Considerarion of sample size effect in codification of wind load and its implication in NBCC

The factored design wind load is calibrated based on structural reliability theory for selected target reliability index by considering a specified (or nominal) reference wind velocity (pressure). The specified wind speeds for different sites are estimated from wind records for a given return period. The length of the record is limited, and the limited sample size leads to the estimated return period value to be uncertain. Quantification of the effect of this uncertainty on the estimated reliability and its potential effect on the calibrated factored design wind load are investigated in the present study. For the analysis, the extreme wind speed is modeled as a Gumbel variate, and the distribution fitting is carried out using the method of moments since this distribution model and the fitting method are employed to develop the reference wind pressure in the current National Building Code of Canada (NBCC). The implication of the results in the context of code making and the NBCC is discussed; recommendations based on the obtained results are given.