A coastal mean sea surface with associated errors in Norway based on new-generation altimetry

Abstract

Abstract The coastal mean sea surface (MSS) has applications within oceanography as well as geodesy. Together with a geoid model, it forms an important component for geodetic mapping of ocean surface currents that are in geostrophic balance. Furthermore, it forms a bridge between open ocean MSS and in situ measurements of mean sea level at or close to land, it contributes to the mapping of the geoid and the marine gravity field, and it is essential for connecting tidal nautical chart datums to physical height systems or global geodetic reference frames. In this study, we determine a coastal MSS with an associated error field for Norway. The MSS is solely based on new-generation altimetry data, i.e., SAR(In) data from Sentinel-3A and CryoSat-2, as well as Ka-band data from SARAL/AltiKa. The data sets partly overlap in time and cover the time period from 2010 to 2017 inclusive. We have chosen these altimeters because they represent evolutions of conventional altimetry, with reduced footprint sizes as a main benefit. This is especially advantageous in the coastal zone, as a smaller footprint reduces the probability of radar pulses being contaminated by energy backscattered from land areas. The satellite missions were harmonized by applying inter-mission biases determined in a regional crossover analysis. Furthermore, in a zone closer to land than 25 km, we have replaced the global ocean tide model with a regional ocean tide model provided by the Norwegian Mapping Authority (NMA). We use an optimal interpolation technique to determine a coastal MSS grid and discuss it in context of the estimated error field. We assess our coastal MSS by comparison to state-of-the-art MSS products along three sections perpendicular to the coast, as well as ellipsoidal mean sea level as observed by an array of permanent tide gauges within the study area. In addition, we assess a higher-resolution version of our MSS in the NMA testbed for vertical datums, by comparison with temporary tide gauges. We find that the coastal MSS outperforms the global MSS models directly at the coast, with standard deviations of differences of ~ 8 cm to the tide gauges, compared to 14–22 cm, obtained with the global MSS models. All MSS models largely agree along three sections perpendicular to the coast, with standard deviations of differences of 2–4 cm. The higher-resolution version of the coastal MSS performs similarly to the coastal MSS in comparison with the temporary tide gauges (standard deviation of differences of ~ 8 cm), but its formal error field also quantifies large uncertainties at the coast and in the fjords, mainly due to the lack of altimetry observations. A trustworthy error field is decisive for the combination of altimetry with other sea-level observations.