Which forecast to believe?

Remember that most of the forecasts coming from unofficial sources (eg not UK Met Office/Météo France/Met Eireann/DWD.) will simply be the USGFS presented under different formats and with no amendment in any way. They generally give reasonable guidance.

Forecasts from National Met services in text form will have the benefit of some human interpretation. They must mention possible hazards and, therefore, may seem a little on the pessimistic side although that is not their intention.

On national Met service sites there are objective forecasts eg http://www.metoffice.gov.uk/public/weather/forecast/weymouth-dorset#?tab=fiveDay&fcTime=1388793600. Whether UK, Météo France or Met Eireann etc, these should be better than the GFS. They will be based upon more detailed models, using more detailed data and finer resolution topography.
 
Last edited:
Bear in mind that gusts can be 40% higher than the average wind speeds in the forecast.

I've taken the higher end of each Force to be on the safe side.

F4 10kts gust 14kts
F5 16kts gust 22.4kts
F6 27kts gust 37.8 kts
F7 33kts gust 46.2 kts

You can see where the problem starts when a F5 goes to a 6.
 
Fantasie 19 said:
Yes... http://www.metcheck.com/HOBBIES/inshore_forecast.asp?zipcode=PO9 1RD
Click to expand...

Almost certainly, this is just another presentation of the GFS. As far as I know, the GFS and the Canadian model are the only two global GRIB outputs that can be obtained at no cost. NOGAPS IS not now available and I do not know if its replacement, NAVGEM, is freely available. Can anyone tell me otherwise?

As a consequence, anyone wanting to produce automated forecasts, as all these various services – MetCheck. XCWeather, Passageweather, WindGuru etc – will use the GFS. As a model, its results rank slightly below the UK which ranks slightly below the GFS. But, for the purposes of producing such forecasts, there will be little to choose between any.

PS. I meant to say "uk ranks slightly below the ECMWF"
 
Last edited:
viago said:
…………………..

best bet is probably mapping the shipping forecast, and keeping an eye on your barometer. a bit old hat but it seems to work. or even just play golf.
Click to expand...

OK for the next few hours. Little use for the next 7 days. What has impressed me over the past three months is that the Met Office is having enough confidence in its forecasts to give 6 or 7 ways warnings.
 
Thanks everyone. The upshot was that the weather was worse than all the forecasts and discretion became the better part of valour.

We ended up parking at Worth Matravers and walking round St Abans Head. It was interesting to see the race from a different perspectve, so all was not wasted.
 
franksingleton said:
Almost certainly, this is just another presentation of the GFS. As far as I know, the GFS and the Canadian model are the only two global GRIB outputs that can be obtained at no cost. NOGAPS IS not now available and I do not know if its replacement, NAVGEM, is freely available. Can anyone tell me otherwise?

As a consequence, anyone wanting to produce automated forecasts, as all these various services – MetCheck. XCWeather, Passageweather, WindGuru etc – will use the GFS. As a model, its results rank slightly below the UK which ranks slightly below the GFS. But, for the purposes of producing such forecasts, there will be little to choose between any.
Click to expand...

I never knew that.. interesting, thanks...

I'm a fan of WindGuru and they do indeed use GFS, but they also use two flavours of "WRF" one is 9Km and one 27Km - always wondered what the difference is between the two?? I'm guessing WRF is also free??
 
franksingleton said:
As a consequence, anyone wanting to produce automated forecasts, as all these various services – MetCheck. XCWeather, Passageweather, WindGuru etc – will use the GFS. As a model, its results rank slightly below the UK which ranks slightly below the GFS. But, for the purposes of producing such forecasts, there will be little to choose between any.
Click to expand...

You seem to have MetCheck using the GFS, which ranks below the UK, which ranks below the GFS ... typo?
 
SimonFa said:
We ended up parking at Worth Matravers and walking round St Abans Head. It was interesting to see the race from a different perspectve, so all was not wasted.
Click to expand...

If you're there and there's somone manning the Coastwatch station, it can be worth the effort of knocking on the door and saying hello. :encouragement:
 
SimonFa said:
Didn't think to knock. We waved but they seemed absorbed in what they were doing. Will do next time, thought.
Click to expand...

When we went in they were probably having a quiet day, though it helped when we told them we sailed out of Poole, and passed by at sea level fairly often.
 
Fantasie 19 said:
I never knew that.. interesting, thanks...

I'm a fan of WindGuru and they do indeed use GFS, but they also use two flavours of "WRF" one is 9Km and one 27Km - always wondered what the difference is between the two?? I'm guessing WRF is also free??
Click to expand...

The WRF software is freely available.

Windguru, Windfinder and, as far as I know, most of the "unofficial" modellers use it. They should be able to model some topographic effects on a scala of about 5 grid lengths. They should be able to model weather on the same scale. NOTE many people confuse grid length with resolution.

What none of these people tell you is that they do not use any weather data to initialize their models - apart from the GFS output on a 0.5 degree grid ie about 50 Kim.

When using Windguru, remember that their forecasts are interpolations between grid points.
 
franksingleton said:
The WRF software is freely available.

Windguru, Windfinder and, as far as I know, most of the "unofficial" modellers use it. They should be able to model some topographic effects on a scala of about 5 grid lengths. They should be able to model weather on the same scale. NOTE many people confuse grid length with resolution.

What none of these people tell you is that they do not use any weather data to initialize their models - apart from the GFS output on a 0.5 degree grid ie about 50 Kim.

When using Windguru, remember that their forecasts are interpolations between grid points.
Click to expand...

Thanks... so in laymans terms (I'm a dullard) what is the significance of the 9Km and 27Km variations?? Is it purely to do with distance from the location selected? Or level of detail/certainty??
 
Fantasie 19 said:
Thanks... so in laymans terms (I'm a dullard) what is the significance of the 9Km and 27Km variations?? Is it purely to do with distance from the location selected? Or level of detail/certainty??
Click to expand...

Essentially, level of detail. In principle, a numerical weather prediction model can only define weather on a scale of about 5 grid lengths. Because of that the mathematics filters out features smaller than that size in order to minimise spurious developments.

In practice, the ability of models is limited by the initial analysis. As far as I know, having asked several directly, none of the "unofficial" firms (eg Windfinder, Windguru, Predictwind etc) uses observational data to produce a detailed analysis. They rely upon the GFS output on a 0.5 degree (~50 m) grid. Therefore, their ability to define weather on the 5 x grid length scale is compromised.
 
franksingleton said:
Essentially, level of detail. In principle, a numerical weather prediction model can only define weather on a scale of about 5 grid lengths. Because of that the mathematics filters out features smaller than that size in order to minimise spurious developments.

In practice, the ability of models is limited by the initial analysis. As far as I know, having asked several directly, none of the "unofficial" firms (eg Windfinder, Windguru, Predictwind etc) uses observational data to produce a detailed analysis. They rely upon the GFS output on a 0.5 degree (~50 m) grid. Therefore, their ability to define weather on the 5 x grid length scale is compromised.
Click to expand...

Thanks Frank - feel like I've learnt something... appreciated
 
You must log in or register to reply here.

Share this page

Top