National geospatial infrastructures are evolving from classic spatial data infrastructures into connected geospatial ecosystems in which authoritative reference data, Earth observation (EO), sensor observations and analytical services are combined in operational workflows. This paper proposes a framework for a Slovenian national geospatial digital twin grounded in the emerging Geo Slovenija ecosystem and demonstrates one validated flood-related component through a pilot application in the Lake Cerknica karst polje. The framework links authoritative cadastral and topographic data, terrain and 3D data, EO products and hydrometeorological observations in a web-based 2D/3D environment. The flood pilot reconstructs daily inundation states from a digital terrain model and a long-term gauge record (1954–2025), while independent EO-based benchmarking is performed against 26 PlanetScope scenes acquired between 2023 and 2025. The results show a distinct non-linear water-level-area relation and moderate to good agreement with EO-derived water masks (mean IoU = 0.62 +/- 0.26; mean F1 = 0.75 +/- 0.22), with clearly stronger and more stable performance during stable lake phases. The paper contributes a national-scale architectural perspective and a computationally lightweight, validated flood-state reconstruction workflow that can serve as a transferable building block for hazard monitoring and decision support in Slovenia.