Coherency-based Inland Water Body Detection with Spire Grazing Angle GNSS-R Data using the Strongest EPL Channel

Mapping inland water bodies (IWB) is crucial for understanding local hydrology, ecosystems, and water management, and satellite-based GNSS-R offers good spatiotemporal resolution in remote areas for IWB mapping. This presentation compares the use of the strongest return among the early, prompt, and late (EPL) channels with that of the prompt-only channel for coherence-based IWB detection. In this study we use the 2021 Spire Global grazing angle GNSS-R data set over high latitude regions.

The GNSS-R technique fundamentally views the reflected GNSS signal as a signal of opportunity, making the reflection locations and surface characteristics initially unknown. An open loop model is typically adopted for tracking the reflected signal, where the expected Doppler and code delay of the reflected GNSS signal is calculated by finding the minimum path length between the positions of the GNSS satellite, the specular reflection point over Earth surface, and the Low-Earth-Orbit satellite using either a local Digital Elevation Model (DEM) or simply the WGS84 ellipsoid. The predicted Doppler and delay are applied to the reflected signal intermediate frequency (IF) data to generate EPL correlators. The prompt correlator channel is traditionally used to calculate the received signal SNR and carrier phases. SNR can be used to differentiate between a dominantly coherent or noncoherent reflection, which plays a crucial role in Inland Water Body (IWB) detection as signal coherency is associated with the presence of water. However, if the predicted delay and Doppler is incorrect, the prompt correlator may not always be the highest correlator in the EPL structure (Figure 1). Preliminary results show that SNR difference between the EPL-highest and the prompt-only channel can be as great as 34 v/v, which can result in missed coherent returns and, in turn, a failure to detect IWBs using GNSS-R.