Speaker: Jia Hu, Postdoctoral Researcher, Department of Statistics, Iowa State University

Title: A Transferable Local Climate Zone Mapping Framework for Urban Heat Island Studies

Abstract: Local climate zone (LCZ) provides a detailed classification scheme for building types and land cover types for urban areas globally, and offers a unified standard for block-scale urban heat island (UHI) studies. However, the existing LCZ classification methods are limited in their applicability to global cities due to the need for a large number of training samples or data with high spatial resolution. In this study, we developed a more universally applicable LCZ classification framework by coupling multi-source remote sensing data and urban GIS data with high spatial resolution, Google Earth Engine platform and random forest model. We generated the 120-meter spatial resolution LCZ maps for 32 major cities in China using the proposed method. Combining Landsat 8 data and land surface temperature (LST) retrieval algorithm, the LST of these cities was obtained. Using dense trees (LCZ A) as the background temperature reference, UHI intensity of different LCZ types in diverse cities were calculated and compared. Our proposed LCZ classification framework can be applied to urban surface classification and UHI studies in any global cities, providing a reference for urban planning and sustainable urban environmental development.