Discovering morphological impact discrepancies on thermal environment among urban functional zones using essential urban land use categories and machine learning

Abstract

The urban heat island effect has undermined the livability of cities. Previous studies have shown that Urban Functional Zones (UFZs) exert impacts on the thermal environment. However, the classification methods of UFZs haven't been standardized, which limits the generation of universal conclusions. This study employs the essential urban land use categories (EULUC) mapping method to delineate UFZs in the central area of Tianjin, China, and utilized multiple regression methods to analyze the correlation between urban morphology indicators and land surface temperature (LST) in summer and winter, and explored the differences in LST and its influencing factors across UFZs. The principal findings are as follows: (1) The EULUC classification system represents an effective method for delineating UFZs, with a classification accuracy of up to 92.8 %. (2) Both in summer and winter, commercial zones demonstrate the highest levels of LST, while open space zones exhibit the lowest levels. (3) The impact of urban morphology indicators on LST varies across UFZs, and these indicators demonstrate threshold effects. LST in residential zones and commercial zones was found to be most strongly influenced by urban morphology indicators, among which building footprint density was the most significant indicator in both summer and winter. This study presents a methodological framework for associating UFZs with thermal environments based on the EULUC, aiming at providing a practical guide for the spatial configuration of UFZs and the urban morphology indicators setting of different UFZs in planning practice.