Voluntary Observing Ships (VOS) are a vital part of the global weather observation system, providing essential meteorological and oceanographic observations at sea. Weather stations managed by National Meteorological Services and installed on board VOS typically record and transmit data on surface wind speed and direction, air temperature, humidity, sea surface temperature (SST), atmospheric sea level pressure (SLP), cloud characteristics, visibility, wave and swell parameters, and overall present weather status. The VOS data are crucial for real-time weather forecasting and historical climate studies, especially in data-sparse regions like the Arctic. They contribute significantly to our understanding of atmosphere-ocean interactions, which is essential for addressing global warming and improving long-range weather forecasts.
There are two main types of VOS: conventional VOS, where sailors make observations and record data in meteorological logbooks, and VOS with automated systems that use electronic logbooks for streamlined reporting.
The NMS of the Netherlands (KNMI) has developed the Turbowin software package to enhance data acquisition, quality and quantity from conventional VOS.
Turbowin+ software
The number of VOS with automated weather stations (AWS) has increased, resulting in more frequent marine observations, usually with hourly transmissions. In 2024, more than 265 operational AWS units have reported data on the Global Telecommunication System (GTS) from ships recruited by several NMSs. A common AWS, called EUCAWS, was developed in 2017 under E-SurfMar and has been installed by national meteorological services on several VOS since then.
In 2022-2023, E-SurfMar developed a new mini automatic weather station for installation on board vessels. This mini station offers a wide range of possibilities for connecting sensors and sending messages. Power consumption is low, installation is easy, and its cost is low. This station is now operational and about ten have already been installed on European vessels.
Moored buoys are anchored at fixed locations and regularly collect observations from various atmospheric and oceanographic sensors. These buoys measure a wide range of parameters crucial for operational applications, including both meteorological and oceanographic data. The meteorological parameters typically measured by moored buoys are air temperature, humidity, atmospheric pressure, and wind speed and direction. Oceanographic measurements generally include sea surface temperature and, in some cases, ocean currents and salinity. Sea state (the general condition of the ocean surface with respect to wind waves and swell at a specific location and time), is also of significant interest for many applications.
Wave parameters are measured by tracking the buoy’s displacement, assuming it responds to waves without affecting the measurement. Modern moored buoys transmit wave parameters and spectral data through the Global Telecommunication System (GTS). The quality of the data from these buoys is considered amongst the finest available from surface marine stations. These measurements are vital for weather forecasting, maritime safety and scientific research, whilst also serving as crucial reference points for calibrating and verifying satellite measurements and numerical atmospheric and ocean models. Moored buoys can be fitted with additional sensors to meet specific research or operational requirements. Some are capable of measuring parameters throughout the water column, which is the vertical section of water from surface to seabed. This provides valuable data on subsurface conditions at various depths, offering a comprehensive view of oceanic processes from top to bottom.
More than 150 E-SurfMar drifting buoys are permanently operating in the EUMETNET Composite Observing System (EUCOS) area. E-SurfMar relies on partner institutions in The Netherlands, Portugal, UK, France, Italy, Germany, Canada and the United States to deploy and maintain the drifting buoy network.
Deployments are carried out using various vessels, including container ships, expedition sailboats, and racing yachts. Our goal is to optimise our networks for consistent spatial and temporal coverage. Drifting buoys, designed to track surface currents, have a hemispheric shape and include a holey sock drogue (sea anchor) at a standard depth of 15 metres below the surface. These buoys use GPS to measure their position and transmit surface air pressure and sea surface temperature data in real time via the Iridium satellite communication system. By analysing multiple reports from different positions, we can estimate near-surface ocean currents. The average operational lifespan of a drifting buoy is approximately 600 days. While sea surface temperature (SST) is important, we primarily focus on atmospheric surface pressure, as this parameter cannot be measured by satellites.
A drifting buoy with the drogue (credit Lagrangian drifting laboratory)
Over the years, the OPERA community has produced an impressive library of documents, recommendations, and definitions. Many of the early technical documents are now more of historical value.
