18 Feb 2025 - 18 Feb 2025

Feb 18, 2025, 3:00 PM - Feb 20, 2025, 5:30 PM CET

Online Event: link

Speaker: Guillem Chust, Leader of MISSION ATLANTIC WP3: Pelagic Mapping: ecosystem, resources and pressures

Species distribution models (SDMs) are widely used as a tool for understanding species spatial ecology. They link species occurrence or abundance with environmental features of the location, via statistical modelling. According to ecological niche theory, species response curves are unimodal with respect to environmental gradients. A general problem with most habitat modelling approaches is that the estimated response curves can display biologically implausible shapes that do not respect ecological niche theory. At this workshop, participants will use shape-constrained generalized additive models (SC-GAMs) that allow imposing concavity constraints in the linear predictor scale and avoid overfitting. SC-GAMs are based on the same statistical framework as GLMs and GAMs regression methods, but they allow us to incorporate monotonicity and concavity shape-constraints in the component functions of the linear predictor of the GAMs. Imposing concavity constraints should be an effective alternative to fitting nonsymmetric parametric response curves, while retaining the unimodality constraint, required by ecological niche theory, for direct variables and limiting factors. We will guide participants building SC-GAMs from the beginning retrieving the occurrence data and environmental data from global public datasets. We will clean raw data removing outliers and select the environmental variables using a step-forward function. We will follow a R tutorial developed in AZTI (https://gam-niche.azti.es/). Depth of knowledge needed by attendees/target audience: This workshop is intended to anyone who wants to broaden their understanding of SDMs and improve their models’ realism. The workshop requires basic experience using and modifying R scripts; basic knowledge about species distribution (or ecological niche) models; a computer with recent versions of R and RStudio already installed; a good enough internet connection for live video sessions, and preferably a webcam for enhanced interactivity. Mode of organization: Online, including 3 sessions of 2.5 hours

Check out the GAM-NICHE Tool!