P-Band Radar Retrieval of Root-Zone Soil Moisture: Airmoss Methodology and Error Assessment.

P-BAND RADAR RETRIEVAL OF ROOT-ZONE SOIL MOISTURE: AIRMOSS METHODOLOGY AND ERROR ASSESSMENT

Alireza Tabatabaeenejad, Richard H. Chen, and Mahta Moghaddam

Department of Electrical Engineering

University of Southern California

Los Angeles, California, USA

The goal of NASA's Airborne Microwave Observatory of Subcanopy and Subsurface (AirMOSS) has been to improve the estimates of the North American Net Ecosystem Exchange (NEE) by providing high- resolution observations of root-zone soil moisture (RZSM) over nine regions representative of the major North American biomes, estimating the impact of RZSM on regional carbon fluxes, and integrating the measurement-constrained estimates of regional carbon fluxes to the continental scale of North America.

Several empirical and semi-empirical methods exist to estimate surface and root-zone soil moisture from radar data. We have used a fully-analytical approach that integrates a discrete scatterer model with layered-soil scattering models. We present the soil moisture profile as az2 + bz + c, where z is the depth and a, b, and c are the coefficients to be retrieved. The details of our methodology have previously been reported.

The uncertainty in the retrieval results are associated with radar calibration, vegetation and soil parameterizations, and the scattering and inversion models. We are currently performing a comprehensive assessment of retrieval errors over all sites. The behavior and location of each in-situ soil moisture probe have to be studied before the corresponding data are used in the calculation of error. Biases present in the RZSM products should also be removed for each site before the final error calculation for the mission.

We will report an overview of the AirMOSS algorithms and the measures taken over the course of the mission to increase the accuracy of RZSM products. We will then present an overall assessment of the retrieval error from the flights over BERMS, Metolius, MOISST, Tonzi Ranch, and Walnut Gulch. The error will be reported in terms of RMSE over all qualified validation points, all sites, and all dates. We will also report the overall error as a function of ‘threshold depth.'