Sea Ice Exchange across the Canadian Arctic Archipelago from Microwave Satellites Sea Ice Exchange across the Canadian Arctic Archipelago from Microwave Satellites T. A. Agnew 1, J. Vandeweghe 2, and A. Lambe 2 1 Environment Canada, Toronto, Canada 2 Engineering Department, University of Toronto, Toronto, Canada

Microwave Satellite Data Advanced Microwave Scanning Radiometer (AMSR): 4-year period September 2002 to June RADARSAT: for the earlier period ( ) by Ron Kwok From these satellites you can get sea-ice area flux, to obtain ice volume fluxes you need contemporaneous ice thickness which is almost never available, use proxy estimates of ice thickness from bore hole estimates in the region, ULS (submarine, moored sites) and general knowledge of ice types in the region. AMSR provides daily estimates of ice motion every 6 pixels (37.5 km) and ice concentration at resolution of 12 km. However there is no data for July and August because ice motions are unreliable due to atmospheric moisture (10-month average Sept - June) RADARSAT provides 3-day ice motions on a 5 km grid but only for northern parts of the Archipelago. Define a number of flux gates surrounding the Archipelago and calculate the sea ice area flux

Corvalis 2006 Sea ice export into the North Atlantic

Fluxgates surrounding the Archipelago } QEI-S } QEI-N

Corvalis 2006 Error in Daily Sea-ice Area Flux 1. Error in u i is 4.64 km/day, error in c i is 10%, error in c i u i is σ f = 4.74 km/day 2. For example, AG gate is L = 180 km, N s = 180 km/37.5 km = 5, error daily flux across the gate is σ T = σ f L/(N s ) ½ = 380 km Monthly error: σ m = σ T (N d ) ½ = 2100 km 2 /month Estimating Ice-area Flux across the Gate using AMSR 1. Obtain an independent ice motion estimate every 6 pixels (37.5 km ) which is interpolate to an estimate of ice motion u i at each 25 km intervals Δx along each fluxgate and similarly for concentration c i. 2. Daily Ice Area Flux = ∑ c i u i Δx SeparationObservations Mean Difference (km/day) SD of Difference (km/day) Mean Buoy Displacement (km/day) 1-day *(5 %) Sea Ice Motion Error Compare with Arctic Drifting Buoys

Corvalis 2006 Area Flux Error Estimate Gate Name Gate Symbol L (km) σ T (day) (km 2 ) σ (month) (km 2 ) σ (10-month) (km 2 ) Amundsen Gulf AG M’Clure Strait MS QEI South QEI-S~ QEI North QEI-N Lancaster Sound LS Jones Sound JS Smith Sound SS

Corvalis 2006 * Negative means export toward the Arctic Ocean **thickness from Melling (2002) Cumulative 10-month Area Fluxes (10 3 km 2 ) each year YearAGMSQEI-SQEI-NLSJSSS Average Estimated ice thickness (m) ** Average volume flux (km 3 )

Corvalis 2006 Average Monthly Area Flux for the Northern Gates Negative means transport toward Arctic Ocean

AO index and Atmospheric Circulation over the Arctic Ocean Curtesy of Ignatias Rigor

Corvalis 2006 Pressure Gradient-Area Flux Negative means transport toward Arctic Ocean Corr =.74 Corr =.77

Corvalis 2006 *Negative means export into the Arctic Ocean + Kwok (2005, 2006) **Melling (2002) RADARSAT and AMSR-E Comparison for Gates facing the Arctic Ocean Sensor\Gate AG (10 3 km 2 ) MS (10 3 km 2 ) QEI-S (10 3 km 2 ) QEI-N (10 3 km 2 ) NS (10 3 km 2 ) AMSR-E (No July, Aug) - 14 * -5 * 306 RADARSAT * -20 * 6233 Estimated ice thickness (m) 1 to ** 3 Avge. Volume (km 3 yr -1 )

Corvalis 2006 * Positive means export into Baffin Bay **thickness from Melling (2002) Cumulative 10-month Area Fluxes (10 3 km 2 ) each year YearAGMSQEI-SQEI-NLSJSSS Average Estimated ice thickness (m) ** Average volume flux (km 3 )

Corvalis 2006 Average Monthly Area Flux for the Southern Gates Positive means transport toward Baffin Bay

Corvalis 2006 *Negative means net loss of sea ice Net Change in Ice Area for the CAA (Net = AG+MS+QEI –LS–JS) Month\Year 2003 (10 3 km 2) 2004 (10 3 km 2) 2005 (10 3 km 2) 2006 (10 3 km 2) Average (10 3 km 2) Sep Oct Nov Dec Jan Feb Mar Apr May Jun month Average

