RestoreAI: A Decision Support Framework for Optimizing Wildlife Habitat Restoration Strategies Using Ecological Data

Abstract

The degradation of habitats continues to be among the major causes of decline in wildlife populations but the restoration process often suffers from incomplete ecological information and inconsistent expert judgement. In this paper, RestoreAI is introduced as an ecological data-driven decision support framework that combines multi-source ecological data and suitability ranking model to inform wildlife habitat restoration. RestoreAI includes three layers: ecological data collection, suitability ranking through a modified weighted Analytic Hierarchy Process (AHP) and species distribution results, and adaptive decision support where a resource-constrained restoration plan with a monitoring component is produced. RestoreAI is illustrated conceptually by applying it to a degraded tropical dry              forest-grassland mosaic system with a focus on vegetation, soil, species occurrence, and connectivity data at 40 candidate locations. It will be evaluated using a retrospective assessment of the rankings generated through RestoreAI compared to the rankings given by experts in terms of rank correlation and tracking simulated restoration outcomes. Based on similar studies in AI-assisted conservation and restoration scoring, it can be expected that RestoreAI will produce prioritization performance at least comparable to that of the expert-only prioritization with significantly reduced planning time.