A transferable approach to assessing green infrastructure types (GITs) and their effects on surface urban heat islands with multi-source geospatial data

Urban green infrastructure (GI) is essential for mitigating surface urban heat islands (SUHIs) and strengthening urban resilience to climate change, thereby contributing to the achievement of sustainable development goals in urban areas. A 'green infrastructure types' (GITs) scheme was recently developed to examine the role of amount, composition, and configuration of GI in providing effective thermal cooling in Australia. However, the GIT scheme has not been applied to an urban environment in other countries, so its general suitability for SUHI assessment needs to be further investigated. Taking the urban core area of Beijing as a case study, a multi-level classification method was developed in this study for GIT mapping using bi-seasonal high-resolution Gaofen-1 satellite imagery, time-series Sentinel-2 A/B and SDGSAT-1 satellite data, and open vector datasets. The time series of 30 m land surface temperature (LST) data was created by blending high temporal resolution MODIS and Landsat data. Statistical analysis was performed to identify factors that may influence the cooling effect of GI. The results demonstrate significant LST differences among the GITs in summer. The aquatic GITs with a water fraction >50% (5.6–7.6 °C) and pervious GITs with a high tree fraction (5.4–5.6 °C) provided the largest cooling effect. The cooling capacity of mixed and pervious GITs was found closely related to the proportion of pervious surface and woody vegetation. Increasing the area of irrigated grass from low (28%) to medium (61%) in mixed surfaces produced only a marginal effect on surrounding LSTs. The GIT mapping approach combined with SUHI analysis provides a transferable and promising framework for examining the cooling effect of GI at the neighborhood level in a consistent manner.