Developing a Stabilized Ensemble Kalman Filter for integrating daily GRACE/GRACE-FO data into process models (S-ENKF)

最近的研究建议使用集合卡尔曼滤波（EnKF）方法，将重力恢复和气候实验（GRACE）任务获得的全球质量再分布数据与水文模型相结合。这将为洪水（即流量）和干旱预报应用提供巨大的潜力，特别是当应用于次月分辨率的GRACE解决方案时。然而，到目前为止，结果并未达到理论预期，其原因可能在于GRACE横向和垂直分解问题固有的不良条件，以及用于最优估计的有限大小集合方法的数值特性。在这里，我们建议发展一个稳定的EnKF；通过实施/裁剪正则化和最优加权技术，综合气象学和大地测量学方法。稳定的方法将用于整合每日GRACE数据与0.5度空间分辨率的水文模型。该方法将应用于恒河-布拉马普特拉河-梅克纳河地区的实际数据，并通过模拟和使用独立数据进行评估。

Recent studies have suggested integrating global mass redistribution data obtained from the Gravity Recovery and Climate Experiment (GRACE) mission with hydrological models, using the Ensemble Kalman Filter (EnKF) approach. This would offer great potential for flood (i.e. discharge) and drought forecasting applications, in particular when applied to GRACE solutions at sub-monthly resolution. However, results so far have not met theoretical expectations and the reason for this is likely to be found in the inherent ill-conditioning of the GRACE lateral and vertical disaggregation problem, combined with the numerical peculiarities of limited-size ensemble approaches for optimal estimation. Here, we suggest developing a stabilized EnKF; through implementing/tailoring regularization and optimal weighting techniques, synthesizing methods from meteorology and from geodesy. The stabilized method will be applied to integrate daily GRACE data with a hydrological model at 0.5-degree spatial resolution. The approach will be applied to real data over the Ganges-Brahmaputra-Meghna region, and evaluated though simulations as well as using independent data.