MFSPP-WP3000 : Planning


SPACE OCEANOGRAPHY MFSPP WP 3000 Near Real Time Remote Sensing Data Collection and Analysis
Objectives Participants Planning Task Structure Data products Data processing Publications Links

TASK 3100 : Develop and test algorithms to process NRTaltimeter data

Forecasting activities will require (at least for intra-seasonal forecasting) near real time altimeter data (the so-called FDP or IGDR altimetric data) (i.e. data collected within a few days). These are less accurate than delayed time data (GDR data).
For example, the accuracy of T/P IGDR orbits (T/P near real time data sets which are available within 3 days) is 5 cm rms for about 2 cm for T/P GDRs. ERS-1/2 have near real time data (Fast Delivery Products or FDPs) which are available within 3 hours but they do not contain any orbit. ERS-1/2 FDPs are also less accurate than ERS-1/2 OPRs because they have different instrumental processing and have less precise altimetric corrections. This means that a special processing will have to be applied to NRT data (in particular to remove the orbit error) and the final accuracy of altimetric NRT data must be assessed.

To prepare forecasting activities in the Mediterranean, the following work is thus proposed :

Subtask 3110 : Orbit error reduction and mapping

The relatively large orbit error in the NRT data and the long wavelength error due to inaccurate inverse barometer correction (Le Traon and Gauzelin, 1997; Ayoub et al., 1998) need to be reduced. Based on the mapping method described in Le Traon et al. (1998), a local inverse method was developed and tested to estimate the long wavelength errors along each satellite track. The method is similar to optimal or objective analysis but in this case the field to be mapped is the long wavelength error. Several tests were performed with different choice of parameters (a priori noise variance, correlation functions). Results were shown to be quite robust to these a priori choices. One year of T/P and ERS-2 delayed time data was finally processed and results were analyzed. The method was shown to reduce quite well the biases between neighboring tracks. The biases have a typical rms amplitudes of 3 cm for T/P and 5 cm for ERS. Similar tests were also performed for the mapping procedure. Final choice of parameters for the orbit reduction and mapping are given in processing .

Subtask 3120 : Compute a precise mean sea surface.

A high resolution and precise mean sea surface (MSS) was constructed using TOPEX/POSEIDON , ERS-1 (35 day repeat cycle and 168 day repeat cycle), ERS-2 (35-day repeat cycle) and GEOSAT Expact Repeat Mission data. The period of reference which corresponds to a 5 year mean, starts in January 1993 and ends in December 1997. TOPEX/POSEIDON mean profile (which is very precise given the low T/P orbit error) is used as a reference for the other missions. Standard (35 days) and geodetic (168 days) ERS-1/2 data were first corrected by T/P Sea Level Anomaly to reduce the problem of oceanic variability aliasing.

Subtask 3130 : Compare with results obtained with off-line data sets

The comparison with delayed time data will allow us to quantify the accuracy of NRT Sea Level Anomaly. The comparison will be performed over a 3 month period. This will be done for the year 1998 (T/P and ERS-2 data). The comparison will also be used to optimise the NRT processing strategy.

( subtask 3110).

TASK 3200 : Develop a Near Real Time SST product

SST products available up to now operationally and in quasi real-time are derived from AVHRR /2 radiometers flying onboard NOAA polar orbiting satellites (NOAA-11, NOAA-13, etc.) with a maximal horizontal resolution of 1.1 km at nadir. The maximal observation frequency at a given location is 2 times daily by one satellite, assuming clear sky conditions. AVHRR SST will the product used for MFSPP although, in the longer run, other products may be envisioned (e.g. ATSR , MSCG). AVHRR SST for the Mediterranean sea is already operationally produced and validated with in-situ measurements by participant xx (CMS). The existing product lacks, however, a small portion of the eastern Mediterreanean sea. It will be provided by participant 5 ( ENEA). An optimal strategy for SST data processing will be defined : cloud detection, validation, weekly means calculation (space and time integration of measurements).

