I'm curious as to what kind of wave model data you use from ECMWF and NCEP? Are you using the raw model data or are you applying an algorithm before displaying the model data on windy.com? I noticed a few times that the wave model data on windy.com does not exactly match up with the raw wave model data from ECMWF I can find elsewhere. Find an image from windy.com from earlier today below. You can see that the underlying map shows a southerly primary swell, however the wave rose and the table shows a northerly primary swell. Is that perhaps just a glitch or known bug?