Methodological improvements for deriving long-term time series of coastal sea level by GNSS-R

While sea level variations at coastal sites can be derived from Signal-to-Noise Ratio (SNR) measurements in GNSS-R, the presence of noise, signal interruptions and unmodelled geophysical contributions still corrupt the quality of the estimates. We propose improvements in the treatment of raw SNR records for obtaining much precise sea level. We implement correction of the atmospheric delays, as well as filtering of loading displacements for producing sea level time series over several years. We also propose empirical corrections on a priori fitting parameters to absorb systematic effects from satellite elevation that spoil the sea level time series. Water height adjustment from periodogram of the windowed SNR signal requires parameters that have been fixed so far, e.g. the width of the analyzing window - or equivalently the number of SNR periods used in the adjustment. In particular, tuning of this latter critical parameter is made versus the receiving antenna height.