A coastally improved global dataset of wet tropospheric corrections for satellite altimetry

The accuracy of satellite radar altimetry (RA) is known to deteriorate towards the coastal regions due to several reasons, amongst which the improper account for the wet path delay (WPD) can be pointed out. The most accurate WPDs for RA are derived from the on-board microwave radiometer (MWR) radiance measurements, acquired simultaneously as the altimeter ranges. In the coastal zone, however, the signal coming from the surrounding land contaminates these measurements and the water vapour retrieval from the MWR fails. As meteorological models do not handle coastal atmospheric variability correctly yet, the altimeter measurements are rejected whenever MWR observations are absent or invalid. The need to solve this RA issue in the coastal zone, simultaneously responding to the growing demand for data in these regions, motivated the development of the GNSS (Global Navigation Satellite System) derived Path Delay (GPD) algorithm. GPD combines WPD from several sources through objective analysis (OA) to estimate the WPD or the corresponding RA correction accounting for this effect, the wet tropospheric correction (WTC), for all along-track altimeter points for which this correction has been set as invalid or is not defined. The current GPD version (GPD Plus, GPD C) uses as data sources WPD from coastal and island GNSS stations, from satellites carrying microwave radiometers, and from valid on-board MWR measurements. GPD C has been tuned to be applied to all, past and operational, RA missions, with or without an on-board MWR. The long-term stability of the WTC dataset is ensured by its inter-calibration with respect to the Special Sensor Microwave Imager (SSM/I) and SSM/I Sounder (SSMIS). The dataset is available for the TOPEX/Poseidon (T/P); Jason-1 and Jason-2 (NASA and CNES); Jason-3 (NASA and EUMETSAT); ERS-1, ERS-2, Envisat and CryoSat-2 (ESA); SARAL/AltiKa (ISRO and CNES); and GFO (US Navy) RA missions. The GPD C WTC for Sentinel-3 (ESA and EUMET-SAT) shall be released soon. The present paper describes the GPD C database and its assessment through statistical analyses of sea level anomaly (SLA) datasets, calculated with GPD C, the ECMWF Reanalysis Interim (ERA-Interim) model or MWR-derived WTCs. Global results, as well as results for three regions (the North American and European coasts and the Indonesia region), are presented for ESA's recent Envisat Full Mission Reprocessing (FMR) V3.0. Global results show that the GPD C WTC leads to a reduction in the SLA variance of 1-2 cm(2) in the coastal zones, when used instead of the ERA WTC, which is one of the WTCs available in these products and can be adopted when the MWR-derived WTC is absent or invalid. The improvement of the GPD C WTC over the ERA WTC is maximal over the tropical oceans, particularly in the Pacific Ocean, showing that the model-derived WTC is not able to capture the full variability in the WPD field yet. The statistical assessment of GPD C for the North American coast shows a reduction in SLA variance, when compared to the use of the ERA-derived WTC, of 1.2 cm(2), on average, for the whole range of distances from the coast considered (0-200 km). Similar results are obtained for the European coasts. For the Indonesia region, the use of the GPD C WTC instead of that from ERA leads to an improvement, on average, on the order of 2.2 cm(2) for distances from the coast of up to 100 km. Similar results have been obtained for the remaining missions, particularly for those from ESA. Additionally, GPD+ recovers the WTC for a significant number of along-track altimeter points with missing or invalid MWR-derived WTCs, due to land, rain and ice contamination and instrument malfunctioning, which otherwise would be rejected. Consequently, the GPD+ database has been chosen as the reference WTC in the Sea Level Climate Change Initiative (CCI) products; GPD+ has also been adopted as the reference in CryoSat-2 Level-2 Geophysical Ocean Products (GOP). Strategies to further improve the methodology, therefore enhancing the quality of the database, are also discussed. The GPD+ dataset is archived on the home page of the Satellite Altimetry Group, University of Porto, publicly available at the repository https://doi.org/10.23831/FCUP_UPORTO_GPDPlus_v1.0 (Fernandes et al., 2019).