Historic wind data

[scruff]

I'm contemplating sailing from Oban to the Azores in a couple of years. Is there a resource out there which has the average wind direction & force, week by week going back X years for the Northern Atlantic? My google powers are evading me today.

It'd be good to play around to work out the "best" time of year and routing to go in order to avoid / minimise beating for 1,500 miles...

thanks

 

[GHA]

The pilot charts which give monthly averages are available here >>
https://msi.nga.mil/NGAPortal/MSI.portal?_nfpb=true&_pageLabel=msi_portal_page_62&pubCode=0003

And 2 weekly averages from satellitte measurements here >>
http://numbat.coas.oregonstate.edu/cogow/0101/66_top_right.html

 

[jdc]

There is a real problem defining 'average' for wind, and then even more difficulty relating this to sailing conditions. Consider two cases of 10 day periods: Case A has 9 days out of 10 with 9 knots and 1 day in 10 with 39 knots.
Case B has 10 days at a constant 12 knots.

They have identical 'average' wind strength, but case A will be boringly slow sailing, mixed with unpleasant gales, whereas case B will be lovely conditions for a fast and safe passage. So what use is a simple arithmetic mean? Besides, what does average mean in a vector quantity? Is it the mean of sqrt(U² + V²), or sqrt(the mean of (U² + V²)) or (bonkers) the means of U and V separately? Remember that wind is both forecast and recorded as a U and a V component.

Hence I think average wind strength is both problematic to define, and probably useless anyway.

Average wind direction is even more problematic: to reduce ad-absurdum suppose the wind is 359 degrees half the time, and 1 degree the rest. Is that an average of 180 degrees, ie South? Clearly that would be wrong, so maybe we have to make it an average of -1 degree and +1 degree. Much better, the average is now 0 - spot on. But if doing this, what if the real wind were 50% of the time 90 degrees (ie E) and 50% at 270 (or -90), ie W? The 'average' is then either 180 (S), or 0 (N), both of which are clearly wrong. You can't 'average' direction, but you can plot a distribution.

What could one do? Fairly simple to calculate is the mean square of wind speed. This relates to the power you might hope to extract from a turbine, so is interesting but probably not that useful for sailing (possibly correlates well with fatigue life in rigging and sails?). Better, one could in principle take the cardioids for your boat for a wide range of wind speeds and apply these to actual historic wind, assuming it's sampled frequently enough, and calculate many, many, voyage times from which one could get a mean time for the voyage (and variance, would probably be just as interesting). But I very much doubt that you'll have the data to perform this: if you do, I'll be jolly impressed - please share the results!

A simpler thing to calculate would be the proportion of the time the wind is in a particular range of directions (say the ones you wish to avoid) in any rolling 10 day period. This might be what you're after, although as someone who has gone to the Azores a few times I'm not convinced that head-winds are to be avoided so much as calms!

So I'd say you have to be very clear about what you are trying to get a handle on and the routing charts, as GHA suggested, are probably still the best source of relevant info as they show percentage of calms, percentage of gales and a distribution (as a rose) of wind strength and directions.

 

[DFL1010]

The pilot charts that GHA are the best for a general overview. There are many flaws/quirks in collection and presentation for precise work, but they are a very good way of getting the gist of it.

The other option is to download historical forecasts in grib format.
NOMADS (https://nomads.ncep.noaa.gov/) is the place to go for GFS ones; I wouldn't worry too much about which model you use, as you'll only be using the first few hours of each forecast before the next one takes over. Also, NOMADS is quite easy to get bulk downloads by means of a simple script, and then cat them for a whole passage file.

Once you've got your grib, you can use some routing software. Two free ones of note are opencpn with a plugin (https://www.opencpn.org/OpenCPN/plugins/weatherroute.html), or qtVlm (https://www.meltemus.com/index.php/en/).

Either would be more than sufficient for this job.

You will need some polars for your boat - what type is she? There's a very useful tool (https://jieter.github.io/orc-data/site/) that has ORC-rated boats' polars; you might be able to find something usefully similar. If not say and I'll see what else I can find in my bookmarks.