Weather forecasts - anyone keep tabs on accuracy?

[estarzinger]

This is a quote from the metoffice website saying exactly what I was trying to say . . . This suggests exactly the (confidence/consistency/certainty) analysis I was suggesting was important for cruisers to understand and do when looking at/for weather windows and routing:

"One to two days ahead the uncertainty is usually in the local details of the weather, but several days ahead the whole weather pattern can often be uncertain.

The spread of different forecasts, for a given time, gives the forecaster an indication of the uncertainty of the forecast. Confidence can also be estimated by the consistency between the latest model forecast, earlier forecasts and ensemble forecasts. If the model is consistent then confidence may be high but if it suddenly changes then confidence falls rapidly. In these situations the solutions of other models may be crucial.

Research is ongoing at the Met Office on the use of ensemble forecasts to estimate the inevitable uncertainty associated with weather forecasting and in how to best communicate that uncertainty to forecast users."

 

[franksingleton]

.............
. . . This suggests exactly the (confidence/consistency/certainty) analysis I was suggesting was important for cruisers to understand and do when looking at/for weather windows and routing:..........

I can buy all of that. On various threads on these forums, I have given the same message. At http://weather.mailasail.com/Franks-Weather/How-To-Use-Marine-Weather-Forecasts, written, initially, many years ago, I make that very point about looking for consistency in forecasts. Anyone who has had the misfortune to hear me give talks will have heard the same message.

I have a number of truisms that, again, I trot out from time to time.

1. As long as you keep the theologians away, the only real certainty in any forecast is the date.

2. Weather does not know itself to within one
Beaufort force.

3. Plan ahead for the next 5 or 6 days – but keep updating your plans.

4. There are always uncertainties because of the quality of the data and the models.

5. Chaos is a real effect.


I try not to use the word accuracy preferring to think of forecasts as useful or not, as the case may be. In the example quoted earlier, 37 May last year between Cedeira and Viveiro nobody would claim that a forecast of F 4 by AEMet was good, nor my best guess of F 6. However, the forecasts did tell me that we would have to remain in A Coruna or Cedeira for at least a week. They told me that I would be able to get to Cedeira but would probably have to remain there for several days. In short, I was able to make informed decisions rather than take pot luck.
I shudder when I hear providers claim

…………detailed marine forecast

Outstanding accuracy

Worldwide forecasts using a 1km resolution

……….showing detail every hour!

Superior forecasting …………

Detailed ……. information for your local area

……highest quality weather forecast data

.............precise …..… accuracy of one nautical mile


Scour the web and you will see all those, or something very similar.


I also shudder when I hear sailors claim that this or that service is always accurate – especially when they are comparing with another service which has the same information in a different package.

This all may seem cynical. It is not, I am a realist, I am pragmatic, I try to use intelligence and common sense. I preach that message to others.

Having said all that, my experience after first using GRIbs in 2001, is that NWP has improved reasonably steadily. Using what I say on my site, you can now plan ahead with some confidence up to around 5 or 6 days ahead. I do look at GRIbs to 8 days for interest. In another couple of years, we may well find that the day 8 is as good as the current days 6 or 7. I live in eternal hope. But, I would also expect that to be getting near the limit for useful-to-sailors forecasts.

 

[estarzinger]

I like the idea of getting an actual (grib) to compare against a forecast. Is this facility available on Zygrib or any other viewer, I seem to be struggling trying to find actuals, perhaps I am looking in the wrong place.

NOAA has a big archive of historical global grib forecasts (2010 and 2011) at:

http://nomads.ncdc.noaa.gov/data/gfs-avn-hi/

and back to 2003 at:

http://nomads.ncdc.noaa.gov/data/gfs/

So, if you want to go to some work, you can compare 120hr forecasts to T0 'actuals' to your heart's content.

We use these archives to get the detailed weather pictures when we write articles about our passages (because on board, to limit file size, we are usually getting smaller areas or lower resolution gribs).

 

[franksingleton]

NOAA has a big archive of historical global grib forecasts (2010 and 2011) at:

http://nomads.ncdc.noaa.gov/data/gfs-avn-hi/

and back to 2003 at:

http://nomads.ncdc.noaa.gov/data/gfs/

So, if you want to go to some work, you can compare 120hr forecasts to T0 'actuals' to your heart's content.

We use these archives to get the detailed weather pictures when we write articles about our passages (because on board, to limit file size, we are usually getting smaller areas or lower resolution gribs).

It would help me - and othersm. probably - if you could give an idiot's guide to extracting archived data from these sites.

