If we have measured the wind speed exactly at hub height over a long period at the exact spot where a wind turbine will be standing we can make very exact predictions of energy production. Usually, however, we have to recalculate wind measurements made somewhere else in the area. In practice,
that can be done with great accuracy, except in cases with very complex terrain (i.e. very hilly, uneven terrain).
Just like we use a wind rose to map the amount of wind energy coming from different directions, we use a roughness rose to describe the roughness of the terrain in different directions from a prospective wind turbine site.
Normally, the compass is divided into 12 sectors of 30 degrees each, like in the picture to the left, but other divisions are possible. In any case, they should match our wind rose, of course.