OpenIFS is an easy-to-use version of the ECMWF IFS (Integrated Forecasting System) developed by the European Centre for Medium-Range Weather Forecasts from medium range to seasonal timescales. OpenIFS aims to develop and promote research, teaching, and training on numerical weather prediction (NWP) and NWP-related topics with academic and research institutions.
OpenIFS provide global meteorological forecasts aiming to show how the weather is most likely to evolve at globe scale. It contains an atmospheric model capable to run at various resolutions (horizontal and temporal) appropriate to the forecast length. At the Barcelona Supercomputing Center, scientists are using the OpenIFS as the atmospheric component of a coupled climate model, EC-Earth. These complex climatemodels can help them to predict the climate for the next months and seasons.
High-resolution models produce large quantities of data with stressing the storage resources making the I/O phase a critical performance bottleneck. In the original OpenIFS version distributed by ECMWF, the forecasting results are gathered on a master process and written as GRIB files. However, such an approach does not scale when increasing the resolution and thus the original architecture was replaced with a distributed approach. Now each worker independently writes its subdomain results into a standard release of a NoSQL database, without additional synchronizations. Currently, the Barcelona Supercomputing Center are replacing this standard release with their own version that includes their distributed multidimensional system, Qbeast.
Qbeast can simplify the visualization of the results. Indeed, to inspect the forecast of a particular region all data needs to be stored, organized and then filtered. Qbeast takes care of organizing and storing the data with a space-driven approach that improve analytics and the retrieval of information. Also, thank its efficient data sampling capabilities, scientists can interactively explore the results using graphical representations that can increase or decrease the level of details at will.
Qbeast and the developing of scientific applications that use its technology are partially funded by the EIC FET-Open programme within the project Quake (801331).
