| 基于自动站资料的WRF-EnSRF陆面同化系统的效果检验及应用 |
| |
| 引用本文： | 闵锦忠,车渌,郭亚凯.基于自动站资料的WRF-EnSRF陆面同化系统的效果检验及应用[J].大气科学学报,2016,39(3):318-328. |
| |
| 作者姓名： | 闵锦忠 车渌 郭亚凯 |
| |
| 作者单位： | 南京信息工程大学 气象灾害教育部重点实验室, 江苏 南京 210044;南京信息工程大学 大气科学学院, 江苏 南京 210044;南京信息工程大学 大气科学学院, 江苏 南京 210044 |
| |
| 基金项目： | 国家重点基础研究发展计划(937计划)项目(2013CB430102);国家自然科学基金重点项目(2014g109);江苏省高校自然科学研究计划项目(20110057) |
| |
| 摘 要： | 基于集合平方根滤波方法(En SRF)同化方法和NOAH陆面模式的WRF-En SRF陆面同化系统,同化了江苏省70个自动站资料进行试验,研究加入不同的同化资料(地表温度、10 cm土壤温度、20 cm土壤温度)及初始扰动强度的大小对陆面数据同化系统性能的影响,以及对不同区域(降水大值区和降水小值区)的分析场进行效果对比,并且检验了同化系统在一次典型的梅雨锋暴雨的同化效果,证明了这个系统的有效性和可行性。对于资料选取试验,比较全场平均的同化时刻分析场模拟观测相对真实观测的均方根误差可以得到:同化地表温度资料并且初始扰动强度1 K的时候同化效果最理想。对于选定的降水大值区和降水小值区来讲,降水大值区的土壤温度和土壤湿度分析场更加接近于真实场。运用于一次梅雨锋暴雨的同化实验,对于最后一个同化时次的分析场作为背景场做集合预报,最终证明预报结果是有效的。土壤温度、土壤湿度、地表温度和近地面风场的预报结果都较用NCEP再分析资料直接做预报作为控制试验的结果有不同程度的改进。这说明该系统应用于实际同化中的性能较为良好,可以应用于实际土壤湿度与温度的预报。
|
| 关 键 词： | 陆面资料同化 集合均方根 滤波 自动站资料 江苏省 |
| 收稿时间： | 2014/4/4 0:00:00 |
| 修稿时间： | 2014/6/3 0:00:00 |
Testing and application of a land data assimilation system using automatic weather station data |
| |
| MIN Jinzhong,CHE Lu and GUO Yakai.Testing and application of a land data assimilation system using automatic weather station data[J].大气科学学报,2016,39(3):318-328. |
| |
| Authors: | MIN Jinzhong CHE Lu and GUO Yakai |
| |
| Institution: | Key Laboratory of Meteorological Disaster, Ministry of Education(KLME), Nanjing University of Information Science & Technology, Nanjing 210044, China;School of Atmospheric Sciences, Nanjing University of Information Science & Technology, Nanjing 210044, China;School of Atmospheric Sciences, Nanjing University of Information Science & Technology, Nanjing 210044, China |
| |
| Abstract: | To date,land surface data assimilation systems have tended to focus on the four-dimensional variational or ensemble Kalman filter assimilation methods using remote sensing data.Considering the advantages of the ensemble square-root filter(EnSRF)over the ensemble Kalman filter,we adopt this assimilation method,using observations without disturbance,in the present study.Following the initial construction of WRF-EnSRF in the WRF model,the WRF-EnSRF land surface data assimilation system was preliminarily constructed,along with a system for different data sources,such as radar,satellites and automatic weather stations(AWSs).Considering the latter of these data sources(i.e.,AWSs),there are many advantages over other sources,such as the dense distribution of sites,real-time recording,the convenience of the observational data as model variables,and so on.The present work begins by testing the WRF-EnSRF land data assimilation system,and then uses AWS data to complete the WRF-EnSRF land data assimilation.Finally,the validity and feasibility of the system is verified.The EnSRF algorithm used in this paper assumes that when the observation error is not related,assimilate the observational data in an orderly way;namely,assimilate each piece of observational data one-by-one.When a piece of observational data is assimilated,the analysis field will be used as the new background field for assimilating the next piece of data.With time,all observational data are assimilated and the system sets the analysis field as the next period''s initial field and then carries out the next period''s ensemble forecast.The system then analyzes the next period''s observational data,and the process recycles.This study uses the comparatively mature Noah land surface model of the WRF model.This model includes a four-layer soil module and a one-layer vegetation module,and can forecast soil humidity and temperature.Its initial temperature field and humidity field are both from the information provided by the large-scale field after interpolation.Based on the simulation of the actual example(A Mei-yu front rainstorm),the data of 70 AWSs in Jiangsu Province are assimilated.Using different observational data,this paper discusses what kind of data has the best effect.Furthermore,the paper discusses the accuracy of the analysis field in different areas.The main results can be summarized as follows:(1)The first test uses different kinds of observational data:10 cm depth soil temperature;20 cm depth soil temperature;and surface temperature.The test uses these data by permutation and combination,with or without initial disturbance intensity,enabling it to identify which kind of data is best for assimilation.The paper reports the root-mean-square error(RMSE)of the variables and pictures the analysis field of each scheme.The data selection test results,comprising the RMSE of the analysis field and the true field,show that when assimilating surface temperature data,and the initial disturbance intensity is 1 K,the assimilation effect is as expected.(2)To discuss the accuracy of the analysis field in different areas,we choose two areas for comparison:one is a high-precipitation area (32-33°N,119.5-120.5°E),and the other a low-precipitation area (33.5-34.5°N,118.5-119.5°E).Compared to the low-precipitation area,the analysis field of soil temperature and soil humidity in the high-precipitation area is closer to the true field.However,the analysis field of surface temperature in both the high-and low-precipitation area is not ideal.For both areas,the assimilation effect is closer to the real value as soil depth deepens.(3)After assimilation of surface temperature and 10 cm depth soil temperature data at the same time,we choose the analysis field of the last assimilation time as the background field,and then perform the ensemble forecast.In the end,the forecast result is proven to be effective.So,in conclusion,the result of the WRF-EnSRF land data assimilation system using AWS data applied to a practical case is satisfactory,and the analysis field and ensemble forecast are accurate.The prediction results for the soil temperature field,soil moisture field,surface temperature field and the surface layer wind field all have different degrees of improvement compared with the control test.Overall,this study shows that the performance of the system,as applied in actual assimilation,is relatively good,and the system can be used in the forecasting of soil humidity and temperature. |
| |
| Keywords: | Land data assimilation ensemble square-root filter automatic weather station data Jiangsu Province |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《大气科学学报》浏览原始摘要信息 |
| 点击此处可从《大气科学学报》下载免费的PDF全文 |
|
