Study and Application of Methods of Soil Moisture Content Data Assimilation for Regions Partially Lacking of Observation Data

Abstract

Data assimilation is always considered as the reliable method to obtain space-time continuous soil moisture content, which is of great significance for regional water resource management. Effective observation data input plays an important role in maintaining steady output of assimilation error and improving the capacity of data assimilation. During the process of data assimilation with remote sensing for observation on regional scale, it is hard to guarantee the completeness of regional observation data. Sometimes, some parts of research region have no remote sensing observation, which leads that region of the modelling and data assimilation has no coordinate with region of remote sensing observation. In the part of region without remote sensing observation (lacking-observation region), data assimilation can’t be performed like in the part of region with remote sensing observation (observation region) because there no observation data to input but modelling data to the data assimilation system. And then it may influence the precision of data assimilation results in lacking-observation region. Thereby, we should promote the completeness of regional observation data and maybe fill the remote sensing images with some mathematical and statistical methods, one of which is that extend the differences(model errors) between modelling results and remote sensing observation data in the observation region to the lacking-observation region. This research used geostatistical methods to extend model errors from observation region to lacking-observation region. Then we calculate data assimilation results in lacking-observation region combined with land surface modelling results in lacking-observation region, thus to obtain the complete data assimilation results in whole research region. It’s shown that this method can truly improve data assimilation capacity in the overall research region. Meanwhile, this method can improve the data assimilation effectiveness in current period and also the subsequent period. It shows this method has active effects on maintaining the continuous and steady data assimilation result. Thus, this method can be an effective means to improve regional data assimilation capacity when lack of regional observation data.