Conventionally, the 50-year wind is calculated on basis of the annual maxima of consecutive 10-min averages. Very often, however, the averages are saved with a temporal spacing of several hours. We call it disjunct sampling. It may also happen that the wind speeds are averaged over a longer time period before being saved. In either case, the extreme wind will be underestimated. This paper investigates the effects of the disjunct sampling interval and the averaging time on the attenuation of the extreme wind estimation by means of a simple theoretical approach as well as measurements. The measurements include climates dominated by extratropical lows, as well as more complicated, wind climate types. For both, the investigations are done for omni-directional and sector-wise situations. The theory assumes the time series is a Gaussian Markov chain and it performs adequately for sites in the extratropical regions. The sector-wise situation is more complicated. The attenuation of the extreme mean winds for a sector shows a dependency on the frequency of occurrence from that sector.
Journal of Wind Engineering and Industrial Aerodynamics, 2006, Vol 94, p. 581-602