Wind and Weather

[Neo]

Can anyone suggest a good source for wind and tidal current forecasts?Currently, I use the free membership at www.sailflow.com. But recently I heard that the GRIB data I had been downloading from another site was not relaiable, and that got me thinking. Thanks.

[matauwhi]

G'day, mate. I haven't been having any trouble with GRIB data. Cheers

[sailingmonica]

A friend of mine who is cruising since 2005 highly recommends (and uses) www.windfinder.com.

[imagine2frolic]

I was recently introduced to UGRIB.com, and with 2 ventures attempting to leave Panama, and sail to Hawaii. It has been completely accurate.......i2f

[Liam Wald]

PassageWeather - Sailing Weather - Marine Weather Forecasts for Sailors and Adventurers


Try this site and see if you like it. I find it very helpful.

[imagine2frolic]

We used Passage Weather alongside UGRIB in both our trips. We found the UGRIB spot on, and P.W. weather lacking a wee bit......i2f

[billreilly]

Originally Posted by imagine2frolic

We used Passage Weather alongside UGRIB in both our trips. We found the UGRIB spot on, and P.W. weather lacking a wee bit......i2f

If you found that the UGRIB forecasts were spot on, then the PassageWeather forecasts must have been spot on as well, because they both use EXACTLY the same weather data... from the GFS model...

It would be impossible for the UGRIB forecasts to be more accurate...

Bill

[franksingleton]

The only readily available GRIBs all come from the US GFS. Alternatives are NOGAPS and the Canadian model. These are unlikely to be better than the GFS.

Cheapest way of accession is via email from Saildocs or MailASail. Good for emails over cell phone, Iridium or over HF.

With Internet access, zyGrib is reliable, fast to download and has many options. It is a better service, better topography, chart projections than UGrib. Particularly useful is the CAPE parameter - an indicator of thundery activity.


PassageWeather, Windfinder, Weatheronline etc require broadband access and all suffer from problems in saving doe future reference.


These all give the same GFS data. None is better than another.


Meso-scale forecasts are another matter. All the models in use are capable of predicting small scale weather. However, there are always problems because of uncertainties in analysis and the short lifetimes of small weather features.

The larger Met services eg the US, UK, France are the only organisations that use good detailed meso-scale analyses, as far as I know. The US COAMPS is only available in a few areas – around the US and in the Mediterranean. It is available from Saildocs via email... The UK meso-scale model is available on Weatheronline but not in GRIB file form.

Weatheronline is the only, to my knowledge,useful ie easily usable source of output from the UK Met Office numerical weather prediction model. This is comparable to the US GFS. Them, they act jointly in providing the two World Aviation Forecast centres services.

The major Met services use these in a probabilistic mode by running model ensembles. That is they put small variations into the data and run the model many times. It is unsafe to use any meso-scale model in a deterministic sense.

My experience, broadly speaking, is that the GFS give generally reliable guidance to 3 days ahead. Day 4 is rarely dangerously misleading. After that there is skill but it decreases rapidly. I would not like to make a critical decision on the basis of a five day forecast. We plan on a 3 day basis with a good long look at 4 days.


To see more, go to Frank Singleton's Weather and Sailing Pages / Franks-Weather | The Weather Window.

[Hud3]

Another thing to keep in mind about GRIBs is that the GFS-based data is presented in 1/2 degree grids, so a single data point represents a ~30 nm square. For that reason I've always looked at GRIBs as a trend forecast, not something that's going to give you an accurate prediction for the specific location of your boat. There're pretty useful in conjunction with other sources of weather info.

[franksingleton]

You are quite right to draw attention to the limitations of the, nevertheless, useful GRIB products.


The two leading countries, US and UK both run global models on a 25 km grid at 70 levels. The UK since last February/March and the US since June 2010. That means that their models can only really describe weather on a scale of about 100 to 120 km. Any Grid can only define weather and topographical features on a scale of about 4 or 5 grid lengths. Models always have to filter out features of smaller size as being computational noise.


That means, as I say on my website, that all global models eg the GFS the UK, NOGAPS etc can and do only give predictions of that large scale weather features and not local detail. The limitations in time ahead that they are useful to us sailors is a result of chaos in the atmosphere. I sometimes see claims that one or other of the models is the best. The models are very similar to each other in configuration and in the way that they model the physics of the atmosphere. The various countries liaise closely through WMO and, in Europe, through membership of the ECMWF and the HIRLAM group. Each modelling group knows what others are doing. It is a very open science. The GFS is as good as any other. Its chief merit is its availability’s

Differences in forecasts on particular occasions are likely to be due to differences in the starting analyses. There is no unique best analysis and always small uncertainties; on some occasions one model may do better than others. Chaos rules. That is why, on the large scale also, the Met services run ensembles. That is why, the UK Met Office radio and TV forecasters will sometimes be very sure about the next 5 to 7 days and on other occasions much less sure. That is why, for us, deterministic forecasts should, in my experience, only be used to 3 or 4 days ahead.


On the smaller scales of meso-scale models, chaos effects become relevant much earlier. A large thunderstorm cloud may have a lifetime of only a few hours and its initial formation cannot be predicted.


An exercise that I strongly recommend that users of GRIB forecasts should take a systematic look at what they can do. Before each talk that I give, I take, at random, a seven day forecast and compare each forecast at 24 hour intervals with the actual analysis at t hat time. I am confident that my assertions will be borne out. You will also see that GRIBs underestimate the wind strength. I add one Beaufort force as a first guess.


GRIBs should always be used in addition to the forecasts produced by a human being, whether in chart or worded form. The forecaster knows about deficiencies in the models. He or she can and does look at output from other centres, especially beyond the first 2 or 3 days.


Sorry for going on at length.

[mbianka]

I've had some unpleasant surprises during recent cruises relying on the reported forecasts. So I've decided to make it a practice of checking the data of NOAA (here in the states) and other research weather buoys in the waters I plan to be sailing in before heading out (in additon to checking the forecasts). Gives me a heads up on what is actually happening vs what someone predicted in the past. If you can get access to them on line I think they really help in passage making.

[franksingleton]

Fine for what is happening now. Not for what may happen in two or three day's time.