On estimation and attribution of long-term temperature trends in the thermosphere

Abstract

Recent analyses of long-term time series of ion temperature from two midlatitude incoherent-scatter radars have revealed very strong cooling, which is substantially greater than predicted by models for neutral temperature. There is also an indication that the cooling has substantially accelerated after a breakpoint around 1979 when the ozone hole was discovered. This has prompted a hypothesis that the accelerated cooling might have resulted from the ozone depletion and associated reduction in daytime radiative heating in the stratosphere. A lively discussion on relative roles of different cooling mechanisms has followed. The purpose of this note is to contribute to this discussion from a theoretical and modeling perspective. In particular, a possible misinterpretation of the modeling results behind the ozone hypothesis is clarified. It is also shown that model predictions of neutral temperature trends in the thermosphere agree well with, and hence are tightly constrained by, independent observations including trends in heights of ionospheric layers and in neutral density from satellite drag. However, they are up to an order of magnitude smaller than the observational estimates of trends in ion temperature. These widely different results cannot be quantitatively reconciled regardless of what known cooling mechanisms are invoked. This stark discrepancy should be addressed on the data analysis and theoretical fronts.