Mass loss from the Antarctic Ice Sheet to the ocean has increased in recent decades, largely because the thinning of its floating ice shelves has allowed the outflow of grounded ice to accelerate 1,2. Enhanced basal melting of the ice shelves is thought to be the ultimate driver of change 2,3 , motivating a recent focus on the processes that control ocean heat transport onto and across the seabed of the Antarctic continental shelf towards the ice 4-6. However, the shoreward heat flux typically far exceeds that required to match observed melt rates 2,7,8 , suggesting other critical controls. Here we show that the depth-independent (barotropic) component of the flow towards an ice shelf is blocked by the dramatic step shape of the ice front, and that only the depth-varying (baroclinic) component, typically much smaller, can enter the sub-ice cavity. Our results arise from direct observations of the Getz Ice Shelf system and laboratory experiments on a rotating platform. A similar blocking of the barotropic component may occur in other areas with comparable ice-bathymetry configurations, which may explain why changes in the density structure of the water column have been found to be a better indicator of basal melt rate variability than the heat transported onto the continental shelf 9. Representing the step topography of the ice front accurately in models is thus important for simulating the ocean heat fluxes and induced melt rates. Main text: The fate of the Antarctic Ice Sheet is the greatest remaining uncertainty when predicting future sea level 10. Estimates of its contribution to global sea-level rise range from none to a catastrophic > 5 cm/year 10-12 (4 m by the year 2100). The ice sheet drains into the ocean where it terminates in floating ice shelves, overlying vast sub-ice cavities. These buttress the flow of the ice sheet, regulating the speed at which it flows into the ocean 13. Rapid thinning of ice shelves in coastal regions with warm ocean water on the continental shelf is accelerating the outflow from the ice sheet 1,2. The perceived reason-although rarely observed directly 14is that ocean currents deliver more warm water to the ice shelf cavities, causing increased basal melt. These currents originate in a reservoir of warm and salty water, known as Circumpolar Deep Water (CDW) 15 , residing at 300-1000 m depth in the Southern Ocean. Substantial amounts of dense CDW are carried onto the continental shelf by various mechanisms 4-7,16 , but only a fraction of this is needed to explain observed basal melt rates 17. The CDW flows southward in deep troughs that crosscut the continental shelf 4,18-21. The currents are steered by the bathymetry and move with shallower water to the left of the flow direction 22-24 so southward transport occurs along the eastern, and northward on western,

Abstract. Rapid retreat of Greenland's marine-terminating glaciers coincides with regional warming trends, which have broadly been used to explain these rapid changes. However, outlet glaciers within similar climate regimes experience widely contrasting retreat patterns, suggesting that the local fjord geometry could be an important additional factor. To assess the relative role of climate and fjord geometry, we use the retreat history of Jakobshavn Isbræ, West Greenland, since the Little Ice Age (LIA) maximum in 1850 as a baseline for the parameterization of a depth- and width-integrated ice flow model. The impact of fjord geometry is isolated by using a linearly increasing climate forcing since the LIA and testing a range of simplified geometries. We find that the total length of retreat is determined by external factors – such as hydrofracturing, submarine melt and buttressing by sea ice – whereas the retreat pattern is governed by the fjord geometry. Narrow and shallow areas provide pinning points and cause delayed but rapid retreat without additional climate warming, after decades of grounding line stability. We suggest that these geometric pinning points may be used to locate potential sites for moraine formation and to predict the long-term response of the glacier. As a consequence, to assess the impact of climate on the retreat history of a glacier, each system has to be analyzed with knowledge of its historic retreat and the local fjord geometry.

<p>Shoreward oceanic heat flux in deep channels on the continental shelf typically far exceeds that required to match observed ice shelf melt rates, suggesting other critical controls.  IN the present study we study the depth-independent (barotropic) and the density-driven (baroclinic) components of the flow of warm ocean water towards an ice shelf. Using observations from the Getz Ice Shelf system as well as geophysical laboratory experiments on a rotating platform, it is shown that the dramatic step shape of the ice front blocks the barotropic component, and that only the baroclinic component, typically much smaller, can enter the sub-ice cavity.  A similar blocking of the barotropic component may occur in other areas with comparable ice-bathymetry configurations, which may explain why changes in the density structure of the water column have been found to be a better indicator of basal melt rate variability than the heat transported onto the continental shelf. Representing the step topography of the ice front accurately in models is thus important for simulating the ocean heat fluxes and induced melt rates.</p>