Geo-AI for Biodiversity Conservation: The Northern Bald Ibis Case Study

The Northern Bald Ibis (Geronticus eremita), Critically Endangered, survives only in a narrow coastal zone of southwestern Morocco, making it highly vulnerable to habitat loss, degradation, and climate change. Identifying key environmental drivers and suitable habitats is therefore urgent. This study applied an integrated Geo-AI framework combining geomatics, multi-source remote sensing via Google Earth Engine, the Maximum Entropy algorithm, and 1,085 field-based occurrence records. From 35 initial environmental variables, six were retained after multicollinearity analysis and ecological validation: summer land surface temperature, maximum and minimum air temperatures, precipitation of the wettest month, altitude, and winter vegetation cover. Model performance was excellent (AUC = 0.973 ± 0.029; TSS > 0.85). Summer land surface temperature was the most influential variable, with a sum of permutation contribution of predictors index (SPCPI = 74.8), followed by altitude (38.4) and the minimum temperature of the coldest month (35.0), highlighting thermal constraints. Response curves indicated a preference for warm, dry conditions, moderate vegetation, mild winters, and elevations below 1,200 m. High-suitability areas (probability > 0.6) are concentrated in Souss-Massa and Marrakech-Safi, overlapping current and historical colonies, including those in Fes-Meknes, while moderate-suitability corridors suggest dispersal pathways. Future high-emission scenarios indicate increasing habitat fragmentation, underscoring species vulnerability. These results provide guidance for conservation planning, emphasizing coastal habitat protection, ecological corridor maintenance, and sustainable management of semi-arid and mid-altitude areas. This study highlights the value of integrating remote sensing, machine learning, and ecological knowledge to support evidence-based conservation of this emblematic species.

Pagina's: 70, Paperback, Eliva Press