摘要

Urban heat island (UHI) effect is an important impact factor of the regional climate and ecological environment. How to observe and analyse the spatial distribution of UHI has become an important issue of urban environmental research. In this paper, the near-surface air temperature of Beijing was derived based on the Landsat/TM satellite imagery on 26 July 2011 to study the near-surface UHI. A statistical model at 195-m window size was established to estimate the air temperature, using land surface temperature, normalized difference vegetation index, altitude, and surface albedo as independent variables. The mean absolute error (MAE) of the model was 0.87 degrees C, and the R-2 was 0.66, indicating that the method can be used to effectively estimate the air temperature. The air temperature distribution obtained from remote sensing revealed that the UHI effect in Beijing was very significant and showed a concentrated pattern. The heat island intensity was stronger in the southern part than in the northern part of the city. In addition, the relationship between the air temperature and impervious surfaces was analysed. The air temperature increased with increasing impervious surface coverage, and the rate of change depended on the impervious surface coverage. When the impervious surface coverage was below 40 per cent, the temperature increased rapidly with increasing impervious surface coverage, and when the impervious surface coverage was above 40 per cent, the temperature increased slowly. This study provides a new approach to monitor near-surface UHI and reveals its relationship with impervious surface, providing a scientific reference for urban planning and environmental assessment.