Hydrologic Model-based Disaggregation Algorithm
Soil moisture estimation using microwave remote sensing data is limited by several factors, with one of the most important being the large (> 25 x 25 km) footprints of microwave radiometers. This is adequate for climate modeling but inadequate for many other applications such as agriculture or initializing mesoscale weather models. We believe that there is much valuable ancillary information that can be used to disaggregate microwave brightness temperatures or soil moisture estimates. Toward this end, a simple algorithm has been developed to produce a disaggregated soil moisture field that matches a surface hydrology model’s moisture pattern in a relative sense, and retains the footprint-mean brightness temperature from an aircraft or satellite sensor. The assumption in developing this algorithm is that the model’s spatial pattern, being controlled by rainfall, soil and vegetation properties, and topography, is relative accurate, but the overall mean moisture may be biased with respect to observations due to model drift caused by inaccurate simulation of infiltration, runoff and evapotranspiration. However, soil moisture derived from satellite microwave data are not subject to such drift and should be more accurate in a mean sense over the large sensor footprint.
An example of an application of the disaggregation method is compared with DisaggNet in the figure at right. DisaggNet results are very similar to hydrologic model-derived (SHEELS) soil moisture, but there is a slight bias with respect to soil moisture retrieved from remotely sensed brightness temperature observations with the Electronically Steered Thinned Array Radiometer (ESTAR). Overall, a significant amount of the sub-grid scale variability present in the ESTAR soil moisture image (at 0.8 km resolution) is reconstructed by both disaggregation schemes. The soil moisture spatial pattern is closely related to the pattern of soil types within the watershed. Information at the model scale (0.8 km) is only known to the schemes in the form of rainfall inputs, soils, and topography; ESTAR emissivity data are averaged to 12.8 km before being input to the schemes.
|Fig. 1. Comparison of 0-10 cm soil moisture from SHEELS, DisaggNet, the model-based disaggregation scheme, and ESTAR. Simulated remote sensing inputs for DisaggNet and the model-based disaggregation scheme are at 12.8 km resolution (this can be seen in the large ‘blocks’ in the left panels).
This research was supported by NASA through grant no. NCCW-0084 to Alabama A&M University , Center for Hydrology, Soil Climatology and Remote Sensing. This work was conducted in collaboration with Dr. Marius Schamschula, Alabama A&M University .