High Spatial Resolution Soil Moisture Mapping Using a Lobe Differencing Correlation Radiometer on a Small Unmanned Aerial System

Abstract

A persistent challenge in the measurement of soil moisture from satellites stems from the inherently low spatial resolution using active or passive system. The lobe differencing correlation radiometer (LDCR) on a small unmanned aerial system (sUAS) is shown to provide a capability to measure soil moisture at high spatial resolution for a range of scientific and operational purposes. Flight tests of LDCR on a fixed wing sUAS were performed at the Canton, Oklahoma Soilscape site in September 2015, and Irrigation Research Foundation (IRF) in Yuma, Colorado, in June 2016. The LDCR design and performance are discussed, and the calibration using both preflight lab test data and in-flight data over a calm pond was performed to calibrate the radiometer. Radio frequency interference (RFI) from the sUAS platform was observed and mitigated. The LDCR sampling processes are detailed and an implementation of the $\tau -\omega $ vegetation correction model along with a semiempirical surface roughness correction model incorporating a full-domain soil moisture mapping algorithm is presented. The algorithm uses a weakly nonlinear observation operator suitable for irregular sUAS flight trajectories that maps volumetric soil moisture (VSM) on a user-defined product grid from the sUAS sampling grid. Using LDCR radiometric and thermal measurements, along with Landsat-based vegetation water content (VWC) and soil texture information, soil moisture was mapped at decameter spatial resolution. The retrieved VSM data are favorably compared with in situ VSM measurements and irrigation records. A method for determining LDCR VSM estimation errors is developed to quantify the mapping algorithm accuracy and assess the impact of error sources.