Moderate melting of a few meters per year occurs adjacent EX-527 to the ice front, especially between 1.5°W and 0°E where the ice shelf overhangs the continental shelf break. Enhanced melting in this region was inferred from oceanographic observations (Walkden et al., 2009), and recently this feature, which is consistently seen in modeling studies (Nicholls et al., 2008 and Smedsrud et al., 2006), has also been confirmed by remote sensing based (Rignot et al.,

2013) and in situ measurements (Langley et al., in preparation) of basal melting at the FIS. While errors in our simulations are likely to be introduced by the artificially enhanced minimum water column thickness of 100 m at the grounding line, the simulated maximum melt rate of about 15 m year−1 in the southernmost part of Jutulstraumen is in good agreement with estimates from glaciological mass flux divergence estimates in this location (Humbert, 2012). Test runs with a more realistic, but numerically less stable water column thickness of 50 m showed only minor variations of the simulated melt rates under forcing conditions similar to the ANN-100 experiment. Smaller areas of net freezing are also observed, mostly in regions where the buoyant ISW ascends along steeper parts of the ice base and becomes supercooled as it reaches shallower depth. But the amount of freezing contributes

less than 5% to the total basal mass balance Angiogenesis inhibitor in the ANN-100 experiment, suggesting that no substantial accretion of marine ice occurs beneath the FIS. However, freezing processes are incomplete, with no frazil ice processes being included in the model. The ANN-100 experiment features a seasonality of basal melt rates that suggests a distinct contribution of melting at different depths beneath the FIS. In order to illustrate this, Fig. 7(b) and (c) show the vertical distribution of ice shelf area and the basal melting contribution in various experiments. While the details of the depth-dependent melting response to

different model forcings will be discussed in Section 5.2, the histogram CYTH4 of horizontal ice shelf area as a function of depth,2 shown by the dashed curve (left axis) in Fig. 7(b), reveals large areas of shallow ice at about 250 m depth and large areas of deep ice at about 350 m depth, with a natural separation at the local minimum of the curve at 300 m depth. As indicated by the thick 300 m contour in Fig. 7(a), this pronounced bi-modal distribution reflects the difference between the thicker body of the eastern FIS and the Jutulstraumen keel, and the large area of shallow ice in the central and western part of the FIS. Fig. 5 suggests that the melt rates within these two different portions of the FIS are controlled by the varying amounts of ASW and WDW that enter the cavity at different times of the year in the ANN-100 experiment. In addition to the synthetic mooring data in Fig. 5(a), Fig.