GPOD SENTINEL-3 SARvatore Software Prototype User Manual

Introduction

The G-POD Service, SARvatore (SAR Versatile Altimetric Toolkit for Ocean Research & Exploitation) for Sentinel-3, is a web platform that provides the capability to process on line and on demand Sentinel-3 SRAL L1A data products in SAR mode until Level-2 geophysical data products.

The service is based on Sentinel-3 SARvatore Processor Prototype, heritage of the CryoSat-2 SARvatore Processor Prototype.

The data products in input to the service are the Sentinel-3 SRAL L1A data products. The manifest of these products are regurarly ingested in the G-POD storage facility so that they are made available in the G-POD data catalogue in order to be processed on line.

The Sentinel-3 SAR Processor has two stages: L1b & L2.

The Processors will make use of the G-POD (Grid-Processing On Demand) distributed computing platform to deliver timely the output data products.

These output data products are generated in standard NetCDF format (using CF Convention), and they are compatible with BRAT (Basic Radar Altimetry Toolbox) and other NetCDF tools.

Using the G-POD graphic interface, it is possible to easily select the geographical area of interest along with the time of interest. After the task submission, it is possible to follow, in real time, the status of the processing task.

spazi

spazio

You find more information about how to use the generic G-POD graphic interface at the G-POD wiki page available here.

The on line SAR service is available here.

List of option entries in the G-POD graphic interface

The processor prototype is versatile in the sense that the users can customize and adapt the processing, according their specific requirements, setting a list of configurable options.

In the G-POD web interface, user can enter easily this list of processing options via a series of drop-down menu. The configurable options are divided according to the processing level they refer to (L1b and L2). Here, you find them listed with a basic description of their significance.

spa

L1b Processing Options:

Option Name	Option Value	Option Description
Data Posting Rate 20Hz/80Hz	- 20 Hz - 80 Hz	User can decide here the posting rate (20 Hz or 80 Hz) of the L2 data published in output
Hamming weighting Window	- Apply only in coastal zone - Yes, apply it - No. do not apply it	User can decide here whether to apply a Hamming weighting window on the SAR burst data, do not apply it, or to apply it only for surface location in coastal area (more info at REF1)
Exact Beam-Forming	- Approximated - Exact	User can decide here whether to operate an exact Beam-Forming or an approximated Beam-Forming (more info at REF1)
FFT Zero-Padding	- Yes, apply Zero-Padding - No, don't apply Zero- Padding	User can decide here whether to operate the Zero-Padding prior to the range FFT (section 4.8 in REF1). Zero-Padding is indicated for coastal zone and sea-ice analysis
Radar Receiving Window Size	- 128 Range Bins - 256 Range Bins	User can select here the size of the radar receiving window: 128 range bins (standard) or 256 range bins (extended). Extended window is indicated for coastal zone analysis
Antenna Pattern Compensation	- NO -YES	User can decide here whether to apply the Antenna Pattern Compensation on Stack Data prior to make the multilooking
Dump SAR Stack Data in output	-NO -YES, with only power -YES, with power and phase	User can command here to post in output SARStack Data Products for each selected pass. Users can select whether to dump just the power of the Stack data or the power and phase. Beware: SAR Stack Data Products are quite bulky (usually 2 GB for 10 minute of data); we recommend to process SAR Stack Data singuarly or max 5 passes at the time: contact G-POD team for massive production of SAR Stack Data

spazio

L2 Processing Options:

Option Name	Option Value	Option Description
Restrict the re-tracking on specific surfaces	- Process all - Process only open sea points - Process only water points	User can decide here whether to process the whole pass, only points in open sea or only water points (points in open sea, costal zone and inland water)
PTR width alphap parameter	- LUT - Constant	User can decide here whether to use a LUT (Look-Up Table) or a constant for PTR (Point Target Response) alphap parameter
SAMOSA Model Generation	- Use SAMOSA2 - Use SAMOSA3 - Use SAMOSA+	User can decide here which SAMOSA generation model to use in the processing (SAMOSA2, SAMOSA3, SAMOSA+). The SAMOSA 3 is a truncated version of SAMOSA 2 (only zero-order term); more info at REF2. SAMOSA+ is a enhancement of the SAMOSA2 solution providing better performances over inland water, coastal zones and sea ice; more info at REF3.
Dump RIP in output	-NO -YES	Flag to append Range Integrated Power (RIP) in the output netCDF data product
Dump SAR Echo Waveforms in Output	- NO -YES	User can command here to attach in the output product the SAR Echo Waveforms
Single-Look or Multi-Look Model	- Multi-look - Single-look	User can decide here to retrack the waveform with the Multi-look waveform model or the single-look waveform model. Single-look waveform model is an acceptable approximation over inland water or sea-ice scenario. The Single-Look option is quicker than Multi-Look option.
Choose the Default Tide Model	-FES2014b -TPXO8 -TPXO9	User can decide here which tide model to use as default in the product between FES2014b, TPXO8-ATLAS, TPXO9-ATLAS
Choose the Default Mean Sea Surface Model	-DTU18 -DTU15 -CLS-CNES15	User can decide here which mean sea surface model to use as default in the product between DTU18, DYU15, CLS-CNES15

spazio

Adopted Conventions

The Reference Time for the TAI Datation is 01/01/2000 00:00:00
The Vertical Datum for altitude reference is the WGS84 Ellipsoid
The Curve Best-Fitting Scheme, used in the re-tracking stage, is a

Bounded Levenberg-Marquardt Least-Squares Estimation Algorithm (LEVMAR-LSE).

