Accuracy of Grib weather data

ParaHandy

ParaHandy

Well-Known Member
Joined
18 Nov 2001
Messages
5,210
Visit site
It's peeing down here and I've nothing better to do than compare the Sandettie Light ship actual wind data with the Grib predictions. The error's worryingly big and I'm sure was never this bad 8 years ago when similarily stuck for something to do.

They issue these things every 6 hours (from NOAAA?) about 5 hours after the forecast date; so the first Grib file might be for 0600 and you get this at 11 so the blighters have at least got the chance to look out the window and check they've got one Grib right ...

Seriously, I don't recall it ever being as bad as this?
 
I don't use raw GRIB data so this might be a daft question.

What altitude is the GRIB data for and is Sandettie Lightship at that height?
 
ParaHandy said:
It's peeing down here and I've nothing better to do than compare the Sandettie Light ship actual wind data with the Grib predictions. The error's worryingly big and I'm sure was never this bad 8 years ago when similarily stuck for something to do.

They issue these things every 6 hours (from NOAAA?) about 5 hours after the forecast date; so the first Grib file might be for 0600 and you get this at 11 so the blighters have at least got the chance to look out the window and check they've got one Grib right ...

Seriously, I don't recall it ever being as bad as this?
Click to expand...

We had 50kts in Chi this weekend Consistent with the shipping forecast but far more than gribs or wind guru suggested.

I've got my own guesses on why this was but I hope a definitive answer crops up in this thread.
 
ParaHandy said:
It's peeing down here and I've nothing better to do than compare the Sandettie Light ship actual wind data with the Grib predictions. The error's worryingly big and I'm sure was never this bad 8 years ago when similarily stuck for something to do.

They issue these things every 6 hours (from NOAAA?) about 5 hours after the forecast date; so the first Grib file might be for 0600 and you get this at 11 so the blighters have at least got the chance to look out the window and check they've got one Grib right ...

Seriously, I don't recall it ever being as bad as this?
Click to expand...

The "grib predictions" are presumably the GFS model, which you download in GRIB format? Grib is just a file format. The GFS always seems to predict about ten or twenty percent low in wind speed and is particularly bad at seeing local effects such as the acceleration through the straits of Dover or in the Western Solent. You get used to it.

Your remark about looking out of the window was obviously a joke but as I'm sure you know there is no human intervention in the US model, unlike the Met Office forecast.
 
Last weekend a good example of tracking a vigorous fast moving depression up from Biscay and away over Anglia to the North Sea, using nothing more than the Met offices' links to 'Latest observations' . Downloadable and 24hr 'at a glance' for each station are all right there, if you so wish.
Doesn't answer the OPs Grib file discrepancies, sorry.

Boring answer: it is wonderful to me ( and prudent ) to be able to compare in real time the predicted and the actual.
 
Gribs sometimes offer windspeed at various altitudes, but the default is the same as many other forecasts, 10 metres.

And, yes, gribs offer under-predict windspeed, especially above 15knots or so. Frank Singleton (who knows a bit about this stuff and will no doubt be along later) suggests one Beaufort point is typical, and two far from uncommon. In other words, if you see 20knots forecast, be prepared for 30.
 
The Sandettie anenometer is at a height of 10mtrs and should be comparable with Grib data.

The errors (for the last 4 days) are never less than 1 and quite often 2 Beaufort divisions and the Grib always predicts less than Sandettie. The Grib source is via WeatherTrack on iPAD.

However, the Greenwich Lightship which is west of Sandettie and in relatively more open water is inaccurate at higher wind speeds, within 1 Beaufort division less, and quite accurate at low wind speeds.

I'm thinking it could be the position of the lightships in relation to the land - that seems at first sight obvious - but might have as much to do with Sandettie's proximity to the centre of the low, the position of which they're not getting accurate enough.
 
oldvarnish said:
It's very difficult for computer models to deal with coastal waters because of local effects. But on the high seas I have found the above to be roughly true.
Click to expand...

Same here - actual is usually 5-10 kts above GRIB wind arrows once forecast is above 15kts or so. Then it's about how many reefs.

However of particular importance is wind direction which is about VMG and comfort - GRIBs seem to be OK on this.
 
oldvarnish said:
It's very difficult for computer models to deal with coastal waters because of local effects. But on the high seas I have found the above to be roughly true.
Click to expand...

My experience somewhat supports that, and it's entirely to be expected since local/topographic effects are rarely modelled in grib forecasts. But even in the Atlantic significant wind speed are often under-stated. This is the opposite of course, of Met Office/Shipping Forecasts which tend to err on the side of caution.
 
I tend not to worry about the wind speed but the direction is generally reasonably accurate and therefore valuable, don't you think?
I used to get the wind forecasts from the Coastguard but are there any left now and will the chappie in Lisbon or wherever we have subbed this out to, really know what is happening in the Sound of Islay?
One thing is for sure - none of them will sound as bodice rippin' downright seductive as the lassie fae Stornoway. I wonder what happened to her.......
 
claymore said:
One thing is for sure - none of them will sound as bodice rippin' downright seductive as the lassie fae Stornoway. I wonder what happened to her.......
Click to expand...
The thought of you drooling into the mouthpiece of yon valve powered radio and electrocuting yersel .. what a way to go
 
macd said:
Frank Singleton (who knows a bit about this stuff and will no doubt be along later) suggests one Beaufort point is typical, and two far from uncommon.
Click to expand...

I think I understand the resolution (and other) limitations of GFS, but I don't see why the error would always be an under-estimate or why it would be worse with a stronger wind. It seems generally agreed to be the case, but I can't see why. Also I don't see why if it's accepted that GFS is 1 force under over 15 knots they don't just add a 1 force offset under those circumstances.

Frank Singleton has said this but it hasn't helped: Try my http://weather.mailasail.com/Franks-Weather/Grid-Length-Resolution with the example of the Isle of Man. Or simply draw a shape on paper, see how many grid points you need to define the shape reasonably well. Compare the ratio of grid spacing to the size of the shape. Do simple Y or N for inside the shape or outside.

I'm going to hit google, in the meantime can someone spell it out to me?
 
Last edited:
I must admit I rarely place that much faith in a single GRIB based forecast. On the other hand if you the same or similar forecast for your period of interest for the preceding 3 days then I place some faith in the forecast. If you compare different GRIB based forecast services you will often see different opinions
 
I use GRIBs as a 'broad outline'. Up to three days out they're a pretty good indication of strength and direction although strength always seems a bit underplayed. More so in open waters than coastal.

You need to use them over a period of time and you get to know their virtues and their faults. Once you've done that, they can be quite valuable.

But using GRIBs alone you are unlikely to make as good as a forecast as one made by a professional with skill and long experience.
 
maxi77 said:
If you compare different GRIB based forecast services you will often see different opinions
Click to expand...
I thought in the UK they were all the same forecast now. NOAA dish it out free from an ftp address every 6 hours? There was a long time ago an internet forecaster who gained quite a good reputation on here but there was a copyright dispute when the company split up into two, both with the same name. I would have thought it not that easy to, say, shift the centre of a low on NOAA's data.

The French Meteo site has grib data which being French, won't have been touched by an American hand. I heard it was quite good.
 
I tried for a few days a while ago to log actuals with predictions +24, 48, 72,96 & 120 from a synoptic & gribs against sandetti. But gave up really quickly :o Not an easy task, it was to see how close I could get from the synoptics. Which wasn't too bad for couple of days, but not great.

But one thing did come out of it, if you don't write it all down then forget it. Memory just isn't up to the job. You see what you want.
 
You must log in or register to reply here.

Share this page

Top