Bayesian Inference of Physical Parameters at Langjökull

• Activity Number: 169 - Advanced Bayesian Topics (Part 2)
• Type: Contributed
• Date/Time: Tuesday, August 10, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Giri Gopalan* and Birgir Hrafnkelsson and Guðfinna Aðalgeirsdóttir and Finnur Pálsson
• Companies: California Polytechnic State University and University of Iceland and University of Iceland and University of Iceland
• Keywords: Glaciers; Physical parameters

Abstract

Glacial surface velocity is modeled as a sum of a gravitational deformational component and a basal sliding component, which both depend on important physical parameters. Using a shallow ice approximation physics model for glacial surface velocity, we conduct Bayesian inference to infer these parameters at Langjökull, a prominent Icelandic glacier. The posterior estimate for an ice-softness parameter is comparable with recommended values, and spatial patterns in basal sliding also coincide with previous studies of Langjökull. Residual analysis suggests that a normal data distribution for surface velocity is not a good fit, an important finding since many Bayesian statistical models of surface velocity make this assumption. Gibbs sampling is used for Bayesian computation.