Aviation and the Global Atmosphere

	Other reports in this collection

6.3.6. Uncertainties

Assignment of formal uncertainty-or the likely (2/3 probability) interval about the best value-to radiative forcing caused by aircraft perturbations is difficult. For well-mixed gases (e.g., CO2, CH4) or for well-defined tropospheric perturbations (subsonic O3), there is small uncertainty in calculated RF. In these cases, the overall uncertainty interval lies with calculating the perturbation itself: 25% for CO2, a factor of 2 for O3, and a factor of 3 for CH4. For perturbations to stratospheric ozone and water, there is much greater uncertainty in calculating RF, especially because in these cases radiative forcing at the tropopause can be substantially different after stratospheric temperatures adjust. In addition, HSCT-induced ozone and water vapor perturbations-with large variations in the lower stratosphere-present a much more difficult calculation of RF where the placement of the modeled tropopause can lead to additional uncertainty.

As an example of the uncertainty in calculating RF values from an adopted ozone perturbation, the NASA-1992 tropospheric ozone perturbation was calculated by several groups, as shown in Table 6-4. The instantaneous RF at the tropopause is consistent across the models, and the stratospheric adjustment (calculated by two groups) is consistently 0.001 to 0.002 W m-2 less. For the HSCT(500) water vapor perturbation, the two groups have significantly greater disagreement, and the correction following stratospheric adjustment is a large fraction of tropopause instantaneous RF. This water vapor perturbation is the result of averaging six model results, and an additional RF is calculated using the water vapor perturbation calculated with a 3-D model that lies at the lower end of this ensemble (Grossman*; see Table 6-4). This table highlights the robustness of calculated RF for tropospheric perturbations and the much greater uncertainty in deriving climate forcing for stratospheric changes.

Uncertainty ranges about the best values are given in Table 6-1 for the NASA-1992* and FESGa (tech 1) 2050 subsonic scenarios and for some components of supersonic scenarios HSCT(500) and HSCT(1000). These intervals are intended to represent the same probability range (67% likelihood), but there is no uniform statistical model (e.g., Gaussian) for all of them, nor are the individual RF contributions fully independent; hence, these ranges cannot be combined into a confidence interval on total RF.