Windguru: how good?

For those of us who sail outside of the UK, the Met Office does not help!

I am currently based in Croatia. I use PredictWind as it has two algorithms - so I choose the one I like the best. Allegedly used by VOR and Americas Cup, so cannot be that inaccurate

TudorSailor
 
I have used Windguru for a few years now for forecasts for Sweden, Finland and Estonia.

As someone said 'Good compared to what?'.

Compared to not being able to get navtex beacuse I'm behind an island, or forgetting to listen to Swedish or Finnish VHF forecasts because I was busy, Windguru is excellent.

Looking at it from my desktop computer in the English Midlands, trying to decide whether i wish i was on the boat or am pleased to be at home, again excellent.

Compared to Swedish official sea area forecasts on VHF or internet, or the Finnish equivalent, probably it isn't so good as there's no human forecaster intermediation, but it's a second string. If two forecasts say the same, then perhaps I can have more confidence in them. Like the pollsters before the last general election . . .

I use Windguru or one of the grib mapping programs on my tablet as the fancy takes me. If I can get Windguru I can get local sea forecasts on the internet too.

One important difference is that official sea forecasts don't try predicting a long way ahead. No doubt if they did they'd be not much more reliable than Windguru. It isn't fair to Windguru to compare its forecast (that is, a computer guess) about the wind strength seven days out with Sven Tonderstorm's forecast from the national Met Office for 12 to 24 hours ahead. They aren't trying to do the same thing.
 
I have used Windguru for a few years now for forecasts for Sweden, Finland and Estonia.

As someone said 'Good compared to what?'.

Compared to not being able to get navtex beacuse I'm behind an island, or forgetting to listen to Swedish or Finnish VHF forecasts because I was busy, Windguru is excellent.

Looking at it from my desktop computer in the English Midlands, trying to decide whether i wish i was on the boat or am pleased to be at home, again excellent.

Compared to Swedish official sea area forecasts on VHF or internet, or the Finnish equivalent, probably it isn't so good as there's no human forecaster intermediation, but it's a second string. If two forecasts say the same, then perhaps I can have more confidence in them. Like the pollsters before the last general election . . .

I use Windguru or one of the grib mapping programs on my tablet as the fancy takes me. If I can get Windguru I can get local sea forecasts on the internet too.

One important difference is that official sea forecasts don't try predicting a long way ahead. No doubt if they did they'd be not much more reliable than Windguru. It isn't fair to Windguru to compare its forecast (that is, a computer guess) about the wind strength seven days out with Sven Tonderstorm's forecast from the national Met Office for 12 to 24 hours ahead. They aren't trying to do the same thing.


Your approach is sensible and pragmatic. I always like to remind sailors that the GMDSS forecasts are a “must have.” Always see what they say. How you use the information is a matter of your own judgement.

Automated forecasts have their uses. They can help to explain the necessarily brief GMDSS texts. To me, their main value is that longer look. I am currently trying to look ahead for the next week and have to make some fairly critical decisions.

Where Windguru and other services are just representations of the GFS then they cannot be better. They will be as good.

When other organisations (ie not National Met Services) run other models, often of finer scale, they are unlikely to be better than the GFS and may well be not as good. Primarily, this is due to lack of data to produce the finer scale analysis. Partly it is a question of what they do around the boundaries of their forecasts areas. This is particularly the case when the weather is mobile. Partly it comes down to predictability in a deterministic sense of detail smaller than can be represented by the GFS.
 
I normally find windguru very good. However, I am currently in Crinan, sheltering from what windguru, xcweathrt, the shipping forecast and the inshore waters forecast all agreed would be an easterly gale. For most of the day we have had gentle westerlies. It's flat calm right now. Windguru says it's gusting to 24kt.
 
I usually use XC Weather rather than Windguru as I prefer the layout. However I treat them all with a pinch of salt. I have been caught like JumbleDuck sheltering from non-existent bad weather and also was once persuaded to go from Loch Fyne to Campbelltown on the basis of a prediction of light south-easterlies and fair weather for rounding the Mull next day. We woke up to near-gale westerlies and pouring rain.

By the way Jumblie, here (mid-Renfrewshire) it was easterly all day today, varying from SE to NE and about 3 to 4. Heavy showers though so I am glad that we curtailed our sail yesterday to a day-sail.
 
I wonder if it's the weather that's unreliable, rather than the forecasts?

Round here in the Midlands people relaying the land weather forecast in the office or pub say 'They've given out . . .' So for example;
Q: Anyone heard the forecast as i'm planning a barbecue this evening?
A: They've given out rain.

I love this way of describing it. It always sounds as if a guy at the Met Office has allocated us his personal choice of weather for the day.
 
I
I wonder if it's the weather that's unreliable, rather than the forecasts?

Round here in the Midlands people relaying the land weather foretcast in the office or pub say 'They've given out . . .' So for example;
Q: Anyone heard the forecast as i'm planning a barbecue this evening?
A: They've given out rain.

I love this way of describing it. It always sounds as if a guy at the Met Office has allocated us his personal choice of weather for the day.



One of my more pithy statements is that "Weather does not know itself to within one Beaufort force." There are limits to deterministic predictability. For planing ahead, my technique is to keep looking at the forecast for the next 7 to 10 dYs. Look for consistency 6 or 7 days ahead. That is a good guide to reliability of the general pattern.
Detail is another mater. Do not believe any one who overs you good detailed forecasts, whether for one or seven days ahead.
 
Top