Evaluating Three Modelling Frameworks for Assessing Changes in Fin Whale Distribution in the Mediterranean Sea

• DOI: 10.1002/ece3.71007
• Publication Date: 2025-03-07T14:02:18Z
• Authors: Francesca Grossi; Elliott L. Hazen; Giulio De Leo; Léa David; Nathalie Di‐Méglio; Antonella Arcangeli; Eugenia Pasanisi; Ilaria Campana; Miriam Paraboschi; Alberto Castelli; Massimiliano Rosso; Aurelie Moulins; Paola Tepsich
• Keywords: Generalized additive model; Species distribution; Marine mammal; Predictive power

ABSTRACT Understanding the habitat of highly migratory species is aided by using distribution models to identify species‐habitat relationships and inform conservation management plans. While Generalized Additive Models (GAMs) are commonly used in ecology, particularly modeling marine mammals, there remains a debate between (presence/absence) versus density (# individuals). Our study assesses performance predictive capabilities GAMs compared boosted regression trees (BRTs) for both fin whale suitability alongside Hurdle treating presence/absence as two‐stage process address challenge zero‐inflated data. Fin data were collected from 2008 2022 along fixed transects crossing NW Mediterranean Sea during summer period. Data analyzed traditional line transect methodology, obtaining Effective Area monitored. Based on existing literature, we select various covariates, either static nature, such bathymetry slope, or variable time, example, SST, MLD, Chl concentration, EKE, FSLE. We explanatory power skill different techniques. results show that all performed well distinguishing presences absences but, while presence patterns similar, their dependencies environmental factors can vary depending chosen model. Bathymetry was most important models, followed SST chlorophyll recorded 2 months before sighting. This underscores role SDMs play mammal efforts emphasizes importance selecting appropriate It also quantifies relationship variables an understudied area, providing solid foundation informed decision making spatial management.