Selected publications of OPERA 6 (2024 -2028)
ODIM version 2.41 (2024 -for older versions, see software documentation)
Selected publications of OPERA 5 (2019-2023)
– Best practices for calibration and monitoring 25 October 2020
– ODIM version 2.4 (2021)
– OPERA-4: On the coexistence of weather radars and wind turbines, 25 February 2022
Selected publications of OPERA 4 (2013-2018)
– OPERA the Radar Project, article in Atmosphere, 12 June 2019 (https://www.mdpi.com/478188)
– Frequently asked questions about radars, 26 January 2016
– Best practices for calibration and monitoring 19 January 2019
– ODIM version 2.3 (9 January 2019)
– BUFR format and graphical format for polar volume data submitted to and composites produced by the OPERA, version dated 12 March 2012
– The Operational Weather Radar Network in Europe – article in Bull.Am.Met.Soc. 2014 (https://journals.ametsoc.org/view/journals/bams/95/6/bams-d-12-00216.1.xml)
– The Threat to Weather Radars by Wireless Technology – article in Bull.Am.Met.Soc. July 2016 (https://journals.ametsoc.org/view/journals/bams/97/7/bams-d-15-00048.1.xml)
– Maintenance keeps radars running – article in Bull.Am.Met.Soc. 2017 (https://journals.ametsoc.org/view/journals/bams/98/9/bams-d-16-0095.1.xml#:~:text=Most%20radars%20are%20equipped%20with,%2C%20the%20emergency%20power%20generator).
– Quality of the composites – report of a study
– Monitoring rain rates – report of a study
– Daily convective precipitation – report of a study
– Cost-benefit of dual-pol exchange – report of survey in 2016
– Improvement of rain rates using water phase (snow or rain) – report of a study
Selected Deliverables of OPERA 3 (2007-2012)
– Status of use of polarimetric C and S band radars (2012)
– Status of use of polarimetric X band radars (2012)
– Radar site selection and protection (2010)
– Impact study of radar observations by wind turbines (2010)
– Statement on wind turbines (2010)
– Statement of the OPERA group on processing RLAN interferences (2009)
– Recommendation on coexistence with 5 GHz RLAN (2008)
– Harmonised production practices for volume data, low-level reflectivity, and wind profiles (2008)
Over the year, the OPERA community has developed several software packages to exchange radar data from different sources and manufacturers.
Data exchange in OPERA is currently conducted using files in the Hierarchical Data Format version 5 (HDF5), following the internally developed OPERA Data Information Model (ODIM). ODIM has been widely adopted and implemented both within and beyond the OPERA community, and it is continuously updated.
Software documentation:
– ODIM version 2.41 (2024)
– ODIM version 2.4 (2021)
– ODIM version 2.3 (2019)
– ODIM version 2.2 (2014)
– ODIM version 2.1
– ODIM version 2.01
Example file of ODIM 2.4: T_PAZA41_C_LPMG_20241205100002.h5.gz
Between 2019 and 2023, the OPERA data centre (ODYSSEY) was gradually upgraded with three new production lines (CUMULUS/STRATUS, CIRRUS, and NIMBUS).
Figure: The schematic structure of the new OPERA data production.
The three new production lines are:
• CUMULUS/STRATUS, a real-time data hub providing the incoming data to members in close to real-time for their own processing (e.g. production of regional composites or NWP assimilation) and to the other two production lines (CIRRUS and NIMBUS). This production line has been operational since mid-2021.
• CIRRUS produces a 5-minute instantaneous maximum reflectivity composite (dBZ) for current weather, severe weather alerts and aviation needs. It replaced the old ODYSSEY maximum reflectivity composite with improved spatial (now with 1 km gridding) and temporal resolution (from 15 minutes to 5 minutes). The production has been operational since early 2024. Documentation on the CIRRUS maximum reflectivity product and its differences compared to the ODYSSEY maximum reflectivity product can be found here:
– OPERA_Max Reflectivity_Product Sheet_Ed-2.0.pdf
– OPERA5_Report_Cirrus-vs-Odyssey_Ed-2.0.pdf
• NIMBUS produces the quality-controlled products centrally: instantaneous rain rate (mm/h) and 1-hour accumulation (mm) composites, wind profiles (expected in late 2025), and the quality-controlled volume data for NWP assimilation. It replaces the old ODYSSEY production for the above-mentioned products. The NIMBUS production has been operational since mid-2024. Documentation regarding the NIMBUS composites and their differences compared to the corresponding ODYSSEY can be found here:
– NIMBUS_datasheet_composites_1.0_13062024.pdf
– NIMBUS_composite_vs_ODYSSEY_1.0_19062024.pdf
OPERA support can be reached by email at support.opera[at]eumetnet.eu. Responses from the service desk are provided on a best-effort basis.
Access to data
EUMETNET members may use the composites for their official duties without a separate license. Their Numerical Weather Prediction (NWP) modelers obtain the quality-controlled volume data directly from the NIMBUS production line.
