Figure 5 . a . The optimized ice rigidity\(\overset{\overline{}}{B}\) through inversion in ISSM. b . The modified \(\overset{\overline{}}{B}\) used for the non-rift scenario.b . The modified \(\overset{\overline{}}{B}\) used for the enhanced weakening scenario.
Non-rift scenario with different front geometry conditions.This scenario aims to examine the dynamic responses of ice shelf velocity and stress to front geometry change. We controlled for rift effects by replacing the low values of \(\overset{\overline{}}{B}\) near Gipps Ice Rise with the values of the adjacent intact areas (Figure 5b). We tested nine front geometry conditions (Figure 6), including four historical front locations (F1–F4), and five hypothetical front locations (H1–H5) under future retreat scenarios. H1–H5 were determined by assuming a radical retreat scenario, in which the new ice front reaches the compressive arch. The compressive arch is the boundary where the second principal stress changes from negative to positive, and has been used as a critical boundary to predict ice shelf instability (Doake et al., 1998; Kulessa et al., 2014). It has been suggested that ice shelf retreat will become irreversible once the ice front retreats behind the compressive arch. H1 (Figure 6e) is the front location where the retreat reaches the current compressive arch (shown in Figure 6d); H2 is the front location where the retreat reaches the compressive arch for the geometry condition H1, and similarly for H3. Once the ice front reaches H3 or exceeds H3, the compressive arch would no longer exist. H4 is the front condition if the H3 front continues retreating another 1 km, and H5 is the front condition if H4 keeps retreating by ~20 km (Figure 6g). H4 and H5 were used to test how sensitive the ice shelf would be to changes in the ice shelf front when the compressive arch is absent.
Rift-scenario 1 with original\(\overset{\overline{}}{\mathbf{B}}\). Through two rift scenarios, we aim to test the hypothesis based on the observations: the weakening due to the rift R2 was crucial for the propagation of R1, which resulted in the 2017 calving event. In this scenario, we used the original ice rigidity \(\overset{\overline{}}{B}\) (Figure 5a), which includes, to a certain extent, the R2-induced weakening.
Rift-scenario 2 with enhanced weakening. We considered a weakening-enhanced scenario by further decreasing the rigidity\(\overset{\overline{}}{B}\) to a minimum value of 0.1×108Pa s1/3 over the rifted area (Figure 5c).