Scaling issues in local productivity hotspots in marine ecosystems using remote sensing data: A case study in the Gulf of Maine

Abstract:

Productivity hotspots are traditionally defined as concentrations of relatively high biomass compared to global reference values. These hotspots often signal atypical processes occurring in a location, and identifying them is a great first step at understanding the complexity inherent in the system. However, identifying local hotspots can be difficult when an overarching global pattern (i.e. spatial autocorrelation) already exists. This problem is particularly apparent in marine ecosystems because values of productivity in near-shore areas are consistently higher than those of the open ocean due to oceanographic processes such as upwelling. In such cases, if the global reference layer used to detect hotspots is too wide, hotspots may be only identified near the coast while missing known concentrations of organisms in offshore waters. On the other hand, if the global reference layer is too small, every single location may be considered a hotspot. We applied spatial and traditional statistics to remote sensing data to determine the optimal reference global spatial scale for identifying marine productivity hotspots in the Gulf of Maine. Our iterative process measured Getis and Ord’s local G* statistic at different global scales until the variance of each hotspot was maximized. We tested this process with different full resolution MERIS chlorophyll layers (300m spatial resolution) for the whole Gulf of Maine. We concluded that the optimal global scale depends on the time of the year the remote sensing data was collected, particularly when coinciding with known seasonal phytoplankton blooms. The hotspots found through this process were also spatially heterogeneous in size, with bigger hotspots in areas offshore than in locations inshore. These results may be instructive for both managers and fisheries researchers as they adapt their fisheries management policies and methods to an ecosystem based approach (EBM).