The impact of bed elevation resolution on Thwaites Glacier ice dynamics

• Main Authors: McCormack, F, Galton-Fenzi, B, Seroussi, H, Roberts, J
• Format: Conference Object
• Language: English
• Published: . 2018
• Subjects: Earth Sciences; Physical Geography and Environmental Geoscience; Glaciology; Antarctic; West Antarctica; Thwaites Glacier; Antarc*; Antarctica; Ice Sheet
• Online Access: https://www.waisworkshop.org/workshop-2018#Agenda; http://ecite.utas.edu.au/135522

Bed elevation is one of the most important datasets in determining the dynamic evolution of theAntarctic ice sheet, being a strong control on ice flow, the evolution of the grounding line, and apredictor for marine ice sheet instability. Concerted efforts through international collaborationssuch as ICECAP (International Collaboration for Exploration of the Cryosphere throughAerogeophysical Profiling) and NASAs Operation IceBridge have made substantial progress inincreasing the resolution of mapped Antarctic bed elevation, particularly on the continentalmargins. This progress has facilitated improved accuracy in numerical modelling of the Antarcticice sheet and constrained uncertainty in projections of future ice loss. Nevertheless, fine scalemapping of bed elevation is costly and there is a case for targeting the collection of suchobservations to get the most out of ice sheet model simulations. Here, we address the question ofhow much resolution in bed elevation is sufficient to get consistent model simulations ofAntarctic ice sheet dynamics and their evolution. We use the Ice Sheet System Model and asynthetic, high resolution (250 m) bed elevation dataset to perform simulations of the ThwaitesGlacier basin in West Antarctica. We subsample the bed elevation dataset at 500m, 1km, 2km,4km, 8km, 16km, and 32km resolutions to investigate the sensitivity of Thwaites Glacierdynamics to the resolution of the underlying bed elevation. The modeled velocities converge forincreasing bed elevation resolution and for most of the basin the differences between the 250 mand 500 m simulation velocities are within 5%, which is within the bounds of uncertaintyassociated with the velocity datasets used to initialize our model simulations. Our results indicatethat a bed elevation of 500 m resolution is sufficient in simulating ice dynamics (velocities, basalshear stresses, strain rates) consistent with those using the higher resolution bed elevation data.This result has implications for future fieldwork mapping Thwaites Glacier bed elevations, andhas the potential to inform campaigns on other Antarctic glaciers with similar mechanismscontrolling ice flow.