Changes in species distributions have been widely associated with climate change. Understanding how ocean conditions influence marine fish distributions is critical for elucidating the role of climate in ecosystem change and forecasting how fish may be distributed in the future. Species distribution modeling (SDM) can enable estimation of the likelihood of encountering species in space or time as a function of environmental conditions. Traditional SDM approaches are applied to species data collected through standardized methods that include both presence and absence records, but are incapable of using presence-only data, such as those collected from fisheries or citizen science programs. Maximum entropy (MaxEnt) models are promising tools as they can predict species distributions from incomplete information (presence-only data). We used MaxEnt models to relate the occurrence records of seven marine fish species (Atlantic herring, Atlantic mackerel, butterfish, silver hake, spiny dogfish, Atlantic long-fin squid, and short-fin squid) from NOAA Northeast Fisheries Observer Program to environmental conditions. Environmental variables derived from remote sensing, such as sea surface temperature (SST), Chlorophyll-a, bathymetry, North Atlantic oscillation (NAO), and Atlantic multidecadal oscillation (AMO), were matched with species occurrence data to develop MaxEnt modeling of the fish species' spatial and temporal distributions in Northeast Shelf area.
Results/Conclusions
We developed maps of habitat suitability for the seven marine fish species and assessed the relative influence of environmental factors on their distributions at spatial and temporal scales. Overall, SST and Chlorophyll-a variables had the greatest influence on monthly distributions of these marine fish species, with bathymetry having a moderate influence and climate indices (NAO and AMO) having very little influence on their distributions. Across months, Atlantic herring distribution was most related to maximum SST values, Atlantic mackerel and butterfish distributions were most related to previous month's SST values, silver hake distribution was most related to previous month's Chlorophyll-a values and bathymetry, spiny dogfish distribution was most related to minimum SST values, and Atlantic long-fin squid and short-fin squid distributions were most related to previous month's Chlorophyll-a values. All seven species' distributions were most affected by previous month's Chlorophyll-a values in summer months, which can indirectly indicate the accumulative impact of food availability on these marine fish species when sunlight intensity increases. These MaxEnt models have the potential to provide hindcasts of where species might have been in the past in relation to historical environmental conditions, nowcasts in relation to current conditions, or forecasts of future species distributions.