Resumen: Environmental data generated by observation infrastructures and models is widely heterogeneous in both structure and semantics. The design and implementation of an ad hoc data model for each new dataset is costly and creates barriers for data integration. On the other hand, designing a single data model that supports any kind of environmental data has shown to be a complex task, and the resulting tools do not provide the required efficiency. In this paper, a new data modeling framework is proposed that enables the reuse of generic structures among different application domains and specific applications. The framework considers four levels of abstraction for the data models. Levels 1 and 2 provide general data model structures for environmental data, based on those defined by the Observations and Measurements (O&M;) standard of the Open Geospatial Consortium (OGC). Level 3 incorporates generic data models for different application areas, whereas specific application models are designed at Level 4, reusing structures of the previous levels. Various use cases were implemented to illustrate the capabilities of the framework. A performance evaluation using six datasets of three different use cases has shown that the query response times achieved over the structures of Level 4 are very good compared to both ad hoc models and to a direct implementation of O&M; in a Sensor Observation Service (SOS) tool. A qualitative evaluation shows that the framework fulfills a collection of general requirements not supported by any other existing solution. Idioma: Inglés DOI: 10.1016/j.envsoft.2024.106248 Año: 2025 Publicado en: ENVIRONMENTAL MODELLING & SOFTWARE 183 (2025), 106248 [23 pp.] ISSN: 1364-8152 Tipo y forma: Artículo (Versión definitiva) Área (Departamento): Área Lenguajes y Sistemas Inf. (Dpto. Informát.Ingenie.Sistms.)