Emulation of Sub-Ice Shelf Melt Forcing Using High-Fidelity E3SM Simulations
Basal melting of Antarctica’s fringing ice shelves has been identified as a significant source of uncertainty in sea level rise projections. Part of this uncertainty derives from internal variability of the coupled climate system. The uncertainty this variability imparts on ice sheet evolution has not been previously explored or quantified because of the prohibitive computational expense of running the necessary large climate model ensembles. Using data from a single, long E3SM v1 Cryosphere simulation, we develop and demonstrate techniques to generate independent realizations of ice shelf basal melting that are statistically consistent with high-fidelity simulations. These realizations maintain the regional characteristics of individual Antarctic ice shelves and local characteristics within individual ice shelves (e.g., melt rates at grounding lines versus ice shelf fronts).
These new methods allow for efficient sampling and generation of a wide range of potential climate-forcing time series at a minute fraction of the cost of generating similar information via climate model ensembles. In the current example, the generated forcing time series can be used to explore and quantify the uncertainty in Antarctic ice sheet evolution that arises from internal climate variability.
The evolution of mass loss from Antarctica has been identified as a major source of uncertainty in future sea level projections. Ice shelves that fringe the continent are particularly susceptible to warming oceans, with some already thinning due to increased basal melting. Full-physics-based climate model simulations are too costly for exploring the impacts of climate variability on ice shelf melting and the ice sheet’s response. We devised a computationally efficient method for generating numerous realistic realizations of sub-ice shelf melting from a single high-fidelity climate model simulation.
Our method is applicable to other studies where the output from climate models is necessary to force a physical model of, e.g., oceans, ecosystems, the solid earth, or other components of the Earth system. These methods can significantly reduce the computational cost of generating large ensembles of climate forcing, which is a necessary step in robustly quantifying the uncertainty in Earth system model projections of future climate change.