Watching ENSO Through Sea Surface Salinity
  • Dashboard
  • Methods
  • Acronyms
  • Update log
  • SMOS home
  • LOCEAN

On this page

  • Scientific basis
  • Data source
  • Index definition
  • Comparison indices
  • Uncertainty and limitations

Methods

Scientific basis

This demonstrator follows the idea that the eastern edge of the western Pacific fresh pool can be tracked through a salinity isohaline near the equator. The 34.8-pss isohaline is used as a proxy for this boundary.

Data source

The page is based on CCI+SSS satellite sea surface salinity fields, produced in the framework of ESA CCI+SSS and its transition toward ECMWF C3S.

For the Climate Data Record period, the demonstrator uses the soon-to-be-released CCI+SSS version 6 product, covering 2010–2024. An interim Climate Data Record updated in near-real time is currently in development and is expected to become operational at the end of 2026.

In the meantime, the near-real-time extension currently used for the latest months is based on the CATDS SMOS+SMAP product, which is also updated in near-real time.

Index definition

The experimental SSS-based index is based on the longitudinal position of the 34.8-pss isohaline in the equatorial Pacific.

Configuration used here:

Isohaline: 34.8 pss
Latitude band: 2°S–2°N
Longitude range: 128°E–270°E
Temporal resolution: monthly
Reference period: 2010–2024
Product version: CCI+SSS v6.3 for the CDR period, with CATDS SMOS+SMAP for the current near-real-time extension

Comparison indices

The SSS-based index is compared with NOAA PSL MEI.v2, NOAA PSL HadISST Niño-3.4, a C3S-derived Niño 3.4 index, and Monthly relative Niño-3.4 (ERSSTv6, 1991–2020) from NOAA CPC. Additional reference indicators may be considered as the analysis develops.

MEI.v2 represents overlapping two-month seasons, identified by their ending month, with a 1980–2018 reference period. The latest value is September 2026 (August–September 2026). HadISST Niño-3.4 uses monthly SST anomalies relative to 1981–2010 and extends through July 2026. Monthly relative Niño-3.4 uses ERSSTv6 with a 1991–2020 reference period and extends through September 2026; these are individual monthly values, not a three-month running mean.

The NOAA comparisons use the exact timestamp intersection with first-of-month S34.8 positions, starting on 1 February 2010. No interpolation, date shifting, or additional smoothing is applied. The matched comparisons end in September 2026 for MEI.v2 and monthly relative Niño-3.4 (200 samples each), and July 2026 for HadISST Niño-3.4 (198 samples). Their Pearson correlations with S34.8 are 0.850, 0.847, and 0.817, respectively.

The C3S Niño 3.4 index is derived from SST anomalies in the Niño 3.4 region in the C3S combined sea surface and sea ice surface temperature dataset, using a fixed 1991–2020 climatology. The C3S comparison uses twice-monthly sampling on the 1st and 15th and extends through September 2026 (Pearson correlation 0.835; n = 400). The latest September SST estimate is provisional.

Uncertainty and limitations

Uncertainty of the S34.8 position. The shaded ranges shown in the S34.8 time series are estimated using 1,000 Monte Carlo realizations, in which satellite salinity values are perturbed according to assumed random measurement errors. The displayed limits correspond to the 1st and 99th percentiles of the successfully identified isohaline positions. These ranges characterize the sensitivity of the indicator to salinity errors under the adopted tracking procedure. They do not account for all possible sources of uncertainty, particularly ambiguity between multiple 34.8 crossings, which can occur near the western Pacific islands during strong La Niña conditions.

For the recent CATDS-derived extension, the assumed error amplitudes are estimated from historical CCI salinity errors where suitable product error estimates are unavailable.

Several methodological choices influence the index:

  • the salinity product version;
  • the selected latitude band;
  • the isohaline value;
  • uncertainty in satellite retrievals;
  • treatment of near-real-time data;
  • definition of thresholds.

The index should therefore be interpreted as a complementary indicator of ENSO-related salinity variability.