基于微波植被指数的L波段多角度数据反演土壤水分算法研究

ESA's SMOS (Soil Moisture and Ocean Salinity) is the first ever space-borne synthetic aperture passive microwave instrument in the world. Its state-of-the-art synthetic aperture technology and L-band multi-angle observation capabilities has brought new opportunities for soil moisture estimation, but also raised many problems that need further studies. Aiming at the critical issue of vegetation effects, this project proposes a method using microwave vegetation indices (MVIs) to improve the accuracy of soil moisture products. In this study, a processing chain is first developed to filters out outliers caused by radio frequency interference and reduce the bias found in the brightness temperature observations through threshold determination, filtering and compensation methods. An objective function is then used to fit multi-angle data for the optimization of SMOS data. And then taking advantage of the optimized multi-angle data, MVIs are developed and calculated in order to achieve the purpose of separation of vegetation and soil signals. By theoretical analysis, physical relations between MVIs and vegetation optical thickness are established for a more effective vegetation effects correction method, which forms a SMOS soil moisture inversion algorithm with more physical mechanism. During meantime, ground-based observation experiment is carried out for typical crop to obtain continuous microwave radiation data and corresponding soil, vegetation parameters. Combined with a wider range of soil moisture observation networks, the new algorithm is validated and analyzed. And a comprehensive assessment of the improvements of the SMOS soil moisture product is achieved by comparison with other inversion algorithms. This research could further enhance the application potential of SMOS data in agricultural production and drought and flood monitoring.

SMOS作为一颗具有划时代意义的地球观测卫星，其先进的合成孔径技术和L波段多角度观测能力为土壤水分反演带来了新的发展机遇，但同时也涌现诸多问题有待于深入研究。本项目围绕植被影响校正这一关键问题，提出使用微波植被指数的方法来改善土壤水分产品精度，即首先通过阈值判定、滤波补偿等方法降低亮温观测偏差和射频干扰，并建立目标函数对多角度数据进行回归，实现对于SMOS亮温数据的优化；接着利用优化的多角度数据发展微波植被指数计算方法，以达到分离植被和土壤信号的目的，理论分析微波植被指数同植被光学厚度之间的物理关系，建立更为有效的植被影响校正方法，形成物理机制更强的SMOS土壤水分反演算法；同时针对典型农作物开展地面观测试验，并结合更为广泛的地面观测数据，进行算法验证和误差分析，通过与其它反演算法进行比较，全面评估SMOS土壤水分的改善效果，以期进一步提高SMOS数据在农业生产和旱涝监测等方面的应用价值。

SMOS作为一颗具有划时代意义的地球观测卫星，其先进的合成孔径技术和L波段多角度观测能力为土壤水分反演带来了新的发展机遇，但同时也涌现诸多问题有待于深入研究。本项目（1）首先根据不同极化下的地表辐射亮温特征，建立了适用于多种地表类型的目标函数对多角度数据进行回归，拟合得到的亮温数据的波动和极化特征背离理论预期等现象明显得到抑制；（2）分析研究地表粗糙度效应随极化和角度的变化规律，寻找最适用的自相关函数类型和最优化的粗糙度参数化方案，建立了不同角度下的土壤微波辐射计算模型；（3）围绕植被影响校正这一关键问题，提出使用微波植被指数的方法来改善土壤水分产品精度，通过理论分析微波植被指数同植被光学厚度之间的物理关系，以达到分离植被和土壤信号的目的，建立了更为有效的植被影响校正方法，形成了新的物理机制更强的SMOS土壤水分反演算法，有效提高了SMOS数据在农业生产和旱涝监测等方面的应用价值。