Investigating the Impact of Soil Texture on Satellite Retrieved Soil Moisture.

Near-surface soil moisture is currently being measured globally by both the Soil Moisture Ocean Salinity (SMOS) and Soil Moisture Active Passive (SMAP) satellites. Accurate retrieval of soil moisture, necessary to improve weather and climate forecasts, is dependent on the parameterization of the emitting surface. Our ability to quantify important land surface characteristics that affect soil moisture retrieval (e.g., soil texture, land cover, and surface roughness) improves over time and subsequently the auxiliary data files that specify these effects during retrieval are updated periodically. In the interest of inter-mission comparability, the SMOS soil texture map has been modified to use the same database utilized by SMAP (as of L2SM v.620, operational since 5 May 2015). This change resulted in enhanced spatial resolution of the auxiliary soils data utilized by SMOS, from a 5' x 5' (9.25km x 9.25km at the equator) grid to 4km x 4km.

We have compared two years of SMOS soil moisture (retrieved using the original SMOS soil map) with the soil moisture from a 20-node in situ soil moisture network in the South Fork of the Iowa River (SFIR) watershed. We found that SMOS consistently retrieves a drier soil moisture than the in situ network average. The updated SMOS soil texture map has a much finer spatial resolution and therefore a much higher spatial variability. However, the impact of the new soils map on the SMOS dry bias in the SFIR cannot be directly observed by simply comparing retrievals processed using the differing soil texture maps due to the presence of numerous other attributes that affect the retrieval process. We will begin our assessment of the impact of the new soil texture map on retrieved soil moisture by completing a model sensitivity analysis. We will determine how strongly soil texture affects soil moisture retrieval for the two types of vegetation, corn and soybean, that dominate the SFIR. We will also investigate the effect of spatial scale on soil moisture retrieval, by comparing retrieved soil moisture using a satellite footprint average clay content versus averaging retrievals made a smaller spatial scales. Additionally, we will compare the 'old' SMOS textural map and the 'new' SMOS/SMAP map to known clay content from USDA-NRCS gridded data sets over the SFIR. Finally, we will compare SMAP soil moisture to the SFIR network and determine if any bias exists.