Climatic records of wind speed are analyzed for eight long-term monitoring locations in the Beaufort/Chukchi coastal region of the Arctic, over the period from 1979 to 2009. Data for this study originated from three national observing networks throughout the Beaufort/Chukchi region and were uniformly quality controlled using a combination of automated and manual checks. The climatology of wind speed was developed over the study period and includes analyses of mean wind speed, extremes, and frequency distributions. Trends in monthly and annual wind speed were examined. The climatology illustrates strong distinctions between coastal and interior locations, strong seasonal characteristics, and diurnal cycles in wind speed. Negative trends in wind speed were apparent at several locations, particularly at locations in Alaska. Many of these trends were statistically significant (95% CI). To assess the impact of artificial changes on trends, station data were homogenized with respect to metadata records. Trends were recalculated and the effects were discussed. Possible drivers of wind speed trends were also discussed, including multi-decadal climate signals. Results from the climatology and trend analyses were compared with output from the North American Regional Reanalysis (NARR). Negative biases in wind speed were apparent at most locations throughout the year in NARR. Trends in NARR wind speed opposed in situ trends, at most locations. Improved knowledge of wind speed records and reanalysis output is important in this rapidly changing region.

No abstract

No abstract

ABSTRACT. Meteorological observations from more than 250 stations in the Beaufort and Chukchi Sea coastal, interior, and offshore regions were gathered and quality-controlled for the period 1979 through 2009. These stations represent many different observing networks that operate in the region for the purposes of aviation, fire weather, coastal weather, climate, surface radiation, and hydrology and report data hourly or sub-hourly. A unified data quality control (QC) has been applied to these multi-resource data, incorporating three main QC procedures: the threshold test (identifying instances of an observation falling outside of a normal range); the step change test (identifying consecutive values that are excessively different); and the persistence test (flagging instances of excessively high or low variability in the observations). Methods previously developed for daily data QC do not work well for hourly data because they flag too many data entries. Improvements were developed to obtain the proper limits for hourly data QC. These QC procedures are able to identify the suspect data while producing far fewer Type I errors (the erroneous flagging of valid data). The fraction of flagged data for the entire database illustrates that the persistence test was failed the most often (1.34%), followed by the threshold (0.99%) and step change tests (0.02%).Comparisons based on neighboring stations were not performed for the database; however, correlations between nearby stations show promise, indicating that this type of check may be a viable option in such cases. This integrated high temporal resolution dataset will be invaluable for weather and climate analysis, as well as regional modeling applications, in an area that is undergoing significant climatic change.Key words: western Arctic, meteorological observations, data quality, automated quality control, Beaufort Sea, Chukchi Sea, Alaska RÉSUMÉ. Des observations météorologiques provenant de plus de 250 stations des régions côtières, intérieures et extracôtières de la mer de Beaufort et de la mer des Tchouktches ont été recueillies pendant la période allant de 1979 à 2009, puis elles ont fait l'objet d'un contrôle de la qualité. Ces stations relèvent de plusieurs réseaux d'observation différents qui existent dans la région à des fins d'aviation, de météorologie forestière, de météorologie côtière, de climat, de rayonnement de surface et d'hydrologie, et elles fournissent des données horaires ou subhoraires. Un contrôle de la qualité (CQ) unifié des données a été appliqué à ces données provenant de sources multiples en faisant appel à trois méthodes principales de CQ, soit le test d'acceptabilité (qui a permis de déterminer dans quels cas une observation ne faisait pas partie de la gamme normale); le test de la variation discrète (qui a permis de détecter les valeurs consécutives qui sont excessivement différentes); et le test de la persistance (qui a permis de repérer les cas de variabilité excessivement élevée ou basse). Les anciennes méthodes de CQ des données quo...