<![CDATA[Accurate prediction of species distribution is crucial for biodiversity conservation, particularly amid widespread habitat loss. As one of the largest islands and heavily developed landmasses in the Indonesian archipelago, Java is facing rapid habitat deterioration due to environmental changes. While most studies have historically focused on the protection species within protected areas, species distribution outside protected areas remain largely understudied. Consequently, metrics of citizen science are necessary to quantify the contemporary requirements of species, and to predict their potential distributions. We therefore developed habitat suitability models for nine species of the kingfishers (family Alcedinidae) distributed broadly across Java. Each model evaluates species occurrence data derived from citizen science platforms. Species distribution models of each species were constructed using bioclimatic, physiographic, and hydrologic variables through the random forest machine learning model. The models showed moderate to high predictive performance, with AUC-ROC values ranging from 0.75–0.93 and cross-validated AUC values ranging from 0.74 to 0.92, indicating reliable discrimination ability across species. Annual rainfall, elevation, vegetation, and proximity to water bodies are key factors influencing the distribution of kingfishers. The study highlights that potential habitats for all the studied species are located across Java, extending beyond protected forest areas to most other types of land uses. Conservation efforts should focus on preserving critical habitat features and implementing landscape-level management strategies, particularly for endemic and rare species.]]>
Copyrights © 2027
