways in bivalves interact with coastal nutrient cycling are filtration, growth, excretion and biodeposits. It is complex process that closed with local environment and culture density, the excretion and biodeposits regeneration of nutrient could stimulate the growth of phytoplankton (Ferreira et al., 2014). It will overestimate the nutrient loads of bivalve and even led to the opposite of positive and negative results, if only focus on excretion and biodeposition process. Therefore, we think it is better not to consider the process and just look at the result. Wang et al. (2020) estimated the nutrient loads of bivalves by the approach of “growth-excretion-biodeposition”, which is one of the reasons that results are significantly higher than this article.

Water quality was a critical component of health and welfare for marine organisms. Nutrient concentration and ratio could affect the composition, structure, dominance, growth, and biomass of planktonic communities and in farther the food web dynamics in the ecosystem (Oh et al., 2015; Price et al., 2015; Dempster et al., 2011). These results demonstrated mariculture may lead to changing element ratios (Table 5). The effects of bivalve mariculture on nutrient fluxes were complex (Sarà et al., 2011). If only consider part of their physiology, the results were different. For example, bivalves consume P and N in phytoplankton, and sometimes these nutrient remain undigested and accumulate in the benthic environment as biodeposits (Newell, 2004). Contrarily, bivalves can accelerate the N cycle via ammonia excretion (Dame, 1996) and benthic remineralization (Grant et al., 1995). These results showed that bivalve and seaweed culture could remove N and P from the coast of China, and the removed N:P were 35:1 and 12:1 for bivalves and seaweed farming, respectively (Table 3). As a result, bivalves farming could decrease the N:P of the water body by harvest which is consistent with report (Nixon et al, 1996). In contrast, seaweed farming could increase the N: P of water. The N and P released into the water body could pass through a complex biogeochemical process, including transport and transformation process, dictated the distribution and final fate. In terms of the total N and P, mariculture not only affected the N and P concentration in the receiving system but also affected its molar ratio. Different farming types had different effects. These results demonstrated that integration of different types of culture species could help to regulate the amounts of N and P, and N/P ratio, and reduce the environmental impact of mariculture.

It was reported that causal linkages between farming and eutrophication at the bay scale had been difficult to establish (Price et al., 2015). In this study, we got the same results. From the perspective of each region, the N and P loads of mariculture were distributed in southern regions (including Hainan, Guangxi and Guangdong) (Fig. 2), where the water quality was in the state of first-class water quality (Ministry of Ecology and Environment of People’s Republic of China, 2020). Rivers were the main source of N and P loads (Fig. 3), especially the Changjiang River and the Huanghe River Estuary were the severe eutrophication area in China, although mariculture in these areas was a negative discharge of N and P. We suggested increasing the amount of seaweed and bivalve farming in the areas of the river estuary.

The N and P loads of mariculture were negative from a national perspective, however, some regions such as Hainan and Guangxi had relatively large positive loads, and the N:P of some regions seriously deviated from the Redfield ratio (16:1) (e.g., in Zhejiang, the total N: P ratio was 81), which should be paid more attention. Through the large-scale concerns and understanding, it is recommended to control the N and P loads and N: P ratio through integrated farming of different trophic species. For Hainan and Guangxi, we suggested decreasing the amount of fed-type culture and appropriately increasing the extractive culture species. Generally, the development of China’s mariculture was relatively sustainable, and the mode and structural characteristics might be a good reference for other countries.

Although China was the largest and one of the fastest-growing mariculture countries in the world, the N and P loads from mariculture in coastal regions were negative from the national perspective. On the province scale, the nutrient loads which came from fed culture types were detrimental in Hainan and Guangxi; but in Shandong, Liaoning and Fujian, mariculture could remove N, P and C via bivalve and seaweed farming. Balancing different types of culture species could help to regulate the amount of N, P, and N: P ratio and reduce the environmental impact of mariculture. This article calculated overall results about N and P emissions from mariculture, and further research on local area effects (i.e., mariculture leads to the local accumulation of nutrient in sediment, and bivalves culture accelerates the nutrient cycle, etc.) is needed in the future.