New Weather and Wave Models Now Available
New Weather and Wave Models Now Available
MeteoFrance WAM, GFS Waves, KNMI, DMI, UKMO
In the past three months, Open-Meteo has expanded its range of weather and wave models.
Weather models: HARMONIE from KNMI and DMI
Both KNMI and DMI have launched updated weather forecasts based on the HARMONIE model. This model is developed and operated under the ‘United Weather Centres West’ (UWC-West) initiative, a collaboration between the National Meteorological Services of Ireland, Denmark, Iceland, and the Netherlands. The system is powered entirely by Iceland’s renewable energy sources, such as hydropower and geothermal energy, ensuring low operational costs and a minimal carbon footprint on the path to net-zero.
The HARMONIE model provides coverage across Central and Northern Europe, with a high resolution of 2 km and a forecast range of 2.5 days. It's designed to focus on localized but impactful weather events, such as extreme precipitation (e.g., cloudbursts) and fast-developing storm systems. This makes it particularly useful for forecasting thunderstorms in the short term. The model's boundary conditions are supplied by the ECMWF IFS model. To offer a consistent 10-day forecast, Open-Meteo combines the first 2.5 days from HARMONIE with ECMWF IFS predictions.
The HARMONIE model is supplied by the Danish and Dutch meteorological services, DMI and KNMI, with each offering slightly different configurations. For instance, KNMI covers a smaller region, focusing on the Netherlands and Belgium, but updates the forecast every hour. In contrast, DMI provides broader coverage but updates every three hours.
Both versions offer a wide range of weather variables, including solar radiation and upper-level wind forecasts, making them particularly suitable for forecasting renewable energy production. Additionally, the KNMI and DMI models are integrated into Open-Meteo’s “Best-Match” model for forecasts in the Netherlands, Belgium, and Northern Europe.
Detailed information for both models can be found in the documentation for DMI and KNMI.
Weather models: UK MetOffice
The UK MetOffice has once again made its weather forecasts available as open data! They offer a global weather model with a resolution of 0.09° (~10 km) and a 7-day forecast. In addition, the UKV local area model provides a higher resolution of 2 km for Great Britain and also covers parts of mainland Europe.
The global model is computed every 6 hours, while the local model updates hourly but offers only a 2-day forecast. As with other models, Open-Meteo merges both to deliver a seamless 7-day forecast.
However, a notable limitation is the 4-hour delay in distributing the open data, which slightly impacts forecast accuracy. For this reason, UK MetOffice models are not automatically included in Open-Meteo's "Best-Match" model, which instead uses the ICON model from the German Weather Service.
Users have the option to manually select UK MetOffice models and compare forecasts using the Weather Forecast API. However, Great Britain is already well-covered by a variety of weather models from different independent weather services. This diversity of models allows for improved uncertainty analysis, especially for critical applications. You can personally compare the 11 most prominent weather models available for London.
Wave Models: MFWAM, GFS Waves, ECMWF WAM
The Marine Forecast API has been expanded with new wave models from MeteoFrance, provided through the Copernicus program, as well as models from ECMWF and the U.S. weather service NOAA.
Historical data for MeteoFrance's MFWAM and Currents models has been integrated starting from October 2021. While GFS Waves also offer public archives, only data from June 2024 onward is currently available.
Both historical data and forecasts from all wave models can be directly compared using the Marine Forecast API.
ECMWF Open-Data Survey
Lastly, ECMWF is seeking your feedback to help shape the future of their open-data offerings. As one of the leading innovators in weather forecasting, ECMWF is widely regarded as providing the best medium-range forecasts. While many national weather services operate their own models, they still rely on ECMWF for model assimilation and boundary conditions. Additionally, through the Copernicus program, ECMWF supplies historical reanalysis weather data, which is vital for climate research and mitigation efforts.
Although ECMWF is funded by European member states, its high-resolution forecasts at 9 km are available only to paying customers, with strict limitations on redistribution. However, in recent years, ECMWF has increasingly moved toward open data, offering forecasts at a global resolution of 0.4°, and since early 2024, at a 0.25° global resolution.
This shift toward open data is being driven in part by the European Commission, which now mandates that European weather services make high-value datasets, such as numerical forecasts, radar, and satellite data, available to the public under open-data directive.
While offering open-data creates significant opportunities, it also comes with considerable costs for servers, bandwidth, electricity, and the specialized staff needed to maintain the infrastructure. Weather services now face the challenge of balancing easy access with managing these costs. Understanding which datasets are most valuable and how users interact with the data is important in shaping the future of ECMWF’s open-data distribution.
Open-Meteo encourages everyone to participate in the ECMWF open-data survey: ECMWF Open-Data Survey 2024.
Open-Meteo already integrates all available open-data IFS forecasts from ECMWF at the current 0.25° resolution, as well as high-resolution 9 km IFS data for historical weather, with a 48-hour delay to comply with open-data requirements. We are excited about the future, anticipating greater access to ECMWF forecasts, including more weather variables, higher spatial and temporal resolution, extended-range forecasts of up to a year, and high-resolution seasonal predictions to better anticipate droughts or storms. We are proud to help make weather data more accessible to the public.
If you have 5 minutes to spare, please consider participating in the ECMWF open-data survey. Your feedback is greatly appreciated!