Corvalis 2006 Monthly Ice Area Budget

Corvalis 2006 Monthly Sea Ice Volume Budget

Corvalis 2006 *Negative means net loss of sea ice Net Volume Advection (Net = AG+MS+QEI –LS–JS) Month\Year 2003 (10 3 km 2) 2004 (10 3 km 2) 2005 (10 3 km 2) 2006 (10 3 km 2) Average (10 3 km 2) Sep Oct Nov Dec Jan Feb Mar Apr May Jun month Average

Corvalis 2006 Advection versus Melt/Growth (growth – melt) (import–export)

Corvalis 2006 Amundsen Gulf and M’Clure Strait are net exporters of sea ice into the Arctic Ocean (fall and winter) and QEI north and south gates import ice from the Arctic Ocean. The monthly variability depends on large scale circulation (strength and location of the Beaufort Sea High). Combining both satellite results, there is net export of about 50 km 3 yr -1 into the Arctic Ocean and there is a export of 100 km 3 yr -1 into Baffin Bay. This ice is generated from within the Archipelago itself. The net ice advection each month is about 10% of the monthly change due to ice growth/melt. These fluxes are considerably less than the flux of sea ice through Fram Strait (~ 2800 km 3 yr -1 ). Conclusions

Corvalis 2006

Fluxgates across Baffin, Davis and Fram

Corvalis 2006 Note: Positive means flux southward or toward Baffin Bay ** Transport for 10-months of the year Cumulative 10-month area fluxes in 10 3 km 2 each year Year\Gate Baffin Bay Davis Strait Fram Strait Cornwallis Island Barrow Strait Lancaster Sound Average 645 ± ± 7631 ± 715 ± 427 ± 449 ± 4 Estimated ice thickness (m) Average volume flux (km 3 ) **

Corvalis 2006 Monthly Ice Area Flux Positive means transport south

Corvalis 2006 Pressure Gradient-Area Flux Negative means transport toward Arctic Ocean

Fluxgates across Cornwallis, Barrow and Lancaster

Corvalis 2006 Monthly Ice Area Flux Positive means transport south

Corvalis 2006 Note: Positive means flux southward or toward Baffin Bay ** Transport for 10-months of the year Cumulative 10-month area fluxes in 10 3 km 2 each year Year\Gate Baffin Bay Davis Strait Fram Strait Cornwallis Island Barrow Strait Lancaster Sound Average 645 ± ± 7631 ± 715 ± 427 ± 449 ± 4 Estimated ice thickness (m) Average volume flux (km 3 ) **

Corvalis 2006

Modeling Sea Ice Formation/Transport in Baffin Bay Model daily ice transport, ice production and salt re- distribution using sea ice cover from AMSR rather than estimating the surface energy budget and then melting or growing the ice thermodynamically. Previous days ice cover, advect it by the ice motion, compare with the next days ice cover, ice is grown/melted to make the advected ice field match the next days ice cover (Pedersen and Coon, 2004) We obtain each day from AMSR, sea ice concentration, sea ice type, sea ice thickness, sea ice motion Advect the sea ice using a trajectory method to reduce numerical diffusion problems

Corvalis 2006 Net Freezing 2003/ /05 Bathymetry

Corvalis 2006 Monthly Net Freezing/melting

Corvalis 2006 Salt Redistribution 2003/ /05

Corvalis 2006 * Negative means export toward the Arctic Ocean **Melling (2002) Cumulative 10-month Volume Flux in km 3 each year YearAGMSQEI-SQEI-NLSJSSS Average

Microwave Satellites Penetrate most cloud and do not require solar illumination RADARSAT resolution of 150 m and repeat coverage is 3 days ( ) RADARSAT relies on surface scattering so you cannot get a contemporaneous ice thickness directly, must rely on proxy estimates of ice thickness Advance Microwave Scanning Radiometer for EOS (AMSR-E) daily repeat coverage but resolution 6 to 12 km and no ice motion estimates in July and August ( ) AMSR is multi-channel/polarization and so has some skill in estimating thickness of FYI Define a number of flux gates surrounding the Archipelago and calculate the sea ice area flux

Corvalis 2006 Monthly Net Ice Volume Flux and Melt/Growth

Corvalis 2006 Monthly Net Ice Area Flux and Melt/Growth

Corvalis 2006 AO Index