Multivariate approaches in species distribution modelling: Application to native fish species in Mediterranean Rivers

Abstract

Worldwide, human activities such as water withdrawals, storing, hydropeaking and climate change are negatively impacting freshwater biota. To evaluate the threats posed by these phenomena the development of scientifically sophisticated tools has now become a fundamental area of research within the scientific community. The physical habitat simulation approach has proved to be adequate to evaluate the effect of management alternatives, restoration actions and climate change. The physical habitat methods, which can be implemented at different spatial scales, assess the quantity and quality (i.e. suitability) of the physical habitat for different flows integrating hydrological, hydraulic and biological data (i.e. habitat suitability models). The development of the habitat suitability models can manifold be approached, although machine learning techniques steadily gained adepts over the classical methods and thus they became common practice.