Datasets

A subset of ECMWF real-time forecast data from the IFS and AIFS models is made available to the public free of charge. Their use is governed by the Creative Commons CC-BY-4.0 licence and the ECMWF Terms of Use.

Ensemble forecast runs produced by the ECMWF Artificial Intelligence Forecasting System (AIFS) Ensemble model.

More information can be found on the implementation page. 

Product description

On 12 May 2026, a new forecast stream will be produced by ECMWF's operational Artificial Intelligence Forecasting System ensemble model (AIFS ENS).

The new forecast stream is for wave forecast runs, marking ECMWF's first operational data-driven wave forecasts.

Product description

Single forecast runs produced by the ECMWF Artificial Intelligence Forecasting System (AIFS) deterministic model.

On 12 May 2026, a new forecast stream will be produced by ECMWF's operational Artificial Intelligence Forecasting System deterministic model (AIFS Single).

The new forecast stream is for wave forecast runs, marking ECMWF's first operational data-driven wave forecasts.

Product description

Ensemble (ENS) of forecasts providing an estimate of the reliability of a single forecast. Currently there are 50 perturbed members in the ensemble.

ENS offers  "High Frequency products"  until step 144:

This licence provides access to ECMWF interactive ecCharts tool to visualise analysis and forecast products  Please note that access to the ecCharts is for End User use only (internal).  

ECMWF ecCharts web service is available and we are pleased to be able to offer the service for evaluation to customers of ECMWF's web products.

The sub-seasonal products comprise ensembles of individual forecasts up to 46 days and post-processed products of average conditions (e.g. weekly averages) and the associated uncertainty.

The purchase of the "Basic Set" +72, +96, +120, +144, +168 hrs is a mandatory prerequisite for the purchase of time steps in the range 12 to 66 hours.

The following sub-sets are available from the sub-seasonal forecast (46 days):

The DestinE Digital Twin for Climate Change Adaptation (Climate DT) supports adaptation activities through the provision of innovative climate information on multi-decadal timescales, at scales at which the impacts of climate change are observed.

SEAS comprises ensembles of individual forecasts coupled to an ocean model and post-processed products of average conditions (e.g. monthly averages) with the associated uncertainty. Products are available up to 7 months ahead.

The following sub-sets are available:

Field computed from data of the daily individual forecast runs (section V-v) and averaged over all ensemble members. The fields are provided in GRIB code.

The DestinE Digital Twin for Weather-Induced Extremes (Extremes DT) supports rapid decision-making in response to meteorological, hydrological and air quality extremes.

These seasonal data comprises ensembles of individual forecasts coupled to an ocean model and post-processed products of average conditions (e.g. monthly averages) with the associated uncertainty. Products are available for forecast month 8-13, each quarter (February, May, August, November).

The following sub-sets are available:

Single prediction that uses

IFS Medium range Control forecast offers "High Frequency products"  

Not all post-processed Products are available at 06/18 runs or in hourly steps.

IFS Medium-range Control forecast (Wave) is coupled to the IFS Medium-range Control forecast and it provides direct output "High Frequency products":

The model computes between the latitudes -78 and 90, the output grid is artificially extended to -90.

Single prediction that uses

HRES Direct model output Products offers "High Frequency products"  

Not all post-processed Products are available at 06/18 runs or in hourly steps.

Products based on an ensemble of forecasts providing an estimate of the reliability of a single forecast.

The following sub-sets are available from the ENS-WAM:

The AI Weather Quest (AI WQ), organised by ECMWF, is an ambitious international competition designed to harness artificial intelligence (AI) and machine learning (ML) in advancing weather forecasting.

15-member coupled IFS (cycle 43R1) extended-range reforecast experiment covering the period 1989-2015. The atmosphere is configured with 91 vertical levels and uses the Tco399 cubic octahedral reduced Gaussian grid. The IFS is coupled hourly to the 75 level version of the NEMO v3.4 ocean model and the LIM2 sea-ice model, both of which use the ORCA025 tripolar grid. Coupling follows the implementation used in ECMWF operational forecasts.

15-member coupled IFS (cycle 43R1) extended-range reforecast experiment covering the period 1989-2015 with bias-corrected sea-surface temperatures (SSTs) in the North Atlantic region. This experiment can be compared with gkzp, which is the relevant control without bias-correction. The atmosphere is configured with 91 vertical levels and uses the Tco399 cubic octahedral reduced Gaussian grid.

HRES-WAM (High RESolution WAve Model) is coupled to the atmospheric model (HRES) and it is a Direct model output Products offers "High Frequency products":

The  model computes between the latitudes -78 and 90, the output grid is artificially extended to -90.

An eddy-permitting ocean reanalysis spanning the period 1979–2012. Includes increased horizontal and vertical resolution, an prognostic sea-ice component, new versions of the ocean and data assimilation system, revised surface fluxes, new version and treatment of satellite sea surface height data, and assimilation of sea-ice concentration, among others.

Uses a sophisticated data assimilation methodology which includes a model bias correction. The ocean model used is forced by atmospheric daily surface fluxes, relaxed to SST and bias corrected.

The ECMWF OCEAN5 system is a new global eddy-permitting ocean-sea ice ensemble reanalysis analysis system. This Technical Memorandum gives a full description of the OCEAN5 system, with the focus on its Behind-Real-Time (BRT) component, the reanalysis product ORAS5. The OCEAN5 Real-Time (RT) component includes all upgrades developed for ORAS5 and runs daily using the latest observations and forcing fields from the operational Numerical Weather Prediction (NWP).

 CAMS Global atmospheric composition forecast production system is used to produce the daily forecasts of pollutants, aerosols and greenhouse gases across the globe. Satellite observations of atmospheric composition are merged with a detailed computer simulation of the atmosphere using a method called data assimilation. The resulting analyses, i.e.

WMO Recommended, previously known as WMO Additional, is available to the public and governed by the Creative Commons CC-4.0-BY licence and also ECMWF Terms Of Use.