Descrizione del progetto
Intelligenza artificiale per previsioni stagionali avanzate
Le previsioni stagionali sono strumenti per le previsioni meteorologiche che aiutano la prevenzione dei rischi legati a condizioni atmosferiche estreme. Se da un lato i progressi compiuti di recente nei metodi statistici e nella modellizzazione numerica hanno migliorato le prestazioni delle previsioni stagionali, la loro utilità rimane spesso limitata, specialmente nelle medie latitudini. Il progetto ARTIST, finanziato dall’UE, migliorerà le conoscenze nell’ambito della prevedibilità del clima su scala stagionale, con l’obiettivo di aumentare le prestazioni dei sistemi di previsione esistenti. Il progetto metterà a punto un modello ibrido statistico e dinamico, sintetizzando un sistema dinamico di previsione stagionale d’avanguardia e un modello statistico basato su tecniche avanzate di apprendimento automatico, concentrandosi sulla previsione stagionale delle temperature estreme in Europa. Questo modello ibrido combinerà il fondamento teorico e l’interpretabilità della modellizzazione fisica con le relazioni predittive spazio-temporali identificate dall’intelligenza artificiale.
Obiettivo
Seasonal Forecasts are critical tools for early-warning decision support systems, that can help reduce the related risk associated with hot or cold weather and other events that can strongly affect a multitude of socio-economic sectors. Recent advances in both statistical approaches and numerical modeling have improved the skill of Seasonal Forecasts. However, especially in mid-latitudes, they are still affected by large uncertainties that make their application often complicated.
The ARTIST project aims at improving our knowledge of climate predictability at the seasonal time-scale, focusing on the role of unexplored drivers, to finally enhance the performance of current prediction systems. This effort is meant to reduce uncertainties and make forecasts efficiently usable by regional met-services and private bodies. A statistical/dynamical hybrid model will be designed through the synthesis of (a) a cutting-edge dynamical Seasonal Prediction System and (b) a statistical model based on advanced Machine Learning (ML) techniques. Such a hybrid approach may become critical to improve climate forecasts, because it combines the theoretical foundation and interpretability of physical modeling with the power of Artificial Intelligence (AI), that can reveal unknown or disregarded spatio-temporal features.
ARTIST will focus on seasonal prediction of temperature hot/cold extremes in Europe, but its scalable nature can make it applicable across a wide range of variables and geographical areas. Besides the employment of AI, a strength of the action stands in the use of local land surface predictors to instruct the empirical model.
The fellowship, which includes a variety of training activities, will be mainly conducted at the Barcelona Supercomputing Centre (Spain), a world-renowned institute for climate predictions and applications. A secondment period is projected at the Max Planck Institute for BGC (Germany), prominent in land studies and ML employment in earth science.
Campo scientifico
Programma(i)
Argomento(i)
Meccanismo di finanziamento
MSCA-IF-EF-ST - Standard EFCoordinatore
08034 Barcelona
Spagna