The seasonal pattern in Fig. 8(a) and (b) also shows that the ASW and the MWDW both reside for several months beneath the FIS, potentially affecting basal melting far inside

the cavity. The MWDW, entering the cavity at the main sill in Fig. 8(b), is advected along topographic (f/Hf/H) contours further into the cavity, appearing as a warmer bottom layer (green) at the southernmost end of the cross-section in Fig. 8(a), and eventually causes melting of deep ice of Jutulstraumen. The evolution of the ASW, entering in the upper part of the cavity in Fig. 8(a), is shown by the thickened and more stratified layer of cold ISW (magenta) at the southern end in Fig. 8(b). A water mass analysis (not shown) reveals that the buoyant upper Etoposide portion of this ISW layer is formed by surface water which entered the cavity during the previous

summer and has expended its available heat for melting. Thus, our simulations confirm the hypothesis of Hattermann et al. (2012) that ASW can travel far into the ice shelf cavity, after initially being subducted beneath the ice front. An overview of the horizontal current strength and direction is presented in the lower panels of Fig. 8. A dominant feature of the sub-ice shelf circulation is the presence of counter-rotating, topographically constrained flows in the upper and lower water column of the central basin. At depth, the model shows a clockwise flow steered by the bottom topography, while in the upper part of the water column a counter-clockwise flow along ice BAY 73-4506 mouse draft contours is observed. We find that the different circulation patterns in the upper and lower parts of the cavity are a direct result of the enhanced stratification due to the presence of ASW. This can be seen by comparing the results from the ANN-100 experiment (Fig. 8(c) and (e)) to the circulation in the initial simulation (Fig. 8(d) and

(f)), which uses the WIN-100 forcing where no ASW is included in the model. In contrast to the vertically sheared currents described ID-8 above, the constant winter scenario shows a narrow but fast-flowing, topographically steered barotropic jet, with much larger current speeds in the upper part of the water column than observed in the ANN-100 experiment. Also the seasonal variability in the ANN-100 experiment (not shown) reveals stronger and more barotropic sub-shelf currents near the ice base during late winter and spring when the upper ocean stratification is weak. The analysis of the ANN-100 experiment thus, reveals several effects of ASW on the cavity ventilation and associated basal melting. In particular, the pronounced seasonality of the MWDW inflow at depth, which occurs in the absence of any variability of the wind forcing, is an interesting result implying a direct link between upper ocean hydrographic conditions and the deep ocean heat fluxes. In fact, without ASW in the model, no MWDW enters the cavity, as can be seen from the last six months of the constant winter initial simulation in Fig. 5(a).