Modelling the middle atmosphere and its sensitivity to climate change

The Earth's middle atmosphere at about 10-100 km has shown a substantial sensitivity to human activities. First, the ozone layer has been reduced since the the early 1980s due to man-made emissions of halogenated hydrocarbons. Second, the middle atmosphere has been identified as a region showing clear evidence of climate change due to increased emissions of greenhouse gases. While increased CO2 abundances are expected to lead to a warmer climate near the Earth's surface, observations show that the middle atmosphere has been cooling by up to 2-3 degrees per decade over the past few decades. This is partly due to CO2 increases and partly due to ozone depletion. Predicting the future development of the middle atmosphere is problematic because of strong feedbacks between temperature and ozone. Ozone absorbs solar ultraviolet radiation and thus warms middle atmosphere, and also, ozone chemistry is temperature dependent, so that temperature changes are modulated by ozone changes. This thesis examines the middle atmospheric response to a doubling of the atmospheric CO2 content using a coupled chemistry-climate model. The effects can be separated in the intrinsic CO2-induced radiative response, the radiative feedback through ozone changes and the response due to changes in the climate of the underlying atmosphere and surface. The results show, as expected, a substantial cooling throughout the middle atmosphere, mainly due to the radiative impact of the CO2 increase. Model simulations with and without coupled chemistry show that the ozone feedback reduces the temperature response by up to 40%. Further analyses show that the ozone changes are caused primarily by the temperature dependency of the reaction O+O2+M->O3+M. The impact of changes in the surface climate on the middle atmosphere is generally small. In particular, no noticeable change in upward propagating planetary wave flux from the lower atmosphere is found. The temperature response in the polar regions is non-robust and thus, for the model used here, polar ozone loss does not appear to be sensitive to climate change in the lower atmosphere as has been suggested recently. The large interannual variability in the polar regions suggests that simulations longer than 30 years will be necessary for further analysis of the effects in this region. The thesis also addresses the long-standing dilemma that models tend to underestimate the ozone concentration at altitudes 40-75 km, which has important implications for climate change studies in this region. A photochemical box model is used to examine the photochemical aspects of this problem. At 40-55 km, the model reproduces satellite observations to within 10%, thus showing a substantial reduction in the ozone deficit problem. At 60-75 km, however, the model underestimates the observations by up to 35%, suggesting a significant lack of understanding of the chemistry and radiation in this region