This is an old revision of the document!
Go back to the main core experiment design page.
Go back to the main working group page.
Please note, this page is a work in progress and is not ready for discussion yet.
Please use the Discussion section below to specifically comment on the choice of ice sheet reconstructions for the core experiment.
For the core experiment, there is a choice of two global ice sheet reconstructions:
Please one of these reconstructions for your Last Glacial Maximum (LGM) experiment, as per the LGM working group requirements, and continue to use the same reconstruction through the transient last deglaciation core simulation.
Those groups that are able may wish to carry out two simulations; one with each ice sheet reconstruction.
The ice mask in this reconstruction is fractional. For the purpose of the animations (below), we have used > 80 % ice cover per grid cell. The timestep is 500 years.[1-2] 1)
The information in this section was provided directly by Dick Peltier et al., October 2014:
Time dependent ice-equivalent contribution to eustatic sea level rise [relative to present day] from each of the primary geographical regions from which grounded ice loss occurred during the [last] deglaciation process. 2)
|Ice-equivalent contribution to eustatic sea level rise (m) 3)|
|Final ICE-4G||ICE-5G v1.2||ICE6G_C|
|N. America (incl. Inuit area)||54.92||83.71||87.01|
|Greenland & Iceland||5.43||2.45||2.39|
|N. America (incl. Innuit area)||64.24||81.47||78.82|
|Greenland & Iceland||6.38||2.49||2.41|
Dick Peltier has suggested that ICE6G_C topographies could be provided as smoothed fields:
…we could…provide these [ICE6G_C] topographies in the form of the smooth fields obtained by projecting them onto the set of spherical harmonics employed in a 1 degree by 1 degree model in the CMIP5 class. The results you obtain when you do this are illustrated in [Peltier and Vettoretti (2014)]. Would you rather have these fields in the smoothed form actually seen by such a climate model? This might be a good idea since some groups may be employing grid point models and these groups would probably appreciate being given a smooth topography field to start with.
The ice mask in this reconstruction is based on 100 % ice or no ice. The timestep is currently 1000 years, but could be provided at 500 years, if desired. [4-7]4)
Provided by Lev Tarasov, October 2014:
The data set has surface elevation (ice if present, otherwise ground) relative to contemporaneous sealevel, so the land/seal mask is the 0 elevation contour. It also has an ice mask.
…[It does not currently include] a floating ice mask. I can easily add that later if someone is modelling sub ice shelf circulation…
The Eurasian (EA) and North American (NA) components are from Bayesian calibrations of a glaciological model. The Antarctic (ANT) component is from the recently published scored ensemble of 3344 model runs. The Greenland (GR) component is my old hand-tuned GrB model. The constraint data sets for these models includes RSL (all), marine limits (NA), present-day vertical velocities of the solid earth (NA and EA), geologically inferred deglacial margin chronologies (NA and EA), strandline proxies for pro-glacial lake levels (NA), ice core vertical temperature profiles (GR), cosmo data for constraining past ice thickness (ANT), and present-day ice configuration (ANT and GR).
Details on the constraint data sets for the NA, GR, and ANT components are in the above refs. The Eurasian component is in the process of completion and uses the geologically inferred DATED deglacial ice margin chronology which includes max/min uncertainty isochrones for each timeslice. Each of these glaciological models employed fully-coupled visco-elastic isostatic adjustment of the solid earth.
The ANT model uses the dynamical core of the Pennstate model that includes shallow-shelf ice physics (the other 3 components [have] just the shallow ice approximation).
These 4 components have been combined under GIA post-processing for a near-gravitationally self consistent solution (the approximation is explained in my 2004 QSR paper and has been tested against complete GIA solutions). The global combined solution includes ICE5-G components for Patagonia and Iceland for topography (but they were not transfered to the ice mask). There is likely a 10-15 m eustatic equivalent shortfall of LGM ice, the “missing ice” issue that has long challenged GIA-based deglacial reconstructions (uncertainty associated with proxies, tidal changes, and earth viscosity structure).
Although ice meltwater fluxes should not be prescribed for the Core simulation, groups may wish to run alternative simulations with meltwater fluxes to the oceans.
A timeseries of global meltwater routing (river mouth discharge) that is consistent with Lev Tarasov's ice sheet reconstruction will be provided.
See the alternative full transient simulations page for more info. [not live yet]
Please think about the following points and add any comments on these or any other aspects of the experiment design to the discussion section below: [Topics will be added here as they are raised below or by email.]