Verification of the Ra2/mwr Level 2 Geophysical Reference Processors

The RA2/MWR Level 2 processing includes some algorithms designed specifically to extract geophysical measurements over different types of surfaces (ocean, sea-ice, continental ice, etc.). In order to do this, four specialised retrackers have been developed and are continuously run in parallel (over all surfaces): __ ocean retracker: optimised for ocean surfaces (operating on Ku and S bands), and based on a fit of the waveforms with the Hayne model __ ice1 retracker: optimised for general continental ice sheets. It employs an Offset Center-o f-Gravity parametrisation of the Ku/S band echoes to calculate tracker offsets. This was used for ERS and will ensure measurement continuity. __ ice 2 retracker: optimised for ocean-like (Ku and S) echoes from continental ice-sheet interior, and based on the Brown model, __ sea-ice retracker: optimised to estimate a tracker offset for narrow peaked (sea-ice) echoes. It operates only on Ku band These algorithms were prototyped within CLS, LEGOS and MSSL (referred to as ESLs, or Expert Support Laboratories), and once verified, became reference processors, that were used to validate the Near Real Time and OffLine ground operational processors (referred to as IPF and F-PAC). The aim of this paper is to summarise the main activities performed, since launch, by the different ESLs, and supported by ESA, in order to fully verify the Level 2 algorithms, and to validate the Level 2 products that are delivered to the end-users. 1 OCEAN RETRACKER VERIFICATION The RA2/MWR Level 2 processing includes some algorithms designed specifically to extract geophysical measurements over ocean surfaces. The validation of the ocean retracking of the RA2 waveforms over ocean surfaces, and the optimisation of the ocean retracking performance will be dealt within this section. In the frame of these activities, the outputs of the ocean retracking have been analysed and compared with other parameters such as level 1b parameters or outputs of the ice2 retracking algorithm. A tuning of the ocean retracking algorithm has been then performed to optimise its performance as far as accuracy and CPU time are concerned The results of the internal verification activities, that will be shown in the first next three subsections, were obtained analysing the ocean retracking outputs (epoch, composite sigma and amplitude) for a single real orbit (orbit no. 1815) that contained 24 minutes of consecutive measurements over the Pacific ocean, with waveheights between 1 and 4 meters. The ocean/ice2 reference processor (referred to as LOP) was then further validated by using a complete cycle of Level 2 off-line data coming from the F-PAC. This will be explained in the last subsection.