Interim Review of the Subsonic Assessment Project: Management, Science, and Goals (1997)

Much of the SASS-sponsored research to date has focused on aviation emissions' chemical effect on upper-tropospheric ozone (O₃), and on how much such emissions add to the number of atmospheric aerosol particles. One of the most important potential consequences of aircraft emissions' effects on the troposphere is the possible alteration of climate through perturbations of ozone and aerosol concentrations. Ozone is highly reactive chemically, and is also a strong specific greenhouse gas that absorbs atmospheric radiation in the troposhere. Aerosols not only absorb radiation, but also can change the direction of light propagation (usually in anisotropic ways), commonly referred to as scattering. They have indirect effects as well, since they can serve as nuclei promoting cloud formation and as surfaces for heterogeneous reactions. Both ozone and aerosols are discussed in greater detail below.

In 1990 commercial aircraft consumed roughly 170 million tons of fuel per year (about 3 percent of the total fossil fuel burned), and that consumption is expected to increase at a rate of 2 to 3 percent per year (WMO, 1995). Moreover, about 60 percent of the aircraft exhaust is emitted into the upper troposphere, a region that otherwise receives only weak, attenuated input from anthropogenic or natural emissions at the Earth's surface.

Aircraft emissions already contribute significantly to the NO_x. budget of the upper troposphere, up to 50 percent in the most heavily traveled corridors—a major perturbation. Because NO_x catalyzes the formation of O₃ by the slow

photochemical oxidation of CH₄ , CO, and possibly other volatile organic compounds (VOCs), the possibility of an increase of O₃ in the upper troposphere has become an immediate concern. Various groups have presented model calculations of the expected O₃ increase, but at present the magnitudes of the numbers resulting from such calculations should be considered highly uncertain.

The concentration of ozone in the upper troposphere is the result of transport of ozone from the stratosphere plus net production in the troposphere (Roelofs and Lelieveld, 1995). The chemical ozone-formation rate depends on the local NO_x concentration in a highly nonlinear fashion. The contribution of NO_x from aircraft could thus lead to an increase or decrease in the local rate of O₃ formation, depending on the NO_x concentration already present. When averaged zonally, O₃ production prevails, and because of ozone's long lifetime in the upper troposphere, current models predict an increase in O₃ as a result of aircraft emissions everywhere in the upper troposphere (Ehhalt and Rohrer, 1995; Brasseur et al., 1996).

The unperturbed NO_x concentration field, however, is poorly known and poorly understood: Available measurements are insufficient to characterize well the background distribution of NO_x , and other source processes that introduce NO_x into the upper troposphere (namely, lightning, convective transport from the boundary layer, and input from the stratosphere) are insufficiently quantified to permit reliance on model calculations of background NO_x . The resulting uncertainties propagate to the averaged 0₃ production rate.

Local formation of 0₃ also depends on temperature and on O₃ , H₂ O, CO, and VOC concentrations. The latter dependence arises because the local formation of ozone in the upper troposphere and lower stratosphere is also catalyzed by oxides of hydrogen, or HO_x (OH and HO₂ ). HO_x is produced primarily from ozone, ultraviolet radiation (which varies with the stratospheric total-ozone column), and water, and is maintained by reactions with CO and VOCs. Removing current uncertainties in predicting impacts of aircraft exhaust on upper-tropospheric O₃ will require substantial research that quantifies the processes governing upper-tropospheric catalysts (NO_x and HO_x ) and reactive trace gases (globally, mainly CO, and regionally, VOCs as well).

Furthermore, in view of the longitudinal variations of background and aircraft-emitted NO_x , and of the non-linear response of local ozone production to NO_x , PAEAN believes that realistic assessments of O₃ changes can be made only with three-dimensional chemical-transport models of both the troposphere and stratosphere that effectively represent the salient features of horizontal and vertical transport as well as the pertinent chemical reactions. Defining activities to test and evaluate such a model must be a primary, although difficult, task of SASS.

PAEAN's preliminary recommendations in this area are to:

• Study the sensitivity of the NO_x , HO_x , H₂ O, CO, VOC, and O₃ budgets in the upper troposphere and lowermost stratosphere to the transport and chemical processes thought to be most important to them.
• Provide a quantitative analysis of the current uncertainties in the NO_x , HO_x , H₂ O, CO, VOC, and O₃ budgets in the upper troposphere and lowermost stratosphere resulting from these processes.
• Set targets for the uncertainty levels that SASS research should be able to achieve, and use them as guides for prioritizing research in future field studies and AEAP's Global Modeling Initiative (GMI).

