SLSTR A Spot

New Sentinel-3 SLSTR Level 1 and Level 2 SST Processing Baselines released soon

 

SLSTR A Spot
SLSTR A Spot

New Processing Baselines (PB), SL__L1_.004.05.00 for Sentinel-3 SLSTR Level 1 (L1) RBT products and SL__SST.003.07.00 for Sentinel-3 SLSTR Level 2 (L2) SST products, are planned to be released on 12 December 2022.

Last Updated

24 November 2022

Published on

24 November 2022

The update includes a new version of the SLSTR L1 RBT Instrument Processing Facility (IPF), v06.20, and a new version of the SLSTR L2 SST IPF, v06.22. The baseline collection numbers of the SLSTR L1 RBT products (004) and SLSTR L2 SST products (003) are unchanged.

The main reasons for the SLSTR L2 SST update are the inclusion of revised SST retrievals coefficients along with their associated uncertainties, as well as a revised quality level scheme and updated SSES, while the SLSTR L1 RBT update resolves only minor issues.

SLSTR L1 PB SL__L1_.004.05 changes

  • Fixing an incorrect dimension in the bayes_orphan datasets
  • Addition of a sentinel-safe:resource to the SL1 manifest file
  • Update to the IPF to fix a potential error in generating a VISCAL auxiliary file
  • Update to the IPF to ensure a gracious exit during lunar calibration processing.

There is no impact from these updates on the quality of L2 SST or L2 NRT Aerosols & NRT Fires products.

SLSTR L2 SST PB SL__SST.003.07 changes

  • Revised SST coefficients, uncertainties, quality level flagging and SSES
  • Addition of degradation flags to SL2 manifest file copied from SL1
  • Fixing an incorrect dimension in the bayes_orphan datasets

Further details of changes and impacts for SLSTR L2 SST

The main change in these updates is the introduction of revised SST coefficients. The new coefficients address (1) limitations with the current coefficients including a limitation of the N2 retrieval in high-water vapour loading and (2) dedicated SLSTR-B coefficients.

Current users will see little impact on their analysis of the changes, especially those users only using dual-view quality level 5 data as a reference sensor. A map of daily daytime and nighttime Level 3 differences for all quality level 5 data (dual-view and single-view) is shown in Figure 1 for SLSTR-A and in Figure 2 for SLSTR-B.

The revised SST coefficients mainly affect users of single-view and data for quality levels other than 5 where impacts will be seen in three main areas:

1. An increased amount of daytime quality level 5 N2 data. The revised coefficients have removed the high bias seen for N2 retrievals with high water vapour loading (> 35kg/m2). Therefore, daytime WST products will now have significantly more data flagged as quality level 5 (see Figure 3) where the N2 retrieval was selected as the best available. Users should be aware that quality levels are applied per retrieval type to remain within the GHRSST GDS 2.0 specification. Quality level 5 N2 data are therefore best data available for the N2 retrieval but are not necessarily the same quality as equivalent dual-view data. As such we do not recommend using single view retrievals if using SLSTR as a reference sensor.

2. Revised theoretical uncertainties. The SLSTR SL_2_WST product is provided with an experimental field that contains a theoretical uncertainty derived from knowledge of the retrieval and the instrument. Previously the uncertainties were validating well in the mean but were not properly representative at the pixel level. An updated algorithm is implemented that results in improved estimates across the full swath (see Figure 4). It is noted that this value is a better estimate of the uncertainty of an SLSTR SST pixel than the SSES standard deviation provided in the SL_2_WST product.

3. Updated quality level flagging. The current quality level flagging scheme is non-optimal as the dust index test used to separate data between quality level 4 and quality level 5 is not working as expected. This is because the test is picking up some failures of the cloud screening in the dual-view and so the test has been down-graded meaning quality level 3 and quality level 4 dual-view data can now be used for qualitative analysis. In addition, the same test when applied to single-view retrievals was not optimal and flagging too much good data as the impacts of aerosol and water vapour are hard to separate given the SLSTR spectral channels. As such the test has been removed meaning quality level 3 and quality level 4 single-view data can now also be used for qualitative analysis.

A reminder that sea surface temperature users should:

  • Only use quality level 5 dual-view data to ensure the best quality sea surface temperature observations for use as a reference dataset.
  • Use single-view (all quality levels) and quality level 3 and 4 dual-view data only for qualitative analysis.
  • Ensure they always apply the SSES to minimise the impact of any changes.
  • Consider using the revised theoretical uncertainties for weighting observations rather than the simple SSES standard deviation values.

Updated Product Notices will be available soon from the Sea Surface Temperatures services page. A journal article on the revised coefficients is also in preparation.

Daily SLSTR-A SL_2_WST Level 3 SST differences maps for day and night
Figure 1: Daily SLSTR-A SL_2_WST Level 3 SST differences maps for daytime (upper) and night-time (lower) showing the relative change in SST for pixels currently flagged as quality level 5 for the revised SST coefficients.
Daily SLSTR-B SL_2_WST Level 3 SST differences maps for daytime (upper) and nighttime
Figure 2: Daily SLSTR-B SL_2_WST Level 3 SST differences maps for daytime (upper) and nighttime (lower) showing the relative change in SST for pixels currently flagged as quality level 5 for the revised SST coefficients.
Daily daytime SLSTR-A SL_2_WST Level 3 SST maps for the current PB 003.06.01 data (upper) and new PB 003.07.00 data
Figure 3: Daily daytime SLSTR-A SL_2_WST Level 3 SST maps for the current PB 003.06.01 data (upper) and new PB 003.07.00 data (lower) for quality level 5. The impact of the new SST coefficients can be seen by the increased coverage in the lower plot in areas where the N2 retrieval has been selected in the SL_2_WST product towards the edges of the swath.
Daily daytime SLSTR-A SL_2_WST Level 3 SST theoretical uncertainty maps for the current PB 003.06.01 data (upper) and new PB 003.07.00 data
Figure 4: Daily daytime SLSTR-A SL_2_WST Level 3 SST theoretical uncertainty maps for the current PB 003.06.01 data (upper) and new PB 003.07.00 data (lower) for quality level 5 pixels. The impact of the revised uncertainties can be seen in the lower by the increase in values towards the edges of the swath for the N2 retrieval as the atmospheric path length increases at higher satellite zenith angles. The theoretical uncertainty is provided as an experimental field in the GHRSST format SL_2_WST product.

For more information, contact our User Service Helpdesk.