Soil Moisture Estimation Based on Probabilistic Inversion over Heterogeneous Vegetated Fields Using Airborne PLMR Brightness Temperature
Chunfeng Ma1, Xin Li1,2,3, and Shuguo Wang4,11Key Laboratory of Remote Sensing of Gansu Province, Northwest Institute of Eco-Environment and Resources, Heihe Remote Sensing Experimental Research Station, Chinese Academy of Sciences, Lanzhou, 730000, China 2CAS Center for Excellence in Tibetan Plateau Earth Sciences, Beijing, 100101, China 3University of Chinese Academy of Sciences, Beijing 100049, China 4Jiangsu Normal University, No. 101 Shanghai Road, Xuzhou, P. R. China
Received: 23 Jan 2017 – Accepted for review: 11 Feb 2017 – Discussion started: 13 Feb 2017
Abstract. The L-band radiometer has demonstrated its strong ability of estimating soil moisture (SM) over vegetated surface. However, the present SM products derived from satellite radiometers are hardly directly used in heterogeneous oasis. This may be attributed to the coarse spatial resolution of the satellite-based observations and the defects of the point-estimation algorithms which cannot quantify the uncertainty in SM inversion. This paper presents a SM estimation based on the combination of Bayesian probabilistic inversion with high resolution airborne radiometer observations. The overall objective is to quantify the uncertainty in SM estimation and to provide a desirable SM estimation. Two retrieval strategies (2P and 3P strategy) are performed based on 6-channel Polarimetric L-band Multi-beam Radiometer (PLMR) observations collected during Heihe Watershed Allied Telemetry Experiment Research, and the ground measurements are used to validate the results. The main findings contain: (1) Accurate SM estimates with RMSE and ubRMSE less than 0.03 m3/m3, exceeding that of 0.04 m3/m3 of Soil Moisture Active and Passive (SMAP) and Soil Moisture Ocean Salinity (SMOS) missions target accuracies, are obtained. (2) The uncertainty in SM inversion is quantified with the value less than 0.095 m3/m3. (3) The Bayesian PI improves the simulation performance of forward model via maximum likelihood estimation. (4) The 3P and 2P strategies result in different SM inversion uncertainties. Overall, the present Bayesian PI combining multi-channel L-band observations has obtained a desirable SM estimation and uncertainty quantification, which may offer an insight into the future SM inversion based on passive microwave remote sensing.
Ma, C., Li, X., and Wang, S.: Soil Moisture Estimation Based on Probabilistic Inversion over Heterogeneous Vegetated Fields Using Airborne PLMR Brightness Temperature, Hydrol. Earth Syst. Sci. Discuss., doi:10.5194/hess-2017-34, 2017.