Designing of high-performance species habitat suitability index model

Habitat Suitability Index (HSI) models are widely used in wildlife management to assess species-environment relationships and inform conservation strategies. However, traditional HSI models often rely on simplistic weighting schemes that may inadequately capture the complexities of species-habitat interactions, particularly under climate change. This study presents an enhanced HSI model that addresses these limitations by integrating multicollinearity analysis to exclude highly correlated variables and applying a Random Forest (RF) for variable selection and weighting. The model was validated using datasets from the Northwest Pacific Ommastrephes bartramii and Southwest Atlantic Illex argentinus fisheries. Results show the proposed model significantly outperforms conventional approaches in predicting species distribution, with improved precision and the identification of key environmental drivers. This refined HSI model would offer greater interpretability, supporting more informed decision-making in marine spatial planning and fisheries management.