Dynamic changes in American lobster suitable habitat distribution on the Northeast U.S. Shelf linked to oceanographic conditions

Abstract

AbstractAmerican lobster (Homarus americanus) supports one of the most valuable regional fisheries in the United States, with its abundance and distribution profoundly influenced by environmental conditions. To explain how lobster distribution has changed over time and assess the role of environmental variables on these changes, we used random forest classification and regression tree models to estimate occupancy and biomass in two seasonal periods. The occupancy models were fit to static and dynamic variables, which yielded model fits with AUC scores of 0.80 and 0.78 for spring and fall, respectively. Biomass models were fit with the same data and resulted in models explaining 61% and 63% of the spring and fall biomass variance, respectively. Significant variables scored in the formation of the regression trees were secondary productivity (i.e., zooplankton), bathymetry characteristics, and temperature. American lobster suitable habitat has changed regionally; habitat has increased in the Gulf of Maine and declined in Southern New England. There is also evidence of declining habitat along the inshore margin of the Gulf of Maine, which has been accompanied by a shift in occupancy probability offshore. Habitat suitability results from the random forest models provide insights on the structure and function of lobster habitat and context to understand recent population trends.