GeoSPARQL query support for scientific raster array data

This paper presents the design of a GeoSPARQL query processing solution for scientific raster array data, called GeoLD. The solution enables the implementation of SPARQL endpoints on top of OGC standard Web Coverage Processing Services (WCPS). Thus, the semantic querying of scientific raster data is supported without the need of specific raster array functions in the language. To achieve this, first Coverage to RDF mapping solutions were defined, based on the well-known W3C standard mappings for relational data. Next, the SPARQL algebra is extended with a new operator that delegates part of the GeoSPARQL query in WCPS services. Query optimization replaces those parts of the SPARQL query plan that may be delegated to a WCPS service by instances of such new WCPS operator. A first prototype has been implemented by extending the ARQ SPARQL query engine of Apache Jena. Petascope was used as the WCPS implementation on top of the Rasdaman raster array database. An initial evaluation with real meteorological data shows, as it was initially expected, that the approach outperforms an existing reference relational database based GeoSPARQL implementation The work of Shahed Bassam Almobydeen was partially funded by European Union under the Erasmus Mundus Peace II mobility program. The work of José R.R. Viqueira was partially funded by Xunta de Galicia, Spain under the Project ED431B 2021/16, by the TRAFAIR EU project 2017-EU-IA-0167, co-financed by the Connecting Europe Facility, by the EU RADAR-ON-RAIA project (0461_RADAR_ON_RAIA_1_E), co-financed by the European Regional Development Fund (ERDF) through the Iterreg V-A Spain-Portugal program (POCTEP) 2014–2020 and by project MAGIST-ELA PID2019-105221RB-C42, funded by Spanish Ministry of Economy and Competitiveness, Spain. The work of Manuel Lama was partially funded by the Spanish Ministry for Science, Innovation and Universities under the project TIN2017-84796-C2-1-R SI