Advancing Subseasonal PredIctions at Reduced computational Effort

Descrizione del progetto

Previsioni meteorologiche più precise ed efficienti

I modelli meteorologici computerizzati forniscono previsioni ragionevolmente accurate. Mentre le previsioni stagionali offrono indicazioni sugli schemi meteorologici su larga scala con uno o più mesi di anticipo, quelle sub-stagionali vengono eseguite da due settimane a due mesi in anticipo e offrono approfondimenti preziosi per il processo decisionale di tutte le attività sociali ed economiche. Muovendo da tale presupposto, il progetto ASPIRE, finanziato dal CER, concentrerà l’attenzione sulla prevedibilità intrinseca delle modalità di variabilità convettiva tropicale. L’approccio adottato sarà quello dell’interdisciplinarietà, per sfruttare al meglio la prevedibilità intrinseca di tali modalità e quantificare il valore aggiunto della risoluzione su scala chilometrica locale ai tropici. Utilizzando modelli basati sull’apprendimento automatico, che simulano l’effetto di una risoluzione su scala chilometrica, ASPIRE mira a ridurre i costi computazionali a lungo termine.

Obiettivo

The new frontier for weather prediction is the so-called subseasonal time scale of two weeks to two months ahead. To take preventive measures at an early stage, reliable forecasts on this time scale are becoming increasingly important for multiple socio-economic sectors. Subseasonal predictability can be gained from recurring patterns in the Earth system. ASPIRE will focus on one of these, namely modes of tropical convective variability. Long-standing systematic errors due to the parametrization of processes in numerical weather prediction models prevent the predictability of these modes from being exploited. Simply running models at a resolution high enough to resolve tropical convection is not feasible due to high computational costs. Taking advantage of three recent developments, ASPIRE will explore new ways to better exploit the intrinsic predictability of tropical convective modes without exhausting the currently available computing resources.
The uniqueness of ASPIRE is its cross-disciplinary approach that builds on my experience in atmospheric dynamics and predictability, numerical modeling, and machine learning (ML). First, ASPIRE will identify the source regions and pathways of tropical forecast errors that prevent the intrinsic predictability from being exploited using a new set of subseasonal ensemble hindcasts. Second, ASPIRE will quantify for the first time the added value of locally confined kilometer-scale resolution in the source regions identified before, and generate probabilistic predictions from deterministic forecasts through ML-based post-processing. Third, to enable simulations at kilometer-scale resolution in operations, ASPIRE will develop ML approaches that emulate the integrated effect of the resolved convection in the tropics at substantially reduced costs. If successful, this approach would be a breakthrough towards improved operational weather forecasts at substantially lower computational costs, for a global socio-economic benefit.