水文模型模拟预报的多源数据同化方法及应用研究进展

介绍了水文遥感数据同化中常用的数据同化方法，总结了变分和顺序两类数据同化常用方法的优势与不足；以土壤湿度与径流两个水文变量的数据同化研究为重点，探讨了土壤湿度、径流、降水、蒸散发、积雪等多源数据在水文模型模拟预报中的同化研究进展及其在同化应用中存在的问题；最后，从数据同化方法、多源数据同化应用方面总结了水文遥感数据同化的未来发展方向，提出遥感、地面等多源数据的同化在改进水文模型模拟预报方面的应用潜力将会随着遥感观测技术与反演方法的改进、水文模型结构的完善以及数据同化方法的优化而不断增大，多源数据在水文模型模拟预报中的综合应用将是水文遥感数据同化发展的必然趋势。

Advances in multi-source data assimilation approach and application in simulation and forecast of hydrological model

In this study, the approaches adopted in the assimilation of hydrological remote sensing data were introduced. Emphasis was placed on the advantages and disadvantages of the commonly used methods in both variational and sequential data assimilation. Focusing on the assimilation of soil moisture and streamflow, the application progress of multi source observations including the soil moisture, streamflow, precipitation, evapotranspiration and snow cover in hydrological simulation and forecast was analyzed in detail. On this basis, the issues existed in the application of multi source observations in hydrological modeling were explored. Finally, future development trends of hydrological remote sensing data assimilation were proposed in terms of both data assimilation method and multi source data assimilation application. It is believed that the potential of remote sensing and in situ multi source data assimilation in improving the hydrological simulation and forecast will increase with the improvement of remote sensing observations and reversion techniques, the refinement of hydrological model structure and the optimization of data assimilation approaches. The integrated application of multi source observations in hydrological simulation and forecast will be the inevitable trend in the development of hydrological data assimilation.