Machine Learning Techniques for Downscaling SMOS Satellite Soil Moisture Using MODIS Land Surface Temperature for Hydrological Application
Machine Learning Techniques for Downscaling SMOS Satellite Soil Moisture Using MODIS Land Surface Temperature for Hydrological Application
复制标题
DOI：
10.1007/s11269-013-0337-9
复制
发表时间：
2013-04
影响因子：
4.3
通讯作者：
P. Srivastava;Dawei Han;M. Ramirez;T. Islam
中科院分区：
文献类型：
--
作者：
P. Srivastava;Dawei Han;M. Ramirez;T. Islam
许多水文现象和应用，如干旱，洪水，灌溉管理和调度需要高分辨率的卫星土壤水分数据在当地/区域尺度。降尺度是将粗糙域卫星数据转换为更精细空间分辨率的一个非常重要的过程。三个人工智能技术沿着与广义线性模型（GLM）被用来提高空间分辨率的土壤水分和海洋盐度（SMOS）派生的土壤水分，这是目前在一个非常粗糙的尺度~40公里。采用人工神经网络（ANN）、支持向量机（SVM）、相关向量机（RVM）和广义线性模型对中分辨率成像光谱仪（MODIS）反演的地表温度（LST）与SMOS反演的土壤湿度进行了集成。土壤水分亏缺（SMD）来自一个水文模型称为PDM（概率分布模型）用于降尺度性能评价。利用白天和夜间平均PDM SMD数据对白天和夜间的MODIS LST差异进行了统计评价，以选择有效的MODIS产品。所有的降尺度算法的准确性和鲁棒性进行了讨论，在他们的假设和适用性。R2、% Bias和RMSE等统计性能指标表明，（R2= 0.751，% Bias = −0.628andRMSE = 0.011），RVM（R2= 0.691，% Bias = 1.009，RMSE = 0.013），SVM（R2= 0.698，% Bias = 2.370，RMSE = 0.013）和GLM（R2= 0.698，%Bias = 1.009andRMSE = 0.013）算法在整体上相对更巧妙地降低土壤水分的变异性，与非对降尺度数据的预测结果表明，ANN算法的预测效果优于其它算法，其R2= 0.418，RMSE = 0.017。与生长和非生长季节相关的其他尝试已在本研究中使用，以揭示基于季节的降尺度甚至比具有相当高的性能统计的连续时间序列更好。
Many hydrologic phenomena and applications such as drought, flood, irrigation management and scheduling needs high resolution satellite soil moisture data at a local/regional scale. Downscaling is a very important process to convert a coarse domain satellite data to a finer spatial resolution. Three artificial intelligence techniques along with the generalized linear model (GLM) are used to improve the spatial resolution of Soil Moisture and Ocean Salinity (SMOS) derived soil moisture, which is currently available at a very coarse scale of ~40 Km. Artificial neural network (ANN), support vector machine, relevance vector machine and generalized linear models are chosen for this study to integrate the Moderate Resolution Imaging Spectroradiometer (MODIS) Land Surface Temperature (LST) with the SMOS derived soil moisture. Soil moisture deficit (SMD) derived from a hydrological model called PDM (Probability Distribution Model) is used for the downscaling performance evaluation. The statistical evaluation has also been made with the day-time and night-time MODIS LST differences with the mean day and night-time PDM SMD data for the selection of effective MODIS products. The accuracy and robustness of all the downscaling algorithms are discussed in terms of their assumptions and applicability. The statistical performance indices such asR2,%BiasandRMSEindicates that the ANN (R2= 0.751,%Bias = −0.628andRMSE = 0.011), RVM (R2= 0.691,%Bias = 1.009andRMSE = 0.013), SVM (R2= 0.698,%Bias = 2.370andRMSE = 0.013) and GLM (R2= 0.698,%Bias = 1.009andRMSE = 0.013) algorithms on the whole are relatively more skillful to downscale the variability of the soil moisture in comparison to the non-downscaled data (R2= 0.418andRMSE = 0.017) with the outperformance of ANN algorithm. The other attempts related to growing and non-growing seasons have been used in this study to reveal that season based downscaling is even better than continuous time series with fairly high performance statistics.