Indications of long-term changes in Middle atmosphere transports

Upper mesosphere temperatures are derived from Meinel OH* emissions. They show large variations of various kinds. Annual mean temperatures measured at Wuppertal (51°N, 7°E) since 1981 do not show a significant long-term trend: the value obtained is 0.02 K ± 0.09 K per year. Monthly trends, on the contrary, are fairly substantial. To take a new and different approach to middle atmosphere long-term behavior, two new parameters are suggested for analysis: the short-period variance σ ² of the temperature, and the Equivalent Summer Duration (ESD). 1.) The variance σ ² itself shows long-term variations, which are correlated with temperature fluctuations at some lower altitude. This behavior is in line with the "downward control principle" and thus indicative of changes in atmospheric transports. 2.) The ESD parameter is defined in analogy to the duration of the vegetation period on the ground. A threshold temperature of 200 K at 87 km is chosen to define the beginning and end of summer. This yields an ESD on the order of 150 days. The ESD values at Wuppertal show a solar cycle effect which needs to be corrected for. Sensitivity of OH temperatures to solar activity at Wuppertal has been analyzed for this purpose. A sensitivity of 0.03 K ± 0.016 K per F _10.7cm flux unit was obtained. The corrected ESD values at the mesopause show a substantial increase of 18 days (11%) in 20 years (extrapolated). Respective vegetation period data have been published recently and show an increase of 20 days (13%) in 20 years (Zhou et al. 2001). They are in line with phenological data presented here. In the stratosphere summer length is estimated from the spring/autumn circulation reversals. Also here substantial changes are found, with a (negative) gradient of 19 days (13%) in 20 years. Considerable changes in middle atmosphere dynamics are thus indicated even though the statistics of some of the data are still limited. The mesopause ESD increases are compatible with a long-term change of the semi-annual oscillation (SAO). The stratospheric data suggest that the variations seen could be (in part) a climate fluctuation rather than a manmade trend, as data from two earlier decades show behavior that is opposite.