1. Progress and Assessment of the Arctic subpolar gyre State Estimate (ASTE) An T. Nguyen, Patrick Heimbach, Ayan Chaudhuri, Gael Forget, Rui M. Ponte, and Carl Wunsch MIT & AER Estimating the Circulation and Climate of the Ocean http://ecco-group.org

2. Rationale Produce a regional eddy-permitting state estimate focused on: –Arctic ocean –North Atlantic, especially subpolar gyre Special emphasis on: –Arctic sea ice-ocean interactions –a dynamically consistent Arctic hydrography –understanding of the planned subpolar gyre mooring array (OSNAP) –Arctic-subarctic exchanges –Subpolar-subtropical exchanges

3. ECCO-production version 4 (ECCO-v4) ECCO MIT JPL effort Focus: State estimation + Climate analysis Assimilate a large suite of existing in-situ & satellite data using the adjoint MITgcm, global coupled ocean and sea-ice Grid: lat-lon-cap: in lon/lat -90 to 60 o N, polar projection 60-90 o N hierarchy of grid spacing derived from a 1/48 o parent grid. 40 (Arctic) to 110km (equator) 50 vertical levels Atmospheric forcing: ERA-interim Model Controls: initial conditions, atmospheric state, 3-D ocean parameters Production period: 1992-2011 Current status: iteration 9

4. Data: ECCO-v4

5. ECCO-v4 iter-9: hydrography cost reduction ΔS at 100mΔS at 300mΔS at 1000m -0.2 0.2 Iteration 0 ΔS at 100m ΔS at 300m ΔS at 1000m -0.2 0.2 Iteration 9

6. Nested within ECCO-v4, which fits observations well Arctic Ocean, Nordic Seas, Pacific Ocean north of 47.5 o N, Atlantic Ocean north of 32.5 o S lat-lon-cap grid 14 km in the Arctic MITgcm coupled ocean and sea-ice Optimization period: 1992-1997, 1992-2012 ASTE set up (1) 5000 4000 3000 2000 1000 0 Depth [km] Bathymetry: - Smith and Sandwell v14.1 and IBCAO v2 (2km, north of 64 o N). - Special treatments applied to ensure proper depth in important channels in the Nordic, Chukchi and Bering Seas, Caribbean Sea. 90x90 grid points

7. ASTE set up (2) Initial conditions: ECCO-v4 iter-9 (ocean), Polar Science Center (ice, Jan 1992) Open boundary conditions: ECCO-v4 iter-9; Atlantic, Pacific, Gibraltar Strait. Control variables: - initial conditions - time-varying surface atmospheric state, - time-varying open boundary conditions, - 3-D ocean parameters.

8. Data: high-latitude dedicated data set

9. ITP locations Nordic Sea ICESat, Oct 2007ICE velocity, Dec 1992

10. Uncertainty: hydrography [Forget and Wunsch., 2007] updating for polar region in progress. Measurement errors Model representation errors Spatial/temporal dependency? Three important ingredients are needed for optimization: 1. the model 2. the data 3. "useful"/credible uncertainty estimates

11. Uncertainty: Atmospheric forcing: [Chaudhuri et al., 2013a, 2013b submitted]  DLW [W/m2]  DSW [W/m2]  Tair2m [degC]

12. Uncertainty: Sea-ice: Reprocessed Sea-ice concentration ftp://ftp.dmi.dk/pub/Users/Leif.Toudal/SICCI/Documents/, osisaf.met.no/docs/pum_seaicereproc_ss2_v1p3.pdf Sep 2005 Ice concentration Total uncertainty

13. Uncertainty: Sea-ice: Obs Model, iter0 Model, iter33 Fenty, 2010, Ph.D. thesis Fenty & Heimbach, JPO, 2013a,b

14. Boundary conditions: Atmospheric “corrected” ERA-interim: JRA25run minus ICESatERArun minus ICESat ICESat thickness [m], Oct 2007 15-30W/m 2 Optimized sea-ice parameters [Nguyen et al, 2011] correction: 5% reduction in dlw 70-90 o N  7W/m 2 (winter) to 15W/m 2 (summer) (DRAKKAR: 1-1.5degC reduction)

15. ASTE iter-0: Sea-ice – thickness and concentration ERArun minus ICESat ERAdlwcorrected minus ICESat ICESat thickness [m], Oct 2007 SSMI, Sept, 2008 ERArun, Sept, 2008 ERAdlwcorrected, Sept, 2008

16. In progress / Outlook Incorporate ECCO-v4 iter-9 atmospheric adjustments at low latitude Finalizing uncertainty fields Short optimization period 1992-1997 to adjust initial conditions Optimization Add to cost function: - Fram Strait, Bering Strait CTD/mooring data, - sea-ice thickness+velocity Long-term perspective is the development of a hierarchy of high-resolution state estimates from globally- constrained coarse-resolution solutions to regional very high-resolution solutions. Thank you

17. ECCO-v4 iter-9: hydrography cost reduction Δθ at 100m Δθ at 300mΔθ at 1000m 1 Iteration 0 Δθ at 100mΔθ at 300mΔθ at 1000m 1 Iteration 9

18. Optimization convergence 1996-1997 Fenty, 2010, Ph.D. thesis Fenty & Heimbach, JPO, 2013a,b

19. Iteration 33 Uncertainty: hydrography [Forget and Wunsch., 2007] updating for polar region in progress. Iteration 0 Fenty, 2010, Ph.D. thesis Fenty & Heimbach, JPO, 2013a,b