Leveraging Ml Approaches for Scaling Climate Data in an Atmospheric Urban Digital Twin Framework
Abstract
This chapter provides foundational information that explores the intersection of Machine Learning (ML) and meteorological modeling with a particular focus on urban climate applications. the meteorological models, short-term numerical weather prediction and for large-scale climate, have been built on the foundation of classical mechanics, physics, thermodynamical and dynamical equations of motion, and energy and mass conservation. These models need to approximate, also termed parameterize, the multiscale processes that cannot be explicitly resolved. for such applications, ML approaches have been growing both to link physics-Based approaches and to develop "fast" surrogate models. This chapter introduces some of the terminologies and applications for large-scale (global) to city-scale high spatio-temporal meteorological modeling and data mapping. as urban environments become increasingly complex, the need for localized, high-resolution climate information becomes coveted for many applications. Yet neighborhood scale simulations within city blocks remain a challenge for the traditional dynamical models. ML offers a solution enhancing the granularity and usability of climate information in these intricate settings. Central to this discussion is the innovative concept of the atmospheric urban digital twins. These digital representations of urban environments link the multifaceted climatic interactions within cities, providing a comprehensive framework for understanding and predicting urban climate phenomena. Highlighting the rapid evolution of ML-Based climate analysis, this chapter delves into some practical applications, such as the use of advanced neural network architectures for downscaling. Such ML techniques are relatively novel in the realm of urban climate studies and help refine climate data to yield high-resolution insights tailored to place-Based analysis and decisions. the chapter concludes by highlighting real-world applications and the transformative potential of integrating ML into urban climate studies.
Recommended Citation
M. Singh and D. Niyogi, "Leveraging Ml Approaches for Scaling Climate Data in an Atmospheric Urban Digital Twin Framework," Advances in Machine Learning and Image Analysis for GeoAI, pp. 315 - 346, Elsevier, Jan 2024.
The definitive version is available at https://doi.org/10.1016/B978-0-44-319077-3.00019-5
Department(s)
Biological Sciences
Keywords and Phrases
Atmospheric urban digital twin; High-resolution climate data; Machine learning (ML); Urban climate; Urban downscaling
International Standard Book Number (ISBN)
978-044319077-3;978-044319078-0
Document Type
Article - Journal
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2025 Elsevier, All rights reserved.
Publication Date
01 Jan 2024