Climate forecast: Sub-seasonal
A global dataset at 0.25° or 0.10° resolution providing bias-corrected sub-seasonal forecasts with a 30-day outlook. The dataset is based on the ECMWF extended range forecast and post-processed to ensure consistency with observed climate statistics.
Specifications
Section titled “Specifications”| Specification | Description |
|---|---|
| Experiment | Sub-seasonal forecast |
| Forecast system | Extended range forecast provided by the European Centre for Medium-Range Weather Forecasts (ECMWF), bias-corrected against observations. It is composed of 100 forecast members. |
| Reference dataset | ERA5/ERA5-Land reanalyses |
| Spatial resolution | 0.25°x0.25° Global or 0.10°x0.10° global over continents |
| Temporal resolution | Daily values over a 30-day forecast horizon |
| Variables | Four essential variables (others available on demand): Mean 2m temperature, Maximum 2m temperature, Minimum 2m temperature, Total precipitation |
| Delivery | Once a month on the 15th. |
| Format | NetCDF, other formats possible depending on data volume |
Supplementary information
Section titled “Supplementary information”The sub-seasonal forecast dataset is statistically bias-corrected using a quantile-mapping approach based on the CDF-t (Cumulative Distribution Function transform) method, applied independently for each lead week and grid point.
Key methodological characteristics:
- Probabilistic bias correction: all ensemble members are treated as equally likely and jointly used to estimate forecast probability distributions.
- Lead-time-dependent correction: bias correction is performed separately for each lead week (weeks 1 to 4), on 20 years of ECMWF reforecasts, to account for forecast drift.
- Non-stationarity handling: the CDF-t method allows model and observational distributions to evolve over time, avoiding the stationarity assumption of standard bias-correction approaches.
- ERA5-consistent climatology: forecast probabilities are computed relative to ERA5-based climatologies derived from the hindcast period.
The reference for the original sub-seasonal forecast (ECMWF extended range) is ecmwf.int/en/forecasts/datasets/set-vi
Availability
Section titled “Availability”The data can be delivered according to your specifications, contact us for a quote.
See all our seasonal and sub-seasonal forecast data.