TASK 3300 : NRT data processing before and during the TOP

Subtask 3310 : NRT altimeter data

T/P and ERS-2 NRT data will be acquired and processed before and during the TOP . The processing period will cover the last year of TOP-2 (i.e. from T0 + 6 months to T0 + 15 months), the TOP-1 and the TOP. This will allow us to drive the model up to the TOP. The processing will be in delayed mode for the TOP-2 and in near real time for the TOP-1 and TOP. The actual delay for the near real time data acquisition and processing will depend on data availability and results from task 3100. It should be around one week.

The NRT Data products and processing will include :

Acquisition of ERS-2 FDPs and T/P IGDRs through the GTS system and the AVISO ftp site, respectively.
Organisation of the data in chronological order and in half orbits like in the GDRs/ OPRs.
Addition of appropriate orbit fields for ERS-2 (the JGM-3 orbits computed in near real time by Delft University for NOAA).
Addition of improved altimetric corrections (dry and wet tropospheric corrections based on meteorological model outputs for ERS-2, same tide model for T/P and ERS-2).

These first steps will use the global NRT altimeter Data products module developed by participant 1 ( CLS). The next steps specific to MFSPP will be :

Correction of orbit error, extraction of along-track Sea Level Anomaly and mapping of Sea Level Anomaly using inverse methods developed in task 3100. Corrected along-track data and weekly maps at the model resolution (1/8°x1/8°) will be produced
Quality control of result and distribution through the MFSPP remote sensing data WWW server (see task 3400).
A detailed analysis of altimeter data and maps will be performed during the 3-month operational TOP period.
Comparison with delayed time data will be made when these data are available. This will allow an a posteriori NRT data quality control.

Subtask 3320 : Sea Surface Temperature

SST products will be weekly maps at the model resolution (1/8°x1/8°) obtained from night time data (available with a delay of 2 days) for the last year of the TOP -2 (i.e. from T0 + 6 months to T0 + 15 months), the TOP-1 and the TOP. In addition during the TOP (3 months), full resolution SST maps will be distributed. SST data will be validated against in-situ data. All products will be checked and analyzed before delivery and will be made available on a WWW server (numerical data and images) and an FTP site (see task 3400).

A detailed analysis of SST data during the TOP period will also be performed (large scale signals, identification of mesoscale structures, etc). It will be used for model validation.

Subtask 3330 : Ocean color

SeaWIFS data will be available using a reception station similar to that used for AVHRR data (HRPT station). Participant 3 ( ACRI) will own such a station and be in a position for delivering SeaWIFS ocean color data during the TOP -1 and the TOP (6 months). The following products will be delivered approximately every week in near real time :

Mediterranean sea chlorophyll maps [weekly means at the model resolution (1/8°x1/8°)].
Primary production maps (photosynthetic carbon fixation) [weekly means at the model resolution)].
Full resolution chlorophyll maps for selected areas of the Mediterranean sea (TOP period only).

All products will be checked and analyzed before delivery and will be made available on a WWW server (numerical data and images) and an FTP site (see task 3400).

TASK 3400 : Remote sensing data management

Remote sensing data will have to be gathered and quality controlled in near real time before they are used in the assimilation/forecasting system. The TOP experiment will be a good opportunity for experimenting with the MFS remote sensing data flow. Remote sensing data will be distributed through a dedicated FTP/World Wide Web server.

Subtask 3410 : Development of a MFSPP remote sensing WWW server

The MFSPP FTP/WWW server will contribute to online information and data exchange between the MFS participants but also with the wider MFS products user community. It will be based on a decentralised architecture, i.e. the relevant data sets (altimetry, SST, ocean color) will be available via a link to the different participant home pages.

The MFSPP remote sensing server will feature:

A description of the MFSPP remote sensing WP.
A catalogue of data sets (including a description of data content, information on data availability and means to obtain the data).
Maps and animations based on MFSPP products (altimetry, SST, Ocean color).

Subtask 3420 : Data distribution

Each responsible of a data set (altimetry, SST, ocean color) will make it available through the MFSPP remote sensing server and an FTP site. He will be responsible for the quality control of the corresponding data sets. During the TOP-2, altimetry and SST data will be distributed in delayed time. During the TOP-1 and TOP, altimetry, SST and ocean color data will be distributed in near real time (e.g. every week).

The task corresponds to the coordination of these data flows, the regular updating of the different WWW home pages (new maps, short description of maps, etc) and the interfacing with MFSPP users (user service).