The determination of accurate design temperature values for a particular location depends on having sufficient long-term quality data for the location. The availability of these data often dictates which locations are included in tables of design conditions. Thus, there is a need to determine the minimum number of years of data needed to determine reasonable esti-mates of design conditions or, if long-term data have already been analyzed, to determine the expected variability in the design values calculated based upon the amount of data used. The variability of the estimation of design conditions for 16 North American locations was investigated by comparing the difference in design conditions obtained from long-term (30 year) period-of-record (POR) data to those determined from subsets of the long-term POR data. The dry-bulb, dew-point, and wet-bulb temperature design conditions were analyzed for 1- to 15-year POR weather data subsets. The analysis demonstrated that if the estimate of the design values were determined using a one-year POR the aver-age of the annually determined design heating values would be slightly warmer (less extreme temperature) and the design cooling values would be slightly cooler (less extreme temper-ature) than using the 30-year POR. However, the range of differences was considerable and demonstrates the masking of variation in design temperatures caused by determining aver-ages for different locations. The 16-location mean and stan-dard deviation of the range in the errors in estimating the design temperatures between a one-year POR and the thirty-year POR indicate that the 99.6% design dry-bulb temperature heating could vary by 10.7±3.7°C (19.3±6.7°F) depending upon which year or location was selected. A POR of at least 12 years must be used to reduce the mean range of differences in design DB, DP, and WB cooling to less than 1°C. In order to reduce the mean range of differences in heating design dry-bulb temperatures to within 2°C, a POR of at least 13 years must be used. However, a considerable range in design condi-tions can still exist with individual stations and depends upon the climate.

Units: Dual