A Stochastic Model for Diagnosing Subtropical Humidity Dynamics with Stable Isotopologues of Water Vapor

The humidity of the free troposphere can be modeled, to first order, in terms of cold-point dehydration, followed by moistening via mixing with boundary layer air. The relative balance between these processes is of prime interest for understanding interannual variability of humidity and for understanding the water vapor feedback. Measurements of water vapor isotopic composition can provide quantitative constraints on these processes. The authors developed a stochastic model that parameterizes water vapor isotopic composition in terms of these processes and fit the model parameters to data from the Chajnantor Plateau, Chile (23 degrees S). For August-November 2012, the average mixing ratio was 1680 ppmv, with mean water vapor delta D of -234 parts per thousand and mean deuterium excess of 21 parts per thousand. The data were best fit by an asymmetric last-saturation distribution with mean last-saturation mixing ratio r(s) of 391 (+45, -75) ppmv, a median r(s) of 368 (+45, -75) ppmv, and a mean mixing fraction between the freeze-dried air and moist boundary layer air of [GRAPHICS] . Measurements from August to November 2014 had an average mixing ratio of 2210 ppmv, an average delta D of -220 parts per thousand, and an average deuterium excess of 14 parts per thousand. The last-saturation distribution for this period was less skewed than for 2012, with an average r(s) of 520 (+42, -75) ppmv and a median r(s) of 507 (+25, -75) ppmv. The mean mixing fraction for 2014 was [GRAPHICS] . The results show that the moistening in 2014, relative to 2012, requires increases in both the last-saturation mixing ratio and the postcondensation moistening and illustrate the utility of isotopic measurements for constraining the processes governing subtropical humidity.