Polar Seminar: Anders Damsgaard, "Numerical modeling of subglacial sediment and water: Viscous creep and plastic failure"


 
04/22/2015 - 3:30pm to 4:30pm

Wednesday, April 22nd, 3:30PM
Location: Revelle 4301
Anders Damsgaard
(soon to be) Green Postdoctoral Scholar, IGPP, Scripps Institution of Oceanography, La Jolla, CA

Numerical modeling of subglacial sediment and water: Viscous creep and plastic failure

Most glaciers and ice sheets reside on melt water-saturated deformable granular materials, consisting of either reworked older sedimentary deposits or mobilized wear products from erosion, collectively termed till. The mechanical behavior of till affects glacier movement and can dramatically change during melting events where water is routed to the glacier bed. The rheology of till has been discussed extensively and has been studied both by in-situ measurements with instruments placed in the glacier bed and by standard geotechnical laboratory studies of till samples. Several studies have highlighted the non-linear strength at the plastic yield limit while only little attention has been given to the viscous pre-failure creep of till. The switch between these two states is a promising explanation for observed large-magnitude glacier flow velocity fluctuations caused by moderate transient forcings. However, the precise micro-mechanics and internal structural changes in a granular material shifting from one state to the other remain unknown. 

In this talk I'll present a new computational model for performing deformation experiments on the subglacial sedimentary bed. The grains in the granular phase are simulated individually by the discrete element method. The grains are fully two-way coupled to the meltwater in the pores, which is treated as a compressible Newtonian fluid flowing per Darcy's law. The computations are solved by utilizing the massively parallel arithmetic capabilities of modern graphics cards (GPUs). The computational approach allows for detailed studies during progressive deformation, e.g. during temporally evolving subglacial stresses or water pressures. The findings improve our understanding of flow velocity records from glacial systems whose stresses are dynamically evolving due to variable melt water input or tides.

For more information on this event, contact: 
Matthew Siegfried
Event Calendar: 
Location: 
Revelle 4301
sharknado