Field-scale root-zone soil moisture : spatio-temporal variability and mean estimation

This thesis is focused around improving soil moisture estimates of spatial variability and mean at the field scale, which are useful for many different applications. The objectives were: (1) examine soil moisture spatial patterns and variability within field scale, and (2) compare field-scale soil moisture determination methods. An observational study was conducted, in which soil moisture was monitored over a 500 m by 500 m area during two and a half growing seasons at a prairie pasture in central Saskatchewan. Analysis of the spatial patterns of root-zone soil moisture revealed two distinct spatial patterns representing wet (spring) and dry (fall) periods. The relationship between spatial variability and mean soil moisture was found to follow an unusual concave trend, where variability was smallest at mid-range moisture contents. These spatial variability characteristics are a result of differences in participation level. Some locations were non-participating having only small moisture changes over the growing season, while others were dynamic having large changes. At the pasture site, these participation differences are a result of high soil heterogeneity, which may be characteristic of Solonetzic soils. In the second part of this thesis, methods to determine mean field-scale root-zone soil moisture were evaluated. The cosmic-ray neutron probe has the most potential for providing field-scale measurements. However, these measurements are only representative of moisture in the top 20 cm of soil, and need to be scaled up in order to represent the entire root-zone (0–110 cm). The three scaling methods applied to obtain lower root-zone soil moisture were: (1) a single time stable location, (2) representative landscape unit, where a single monitoring profile for each vegetation type was used, and (3) modeling by exponential filter. The representative landscape unit approach estimated soil moisture changes well, but not volumetric water content. The time stability method performed the best, followed by the exponential filter. However, the exponential filter has more potential, as the time stability method is difficult to apply to other field sites; particularly those without existing soil moisture instrumentation, due to its calibration requirements. The findings of this thesis make a contribution to the large body of existing literature on soil moisture variability and scaling. Suggestions for future research are provided