Why NWP models aren’t very good beyond T+144

I download a lot of NWP images from the internet every day, so I thought that I would come up with a way of looking at the evolution of a particular feature by displaying forecast charts for each run of a particular model, looking first at T+240 right though to T+24. Here’s how the ECMWF model forecast today’s vigorous low. As you can see it really wasn’t till the T+144 that the model got a firm grasp on how this feature would eventually look.

2 thoughts on “Why NWP models aren’t very good beyond T+144”

  1. Robbie Royson

    Not really sure this is telling me why NWP models aren’t very good beyond T+144. And in any case, it depends what you’re wanting to know. If you’re not happy unless it’s got the weather spot on at Strathpeffer, then don’t even bother looking at T+0 sometimes. However, if you’re content with the knowledge that the UK as a whole is likely to be under a cyclonic, unsettled regime tomorrow, then even the forecast at T+192 was pretty decent.

    1. I did a survey a number of years ago and verified a lot of NWP forecasts at different timestep against the eventual analysis and always realised that beyond T+120 you were entering the land of science fiction.
      I suppose this bit of functionality I have now added to compare NWP forecasts for a specific day against multiple earlier runs has just reminded me of that fact.

Comments are closed.

Scroll to Top
%d bloggers like this: