Error characterization methods for surface soil moisture products from remote sensing

To support the operational use of Synthetic Aperture Radar (SAR) earth observation systems, the European Space Agency (ESA) is developing Sentinel-1 radar satellites operating in C-band. Much like its SAR predecessors (Earth Resource Satellite, ENVISAT, and RADARSAT), the Sentinel-1 will operate at a medium spatial resolution (ranging from 5 to 40 m), but with a greatly improved revisit period, especially over Europe ( 2 days). Given the planned high temporal sampling and the operational configuration Sentinel-1 is expected to be beneficial for operational monitoring of dynamic processes in hydrology and phenology. The benefit of a C-band SAR monitoring service in hydrology has already been demonstrated within the scope of the Soil Moisture for Hydrometeorologic Applications (SHARE) project using data from the Global Mode (GM) of the Advanced Synthetic Aperture Radar (ASAR). To fully exploit the potential of the SAR soil moisture products, well characterized error needs to be provided with the products. Understanding errors of remotely sensed surface soil moisture (SSM) datasets was indispensible for their application in models, for extractions of blended SSM products, as well as for their usage in evaluation of other soil moisture datasets. This thesis has several objectives. First, it provides the basics and state of the art methods for evaluating measures of SSM, including both the standard (e.g. Root Mean Square Error, Correlation coefficient) and the advanced (e.g. Error propagation, Triple collocation) evaluation measures. A summary of applications of soil moisture datasets is presented and evaluation measures are suggested for each application according to its requirement on the dataset quality. The evaluation of the Advanced Synthetic Aperture Radar (ASAR) Global Mode (GM) SSM using the standard and advanced evaluation measures comprises a second objective of the work. To achieve the second objective, the data from the Australian Water Assessment System (AWRA-L) hydrological model, OzNET in-situ stations, and several other coarse resolution data sources were used. The results are combined to provide an exhaustive estimate of all qualities of the ASAR GM SSM product. The third objective is to provide guidance on appropriate evaluation methodology applicable to any SSM product. For this purpose the results of the ASAR GM evaluation analyzed are discussed from a general perspective and restructured to answer scientific questions identified in the introductory part of the thesis. These include:  Can we apply the evaluation requirements from comparable missions such as SMOS and SMAP to ASAR GM SSM?  How does spatial resolution influence error estimates?  Is there a single measure to describe the quality of SSM data?  What is the quality and what are the limitations of ASAR GM SSM?  Learning from ASAR GM SSM errors for Sentinel-1 The findings and suggestions originating from the discussion are transferable to other satellite-derived soil moisture data. Of special interest is its transfer to data from the planned Sentinel-1 SAR sensor that shares similar technical characteristics but has an improved retrieval error comparable to the ASAR GM sensor. The operationally available medium resolution soil moisture from Sentinel-1 with a wellcharacterized error is likely to yield benefits for modelling and monitoring of land surface-atmosphere fluxes, crop growth and water balance applications.