by Prof. Sandy Harrison
This week we have finally submitted the manuscript which describes the ice sheet boundary conditions used in the CMIP5/PMIP3 experiments for the Last Glacial Maximum (LGM). When the protocol for this experiment was set up – nearly five years ago – there were three candidate ice sheets that could have been used: (ICE-6G v2.0: Argus and Peltier, 2010: GLAC-1a: Tarasov et al., 2012; ANU: Lambeck et al., 2010). The problem is that although the margins of the ice sheets are relatively well known, from moraines and other glacio-geomorphic evidence, there are no direct observations on the form and height of the individual ice sheets, so this has to be reconstructed through inversion of a glacio-isostatic model constrained by observations such a relative sea level, post-glacial rebound rates, and mantle properties. And when you have three models, you introduce model-based uncertainties, and inversions are highly sensitive to the type and amount of data used as a constraint. If there were time (and worlds enough), the PMIP modellers would have run three separate LGM experiments with the three different ice sheets but it was decided to create a composite ice sheet through combining the information from the three ice sheets. Designing and creating that composite was a time-consuming job, describing why and how it was done (as the delayed submission shows) even more time-consuming – and of course, no-one is entirely happy with the composite “Frankenstein” ice sheet. But as we demonstrate in the submitted manuscript, the good news is that the differences between the impact of the composite and the individual candidate ice sheets on climate outside the regions covered by ice and global climate is trivial and much less than the differences produced by different climate models using the same ice sheet. And the benefit of only asking people to run one experiment was that many more modeling groups ran an LGM simulation than would have done otherwise. So it seems that compromise brings perils yes, but also benefits. If you want to find out more about our Frankenstein ice sheet, please check out the paper online (Abe-Ouchi et al.) at Geoscientific Model Development.
This week we have finally submitted the manuscript which describes the ice sheet boundary conditions used in the CMIP5/PMIP3 experiments for the Last Glacial Maximum (LGM). When the protocol for this experiment was set up – nearly five years ago – there were three candidate ice sheets that could have been used: (ICE-6G v2.0: Argus and Peltier, 2010: GLAC-1a: Tarasov et al., 2012; ANU: Lambeck et al., 2010). The problem is that although the margins of the ice sheets are relatively well known, from moraines and other glacio-geomorphic evidence, there are no direct observations on the form and height of the individual ice sheets, so this has to be reconstructed through inversion of a glacio-isostatic model constrained by observations such a relative sea level, post-glacial rebound rates, and mantle properties. And when you have three models, you introduce model-based uncertainties, and inversions are highly sensitive to the type and amount of data used as a constraint. If there were time (and worlds enough), the PMIP modellers would have run three separate LGM experiments with the three different ice sheets but it was decided to create a composite ice sheet through combining the information from the three ice sheets. Designing and creating that composite was a time-consuming job, describing why and how it was done (as the delayed submission shows) even more time-consuming – and of course, no-one is entirely happy with the composite “Frankenstein” ice sheet. But as we demonstrate in the submitted manuscript, the good news is that the differences between the impact of the composite and the individual candidate ice sheets on climate outside the regions covered by ice and global climate is trivial and much less than the differences produced by different climate models using the same ice sheet. And the benefit of only asking people to run one experiment was that many more modeling groups ran an LGM simulation than would have done otherwise. So it seems that compromise brings perils yes, but also benefits. If you want to find out more about our Frankenstein ice sheet, please check out the paper online (Abe-Ouchi et al.) at Geoscientific Model Development.