Correlation coefficients for the LGMGHG experiment, where only the greenhouse gases were changed to LGM levels but all other forcings were kept the same as the control run, are shown in Table 2. As already discussed in Sect. 3.1 the LGMGHG temperature changes are more strongly correlated with the LGM temperature changes than the 2xCO2 temperature changes. As is apparent from Fig. 3 there is almost as much scatter in the LGMGHG vs 2×CO2 (red) temperatures as there is for the LGM vs 2×CO2 temperatures (blue). This is a somewhat surprising result which implies that the uncertainty in the response to the ice sheet does not outweigh that due only to the nonlinearity in the response to increasing versus decreasing GHG levels. Looking at the dot-dashed line in Subplot A of Fig. 4 which shows the correlation between the magnitudes of the global temperature change for 2×CO2 and zonal temperature change for LGMGHG, the line more closely follows the LGM zonal variation except north of about 30 oN where the correlation is more like the 2×CO2 zonal variation. So, while the north of the northern hemisphere is largely influenced by the ice sheets at the LGM, it seems that uncertainty in the influence of the ice sheet does not have a clear influence on the rest of the globe and therefore it must be nonlinearity of the response to differing GHG levels across the range tested that produces a large part of the observed scatter in the relationship between LGM and 2×CO2 climates. The standard estimate for the radiative forcing due to changes in greenhouse gas levels at the LGM is approximately –2.8Wm−2, whereas that due to doubling CO2 is +3.7Wm−2 (Houghton et al., 2001), a ratio of –0.76. The radiative forcing has been calculated in two version of the MIROC model (at T42 resolution, with different atmospheric parameterisations) and found to be in close agreement with this ratio, even though the absolute values were marginally different in each case (Yokohata et al., 2005). Since we did not alter any parameters directly relating to radiative transfer in our experiments, we expect this relationship to hold across our ensemble and although we have not performed a precise calculation, there is no sign of a correlation between the initial radiative imbalance (when GHG levels are changed abruptly) and the equilibrium temperature change. If the models exhibited a constant sensitivity across this range, the magnitude of the LGMGHG global temperature changes would be 76% of that for 2×CO2. In Fig. 5 we show the histogram of the global temperature changes plotted as a ratio T2(CTRLLGMGHG)/ T2(2xCO2-CTRL). Also shown, with a red line, is the 0.76 value corresponding to equal sensitivity. The line on Fig. 3 also shows the expected results for the LGMGHG and CO2 experiments if the response to increased and decreased GHG concentrations were linear.
For the median of the ensemble, the difference between the value of climate sensitivity and T2(CTRL-LGMGHG)/0.76 is 0.62oC. Furthermore, close to 80% of the ensemble have a smaller magnitude of temperature change when greenhouse gases are decreased rather than increased. This result is rather similar to that obtained by Hansen et al. (2005), but other models may differ and this conclusion will remain rather tentative until others undertake similar investigations.
Answers to all your Biology Questions
