Difference between revisions of "Participating Models"

From Interactive System for Ice sheet Simulation
Jump to: navigation, search
(Table 1. Characteristics of Various Whole Ice-Sheet Models)
Line 1: Line 1:
==Table 1. Characteristics of Various Whole Ice-Sheet Models==
+
==[[Table 1]]. Characteristics of Various Whole Ice-Sheet Models==
  
  

Revision as of 11:43, 30 November 2009

Table 1. Characteristics of Various Whole Ice-Sheet Models

Characteristics Glimmer CISM est. Q2 2009 PISM MAINE GLAM
Domain
Flowline (1d or 2D); Plan view (2D or 3D)
3D 3D 3D map-plane 3D
SPACING
Average grid spacing; Adaptive grid
uniform grid > 10 km uniform grid > 2.5 km adjustable; non-adaptive grid can run whole ant at 20 km (70,000 nodes) with embed for higher res uniform grid (resolution limited by desired model run-time)
GRID
Finite-difference; Finite-element; Eulerian/Lagrangian
FD FD + incremental remapping scheme Eulerian; finite-difference Finite Element quadrilaterals FD
FLOW APPROXIMATION
Shallow ice; Shelfy-Stream; Higher-order; Full Stokes; other
SIA SIA + Price/Payne 1st Order + Pattyn/Johnson 1st Order Hybrid model: Shallow ice (SIA) + Shelfy-Stream (Schoof) shallow ice 1st – order SIA (e.g. Pattyn/Blatter models)
THERMODYNAMICS
Thermomechanical; Polythermal
Thermomechanical Thermomechanical + Polythermal of Greve Thermomechanical + Polythermal (Enthalpy Formulation) Thermomechanical (1D columns with explicit vertical advection and diffusion, with horizontal advection as an additions “source” (negative heat)) thermomechanical
BASAL SLIDING
Weertman sliding law; Coulomb plastic sliding law; Budd-type sliding law
Proportional to driving stress and inverse water layer thickness Flexible with linear and plasitc till being the most prominent Coulomb plastic or Weertman Weertman modified with a lubrication factor proportional to “amount” of water at the bed linear-viscous sliding law (“B2” param.) w/ iteration for plastic bed sliding
HYDROLOGY
Surface, internal, basal water treatments
Conservative steady state basal water routing Surface and basal water treatments basal meltwater model: controls bed strength Basal melt water from thermo-calc used a source too diffusive-advective continuity model for basal water local basal water production and storage; sub-model to link production rate to plastic till yield strength
SURFACE MASS BALANCE
Positive degree day; Surface energy balance; empirical method; other
Positive degree day Surface energy balance + downscaling of GCM data based upon elevation classes Positive degree day Mean annual temp from latitudinal and elevation lapse rates, accumulation from MAT, ablation from PDD with lat-dependent amplitude around MAT PDD scheme
CALVING
Calving "law"; calving mechanics
Heuristic Improved calving law based on stress/strain rates? Fixed calving front Longitudinal extension at unbuttressed grounding line yields thinning rate at GL added to local mass balance, modified by "Weertman" parameter (1-no buttressing, 0-full buttressing) Fixed calving front
SPIN-UP/INITIALIZATION
# glacial cycles; req'd initial fields
One glacial cycle at minimum, preferable to do 2-3 Concerns about performance make this a problem. HO physics may prevent long initialization periods. May have do some hybrid SIA/HO spin up, or find very powerful computers. arbitrary number of glacial cycles. Surface elevation, bedrock elevation, geothermal flux usually a glacial cycle, but 50Ka is usually enough HO solver performance constraints make a simple steady-state spin up the most practical choice
OTHER
Explicit mass conservation
Well documented problems (see EISMINT II) papers. We appear to be no worse than other comparable models. While this error relates to numerics, other errors arising in PDD schemes are unavoidable until a better (surface energy balance) scheme is used. As before, again, placing some hope in better advection schemes. yes, adaptive time-step yes Choice of solvers for mass conservation when using HO dynamics (e.g. Bueler explicit scheme, incremental remapping scheme)
Grounding line migration Major problem area. Currently I think it should be forced as part of the experimental setup. As with before, but finer grids and incremental remapping scheme offer some hope.. yes GL determined by location surface drops below flotation height Currently in 2d (x,z) plane only, using fine (~1km) grid spacing and grounding line interpolation ala Pattyn et al. (2006, JGR v.111)

Table 1 (Cont). Characteristics of Various Whole Ice-Sheet Models

Characteristics PENN STATE SICOPOLIS Chicago PENN STATE 2-D
Domain
Flowline (1d or 2D); Plan view (2D or 3D)
3D Greenland, Antarctica or northern/southern hemisphere from tropics to the Pole 3D flowline and plan view mode flowline, 2-d
SPACING
Average grid spacing; Adaptive grid
40 km, or nested 10 km to 5 km 10 km for Greenland; 20 km for Antarctica variable resolution average, typically ~10 km Depends on the simulation… for whole ice sheet, average ~10 km; adaptive
GRID
Finite-difference; Finite-element; Eulerian/Lagrangian
finite difference Finite difference; stereographic projection or lat/log grid finite difference in horizontal/spectral in vertical, semi-Lagrangian advection Finite Element; Eulerian
FLOW APPROXIMATION
Shallow ice; Shelfy-Stream; Higher-order; Full Stokes; other
Heuristic combination of shallow ice and shelfy stream Shallow ice Other SIA ready to go, Higher-order ready in isothermal mode
THERMODYNAMICS
Thermomechanical; Polythermal
Thermomechanical Thermomechanical; can be ran as polythermal (but I did not run it myself as polythermal) Thermomechanical SIA: thermomechanical; Higher-order: presently isothermal
BASAL SLIDING Weertman sliding law; Coulomb plastic sliding law; Budd-type sliding law Weertman sliding law Weertman sliding law adjustable, tested with Weertman, Budd-type and Mohr-Coulomb SIA: Weertman linear viscous; other powers available for Higher-order
HYDROLOGY
Surface, internal, basal water treatments
No sliding when base below melt point Vertical penetration of surface meltwater to the bed; acceleration of basal sliding No sliding when base below melting point, does not do any accounting of basal water Through parameterizations/budgeting
SURFACE MASS BALANCE
Positive degree day; Surface energy balance; empirical method; other
Positive degree day Positive degree day, but Miren Vizcaino ran it also with a of surface energy balance Positive degree day PDD
CALVING
Calving "law"; calving mechanics
ocean sub-ice melt rate prescribed No explicit calving (under construction) Several mean-field "calving laws"implemented, heuristic implementation of rift initiation and propagation Will be added; presently thickness related
SPIN-UP/INITIALIZATION
# glacial cycles; req'd initial fields
can be run for many glacial cycles glacial cycles or topography + T(z) Can be run over many glacial cycles, Initial parameters: surface elevation, bedrock elevation, geothermal flux, location/depth of sediment This depends on the experiment as stated in my earlier email. If thermal profile and melt/freeze boundaries matter to the simulation, 250kyr of thermal spin-up and then one or two glacial cycles; need bed elevation, surface elevation, surface temp (mean annual and summer average), surface accum, geothermal flux, upper mantle viscosity, some sense of distribution of basal friction coeff, sea level fluctuation, and any glacial geologic constraints on reconstructions
OTHER
Explicit mass conservation
conserves mass no no diffusion formulation for SIA; advection for higher order
Grounding line migration yes no (under construction) yes - with concerns about stability yes