BBC coastal forecast vs inshore waters forecast????

[Elemental]

When describing GRIBs franksingleton wrote

They will tend to under-estimate wind strengths - often by about one Beaufort or 20%.

I'm curious. This is pretty much my experience too. I'm left wondering why, if this is a known attribute of the model, they don't adjust the model to get it right? I know weather sites can takes the model data as source and use it as a part of their analysis. But why don't they just update the model?

 

[franksingleton]

When describing GRIBs franksingleton wrote

I'm curious. This is pretty much my experience too. I'm left wondering why, if this is a known attribute of the model, they don't adjust the model to get it right? I know whether sites can takes the model data as source and use it as a part of their analysis. But why don't they just update the model?

Numerical weather prediction operates on a 3D grid of points. The UK, NOAA GFS and some other global models use a grid spacing of about 25 km in the horizontal and 70 unevenly spaced levels in the vertical. If you think about it, and do a little experimentation, you will see that any grid can only define features on a scale of about 5 grid lengths. As a result, the mathematics has to filter out anything smaller that that ie smaller than about 130 km for the GFS. Otherwise, there would be computational noise.

The result is a general smoothing. Within the grid there will be parts with winds more and some less than the average. A general fudge to increase all winds by 20%, or whatever, would give too strong values in some places. Of course, you have to remember that GRIB files are output from numerical weather prediction programs intended for their own use by national weather services. We get them but they are not designed for our use. How we use GRIBs and what we do with them is our business and not their main concern. It is up to us what we do and how we use them.