Extended Data Fig. 1: Ensemble observational targets.
From: The Paris Climate Agreement and future sea-level rise from Antarctica
196 simulations (grey lines), each using a unique combination of hydrofracturing and ice-cliff calving parameters (Extended Data Table 1) compared with observations (blue dashed boxes). Solid blue lines show simulations without hydrofracturing and ice-cliff calving. Red lines show simulations with maximum parameter values in our main ensemble. Additional simulations (black lines) allow ice-cliff calving rates of up to 26 km yr−1, twice the maximum value used in our main ensembles. The vertical heights of the blue boxes represent the likely range of observations. Changes in ice mass above floatation are shown in equivalent GMSL. a, Simulated annual contributions to GMSL in the RCP8.5 ensemble compared with the 1992–2017 IMBIE4 observational average (0.15–0.46 mm yr−1; dashed blue box). b, LIG ensemble simulations from 130 to 125 kyr ago. The height of the dashed blue box shows the LIG target range (3.1–6.1 m), the width represents ~1,000-yr age uncertainty34. c, The same LIG simulations as in b, showing the rate of GMSL change contributed by Antarctica, smoothed over a 25-yr window. The peak in the early LIG is mainly caused by marine-based ice loss in West Antarctica. d, The same as b, except for warmer mid-Pliocene conditions. Maximum ice loss is compared with observational estimates of 11–21 m (refs. 35,36; blue dashed lines). Note the saturation of the simulated GMSL values near the top of the LIG and Pliocene ensemble range, and the failure of the model to produce realistic LIG or Pliocene sea levels without hydrofracturing and ice-cliff calving enabled (blue lines).
