Past Events

Seminar - Calibration and hypothesis testing for a model of avalanche behaviour

In avalanche dynamics, 'rheology' is used to denote the closure scheme of the equations of motion, in terms of a relation between internal forces and deformation of the flowing snow.  One such rheology is the Herschel-Bulkley (HB) model, which has four parameters that are somewhat empirical, and must be calibrated by observations. 

The statistical approach to calibration involves formulating a probabilistic description of the relationship between the HB model and observations from avalanche experiments, and then use Maximum Likelihood to estimate the parameters.  This approach also allows us to assess parametric uncertainty, and to test the hypothesis that the parameters are stable across different environmental conditions.  Exactly the same techniques can be applied across a very wide range of model-based inferences.  This is joint work with Hannah Cooper and Martin Kern.

Dr Jonathan Rougier is from the Department of Mathematics, University of Bristol.

Contact brett.cherry@durham.ac.uk for more information about this event.

Download this event in iCalendar format