Is a 3-Dimensional Stress Balance Ice-Stream Model Really Better Than a 2-Dimensional "Reduced Order" Ice-Stream Model?

With growing observational awareness of numerous ice-stream processes occurring on short time and spatial scales, e.g., sub-ice-stream lake volume changes and grounding-line sediment wedge build-up, the question of how well models based on "reduced-order" dynamics can simulate ice-stream behavior becomes paramount. Reduced-order models of ice-streams are typically 2-dimensional, and capture only the largest-magnitude terms in the stress tensor (with other terms being constrained by various assumptions). In predicting the overall magnitude and large-scale pattern of ice-stream flow, the reduced-order models appear to be adequate. Efforts underway in the Glaciological Community to create 3-dimensional models of the "full" ice-stream stress balance, which relax the assumptions associated with reduced-order models, suggest that a cost/benefit analysis should be done to determine how likely these efforts will be fruitful. To assess the overall benefits of full 3-dimensional models in relation to the simpler 2-dimensional counterparts, we present model solutions of the full Stokes equations for ice-stream flow over a variety of basal perturbations (e.g., a sticky spot, a subglacial lake, a grounding line). We also present the solutions derived from reduced 2-dimensional models, and compare the two solutions to estimate effects of simplifications and neglected terms, as well as to advise on what circumstances 3-dimensional models are preferable to 2-dimensional models.