ECMWF accuracy degrade?

[Diatom]

Feedback FYI:
Lately (at least a month) I note a distinct disconnect between forecast/actual winds/waves in my area (Salt Spring Island BC Canada; ECMWF model). Concerning in both directions: dead-calm forecasts are stormy, horrific winds/wave predictions are smooth as glass. Use: 30' boat Georgia Strait crossing judgement. This follows a year of "pretty fair" forecasts with reasonable correlation between prediction/actual using ECMWF.

Weather during the period of concern has been seasonally atypical - "March weather in May/June".

[idefix37]

@Diatom

I do not think that the ECMWF model is degraded. But the type of weather you've had recently was probably hard to predict for this model. You're talking about stormy weather. The consequences of convection are not easy to predict for a global model.
Have you checked the higher resolution models like NAM, or better, HRRR that are available on your island?
And concerning the wave forecast it is powered by the wind data from the model, so the problem is linked.

From time to time, you could check the reported wind vs the forecast at this airport weather station to get a better idea about the prediction errors.