Numerical study of urbanization effect on a heavy storm event in Beijing

Abstract:

In the past few decades, Great Beijing area has experienced rapid and widespread urbanization, which has substantially modified the land surface physical characteristics and affected urban regional weather and climate.

A single layer urban canopy module has been developed based on the Community Land Surface Model Urban Module (CLMU) with several improvements: the energy balances on the five surface conditions (building roof, sun side and shaded side wall, pervious and impervious land surfaces) are considered separately. A method to calculate sky view factor is developed based on the physically process while most urban models simply provide an empirical value. This method improves the solar and long wave radiation simulation on each surface. The latent heat flux on both wall and impervious land is calculated; the anthropogenic heat is considered in terms of industrial production, domestic wastes, vehicles and air conditions.

The urban effect on summer convective precipitation over Beijing was investigated by simulating a heavy storm event on July 21^st 2012, when precipitation of averagely 164 mm was brought to Beijing within 6 hours, which is the heaviest during the past 60 years’ record in the region. Numerical simulating experiment was set up by coupling Weather Research and Forecast (WRF)/SSiB3 model with the Modified CLMU (MCLMU). Several control cases without MCLMU were set up. The horizontal resolution in the inner domains was set to be 2 km.

While all of the control results drastically underestimate the urban precipitation, the result of WRF/SSiB3/MCLMU is much closer to the observation. Sensitive experiments show that although large areas of impervious surfaces restrain the surface latent heat flux in urban, the anthropogenic heat and enhanced sensible heat flux warmed up the lower atmospheric layer then strengthen the vertical stratification instability, which is the key factor for heavy storm while moisture is sufficient.