The use of satellite-derived heterogeneous surface soil moisture for numerical weather prediction

This work's general scientific objective is to develop and apply a coupled satellite-model data assimilation approach to observe heterogeneous soil moisture effects on the mesoscale for use in understanding and quantifying the processes that have an impact on the preconvective mesoscale environment and affect initial convective cloud development related to surface-forced circulations. Several satellite-based methods are used to investigate the feasibility of retrieving surface wetness information. Passive microwave satellite data are used as a subjective indicator of surface wetness, while a quantitative satellite data assimilation method is developed to assimilate observational infrared heating rates into an atmospheric mesoscale model to retrieve model soil moisture. The data assimilation method employs a prognostic soil model with explicit bare soil and vegetation surface components.;Atmospheric-corrected microwave surface emittance results are shown to enhance the use of the microwave datasets for determining land surface characteristics, especially in regards to analysis of the data's frequency dependencies. Several problems that affect the use of the microwave brightness temperature data were examined, including natural characteristics of the spatial and temporal variability of the microwave background signature, and sub-field of view effects.;The microwave surface emittance was found to be sensitive to numerous rain events captured in the dataset. The relationship of the microwave surface emittance to a vegetation index is shown to be highly variable for all but the sparsest vegetation amounts when analyzed at high spatial resolutions. The complexity of the microwave surface emittance versus vegetation relationship limits any future quantitative use of the microwave surface emittance for direct soil moisture retrieval; however, the potential application of the microwave surface emittance for flood monitoring purposes and trafficability indices is high for non-forested regions.;In a case study, the data assimilation method is successful at retrieving realistic representations of the heterogeneous soil moisture. However, limitations are found regarding the ability to retrieve extreme dry or wet events using the current model surface parameterization. This has implications on the ability of the retrieved soil moisture values to affect the atmospheric model's forecast. Instrument noise is not found to be a major factor in the data assimilation method's performance.