A high-resolution monitoring approach of canopy urban heat island using a random forest model and multi-platform observations

[摘要] Due to rapid urbanization and intense human activities, the urbanheat island (UHI) effect has become a more concerning climatic andenvironmental issue. A high-spatial-resolution canopy UHI monitoring methodwould help better understand the urban thermal environment. Taking the cityof Nanjing in China as an example, we propose a method for evaluating canopyUHI intensity (CUHII) at high resolution by using remote sensing data andmachine learning with a random forest (RF) model. Firstly, the observedenvironmental parameters, e.g., surface albedo, land use/land cover,impervious surface, and anthropogenic heat flux (AHF), around denselydistributed meteorological stations were extracted from satellite images.These parameters were used as independent variables to construct an RF modelfor predicting air temperature. The correlation coefficient between thepredicted and observed air temperature in the test set was 0.73, and theaverage root-mean-square error was 0.72  ∘ C. Then, the spatialdistribution of CUHII was evaluated at 30 m resolution based on the outputof the RF model. We found that wind speed was negatively correlated withCUHII, and wind direction was strongly correlated with the CUHII offsetdirection. The CUHII reduced with the distance to the city center, due tothe decreasing proportion of built-up areas and reduced AHF in the samedirection. The RF model framework developed for real-time monitoring andassessment of high spatial and temporal resolution (30 m and 1 h) CUHIIprovides scientific support for studying the changes and causes of CUHII,as well as the spatial pattern of urban thermal environments.

[学科分类] 内科医学