background
where does data come from?
Sentinel-3
Lessons were learnt from the CryoSat-2 and Jason-2 missions and a new L0/L1 Delay Doppler Ground Processor Prototype (GPP) was defined and implemented. This GPP was developed, tested and validated by isardSAT Cat, based on the algorithms defined by CLS (see ADAS). LAter, this definition, together with the lessons learned from the GPP implementation, were used to implement the Sentinel-3 Instrument Processing Facility by ACRI ST and CLS companies (see S3 IPF DPM).
CryoSat
The first Delay-Doppler Processor used with Altimeter data from a European Earth Observation satellite was a Ground Processor Prototype (GPP) developed by UCL for the CryoSat mission. This GPP was used for validating CryoSat products and the end-to-end chain performances (Wingham, et al., 2005). Later, this GPP was operation adapted by Aresys, becoming the SAR and SARin chains of the Instrument Processing Facility 1 of the CryoSat-2 mission (Cullen et al.).
In CryoSat-2, new intermediate lower-level products such as Full Bit Rate (FBR) and calibrated FBR (C-FBR) products, equivalent the L1A products in Sentinel-3 and Sentinel-6 , were introduced with the main purpose of testing, debugging and internal verification of the processor algorithms. FBR products and C-FBR products were not publicly available; hence a limited number of studies were carried out. The main goal of the CryoSat-2 L1 SAR Delay-Doppler processor was to produce L1B products to be injected into the L2 processor chain.
In addition to that, some calibration and validation activities also needed customised L1 algorithms different from the baseline algorithms provided in the L1B GPP. For instance, the calibration activities developed by isardSAT implemented a new algorithm to steer the beam to the precise transponder position in order to improve the absolute range calibration method.
Sentinel-6
the new Sentinel-6/Jason-CS mission will provide continuity for previous Jason conventional altimetry missions, but with a new SAR Altimeter (Poseidon-4) on-board. The main differences between the Sentinel-3 altimeter (SRAL) and the Sentinel-6 altimeter (Poseidon-4) are:
- Improved digital and radio frequency hardware.
- Open burst Ku-band pulse transmission (interleaved mode): This performs a near continuous transmission of Ku-band pulses. It will allow simultaneous processing of the measurements to obtain High Resolution along-track (HR or SAR) and Low Resolution along-track (LR or LRM) data.
- As with previous satellite RAs, the P-4 transmits C-band pulses in order to retrieve a correction for ionospheric path delay.
- Implementation of On-board "Range Migration Correction" (RMC) processing in order to reduce the amount of data to download to ground.
The Ground Processor Prototype Processor of the Sentinel-6 mission altimeter, Poseidon-4, is being developed by isardSAT under an ESTEC/ESA contract that started in 2011 and will last until the end of the satellite commissioning phase. This processor includes a lot of new features partly thanks to the experience gained with the CryoSat-2 data, and partly thanks to the many studies carried out during the course of the project. It should be noted that this development has been taking advantage of simulated data produced by the ESTEC Sentinel-6/Jason-CS mission performance simulator. This simulated data has proven to be extremely useful, interesting and crucial for the understanding of the details of Delay-Doppler Processing, which is still relatively new in the worldwide altimetry experience. The fact that we had to face a new configuration has also invited us to review some of the theory, coming up with new methodologies in different aspects of the processing. Some of these improvements and new methods are shown in the sample test data sets and corresponding documentation (UDD, isardSAT, 2014), delivered in October 2014. Another major improvement, which is a new and revolutionary concept for the stack building was presented at the OSTST that year (ACDC, Ray et al. 2014). The application of this methodology will impact both the DDP and the retracking method, by not only significantly improving the performances, particularly at low level SWH, but also reducing the run-time. This new method has also been chosen as a demonstration of the usefulness of using lower level of data, in this case starting from the L1A and producing the L1B-S.
The Sentinel-6/Jason-CS GPP has an important configuration capability: The ability to study the different possible solutions and how they perform under different circumstances. It is this configuration capability that we will try to reproduce here for Sentinel-3.