Towards an improved Sea Surface Temperature Climate Data Record from the MeteoSat Second Generation mission (2004-2023)

In the context of the OSI SAF Visiting Scientist Program, Marouan Bouali from ORBTY Ltda., São José dos Campos, Brazil, worked on several areas of improvement for an upcoming Sea Surface Temperature Climate Data Record from the Meteosat Second Generation (MSG) mission. This work took place in 2024 and was supervised by Stéphane Saux Picart (Météo-France/CNRS, Lannion, France).

Objectives and framework of the study

In 2018, the Sea Surface Temperature (SST) team of OSI SAF released a Climate Data Record based on the measurement made with the Spinning Enhanced Visible and Infrared Imager (SEVIRI) on board MSG satellites. It covers the period from 2004-2012. OSI SAF plans to create a CDR for the entire MSG mission (2004-2023). In this context, we explored three areas of improvement in the current processing chain for the upcoming climate data record, which consisted in:

• Improving the consistency of SST gradients in OSI SAF products from MSG SEVIRI via the use of the Multidimensional Dynamic Data Fusion System (M3DFS)
• Deriving new climatologies of SST gradients to be used in the future for operational Near Real Time and climate data record SST products.
• Developing and testing new methodologies for the harmonization of the upcoming OSI SAF MSG climate data record SST (2004-2024).

Figure 1: Daily Sea Surface Temperature gradient magnitudes for February 27th, 2007 over the Northwestern coast of Africa | TA: derived from time-averaged | M3DFS derived from Multidimensional Dynamics Data Fusion.

Report conclusions

This Visiting Scientist Activity, aimed mainly as diagnostic analysis of the temporal stability of MSG-1/MSG-2 measurements. It had significants results on three directions:

• The M3DFS was tested using high temporal resolution MSG SEVIRI data over one year (2007), to generate daily SST (level 3C). The statistical validation based on independent in situ measurements provided by NOAA’s iQuam system, have shown similar performance to that obtained with conventional time-averaging approach. More importantly however, analysis of the spatio-temporal characteristics of gradient magnitudes derived from M3DFS SST, have shown significant improvement both in the spatial distribution and the temporal variability of SST gradients (see Figures 1 and 2). The artifacts contribution to the overall gradients represent approximately 15% over the entire SEVIRI disk and can go up to 50% over specific regions with no significantly intense ocean surface dynamics like the Mediterranean Sea. The application of the M3DFS allows to remove the contribution of these artifacts to the overall gradient magnitude.

• The currently quality control of the SEVIRI SST is based, on a comparison of the SST gradient magnitude with monthly climatologies covering the period 1985-1996 (Andersen and Belkin, 2006). Analysis of this climatology has shown several limitations related to data gaps and extended regions with identical values. To mitigate the impact of such limitations on future near real time and reprocessed OSI SAF SST products, two independent climatologies of SST maximum gradients have been derived from SEVIRI (2004-2012) and from Aqua MODIS (2003-2023). These new climatologies provide a more consistent spatial distribution of SST maximum gradients and will serve as a basis for improved SST quality control and discrimination between clouds and oceanic fronts.

• Preliminary diagnostic analysis of brightness temperatures derived from Aqua MODIS and SEVIRI was conducted using a match-ray technique to generate a matchup dataset. Comparison of brightness temperatures was done at the beginning and end of both MSG-1 and MSG-2 missions. Results derived from a linear regression fit model, indicate very good agreement between those from Aqua MODIS and MSG SEVIRI. These results likely indicate that no major drift in the calibration of corresponding SEVIRI’s thermal channels has occurred throughout the two satellite missions. Pending the generation of a larger satellite-satellite matchup dataset, the MODIS mission onboard Aqua and Terra provides a major potential to adjust the brightness temperatures from SEVIRI and ensure a long term homogeneous climate data record of SST across all MSG missions.

Figure 2: Time series of SST gradient magnitudes derived from time-averaging and M3DFS Level 3C daily SST (left) and corresponding absolute differences (right) over the entire SEVIRI Disk for 2007.

Benefits for the SAF

• The M3DFS has shown promising results when applied to MSG SEVIRI SST and provides a wide range of potential applications. In the context of the upcoming SST climate data record, this includes the generation of daily or monthly climatologies of SST. These climatologies can as well be used in the Level 2 SST retrieval scheme to improve the spatio-temporal consistency of SST gradients derived from the MeteosatSG missions.
• New monthly climatologies of maximum SST gradient magnitudes derived from the MeteosatSG1 and MeteosatSG2 missions (2004-2012) as well as from the Aqua MODIS mission (2003-2023) can provide significant improvement in the quality control of SST, particularly with respect to the misclassification of ocean sharp fronts as clouds.
• The harmonization of the SST climate data record can potentially rely on the overlapping MODIS missions to adjust SEVIRI measures prior to SST retrieval and ensure long-term stability of the climate data record.

Report on this study

Towards improved Climate Data Record Sea Surface Temperature from the Meteosat Second Generation mission (2004-2023)

Authors

Marouan Bouali - Orbty Ltda.

Stéphane Saux Picart - CNRM - Météo-France/CNRS