Current technology developments at Finnish Meteorological Institute

Current technology developments at Finnish Meteorological Institute 

Drawing from initiatives like the European Union Open Data Directive and the WMO WIS2 FMI has formed a consortium which received EU funding to build new technology to foster data sharing across Europe. The RODEO project, a collaboration of 11 meteorological services in Europe as well as ECMWF and EUMETNET, started in 2023 aims to build a one-stop shop for open European meteorological data. The solution will be fully compliant with the WMO WIS2 specification and in a sense be the “WIS2 in Europe”.  RODEO will however additionally include more frequent surface observations, provide a single API gateway to access data from at least all EUMETNET countries and host a visualization tool for viewing the data (Geoweb). Data will be provided between member states and to the public, although some data may be of restricted use. Technically the API will be OGC API EDR. Several implementations of EDR including FMI’s own have already been developed, at least in the Nordic countries. The resulting open-source system will be named MeteoGate and is planned to be in operative use as of 2025 running on ECMWF and EUMETSAT EWC instances. 

FMI’s main forecaster tool, SmartMet Workstation, has been around for over two decades now and it still is without a match with its visualization capabilities. However, being based on an architecture that demands data to be physically on the same computer, it is becoming increasingly difficult to cope with growing numerical weather model sizes. However, the open-source successor Geoweb is on good track to become the next number one forecasting tool. Geoweb works in the browser and utilizes standard OGC interfaces (currently WMS, API Records and API EDR) for data retrieval and does not require local data. This gives flexibility for the backends to run close to the data. This software co-operation originally started by KNMI is now a joint initiative as FMI and Met Norway joined the development. In addition to visualizing data Geoweb will feature for example weather warning and aviation domain functionality as well as TAF and SIGMET production. 

In addition to utilizing AI assistants in software development, AI and machine learning methods are used for postprocessing weather model data in many different use cases such as calibrating the ECMWF and MEPS models with various methods. We are also collaborating with ECMWF, Met Norway and other EUMETNET members to enhance weather forecast quality across Northern Europe, focusing on nowcasting and medium-term predictions. Here the ECMWF Anemoi framework is used to develop and operationalize the stretched grid model (Bris) by Met Norway.