Classifying Method For Beijing Land Use And Improvements To The Modeling Of Urban Meteorology

With the rapid development of urbanization,the influences of city land use on weather and climate increased significantly.There is a need to improve the numerical simulation capability of urban effect in high resolution numerical weather prediction systems.Refined city underlying surface information is one of the important foundations for the urban effect simulation.This study set up a refined underlying surface database in Beijing based on the local climate zone(LCZ)city refined classification method.In order to achieve the target,high resolution remote sensing images as well as the geographic information system(GIS)software are used to carry on the reasonable classification of urban land use types.Then the refined underlying surface database is applied to the meso-scale numerical model Weather Research and Forecasting Model(WRF)to test the influence on the numerical simulations with the refined underlying surface database in a fine-weather day case.Currently,the widely-used underlying surface database of Beijing is the three categories classified by different impervious area percentage thresholds.After field inspections on Beijing actual buildings parameters,it can be concluded that the refined LCZ database showed obviously better results in urban buildings classification than the old surface database.Building heights and rates of impervious surface distributions were more consistent with the actual situations.The simulated results of WRF model that applied the refined underlying surface database were also improved,especially for the temperature simulation using the single-layer urban canopy model,LCZ simulation described the characteristics of buildings and vegetation more accurately on the scale of the community.The sensible heat fluxs and latent heat fluxs were more in line with the actual measurement,and the maximum errors were significantly reduced.Diurnal variations and spatial distributions of 2 m temperature were more consistent with the observations and solve the problem of model temperature simulation of the most low temperature in summer higher than actual effectively,the root mean square error decreased from?2.81? to 1.82?.The results of multi-layer urban canopy model were also improved.In LCZ different underlying surface types,Compact low-rise,Open mid-rise,Open low-rise and Light weight low-rise these four types of land use significantly optimized the temperature simulation results.It reflected in the absolute deviation is under 1.49 ? and the root mean square is error under 1.98 ?.The optimization of simulation results shows that it is better to supplement the vacancy of the type of default land use,and the more reasonable and more precise classification of the underlying surface can obviously improve the temperature simulation accuracy.On the spatial distribution of 10 m wind speed simulation,single and multi-layer urban canopy model simulation results can better present the wind speed distribution in the city due to the construction parameters in LCZ refined underlying surface more accord with the actual characteristics of buildings.The results of test confirmed that by applying as the basic data of urban effect research and high resolution numerical weather forecast,the refined underlying surface database of Beijing can improve the precision of numerical forecast simulation effectively.