Spatial Analysis of Remotely Sensed Soil Moisture Data

Soil moisture plays a key role in various fields such as agriculture, meteorology, ecology, hydrology, and geotechnical engineering. A critical problem in the measurement of soil moisture through remote sensing methods is representing the spatial heterogeneity without distortion. In this study, a geostatistical technique, kriging, is used to characterize the spatial structure of the Little Washita soil moisture data measured by passive microwave remote sensing as part of the 1992 Washita study. Second-order statistical metrics verify the adequacy of kriged estimates against the observed soil moisture field. The results show that sampling of a systematic set of points produces a smoother soil moisture field compared to a random sampling of data points. To check the efficacy of kriging in estimating the unknown soil moisture values and related surface fluxes, evaporation is calculated. There appears to be a good correspondence between the evaporation values calculated from the original data and the kriged data. The frequency scaling ratio is used to identify the spatial coherence between the observed and the estimated soil moisture fields.