BEGIN:VCALENDAR VERSION:2.0 PRODID:-//pretalx//talks.osgeo.org//foss4g-2026//talk//CVUMUK BEGIN:VTIMEZONE TZID:JST BEGIN:STANDARD DTSTART:20000101T000000 RRULE:FREQ=YEARLY;BYMONTH=1 TZNAME:JST TZOFFSETFROM:+0900 TZOFFSETTO:+0900 END:STANDARD END:VTIMEZONE BEGIN:VEVENT UID:pretalx-foss4g-2026-CVUMUK@talks.osgeo.org DTSTART;TZID=JST:20260901T160000 DTEND;TZID=JST:20260901T163000 DESCRIPTION:Motivation and Problem Context\n\nAdvances in generative artifi cial intelligence are enabling new applications across many industry secto rs. Large language models (LLMs) have the potential to bring geospatial de cision support to non-technical professionals by allowing users to explore spatial questions through natural language. Compared with traditional geo spatial workflows—characterized by manual layer selection and preprocess ing\, static maps\, delayed updates\, limited model transparency\, fragmen ted analytical processes\, and results that are difficult to reproduce—G eoAI-enabled approaches offer the promise of on-demand\, ad hoc spatial an alysis. By reducing the burden of integrating knowledge across multiple da ta silos\, such approaches can broaden access to geospatial insight while maintaining essential principles of integrity\, provenance\, and trust (IP T).\n\nHowever\, LLMs lack intrinsic geospatial awareness and do not maint ain structured knowledge of geographic features or their domain-specific r elationships. As a result\, they cannot natively reason about how disrupti ons propagate through interconnected spatial systems—such as transportat ion networks\, health service accessibility\, or school catchments—durin g events like floods\, wildfires\, or other natural hazards. While LLMs ex cel at synthesizing textual information\, they lack an internal representa tion of the spatial networks and semantic relationships required for geosp atial reasoning.\n\nOvercoming this limitation requires a sustainable and open approach for making the wealth of geoinformatics data interoperable a nd accessible to retrieval-augmented generation (RAG) workflows that can p rovide geospatial grounding for LLMs. Rather than relying on custom point- to-point integrations between individual systems\, interoperability should occur at the point where data are published. Such an approach would act a s a force multiplier for data reuse by eliminating bespoke integrations an d enabling on-demand data fusion across domains.\n\nDiscrete Global Grid S ystems (DGGS)\, recently standardized by the Open Geospatial Consortium (O GC)\, provide an important step in this direction by enabling aggregate an d statistical datasets to be aligned to a common geographic reference fram ework through standardized zone identifiers. When datasets across differen t domains are published using the same DGGS reference system\, they can be integrated directly and consistently across organizational and disciplina ry boundaries.\n\nSpatial Knowledge Graphs (SKGs) provide a complementary mechanism for representing geographic entities and their semantic relation ships\, enabling machines to discover\, retrieve\, and reason over geospat ial context across networks of features. However\, while DGGS enables inte roperability for statistical and aggregated data\, no widely adopted stand ard currently exists for publishing SKGs in a way that supports interopera bility by common geography across independent organizations. Without such a mechanism\, spatial knowledge graphs remain difficult to integrate dynam ically across domains and data providers.\n\nProposed Approach\n\nThis pap er outlines a metamodel that enables disparate organizations to independen tly publish interoperable spatial knowledge graphs (iSKGs) anchored to com mon geographies. By adopting a federated architecture inspired by the Data Mesh paradigm\, independently managed iSKGs can be combined into a broade r Spatial Knowledge Mesh\, allowing networks of geographically referenced knowledge to be discovered and integrated on demand. In this architecture\ , organizations publish domain-specific knowledge graphs that reference sh ared geographic abstractions while maintaining their own governance and pr ovenance.\n\nThe proposed metamodel further supports the propagation of ch anges across dependent knowledge graphs\, enabling updates to flow through the mesh and ensuring that downstream users—particularly those operatin g in lower-resource environments—can maintain up-to-date geospatial awar eness. When combined with DGGS-based data integration and retrieval-augmen ted generation workflows\, the resulting Spatial Knowledge Mesh provides t he foundation for Geo-GraphRAG pipelines that allow LLM-based systems to r eason over interconnected spatial systems with explicit semantic context.\ n\nLLMs lack geospatial awareness because they do not maintain a live mode l of the Earth or structured graphs of geographic features and their topol ogical relationships. Instead\, they learn statistical associations betwee n words rather than networks linking roads\, facilities\, and populations\ , and therefore cannot determine which specific roads connect communities to hospitals without structured geospatial data at runtime. They also lack domain-specific geospatial semantics\, spatial reasoning capabilities\, a nd up-to-date local knowledge as infrastructure and hazards evolve.\n\nSpa tial Knowledge Graphs address these limitations by representing geographic features as interconnected networks of geo-objects and their semantic rel ationships. In an SKG\, each node represents a geographic feature\, while edges capture topological relationships or analytical findings derived fro m geoinformatics data and expressed through semantic labels. Together with a geo-ontology defining the graph schema\, these relationships create a s tructured representation of spatial systems that can be traversed and quer ied. This structure enables Geo-GraphRAG pipelines in which LLMs translate natural-language questions into graph queries\, such as GeoSPARQL or Cyph er\, enabling geospatial insights to be retrieved with full transparency a nd traceability.\n\nTo support interoperability across organizations\, the metamodel enables geo-ontologies—comprising definitions of geo-object c lasses and semantic relationships—to be published as first-order graphs that provide shared schemas for interoperable graph creation. Integrations between iSKGs can also be published as reusable first-order entities. Pro venance metadata records both the original source of attribute values and geometries and the publishing organization\, supporting traceability and i ntegrity using existing standards such as GeoDCAT. iSKGs can be published with explicit periods of validity and version identifiers to capture tempo ral evolution. Open-source software implementing this metamodel has been d eveloped to support the creation of geo-ontologies and iSKGs\, as well as graph operations such as pull\, merge\, and change detection across federa ted knowledge graphs.\n\nThis work has been supported by the U.S. Army Cor ps of Engineers (USACE) Civil Works Division and\, through the OGC\, by Na tural Resources Canada and the United States Geological Survey as part of a collaborative research effort to integrate geospatial awareness into LLM s for disaster response and resilience. The proposed metamodel extends OGC Building Blocks—which model dependencies between specifications and pro mote reuse across geospatial standards—by adding support for change prop agation\, temporal alignment\, and spatial knowledge graph interoperabilit y.\n\nExpected Impact\n\nBy providing a standard approach and open-source software for publishing geoinformatics data as interoperable geospatial kn owledge networks accessible to large language models\, this work aims to a ct as a force multiplier for positive impact. In particular\, it seeks to enable lower-resourced settings to develop bespoke GeoAI solutions by prov iding access to integrated geospatial knowledge infrastructures optimized for use with LLM-based systems. DTSTAMP:20260718T071930Z LOCATION:Cosmos1 SUMMARY:Towards a Spatial Knowledge Mesh: A Metamodel for Federated and Int eroperable Spatial Knowledge Graphs to Enable Geospatial Awareness\, Integ rity\, Provenance\, and Trust in Large Language Models - Nathan McEachen URL:https://talks.osgeo.org/foss4g-2026/talk/CVUMUK/ END:VEVENT END:VCALENDAR