| A new soil moisture data set has been developed and used to investigate spatial and temporal variations of soil moisture fields in different climatic regions, and includes actual in situ observations from Illinois, Iowa, Russia, China, and Mongolia. The data from China, Mongolia, and Iowa have been validated and made available to the scientific community for the first time. The data are analyzed using empirical techniques, as established in Russia by Gandin and Kagan, to estimate spatial and temporal autocorrelations.; The theory of temporal autocorrelation of soil moisture proposed by Manabe and Delworth gives guidance on the statistical structure of soil moisture, and allows independent validation of empirical estimates of scales of temporal autocorrelation of soil moisture. I find that the temporal autocorrelation may be described as an exponential function. There is a good agreement with the theory of Manabe and Delworth. For the upper 1 m of soil, temporal scales are on the order of 2 to 3 months for all of the locations, even though they are from different climate regimes. Scales are longer for layers deeper than 1 m.; Previous results by Meshcherskaya, Vinnikov, and Robock give guidance for studying the spatial autocorrelation of soil moisture. I find that the spatial autocorrelation may also be expressed as an exponential function and is on the order of several hundreds of kilometers. Again, the scales are in same range for the different regions. The scales of spatial autocorrelation of soil moisture resemble the scales of spatial autocorrelation of precipitation.; Small-scale variations of soil moisture are related to heterogeneity of soils, infiltration, topography, and vegetation. Here I separate the variance into its small-scale and large-scale components, and estimate the magnitudes of these different scales.; Both temporal and spatial scales of autocorrelation of soil moisture exhibit seasonal variation, due to the component of soil moisture variation produced by atmospheric forcing.; Examples of several possible applications of the spatial and temporal structure of soil moisture variations are discussed, including evaluation of land-surface models and satellite remote sensing of soil moisture. |
