Anticipated trends in atmospheric water vapor and the challenge of quantifying them

Various models agree that water vapor is anticipated to increase throughout the troposphere and into the stratosphere as a result of anticipated increases in temperature during the current century. International efforts such as the Network for the Detection of Atmospheric Composition Change (NDACC) and the GCOS Reference Upper Air Network (GRUAN) have tasked themselves with establishing networks to measure atmospheric water vapor with sufficient accuracy to monitor these trends. But how long will these networks have to measure water vapor and with what accuracy to reveal the trends with statistical confidence? Statistical tools exist for estimating the time to reveal a trend from long term datasets. For example, Weatherhead et al. 1998 present such a formulation where the time to reveal a trend is a function of the magnitude of the trend, the length of time the parameter is measured, and the variance and autocorrelation of the data. We use this formulation and estimates of anticipated trends in water vapor based on coupled climate models to estimate the time required to measure trends in water vapor with statistical robustness. The results of this study have implications for such networks as NDACC and GRUAN in terms of measurement technology, the location of measurement sites and the calibration requirements of the instruments.