Assessment of the ECCO2 optimized solution in the Arctic An T. Nguyen, R. Kwok, D. Menemenlis JPL/Caltech ECCO-2 Team Meeting, MIT Sep 23-24, 2008

Outline:  Optimized solution in the Arctic  Ocean: hydrography fresh water & heat content heat & volume fluxes  Sea ice: Extent  summer 2007 Velocities & fluxes Thickness Volume  Summary Arctic solution total cost reduction Greenland Sea

Regional optimized solution:  Data: Sea-ice velocity, sea-ice thickness, CTD profiles, Initial conditions, Forcings  Parameters: salt-plume parameters air / sea-ice drag coefficients sea-ice and ocean albedos sea-ice strength and H0 initial conditions freshwater content GMredi

Sea ice velocity

Sea ice extent

c37 : NCEP c47 : c37, WOCE climatology, without specifying net shortwave radiation, bug fixes in growth.F, including Peter Winsor's new monthly Arctic Ocean runoff c55 : first blend integration (c49), more realistic initial sea-ice conditions from a PSC estimate SEAICEadvSnow=.true., SEAICEuseFlooding=.TRUE. c62 : c55, but with blend forcing and blend wind  NCEP at high lat arc1 : First integration of Arctic domain, open boundary, NCEP wind, c55 param Passive Microwave data: Sea Ice flux: Fram Strait greenland Annual c37 c47 c55 c62 arc1 Mean(  ) Std(  ) %  model – data

Sea Ice flux: comparison

Sea ice extent: Arctic Ocean

Sea ice thickness Apr May Aug Apr Sep Aug Oct Arctic sea ice: too thin Mean STD Apr-May: Aug-Sep: Oct: greenland siberia alaska ULS data from submarine:

Sea Ice thickness: ULS data at Fram Strait: greenland

Sea ice volume Note: c37: specifying net shortwave radiation c47: remove net shortwave radiation specification and fix growth.F  reduces sea-ice volume c55 & c62: starts with higher initial sea-ice thickness  increases volume c57: increase dry-ice & dry-snow albedos  increases ice-thickness & volume c56: increase wet-ice & wet-snow albedos  higher increase in volume Dropping too fast? Positive trend 96-02

Sea ice velocity Expect: ice velocity direction to the right of wind at ~ o Model: Turns to the right, but not enough Sometimes: parallel to the wind dir turning angles correlate with |  vel|

Sea ice velocity Angle difference between data and model ice vel Velocity residuals (mm/yr) Sensitivity experimen t %cube Oct-Dec Jan-May Mar-May %NCEP? %NCEP %NCEP %NCEP %75%CORE %NCEP %NCEP Consistently high, essentially // to the wind dir

Summary-1: Sea Ice 1.Sea Ice flux:  over-estimates net ice export across Fram Strait by 40% to 60% 2.Sea ice thickness/extent/volume:  too much ice extent in winter, too little in summer  ice thickness small relative to ULS data in Arctic ocean  closer to data in Greenland sea  ice volume too small, with positive trend! 3.Sea ice velocity:  parallel to wind direction, consistently to the left of data  turning angles correlate with magnitude of  vel

Arctic ocean (5): Fresh water volume: Expect: ~7.5x10 4 km 3 Too low

Arctic ocean : freshwater input through Bering Strait Expect: ~0.053Sv [Aagaard and Carmack, 1989]

Arctic ocean: SST and Dynamic Topography:

Summary-2: Arctic ocean 1.Fresh water volume: Between 60,000km 3 to 67,000km 3 2.Fresh water input through Bering Strait: ~ 0.031Sv 3.Arctic domain:  SST does not drift  ETAN drifting  Ice velocity across Bering Strait: significantly larger than both c47 and c55  ice import increases by 8% to 10% (NOT SHOWN HERE)

END

List of things to do (04/13/07): 1)Fix drifts in Arctic domain 2)Fix vertical mixing  fix ice thickness + ice vol + fresh water vol? 3)Jinlun’s comments: Initial condition for ice thickness: in the wrong season? Ice extent problem in summer: thermodynamics Tear-drop vs. elliptical yield curves: might not solve discrepancies of vel near coast Need to check if stresses lie on the yield curve

Arctic ice thickness PDF too thin, lacking modes

Sea Ice strength