Assessment of Fuzzy Logic to Enhance Species Distribution Modelling

Abstract

Abstract Fuzzy logic presents a promising approach for Species Distribution Modelling by generating a commensurable value termed ‘favourability’. Departing from conventional value ‘probability’, ‘favourability’ remains robust regardless of species prevalence, enabling across species comparisons despite varying prevalence. Such comparisons facilitate the interpretation of cryptic species, which have intricate distribution data to assign. This study generated environmental favourability values for two borers within a cryptic beetle species complex: Euwallacea fornicatus and Euwallacea perbrevis in Australia. This research delved into biogeographic relationship analyses fuzzy intersection and potential biotic interaction of these closely related borers, highlighting a notably favourable distribution pattern for Euwallacea fornicatus in Queensland. To evaluate the model’s performance, this paper utilized commonly employed evaluation metrics (Area under the receiver operating characteristic curve, True statistical skill, Correct classification rate), alongside fuzzy entropy value and the Hosmer-Lemeshow test to assess the model reliability. This study validates the efficacy of fuzzy logic in species distribution modelling and showcases its utility in assessing habitat suitability for closely related species through the utilization of a more informative value – favourability. This value emerges as a valuable refinement to Species distribution models, enabling the assessment of differences and similarities among species’ distribution areas alongside the species’ environmental correlates.