Validation of remotely-sensed soil moisture observations for bare soil at 1.4 GHz:A quantitative approach through radiative transfer models to characterize abrupt transitions caused by a ponding event in an agricultural field,modifications to radiative transfer models,and a mobile groundbased system

摘要

Soil moisture controls the physical processes that exchange mass and energy between the atmosphere and the land surface in the hydrologic cycle. Improved observations of soil moisture may lead to dramatic improvements in weather forecasting, seasonal climateprediction, and our understanding of the physical, chemical and biological processes that occur within the soil. Recent advances in remote sensing have shown that microwave radiometry is a suitable approach to retrieve soil moisture. However, the quantitativeaspects of remotely-sensed soil moisture observations are not well-known, and validation of remotely-sensed measurements is an important challenge. In this dissertation, we describe efforts made at Iowa State University to establish the framework needed for thevalidation of remotely-sensed soil moisture observations. In the process of developing this framework, we engineered new tools that can be used by both our research group and the wider remote sensing community, and we discovered new science. The first tool is a direct-sampling digital L-band radiometer system. This radiometer system is the world\u27s first truly mobile ground-based system. The other tools are radiative transfer models that have been modified in order to be applied to the most general remote sensing situations. An incoherent radiative transfer model was modified to include the contributions of a semi-infinite layer, and a coherent radiative transfer model was modified to account for abrupt transitions in the electrical properties of a medium. The models were verified against each other and the code was written in a user-friendly format. We demonstrated the use of these tools in determining the effect of the transient ponding of water in an agricultural field on the remote sensing signal. We found that ponding was responsible for a 40 K change in the L-band horizontally-polarized brightness temperature. We were able to model this change with both modified coherent and incoherent radiative transfer models. Finally we gave an example of how these tools could be used to quantitatively compare remote sensing observations with models.