Sample size effect on the reliability and calibration of design wind load

The small sample size of the annual maximum wind speed at a wind speed recording site causes uncertainty in the estimated return period of wind speed. The influence of this uncertainty on the estimated return period values of annual maximum wind speed and structural reliability is quantified in this study. The results show that the effect of statistical uncertainty due to limited sample size on the estimated return period value of the wind speed and structural reliability is largely reduced if the sample size is increased from 10 to 20. The implication of the results is discussed in terms of two alternative options to specify the factored design wind load: (i) a low return period for selecting nominal wind speed combined with a large wind load factor and (ii) a high return period for selecting nominal wind speed combined with unity wind load factor. It was concluded that option (i) is preferred to cope with sample size effect if the design in a region with statistically homogeneous wind climate; for code making with spatially varying coefficient of variation of annual maximum wind speed, the analysis results support the use of option (ii) since it leads to improved consistency of reliability.