Aviation and the Global Atmosphere

	Other reports in this collection

4.2. Model Studies of Subsonic Aircraft

In this section, we discuss the results of global 3-D CTMs used to assess the effects of subsonic aircraft on atmospheric concentrations of O3, NOx, and OH. The models differ in their formulations of vertical and horizontal resolution, transport, boundary conditions, and chemistry. Therefore, a wide range of results is to be expected. A short presentation of models and assumptions follows.

4.2.1. Models Used in Subsonic Aircraft Assessment

Table 4-1 lists the six CTMs used and the names of the associated investigators.1 Readers are referred to a Technical Report on Subsonic Aircraft Effects, which is presently available over the Internet for additional details.

4.2.1.1. Off-Line vs. On-Line Models

All of the models in Table 4-1, except the ECHAm³/CHEM model, are off-line models; that is, they are driven using meteorological fields derived either from GCMs or from analysis of observations. The temporal resolution of the various meteorological fields used to drive the models ranges from 40 minutes to a day. One exception is the IMAGES/ BISA model, which uses monthly-average meteorological fields and includes a parameterization to account for shorter-term variability in transport. In all of the off-line models, 1 year of wind fields is recycled in multiyear simulations to get the steady-state atmosphere. On-line calculations provide the potential capability of examining chemistry-climate interactions when model-calculated fields are used in radiation calculations. In ECHAm³/CHEM, the evolution of chemical fields is calculated on-line in a GCM, but the calculated chemical fields do not feed back into the dynamic calculations in this application. The model, therefore, operates in a similar way to the off-line models.

4.2.1.2. Model Resolution

Typically, these models have horizontal resolutions of 3-6�, with the exception of the UiO model, which has a horizontal resolution of 8�x10�. In the vertical dimension, the IMAGES/ BISA model has 25 levels; the Tm³/KNMI and ECHAm³/ CHEM models have 19 levels; and the HARVARD, UKMO, and UiO models have nine levels. Four of the models (ECHAm³/CHEM, HARVARD, Tm³/KNMI, UiO) have a top layer located at 10 mb; the IMAGES/BISA and UKMO models have top layers located at 50 mb and 100 mb, respectively. Because vertical model levels are defined on sigma coordinates and not on pressure coordinates, the number of model levels between fixed pressure surfaces can vary in time. Between the surface and 850 mb, the HARVARD, UiO, and UKMO models have about two vertical levels; the ECHAm³/CHEM and Tm³/KNMI models have about five vertical levels, and the IMAGES/BISA model has eight vertical levels. In the UT/LS region between 100 and 300 mb, the HARVARD model has one vertical level, the UKMO model has about two levels, and the other four models have about four vertical levels.

4.2.1.3. Coupling to the Stratosphere

With the exception of the ECHAm³/CHEM model, the models have little or no representation of explicit stratospheric chemistry. Instead, either the cross-tropopause fluxes of O3 and NOy are specified or the mixing ratios of these species are specified in the LS based on observations. In the Tm³/KNMI, UiO, and IMAGES/BISA models, however, the upper boundaries are higher in an attempt to minimize their influence on regions of maximum perturbations by aircraft. It should be noted that this condition may not be satisfied for the IMAGES/BISA model because the model top is at 50 mb.

4.2.1.4. Tropospheric NOx Sources

All of the models include anthropogenic and biogenic tropospheric NOx sources. For present-day conditions, the magnitudes of surface NOx sources in the various models are ~21 Tg nitrogen (N) yr^-1 from surface-based fossil-fuel combustion, 5-12 Tg N yr^-1 from biomass burning, and 4-6 Tg N yr^-1 from soils. The present-day magnitude of the lightning source is 5 Tg N yr^-1 in the IMAGES/BISA, Tm³/KNMI, UKMO, and UiO models; 4 Tg N yr^-1 in the ECHAm³/CHEM model (increased to 5 Tg N yr^-1 in the 2015 and 2050 simulations); and 3 Tg N yr^-1 in the HARVARD model. It should be noted, however, that the simulated impact of lightning on NOy species in the troposphere can differ from model to model even if the magnitude of the lightning source of NOx is the same, as a result of differences in factors such as duration and intensity of convective events, land/ocean differences in convection, height of NOx emissions, and so forth. Sensitivity tests were run to evaluate the effect of this lightning assumption on calculated aircraft perturbation.

4.2.1.5. Tropospheric Chemistry

Most of the models include a comprehensive description of the CH4-CO-NOx-HOx-O3 chemical system. With the exception of ECHAm³/CHEM and Tm³/KNMI, the models include representations of NMHC chemistry. However, the details of NMHC chemistry differ significantly from model to model. The ECHAm³/CHEM model includes a stratosphere with a chemistry scheme more suited to the stratosphere, however, it does not include some of the species that are important for tropospheric chemistry.

