A computationally efficient model for the Greenland ice sheet
We present a one-dimensional model of the Greenland Ice Sheet (GIS) for use in analysis of future sea level rise. Simulations using complex three-dimensional models suggest that the GIS may respond in a nonlinear manner to anthropogenic climate forcing and cause potentially nontrivial sea level rise...
| Main Authors: | , , , , |
|---|---|
| Format: | Text |
| Language: | English |
| Published: |
2018
|
| Subjects: | |
| Online Access: | https://doi.org/10.5194/tcd-6-2751-2012 https://tc.copernicus.org/preprints/tc-2012-86/ |
| Summary: | We present a one-dimensional model of the Greenland Ice Sheet (GIS) for use in analysis of future sea level rise. Simulations using complex three-dimensional models suggest that the GIS may respond in a nonlinear manner to anthropogenic climate forcing and cause potentially nontrivial sea level rise. These GIS projections are, however, deeply uncertain. Analyzing these uncertainties is complicated by the substantial computational demand of the current generation of complex three-dimensional GIS models. As a result, it is typically computationally infeasible to perform the large number of model evaluations required to carefully explore a multi-dimensional parameter space, to fuse models with observational constraints, or to assess risk-management strategies in Integrated Assessment Models (IAMs) of climate change. Here we introduce GLISTEN (GreenLand Ice Sheet ENhanced), a computationally efficient, mechanistically based, one-dimensional flow-line model of GIS mass balance capable of reproducing key instrumental and paleo-observations as well as emulating more complex models. GLISTEN is based on a simple model developed by Pattyn (2006). We have updated and extended this original model by improving its computational functionality and representation of physical processes such as precipitation, ablation, and basal sliding. The computational efficiency of GLISTEN enables a systematic and extensive analysis of the GIS behavior across a wide range of relevant parameters and can be used to represent a potential GIS threshold response in IAMs. We demonstrate the utility of GLISTEN by performing a pre-calibration and analysis. We find that the added representation of processes in GLISTEN, along with pre-calibration of the model, considerably improves the hindcast skill of paleo-observations. |
|---|