AI Weather Forecasts from an Operational Forecasting Perspective

Oral Presentation 

The rainbow each individual sees is wonderfully unique. Your eyes are seeing a slightly different refraction than mine even if we're standing next to one another, and we each observe colour differently. 

In the same way, our view of AI weather forcasts will be different: how we see them will likely depend on our scientific background, and personal AI experiences. 

Mine is coloured by my doctoral thesis on distinguishing initial condition error (chaos) from model error in weather forecasts, and by being an operational weather forecaster for almost two decades. During this talk I will share the refraction of AI weather forecasts I see, by showing examples from actual forecast situations where AI models outperformed traditional NWP output, where AI output changed (flipped) with the NWP models, and where the two approaches differed but both had their own errors. 

The limitations of the current AI setup, and how they could improve over the next years, will also be discussed.