Abstract
Abstract
Basal melting is a major component of Antarctic ice mass loss. It is mainly driven by the circulation of water in ice shelf cavities. One approximation of this circulation is by models of buoyant plumes rising along the ice shelf base. For large-scale ice sheet models, a local melt parameterization of plume models that can be applied at an ice shelf-wide scale is desirable, and such a parameterization has recently been proposed and shown to perform well compared to a spatially resolved plume model when the ice shelf profile is fixed. In this study, we test the local parameterization for a dynamically evolving ice shelf, which allows for feedbacks between basal slope and melt rate. We find that these feedbacks enhance melt for steep basal slopes and reduce melt for small slopes. This decreases the parameterization performance, and under certain conditions the parameterization introduces an unphysical melt discontinuity. This can lead to large differences in the predicted melt rate and ice shelf length compared to the nonlocal plume model. More generally, our findings suggest that i) local parameterizations should be applied with caution in coupled ice–ocean frameworks, and ii) melt parameterizations should be evaluated in coupled settings to avoid unwanted behavior.