 

[Blue5]

It would help me - and othersm. probably - if you could give an idiot's guide to extracting archived data from these sites.

+1, didn't like to mention it in case it was my pc skills

 

[estarzinger]

It would help me - and othersm. probably - if you could give an idiot's guide to extracting archived data from these sites.

I am not doing anything clever.

Some of the directories are empty - not sure why. But most of the directories have files in them - the files are big - a complete global grib - around 28mb. They have two types of files. I get the .grb ones. I just click on them and 'save as' to my desk top. Viewfax will choke on that size grib, but zygrib can open them and zoom in. What problem are you having?

Just for fun . . . I just went and got an example from our last passage situation, that I mentioned in a post above:

This was the 5 day forecast (120hr forecast showing Nov 6 000z) when we departed - looks pretty good.
View attachment 15746

And this is the actual (oohr for Nov 6 000z) - TS developed and just a bit windier, and different routing to the bvi required!
View attachment 15749

 

[mcframe]

In case anyone is interested, there are some places on a one day course at CA House, Limehouse Basin, on Sunday 29 January. This is in the list of courses run by the CA - http://www.cruising.org.uk/events/cahcourses.

Booked now - see you there

 

[franksingleton]

Booked now - see you there

Thanks. please let me know who you are on Sunday.

 

[franksingleton]

Last June we were in the Southern Rias wanting to be in A Coruna on or by the 19th. This was to get two people to a railhead to Madrid to fly back to he USA.

Whilst in Bayona, the 170 forecast for noon on June 19 was this

In Ria Muros, the 84 hour forecast was

We went o to Camarinas fully confident that in 3 days time we woul be able to get to A Coruna. The 12 forecast for noon on June 19 was

I would not normall expect a 170 hour forecast to be as good as that but it makes a good exeample of what the GFS can do, even though it will not always do as well.

 

[Bosun Higgs]

Would you cross the Channel on a persistence forecast? Would that tell you about the norther gales that kept you in Cherbourg when you should have been back at work?

I didnt say that the "persistance" method was accurate, just that it was a bit better than your ex employer. The implication of what I wrote was that I wait for reasonably settled weather if doing a long passage. Like you, I dont have to be "back at work" and indeed the only times I have been caught out were in those bad old days when I did have to be back and had to trust the forecast or my guess and chance it.

You may well view the forecasts with a weather man's idea of what "accurate" means. I don't. If the forecast says SW 4 then thats what I expect and SW3 or W4 or even SSW4 are wrong.

 

[franksingleton]

I didnt say that the "persistance" method was accurate, just that it was a bit better than your ex employer..

That is fair enough. Your post worried me because I have heard sailors say they they believe in, and go on the basis of, actuals.

I like to regard forecasts as useful rather than accurate. If the forecast says that I can go in 1, 2, 3 or whatever time and that is so then dine. Whether the actual winds are precisely right does not matter. I have long said that planning ahead and avoidance of being caught out is what is required.

I think that it is highly unlikely that forecasts of wind will ever meet the demands of some sailors. It comes back to my point that the atmosphere is imprecise.

 

[estarzinger]

Quote from Camper (volvo) navigator yesterday: “The navigators are all scratching their heads as the grib files are far from dealing with the present situation," he said. "The winds are 40 degrees out on the current direction."

 

[franksingleton]

Quote from Camper (volvo) navigator yesterday: “The navigators are all scratching their heads as the grib files are far from dealing with the present situation," he said. "The winds are 40 degrees out on the current direction."

So?

Charlatans apart, no forecaster would claim to be highly accurate. I have said as much many a time. No forecaster will deny that there is always some degree of uncertainty. How much will depend on the particular situation. The current UK Met Office forecast for 6+ days says “Although there is a lot of uncertainty........” On other occasions they will be much less equivocal.

Stan Honey says that the way he works is to use GRIBs to determine the course required to get him where he wants to be in N days ahead. He then updates every 6 hours. That is really the same kind of approach that I advocate for the majority of cruising sailors. Use the latest forecast to see where you should be or can be in 5 or 6 days’ time. Have regard for the consistency of the forecasts, whether GRIBs, Synoptic charts or worded texts. Update, at least, daily .

Without looking at the detailed forecasts, knowing just where they are, what they are actually getting – life is too short – I cannot make any sensible comment on what you quote. Other, that is, than to say that I would be surprised if such a degree of uncertainty never occurred. Conversely, I would be surprised were it frequently the case.

 

[estarzinger]

So?

