Identification of urban thermal properties by combining urban microclimate modeling and thermal infrared satellite data

Abstract

Climate models are increasingly used to predict urban climate, but uncertainties in urban surface properties, such as thermal conductivity and capacity, lead to inaccuracies in comfort indices calculations. The forthcoming TRISHNA satellite mission will provide a unique dataset of thermal infrared imagery (TIR) to study urban area, with a planned spatial resolution of 57 meters and a revisit schedule of three times every eight days.

This study aims to develop a novel method to identify urban thermal properties based on prior knowledge by combining urban microclimate modeling and high spatial resolution satellite TIR, to provide input to microclimate urban models, for outdoor summer comfort assessment. Using a Bayesian approach, the method compares observed satellite-derived land surface temperatures (LST) with simulated LST from a combination of microclimate (Solene micro-climate) and radiative (DART) models. In this present work, as TRISHNA not yet operational, simulated data also serves as the observations.

An ideal street canyon model, representing typical urban structures, was used to validate the method. The approach improved the accuracy of thermal property estimates for horizontal surfaces by up to three times compared to random estimations within prior intervals. Best results were achieved with daytime TIR images acquired during the hottest days.