ICUC12-954, updated on 21 May 2025
https://doi.org/10.5194/icuc12-954
12th International Conference on Urban Climate
© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.
3D Local Climate Zone (LCZ) Modeling and Digital Twin Integration for Advanced Climate Resilience using AI
Dyutisree Halder1, Harshit Harshit2, and Rahul Dev Garg1
Dyutisree Halder et al.
  • 1Indian Institute of Technology Roorkee, Civil engineering (Geospatial Engg. Group), Roorkee, India (dhalder@ce.iitr.ac.in)
  • 2IT University of Copenhagen, Data Science Section, Copenhagen, Denmark (hano@itu.dk)

Local Climate Zones (LCZs) provide a standardized classification of urban morphology and climate interactions, yet traditional 2D representations lack spatial depth for comprehensive urban climate analysis. This research proposes an innovative approach that integrates urban remote sensing , AI-driven 3D modeling, and state-of-the-art Digital Twin workflows to enhance LCZ visualization and climate analysis. The methodology involves leveraging multi-spectral datasets (MODIS, Landsat, Sentinel) to derive urban morphological and climatic parameters such as building height, vegetation cover, land surface temperature (LST), and albedo. Next a Text-to-Mesh AI model is used to convert LCZ descriptions into spatially detailed 3D urban morphologies which means textual descriptions of LCZ characteristics are used to procedurally generate 3D urban morphologies, replacing traditional 2D classified patches with immersive, spatially accurate representations. At Last, these AI-generated models are embedded into a georeferenced digital twin framework to evaluate and analyze for its usability and consistency, based on scale generalization, and real-world validation of these 3D models for climate impact analysis. This open-data-driven approach enables improved climate visualization, supporting urban planners and policymakers in climate adaptation strategies. Our attempt is to showcase the enhanced interpretability and usability of 3D LCZ models in urban climate research. The integration of AI, digital twins, and open climate datasets provides a scalable, innovative tool for understanding and mitigating urban climate challenges. By linking remote sensing-derived land cover, thermal, and vegetation indices with AI-generated urban forms, this study enables an interactive data-driven approach to enhance urban climate zone representation. Potential applications include urban heat island mitigation, microclimate simulations, and climate-resilient urban design strategies.

How to cite: Halder, D., Harshit, H., and Garg, R. D.: 3D Local Climate Zone (LCZ) Modeling and Digital Twin Integration for Advanced Climate Resilience using AI, 12th International Conference on Urban Climate, Rotterdam, The Netherlands, 7–11 Jul 2025, ICUC12-954, https://doi.org/10.5194/icuc12-954, 2025.

Supporters & sponsors