Reliability of Wind Predictions

[barnakiel]

Originally Posted by viben

(...)

look for differences and commonalities? If they all align, the odds are that it is reliable. If they are all over the place, anything could happen...

(...)

Hahaha! Read your own words again.

Whether the results from different models align or not has no bearing on their reliability.

Imagine an alternate situation: Out of six experts, five claim A will happen while one claims B will happen. Now does the five claims bear on what will happen? You may be tempted to answer YES - only if there is a feedback loop - someone who has power over A and B outcome hears the experts and takes advice from them. In all other cases the answer is NO.

For the same reason (although in a way too complex to present here) results of mathematical weather, financial or otherwise models have no bearing on probability of the actual thing happens to be. Whether they agree or disagree ANY can be more reliable than any other. There is no 'team work' nor any 'crowd wisdom' for such models.

Weather forecasting, statistics, probability. Read, read, read then read more. It is all a bit more complex than the common sense.

Cheers,
b.

[rbk]

Not that this the below is a great site specific forecaster but the modelling for the below website and app are pretty spot on for the PNW anyways and is a great tool for predicting weather systems. I draw my own conclusions from watching different pressure systems move around and on the app you can view the 500mb systems in the same format. Takes some playing with to learn all the tools available on it, but it's my go to for weather wind and waves.

https://www.windytv.com/

[Paul L]

Barnakeil
If a group of models that you are using to plan a passage all align reasonably, why would you not put higher likelihood on that prediction than if the models clearly diverge. I'm asking from a practical passage planning point of view.

[Kenomac]

Here's a video shot when we were using Pocketgribs. We now use a PredictWind paid subscription which has been much more accurate. On the day pictured, the forecast predicted zero winds and calm seas. Here we are using just our staysail in 30-35 knots, making 8-9 knots.

It's wise to be ready for the nasty stuff, even when fair winds are predicted.

https://m.youtube.com/watch?v=s0HgonCDFxI

[Paul L]

Originally Posted by Kenomac

Here's a video shot when we were using Pocketgribs. We now use a PredictWind paid subscription which has been much more accurate. On the day pictured, the forecast predicted zero winds and calm seas. Here we are using just our staysail in 30-35 knots, making 8-9 knots.

It's wise to be ready for the nasty stuff, even when fair winds are predicted.

PocketGrib has nothing to do with the prediction. It, like most other GRIB viewers, is only a display program. In the case of PocketGrib it displays the NOAA GFS model data.
Perdictwind also displays the GFS model, as well as some others.

[Stu Jackson]

Originally Posted by viben

Any rational opinions, based on actual use and experience, about the various wind forecasts, models and apps available out there?

We found Passageweather very good coming up the Pacific coast last summer.

But I think the perceived specificity of numerical forecasts is vastly overrated, simply because they are numerical. Used to be: winds 34-45 would be enough, 'cuz most folks knew there were variations not only to be expected of a forecast, but because of local conditions.

Now, it seems if the forecast is 32.3567 to 33.4567, some bozo says they were wrong because it blew 34.

Heck, wasn't that long ago that all we had were barometers. And even now good weather reference material will suggest rapidly dipping barometric pressure indicates duck and cover.

[Hearts Content]

The fact is that they are still guessing and almost never get it right on the Gulf Coast of Florida. You can almond st bet on the opposite.

My favorite: Predicted 15-25 NNE and 5-8' seas. Wind died to zero in the morning and then an unexpected front stirred it to SSE 25-30+ and 10-20' seas. After about an hour of being in this, they were nice enough to change the forecast. . We decided to head in to Marco Island.

Always make sure you and your boat can handle the worst when making passages.

[Paul L]

Originally Posted by Stu Jackson

We found Passageweather very good coming up the Pacific coast last summer.

But I think the perceived specificity of numerical forecasts is vastly overrated, simply because they are numerical. Used to be: winds 34-45 would be enough, 'cuz most folks knew there were variations not only to be expected of a forecast, but because of local conditions.

Now, it seems if the forecast is 32.3567 to 33.4567, some bozo says they were wrong because it blew 34.

Heck, wasn't that long ago that all we had were barometers. And even now good weather reference material will suggest rapidly dipping barometric pressure indicates duck and cover.

Passageweather displays the GFS model data:

Our worldwide Surface Wind (10m above sea level), Surface Pressure, Visibility, Cloud Cover and Precipitation forecast charts are derived from the 0.25 degree GFS (Global Forecast System) model, one of the operational forecast models run at NCEP.

[J Clark H356]

I have found SailFlow is pretty good. None are perfect as the mathematical estimates are just that "estimates".

New Bern coastal area from SailFlow at 11:00 PM 3/15 shows 9 Gusts 16. Actual 9 by Intellicast Storm doesn't show gusts. NW direction forecast and actual the same.

[Paul L]

Originally Posted by J Clark H356

I have found SailFlow is pretty good. None are perfect as the mathematical estimates are just that "estimates".

New Bern coastal area from SailFlow at 11:00 PM 3/15 shows 9 Gusts 16. Actual 9 by Intellicast Storm doesn't show gusts. NW direction forecast and actual the same.

And the models displayed by Sailflow include:

Regional Models

NAM - North American Mesoscale Model - This was the former ETA model and is delivered on WindAlert products with a 5km and 12km resolution and delivers you forecast information for the next 78 hours. It is only available across North America and for the surrounding coastal regions.
Global Models

GFS - Global Forecast System - It is run four times a day (00Z, 06Z, 12Z, 18Z) and produces forecasts up to 16 days in advance. It is run at roughly 50km resolution. However, the spatial and temporal resolution is reduced as the model is further out in time. (The forecast produced beyond day 7 is generally not trustworthy. It produces a forecast for every 3rd hour for the first 180 hours.
CMC - Environment Canada Global Model - This is another model which will give you a forecast for all points on the globe. The resolution is approximately 100km resolution and runs out 138 hours.

[Steve Bean]

"Be ready for the nasty stuff." "Be sure you and your boat can handle the worst." This is how I think:. If the prediction is for "15 to 25 knots" I just add the two together, and get ready for 40!

[barnakiel]

I have noticed the following 'rules':

Open waters, NOAA source:

- forecast 20kts or above: add 5kts,
- forecast 8kts or below: subtract 3kts.

Also, for some odd reason, NOAA models do not give a hoot for tropical waves passing by until they are at Barbados or thereabout. In open waters the wind force on the wave is 'OK' (= too low) BUT the gust force simply disregards the specifics of the wave. So there is still some work to be done at NOAA, much as these guys are getting pretty close to perfection otherwise.

Cheers,
b.

[zeehag]

i use passage. when it is green area coming at me i put out another 30-40 ft chain. whooyaaaahhhh if gods wanted us to know our wind they woulda made em predictable. there are seasonal trends. know these. in southwest coastal mexico, trend is march expect 30 kts. april 35. may is readying for canes and named storms--40 kts..find marina for the rest of summer or hide in a mangrove bank.
eacch area has its own seasonal trends. it helps to know what is happening in your cruising grounds.

[adfrolov]

And what sailors think about ECMWF? Or GFS is enough for rough view and than making own predictions based on geography?

[barnakiel]

Originally Posted by adfrolov

And what sailors think about ECMWF? Or GFS is enough for rough view and than making own predictions based on geography?

The European model is OK. I did not find it better or worse. It is nice to have both sources and compare the info.

I did find the European data more difficult to download and process as not all packages read it. I used the French website to download.

If you know a better (platform independent) way to download, PLS share.

b.