Shellfish toxicity in the Gulf of Maine
Using Bayesian networks to link river inputs with ecosystem indicators in the Gulf of Maine
Funding source: U.S. Environmental Protection Agency
Personnel: M. Borsuk (PI), A. Nair (Dartmouth); A. Thomas, R. Weatherbee (U. of Maine); I. Alameddine (Duke)
Synopsis: Routine monitoring along the coast of the Gulf of Maine (GoM) reveals shellfish toxicity nearly every summer, but at varying locations at varying times. The shellfish beds found to be toxic are then closed to protect public health, resulting in significant commercial and tourist based economic losses. Many previous studies have looked at the relationship between environmental factors and the presence of Alexandrium fundeyense, the dinoflagellate that causes shellfish toxicity in the GoM. Our study instead investigates relationships between environmental factors (including sea-surface temperature (SST) gradients, wind speed and direction, and river flow) and the occurrence of shellfish toxicity.
First, twenty-one years of toxicity data (1985-2005) on Mytilus edulis (blue mussels) and Mya arenaria (softshells) from about 90 monitoring stations along the coastlines of New Hampshire, Maine and Massachusetts were used to identify eight archetypical patterns of seasonal toxicity, using a combination of statistical curve fitting and hierarchical cluster analysis. These patterns explained 94% of the variability across all year/location combinations.
Next, the geographically distributed historical monitoring locations were grouped into five distinct groups for Mytilus and three for Mya, each displaying similar seasonal toxicity patterns across all the years, again using hierarchical clustering.
Finally, the annual toxicity patterns within each spatial group were related to environmental variables using classification trees. Satellite-derived measures of SST were used to construct along-shore and cross-shore gradients as indicators of annual circulation patterns. These variables were then used together with wind and river discharge measures, all at a monthly scale, as independent variables in the classification trees. July average cross-shore SST gradient was found to be the strongest and most consistent correlate with toxicity patterns across all station groupings, followed by June cumulative downwelling wind strength. April river discharge and May cross-shore SST gradient play tertiary roles. These results suggest that the existence of a cross-shore connection between coastal areas and offshore waters is an important factor leading to shellfish toxicity, likely by providing a transport route between offshore Alexandrium populations and near-shore shellfish beds.