High-latitude chemical release wind measurements were carried out in February and March 1978, in March 1985, and in March 1987. In each of the experiments, wind profiles were obtained covering heights in both the E and the F region. Three of the release experiments were carried out on the evening side of the auroral oval and one on the morning side. Two sets of measurements were carried out in disturbed conditions at solar maximum, while the other two were carried out during quiet periods at solar minimum. The spectral thermospheric general circulation model that has been developed at the Danish Meteorological Institute is used to simulate the conditions appropriate to each of the four experiments and detailed comparisons between the model predictions and the measurements are presented. Considering the uncertainties in the various external sources of forcing, such as the plasma convection patterns, the model adequately reproduces the major features of all the wind profiles. However in the E region the relative wind maxima from the model are, in general, above the heights of the observed wind maxima, possibly due to the oversimplified auroral precipitation used in the model, with the electrons being represented by single Maxwellian energy spectra only.

The uncoupled neutral and ionized atmospheric compositions used in the model may also explain part of the unrealistic simulated winds. The upward propagating tides are found to modify the E region winds significantly, even under disturbed conditions when the plasma forcing might be expected to dominate the dynamics. In our results the latter is shown by the sensitivity of the simulated flows to the lower boundary condition which is the imposed tidal oscillation structure. ¿ American Geophysical Union 1993