The most useful assessment of the impact of the aircraft fleet on climate would be a comprehensive prediction of changes to the climate system, including temperature, sea level, frequency of severe weather, and so forth. Such assessment is difficult to achieve given the current state of climate models and the small global forcing of climate attributable to the single sector of aviation chosen for this special report (see discussion in Sections 6.1 and 6.5). Following IPCC (1995, 1996), we choose a single measure of climate change: radiative forcing (RF), which is calculated directly from changes in greenhouse gases, aerosols, and clouds, and which allows ready comparison of the climate impact of different aviation scenarios.
The Earth's climate system is powered by the sun. Our planet intercepts 340 W m-2 of solar radiation averaged over the surface of the globe. About 100 W m-2 is reflected to space, and the remainder-about 240 W m-2-heats the planet. On a global average, the Earth maintains a radiative balance between this solar heating and the cooling from terrestrial infrared radiation that escapes to space. When a particular human activity alters greenhouse gases, particles, or land albedo, such activity results in radiative imbalance. Such an imbalance cannot be maintained for long, and the climate system-primarily the temperature and clouds of the lower atmosphere-adjusts to restore radiative balance. We calculate the global, annual average of radiative imbalance (W m-2) to the atmosphere-land-ocean system caused by anthropogenic perturbations and designate that change radiative forcing. Thus, by this IPCC definition, the RF of the pre-industrial atmosphere is taken to be zero. (Although the term "radiative forcing" has more general meaning in terms of climate, we restrict its use here to the IPCC definition.)
As an example, burning of fossil fuel adds the greenhouse gas CO2 to the atmosphere; this burning is responsible for the increase in atmospheric CO2 from about 280 ppmv in the pre-industrial atmosphere to about 360 ppmv in 1995. Added CO2 increases the infrared opaqueness of the atmosphere, thereby reducing terrestrial cooling with little impact on solar heating. Thus, the radiative imbalance created by adding a greenhouse gas is a positive RF. A positive RF leads to warming of the lower atmosphere in order to increase the terrestrial radiation and restore radiative balance. Radiative imbalances can also occur naturally, as in the case of the massive perturbation to stratospheric aerosols caused by Mt. Pinatubo (Hansen et al., 1996).
Because most of the troposphere is coupled to the surface through convection, climate models typically predict that the land surface, ocean mixed layer, and troposphere together respond to positive RF in general with a relatively uniform increase in temperature. Global mean surface temperature is a first-order measure of what we consider to be "climate," and its change is roughly proportional to RF. The increase in mean surface temperature per unit RF is termed climate sensitivity; it includes feedbacks within the climate system, such as changes in tropospheric water vapor and clouds in a warmer climate. The RF providing the best metric of climate change is the radiative imbalance of this land-ocean-troposphere climate system-that is, the RF integrated at the tropopause.
When radiative perturbation occurs above the tropopause, in the stratosphere (as for most HSCT impacts), this heating/cooling is not rapidly transported into the troposphere, and the imbalance leads mostly to changes in local temperatures that restore the radiative balance within the stratosphere. Such changes in stratospheric temperature, however, alter the tropospheric cooling; for example, warmer stratospheric temperatures lead to a warmer troposphere and climate system. This adjustment of stratospheric temperatures can be an important factor in calculating RF and is denoted "stratosphere-adjusted."
All RF values used in this report refer to "stratosphere-adjusted, tropopause RF" (Shine et al., 1995). For primarily tropospheric perturbations (e.g., CO2 from all aviation, O3 from subsonic aircraft), this quantity can be calculated with reasonable agreement (better than 25%) across models used in this report (see Section 6.3). For specifically stratospheric perturbations (e.g., H2O and O3 perturbations from HSCT aircraft), the definition of the tropopause and the calculation of stratospheric adjustment introduce significant sources of uncertainty in calculated RF.
The concept of radiative forcing (IPCC, 1990, 1992, 1995) is based on climate model calculations that show that there is an approximately linear relationship between global-mean RF at the tropopause and the change in equilibrium global mean surface (air) temperature (�Ts ). In mapping RF to climate change, the complexities of regional and even hemispheric climate change have been compressed into a single quantity-global mean surface temperature. It is clear from climate studies that the climate does not change uniformly: Some regions warm or cool more than others. Furthermore, mean temperature does not provide information about aspects of climate change such as floods, droughts, and severe storms that cause the most damage. In the case of aviation, the radiative imbalance driven by perturbations to contrails, O3, and stratospheric H2O occurs predominantly in northern mid-latitudes and is not globally homogeneously distributed (see Chapters 2, 3, and 4), unlike perturbations driven by increases in CO2 or decreases in CH4. Does this large north-south gradient in the radiative imbalance lead to climate change of a different nature than for well-mixed gases? IPCC (Kattenberg et al., 1996) considered the issue of whether negative RF from fossil-fuel sulfate aerosols (concentrated in industrial regions) would partly cancel positive RF from increases in CO2 (global). Studies generally confirmed that global mean surface warming from both perturbations was additive; that is, it could be estimated from the summed RF. Local RF from sulfate in northern industrial regions was felt globally. Nevertheless, the regional patterns in both cases were significantly different, and obvious cooling (in a globally warming climate) occurred in specific regions of the Northern Hemisphere. Such differences in climate change patterns are critical to the detection of anthropogenic climate change, as reported in Santer et al. (1996). As a further complication of this assessment, aviation's perturbation occurs primarily in the upper troposphere and lower stratosphere, and thus may alter the vertical profile of any future tropospheric warming. Therefore, the patterns of climate change from individual aviation perturbations (e.g., CO2, O3, contrails) would likely differ, but we take their summed RF as a first-order measure of the global mean climate change (see also the discussion in Section 6.5).
The equilibrium change in mean surface air temperature (�Ts(equil)) in response to any particular RF is reached only after more than a century because of the thermal inertia of the climate system (primarily the oceans) and is calculated with long-term integrations of coupled general circulation models (CGCM). A climate sensitivity parameter (l) relates RF to temperature change: �Ts(equil) = l RF. Provided that all types of RF produce the same impact on the climate system (in this case measured by mean temperature), the climate sensitivity parameter derived from a doubled-CO2 calculation can be used to translate other RFs, say from ozone or contrails, into a change in global mean surface air temperature.
For doubled CO2 relative to pre-industrial conditions (+4 W m-2), surface temperature warming ranges from 1.5 to 4.5 K, depending on the modeling of feedback processes included in the CGCM. The recommended value in IPCC (1996) of 2.5 K gives a climate sensitivity of l = 0.6 K/(W m-2). With limited feedbacks (e.g., fixing clouds and surface ocean temperatures), the sensitivity parameter is smaller, and most models produce similar responses. In contrast, when all feedbacks are included, model results are quite different, as a result (for instance) of alternative formulation of clouds. The obvious limitation of this approach is that we get no information about regional climate change. The sensitivity parameter for aircraft-like ozone perturbations is discussed in Section 6.5.
In spite of all these caveats, the radiative forcing of an aviation-induced atmospheric perturbation is still a useful index that allows, to first approximation, the different atmospheric perturbations (e.g., aerosols, cloud changes, ozone, stratospheric water, methane) to be summed and compared in terms of global climate impact.
Other reports in this collection