GLOSSA: a user-friendly R Shiny application for Bayesian machine learning analysis of marine species distribution

Abstract: Species distribution models (SDMs) are one of the most common statistical methods to assess species occupancy and geographic distribution patterns. With the increasing complexity of ecological data, many methodological approaches have been developed, often accessible through command-line interfaces or graphical user interfaces (GUIs). However, few species distribution modeling tools are designed to be well-documented, user-friendly, flexible, and reproducible. Here we introduce GLOSSA, an open-source R package and Shiny app designed for species distribution modeling using species occurrence and environmental data. GLOSSA's user-friendly interface guides users through steps including data uploading, processing, model fitting, spatial and temporal projections, and interactive visualization of results. The app also calculates variable importance, generates response curves with environmental variables, and performs cross-validation. At its core, GLOSSA modeling approach is based on Bayesian Additive Regression Trees (BART), an innovative machine learning method. We present the functionality and versatility of GLOSSA through three case studies, addressing a range of ecological scenarios at regional and global scales. Along with comprehensive documentation, examples, and tutorials, these case studies illustrate how an intuitive graphical interface can make species distribution modeling accessible to a broad audience. GLOSSA stands out as an easy-to-use tool for species distribution modeling, providing an intuitive interface, detailed documentation, flexible modeling, and interactive result exploration and export options. Additionally, its outputs can be used directly to inform marine ecosystem models (MEMs), enhancing its utility in ecological research and applications.