The OPERA products are also available under license to Third parties:
– For national meteorological services not participating in the OPERA programme that want to use the products for their official duties, contact info[at]eumetnet.eu
– For a research and education license, contact info[at]eumetnet.eu.
– For a license to exploit the OPERA products commercially, contact one of the EUMETNET members’ contact points (https://www.ecomet.eu/contact/members) or send an email to the EUMETNET Secretariat using the email address: info[at]eumetnet.eu
Microwave radiometers are passive instruments that measure the natural radiation from the atmosphere. These radiation measurements are presented as “brightness temperatures” in Kelvin units. We use the TROPOe software suite to convert the brightness temperatures into vertical profiles of temperature and humidity. These profiles typically cover the first 2 km of the atmosphere. Both brightness temperatures and profiles are disseminated via the Global Telecommunication System (GTS). In the future, they will also be disseminated via the WIGOS Information System (WIS).
Microwave radiometers at Payerne, Switzerland.
E-Profile’s wind network consists of about 30 radar wind profilers, about 100 weather radars providing wind profiles and new Doppler lidars. All radars send processed vertical wind profile data to E-Profile. According to the operating frequency of the radar, the derived wind profiles cover the atmospheric boundary layer (up to about 3 km), full troposphere (up to about 12 km), or troposphere and lower stratosphere (up to about 20 km). Lidar data are processed centrally using the Doppler lidar toolbox developed by the National Meteorological Service of Germany (DWD). The range of the Doppler lidars depends on the presence of aerosols and is typically limited to the atmospheric boundary layer.
E-Profile maintains a list of recommended stations for data assimilation into numerical weather prediction models based on the EUCOS (EUMETNET Composite Observing System) monitoring results and considering data quality and timeliness.
Please contact e-profile[at]meteoswiss.ch to get the current list of the E-Profile stations.
Wind profiler at Lindenberg, Germany.
Data download (CEDA archive, registration required):
Wind profiler network
The aerosol and cloud network consists of more than 450 automatic lidars (light detection and ranging sensors) and ceilometers. The main purpose of the latter is to measure the height of the cloud base, which is important for planning safe navigation in aviation and understanding atmospheric stability and convection that can result in severe weather or favourable conditions for high-air-pollution episodes. The diverse applicability of ceilometer data explains the high density of the ceilometer network. The automatic lidars and ceilometers are centrally calibrated to create profiles showing how much light is scattered back by particles. That light is related to the size and concentration of particles in the atmosphere. Appropriate calibration techniques enable the estimation of the mass concentration of aerosols and their impact on the radiative transfer in the atmosphere.
An interactive interface to analyse advanced aerosol properties can be found here: https://vprofiles.met.no/.
Data download (CEDA archive, registration required):
Aerosol and cloud profiler network
Ceilometer and automatic lidar at Hohenpeissenberg, Germany.
The figure below shows the active GNSS stations currently providing data to E-GVap
In E-GVap, we aim to densify the network of GNSS receivers, providing us with valuable ZTD data. Several EUMETNET members run global weather forecasting models; for that reason, data from outside Europe is also important.
The coverage for the EUMETNET-derived data network from the European Meteorological Aircraft Derived Data Centre for a typical day is shown on the map below. This shows how tens of millions of observations are collected daily, with the highest concentrations indicated by the red colours.
Total Coverage
Number of derived obs in [0.5×0.5]° – 15/01/2025
The coverage for the EUMETNET AMDAR network for a typical day is shown on the map below. Yellow plots indicate observations taken en-route with red and green plots representing aircraft ascent and descent profiles.
Data Publishers such as National Meteorological Services (NMSs) will need to create and operate (potentially in collaboration) several MeteoGate Local components to exploit the system’s benefits. They can choose the level of maturity of the solutions they want to implement – from manual to fully automated. There are collaborative options that allow the reduction of the costs of developing Local components; most of them are needed for NMSs’ obligations towards WMO and its WIS2.0.
– Persistence to store and allow for shared data to be accessed through APIs
– Data Processing in line with WMO regulations and MeteoGate Policies, Standards and Processes (PSPs) to standardise and productise data so that it can be shared using APIs
– Local Broker to send change notifications to the WMO Global Broker of news and updates to discovery metadata and shared data
– Metadata Management to create, publish and maintain the discovery and provenance metadata to share changes with the Global Discovery Catalogue.
