The Co-evolution of continental ice cover and 3D ground temperature

overthe last glacial cycle

Lev Tarasov Memorial University of Newfoundland

(Thanks to A.S. Dyke, Richard Gyllencreutz, Oystein Lohne, Jan Mangerud, W. R. Peltier, and John Inge

Svendsen)

Outline

MOTIVATION: Ice-sheet evolution What makes the model results meaningful?

PHYSICS: Glacial Systems Model (GSM) DATA BRIDGING DATA and PHYSICS: model

calibration SOME RESULTS

Questions

What did the the northern Hemispheric ice-sheets do?

How do we find out? How did the changing ice-

sheets affect the ground temperature field?

How close to equilibrium is the present-day 3D ground temperature field?

Lot’s of poorly constrained components in the glacial system

Poorly constrained system; model pre(retro)dictions have generally

lacked meaningful error bars => no meaningful interpretation.Solution: 3 components:

#1: Glacial Systems Model (GSM) 3D thermo-

mechanically coupled ice-sheet model, 0.5 * 1.0 (lat/long) model resolution

VM2 viscosity model detailed surface

mass-balance and ice-calving modules

global gravitationally self-consistent RSL solver

fully coupled surface drainage solver

Model thermodynamics

3D ice thermodynamics 3D heat advection and vertical diffusion

Bed-thermal model Vertical diffusion only Non-uniform 20 layer grid spans 3 km Deep geothermal heat flux field from Pollack et al (1993) Simple: scalar thermal conductivity and heat capacity Standard permafrost treatment using effective heat capacity Parametrized TTOP model (based on Smith and

Riseborough, 2002) for near surface ground temperature as a function of air temperature

Large-scale and Long-time scale Horizontal resolution: 1 by 0.5 degrees

Climate forcing

Last Glacial Maximum (LGM) precipitation and temperature from 4 (6 for N. A.) highest resolution Paleo Model Intercomparison Project GCM runs

Mean and EOF fields

Present day observed fields

120 kyr climate forcing(based ss09 chronology)

Embrace uncertainty: Lots of ensemble parameters

3(5 for North America) ice dynamical 13(16) regional precipitation

LGM precipitation EOFs most significant for North America 4(4) temperature 4(4) ice calving 4(2) ice margins 1 model version = 29 (32)

Need constraints -> #2: DATA

Oct/06 margin chronology

Deglacial margin chronology

(Dyke, 2004) 36 time-slices +/- 50 km uncertainty Margin buffer Margin forcing:

2 ensemble parameters control extent of adjustments to surface mass-balance

Calibration minimizes amount of margin forcing

RSL data; site weighting (A. S. Dyke and U. of Toronto RSL database)

Noisy data and non-linear system => need #3: calibration and error bars

Bayesian calibration

Sample over posterior probability distribution for the ensemble parameters given fits to observational data using Markov Chain Monte Carlo (MCMC) methods

Sampling also subject to additional volume and ice thickness constraints

North American calibration also uses strand-line observations

Large ensemble Bayesian calibration

Bayesian neural network integrates over weight space

effectively generate millions of model runs

LGM characteristics

Dynamic ice-sheet streaming

constrained by sub-glacial till availability and thermodynamics

Large Keewatin ice-dome

Validation of glaciological model

RSL fits are overall as good as that of the non-glaciological ICE5-G model

North American model also validated against GRACE observations for present-day rate of mass-change

Surface topopgraphy similar to that infered on the basis of striations and landform orientation

Thermodynamic results (North American are published, Eurasian

are preliminary)

Borehole comparison to observations (93.72W, 51.03N, with error bars !!!)

Caveat, dependence on thermal conductivity (66.64W, 50.21N)

Present-day permafrost distribution

Permafrost depth evolution: impact of surface ice

Present day proximity to equilibrium

European LGM ice-sheet (Problem: less ice-cover)

Kola Borehole Temperature profile

Kola latitudinal transect: Present Proximity to Equilibrium

Kola

Ice area evolution comparison : North America versus Europe

Summary

Glaciological results North America: Large Keewatin ice dome Dynamic ice-sheets

Permafrost/ground temperature: Large variations in permafrost depth over glacial cycle Results suggest southern half of the Laurentide ice-sheet

was often extensively warm-based Significant present-day dis-equilibrium

Impact of ice on present-day ground temperature field: Is small outside of permafrost zone

Bayesian calibration method links data and physics (model) -> rational error bars

Calibration selects a climate forcing that produces a significantly different present day ground thermal field than that from a simple average of PMIP 1 and 2 General Circulation model results

For more results, refer to Tarasov and Peltier, JGR 2007

Request/Plea

Need an online database of deep borehole observations (especially for

Eurasia)!!!!

Probability of warm based ice over last

20kyr for N.A.

Present-day permafrost distribution

Kola Transect: Present-day ice versus no-ice models

Kola borehole time transect

Present ensemble mean permafrost depth for Europe