Evaluation of new CryoSat-2 products over the ocean

Abstract

The CryoSat-2 satellite, primarily dedicated to precise monitoring of the Cryosphere, is demonstrating its capability to provide valuable altimetric data also over the ocean. Here we present the results of a global assessment and validation of the new Geophysical Ocean Product (GOP) distributed by the European Space Agency (ESA) since April 2014, focusing on the sea surface height anomaly (SSHA), the significant wave height (SWH), and the wind speed. Our assessment involves only Low Resolution Mode (LRM) and Pseudo LRM (PLRM) data, since full SAR processing is not already operationally implemented in the GOP. The global assessment is conducted on the basis of measurement noise and along-track spectral and crossover analysis, whereas the validation is performed against a variety of in situ observations such as tide gauges, buoys and Argo floats as well as data from the WaveWatch III (WWIII) model. The performance of the GOP is compared to that of Jason-2 and CryoSat-2 data from the Radar Altimeter Database System (RADS). The mean value of the 20-Hz SSHA noise at 2m SWH is 6.3cm for LRM and 10.2cm for PLRM, and the standard deviation of the crossovers is ~5.4cm. The mean 20-HZ SWH noise over the global oceans is 49.4cm and 69.8cm, for LRM and PLRM respectively. CryoSat-2 and Jason-2 show almost identical performance when SSHAs are validated against tide gauges, with a median correlation and root mean square difference (RMSD) of 0.78 and 7.1cm for the GOP, 0.76 and 7.3cm for Jason-2, and 0.79 and 7.8cm for CryoSat-2 from RADS. The median correlation with Argo-derived steric heights is 0.68 for the GOP, 0.74 for Jason-2, and 0.67 for CryoSat-2 from RADS. However, the correlation shows a strong latitudinal dependence, with higher values at low latitudes (median value larger than 0.80 in the 10°S-10°N band). The median RMSD between the SSHAs and steric heights is 5.3cm for the GOP, 4.6cm for Jason-2, and 5.1cm for CryoSat-2 from RADS. The GOP and Jason-2 show also identical performance when SWHs are compared to buoy data, with a slope and RMS error of 0.98 and 15cm for GOP, 0.97 and 16cm for Jason-2, and 1.05 and 17cm for CryoSat-2 from RADS. On the other hand, the GOP wind speed exhibits a bias of about 2m/s relative to both Jason-2 and to buoy data. Differences between the GOP and WWIII SWH are smaller than 20% of the SWH almost everywhere. In summary the GOP products are fit for oceanographic applications.