Optimization of GPS Interferometric Reflectometry for Remote Sensing

GPS Interferometric Reflectometry (GPS-IR), a passive microwave remote sensing technique utilizing GPS signal as a source of opportunity, characterizes the Earth\u27s surface through a bistatic radar configuration. The key idea of GPS-IR is utilizing a ground-based antenna to coherently receive the direct, or line-of-sight (LOS), signal and the Earth\u27s surface reflected signal simultaneously. The direct and reflected signals create an interference pattern of the Signal-to-Noise Ratio (SNR), which contains the information about the Earth\u27s surface environment. GPS-IR has proven its utility in a variety of environmental remote sensing applications, including the measurements of near-surface soil moisture, coastal sea level, snow depth and snow water equivalent, and vegetation biophysical parameters. A major approach of the GPS-IR technique is using the SNR data provided by the global network of the geodetic GPS stations deployed for tectonic and surveying applications. The geodetic GPS networks provide wide spatial coverage and have no additional cost for this capability expansion. However, the geodetic GPS instruments have intrinsic limitations: the geodetic-quality GPS antennas are designed to suppress the reflected signals, which is counter to the requirement of GPS-IR. As a result, it is desirable to refine and optimize the instrument and realize the full potential of the GPS-IR technique. This dissertation first analyzes the signal characteristics of four available polarizations of the GPS signal, and then discusses how these characteristics are related to and can be used for remote sensing applications of GPS-IR. Two types of antennas, a half-wavelength dipole antenna and a patch antenna, are proposed and fabricated to utilize the desired polarizations. Four field experiments are conducted to assess the feasibility of the design criteria and the performance of the proposed antennas. Three experiments are focused on snow depth measurement. The Table Mountain experiment data shows a more distinct interference pattern of SNR and yields a more precise snow depth retrieval. The Marshall experiment data reveals the effect of the underlying soil medium on snow depth retrievals, which benefits from the improved sensitivity of the dipole antenna to snow depth change. The experimental data of the third snow experiment conducted at a mountain-top location shows that a variant surface tilt angle can result in considerable retrieval errors of snow depth. An algorithm utilizing the estimated surface tilt angle is proposed to calibrate the retrieved snow depth. The calibration algorithm significantly improves the accuracy and precision of snow depth retrievals. The last experiment provides an opportunity to evaluate the dipole antenna as applied to the measurements of vegetation biophysical parameters and near-surface soil moisture. The normalized SNR amplitude shows a negatively linear relationship with in situ measurements of vegetation water content over a range of 0-6 kg/m2, which is much greater than the range from the geodetic antenna data (0-1 kg/m2). The normalized SNR amplitude also shows a positive linear relationship with the near-surface soil moisture measurements, indicating its potential as a soil moisture sensor