The role of supercomputing in weather and climate
The demand for increasingly accurate weather forecasts, together with the scientific development of atmospheric models, has led over the years to an increase in the complexity of the physical parameterizations of the models themselves. The consequence has been an increase in the spatial resolution of the forecasts, thus implying an even greater demand for the processing and storage resources required to run model simulations.
To be able to guarantee a service as complex and critical as weather forecasting, therefore, requires skills and resources of proven reliability and efficiency, and a proven track record. Cineca‘s experience in this area is long-standing. For more than 30 years, Cineca’s supercomputers have been processing and providing detailed weather forecasts for Italy, particularly for the National Civil Protection Department‘s Central Functional Center for Risk Management and its crisis units, in collaboration with the Regional Meteorological Service of Emilia-Romagna, ARPA Piemonte and the Italian Air Force.
Why Mistral
MISTRAL – an acronym for Meteo Italian SupercompuTing poRtAL – was born as a project coordinated by Cineca, carried out in collaboration with key national stakeholders in the weather sector, such as National Civil Protection Department, Arpae, Arpa Piemonte and Dedagroup.
Funded in the context in the Connecting European Facility call for proposals in 2018, the project aimed to build a national open weather data platform to provide national and international citizens, public administrations, and private organizations with weather data from observational networks, historical and real-time analysis and forecasts. This goal has been realized with the bringing online of the Meteo-Hub platform in 2020.
Mistral Meteo-hub is a data distribution service that allows users to access and download weather forecast data or obtain observed ground data from different regions of Italy.
Also available are Radar data made available by the Civil Protection Department and observed data from the MeteoNetwork, a nonprofit amateur network with the mission of spreading knowledge in the field of meteorology and climatology.
Seventeen regions and autonomous provinces (Emilia-Romagna, Piedmont, Latium, Campania, Umbria, Liguria, Autonomous Province of Bolzano, Marche, Sardinia, Veneto, Sicily, Lombardy, Autonomous Province of Trento, Molise, Basilicata, Apulia and Calabria), have joined the project which aims to aggregate, harmonize and post-process all observed weather data from ground stations nationwide. Observed data from ground stations in the regions themselves are made available in open access mode.
The data are available in terms of gridded fields, probabilistic products (such as rainfall forecasts for flood prediction) or point time series from the Italian operational forecast modeling chain and post-processing fields (such as storm probabilities).
International collaborations
The international partner is the European Centre for Medium-Range Weather Forecasts, ECMWF, which is in charge of implementing the Italy Flash Flood service, an application for Italy based on recent advances in post-processing on HPC (High Performance Computing) of ensemble forecasts, and in particular of precipitation to identify flood episodes.
The example in the following video shows the forecast for May 15, 2023 and the following 48h, during which the territory of Emilia-Romagna was affected by the second extreme event in sequence in less than twenty days, with monthly cumulative precipitation exceeding 450 millimeters at various locations. The ongoing event from midnight May 15 to May 17 caused the overflow of 21 rivers and widespread flooding in 37 municipalities.
The goal of Mistral and the Meteo-Hub portal is to facilitate and foster the reuse of data sets by the meteorological community, as well as its inter-area communities, to provide value-added services through the use of HPC resources, and by fostering the creation of new business opportunities.
The MISTRAL project activites have been performed thanks to the INEA/CEF contribution under the grant agreement number: INEA/CEF/ICT/A2017/1567101 – Action No: 2017 IT IA-0144