A molecular view of atmospheric temperature

Atmospheric temperature is measured by calibrated thermometers to yield a traceable record. The sensors employed average over a distribution of molecular velocities which is implicitly assumed to be Maxwellian, yielding a well known relationship from statistical mechanics between macroscopic temperature and mean square molecular velocity. We use statistical multifractal analysis of the intermittency of observed airborne lower stratospheric temperatures together with the observed rate of ozone photodissociation to demonstrate a strong correlation between these two variables over the winter and summer seasons in the Arctic lower stratosphere. We argue that literature molecular dynamics simulations showing the evolution of vortices on very short spatial (10**-8 m) and temporal (10**-12 s) scales provide a plausible explanation, via a self-sustaining interaction between the over-populated molecular high velocity tails and the vortices. The asymmetric, long-tailed PDFs of temperature observed in the winter vortex and in the summer anticyclone are consistent with the argument, as are flights in the same air mass on either side of the terminator, which also provided an integral measure of the effect. The implied ability of translationally hot ozone photofragments to maintain a non-Maxwellian distribution of molecular velocities in all air molecules has many implications. One is that the measured temperature in daylight may depend upon the ozone concentration, a potentially significant finding in view of the increase during the 20th Century in the troposphere by as much as a factor of five, and the later decrease in stratospheric ozone from chlorofluorocarbons. A second is that the overpopulation of high velocity molecules could cause non-Lorentzian absorption in the wings of the infrared spectral line shapes of water vapour, affecting a basic greenhouse gas process. A third is that the high velocity molecules could be accelerating some atmospheric chemical reactions (those with activation energies) and decelerating others with negative temperature dependences (such as radical recombinations). A fourth is that there may be a fundamental link to atmospheric turbulence. We suggest observational, experimental and theoretical approaches to examine and test these possibilities.