Speaker: David al-Attar (University of Cambridge, UK)

Abstract:

When ice sheets melt or grow, sea level varies in a complicated manner in space and time, and this is recorded in relative sea level curves from around the globe. From these observations inferences can be made about the evolution of ice sheets and mantle viscosity. To do this, we must first be able to calculate the sea level changes predicted by a given model, and then can search for those parameters which best-fit the data. A practical approach for solving this problem is to apply gradient-based optimization. In this talk, I will show how the required gradients can be efficiently calculated using the so-called adjoint method. I will also outline some future plans to incorporate ice-dynamics directly into the problem.

***please see attached paper relevant to the talk***