> These weather predictions are from passageweather.com.
> Each row represents a new set of predictions extending 180 hours into the future. In a row, the first plot is the prediction approx. 0 hours into the future, the second plot 3 hours, the third plot 6 hours, and so on. I checked the predictions and saved them every 6 (or sometimes 12) hours. So, there are 6 (or 12) hours between the set of predictions of each row.
> Each column represents the predictions for the same time. So, ideally the predictions down any one column would all be the same. The variation from one row to another in the same column shows the consistency in the predictions. Accuracy is another matter.
> The window of our passage is shown in the gray section, with the most concerning bits in red. (scroll down and to the right)

120 hours out

180 hours out

108 hours out

150 hours out

96 hours out

156 hours out

84 hours out

144 hours out

72 hours out

132 hours out

48 hours out

108 hours out

24 hours out

84 hours out

The day before we leave

12 hours out

72 hours out

The morning we leave

0 hours out

60 hours out

Ideally, if all the predictions made at different points in time for a particular future point (any one column) were consistent then all the plots in the column would all be the same. The following annimations are each for the column above the annimation. Predicting the weather window for a significant event (a strong Mistral) this far out appears to be something of a crap shoot. The thing is, the boat moves only so fast and we need to leave Barcelona days in advance of any potentila Mistral we want to avoid. You judge ...

Here is an annimation of all the predictions for this point in time (Tue 22 May 00 UTC).

Here is an annimation of all the predictions for this point in time (Tue 22 May 06 UTC).

Here is an annimation of all the predictions for this point in time (Tue 22 May 12 UTC).