Frank, for some reason you are being defensive and taking this thread personally. I did not post that quote in any way aimed at you. The thread is about weather forecast accuracy, and I was just point out that the volvo navigators were seeing actual wind that was 40 degrees out from the T0 model output. That's seems to me to be both relevant to the thread discussion and useful for cruisers to understand that the models can be that far out even with 'current conditions'.

 

[franksingleton]

Frank, for some reason you are being defensive and taking this thread personally. .............

I am not taking anything personally at all. Reasonable though it was, some people reading your post may jump to the conclusion that forecasts are not useful. There is so much folk lore, quite understandably, because weather is such an imprecise entity. Since my early days pre-somputer forecasting, through the still developing NWP, there have been improvements that we never thought possible.

I am always trying to get it across, to sailors especially, that there is no such thing as an accurate forecast. Nevertheless, used sensibly, forecasts can usually – never guaranteed – keep most of us out of trouble. I make the assumption that “most of us” are coastal ie not blue water sailors. Coastal sailors can be too far from a port of refuge if caught out at sea and without sea room to ride out bad weather. In many ways, they are more at risk from weather than ocean sailors. It is that paradox that concerns me. I see and hear of too many sailors who are really naive in their use and understanding of what forecasts can and cannot do.

 

[estarzinger]

Nevertheless, used sensibly, forecasts can usually – never guaranteed – keep most of us out of trouble.

I don't think many would disagree with that. BUT I will say I have seen people get into trouble believing the forecast too much, when there were subtle signs both by looking outside and in the forecast itself that it should not be trusted. As you say, often the models provide excellent guidance, but my experience is that just often enough they do not so that one must be watchful and careful.

I am currently writing a report (for US safety authorities) and one of these pages has this - the second most important root cause is a poor understanding of the weather (and how to deal with it):

Root Causes of recent safety incidents:

1. Structural and design factors

Boat does not break
Boat does not stay upside down for a long time​

2. Intense Weather

Forecasting and routing​

Heavy weather tactics
Fatigue and attitude management​

3. Navigation & groundings

Keep head out of boat
Chart plotters not 100% accurate​

4. MOB

Most deaths have apparently been with unconscious MOB or trapped​

5. Emergency voice communication equipment

Sat phone proven extremely helpful in offshore experience – ‘after the fact’​

 

[franksingleton]

I don't think many would disagree with that. BUT I will say I have seen people get into trouble believing the forecast too much, when there were subtle signs both by looking outside and in the forecast itself that it should not be trusted. As you say, often the models provide excellent guidance, but my experience is that just often enough they do not so that one must be watchful and careful.

I am currently writing a report (for US safety authorities) and one of these pages has this - the second most important root cause is a poor understanding of the weather (and how to deal with it):

Root Causes of recent safety incidents:

1. Structural and design factors

Boat does not break
Boat does not stay upside down for a long time​

2. Intense Weather

Forecasting and routing​

Heavy weather tactics
Fatigue and attitude management​

3. Navigation & groundings

Keep head out of boat
Chart plotters not 100% accurate​

4. MOB

Most deaths have apparently been with unconscious MOB or trapped​

5. Emergency voice communication equipment

Sat phone proven extremely helpful in offshore experience – ‘after the fact’​

I agree with what you say. That is my motivation for trying to get the message home that forecasts are rarely "accurate" whatever that means and however defined. Used with care they are useful.

Good luck with your drafting. If any of my site is helpful to you, be my guest!

 

[mcframe]

In case anyone is interested, there are some places on a one day course at CA House, Limehouse Basin, on Sunday 29 January. This is in the list of courses run by the CA - http://www.cruising.org.uk/events/cahcourses.

On the 29th, I am giving 4 talks.

Thanks Frank - and very interesting it was too.

I now realise that trying to work out whether a "Selsey bill to Lyme Regis, F5-F7" is likely to be toward the top or bottom of the range, is a bit of a silly question (at 07:30 on a Sat morning, for a w/e sail in the Solent), and that I should put a bit more prep work in beforehand.

GRIBs are bit small on my iPhone, maybe I've got an excuse to buy an iPad? ;->

 

[franksingleton]

Thanks Frank - and very interesting it was too.
........

Thanks for coming. I am glad that you enjoyed it and hope that it was usful. Being a traveller when cruising, I will still take my laptop. However, I am very impressed by the iPad and the ease with which GRIBs can be obtained and the wealth of other useful information.

 

[Conachair]

Thread rebirth

Found this this morning. Not too bad for a 5 day.

http://www.weatherweb.net/wxwebtvsimonnew.php?ID=265


http://policlimate.com/verf/gfs_ecmwf_nh_f120.png