Enhanced geographic information system-based mapping of local climate zones in Beijing, China

The vague urban-rural dichotomy severely restricts effective comparisons and communications among urban heat island studies. A local climate zone (LCZ) scheme has therefore been developed to classify urban and natural landscapes in a standardized and universal manner. Despite LCZ mapping efforts in worldwide cities, this study attempts to propose an enhanced geographic information system-based workflow to enable the hierarchical classification of LCZs with fewer indicators but higher accuracies while considering supplementary classes and subclasses. Specifically, five morphological and coverage indicators that were easily obtained and well differentiated among LCZs were derived from a city street map and satellite images, and 25 LCZs (including 16 standard, 3 supplementary, and 6 sub-classified zones) were determined at a block-level according to the indicator hierarchy and criteria. The method was performed over Beijing, China, and evaluations by field surveys and google earth images showed a high accuracy with little noise and sharp boundaries, outperforming the widely-used remote sensing-based method of the World Urban Database and Access Portal Tools, particularly in terms of building height and heavy industry. Results also demonstrate that the Beijing core was dominated by open (including extremely open) mid-rise buildings (28.7%) and open low-rise buildings (12.8%), forming an inner-low-middle-high-outer-low annular building-height pattern. Significant land surface temperature differences were detected among the LCZs, where the low-rise and compact LCZs had higher temperatures than the mid-/high-rise and open LCZs during daytime, and subclasses LCZ X-B/C/D (LCZ X-E/F) generated lower (higher) temperatures than their parent classes in May. This method was proposed to augment the LCZ mapping system and further support applications (e.g., urban planning/management and climate/weather modeling) in high-density cities similar to Beijing.