Simulating the Last Glacial Cycle using a Glacial Index and Climate Matrix Method

For simulating ice sheet – climate interactions on multi-millennial time-scales, a set-up that uses a two-way coupled Earth System Model would be ideal. However, running these simulations over multi-millennium time-scales while including ice sheets, is not feasible. Alternatively, ice sheet models can be forced by interpolating climate time-slices, allowing for a transient forcing to an ice sheet model at limited computational costs.

Here, we compare two methods that interpolate between climate time-slices to create a transient forcing for ice sheet simulations. Firstly, we use a glacial index method, in which the climate is linearly interpolated between time-slices based only on prescribed atmospheric CO₂ concentrations. Secondly, we use a climate matrix method in which the interpolation is not only dependent on the prescribed CO₂ concentration, but also on internally generated thickness, volume and albedo. As a result, the climate matrix method captures ice-sheet atmosphere feedbacks.

Here we present ice sheet simulations of the Last Glacial Cycle using IMAU-ICE forced with Last Glacial Maximum (LGM) and Pre-Industrial time-slices. For the time-slices we use the output from nine Paleoclimate Modelling Intercomparison Project Phase III (PMIP3) GCMs. Our aim is to compare and to evaluate the differences in ice sheet evolution and LGM volume and extent resulting from the different PMIP3 models and the interpolation method used for transient simulations.

For most PMIP3 forcings, both the North-American and Eurasian ice sheets build up quicker in the climate matrix method compared to the glacial index method, which is in better agreement with paleo-observations. This is mostly a result from precipitation differences between the two interpolation methods: In the climate matrix method the interpolation of precipitation is dependent on internally generated ice thickness instead of only CO₂. Therefore, when ice thickness is smaller than LGM, the interpolation tends to shift more towards pre-industrial in the climate matrix method compared to the glacial index method. As precipitation is larger during pre-industrial compared to LGM in most Eurasian and North-American regions, this leads to a larger precipitation in the climate matrix method, increasing ice sheet volume. Similarly, the climate matrix method results into warmer temperatures in ice-free areas as the interpolation is dependent on both CO2, albedo and insolation. However, for most PMIP3 models, this ice sheet-temperature feedback does not cancel-out the increased precipitation in the climate matrix method.