Representing and Querying 3D Cadastral Data Using GeoSPARQL: Compatibilities and Gaps in Australia

Abstract. Three-dimensional cadastral representation requires more than explicit 3D geometry, yet it remains unclear whether limitations arise from the published standard, evolving model, or source data. This study addresses that gap by assessing two real industry datasets, a GeoPackage exported to GeoJSON and an IFC2X3 building model, against the 3D Cadastral Survey Data Model (3D CSDM) reference target built on GeoSPARQL 1.1 as its semantic-spatial base. The specific contribution is the classification of standard-, model-, and source-data gaps so findings can be directed to the appropriate community. Across 70 assessed elements, one mapping was clean, 36 were partial, and 33 were absent, showing that neither dataset, in its delivered form, is directly representable in the target without significant transformation or augmentation. GeoSPARQL 1.1 geometric functions ignore Z values in spatial calculations, while the standard lacks solid-geometry classes and three-dimensional topology functions, limiting volumetric representation and 3D spatial querying. The 3D CSDM retains open or profile-dependent points, while both datasets lack survey observation evidence, cadastral survey provenance, and complete structured legal-interest records. Their strengths are complementary: GeoJSON provides clearer parcel-part identifiers and cadastral attributes but only implied 3D geometry, whereas IFC provides explicit 3D geometry but no formal georeferencing and limited cadastral parcel structure. The findings provide criteria for future GeoSPARQL 3D capability, priorities for 3D CSDM refinement, and a practical data-enrichment roadmap for industry.