Metodi e analisi statistiche 2023

This paper aims to investigate the small-scale distribution of benthopelagic fauna within the two cold-water coral provinces of Santa Maria di Leuca and Bari Canyon in the Central Mediterranean. We analyze data from video systems to investigate the coral habitat's role and explore potential environmental factors influencing biodiversity and community dynamics. We propose a modeling approach based on distributional regression models to estimate species abundance while considering different response variables. In the context of intermittently observed data, such as in-situ observations, we use the GLM and GAMLSS frameworks to model count and semi-continuous abundance data. The latter approach provides flexibility in efficiently addressing issues such as overdispersion and zero-inflation, allowing us to simultaneously evaluate the effects of drivers on both location and shape characterizing the considered distributions. Results highlight differences in estimated abundances explained by the effects of different enviromental factors ascribed to data characteristics and model assumptions and can support ecologists in the evaluation of species abundance in the ecosystems under study.