Online climate change projections report 3.2.13 SRES B1 & A1FI
The ensemble simulations of Sections 3.2.4 and 3.2.5 are all driven by future emissions and/or concentrations of anthropogenic forcing agents consistent with the A1B emissions scenario. In order to provide probabilistic projections for the B1 and A1FI scenarios, the 17 member PPE_A1B ensemble was re-run using appropriate time-dependent concentrations of greenhouse gases, and emissions of sulphate aerosol precursors. These ensembles were used to re-calibrate key timescaling statistics (specifically the correction and error terms) for the B1 and A1FI scenarios by comparing the HadCM3 simulations against timescaled estimates derived from corresponding HadSM3 simulations in conjunction with our simple climate model, as described in Section 3.2.4.
Probabilistic projections were then obtained by following the procedure of Section 3.2.12, specifying time series of forcing agents for B1 or A1FI in the simple climate model in step (iii). Apart from the timescaling aspects referred to above, all sources of uncertainty were all assumed to be the same as those specified for the A1B scenario. Some of these sources would clearly be independent of future emissions, such as emulation errors derived from our HadSM3 simulations, or the discrepancy attached to simulations of historical observables. The discrepancy for future projection variables is assumed independent of future emissions as a basic constraint of our experimental design. Further uncertainties could be specified separately for different emissions scenarios in principle, but were not in practice. These include global mean sulphate aerosol forcing, ocean heat uptake efficiency and feedback strengths, and regional downscaling relationships, for which resources to run additional ensemble simulations for B1 and A1FI were not available.
These assumptions are generally likely to be reasonable if global feedback strengths, and regional patterns of change per unit global warming (e.g. Mitchell, 2003), can be assumed independent of the chosen emissions scenario. Results from the latest IPCC assessment suggest that this is a reasonable assumption to leading order (e.g. Figure 10.9 of Meehl et al. 2007), however our assumptions render the results for SRES B1 and A1FI somewhat less robust than those for A1B, particularly for projections in the latter decades of the 21st century, when the applied forcing and simulated response for different SRES scenarios diverges significantly (Figure 2.14).