A Global, Multi-Factor Geospatial Framework for Predicting Species Encounter Probability in Recreational Hunting and Foraging

Petr Salnikov

Traditional species distribution models (SDMs) are often computationally heavy and inaccessible to recreational users such as hunters and foragers. This paper presents a novel, precomputed geospatial framework (WildTracks) designed to calculate real-time species encounter probability. The system utilizes a 150 GB PostgreSQL/PostGIS database containing vectorized species ranges, ecoregions, Köppen climate zones, soil data, and elevation models. By precomputing spatial intersections using GiST indexes and centroid-based ST_Intersects queries, the framework reduces real-time query latency to milliseconds. The final probability is calculated using a multi-factor algorithm that weighs habitat suitability, seasonality, time of day, lunar phases, and terrain. Additionally, the system incorporates a safety module that identifies dangerous look-alike species (e.g., poisonous mushrooms) and hazardous taxa. This approach demonstrates how heavy geospatial data can be optimized for web-based, user-facing applications, providing a scalable model for citizen science and wildlife management.