The geography of evolving concepts: quantifying spatial, temporal, and semantic consequences of ontology changes

Concepts, shaped by human perception and cognition of the world, are constantly evolving. Among these, geographic concepts (e.g. forest, mountain, island) are particularly dynamic and context-dependent. These concepts have a key role in interlinked knowledge systems, providing rich spatio-temporal reasoning in applications such as disaster responseand land use classification. However, their evolving nature makes them susceptible to reinterpretation. Even minor changes in their definition, spatial extent, or temporal scope can cascade through ontology-driven reasoning systems, disrupting data reuse, integration, inference, and decision-making processes. In this work, we aim to quantify how the impact of ontology changes propagates through space, time, and semantics using a multi-dimensional framework. We focus on conceptual changes,rather than changes to individual instances,and assess their downstream impacts within a geospatial ontology. To showcase the impacts of such concept drifts, we construct a maritime legal ontology as a real-world application of our framework. We demonstrate that by adding just one additional criterion to the definition of rocks, 6.26% of statements in our exemplary knowledge graph would require revalidation. By quantifying the impact, we seek to inform more robust research data management practices and ensure the long-term usability of spatio-temporal knowledge systems in evolving domain semantics.