Observation of Soil Moisture Vertical Profiles from GNSS Signal Multi-Path Interferences

Abstract

A new remote sensing approach is developed to estimate soil moisture vertical profile utilizing the Global Navigational Satellite System (GNSS) signal of opportunity. The direct GNSS signal and its reflected counterpart from the air/soil interface interfere with each other, and such multi-path interference generates unique features in the angular pattern of the observed signal to noise ratio (SNR). The SNR angular patterns are sensitive to the variation of the near-surface soil moisture vertical profile and the height of the antenna phase center above the air/soil interface. In this paper, we developed a rigorous mathematical and physical basis to predict the signal received by a GNSS antenna placed above ground at a certain height. The vertical profile of soil moisture is parameterized with three parameters based on the physical solution to the Richards' equation for unsaturated flow in soils. The Fresnel reflection coefficients from the soil with moisture profiles are derived from Maxwell's equations on flat multi-layered media. The Mironov soil permittivity model is used to link the soil permittivity at L-band to its moisture content and clay fraction. The polarization coupling of the GNSS antenna is carefully incorporated in the interference model by characterizing the antenna with both the gain pattern and the phase pattern for both the right-hand circular-polarized (RHCP) and the left-hand circular polarized (LHCP) signals. Besides the rigorous forward physical model, a least-mean-square-error (LMSE) based retrieval algorithm is developed to retrieve the soil moisture profile by matching the received GNSS SNR angular patterns with model predictions. Both synthesized data and measured data are used to test the retrieval algorithm performance.