Quantifying urban heat island intensity and its physical mechanism using WRF/UCM

Reliable quantification of urban heat island intensity (UHII) is crucial for the evaluation of extreme heat waves and the related heat stress. As a powerful approach for the study of urban climate, numericalmodels can simulate urban heat island (UHI) in both high spatial and temporal resolutions. However, accurate quantification of UHII using modelling grid data is still a challenge at present, due to the different criterions for the selection of urban/rural grids. This study simulates the high-resolution UHI in the city of Berlin using the Weather Research and Forecasting Model coupled with Urban Canopy Module. A new method to quantify UHII, which is based on the fitted linear functions of simulated 2-m air temperature (T-2m) using the impervious surface area inWRF grids (ISA(WRF)), was adopted and evaluated. The simulated T-2m matches the observations well, with a correlation coefficient of 0.95 (P < 0.01) and RMSE of 1.76 degrees C. The study area shows a strong UHI at nighttime. The simulated nighttime T-2m increases with the increase in the ISAWRF. The linear functions of simulated nighttime T-2m against ISAWRF are well fitted. The UHII is calculated as the products of the slopes of fitted functions and the largest ISAWRF. The derived UHII shows U-shaped diurnal variations, with high values at nighttime. The difference of simulated surface temperature and sensible heat flux between the impervious surface and the vegetation surface jointly determines the derived UHII. The large difference of surface temperature and the small difference of sensible heat flux between the impervious and the vegetation surface generate the high UHII at nighttime and vice versa during the daytime. The method of ISA(WRF)-based function of T-2m overcomes the problems of traditional methods in arbitrary selecting urban/rural grids. It can be used easily to quantify UHII and to do the comparison study of UHII between different cities. (c) 2018 Elsevier B.V. All rights reserved.