Spatio-temporal dynamics of soil moisture in an agricultural watershed
The spatial and temporal dynamics of surface, 0-0.06m, soil moisture over a 24Ha
agricultural site is examined by using spatially intensive empirical point measurements.
Statistical analyses of the soil moisture indicate a variable spatial structure with
antecedent rainfall and time of the year having the greatest influence on the spatial field.
Static and dynamic indices do not have consistent statistically significant relationships
with the soil moisture fields. The spatial correlation structure has a temporal evolution,
in terms of the amount of spatial variance and the organisation of the spatial soil
moisture field, which is dependent on the mean moisture content. Several interpolation
procedures are considered to analyse the linkage between measurement and modelling
scale. Ordinary kriging provides the best goodness of fit between the computed
realisation and the empirical data, but simplifies the data and a value for the uncertainty
value is not calculable. Conditional simulation allows for the assessment of the
uncertainty, but with a reduced accuracy between the representation and the empirical
data. Comparison of the surface moisture and hydrological flux within the profile
shows that the soil moisture at this depth is disassociated from the hydrological
processes at depth. One-dimensional modelling of the moisture fluxes returns a close
association between the modelled and measured moisture, even with inaccurate
modelling of the fluxes. Thus, limiting the applicability of shallow surface soil
moisture in the internal validation of current generations of hydrological models.