Validation of ALEXI-derived volumetric soil moisture over the Continental United States.

Soil moisture plays a vital role in the partitioning of sensible and latent heat fluxes in the surface energy budget; however, high spatial-resolution observations of soil moisture distributions are difficult to acquire. The ALEXI model contains the two-source land-surface representation of Norman et al, 1995, which partitions surface fluxes and radiometric temperature into canopy and soil contributions based on the fraction of vegetation cover within the scene. Anderson et al. (1997) and Mecikalski (1999) detail the implementation of ALEXI as a regional-scale application over the continental United States. This model relies on remote sensing data to operate, including GOES-derived surface brightness temperature changes, satellite-derived land cover properties, and limited synoptic weather data to operate. This version of the ALEXI algorithm has been run daily on a 10 km resolution grid from the years 2002 to present.

Soil moisture conditions within the soil surface and root-zone layers yield a distinctive thermal signature, where moisture deficiency will lead to surfaces warming more quickly. ALEXI diagnoses a fraction of potential evapotranspiration (fPET) for both the surface layer (0 – 5 cm) and root-zone (5 – 200 cm), given a calculation of the potential ET for soil and canopy components of each pixel scene. In current mesoscale modeling the fraction of potential ET can be directly related to a fraction of available water, which in turn can be used to calculate volumetric soil moisture for a given soil texture. Current land-surface models (LDAS,NLDAS) such as those used in the North American Mesoscale Model (NAM) use antecedent precipitation as the primary component to the calculation of volumetric soil moisture. These models use four layers in their soil model (0-10, 10-40, 40-100, and 100-200 cm), while ALEXI provides a derived volumetric soil moisture for only two layers (0-5 and 5-200 cm). This discrepancy can be solved with a blending of the two layers from ALEXI to provide a reasonable representation of the observed behavior of the four layers used in the land-surface models.

Volumetric soil moisture estimates from ALEXI, NAM (EDAS), and LDAS will be validated against ground-based observations taken over the continental United States during the years of 2003/2004. Results will be quantified through statistical techniques. Initial results from the validation look promising, although a more robust validation is still needed.