Medium-range forecasts

For the medium-range forecasts an ensemble of 52 individual ensemble members are created twice a day. One member is at a higher spatial resolution than the other members (called the HRES at ECMWF), its initial state is the most accurate estimate of the current conditions and it uses the currently best description of the model physics. 

The HRES provides a highly detailed description of future weather and averaged over many forecasts it is the most accurate forecast for a certain period, which is currently estimated as 10 days for large scale properties of the atmosphere. However for any particular forecast it may not be the most skilful member of the ensemble. Also when viewed in isolation it cannot provide an estimate of forecast uncertainty or confidence.

Another member of the ensemble (CNTL: Control forecast) is at a lower spatial resolution than the HRES but at that lower resolution it utilises the most accurate estimate of the current conditions and the currently best description of the model physics. Its significance for the ensemble is that it provides the unperturbed member to which the perturbations for the remainder of the ensemble members are applied.

The perturbed members (50 members) are similar to the CNTL but their initial states and model physics have been perturbed to explore the currently understood range of uncertainty in the observations and the model. They provide a range of possible future weather states. When averaged over many forecasts (although not necessarily for any particular forecast) these have lower skill than either the HRES or the CNTL.  However they do provide an estimate of the forecast uncertainty or confidence.

The CNTL and perturbed members are continued beyond fifteen days at a reduced horizontal resolution.

The sections below highlight the variety of ECMWF medium-range forecast products.

Depending on who you are, you will be able to access different types of products. The linked charts are only displayed if you have correct access rights for that product. Find out about access to our forecasts.


High resolution global model

Plots show different parameters over the whole planet – normally split into six different geographical areas.

Charts of 10-day high-resolution forecasts of mean sea level pressure, wind speed and temperature at low levels, and height of the 500-hPa isobaric surface are available to the public.

Model simulated satellite images may be available slightly later then other high resolution forecast charts due to longer processing time.

ENS charts

Charts of ensemble mean and spread for mean sea level pressure, wind speed and temperature at 850 hPa, and geopotential at 500 hPa are available to the public.

ENS probabilities

Forecast probabilities are computed from ENS for different parameters and relevant thresholds, for example the probability of more than 1, 5, 10 or 20 mm precipitation in a 24-hour period.

Extreme forecast index

The Extreme Forecast Index (EFI), Shift Of Tails (SOT) index, and maps of model climate quantiles are produced for forecasts up to 10 days ahead for 10-metre wind (daily mean), 10-metre wind gusts (daily maximum), 2-metre temperature (daily mean, minimum, maximum), precipitation (daily accumulations), snowfall, significant wave height.

EFI charts

An ‘anomalous weather’ chart summarises EFI information for wind, temperature and precipitation.

ENS clusters and ENS members (‘stamp maps’)

The 'stamp maps' show each ENS member individually so the forecaster can see the full range of possible forecast patterns.

The clustering groups together ENS members with similar flow evolution. For each cluster (group of similar ENS members), the ENS member closest to the centre of the cluster is referred to as the weather scenario (or representative member) for that cluster.

More information about the cluster product is available here.

ENS plumes

Show a collection of curves for the high resolution, control and perturbed forecasts over the next ten days for 850 hPa temperature, total precipitations accumulated over 12 hours and 500 hPa geopotential for different locations in Europe. They are generated on lower ENS resolution and the nearest grid point data is used.

Tropical cyclone activity

Shows the potential tropical cyclone activity at different time ranges during the forecast. This includes both tropical cyclones that are present at analysis time and those which may develop during the forecast.


Meteograms are forecasts for a single place. The Meteogram page allows you to choose a place for the forecast and has information on how it is created.

Severe weather - the Extreme Forecast Index

The Extreme Forecast Index (EFI) was developed at ECMWF as a tool to provide forecasters with general initial guidance on potential extreme weather events. It is constructed specifically to highlight occasions when there is a significant shift in the current ENS towards the extreme of the model climate. The EFI is produced for a number of important weather parameters: 2 metre minimum,  maximum and mean temperatures, total precipitation, snowfall, wind gust and mean wind speed, and for maximum significant wave height.

The complementary Shift Of Tails (SOT) provides information about how extreme an event might be. Positive values of the SOT indicate that at least 10% of the ensemble members are above the 99th percentile of the model climate. The higher the SOT value is, the further this top 10% of the ensemble forecast is beyond the model climate.


Lalaurette F. 2002. Early detection of abnormal weather conditions using a probabilistic extreme forecast index. Q. J. R. Meteorol. Soc. 129: 3037–3057.

Tsonevsky and Richardson, 2012. Application of the new EFI products to a case of early snowfall in Central Europe. ECMWF Newsletter 133, Autumn 2012, p 4.

Petroliagis and Pinson, 2012. Early indication of extreme winds utilising the Extreme Forecast Index. ECMWF Newsletter 132, Summer 2012, pp 13–19.

Zsótér E. 2006. Recent developments in extreme weather forecasting. ECMWF Newsletter 107, Spring 2006, pp 8–17.


Zsoter, E., Pappenberger, F. and Richardson, D., 2014, Sensitivity of model climate to sampling configurations and the impact on the Extreme Forecast Index. Met. Apps. doi: 10.1002/met.1447

Dutra, E., Diamantakis, M., Tsonevsky, I., Zsoter, E., Wetterhall, F., Stockdale, T., Richardson, D. and Pappenberger, F., 2013, The extreme forecast index at the seasonal scale. Atmosph. Sci. Lett., 14: 256–262. doi: 10.1002/asl2.448

Tropical cyclones

We have developed specialised products for tropical cyclone forecasts.

All tropical cyclones present at the initial time are tracked throughout the forecast. As well as the individual forecast tracks, products such as the strike probability maps summarise the evolution in HRES and ENS and highlight the uncertainty in the future position and intensity.

Additional products identify the potential for new tropical cyclones to develop during the forecast.

More information is available on the tropical cyclone page


Find out about how we monitor the accuracy of our forecast and check out the latest performance scores here.

Ocean waves

Here are links to wave-related forecast products available on this site. Linked charts are only displayed if you have correct access rights. Find out about who can access our information.

Ocean wave products can be seen in Charts catalogue

Extra tropical cyclones

These web-based products present the location and behaviour of near-surface synoptic scale features, such as warm and cold fronts and low-pressure systems, in both the HRES and the ensemble. An objective algorithm identifies and tracks cyclonic features through the forecast.

extra-tropical cyclone products page