Investigating the Sensitivity of Spaceborne GNSS-R Measurements to Ocean Surface Winds and Rain

Abstract

Earth remote sensing using reflected Global Navigation Satellite System (GNSS) signals is an emerging trend, especially for ocean surface wind measurements. GNSS-Reflectometry (GNSS-R) measurements of ocean surface scattering cross section are directly related to the surface roughness at scale sizes ranging from small capillary waves to long gravity waves. These roughness scales are predominantly due to swell, surface winds and other meteorological phenomena such as rain. In this study we are interested in understanding and characterizing the impact of these phenomena on GNSS-R signals in order to develop a better understanding of the geophysical parameters retrieved from these measurements. In the first part of this work, we look at GNSS-R measurements made by the NASA Cyclone Global Navigation Satellite System (CYGNSS) for developing an effective wind retrieval model function for GNSS-R measurements. In a fully developed sea state, the wind field has a constant speed and direction. In this case, a single Fully Developed Seas (FDS) Geophysical Model Function (GMF) is constructed which relates the scattering cross-section to the near surface wind speed. However, the sea age and fetch length conditions inside a hurricane are in general not consistent with a fully developed sea state. Therefore, a separate empirical Young Sea Limited Fetch (YSLF) GMF is developed to represent the conditions inside a hurricane. Also, the degree of under development of the seas is not constant inside hurricanes and conditions vary significantly with azimuthal location within the hurricane due to changes in the relative alignment of the storms forward motion and its cyclonic rotation. The azimuthal dependence of the scattering cross-section is modelled and a modified azimuthal YSLF GMF is constructed using measurements by CYGNSS over 19 hurricanes in 2017 and 2018. Next, we study the impact of rain on CYGNSS measurements. At L-band rain has a negligible impact on the transmitted signal in terms of path attenuation. However, there are other effects due to rain, such as changes in surface roughness and rain induced local winds, which can significantly alter the measurements. In this part of the study we propose a 3-fold rain model for GNSS-R signals which accounts for: 1) attenuation; 2) surface effects of rain; and 3) rain induced local winds. The attenuation model suggests a total of 96% or greater transmissivity at L-Band up to 30mm/hr of rain. A perturbation model is used to characterize the other two rain effects. It suggests that rain is accompanied by an overall reduction in the scattering cross-section of the ocean surface and, most importantly, this effect is observed only up to 15 m/s of surface winds, beyond which the gravity capillary waves dominate the scattering in the quasi-specular direction. This work binds together several rain-related phenomena and enhances our overall understanding of rain effects on GNSS-R measurements. Finally, one of the important objectives for the CYGNSS mission is to provide high quality global scale GNSS-R measurements that can reliably be used for ocean science applications. In this part of the work we develop a Neural Network based quality control filter for automated outlier detection for CYGNSS retrieved winds. The primary merit of the proposed Machine Learning (ML) filter is its ability to better account for interactions between the individual engineering, instrument and measurement conditions than can separate threshold quality flags for each one.