基于集合卡尔曼滤波的土壤温湿度同化试验

发布时间：2018-03-26 20:32

本文选题：通用陆面模式　切入点：集合卡尔曼滤波　出处：《气象科学》2016年02期

【摘要】：以通用陆面模式CLM 3.0(Community Land Model 3.0)为模型算子,基于集合卡尔曼滤波(Ensemble Kalman Filter,En KF)发展了一个土壤温湿度同化系统,主要用于改进模式对土壤温湿度和地表水热通量的模拟精度,并考察集合样本数、同化频率及不同观测量的组合对同化效果的影响。该系统同化了FLUXNET两个站点(阿柔和Bondville)不同土壤深度、不同时间频率的土壤温度和湿度数据。通过对阿柔站不同集合样本数的设计,综合考虑计算成本和计算精度,最终将集合样本数设置为40。通过分析三种同化方案对同化频率的敏感性得出,同化土壤温度最为敏感,同时同化土壤温湿度次之,同化土壤湿度最不敏感。对于阿柔站点,同化系统对不同土壤深度温度和湿度的模拟精度均能提高90%,潜热通量的均方根误差由94.0 W·m~(-2)降为46.3 W·m~(-2),感热通量均方根误差由55.9 W·m~(-2)降为24.6 W·m~(-2)。Bondville站点浅层土壤温度的改进在30%左右,深层土壤温度改进达到60%,对土壤湿度的改进均在70%以上,潜热通量和感热通量的均方根误差分别从57.4 W·m~(-2)和54.4 W·m~(-2)降为51.0 W·m~(-2)和42.5 W·m~(-2)。试验结果表明,同化站点土壤温湿度数据对土壤水热状况及通量的模拟改进非常有效,同时也验证了同化土壤水分遥感产品的可行性和必要性。
[Abstract]:Taking CLM 3.0(Community Land Model 3.0) as model operator, a soil temperature and humidity assimilation system is developed based on the set Kalman filter Ensemble Kalman filter en KF, which is mainly used to improve the simulation accuracy of soil temperature and humidity and surface water heat flux. The effects of the number of samples, the frequency of assimilation and the combination of different observations on the assimilation effect were investigated. The system assimilated the different depth of soil at two FLUXNET stations (Amoi Bondville). Soil temperature and moisture data of different time frequency. Through the design of different set samples of Arou station, the calculation cost and accuracy are considered synthetically. By analyzing the sensitivity of the three assimilation schemes to the assimilation frequency, it is concluded that the assimilation soil temperature is the most sensitive, and the assimilation soil temperature and humidity is the second, and the assimilation soil moisture is the least sensitive. The simulation accuracy of assimilation system for different soil depth temperature and humidity can be improved by 90%. The root mean square error of latent heat flux is reduced from 94.0 W / m ~ (-2) to 46.3 W / m ~ (-2), and the mean square root error of sensible heat flux is reduced from 55.9 W / m ~ (-2) to 24.6 W / L soil temperature. The improvement of soil temperature in shallow layer is about 30%. The temperature improvement of deep soil reaches 60%, and the improvement of soil moisture is more than 70%. The root mean square errors of latent heat flux and sensible heat flux are reduced from 57.4 W / m ~ (-2) and 54.4 W / m ~ (-2) to 51.0 W / m ~ (-2) and 42.5 W / m ~ (-2) respectively. The soil temperature and humidity data at the assimilation site are very effective to improve the simulation of soil water and heat status and fluxes. At the same time, the feasibility and necessity of assimilating soil moisture remote sensing products are also verified.
【作者单位】：华东师范大学地理信息科学教育部重点实验室;中国气象局乌鲁木齐沙漠气象研究所;华东师范大学中国科学院对地观测与数字地球科学中心环境遥感与数据同化联合实验室;Department
【基金】：国家自然科学基金资助项目(41101037,13231203804) 华东师范大学地理信息科学教育部重点实验室主任基金(KLGIS2011C06) 中央高校基本科研业务费专项奖金项目
【分类号】：S152

【相似文献】