ECCO2 2008 figures Figure 1: Adjoint sensitivity of cost function (monthly model/data quadratic misfit) to the initial temperature field at the 15-m depth. Units are (°C)–1. The sensitivity is shown after a 332-day adjoint integration. The domain of integration is a global cubed sphere with 18-km horizontal grid spacing, as described in the text. Red patches are regions where the model-data misfit increases and blue patches are regions where the model/data misfit decreases with increasing initial temperature. The adjoint sensitivities are a critical step towards fully constraining this global ocean and sea-ice model with NASA satellite data. Figure 2: Atlantic Water rim currents in the Arctic Ocean from a constrained ECCO2 solution. Background color shows the bathymetry in km. Small arrows represent velocity magnitudes (|u|>0.03m/s in black, 0.01<|u|<0.03m/s in dark gray, and 0.003<|u|<0.01m/s in light gray). Large white solid arrows show total Atlantic Water transports as per scale provided in the legend. Large dashed white arrows are inferred transports based on current strengths. The thick red lines denote the Fram Strait (FA) and the St Ana Trough (SA). A fully constrained ECCO2 solution will permit more accurate estimates of these currents, of their time evolution, and of their impact on Arctic climate than is possible with data or with models alone. Figure 3: Self-assembling marine ecosystem. ECCO2 results are being used to supply boundary conditions for regional studies and to drive biogeochemical, geodetic, acoustic, and electromagnetic models. For example, O. Jahn used ECCO2 results to drive a self-assembling marine ecosystem, a shown in the figure above. Figure 4: Vertical temperature and salinity structures (upper panel) and T/S diagram (lower panel) of the Canadian Basin in August 2003. In the upper panel, the actual CTD observations are shown in light gray, with the data mean shown in dashed heavy black. Additional annotations are a) mixed layer, b) summer Pacific Water source, and c) winter Pacific Water source water. Dashed contours in T/S diagram are density anomalies. Blue lines (A0) are from the global ECCO2 solution while red lines (A1) are from a regional optimization that includes a sub-grid-scale parameterization of salt plumes (Nguyen et al., 2008). Figure 5: Simulated melt rate of all Antarctic ice shelves in November 2002. Units are m of ice per year. Positive values indicate refreezing and negative values indicate melting of the ice shelves. The simulation was carried out by M. Schodlok on a regional, Southern Ocean domain carved out from the global cubed-sphere ECCO2 domain of Figure 1. The mean circumpolar freshwater input into the Southern Ocean from ice shelves between 1992 and 2006 amounts to 59.8±7.4 mSv. This regional study is a step towards incorporating realistic ice-shelf ocean interactions in the global ECCO2 solutions. Figure 6: Snapshot of near-surface (15-m depth) current speed on March 3, 1993 from a quasi-global simulation with 1/16 th-degree horizontal grid spacing (horizontal grid spacing is approximately 6.9 km at the Equator and 1.19 km at +/-80°). High-resolution simulations are being used by the ECCO2 project to estimate model errors and to inform sub-grid-scale parameterizations (Hill et al., 2007). This particular simulation is also being used for testing a new supercomputer (Pleiades) at NASA Ames and in the design of the Surface Water and Ocean Topography (SWOT), a Decadal-Survey- recommended high-resolution altimeter mission. sion.