Mean state and long term variations of temperature in the winter middle atmosphere above northern Scandinavia

Abstract

Stratosphere and mesosphere temperatures were measured during four winter months (November–February) at high latitudes (Andøya, ESRANGE) by means of numerous rocket flights during the Energy Budget Campaign 1980 and the MAP/WINE Campaign 1983–1984. They are compared to ground-based OH ∗ measurements and SSU satellite data. The atmosphere was found to be very active, with several minor and one major stratospheric warming occurring. A harmonic analysis of the temperature oscillations observed is performed and found to be suitable to model the atmospheric disturbances (warmings) to a large extent by superposition of waves with appropriate periods. These periods are of the order of several days and weeks and are thus similar to those of planetary waves. Stratospheric warmings tend to be correlated with mesospheric coolings, and vice versa. This is reproduced by the model, giving details of the phase relationships as they depend on altitude. These are found to be more complicated than just an anticorrelation of the altitude regimes. Strong phase changes occur in narrow altitude layers, with oscillation amplitudes being very small at these places. These ‘quiet layers’ are frequent phenomena and are independently found in the data sets of the two campaigns. They are tentatively interpreted as the nodes of standing waves. The time development of temperature altitude profiles shows strong variations that lead to peculiar features, such as a split stratopause or a near-adiabatic lapse rate in the mesosphere on occasion. The superposition model is able to reproduce these features, too. On one occasion it even shows super-adiabatic temperature gradients in the lower mesosphere for several days. Though this should be taken as an artifact, it nevertheless suggests a considerable contribution of the long period waves to atmospheric turbulence. The many rocket data are also used to determine monthly mean temperature profiles. These are compared to reference atmospheres recently developed for the CIRA ( Barnett and Corney, 1985; Groves, 1985). Fair agreement is found, which is much better than with CIRA (1972). This is not true for February 1984, because of the major warming that occurred late in that month. Before this warming took place, atmospheric preconditioning appears to have been present for more than two months.