Assessing the long-term representativeness of short wind records

Like most meteorological variables, wind speeds at a location may exhibit substantial interannual variability and may also be biased by observational inhomogeneities. Unfortunately, such variability and observational bias is all too often overlooked when utilizing short-term wind records for long-term planning purposes (e.g., dispersion modeling, wind energy, or structure loading). This paper addresses the statistical evaluation of time series and probability distributions from short wind records relative to long-term series. Specifically, this research focuses on evaluating post-1992 Automated Surface Observing System wind records relative to observations from pre-1992 NWS First Order stations using two data sets archived at the National Oceanic and Atmospheric Administration's National Climatic Data Center. Special attention is given to the differences in observational procedures between the two.