Reconstructing climate patterns in Europe during the LGM from inversions of past ice extents

Over 20 ky ago, Earth experienced its last large scale glacial advance. Ice sheets and ice caps covered large areas in both the Northern and the Southern Hemisphere. Conditions over Europe were cold and dry, with both the Scandinavian Peninsula and the British Isles in the north, as well as the Alps and the Pyrenees in central and western Europe, covered by ice. The influence of these ice sheets and ice caps on the atmospheric circulation, and to what point did the ice sheets in the north, and the subsequent expansion of the polar front, control the position of the Westerly winds and the regional precipitation patterns over Europe, remains a topic of debate. Secondly, there is a strong need for additional constraints on the amplitudes of mean temperature and precipitation changes compared to present-day. Although there is a discrepancy between observations and the modeling studies on both issues, it is clear that proper representation of ice surface topography in the models will strongly effect on the outcomes of such paleoclimate reconstructions. The approach presented in this study will be to combine ice flow modeling with the geomorphological evidence these ice caps and ice sheets left in the landscape in the form of moraines and trimlines. This approach enables us to gain insights into past climate conditions, by constraining changes and exploring patterns in temperature and precipitation across a mountain range, thus recovering valuable information on the regional atmospheric circulation over Europe. In order to investigate these questions, I have developed an inverse method to reconstruct the spatially variable mass balance rate of an ice cap, and in return the spatially variable position of the Equilibrium line altitude (ELA), using mapped ice extent and ice thickness data as input. The method is applied to the mountain ranges of the Alps and the Pyrenees, as they are effected by the same regional climate, and the recovered spatial variability in the patterns of the ELA reflects patterns in temperature and precipitation, indicating to the dominant directions of the regional atmospheric circulation. Since inverting ice extent data does not lead to the unique solution, I ran many scenarios with different climatic forcing in order to investigate the parameter space of the problem. In the case of the Alps, all of the presented scenarios recover a clear W-E gradient in the position of the ELA across the mountain range, and a N-S gradient in the west and central part, pointing to a dominantly zonal circulation with Westerlies bringing moisture from the Atlantic. This is supported by the Pyrenees reconstruction, where the method recovers a clear N-S gradient for all scenarios, indicating that the moisture source from the direction of the Atlantic. Finally, I used the change between present day ELA and the recovered LGM one to constrain mean annual temperature and precipitation change. I find that the temperature differences were about 9◦C in the Alps and 6◦C in the Pyrenees, and that climate must have been at least 25% drier in order to reproduce temperature depressions reported by proxy studies (pollen). The presented results show that by using the newly developed approach one can reconstruct valuable paleoclimatic information by using only ice flow modeling and mapped ice extents, enabling the applicability of the method even to areas where the data coverage is sparse.PhD thesis