sea-ice in roms
DESCRIPTION
Sea-Ice in ROMS. Kate Hedstrom , UAF. Outline. Equations Dynamics Thermodynamics Code cppdefs.h SeaIce directory ice.in Examples Future plans. Dynamics. Momentum equations: Viscous-plastic term:. Ice Strength. Linear with thickness: Quadratic with thickness:. More Dynamics. - PowerPoint PPT PresentationTRANSCRIPT
Sea-Ice in ROMS
Kate Hedstrom, UAF
Outline• Equations
– Dynamics– Thermodynamics
• Code– cppdefs.h– SeaIce directory– ice.in
• Examples• Future plans
Dynamics
• Momentum equations:
• Viscous-plastic term:
Ice Strength
• Linear with thickness:
• Quadratic with thickness:
More Dynamics
•Rearrange VP:
•EVP version:
Solution• Stress tensor equation is
timestepped explicitly• Young’s modulus E depends on ice
thickness to keep solution close to VP solution
• Ice velocities then timestepped with air/water stresses, Coriolis
• Since it is all explicit, easy to parallelize (ROMS tiles)
Thermodynamics
• Need to compute all the ice growth/melt terms shown
Evolution of Tracers
• Ice volume is Ah• Ice concentration is A,• Phi is tunable parameter, value
depends on sign of Wao
i
Advection of Tracers
• Advection uses MPDATA, upwind or full – Ice volume,– Ice concentration– Snow volume– Ice enthalpy– Etc.
• Followed by limiter 0 <= A <= 1
More Thermodynamics
• Heat fluxes are computed through the ice and snow
• Temperature is linear in the ice• FT has an oceanic log layer
Ice cppdefs.h
• ICE_MODEL – turn on whole seaice model• ANA_ICE – analytic initial conditions• ICESHELF – ice shelves – something else
again entirely• ICE_ADVECT – turn on advection
– ICE_SMOLAR– ICE_UPWIND
cppdefs.h
• ICE_MOMENTUM – turn on rheology– ICE_EVP
• ICE_MOM_BULK• ICE_SHOREFAST – ice slows as it fills water
column• FASTICE_CLIMATOLOGY – read in locations
where ice doesn’t move• ICE_STRENGTH_QUAD – quadratic ice strength
as function of thickness
cppdefs.h
• ICE_THERMO – turn on thermodynamics– ICE_MK– ICE_BULK_FLUXES
• OUTFLOW_MASK – turn on outflow cells – Reduces ice strength and viscosity in some cells
– Cells specified in set_masks.F
Albedo
• ROMS default has shortwave radiation being net into the water– Preprocess shortwave to account for albedo
• Ice/snow/melt ponds affect albedo, so incoming shortwave must be before albedo corrections– Don’t preprocess shortwave
• Various options exist…
Albedo Options• ALBEDO – needs clouds
– Water only
• ALBEDO_CURVE – function of latitude– Water only
• ALBEDO_FILE– Ice and water
• ALBEDO_CSIM– Ice only
• ICE_ALB_EC92 – Ebert and Curry– Ice only
SeaIce Directory
• seaice.F – calls all the rest• Boundary conditions
– i2d_bc.F, tibc.F, uibc.F, vibc.F• ice_advect.F – advect the ice tracers
– ice_smolar.h– ice_enthalpi.F
• ice_thermo.F – compute heat fluxes, melt/freeze rates
– ice_mk.h
SeaIce Directory
• EVP rheology
– ice_evp.F– ice_evp_sig.F– ice_elastic.F
• ice_frazil.F– Computes frazil ice formation– Called from step3d_t.F– Checks for negative ice growth and crashes
ROMS if found
SeaIce Directory
• ice_limit.F– Clamps on ice fields
• ice_spdiw.F– Compute shear between ice and water in upper
mixed layer
• ice_vbc.F– Sets ice-water and ice-air stresses for ice
momentum equation
SeaIce Directory
• ini_strengthcoef.F– Quadratic ice strength depends on grid spacing– Set 2D coefficient during initialization
• ice_flux_rst.F– Timestepping in ROMS calls seaice before output– Seaice computes heat and momentum fluxes for
ocean– To get perfect restart, we save the fluxes and use
saved fluxes instead of timestepping ice in first step after restart
Boundary Conditions
• LBC options– Mix: clamped on inflow, gradient on outflow– Otherwise much like the ocean
• Outflow cells– Specify in set_masks.F
• Nesting– Offline nesting gives trouble for ice BC– Online nesting is reputed to work
ice.in
• Lice – logical for ice (ignored)• dtice – ignored, uses ocean dt• nevp – number of EVP iterations per step• min_h, min_a, max_a – ice limiters for
thickness and concentration• Hout, Aout for ice variables• Various things I never change – read about
them in the file
Examples
• From most mature to least mature– Northeast Pacific (NEP)– Arctic– Arctic2– Cook Inlet (NWGOA)
• All are ice-ocean only, just one grid
Northeast Pacific
• 11 km resolution• GLOBEC, BEST,
BOEM• Looking at
interannual variability
• First-year ice only in Bering
Ice Timeseries
Ice Concentration
Seth Danielson
Summary of Bering Sea
• Ice extent sensitive to incoming solar radiation (including albedo)
• Once tuned, the model did a good job• Spring melt is too slow (didn’t use
melt pond albedo)
Arctic 1
• Telescoping grid, 5 to 20 km
• Boundary at Bering Strait
• 20 year hindcast for BOEM
• Fastice climatology
Ice Thickness
Seth Danielson
Satellite Model - obs
Ice Concentration
Seth Danielson
Timeseries
Seth Danielson
Summary of Chukchi-Beaufort Seas
• Ice too mobile, getting too thick off Siberia, too thin off Alaska
• Too thin areas grow more ice, making ocean too salty
• Excess ice melt in May-June, not enough later in summer
• Monthly anomalies have skill except for a few problem summers (1988, 2002, 2003)
Arctic 2
• Includes Bering Sea for looking at Bering Strait flow for NPRB
• Includes Canadian archipelago
• Quadratic ice strength
Ice Thickness
Northwest Gulf of Alaska
• 1.5 km resolution• Includes tides and
WET_DRY• Fresh water from
land runoff model• Try out CVMix?• For BOEM
Papers
• Hunke and Dukowicz, An elastic-viscous-plastic model for sea ice dynamics, JPO, 27, 1849-1867 (1997)
• Hunke, Viscous-plastic sea ice dynamics with the EVP model: linearization issues, JCP, 170, 18-38 (2001)
• Mellor and Kantha, An ice-ocean coupled model, JGR, 94, 10,937-10,954 (1989)
Papers
• Hakkinen and Mellor, Modeling the seasonal variability of a coupled Arctic ice-ocean system, JGR, 97, 20,285-20,304 (1992)
• Overland and Pease, Modeling ice dynamics of coastal seas, JGR, 93, 15,619-15,637 (1988)
• Steele, Mellor and McPhee, Role of the molecular sublayer in the melting or freezing of sea ice, JPO, 19, 139-147 (1989)
Plans
• Eight years ago I wrote:– Replacing Mellor-Kantha with CICE 3.1
• Now aiming for CICE 5. What does it offer?– Multiple ice categories and temperature
levels– All the recent sea ice model improvements
to melt ponds, rheology, drag, etc.
• Still missing fast ice, embedded ice
Random Tips
• Circumpolar considerations– GLOBAL_PERIODIC– grid_coords.F (floats, stations)
• Beyond ROMS trunk– BIO_UMAINE, aka CoSiNE
• ARCTIC in 2012 ROMS manual• BERING coming soon
– With WRF