4.2.1.6. Tropospheric Transport

In addition to transport by resolved-scale winds, all models considered here include parameterizations of vertical transport by sub-grid-scale processes such as convection and turbulent mixing in the boundary layer. Again, the manner in which these processes are parameterized differs from model to model. In this context, it is worth noting that four of the models (or their close counterparts) used in this exercise (ECHAm³/CHEM, HARVARD, UKMO, and Tm³/KNMI) were also involved in a model intercomparison exercise sponsored by the World Climate Research Program (WCRP) in 1993 (Jacob et al., 1997). As part of this exercise, each model simulated a scenario in which a fictitious tracer with a 5.5-day e-folding lifetime was emitted in the Northern Hemisphere mid-latitude UT. The vertical gradient in the simulated fields was similar in several of the participating models. However, there were significant inter-model differences in the simulated rates of meridional tracer transport in the UT.

4.2.2. Definition of Scenarios

This section describes the scenarios for aircraft emissions evaluated for this assessment. The premises for current (circa 1992) and future (roughly 2015 and 2050) aircraft fleet emissions, along with descriptions of actual emissions databases, are given in Chapter 9. The assumptions used for the background atmosphere in model calculations of the effects of aircraft emissions on O3 are important and influence the results. In the following sections, we discuss the basis for background atmospheres used in model calculations and the aircraft scenarios evaluated.

4.2.2.1. Background Atmospheres

Boundary conditions for CH4 in the background atmosphere are 1714, 2052, and 2793 ppbv for the years 1992, 2015, and 2050, respectively. These amounts are based on the IPCC IS92a scenario (IPCC, 1992, 1995). Updated projections for future CH4 concentrations (WMO, 1999) are smaller than those assumed here. Recent observations by Dlugokencky et al. (1998) show that CH4 levels currently are leveling off. If this trend continues during the next century, with little or no increase in the CH4 concentration, the increase in background O3 will also be substantially less than that calculated in these studies.

For shorter lived gases-such as CO, NOx, and volatile organic compounds (VOCs)-the participating models use their standard boundary conditions for the 1992 cases. For 2015 and 2050, most model calculations assume that emissions are increased by the same factors at all locations relative to 1992 emissions, as shown in Table 4-2. Such constant increases were necessitated by difficulties in 3-D models to readily change emission inputs for assessment studies.

A special sensitivity study was also conducted for 2050 with the UiO model using a geographically varying emission increase (IPCC, 1995). A summary of these factors is presented in Table 4-3. Such regional differential factors are applied only to energy-related sources; biomass burning factors are applied as in the standard case (using Table 4-2).

4.2.2.2. Aircraft Emission Scenarios

The 3-D aircraft scenarios described in Chapter 9 form the basis for the assessment. The scenarios evaluated by the participating models are summarized in Table 4-4. Summaries of global emissions for these scenarios are given in Tables 9-4 and 9-5. Only a few scenarios are considered for subsonic assessment calculations because computational requirements for the 3-D models are high. The subsonic scenarios in Table 4-4 are generally analyzed relative to corresponding background atmospheres for 1992, 2015, or 2050.

In model calculations, aircraft effluents are put into the models as follows: Gridded fuel burn data (kg fuel/day) are first mapped into the model grid. The amount of material emitted into each grid box is given by the product of the fuel burn and the emission index. The emitted material is put into the grid box at each time step at the equivalent rate. In this approach, we ignore the effect of plume processing and assume that emitted material is instantaneously mixed into the grid box. For the subsonic assessment, NOx is the only aircraft emission considered. Because most models do not calculate the hydrological cycle in the troposphere, emitted water is not calculated. Sulfur, CO, and unburned hydrocarbons are also ignored.

The basic scenarios examine some of the important aspects in understanding the calculated environmental impact of aircraft. However, a number of uncertainties remain in the treatment of chemical and physical processes that may influence the effects from aircraft emissions. Therefore, a series of special sensitivity calculations was designed to investigate the most important of the recognized uncertainties. The subsonic aircraft sensitivity scenarios, as described later, examine uncertainties in the background atmosphere, the treatment of chemical and dynamical processes in the UT and LS, and different analyses of aircraft emissions.

It has not been possible (for practical reasons) for each of the modeling groups to run all of the scenarios set up for these 3-D model studies of aircraft perturbations. Each modeling group has completed a limited number of model simulations.