"Aerosols", properly called "aerosol particles", are liquid or solid particles suspended in air. Ubiquitous throughout the atmosphere, aerosols arise from a variety of natural and anthropogenic processes. Particles larger than about 1 mm diameter are generally dominated by mechanically derived sea salt, dusts, and fly ash, whereas smaller particles (e.g., sulfates and soot) typically arise and grow via gas-to-particle conversion in gas plumes (from, e.g., volcanoes and combustion), in cloud-free air, and through processes occurring within clouds and cloud droplets. The size distributions, compositions, and concentrations of aerosol particles reflect a dynamic balance among source, transport, evolution, and removal mechanisms that vary in both space and time. (For instance, some particles and gases are injected from the atmospheric boundary layer into the upper troposphere and lower stratosphere by deep convection. Because of concurrent removal through precipitation, however, this injection process is not so efficient for particles as for gases with only slight solubility in water, such as NO and NO₂ .) These and other processes result in a "background aerosol'' present in the free troposphere, and it is the nature and significance of perturbations of this background aerosol by aircraft emissions that must be evaluated in the SASS project.

The background aerosol can be highly variable, with large excursions in effective particle size and with mass concentrations ranging over three orders of magnitude (Clarke, 1993). The surfaces of these particles may act as sites for preferred chemical reactions ("heterogeneous" chemical reactions), which can further influence the evolution of the size distribution and its chemistry. In turn, these physical and chemical properties directly affect the interactions of the particles with light and the abilities of the particles to nucleate cloud droplets and ice crystals. Therefore, any "signal" caused by aircraft exhaust must be identified and assessed within the context of this variable background.

Determining the possible climate perturbation becomes very complex when both the links with ozone concentration and the role of aerosols are included.

Assessments of aircraft perturbations of aerosols must evaluate influences on atmospheric chemistry, clear-air radiative transfer, cloud-droplet and ice-crystal nucleation, and the associated optical properties of clouds. Aerosols alter the forward-and backscattering of radiation, as well as lead to absorption in some situations. Particles that scatter light effectively at relative humidities below saturation, yielding direct radiative effects, are usually about 0.1 to 10 μm in diameter; normally they dominate the aerosol mass concentration in the upper troposphere. One subset of the total aerosol population, which can grow into cloud droplets through the condensation of water vapor at typical cloud super-saturations, is called cloud condensation nuclei (CCN). Another subset of particles (not necessarily disjoint), which are most effective as sites for water-vapor deposition leading to ice-crystal formation and growth in cold clouds, is identified as ice nuclei (IN). The roles of IN versus CCN for the formation of ice clouds appear to be temperature dependent. IN are important above-38°C; below-40°C homogeneous nucleation seems to dominate the ice-forming process, depending on the size of the CCN (Sassen and Dodd, 1989; Heymsfield and Sabin, 1989). An increase in these CCN and IN could have important indirect climate consequences (Twomey, 1977; Coakley et al., 1983; NRC, 1996). (Early calculations with aerosols in a radiative-convective model (e.g., Reck, 1975) showed that effects on Earth's surface temperature varied with altitude of the aerosol, but were also dependent on surface albedo and seasonal effects. More recent estimates (Charlson et al., 1990) show that a 1 percent change in average daily cloudiness could lead to a change in surface radiation forcing of about 1 watt per square meter.)

The physical and chemical properties of the particles dictate their effectiveness as CCN and IN. The radiative effects of clouds depend on cloud phase (liquid water, ice, or mixed phase), liquid (and/or ice) path, cloud morphology (three-dimensional effects), and optical properties. The latter in turn depend on microphysical properties, including particle size and shape (ice crystals). The size increase of CCN and IN during growth under cloud conditions dramatically enhances the light scattered and absorbed by these cloud nuclei. Hence, the radiative properties of clouds, for both solar and terrestrial radiation, are closely linked to the concentration and composition of the CCN on which water condenses and the IN on which ice forms (Twomey, 1980; Reck and Hummel, 1981; Twomey et al., 1984; Sassen, 1992; Liou, 1992). It has been suggested (Twomey, 1991; Stamnes et al., 1995) that a consequence of increasing aerosol burden might be a tendency for cloud drops to become more numerous and smaller. This would suppress drizzle (Albrecht, 1989), and lead to more persistent stratus coverage with higher liquid-water content (Feingold et al., 1994, 1996; P. Olsson, personal communication). Cloud-radiation interactions and feedbacks constitute the main focus of the Atmospheric Radiation Measurement (ARM) program, supported by the Department of Energy (Stokes and Schwartz, 1994). The indirect radiative effect (i.e., that associated with contrails and with influences of

aircraft aerosols on natural clouds) is the most uncertain, and potentially the most significant, impact of aircraft-generated aerosols. Both this indirect effect and the direct effect of radiative forcing, as well as their links to potential climate change, have been recognized in a recent National Research Council report (NRC, 1996) that outlines a plan for a research program aimed at this topic.

