Assessing the ability of a stretched-grid deep-learning weather prediction model to capture physical balances

Weather forecasting has traditionally relied on Numerical Weather Prediction (NWP) models, which simulate weather
by solving the governing fluid equations. Recently, the emergence of Deep Learning Weather Prediction (DLWP)
models has opened a new era in weather forecasting, offering a data-driven alternative to classical NWP approaches.
Regional DLWP models such as the stretched-grid model Bris developed by Met Norway, have demonstrated perfor-
mance on par with, or even slightly better than regional NWP models across a range of standard forecast metrics.
By overcoming the coarse horizontal resolution that constrained earlier global data-driven models, the operational
use of regional DLWP systems now appears increasingly promising. Nevertheless, the performance of such models
during extreme events is generally inferior to that of regional NWP models, and comprehensive evaluations of their
ability to generate physically realistic forecasts are still lacking.

Here, we present a study comparing the physical consistency of the deterministic version of Bris with the control
run of the operational MetCoOp Ensemble Prediction System (MEPS) in forecasting the severe extratropical cyclone
Poly, which hit the Netherlands on 5 July 2023. We examine whether Bris accurately represents deviations from
key atmospheric balances and whether it reproduces expected dynamics of the storm. We show that, despite its
relatively good performance in terms of RMSE, Bris struggles to capture important mesoscale features of the event
and that it significantly disrupts several atmospheric balances. This unrealistic disruption is mainly linked to the
fine-scale noise evidenced in its output fields, which leads to incorrect and unrealistic spatial gradients. The analysis
of the amplitude spectra also reveals that, in the stretched-grid DDM, fine-scale noise coexists with a seemingly
competing smoothing of the same meteorological variables at larger scales. This tendency for large-scale smoothing
is commonly observed in DLWP models trained with MSE loss and we show that it has notable consequences when
forecasting extreme windstorms such as Poly. These results raise critical questions for improving AI-based models
to better represent extreme events and how to ensure physical consistency in their predictions.