Currently, I don't see the issue: both the particle simulation and the indicator in the picker show a weak SW wind.
Is there a bug fix, does it not reproduce, or is it dependent on certain conditions like the particle simulation and picker using different information sources? In the latter case, the discrepancy will only occur if the sources differ.

Here an example, the Markermeer in the Netherlands. Not a 180° difference in this case. But the particle animation show a NW wind, whereas the picker shows a NE wind. The later is correct, at least in agrees with Windfinder and other sources (KNMI, buienradar)
This morning I noticed a 180° difference in direction between WindFinder and the Windy particle simulation, even when I selected the same model.
My conclusion is that is something seriously wrong with the particle simulation. This can have potentially have grave consequences.