A Graphical Species Distribution Model of life history connectivity and multi-scale co-existence of marine species

University Of Maryland Center For Environmental Sciences, Cambridge MD

Investigators

Abstract

An understanding of the spatial distribution of marine species within ecologically and economically important regions is essential for testing fundamental ecological theory. It is also crucial for conservation efforts and to forecast potential impacts of climate change. Technologies are increasingly available and affordable for ecologists to sample the fine-scale distribution and habitat condition of ecological communities. Specifically, imaging systems can generate rich and integrated biological and environmental data at multiple spatial scales. However, there is a lack of statistical tools to analyze these multi-scale data, which limits our ability to extract critical information. This project will develop a statistical framework that encompasses the complex dynamics between marine ecological communities and spatial scales. The approach will improve our understanding of how biotic interactions and habitat factors shape the spatial patterns of marine communities, yet will be simple enough to contribute to broad ecological research and guide various actions of conservation and management. The project includes the following activities: develop, evaluate and quantify the performance of a graphical species distribution modeling (GSDM) within an R environment; employ the GSDM framework to leverage the rich multispecies, multiscale data inherent in imaging systems, and available to us from ongoing NSF-funded programs; and facilitate the co-development of an open-source R package by the marine ecological community to inform conservation and management decision making. This project will build capacity in marine biology through developing new species distribution modeling frameworks that overcome the computational challenge of incorporating data from sampling platforms with high spatial resolution. The frameworks will be coded in the R computing language to access the increasingly available, high resolution, multispecies data collected by image-based sampling systems. This approach, leveraging the most recent advances in Graphical and Bayesian network models, will improve the understanding of trophic interactions and life history connectivity within productive marine ecosystems. It will also further the development of generic ecological tools for the research, conservation, and management communities. Target audiences include graduates and the ecological research community of the University System of Maryland, and fisheries managers and agencies through the Goal Implementation Teams of the Chesapeake Bay Program. Feedback from community members will be engaged in a co-development process through online workshops, conference sessions, webinars, and dedicated websites (https://esc.cbl.umces.edu/gsdm). This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.