Short-Term Records and Extreme Wind Speeds

An empirical study was conducted to determine whether design wind speeds can be estimated confidently from short records, i.e., records extending over periods of a few years. The study was based upon the analysis of records of daily, weekly, and monthly largest wind speeds taken at 36 United States weather stations over periods ranging from three to 10 years and upon largest yearly wind speeds recorded at these stations over periods ranging from approximately 25-45 years. Investigations were presented into the type of probability distributions that best fit sets of largest daily, weekly, and monthly wind speeds and into the relationship between winds estimated from short records of largest daily and weekly wind speeds on the one hand, and from 25-year to 45-year records of largest yearly wind speeds on the other hand. The main effort in the paper was focused on the relationship between estimates of 50-year wind speeds based on three-year records of largest monthly speeds and on 25-year to 45-year records of largest yearly speeds. It was found that 50-year speeds can be estimated confidently from three-year records of the largest monthly speeds.