The pressure patterns on the primary mirror of a telescope exposed to wind have been investigated in three different configurations: with the telescope in open air, in the slit of the NTT building and in a cylindrical recess.
While the spatial distribution of mean pressures shows, as expected,
considerable differences in the different types of enclosure,
the patterns of pressure fluctuations are quite similar.
As a consequence their effect on optical aberrations
resulting from dynamic deflections of the mirror surface can be
predicted with reasonable accuracy by a general relationship between
the average rms
of pressure variations on the mirror and the
total rms
surface error.
For an astatically supported mirror with three reacting virtual support
points, this parameterization is given by expression (
):
The optical aberration is found predominantly in the astigmatic mode,
which allows us to relate the rms
surface error to the image spread
angle 
obtaining expression ():
Simple approximate relationships can also be found between the local speed values and the rms of pressure variations on the mirror, but they will likely depend on the enclosure type. In the particular case of the VLT cylindrical enclosure, a parameterization is found between the average speed measured 60 cm (full scale) from the mirror surface and the rms of pressure variations, which does not depend on the azimuth angle.
These relationships will be particularly useful for parametric analysis of the overall wind+seeing effects of the primary mirror and for determining the optimum operation of the venting devices (windscreen, louvers) of the enclosure.
The study results outline the great sensitivity of large primary
mirrors to wind buffeting. For example in the instance of the VLT,
according to equation (),
the set limit of a wavefront error 
of 200 nm implies a
maximum allowed pressure fluctuation 
of 1.3 Pa only.