Scale-dependent assumptions influence habitat suitability estimates for the American lobster (Homarus americanus): Implications for a changing Gulf of Maine

Bioclimate envelope models were evaluated for possible spatial nonstationarity for American lobster (Homarus americanus) in the Gulf of Maine (GOM). Five environmental variables were considered, including bottom water temperature, bottom water salinity, distance offshore, sediment grain size, and latitude. Three models accounting for varying degrees of spatial nonstationarity were developed and evaluated. Spatial scales developed in these models were determined by the patterns of the GOM coastal currents and the associated affects these currents have on the GOM oceanography. This study suggests that incorporating spatial nonstationarity into habitat suitability modeling may lead to improved models. Forecast plots for the 2028-2055 period revealed that traditional modeling techniques applied at coarse spatial scales estimated higher suitability of habitat for ju-venile lobsters than models at smaller spatial scales. For adult lobsters, coarse spatial model yielded higher suitability in both coastal waters in western GOM and farther offshore waters in eastern GOM, and generally estimated lower suitability in coastal eastern GOM waters and some offshore western GOM waters, compared to finer-scale spatial models. These results demonstrate how HSI estimates would vary based on the choice of spatial scales and its consequences. This information could benefit future stakeholders to prepare and adapt at more localized scales for changes that may occur in a region that is susceptible to climate change.

Automated summary

Bioclimate envelope models for American lobster in the Gulf of Maine were evaluated, and incorporating spatial nonstationarity was found to improve model accuracy. The choice of spatial scales had significant impacts on habitat suitability estimates, which is important for stakeholders in a climate-vulnerable region.