A Bayesian approach for Soil Moisture and Optical Depth Retrieval: evaluation on Aquarius/SAC-D observations
13th Specialist Meeting on Microwave Radiometry and Remote Sensing of the Environment (MicroRad), 2014 , vol., no., pp.1,4, 24-27 March 2014 doi: 10.1109/MicroRad.2014.6878896
Abstract: In this paper, the methodology of a novel Bayesian algorithm to retrieve soil moisture (sm) and optical depth (τ ) from passive microwave data is introduced. As a major advantage of this approach, prior knowledge for sm and τ can be directly included within the Bayesian inference framework in order to improve the retrieval. In order to test the methodology presented and contrast its results with existing passive-microwave-based sm retrievals, several algorithms were implemented using Aquarius/SAC-D observations. Algorithms computed include: H- and V-pol Single Channel Algorithm, (SCAH and SCAV respectively) and Microwave Polarization Difference Algorithm (MPDA). Currently available L-band sm products were also incorporated to the study: SCAH for Aquarius (developed by the United States Department of Agriculture) and SMOS Level-2 product (European Space Agency). The analysis was carried out over Pampas Plains, Argentina on an specific date in 2012.Global Land Data Assimilation System sm product was used as benchmark for performance analysis. The Bayesian approach introduced here resulted in the lowest unbiased root mean square error and bias. The main drawback of this approach is that it is highly time consuming, thus making it not suitable for the development of a global near-real time soil moisture product. Efforts were made towards lowering time consumption through Markov Chain Monte Carlo method, although it is still a limiting factor. Nevertheless, the proposed algorithm can provide a framework for evaluation of sm products over limited areas or short time periods.
Index Terms—Aquarius; soil moisture; Bayesian inference; Markov Chain Monte Carlo.
keywords: {Bayes methods; hydrological techniques; moisture; soil; Aquarius/SAC-D observations; Aquarius/SAC-D soil moisture product; Argentina; Bayesian algorithm performance; H-pol Single Channel Algorithm; Pampas Plains; SCAV ubRMSE; SMOS Level-2 optical depth; SMOS Level-2 soil moisture; USDA SCA; V-pol Single Channel Algorithm; benchmark products; microwave polarization difference algorithm; performance metrics; soil moisture retrieval algorithms; Bayes methods; Measurement; Optical polarization; Optical scattering; Optical sensors; Soil moisture; Uncertainty; Aquarius; Bayesian inference; Markov Chain Monte Carlo; soil moisture}