The PAGE (Policy Analysis of Greenhouse Effect) integrated assessment model (IAM) was used in the AVOID programme to estimate the cost of avoided impacts. Estimates of costs of climate impacts depend critically on the ability of the IAM to produce a credible temperature rise for a given emissions trajectory.
This report focuses on a comparison of the temperature response as calculated by the PAGE IAM and by MAGICC, the simple climate model used to estimate the temperature response of the AVOID programme emissions trajectories. The analysis carried out in this report concentrates on the climate model of a recent version of the IAM, PAGE09 V1.5, and examines each part in the calculation of temperature rise beginning from emissions. The key findings of the report show that although there are some similarities in the equations used between PAGE09 V1.5 and MAGICC, the more simple representation of the carbon cycle model and the temperature rise calculation within PAGE09 V1.5 than within MAGICC appeared to increase the disagreement. For example, an aggressive mitigation trajectory results in a more than 1° C difference in temperature rise between PAGE09 V1.5 and MAGICC estimates, with the former simulating the higher temperature rise.
The initial comparison highlighted areas within PAGE09 V1.5 to be amended and a new version PAGE09 V.1.7 has been developed that addresses the representation of the climate carbon cycle feedback and climate sensitivity within PAGE09 V1.5, and produces global mean temperature results which are consistent with those from more detailed models such as MAGICC.
Further work includes recommendations to improve the global temperature calculation within PAGE09 to capture the complex response to a mitigation scenario and a revision of the calculation of the regional temperature rise.
The results of this report are not expected to change the conclusions of earlier work carried out by PAGE2002 in the Stern report and AVOID impact work because the climate sensitivity distribution used in the Stern analysis had a lower mean value than the distribution within PAGE09 V1.5 which translates to lower temperature projections made by PAGE2002 than by PAGE09 V1.5. The conclusions of the Committee on Climate Change 2008 main report is unaffected by this work. However, the use of PAGE2002 in the Technical Appendix to Chapter 2 of the report may have underestimated the benefits of mitigation due to the higher temperature estimate from PAGE2002 than MAGICC for the low emissions scenario (partly due to the relatively strong carbon cycle in PAGE2002 see page 12 of Sura and Golborne, 2008).