El Niño–Southern Oscillation (ENSO) research forecast
Preliminary results using the "Wyrtki" cyclostationary linear inverse model (Wyrtki-CSLIM)
ENSO forecasts and past skill assessments are from the Wyrtki-CSLIM prediction system, which is described in: Wang et al., “ENSO Predictability From Combined Wyrtki and Hasselmann Memory in a Cyclostationary Linear Inverse Model,” Geophysical Research Letters (2026) .
This website includes:
- Real-time forecasts for: (i) the Niño-3.4 index (area-mean SST anomalies over 5°S–5°N, 170°–120°W) and (ii) global gridded SST anomalies, issued from the latest monthly sea surface temperature (SST) and sea surface height (SSH) from the Wyrtki-CSLIM, out to 12 months.
- Retrospective hindcast:the retrospective 6-month lead hindcast of the Niño-3.4 index from the Wyrtki-CSLIM.
- Forecast skill: The Wyrtki-CSLIM's ENSO forecast skill anomaly correlation coefficient (ACC) and root-mean-square error (RMSE).
- Model training predictors: empirical orthogonal functions (EOFs) of SST and SSH used in the Wyrtki-CSLIM.
- Predictability sources: optimal initial conditions linked to ENSO events.
Figures update as new data become available (on the 16th of each month, Hawaii Time). Feedback and comments are welcome.
Wyrtki-CSLIM Niño-3.4 SST anomaly monthly real-time forecast
For additional ENSO outlooks, including statistical and dynamical model forecasts from other groups, see the International Research Institute (IRI) for Climate and Society ENSO forecast page: iri.columbia.edu/our-expertise/climate/forecasts/enso/current/.
Global gridded SST anomaly real-time forecast from the Wyrtki-CSLIM
Retrospective Niño-3.4 index hindcast with observation overlay
Wyrtki-CSLIM ENSO (out-of-sample) forecast skill
Leading EOF patterns of SST and SSH
Predictability source: optimal initial and final evolved conditions from the Wyrtki-CSLIM
Code availability: the Linear Inverse Model (LIM) Python toolbox, along with code to reproduce Wang et al. (preprint in review) results, is available at github.com/uhsealevelcenter/Wyrtki-CSLIM or github.com/WANGYuxinCi/linear-inverse-model
Funding source: This work was primarily supported by NOAA Climate Program Office’s Climate Variability Predictability (CVP) and NOAA Climate Program Office’s Modeling, Analysis, Predictions, and Projections (MAPP) cross-program grant NA23OAR4310454.
Disclaimer and Note Regarding Forward Looking Statements: The Wyrtki-CSLIM results are for research use and informational purposes only, and there is no guarantee of the accuracy of the contents on this website. This website and its affiliated entities expressly disclaim any liability for decisions or actions based on reliance on this information. Furthermore, no responsibility is assumed for any consequential, special, or similar damages resulting from such reliance.
