Downscaling near-surface soil moisture from field to plot scale: A comparative analysis under different environmental conditions

Indirect measurements of field-scale (hectometer grid-size) spatial-average near-surface soil moisture are becoming increasingly available by exploiting new-generation ground-based and satellite sensors. Nonetheless, modeling applications for water resources management require knowledge of plot-scale (1-5 m grid-size) soil moisture by using measurements through spatially-distributed sensor network systems. Since efforts to fulfill such requirements are not always possible due to time and budget constraints, alternative approaches are desirable. In this study, we explore the feasibility of determining spatial-average soil moisture and soil moisture patterns given the knowledge of long-term records of climate forcing data and topographic attributes. A downscaling approach is proposed that couples two different models: the Eco-Hydrological Bucket and Equilibrium Moisture from Topography. This approach helps identify the relative importance of two compound topographic indexes in explaining the spatial variation of soil moisture patterns, indicating valley- and hillslope-dependence controlled by lateral flow and radiative processes, respectively. The integrated model also detects temporal instability if the dominant type of topographic dependence changes with spatial-average soil moisture. Model application was carried out at three sites in different parts of Italy, each characterized by different environmental conditions. Prior calibration was performed by using sparse and sporadic soil moisture values measured by portable time domain reflectometry devices. Cross-site comparisons offer different interpretations in the explained spatial variation of soil moisture patterns, with time-invariant valley-dependence (site in northern Italy) and hillslope-dependence (site in southern Italy). The sources of soil moisture spatial variation at the site in central Italy are time-variant within the year and the seasonal change of topographic dependence can be conveniently correlated to a climate indicator such as the aridity index.