The SAR Power Return Waveform Model, used in the re-tracking stage, is the SAMOSA Model (REF2).

Static bias has been applied to the range, sigma nought and antenna

mispointing measurements. The values of the static biases are reported

in the output NetCDF data products.

The product orbit latency/source is the one coming from the SAR L1A data file used in input.
The orbital altitude has been corrected for a Time Tag Bias. The value of the applied time tag bias is reported in the output NetCDF data products.

The measurements are posted both at rate of 20 Hz and of 1 Hz.
The sea state bias has not been applied to the sea level anomaly measurements.

No a priori data editing has been applied to the 20 Hz measurements.
The misfit between SAR Waveform Model and SAR Waveform Data has been computed as:

sqrt( 1/128*sum(residual)^2 )*100

where residuals are the differences between model's waveform power and

data's waveform power, normalized for the waveform power's maximum value.

Known Limitations of Sentinel-3 GPOD Products

The Sentinel-3 GPOD products feature the following limitations:

- The Dual Frequency Iono Correction (Ku&C) is not computed. Instead, the iono correction from the JPL GIM model is provided in the L2 product.

- The MWR wet tropo correction is not computed. Instead, the wet tropo correction from the ECMWF model is provided in the L2 product.

- The sea state bias solution in the L2 product is the CLS 2012 Jason-2 SSB solution

Known Limitations of Sentinel-3 GPOD System

We recommend to select from the Graphic User Interface the Cluster Element CREODIAS. With this framework, the input data products are fetched and processed in local on the CREODIAS platform (https://creodias.eu/) and not transfered on fly to GPOD platform from a third party repository.

Instead, in case of your choice for Cluster Element ESRIN CE 01 or ITER (GPOD local Clusters), the Sentinel-3 STM L1A data products get transfered on fly from the ESA Copernicus SciHub Sytem to GPOD servers at user's request. Currently, the ESA SciHub System is still in pre-operational phase and hence, with above-said two GPOD local Clusters, becasue constrains to simultaneous data transfers from SciHub are in place or becasue of on fly data transfer errors, data gaps in the output data can occur.

Hence, there are currently limitations in the usability of the Sentinel-3 GPOD service:

- The max number of products in one single task is 250 in order to limit a single task duration at around 3 days

- By default, the users can run just one single task at the time (i.e. one credit per user). Email to eo-gpod at esa.int if you need to run more tasks in parallel

- The GPOD L2 data products are available since a sensing start date of 15 06 2016 for Sentinel-3A and from 11 11 2018 for Sentinel-3B

- Gaps in the product catalogue can occur whether gaps are present originally in the

ESA SciHub System database and Cluster ESRIN CE 01 or ITER are selected. We recommend always to select Cluster CREODIAS to avoid gaps in the data output.

Output Format

The L2 data products generated in output are in standard NetCDF format (CF-convention), fully compatible with BRAT.

The file naming convention for those files is :

RES_INPUT_FILENAME.nc

where:

INPUT_FILENAME is the filename of the SAR L1A data file given in input

(for more details on the input filename, please check the product specification)

The netCDF format is self-explanatory with all the data field significance described in the attributes.

Along with the NetCDF file, it is provided in output a .png picture of the radar data image (Radar Echogram) and a kmz file of the satellite pass ground-track.

Selecting the corresponding option, user can decide to post in output also the SAR Stack Data Products, still in netcdf format.

The naming convention for those stack data products is :

STK_INPUT_FILENAME.nc

They will be held in the STACK folder inside the .tgz output package

How to access to the service

The service is open and free of charge.

In order to be granted the access to the service, you need to have an EO-SSO account (for EO-SSO registration, go here) and afterwards, you need to submit an e-mail to G-POD team (write to eo-gpod@esa.int), requesting the activation of the service for your EO-SSO user account.

Then, the service is available here.

Offline Processing Order

In case of need of massive data processing (months of data with several specific processing baselines), it is highly recommended to request an off line processing order, sending an email to eo-gpod at esa.int with all the details of selected processing baselines and area/time of interest. G-POD team will try to accomodate your request, processing off line the data and delivering to you via ftp.

Contact Us

For any question, bugs and support, please contact us at:

altimetry.info@esa.int

For G-POD specific questions please contact:

eo-gpod@esa.int

Join & Share Forum

In order to express feedbaks, place requests and share information, you can

consider to consult the SARvatore Join & Share Forum here

Acronyms

SAR: Synthetic Aperture Radar

WGS: World Geodetic System

EOP-SER: Earth Observation Programmes: Science, Exploitation, Research

CF: Climate & Forecast

G-POD: Grid Processing On Demand

LUT: Look-Up Table

L1A: Level 1A

L1b: Level 1b

EO-SSO: Earth Observation Single Sign-On

References

REF1: Guidelines for the SAR (Delay-Doppler) L1b Processing, ESA, 2013, available here

REF2 : (SAMOSA Model Paper) Ray, Chris & Martin-Puig, Cristina & Clarizia, M.P. & Ruffini, Giulio & Dinardo, Salvatore & Gommenginger, Christine & Benveniste, Jerome. (2015). SAR Altimeter Backscattered Waveform Model. Geoscience and Remote Sensing, IEEE Transactions on. 53. 911-919. 10.1109/TGRS.2014.2330423.

REF3 : Dinardo, Salvatore & Fenoglio, Luciana & Buchhaupt, Christopher & Becker, Matthias & Scharroo, Remko & Joana Fernandes, M & Benveniste, Jérôme. (2017). Coastal SAR and PLRM altimetry in German Bight and West Baltic Sea. Advances in Space Research. 62. 10.1016/j.asr.2017.12.018.