GRaaS: GeoReasoning as a Service — A Reference Architecture for Ontologically Grounded GeoAI Integrating BFO, GeoRegion, TST, and State-Time Versioning

This concept paper proposes GRaaS (GeoReasoning as a Service), a modular reference architecture for ontologically grounded GeoAI. GRaaS operationalizes a broader “Space as Semantic Services” paradigm in which spatial reference, semantic registration, predicate reasoning, continuity, interaction, and presentation are exposed as interoperable service layers over existing GIS infrastructure. The architecture integrates BFO, GeoRegion, GeoRegistration, TST, State-Time Versioning, GeoLLM, and GeoUI to support inspectable and traceable geospatial reasoning over declared spatial commitments rather than implicit GIS assumptions. The paper positions GRaaS as a transition from domain-specific GIS applications toward reusable geospatial reasoning functions.