摘要
Numerous articles have examined archival weather observations and attributed climate changes on time scales ranging from centuries to decades and in one case even days to human activity. This article gives examples showing how climate variability and sudden changes in instruments affect trend determination. In particular, surface temperature and water vapor pressure trends in North America during 1948-2010 are discussed. Over 1/4 billion hourly observations taken at 309 stations, were first carefully examined for inhomogeneities. Positive and negative steps, for both temperature and water vapor pressure were found to not be evenly distributed in time. Inclusion of such data in a trend analysis would overstate decadal changes in temperature and water vapor. Time series free of such discontinuities show a statistically significant warming has primarily affected the western Arctic, Canadian prairies and the Midwestern U.S. during winter. Increases in water vapor pressure are most pronounced in summer in the eastern U.S. The decadal water vapor pressure trends are somewhat smaller than found in other studies that examined data for far shorter time periods. The claim of a change in the DTR （diurnal temperature range） during the 3 day flight ban following Sept. 11, 2011, is not substantiated. The observed change in the DTR was likely caused by a reduction in cloudiness during the flight ban.
