Retrieval of Subsurface Soil Moisture Profiles from L-Band and P-Band Reflectometry

The use of signals-of-opportunity (SoOp) is becoming one of the mainstream methods for retrieving geophysical properties for environmental studies and operational applications. The pervasive nature of SoOp transmitters and that the rest of the observational hardware system consists principally of a receiver, make this measurement approach quite attractive from the perspective of cost and practicality. Examples of SoOp transmitters include the global navigation satellite system (GNSS) at L-band and the Mobile User Objective System (MUOS) at P-band. NASA has recently invested in an Earth Venture Mission for the use of GNSS reflectometry (the CYGNSS small-sat constellation mission) and MUOS reflectometry (SNOoPI cubesat mission) for various Earth science applications. Although the reflectometry observational systems have several attractive features and have been shown to produce reliable and successful observations in the form of delay-Doppler maps (DDMs), retrieval of land-based variables, such as surface and subsurface profile soil moisture from these maps is still a subject of much ongoing research. There are a number of challenges in the retrieval of soil moisture, including absolute calibration, resolving spatial ambiguities of scattering and reflection points, discerning coherent vs. incoherent contributions, representative and accurate forward scattering models, and accurate inverse scattering algorithms. In this paper, we tackle several of these challenges, and propose a retrieval algorithm for surface-to-root-zone profiles of soil moisture (RZSM). The algorithm uses reflectometry signals at L-band, P-band, and their combinations to obtain accurate estimates of RZSM. First, we use an RZSM profile scattering model that contains both coherent and incoherent scattering contributions. There are multiple options to solve this problem, and our group has previously developed both numerical and approximate analytical methods for this purpose [1]–[3]. Second, we make a correspondence between polarimetric scattering cross sections, reflectivities, and circularly polarized DDM observations so that the model predictions can be related to reflectometry data [4]. The next step is to solve the inverse problem for surface soil moisture and RZSM using single frequencies (L- and P-band) and their combinations if available within a reasonable observation time window. To accomplish the latter task, we use a powerful inversion method recently developed in our group [5], which is a hybrid of global and local optimization methods. We perform a thorough sensitivity analysis to investigate the utility of either frequency alone and in combination for inverting soil moisture at surface and at the root zone. We then show the application of this method to actual SoOp data to the extent available. Recommendations are made for the combined use of GNSS and MUOS for retrieving RZSM. The work reported here focuses on bare surfaces. Vegetated landscapes can be treated similarly, and will be the subject of our follow-on work.