3.1 Period 1: Retreat (2002-2004)
From 2002-2004, terminus retreat of over 10 km occurred, with moderate retreat rates from November 2002 to March 2004, and more rapid retreat from March to November 2004 (Figure 2). During this two year window, the glacier thinned by 60.9 m at the location of the November 2004 ice front.
Satellite-derived, pre-LBIS collapse grounding line positions do not agree with the sea-floor sedimentary record or the updated bathymetry measured post-collapse (Rebesco et al., 2014). Satellite data placed the pre-collapse grounding line near the opening of the fjord (Rack & Rott, 2004), implying that the full extent of retreat observed from 2002-2004 occurred by failure of a grounded ice cliff. Radiocarbon-dated sediment cores show that ice was not grounded there in 2002, and suggest the pre-collapse grounding line was located further inland of the fjord opening (Rebesco et al., 2014). In addition, measured surface elevations from 2002 at the purported grounding line fell far below the floatation threshold (Fig. 1B). This means that a substantial portion of the initial terminus retreat occurred in ice that was at or below the floatation thickness.
Using our floatation criterion, we calculate an updated grounding line position for November 2002. Surface elevation data show a break in slope at ~9.5 km along our reference line, more consistent with calculated grounding lines assuming 0 m of FAC. We calculate retreat rates and cliff heights assuming this value.
While the majority of the Larsen B collapsed during February and March of 2002, a floating ice shelf remained at Crane. Slow terminus retreat began, but had not reached the 2002 grounding line by March of 2004, after 5.30 km of retreat (Figure 2). Retreat continued, and during the following 8 month period, the glacier experienced 4.75 km of additional terminus retreat, beginning 1.47 ± 0.03 km down-flow of the updated 2002 grounding line and halting at the 2004 terminus. We quantify the effective rate of ice loss during the periods preceding and following March 26, 2004 using a calving rate, equal to the sum of the terminus retreat rate and the ice flow speed at the terminus (~1.5 km per year, the nearest available satellite measurement in space and time, and consistent with estimates from Scambos et al., 2004). This results in minimum calving rate estimates of 5.48 km/a for the period ending on March 26, 2004, and 8.50 km/a in the period after, a 55% increase in calving rate over the 2 years following LBIS collapse.
If grounded ice cliff failure occurred at Crane, it must have happened during this final period of retreat. While there are no direct observations of a grounded ice cliff during that period, the first cliff that could have formed (at the 2002 grounding line position) would have existed with a height of 111 m (as measured above the water line; Figure 1B).