Using satellite remote sensing to improve the prediction of scallop condition in their natural environment: Case study for Georges Bank, Canada

Abstract

The continuous, synoptical and high spatio-temporal resolution of thermal and visible satellite observations constitute an asset when characterizing and monitoring biogeochemical cycles in the oceans. In particular, they provide a unique insight into the hydrodynamics of the surface ocean and phytoplankton phenology. This information can be combined with in-situ observations of higher trophic levels to understand the functioning of the marine ecosystem, with potential use for fisheries and aquaculture. In the Gulf of Maine, off the eastern seaboard of North America, Georges Bank is one of the world's most productive natural sea scallop (Placopecten magellanicus) fisheries, which sustains economically valuable fisheries in Canada and the US. In Canada, the relationship between the scallop meat weight and its shell height (known as scallop condition) is used to obtain biomass estimates for stock assessment and thereby influences the fishing quotas. On Georges Bank, for a 100 mm shell height scallop, scallop condition has varied significantly interannually between 10.2 g and 19.5 g. Establishing a relationship between environmental factors and scallop condition would improve the forecasting of scallop biomass and reduce the uncertainty when determining the total allowable catch for the subsequent fishing season. In this study, satellite remote-sensed sea-surface temperature (SST) and chlorophyll-a concentration (Chla) coupled with in-situ measurements of scallop condition (SC) on Georges Bank were used to assess the relationship between SC and satellite-derived environmental conditions between 1985 and 2019 for SST and between 1998 and 2019 for Chla. The inter-annual variation of SC in May was found to be positively correlated (p-value <0.001) with the variation in SST in winter-spring months of the current and previous year of the SC survey. An empirical linear model for the estimation of SC was developed and validated using two independent datasets of satellite SST (r2 = 0.64, MAPD = 7%). Another model using cumulative SST in the previous year only was also provided (r2 = 0.58, MAPD = 7.7%), which could serve as an alternative choice when the SST products in the current year are not yet available for use. There was a weak relationship observed between SC and Chla; this is likely due to the influence of water-column vertical mixing on transport of phytoplankton to the benthic zone where the sea scallop resides. Our study indicates that a cost-effective satellite-derived SST product can be used to improve predictions of SC and provide useful information and decision support tools for managing scallop stocks.