论文部分内容阅读
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。
Based on the Entropy Kalman Filter (En KF), CLM 3.0 (Community Land Model 3.0) is used as a model operator to develop a soil temperature and humidity assimilation system, which is mainly used to improve soil temperature and humidity and Surface water heat flux simulation accuracy, and to examine the collection of samples, assimilation frequency and the combination of different observations on the assimilation effect. The system assimilates data on soil temperature and humidity at different soil depths and frequencies over time at two sites at FLUXNET (Aloft and Bondville). Through the design of different sets of samples in A soft station, considering the calculation cost and calculation precision, the number of the set samples is finally set as 40. By analyzing the sensitivity of three assimilation schemes to assimilation frequency, it is concluded that assimilation of soil temperature is the most sensitive, followed by assimilation of soil temperature and humidity, and assimilation of soil moisture is the least sensitive. For the Alhual site, the simulation accuracy of the assimilation system for different soil depths can be improved by 90%, and the root mean square error of the latent heat flux decreased from 94.0 W · m -2 to 46.3 W · m -1. 2). The root mean square error of sensible heat flux decreased from 55.9 W · m -2 to 24.6 W · m -2. The improvement of shallow soil temperature in Bondville site was about 30%, the improvement of deep soil temperature reached 60%, the improvement of soil moisture was over 70%, and the root mean square error of latent heat flux and sensible heat flux were respectively 57.4 W · m -2 and 54.4 W m -2 decreased to 51.0 W m -2 and 42.5 W m -2, respectively. The experimental results show that the data of soil temperature and humidity at the assimilation site are very effective in simulating soil hydrothermal conditions and fluxes, and also verify the feasibility and necessity of assimilating soil moisture remote sensing products.