Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 6/1/2010
Publication Date: 9/30/2010
Citation: Crow, W.T., Miralles, D., Cosh, M.H. 2010. Evaluating remotely-sensed soil moisture retrievals using triple collocation techniques [abstract]. 2010 IAHS Remote Sensing and Hydrology Symposium. p. 79. Interpretive Summary:
Technical Abstract: The validation is footprint-scale (~40 km) surface soil moisture retrievals from space is complicated by a lack of ground-based soil moisture instrumentation and challenges associated with up-scaling point-scale measurements from such instrumentation. Recent work has demonstrated the potential of enhancing soil moisture validation activities through the application of so-called “triple collocation” techniques which inter-compare three (or more) interdependent soil moisture products in order to acquire the root-mean-square error of each product individually. Such techniques have been widely applied in oceanography as a tool for evaluating remotely-sensed wave height and wind speed retrievals in cases where surface-based instrument is not available or spatially inadequate. This presentation will describe recent application of triple collocation techniques to the soil moisture remote sensing evaluation/validation problem – with a special emphasis an verifying the reliability of triple collocation techniques over areas of very dense soil moisture ground instrumentation. Results will cover the application of triple collocation to the triplet of soil moisture acquired from: 1) passive microwave (AMSR-E) soil moisture retrievals, 2) spatially sparse (one observation per footprint) ground-based instrumentation and 3) a land surface model. As well as a more ambitious case - based on 1) passive microwave (AMSR-E) soil moisture retrievals, 2) scatterometer (ERS) soil moisture retrievals, and 3) a land surface model - where no ground-based soil moisture instrumentation is utilized. Overall, verification results suggest that triple collocation techniques represent a valuable technique for leveraging and/or supplementing limited ground-based resources for validation of remotely-sensed surface soil moisture retrievals.