Effects of spatial scale on assessments of suspension bivalve aquaculture for productivity and environmental impacts

Proper evaluations of productivity and environmental impacts of bivalve aquaculture are necessary to achieve sustainable operations in coastal marine systems. Estimation of organic matter budgets using completely mixed one-box models provides an easy approach to help stakeholders of environmental management and aquaculture fisheries obtain an overview of the interactions between aquaculture and the environment. However, estimations and interpretations are highly dependent on the spatial scale of assessments because the distributions of aquaculture facilities and ecosystem structures are spatially heterogeneous. The objective of the present study was to examine how the spatial scale of assessments for the budgets of particulate organic matter (POM) affects the interpretation of the site-specific food supply for suspension oyster aquaculture and impacts of oyster biodeposition on local bottom environments. Based on the actual distribution of oyster farms in a temperate bay on the northeast coast of Japan, we estimated [food supply]/[food demand] (S/D) for oysters as well as the relative contribution of oyster biodeposition to the total deposition of POM (B_rel) in the bay along a gradient of spatial scales l from 0.5 m to 4000 m. Although food limitation for aquaculture oysters has been reported in the study bay, S/D was greater than 1 (ranging from 16 to 154) over the spatial scale gradient. S/D tended to be relatively high and overestimated for l < 200 m and l > 1400 m due to the nature of completely mixed one-box models, but relatively lower at intermediate scales. B_rel was higher in assessments at farms or smaller scales (e.g., l < 100 m) and was relatively lower at larger scales (e.g., <0.2 at the scale of >500 m). Based on our model-based calculations, the patterns of variation in S/D along the spatial scale gradient are considered to be universal; however, the scale at which S/D values are relatively low varies depending on site-specific conditions, including primary productivity and current velocity. Our findings suggest that estimated budgets of POM need to be carefully interpreted, accounting for model assumptions as well as the influence of the spatial scale of assessments.