A Scale-Wetness-Heterogeneity (SWHET) Cuboid to Validate Footprint Scale Soil Moisture

Abstract:

The validation of footprint scale soil moisture is severely limited because of the scale discrepancy between footprint scale soil moisture and observed ground based data that is typically used to validate soil moisture products. Scale discrepancy poses a problem in validation because of the difference in land-surface heterogeneity encompassed between the two scales. In this study, we propose a data-driven, scale appropriate scheme that can be used to validate footprint scale soil moisture. The designed scheme generates the spatial variance structure of footprint scale moisture redistribution as a function of a scale appropriate dominant physical factor on which soil moisture redistribution depends. The scheme was developed for a variety of heterogeneous conditions found in 3 regions (Arizona, Iowa and Oklahoma) with different hydro-climates and 3 footprint scales (0.8 km, 1.6 km and 3.2 km). Our results indicate that the spatial variance of moisture redistribution can be effectively modelled as a function of the most dominant physical control (linking parameter- α’, Figure 1) with minimum mean R² = 0.8 and maximum mean RMSE = 0.119 in each case. In order to make the validation scheme potentially transferrable to regions with heterogeneity that is different from the data used, the validation scheme was extended by exploiting the relationship between soil moisture, scale and heterogeneity through a conceptualized ‘Scale-Wetness-Heterogeneity (SWHET) cuboid’ (Figure 1). The SWHET cuboid is described by a scale index, wetness index, and a newly defined heterogeneity index which can adequately quantify the land-surface heterogeneity across scales. The proposed SWHET cuboid can potentially serve as a look-up graph to model the spatial variance of footprint scale moisture redistribution as a function of the dominant physical controls. The within region validation of SWHET resulted in rmse values <0.002 in all three regions. The SWHET generated from this study is season and location specific for now but has the potential to be temporally and spatially (across different hydro-climates) transferable.