According to the Working Group on Calibration and Validation (WGCV) of the international Committee on Earth Observation Satellites (CEOS), validation is the process of assessing, by independent means, the quality of the data products derived from the system outputs. For Sentinel-1 verification is only applicable to Level-2 products, as described below.
Level-2 product verification consists of verifying the main quality of the engineering parameters prior to the inversion steps. Ocean Swell Spectra (OSW) verification consists mainly of measuring the performance parameters related to the input imagette or the cross-spectra. Ocean Wind Field (OWI) verification consists of verifying the calibration constant obtained from the transponders against a geophysical calibration constant and deriving the noise equivalent radar cross-section (NESZ). Radial Velocity (RVL) verification is mainly to verify the accuracy of the Doppler anomaly estimation.
The geophysical validation of Sentinel-1 Level-2 products consists of comparing retrieved Level-2 products using a geophysical model function from Sentinel-1 measurements with independent 'equivalent' measurements. This allows assessment of the information content and the quality of the Sentinel-1 Level-2 products.
The geophysical validation of the OSW component consists of characterising the performance of wave spectral parameters derived from the Sentinel-1 SAR with other independent sources (buoys), followed by an estimation of RMS error and bias. The comparison is made possible through the collocation of the buoys with SAR.
OWI validation assesses the radiometric calibration performance for all combinations of mode, swath and polarisation. It also characterises wind retrieval performance as a function of the mode, incidence angle (swath), polarisation and wind conditions:
The RVL component is less mature than OSW and OWI. The experiment performed on ASAR has demonstrated that the operational current extraction requires stringent performance on the accuracy of the Doppler shift anomaly estimation. It has been demonstrated that several biases or trends were present in the ASAR Doppler estimation impeding a straightforward RVL estimation without performing heavy Doppler calibration steps for fixing the biases.