NEWS: PISM and PETSc 3.7
Simulated ice extent and velocity in April (left) and November (right) of 1995.
|The Autumn of break-ups: When Jakobshavn Isbrae lost its floating tongue|
|investigators:||A. Aschwanden, M. Fahnestock, M. Truffer, and R. Motyka|
|venue:||2015 AGU Fall Meeting|
Jakobshavn Isbrae, Greenland's fastest-flowing outlet glacier, lost its floating tongue in 1995, an event which is often attributed to changes in ocean temperature. This poster and movie show the results of PISM simulations of this event, based on a step increase from 180 m/yr to 225 m/yr in sub-shelf melt rate during 1995 (Motyka et al. 2011). The simulations are started from reasonably-detailed observations of the 1985 state of the outlet glacier. A high-resolution HIRHAM5 reanalysis (Langen et al. 2015) is used for the atmospheric 1989–2011 climate. The results show that general patterns are simulated correctly, with ice speeds which almost double after break-up of the floating tongue. The timing of the break-up is too early and too fast, but these simulations do not include the “ice rumple” (Echelmeyer et al. 1991), which may add stability to the floating tongue.
A new paper in the Proceedings of the National Academy of Sciences by J. Feldmann and A. Levermann, of the Potsdam Institute for Climate Impact Research, uses PISM simulations to show that nearly-complete WAIS collapse is triggered by present-day melt rates in the Amundsen Sea. Modeled WAIS deglaciation follows after relatively-short (60–200a) periods in which the present-day sub-shelf (i.e. ocean-caused) melt rates are sustained.
The simulations use conservative assumptions about, and (necessarily) modeling of, the interaction of the ice sheet with the ocean and atmosphere. In particular, subshelf melt rates for the present ice shelf geometry are taken from Finite Element Sea Ice-Ocean Model (FESOM) results. These are then extended to the evolving cavity geometry by a diffusive algorithm into regions below sea level, but with a pressure adjustment using the ice shelf base elevation. This leads to melt rates further inland that are similar to corresponding present-day-cavity-geometry-induced melt rates.
In most other ways this application of PISM is as expected, though at high (5km) resolution and using a full suite of marine ice sheet submodels: 50ka spinup, SIA+SSA model with plastic till, subgrid motion of the calving front, ocean-water stress boundary condition at the calving front, the "eigen-calving" calving law, and an interpolated grounding line.
The results of the simulations are most easily understood by seeing what happens:
This work appeared today, 2 November 2015. It is already featured in commentaries at the Washington Post, The Guardian, and Bloomberg Business News. It is also featured in Nature journal's "News:Explainer", and in Science magazine's "Latest News".
The last of these includes this high-level view from two well-known students of the behavior of Amundsen Sea-sector glaciers:
“This paper does confirm what we hypothesized, that knocking out the Pine Island Glacier and Thwaites takes down the rest of the West Antarctic Ice Sheet,” says Ian Joughin, a glaciologist at the University of Washington, Seattle, who co-authored last year’s Science paper. He adds, however, that the model’s weakness is its [temporal] resolution; it shows the destabilization of the glaciers occurring roughly 60 years from now, whereas present observations suggest that collapse is already underway. As a result, Joughin says, the model’s time scale for the collapse is probably too long. “It’s more likely measured in centuries rather than millennia.”
Indeed, “the jury is still out” on whether Feldmann and Levermann’s study got the time scale right, [Eric] Rignot [of the University of California, Irvine] says. The long-term evolution of an ice sheet “is a very complex modeling problem. Some of the variables controlling the models are not all that well known,” he adds, including forces such as winds, ocean circulation, and how icebergs calve. “There is not a model out there that is getting it right, because they all have caveats. I think the discussion is ongoing, and is only going to be more interesting with time.”
As a result of the buzz around Winkelmann et al. (2015)'s modeling of the effect of full conversion of available fossil fuels in the ground into atmospheric CO2, using PISM for determining ice dynamics/response timescale, on 5 October our local paper the Fairbanks Daily News Miner featured PISM. The content is a bit warped by scientist-to-journalist transmission issues, but we are happy to have local recognition of this UAF-lead project!
PISM is jointly developed at the University of Alaska, Fairbanks (UAF) and the Potsdam Institute for Climate Impact Research (PIK). For more about the team see the UAF Developers and PIK Developers pages.
UAF developers, who are in the Glaciers Group at the GI, are supported by NASA's Modeling, Analysis, and Prediction and Cryospheric Sciences Programs (grants NAG5-11371, NNX09AJ38C, NNX13AM16G, NNX13AK27G) and by the Arctic Region Supercomputing Center.