Through times, regions all over the world are very often subject to change (their names, their belonging, their composition, and their geometries). In this paper, we present a Semantic Matching Algorithm for automatically detecting, describing and publishing in the Linked Open Data Web, rich descriptions of changes occurring in multi-level territorial partitions (e.g., partitions made of major regions, regions and districts levels). We adopt a Linked Data (LD) approach for the semantic descriptions of the changes they undergo, relying on two existing generic ontologies, TSN-Ontology and TSNChange Ontology. The created RDF graphs draw the lineage of each region over time (horizontal reading of the graphs), as well as the propagation of a change event through the partition levels (vertical reading).