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

Area of Science:

• Marine ecology
• Conservation biology
• Quantitative ecology

Background:

• Species distribution models (SDMs) are crucial for understanding highly migratory species' habitats and informing conservation strategies.
• A debate exists regarding whether to model habitat suitability (presence/absence) or density (# individuals), especially for marine mammals.
• Zero-inflated data present a challenge in ecological modeling, necessitating methods like Hurdle Models.

Purpose of the Study:

• To assess and compare the performance of Generalized Additive Models (GAMs) and Boosted Regression Trees (BRTs) for modeling fin whale (Balaenoptera physalus) density and habitat suitability.
• To evaluate Hurdle Models in addressing zero-inflated data for fin whale distribution analysis.
• To identify key environmental covariates influencing fin whale distribution in the NW Mediterranean Sea.

Main Methods:

• Collected fin whale data from 2008-2022 along fixed transects in the NW Mediterranean Sea during summer.
• Employed traditional line transect methodology to calculate the Effective Area Monitored.
• Utilized GAMs, BRTs, and Hurdle Models, incorporating static (bathymetry, slope) and dynamic (SST, MLD, Chl, EKE, FSLE) covariates.

Main Results:

• All models demonstrated good performance in distinguishing presence and absence of fin whales.
• While density and presence patterns were similar, their environmental dependencies varied by model type.
• Bathymetry emerged as the most significant covariate across all models, followed by sea surface temperature (SST) and pre-sighting chlorophyll concentration.

Conclusions:

• The choice of modeling technique can influence the understanding of environmental drivers for fin whale distribution.
• SDMs are valuable tools for marine mammal conservation, providing essential data for spatial management.
• This research quantifies fin whale-environment relationships in a data-limited region, supporting evidence-based conservation efforts.