– Data Access to expose data to MeteoGate users using interactive APIs through the API Gateway
MeteoGate Community components are needed to deliver shared capabilities, e.g. a data catalogue or access management. They will be established and operated by EUMETNET or its members and useable by data publishers and consumers inside and outside the hydro-meteorological community. 
1- A shared Data Catalogue defining and describing data assets and data collections, indexed, linked, and searchable to support data discovery and usage.
2- A Message Broker with a subscription service for notifications about the updates to both discovery metadata and the data.
3- Identity & Access management to control user and system access to platforms and platform services.
4- API management providing the interface between data consumers and underlying data stores, enabling managed access and usage of data and data products.
5- An Insights service reporting on data discovery and publication transactions for Data Publishers to derive management information.
In the Implementing Regulation of the revised Open Data Directive, the European Union has defined a set of High-Value Datasets (HVDs). These are data that have high value for the community. One of the sets defined by the Regulation concerns meteorological data: weather observations, climate time series, warnings, weather radar data and numerical weather prediction (NWP) data.
Following the regulation, these data should be available free of charge, under an open license, in a machine-readable format using Application Programming Interface (API), and bulk downloadable.
Further information on EUMETNET’s response to the High-Value Datasets Regulation can be found under the RODEO page.
The E-AI programme will enhance the value of earth observations, gridded observational datasets, model fields and further datasets by making them AI-ready. The goal is to define and prepare a collection of datasets and develop appropriate interfaces and tools to make these data easily available for development, retraining, and operational use of AI/ML methods. Image sources: ESA (European Space Agency), EUMETSAT, DWD, Reinhardt et al. (https://doi.org/10.1175/AIES-D-22-0025.1) published under CC BY-SA 3.0 IGO
The E-AI programme is structured in three modules: a) Data Curation, b) Analysis, Modelling and Post-Processing, and c) Products and Services. By working together on overarching Workflows, AI/ML can be used effectively across the entire value chain while also addressing necessary technical, administrative, and structural requirements. Moreover, the programme offers Communication and Training opportunities to ensure an efficient flow of information on AI/ML developments within and beyond the E-AI programme.
The CBCF is a collaborative forecast that provides EUROCONTROL’s Network Manager (NM), participating Air Navigation Service Providers (ANSPs), and other aviation stakeholders with strategic planning information about convective weather across European airspace. NM and the attached ATM-units use this information to coordinate with impacted Air Traffic Control Centres (ACCs) across Europe to minimise flight delays and improve aviation safety. The EUMETNET CBCF module comprises:
– the operational framework to produce a collaborative weather chart
– the organisational environment enabling the production and development of the forecast service in line with users’ requirements
– the designation of roles, responsibilities and interfaces amongst all stakeholders
CBCF participating MET ANSPs
CBCF domain in 2025
Base dataset – Low-resolution gridded data
The benchmark base dataset is a low-resolution gridded dataset that covers the European area from 36 to 67° in latitude and from -6 to 17° of longitude, with grid spacing of 0.25°x0.25°.
The benchmark dataset consists of 47 fields, with at least seven requiring post-processing. It includes daily forecasts at midnight, which feature ensemble, control, and high-resolution ECMWF runs, as well as reforecasts twice a week and hourly analysis data. The training set covers 2017 and 2018, while the test set is for 2019. The total size of the gridded benchmark is about 6 terabytes (TB), with the training set being 4.3 TB and the test set 1.6 TB. This dataset provides a solid basis for evaluating and comparing different post-processing methods used by National Meteorological Services.
More details about the base dataset can be found at https://eupp-benchmark.github.io/EUPPBench-doc/files/base_datasets.html.
EUPPBench datasets
In addition to the base dataset, specific datasets for statistical and machine learning over a smaller domain have been designed. The EUPPBench datasets are available over a smaller area within Europe, stored in Zarr format for easy access. The forecasts and observations datasets are already paired together, providing analysis-ready data for post-processing benchmarking purposes.
The EUPPBench datasets also include station data all over the selected domain and ECMWF forecasts and reforecasts from the nearest grid point.
With these datasets, the first benchmark of several post-processing methods was conducted and published in Earth System Science Data (ESSD). More details on the EUPPBench datasets and this first benchmark can be found at https://essd.copernicus.org/articles/15/2635/2023/.
An example of deterministic IFS forecast for 100m u wind and the corresponding gridded observations in the EUPPBench dataset.
38 European National Meteorological Services contribute to the consistent visualisation of weather awareness information, thereby ensuring coherent interpretation of multi-hazard early warnings throughout Europe.
Visit the MeteoAlarm Website for the latest weather-related awareness situation and further information.
