Numerical Weather Prediction (NWP) models and AI to predict extreme weather events
Description
State-of-the-art numerical prediction models (e.g. WRF operative at DICCA) and data-driven AI-based methods are exploited in concert to produce forecasts of extreme weather events with unprecedented accuracy. In doing that, the raw output from NWP models are corrected using meteorological observations coming from both satellite-borne instruments (IMREG system) and lightening detection system (Lampinet network) together with in-situ observations. The former two strategies guaranty a diffuse coverage of the sea areas thus overcoming the problem associated to the sparse character of in-situ measurements. To calibrate NWP outputs, different AI strategies are/will be used, including deep learning strategies such as Convolutionary Neural Network (CNN) coupled to other strategies (e.g. the LSTM strategy) to account for the temporal structure of a sequence of frames associated to evolving-in-time atmospheric states.
Team
Andrea Mazzino, Francesco Ferrari, Mattia Cavaiola, Peter Enos Tuju, Daniele Carnevale
References
P.E. Tuju, F. Ferrari, G. Casciaro, A. Mazzino, The added value of high-resolution downscaling of the ECMWF-EPS for extreme precipitation forecasting, Atmos. Res. 280, 106458 (2022)
F. Ferrari, F. Cassola, P.E. Tuju and A. Mazzino, RANS and LES face to face for forecasting extreme precipitation events in the Liguria region (northwestern Italy), Atmos. Res. 259, 105654 (2021)
F. Ferrari, F. Cassola, P. E. Tuju, A. Stocchino, P. Brotto and A. Mazzino, Impact of Model Resolution and Initial/Boundary Conditions in Forecasting Flood-Causing Precipitations, Atmosphere 11, 592 (2020)
F. Cassola, F. Ferrari, A. Mazzino and M.M. Miglietta, The role of the sea on the flash floods events over Liguria (northwestern Italy), Geophys. Res. Lett. 43, 3534-3542 (2016)
F. Cassola, F. Ferrari and A. Mazzino, Numerical simulations of Mediterranean heavy precipitation events with the WRF model: A verification exercise using different approaches, Atmos. Res. 164-165, 210-225 (2015)