Physical controls of near-surface soil moisture across varying spatial scales in an agricultural landscape during SMEX02

Understanding of near-surface soil moisture variability at different spatial scales and associated dominant physical controls is limited. In the past, soil moisture dynamics studies have been conducted extensively at different spatial scales using both in situ and remote sensing (RS) data in the subhumid Southern Great Plains region, which has mostly pasture and range land cover with rolling topography. Compared to the past efforts, we investigated the space-time characterization of near-surface soil moisture and associated physical controls at multiple scales (field, watershed, and region) in a humid hydroclimatic region with different topography and agricultural land cover. Soil moisture data from two different measurement support scales (theta probe based (point scale) and airborne RS derived; footprint scale, 800 m x 800 m), obtained during the Soil Moisture Experiment 2002 (SMEX02) in Iowa were used. Geostatistical analysis showed the spatial soil moisture correlation lengths varied between 78 m and 307 m (at the field scale), 2044 m and 11,882 m (at the watershed scale), and 19,500 m and 118,500 m (at the regional scale). The correlation length values were usually smaller on wet days than the relatively dry days at the field and watershed scales. The trend was opposite at the regional scale with correlation lengths being larger on wet days. Furthermore, the soil moisture data sets were decomposed into spatial Empirical Orthogonal Function (EOF) patterns, and their relationship with various geophysical parameters (rainfall, topography, soil texture, and vegetation) was examined to determine the dominant control on the near-surface soil moisture variability. At the field scale, the first four EOFs together explained about 81% of the total variability. At the watershed scale, the first two EOFs were dominant explaining about 93% of the total variance, whereas at the regional scale, the primary EOF itself explained more than 70% of the variance. In other words, the complicated dynamics of near-surface soil moisture fields can be described by a few underlying orthogonal spatial structures related to the geophysical attributes of the region. Correlation analysis of the RS soil moisture data showed that rainfall, topography, and soil texture have mixed effects on the variability explained by the dominant EOFs, at both watershed and regional scales, with limited influence of vegetation parameters. The effect of rainfall on the soil moisture variability is higher at the watershed scale compared to the regional scale in Iowa.