As I sit here listening to the first heavy rain in weeks, I decided to share some observations about the wind forecast for the Gulf Stream which are better than anecdotal, if not quite statistically robust. What is important here is the thought process rather than the precise data. The central observation is:
How much a ten day wind forecast varies often escapes us, and forecasts of more than two or three days (text or GRIB) just aren’t very useful for spotting Stream-crossing weather windows. These observations neither transfer nor scale to other situations/areas. [All graphics can be clicked to expand.]
- Since we have been stuck here in our Catch-22, there have been 11 crossable days (blue green and about half yellow — 11 out of 23) — single days sandwiched between bad ones didn’t count.
- Persistent forecast for each of those crossable days was available 3, 4, 6, 2, 0, 0, 1, 0, 0, 9, 0 days ahead — or an average of 2.3 days in advance.
- If the one nine day outlier is removed, the average drops to 2 days.
- Since we prefer to avoid the yellow days and single days, this drops to 5 crossable days with 2-3 days notice.
- Crossable doesn’t necessarily mean enjoyable.
Stare at the picture from a distance. Columns are the successive forecasts. Rows are the evolution of the individual daily forecasts. The last block in a row is the observed wind on forecast day. The open green boxes capture the histories of the evolving crossable windows. One can begin to see windows don’t announce themselves very far ahead.
So what have we seen as to observed winds and predictions? The best windows (in turquoise territory) are two or more days long — either leaving time for swell to abate, or time to get farther along one’s route — beyond the Gulf Stream. [The convention with each of the charts that capture wind direction is N = 0, E= 90 and S= 180 or -180 which ever was least ambiguous when plotting, and W = -90.
Interestingly, the observed winds seem to come closer to the maximum westerly prediction and the predicted winds come closer to the easterly maxes. I’m betting this is a jet-stream driven observation, Especially since a deep south jet has been a strong influencer lately. So can one say if the jet stream is well south, bet on the more westerly predictions? Don’t know, but I think I might since the assumption fails toward safe.
As you can see the data is choppy, and two of the most variable days were associated with a very good window. So wide variability is not cause for writing off a particular day — especially if one is outside the “last two days” zone.
*Somewhat biased downward by short forecast periods or short evolutions at each end of the data field.
There are some features to variability that may be useful (or may be more jet stream). Here we can see the greatest variability was strongly associated with observed winds from the W>S (blue). This doesn’t surprise since these winds are themselves the heralds of change — of approaching Fronts.
However, while the predicted wind (yellow) directions seem to lack correlation with variability, the amplitude of the variability is as great as the observed winds. This tells the story — the best wind directions are the hardest to predict. Or, when it’s forecast to be bad, you can be pretty sure it is going to be bad, and when it’s forecast to be good, go figure.
The widest window (25-28 Jan), the one with one day which had nine solid days of notice, how did the variability play out? Through most of the forecast period, only the 27th looked useful. The other three days offered no notice.
The bottom line. Be prepared and jump when nature says, “frog.”