| Soil hydraulic properties (SHPs) play an important role in land surface models, but these properties are very poorly known at present. Because of the need to define SHPs on a global basis for climate models, and at a very high spatial resolution for weather forecasting, it is not feasible to make field measurements of these SHPs everywhere they are needed. We present a method of identifying SHPs from thermal remote sensing data. We determine SHPs with data from 2002, and verify these SHPs by comparing model output and field observations from 2005. We find that the SHPs we determined decrease errors in land surface model predictions of latent heat flux from 100W/m 2 to 65W/m2. For reference, fitting the land surface model to these same latent heat flux measurements only decreases the error to 50W/m2. We also find that these results are insensitive to errors in other model parameters. For example, changing albedo by 0.1 only changes the predicted saturated conductivity (Ks) by 10% and van Genuchten m by 1%. Changing surface roughness by an order of magnitude or minimum stomatal resistance by 40s/m has less effect. In addition, we determine SHPs in eastern Colorado for an area where we expect there to be a spatial gradient in SHPs based on existing soil maps. We find that our predicted Ks varies in a spatially consistent manner with the expected Ks, and that the predicted alpha and m parameters change with a spatial pattern that matches the landscape transitions. |
