1. Using Global Ocean Models to Project Sea Level Rise Robert Hallberg NOAA / GFDL

2. Plausible 1 m ice-sheet dynamics contribution to sea level rise Sources of uncertainty in 2100 global mean sea level projections: Forcing scenario (~20 cm range at 2100 for GFDL-CM2.1) Ocean heat storage (steric rise) (13 to 32 cm in IPCC AR4) Land ice (except ice sheets) (4 to 19 cm in IPCC AR4) Ice sheet surface mass balance (-10 to 4 cm in IPCC AR4) Dams & land water ±3 cm/century? (Lettenmaier & Milly, Nature Geo. 2009) Antarctic & Greenland Ice sheet dynamics changes – ? Plausible range 20 to 110 cm by 2100 (Pfeffer et al., Science 2008) Projected Global Mean Sea Level Rise Reservoir sizes in Sea Level Rise equivalent: Mountain glaciers & ice caps – 0.3 m Greenland Ice Sheet – 7.3 m West Antarctic Ice Sheet (marine) – 5 m East Antarctic Ice Sheet (land) – 51.6 m (Uniform warming of ocean ~0.5 m °C -1 )

3. “Observed” and Modeled Decadal Mean Sea Surface Height 10-Year average inferred from observations, 1992-2002 A GFDL coupled climate model with 1990 atmospheric CO 2

4. Global SLR from Ocean Thermal Expansion in IPCC 4 th Assessment Report (AR4) Models (2007) Note: Models have no ice-sheet dynamics!

6. 21 st Century Ocean Dynamics SLR Anomalies from the Global Mean in 12 IPCC AR4 Coupled Models Yin, Griffies & Stouffer, J. Climate 2010 (2091-2100 SSH) – (1981-2000 SSH) in m Hatching = Signal exceeds 1.5 standard deviations of ensemble

7. Sea Level Trends in GFDL’s CM2.1 Coupled Model Yin, Griffies & Stouffer, J. Climate 2010 Note: CM2.1 has no ice-sheet dynamics!

8. Sea Level Trends in GFDL’s CM2.1 Coupled Model Yin, Griffies & Stouffer, J. Climate 2010 Note: CM2.1 has no ice-sheet dynamics!

9. 21 st Century Thermosteric and Halosteric Sea Level Rise Anomalies in GFDL’s CM2.1 Yin, Griffies & Stouffer, J. Climate 2010 Local Anomalies of SLR due to Ocean Temperature Changes Local Anomalies of SLR due to Ocean Salinity Changes

10. 1° Coupled Model June SSH Annual Mean SSH Inferred from Obs. 1/8° Global Ocean Model June SSH

13. Ocean Influence on Jakobshavn Glacier, Greenland Holland et al., 2008, Nature Geoscience. 50 km Satellite Image from NASA

14. Regional Sea Level Change from Geoid Changes Gravitational attraction by the Ocean, Ice-sheets, and Soil-moisture mass distributions help shape the Earth’s geoid (resting sea-level). Geoid changes will be an important part of regional sea level rise. <011.40.21.20.40.60.8 Sea-level rise per meter equivalent Greenland melt (m) 1 1 0 Kopp et al., 2009 (sub.)

15. Ice shelves - where water meets ice Ice shelves provide buttressing to ice sheets. Several shelves Antarctica have collapsed. The resulting acceleration of flow into the ocean is significant! Key question: what determines distribution/rate of ice shelf melt? We are actively working to put ice shelves in GFDL’s coupled models. Little, Gnanadesikan and Hallberg, 2008- melt in large ice shelves concentrated in southeast- regardless of bottom slope. Little, Gnanadesikan and Oppenheimer, 2009- melt rate (contours) strongly controlled by ice shelf bottom. Further work (w. Goldberg/Sergienko) indicating that the melting also shapes the bottom! Seafloor slopes up to north Seafloor slopes down to north Flat seafloor

16. What is needed to model ice-sheet dynamics contributions to global mean sea level rise? Bamber et al., Science 2000; C. Rapley, British Antarctic Survey 1.Ice sheet dynamics model, including the ability to simulate the rapidly flowing ice streams. 2.Ice shelf model, including calving of ice-bergs and collapse. 3.Model of the ocean circulation in the ice-shelf cavity. 4.Parameterizations or resolution of the eddy- and tidal- delivery of warm ocean water to the ice shelf cavity. Ocean-weather heat transport Ice Shelf Cold-fresh Warm-salty melts

17. Higher resolution coupled models are coming, but … Higher resolutions are much more expensive to run: 1/8° ≈ 512x 1° New algorithms are needed for the simulations to be acceptable for climate runs.

18. Spurious diapycnal mixing in coupled climate models?

19. Spurious diapycnal mixing increases with resolution?

20. Isopycnal coordinates avoid spurious diapycnal mixing.