HabitatNet: A Spatial Ecological Modeling Framework for Predicting Wildlife Habitat Suitability Under Landscape Transformation

Abstract

Landscapes that experience dramatic transformations due to the urbanization process, agricultural intensification, and deforestation have led to habitat losses among other causes of biodiversity loss around the world. However, the traditional habitat suitability models employ the use of static and low-resolution predictor datasets that fail to adequately reflect the small-scale dynamics inherent in rapidly changing landscapes. This paper introduces HabitatNet, an approach to habitat suitability prediction in changing landscapes through the development of an integrated spatial ecological modeling framework. This framework uses multiple sources of spatial information such as land cover, topography and climatic datasets in the modeling process. HabitatNet can generate continuous suitability predictions for wildlife habitats while being able to account for any changes resulting from the transformation of these landscapes. The proposed framework has been evaluated and applied in the case of several hypothetical land-use change scenarios and the results show that the model produces spatially coherent and interpretable habitat suitability predictions which can reveal habitat vulnerable areas.