Data assimilation in glaciology - Three inverse problems
Résumé
In this talk I will give an overview of inverse modelling for polar ice sheet through several examples. First I will present the basics of ice dynamics, the main physical processes, the sensitive parameters, some retroaction feedbacks, and the available observations. Then I will present some recent examples addressing the following questions: {Paleoclimatology.} How to infer a climatic scenario (i.e. how to reconstruct past polar temperature) from ice volume records? {Study of a surging glacier.} How can we explain that some glaciers, flowing normally at around 0.2 m/day, accelerate brutally to reach 80 m/day? {Sea level rise.} How can we compute projections of the polar ice sheets (the Antarctica and Greenland) contribution to sea level rise? The ingredients used to address these problems (either in idealised frameworks or with real data) include state-of-the-art ice flow models and data assimilation methods: Kalman filtering, adjoint method (optimal control).