The Impacts Of Updated Land Use/Land Cover And Surface Data Assimilation On Precipitation Simulation In Complex Terrain

Initial and underlying surface conditions have significant impacts on the simulations of regional weather or climate with numerical models,especially in complex terrain.How to simulate accuratly regional weather and climate in the complex terrain has always been a challenging issue.Land use and land cover(LUC),as one of important underlying surface parameters,varies with anthropogenic activities.The LUC data used in the current models may not precisely reflect actual scenario,and the difference between the model default and actual LUC may result in an unignorable effect on the simulations of regional weather or climate factors.Surface data assimilation is an effective approach to improve regional weather or climate simulations,and it is another concern on the comprehensive impacts of updated LUC and surface data assimilations.In this regard,the Qinba Mountain area is selected as the research object for its complex terrain conditions.In order to analyze the role of the updated LUC and surface data assimilation on the simulation and attempt to provide an effective approach to obtain more accurate simulated data for regional weather and climate,four sets of four-month numerical experiments from May to August 2015 were conducted with the weather research forcasting(WRF)model,in which ERA5 reanalysis data was used to provide initial and boundary conditions for the WRF model,MODIS 2015 LUC data was used to update the default model data in the updated LUC experiment,and a combination of spectral nudging and three-dimensional variational method was used to assimilate surface observations.Comparing the simulated 2-meter temperature and precipitation from the above four experiments with the observations,the main conclusions were drawn as follows:(1)Using ERA5 reanalysis data as initial and boundary conditions,the spatial pattern of 2-meter temperature and precipitation simulated by the CON experiment are consistent with observations.The CON experiment presents a good simulation ability on 2-meter temperature in the central-eastern part of the Qinba Mountains,but it performs poorly in the western part of the Qinba Mountains.Cold biases exist in the simulated 2-meter temperature in most areas of the Qinba Mountains.The CON experiment has a certain ability in simulating precipitation in the Qinba Mountains,and there is a wet bias in the simulated precipitation in the central-western part of the Qinba Mountains,and a dry bias in the simulated precipitation in the southeastern part of the Qinba Mountains.(2)There is a significant difference between MODIS 2015 and MODIS 2001 LUC data,with 70.4% of the changed LUC area in Qinba Mountain.Compared with MODIS 2001 LUC data,the changes in MODIS 2015 LUC data are mainly reflected by the following three categories: the conversion of mixed forests to deciduous broadleaf forest,the conversion of mixed forests to woody savannas,and the conversion of croplands to woody savannas,with the changed area accounting for 24.1%,17.2%,and 9.0%,respectively.The changes of LUC types cause changes in leaf area index,surface albedo and surface emissivity.The conversion of croplands(mixed forest)to woody savannas results in a decrease in leaf area index/surface emissivity and an increase in surface albedo,while the conversion of mixed forests to deciduous broadleaf forest results in a decrease in leaf area index,surface emissivity and surface albedo.(3)Either updated LUC data or surface data assimilation can effectively reduce the bias of simulated 2-meter temperature and precipitation in the Qinba Mountains.With updating LUC data,the reduced cold biases of simulated 2-meter temperature occur in the Hanjiang River basin,Danjiang River basin and the southtern part of the of Micang-Daba Mountains where the LUC data mainly changed from croplands(mixed forest)to woody savannas.The reduced wet biases of simulated precipitation occur in the southern part of Shaanxi region,eastern part of the Funiu Mountains and the Shennongjia.Although the use of more accurate LUC data makes a positive contribution to the simulation of precipitation and 2-meter temperature in the Qinba Mountains,the improvement effect is not as significant as that of surface data assimilation.(4)The surface data assimilation with updated LUC experiment has the best simulation effect on 2-meter temperature in the central-eastern part of the Qinba Mountains,followed by the experiment of surface data assimilation only,and then the experiment of updated LUC data only.All the results of the above three experiments are better than the reference experiment.As for precipitation,updating LUC data and(or)surface data assimilation have a positive effect on the simulation of precipitation in the Qinba Mountains,but there is variability in the simulation effect of updating LUC or surface data assimilation in different regions.The simulation results of the two surface data assimilation experiments are generally better,which is possibly result from that surface data assimilation may improve the accuracy of the initial conditions,and lead to changes in the circulation field and water vapor flux field,which are more conducive to reducing the simulation bias of precipitation and making it more consistent with observation.The surface data assimilation based on updated LUC data is an effective approach to improve regional weather or climate simulations.