Subseasonal to Seasonal Climate Prediction Techniques

Subseasonal to seasonal (S2S) prediction occupies the forecasting window between weather models and long‐range seasonal outlooks, typically spanning lead times of two weeks to two months. Skillful S2S forecasts support agriculture, water management, disaster preparedness and energy planning by anticipating temperature and precipitation anomalies beyond the medium‐range. Modern approaches rely on dynamical coupled atmosphere–ocean–land models, often run as large ensembles to sample uncertainty in initial conditions and model physics. Statistical and machine‐learning methods have been reintroduced to calibrate dynamical outputs, correct biases and extract precursor signals from large datasets. Key predictability sources include slowly varying ocean states such as El Niño–Southern Oscillation, land‐surface memory via soil moisture and snow cover, and intraseasonal variability exemplified by the Madden–Julian Oscillation. Recent advances aim to refine ensemble calibration, improve representation of teleconnections and harness artificial intelligence to extend forecast skill. Together these techniques are forging a new paradigm in Earth‐system prediction, linking physical understanding with data‐driven innovation to bridge the subseasonal gap.

Research from Nature Portfolio

Recent studies have introduced a deep‐learning subseasonal‐to‐seasonal model trained on decades of reanalysis. This machine‐learning framework generates global daily forecasts up to 42 days for multiple atmospheric and surface variables, yielding superior ensemble means and spread compared with the leading operational system. Its enhanced skill arises from a better capture of forecast uncertainty and an extended predictive range for the Madden–Julian Oscillation, thus improving precipitation and outgoing longwave radiation forecasts. Another investigation assessed extended‐range predictions of a major European heatwave, showing that multi‐model subseasonal ensembles can capture blocking regimes and large‐scale teleconnections up to three weeks ahead. While regional heat‐spell occurrence remains challenging, the work emphasises the value of probabilistic ensemble approaches and identifies key circulation patterns that underpin subseasonal predictability.

Subseasonal to Seasonal Climate Prediction Techniques publication trend

The graph below shows the total number of articles in subseasonal to seasonal climate prediction techniques across all publications each year (not limited to Nature Index journals).

Technical terms

Subseasonal to Seasonal (S2S): Forecast horizon covering lead times from two weeks up to two months.

Ensemble Forecasting: Technique of running multiple model simulations with varied initial conditions or physics to sample forecast uncertainty.

Madden–Julian Oscillation (MJO): A tropical intraseasonal fluctuation in convection and circulation with a 30–60-day period that influences global weather patterns.

Teleconnection: Atmospheric linkage whereby perturbations in one region (for example ENSO) influence climate anomalies in remote areas.

Reanalysis Data: A consistent historical dataset produced by assimilating observations into a fixed model framework to reconstruct past states of the climate system.

References

1. The sub-seasonal to seasonal prediction project (S2S) and the prediction of extreme events. npj Climate and Atmospheric Science (2018).
2. Progress in subseasonal to seasonal prediction through a joint weather and climate community effort. npj Climate and Atmospheric Science (2018).
3. S2S reboot: An argument for greater inclusion of machine learning in subseasonal to seasonal forecasts. Wiley Interdisciplinary Reviews Climate Change (2018).
4. A machine learning model that outperforms conventional global subseasonal forecast models. Nature Communications (2024).
5. Quantifying sources of subseasonal prediction skill in CESM2. npj Climate and Atmospheric Science (2024).
6. Seasonal forecasting skill for the High Mountain Asia region in the Goddard Earth Observing System. Earth System Dynamics (2023).
7. Significant advancement in subseasonal-to-seasonal summer precipitation ensemble forecast skills in China mainland through an innovative hybrid CSG-UNET method. Environmental Research Letters (2024).
8. The 2018 summer heatwaves over northwestern Europe and its extended-range prediction. Scientific Reports (2020).