Modeling marine pelagic fish species spatiotemporal distributions utilizing a maximum entropy approach. (19th June 2018)

• Record Type: Journal Article
• Authors: Wang, Lifei; Kerr, Lisa A.; Record, Nicholas R.; Bridger, Eric; Tupper, Benjamin; Mills, Katherine E.; Armstrong, Edward M.; Pershing, Andrew J.
• Is Part Of: Fisheries oceanography. Volume 27:Number 6(2018)
• Page Start: 571
• Page End: 586
• Publication Date: 2018-06-19
• Subjects: Atlantic herring (Clupea harengus) -- Atlantic mackerel (Scomber scombrus) -- butterfish (Peprilus triacanthus) -- fishery‐dependent data -- marine pelagic fish -- maximum entropy (MaxEnt) -- satellite remote sensing data; Fishery oceanography -- Periodicals; 639.2
• DOI: 10.1111/fog.12279
• Languages: English
• ISSNs: 1054-6006

Abstract

Abstract: Understanding how ocean conditions influence fish distributions is critical for elucidating the role of climate in ecosystem change and forecasting how fish may be distributed in the future. Traditional species distribution models are often applied to scientific‐survey data, which include species presence‐absence information, to predict distributions. Maximum entropy (MaxEnt) models are promising tools as they can be applied to presence‐only data (e.g., data collected from fishermen targeting a specific species or from observers in citizen‐science programs). We used MaxEnt models to relate occurrence records of three marine pelagic fish (Atlantic herring, Atlantic mackerel, and butterfish) in fishery‐dependent data to environmental conditions (sea surface temperature (SST) and chlorophyll‐ a concentration from satellite remote sensing, bathymetry, and climate indices), and evaluated model performance by both cross‐validation and validation using fishery‐independent data. We developed monthly habitat suitability maps for these fish in the Northwest Atlantic Shelf area, and assessed the relative influence of environmental factors on their distributions. Across months, their suitable habitat areas varied with each species exhibiting inshore‐offshore and north‐south movements in response to changing environmental conditions. Overall, SST and chlorophyll‐ a concentration had the greatest influence on the distributions of these fish, with bathymetry having moderate influence and climate indices having little influence. Our application of MaxEnt models enabled us to integrate presence‐only data and high resolution environmental data from satellite remote sensing to describe spatiotemporal distributions of marine pelagic fish. These models were used to hindcast species occurrence in relation to historical environmental conditions to evaluate their predictive performance, and have the potential to provide nowcasts in relation to current conditions or forecasts of species future distributions.