Enugu State CityGML LOD1 3D Building Model: The First Open-Access Semantic 3D City Model for Nigeria and Among the First in Sub-Saharan Africa

Overview

This dataset presents the first city-scale CityGML Level of Detail 1 (LOD1) 3D building model for Enugu State, Nigeria, and to the authors' knowledge, the first openly published semantic 3D city model for any city in Sub-Saharan Africa. The dataset covers the urban extent of Enugu, comprising thousands of building footprints extruded to terrain-referenced heights using a DEM-based methodology. It was produced entirely with open-source tools and is published under an open licence to support urban planning, disaster risk reduction, and geospatial research in African cities.

Methodology

Building footprints were sourced from OpenStreetMap and validated against satellite imagery. Heights were derived from a Digital Elevation Model (DEM) using terrain-referenced extrusion, with relative building heights estimated from contextual analysis. The CityGML conversion and attribute-enrichment pipeline was built in FME Workbench, producing a valid CityGML 2.0 dataset that conforms to OGC standards. The model was validated in 3DCityDB and visualised in CesiumJS as part of a broader smart city platform developed for Enugu.

Significance

The global 'awesome-citygml' registry, the most comprehensive list of open CityGML datasets, covering 21 countries and 66+ cities with over 215 million buildings, yet contains no African city. This dataset begins to address that gap. Enugu is a rapidly urbanising city of approximately 1 million people that experiences recurring flood events. A city-scale 3D model provides a foundational spatial layer for flood modelling, urban heat analysis, solar potential estimation, network analysis, and digital twin development.

Applications

This dataset supports:

• Flood risk modelling and disaster risk reduction

• Urban climate and heat island analysis

• Solar energy potential estimation

• Urban digital twin development

• 3D GIS research and education in African urban contexts

• Benchmarking LOD1 generation methods in data-scarce environments