Resource Request -- Vanderford Glacier Sensitivity Experiments

Project Title:
Understanding the physical processes driving recent changes at Vanderford Glacier, and the sensitivity of the system to different ice sheet model parameter choices

Scientific justification:

The Vanderford Glacier is the fastest retreating glacier in East Antarctica and drains a portion of the Aurora Subglacial Basin (ASB), a marine-based sector of East Antarctica vulnerable to atmospheric and ocean warming. The Vanderford grounding line has retreated over 18 km since 1996 (Picton et al. 2023), conducive with warm modified Circumpolar Deep Water driving high basal melt of the ice shelf. Whilst basal melt is likely a key driver, the influence of other processes (e.g., calving and subglacial processes) on this rapid retreat remains unclear.

Experimental design:

Using the Ice-sheet and Sea-level System Model (ISSM), I will conduct an ensemble of transient (100-year) model experiments to assess the sensitivity of the Vanderford system to recent observed changes and ice sheet model parameterisations (i.e., friction laws). A three-pronged approach will be employed to better understand the Vanderford Glacier by assessing: 1) the relative impact of various physical processes (e.g., basal melting and calving) on mass loss and ground line retreat, 2) the sensitivity of the systems response to external forcing with different friction laws and associated parameter choices, and 3) the timescale over which Vanderford Glacier responds to external forcings. A brief overview of key tasks is provided below:

Initial testing and benchmarking:

  • ISSM compilation and functionality testing.
  • Benchmarking of resource requirements for individual and ensemble run requests on Gadi.

Ice sheet model initialisation:

  • I will use data assimilation techniques, including inversion for friction coefficient on grounded ice and rheology for floating ice, with iteration between mechanical and thermal models as per Zhao et al. (2018), until pseudo-steady state is reached.
  • L-curve analysis is used to determine optimised regularisation terms for both friction and rheology inversions.
  • All model initialisations are completed using the Higher-Order mechanical model on an anisotropic mesh consisting of over 1 million elements spread across 10 vertical layers. Mesh resolution varies from 250 m on floating ice and close to the grounding line, to 20 km in the interior of the domain.
  • Model ice geometry and bedrock is initialised using BedMachine V3 (Morlighem et al. 2020), with a fixed ice front at the largest extent recorded between 1997 and 2021 (Greene et al. 2021).
  • Initial basal melt rates are taken from ISMIP6 (Jourdain et al. 2020).

Ice sheet model transient runs:

  • Select transient runs will be completed in both Higher-order and SSA to assess the variability of model results using different stress-balance approximations. Pending the results of this investigation, remaining transient runs will be completed in Higher-Order (computationally more expensive) or SSA.
  • Transient runs will be conducted for a period of 100-years, with a timestep of 1-month.
  • Experiments will include (in isolation and in combination) a transient ice front (to simulate calving), as well as basal melt parameterisations using the ISMIP6 basal melt protocol (Jourdain et al. 2020), and will be forced with a constant (mean) surface mass balance from RACMO.
  • External forcing will be applied for a 25-year period and the model will be allowed to evolve for an additional 75-years to assess the response time of the system.
  • The grounding line will be allowed to evolve throughout the simulation.

Compute Resource Request:

  • Initial testing and benchmarking: 20 kSU
  • Ice sheet model initialisation process: 40 kSU
  • Transient runs: 60 kSU

Contribution to Cryosphere Working Group:

The experiments proposed here will contribute to our understanding of appropriate ice sheet initialisation techniques, as well as necessary parameterisations for regions of rapid change. The work will provide quantification of uncertainties in a region that has been identified to be vulnerable from warming and where projections of ice mass loss are highly uncertain.

1 Like