热力粗糙度引入精细城市边界层模式的初步应用

A Study of Introducing Temperature Roughness Length into a Fine Urban Boundary Layer Model

运用南京大学城市尺度边界层模式(UBLM)以100 m水平网格距的精细高分辨率,诊断分析小规模城市(水平范围在一二十公里)的气象环境与边界层结构。就高分辨率精细模拟分析研究了两个问题:(1)在模式中引入热力粗糙度,就引入热力粗糙度前后的模拟结果与实测结果进行对比,结果表明引入后的模拟结果与自动气象站的观测结果吻合较好,能够很好地反映出地面气温的日变化规律。冬季,白天不引入热力粗糙度比引入热力粗糙度会高估小规模城市地区感热通量约60 W/m2,而对乡间农作物地区约10 W/m2。(2)以四种不同水平网格距对同一下垫面及气象条件进行诊断分析,结果表明以100 m网格距精细模拟效果明显优于以较大网格距进行的模拟效果。以这样的精细高分辨率模式所获模拟结果分析了小规模城市的边界层特征,表明:冬季白天和夜间小规模城市与乡村气温差分别为0.8℃和0.6℃;风速明显低于乡村地区;湍能高于乡村地区,白天和夜间湍能一般为乡村的3倍左右;白天混合层发展较乡村迅速,且高度高于乡村地区。这样的特征在当今城市规划布局工作中,为特大型城市的卫星城市的建设规划提供了有益的科学实验依据。模拟效果表明以如此高的水平分辨率实施的模式和模拟是可行和有意义的。

The present boundary layer model whose horizontal grid spacing is more than 500 m is not efficient to simulate a town whose horizontal range is fewer than twenty kilometers, as the large horizontal grid spacing cannot satisfactorily distinguish the land surface types from each other. For the present study, the Urban Boundary Layer Model （UBLM） of the multi-scale numerical modeling system with a high resolution （100 m× 100 m） is used to simulate the meteorological environment of such a town. UBLM is a three-dimensional model with non-hydrostatic E-ε turbulence closure scheme, and it has been used to diagnose the meteorological environment and the characteristics of the boundary layer over complex terrains. The UBLM is in terrain-following coordinates, as the horizontal grid spacing is 100 m and the least vertical grid spacing is 5 m. The initial meteorological fields are derived from the results of NJU-Regional Boundary Layer Model （NJU-RBLM） with its horizontal grid spacing being 500 m. Based on the UBLM with a high resolution （100 mN 100 m）, this study diagnoses and analyses the diurnal variation of the meteorological environment and the boundary layer structure of a town in winter. The research methods and results are as follows： 1） the simulated results of surface air temperature, when taking temperature roughness length into account, are in reasonable agreement with the observational results of automatic weather station, and they actually reflect the diurnal variation of the temperature. The relative mean bias of surface air temperature is only 0. 22 and its correlation coefficient is 0. 77. In the daytime of winter, the simulated results without temperature roughness length will overestimate the sensible heat flux by 60 W/m^2 or so in the town, and about 10 W/m^2 or so in the rural area. 2） By comparing four different scales of horizontal grid spacing in the same area with the same mete- orological condition, the diagnostic analysis shows that the results from a fine simulation （100 m horizontal grid spacing） are apparently better than those done with larger-scale simulations. Then, by analyzing the boundary layer structure of the town in winter with the results from high-resolution simulation, the study shows that the temperature differences between the town and the rural area are 0.8℃ in the daytime and 0. 6℃ in the nighttime respectively, and that the wind speed in the town is lower than that in the rural area. Simultaneously, the turbulent kinetic energy of the town is three times that of the rural area, and the evolution of boundary layer of the town is quicker than that of the rural area, with the boundary layer higher than that of the rural area. In the daytime of winter, the height of the boundary layer in the town is about 700 m, while that in the rural area is around 500 rn. In urban planning, therefore, these characteristics of the town provide scientific evidence for the planning of the satellite cities around the megalopolis. The simulation results prove that such high-resolution model and simulation are both practical and feasible.