GRIB speak

[Blue5]

Because of the GRIB spacing of ~27 km for the GFS, it cannot capture detail below about 130 - 150 km size. Although there may be a broad area of strong winds, the strongest will be fairly localised. What you will get, if you could get the 27 km data, would, in effect, be values smoothed to that resolution. The strongest winds would be missed. As it is, we only get the data at about 50 km so the smoothing effect is more marked.

Apologies to Frank for lifting this from another thread but in an effort to really understand the benefit of GRIBs. I enjoy the weather threads but struggle to fully understand some of the jargon used.

We keep hearing about the spacing of Grib data being 27km, 50km, 100km so could you explain what this means and why it is important for us.

Oh, and can you explain it in simple terms please

 

[oldbilbo]

Apologies to Frank for lifting this from another thread but in an effort to really understand the benefit of GRIBs. I enjoy the weather threads but struggle to fully understand some of the jargon used.

We keep hearing about the spacing of Grib data being 27km, 50km, 100km so could you explain what this means and why it is important for us.

Oh, and can you explain it in simple terms please

That's going to be a challenge, but you could usefully try here.


This stuff is IMHO best left to serious geeks, but you could also try here for occasional entertainment. It's past of the Pennsylvania State University output, FWIW.

 

[estarzinger]

also a lot of terrific professional (and free) online Met seminars here.

 

[oldbilbo]

also a lot of terrific professional (and free) online Met seminars here.

Now that's useful!

Something else with which to while away the long dark winter nights, dreaming of colourful tropical sunsets, diurnal variation anomalies, long bands of organised cloud and long low swell.....

Thanks, Evans!

 

[Blue5]

Thanks to Oldbilbo and Evans, that should keep me out of mischief for a while.

 

[tcm]

I enjoy gathering the crew around a computer screen with very latest grib, and explaining that our strategy will be to sail IN BETWEEN the arrows, and hence obviously miss the highest winds.

 

[oldbilbo]

I enjoy gathering the crew around a computer screen with very latest grib, and explaining that our strategy will be to sail IN BETWEEN the arrows, and hence obviously miss the highest winds.

So I have this picture in mind of Mojo being gybed downwind, from Somewhere Off Cape Town to Somewhere Near Barbados, every three or four hours - and the Watch-On-Deck briefed to keep a weather eye out for the next GRIB windlane, with points awarded for spotting a Green one, a Yellow one, or a Red one....

 

[franksingleton]

..............................

We keep hearing about the spacing of Grib data being 27km, 50km, 100km so could you explain what this means and why it is important for us.

Oh, and can you explain it in simple terms please

Can I please ask all that, if you have a query that I might be able to answer, to copy to me by email. It is all too easy for me to miss such queries.

In order to predict the weather anywhere for more than hours ahead, it is necessary to know about the weather everywhere. That is why models have to be run on a global scale. The grid is 3D. Computer power determines the size of the grid used. At present National Met Services run global models with a grid of around 25 km. For the UK, that means 787456 grid points at each of 70 levels from 20 m above the surface up to about 80 km.


The ECMWF uses a horizontal spacing of about 15 km. They can do this because they do not have to issue short period forecasts. They can afford the extra time and effort to analyse and calculate on that scale because their aim is prediction beyond 3 days. The UK, US, French etc NMSs have to concentrate on shorter periods.

The horizontal spacing defines the size of both weather and topography that can be defined. See my Isle of man example at http://weather.mailasail.com/Franks-Weather/Grid-Length-Resolution. The diagrams there show that any grid can only define both on a scale of about 5 grid lengths. For that reason, the mathematics has to filter out any weather feature of that size or smaller in order to avoid artefacts of the maths as opposed to real weather features.

NOAA calculates on a 27 km grid, and 68 levels but only issues the output on a 1/2 degree grid ie about 50 km. They issue the data at far fewer than 68 levels.

To get better detail, it is possible to calculate on smaller areas at a smaller grid length BUT, these models have to be fed continuously from the global grid around their boundaries. Most regional models eg the USN COAMPS, the US NAM, The UK NAE the model used by Windfinder in its “Superforecast,” all run on about 10 km. However, because the global grid is now around 25 km, these regional models give little extra value. The UK is about to retire the NAE.


There are two thrusts in operational (as opposed to seasonal or climate) modelling. One is to reduce global grid lengths further in order to capture smaller detail at the analysis stage and so reduce the effects of chaos for forecasts up to 10 or so days ahead.

The second is to use fine scale models to improve short period forecasts, particularly of heavy rain or snow and cloud cover for fog and icy roads warnings. The UK runs its meso-scale model on a 1.5 km grid. If you look at the App on a Tablet you can see that the area is quite small.

There are several problems with these small scale models:

• They have to be fed by global models and so are dependent on the global model accuracy.

• With any movement, weather that is only resolved on the global scale moves into the area where the initial analysis was on a fine scale.

• They cannot deal with convection- you cannot predict that there will be a shower over the Solent 6 hours ahead, let alone over the nab Tower.

• They are capable of predicting the Solent sea breeze coming up both arms of the Solent. Whether they will do so will depend on getting cloud amounts correct. That is difficult.

• They cannot deal with smaller scale sea breezes such as the Tor bay effect. For these. See http://weather.mailasail.com/Franks-Weather/How-Sea-And-Land-Breezes-Form.


The models can deal with topographical effects if the topography is big enough. They cannot replicate winds around small headlands such as Start Point or Portland Bill. They can improve on strengths through straits such as Dover, Bonifacio, Gibraltar and the North Channel.

Sorry for the length. Does that help?

 

[uxb]

Frank, that is very interesting , thanks for posting I'm off to read your website now.

 

[Blue5]

The ECMWF uses a horizontal spacing of about 15 km. They can do this because they do not have to issue short period forecasts. They can afford the extra time and effort to analyse and calculate on that scale because their aim is prediction beyond 3 days. The UK, US, French etc NMSs have to concentrate on shorter periods.

Thanks Frank, very helpful, have bookmarked your site for an in depth read.

Given your statement above about ECMWF, would you give more weight to ECMWF over GFS for the forecast for say day 7 ahead.

 

[franksingleton]

Thanks Frank, very helpful, have bookmarked your site for an in depth read.

Given your statement above about ECMWF, would you give more weight to ECMWF over GFS for the forecast for say day 7 ahead.

ECMWF should be better than the other models because it is able to use a smaller grid length so reducing uncertainty in the initial analysis. It knows about slightly smaller weather features than the others. Objective tests shows that it out-performs the rest. However, and there is always a “however,” in meteorology, whether it will do so on any particular occasion is another matter. It was put to me by a former colleague that ECMWF is about 12 hours better than the rest, ie the ECMWF 84 hour forecast is as good as the 72 hour from the others.