Tropospheric delays in ground-based GNSS Multipath Reflectometry – experimental evidence from coastal sites

Recent studies have demonstrated the utility of ground based GNSS Multipath Reflectometry (GNSS-MR) for sea level studies. Typical root-mean-square (RMS) differences of GNSS-MR derived sea level time series with respect to nearby tide gauges are on the order of 6 – 40 cm, sufficiently accurate to estimate tidal and secular sea level variations but are possibly biased due to delay of the signal through the troposphere. In this study we investigate the tropospheric effect from more than 20 GNSS coastal sites located from several meters up to 280 m above sea level. We find a bias in the estimated heights that is elevation and height dependent and can reach orders of 1 m for a 90 m site. Without correcting for tropospheric delay we find that GNSS-MR estimated tidal coefficients will be smaller than their true amplitudes by around 2% while phases seem unaffected. Correcting for the tropospheric delay also improves levelling results as a function of reflector height. Correcting for the tropospheric delay in GNSS-MR for sea level studies is therefore highly recommended for all sites no matter the height of the antenna above the sea surface as it manifests as a scale error.