The indirect effect of aerosol can be illustrated by its potential impact on low-level marine stratocumulus clouds. This type of cloud covers large areas of the global ocean, and contributes a significant measure of shortwave cloud forcing (Atkinson and Zhang, 1996). (This forcing is only slightly offset by long-wave forcing, because the albedo difference between the cloud and the ocean is high, whereas the temperature difference is low.) Such clouds could be markedly influenced by an increase in cloud-seeding nuclei from aircraft-engine exhaust. If subsidence causes particles produced aloft to constitute a substantial component of boundary-layer aerosol, it will be important to understand the source and evolution of those particles, including their eventual incorporation into marine stratus.

There are two basic types of aircraft-generated aerosols. Aircraft generate both primary particles (e.g., soot) and secondary particles that are formed though gas-to-particle conversion in aircraft wakes (e.g., sulfates). Freshly formed soot is believed to be hydrophobic (does not take up water easily), making it a poor CCN but possibly a better IN. However, interactions with sulfuric acid and soot already present in aircraft plumes may modify this hydrophobic behavior markedly (Schumann et al., 1996). If small sulfuric acid particles form in the wake, they will take up water vapor easily, but at first may be far too small to be effective either as CCN or IN. Hence, the evolution of particles in the wake (including coagulation, heterogeneous growth, and deposition of hydrophilic sulfates on soot) can result in changes that affect both direct and indirect radiative properties of aircraft-generated particles. The nature of this evolution, both physically and chemically, is likely to depend on the surface area of the aerosol mix in the wake region. Because these aerosol-production processes depend so strongly (in many cases, exponentially) on such ambient conditions as temperature, humidity, and concentrations of various other species, it is clear that aerosol-production data must be obtained from series of "complete" measurements taken behind aircraft flying at normal flight altitudes. Such measurements were recently attempted in a specific location in the SASS project's Subsonic Aircraft: Contrail and Cloud Effects Special Study (SUCCESS) program, as well as under the German Advanced Technologies Testing Aircraft Systems (ATTAS) program (Schumann et al., 1996).

One of the stated targets of the SUCCESS program was contrails, which are clouds that form in the wake of aircraft under certain favorable conditions. The origin, persistence, growth, and decay of contrails are not all well understood, but are expected to depend on the interplay of environmental, meteorological, and wake conditions. Contrails are the most visible wake effects. They can cover an

appreciable fraction of the sky in areas of heavy air traffic, obviously perturbing local cloudiness and its impact on radiation. These readily observable effects, however, may be of less significance globally than emissions that do not result in contrails but contribute to the CCN available for later cloud nucleation. Consequently, we view the study of contrails as important for SASS, but nonetheless only a part of the larger subject of aircraft emissions and their effects on clouds.

The panel's preliminary recommendations in this area are to:

• Designate a team of researchers to review extant data sets (U.S. and other) for the mid-troposphere (e.g., NASA-PEM, NASA GLOBE, NSF ACE-1), to assess the extent to which they provide a consistent picture of the aerosol and gas-phase characteristics of the free troposphere, and its regional variability. Such an assessment is needed to provide a framework into which the results of brief, intensive measurements can be placed.
• Use these data sets and other information to bound current uncertainties and sensitivities of the relationships among clouds, aerosol, and radiative effects.
• Evaluate and prioritize research strategies on the basis of these existing data sets and uncertainties, balancing research needs against realistic appraisals of cost and achievability.
• Expand the current support for miniaturization of gas-phase instrumentation to include aerosol-measurement instrumentation.
• Increase efforts to characterize the size and properties of soot particles emitted under ambient operating conditions.
• Pursue direct interactions with aircraft manufacturers and air carriers to identify a workable joint strategy for using the commercial air fleet as platforms for measurements of both aerosol and gas-phase species.