From cyclic ice streaming to Heinrich-like events: the grow-and-surge instability in the Parallel Ice Sheet Model

Abstract

Abstract. Here we report on a cyclic, physical ice-discharge instability in the Parallel Ice Sheet Model, simulating the flow of a three-dimensional, inherently buttressed ice-sheetshelf system which periodically surges on a millennial timescale. The thermomechanically coupled model on 1 km horizontal resolution includes an enthalpy-based formulation of the thermodynamics, a nonlinear stress-balance-based sliding law and a very simple subglacial hydrology. The simulated unforced surging is characterized by rapid ice streaming through a bed trough, resulting in abrupt discharge of ice across the grounding line which is eventually calved into the ocean. We visualize the central feedbacks that dominate the subsequent phases of ice buildup, surge and stabilization which emerge from the interaction between ice dynamics, thermodynamics and the subglacial till layer. Results from the variation of surface mass balance and basal roughness suggest that ice sheets of medium thickness may be more susceptible to surging than relatively thin or thick ones for which the surge feedback loop is damped. We also investigate the influence of different basal sliding laws (ranging from purely plastic to nonlinear to linear) on possible surging. The presented mechanisms underlying our simulations of self-maintained, periodic ice growth and destabilization may play a role in large-scale ice-sheet surging, such as the surging of the Laurentide Ice Sheet, which is associated with Heinrich events, and ice-stream shutdown and reactivation, such as observed in the Siple Coast region of West Antarctica.

Heinrich events show two-stage climate response in transient glacial simulations

“…5b, d). On land, pollen records from around the Mediterranean and western Europe (e.g., Tzedakis et al, 2004;Fletcher and Sánchez Goñi, 2008;Fletcher et al, 2010) show a correlation between the Dansgaard-Oeschger cycles and tree types and cover. The pollen records indicate relatively cold, arid climates during stadials and relatively warm, humid climates during interstadials.…”

Glacial-cycle simulations of the Antarctic Ice Sheet with the Parallel Ice Sheet Model (PISM) – Part 1: Boundary conditions and climatic forcing

“…Basal friction is interpolated accordingly. Thus, grounding-line migration is reasonably well represented in PISM (compared to full Stokes), even for coarse resolutions (Pattyn et al, 2013;Feldmann et al, 2014). Ice deformation (˙ ) in response to deviatoric stresses τ (and effective stress τ e = τ e (τ )) can be described according to the Glen-Paterson-Budd-Lliboutry-Duval flow law, with enhancement factor E and flow-law exponent n, ij = E • A(T , ω) τ n−1 e τ i,j .…”

“…It uses a hybrid combination of two stress balance approximations for the deformation of the ice, the shallow ice -shallow shelf approximation (SIA-SSA), which guarantees a smooth transition from verticalshear-dominated flow in the interior via sliding-dominated ice-stream flow to fast plug flow in the floating ice shelves (Bueler and Brown, 2009) while neglecting higher-order modes of the flow. Driving stress at the grounding line is discretized using one-sided differences (Feldmann et al, 2014). Using a sub-grid interpolation scheme (Gladstone et al, 2010) the grounding-line location simply results from the flotation condition, without additional flux conditions imposed.…”

“…PISM simulates sub-grid basal friction according to interpolated grounding-line location between grounded ice sheet and floating ice shelf (Gladstone et al, 2010). Hence, grounding-line migration can be reasonably well represented in PISM (compared to full Stokes), even for coarse resolutions (Pattyn et al, 2013;Feldmann et al, 2014). The sensitivity of grounding-line motion also depends on applied boundary conditions.…”

Snowfall versus sub-shelf melt: response of an idealized 3D ice-sheet-shelf system to mass redistribution

“…The idealized bed topography (Fig. 4b) is a superposition of two components and very similar to the one used in (Feldmann and Levermann, 2017): the bed component in x direction, B x (x) = −150 m − s · 0.9 · 10 −3 |x|, with bed-slope scaling factor s, is an inclined plane sloping down towards the ocean ( shelf. Ice is cutoff from the ice shelf and thus calved into the ocean beyond a fixed position (Fig.…”