Thanks. Some further questions if you have the time...

Do you use raw NOAA data or GFS + validation using other meteorological models? GFS afaik underestimates wind speeds, especially for higher speeds, so I would be curious to see how one corrects for that.

As far as wake losses are concerned, I suppose your regular 2.5/5 diameter rule does not apply... but why is that? Is it because wake effects do not move linearily with rated power?

Rien n'est gratuit en ce bas monde. Tout s'expie, le bien comme le mal, se paie tot ou tard. Le bien c'est beaucoup plus cher, forcement. Celine

You do get higher correlation by having a mast to use as a base station. The most expensive ones are complete research stations, which can be used for wave and current measurement and environmental issues. They can run over €3M, which can be shared between projects. Germany has built two in the North Sea and one in the Baltic. FINO 1 was operational since 2003, and FINO 3 since 2009.

You can visit FINO 1, 2, and 3 on the web, starting in english HERE.  You can get live speed data and images. FINO 3 cost €12M which includes years of research and measurement projects. It measures to 105m, with a 15m lightning rod taking it to 120m.

There are less expensive versions primarily aimed and wind and wave measurement. The technology of floating stations is gradually gaining acceptance as well, or at least entering the market, usually LIDAR or SODAR based. Here's a test of one model:

I believe many of the larger projects will need to have a station, because there is no substitute for onsite data. This can help with power curve verification as well as wake analysis, so should prove cost effective, especially if shared.

"Life shrinks or expands in proportion to one's courage." - Anaïs Nin