A Report of Findings and Recommendations Add Picture

Joint Planning & Development Office Environmental Integrated Product Team

Partnership for AiR Transportation Noise and Emissions Reduction

Workshop on the Impacts of Aviation on Climate Change June 7-9, 2006, Boston, MA

August 31,2006 1

Contents Report Authors...... 2 Acknowledgement...... 3 Executive Summary ...... 4 A. Emissions in the UT/LS and Resulting Chemistry Effects ...... 6 B. Contrails and Cirrus...... 7 C. Climate Impacts and Climate Metrics ...... 10 D. Studies for Trade-offs Amongst Aviation Emissions Impacting Climate ...... 12 1. Background...... 13 1.1 An Introduction to the Issues ...... 15 2. State of the Science...... 19 2.1 Emissions in the UT/LS and Resulting Chemistry Effects...... 19 2.2 Contrails and Cirrus...... 20 2.3 Climate Impacts and Climate Metrics ...... 22 3. Uncertainties and Research Gaps ...... 28 3.1 Emissions in the UT/LS and Resulting Chemistry Effects...... 28 3.2 Contrails and Cirrus...... 29 3.3 Climate Impacts and Climate Metrics ...... 37 4. Use of Climate Metrics in Trade-off Studies ...... 38 4.1 Issues with Metrics for Assessing Aviation Climate Impacts ...... 39 4.2 Key Challenges and Gaps in addressing Trade-offs among Aviation Climate Impacts40 4.3 Extending the Trade-off Space ...... 41 5. Research Needs and Prioritized Recommendations ...... 41 5.1 Emissions in the UT/LS and Resulting Chemistry Effects...... 42 5.2 Contrails and Cirrus...... 46 5.3 Climate Impacts and Climate Metrics ...... 52 5.4 Research for Trade-off Studies...... 54 6. Next steps...... 55 References ...... 56 Appendices...... 59 Appendix 1 Workhop Agenda...... 60 Appendix 2 Workshop Participants/Authors for Each Subgroup ...... 62 Appendix 3 Other Attendees at the Workshop...... 63

Report Authors*

Workshop Chair Donald Wuebbles University of Illinois-Urbana Champaign

Subgroup Leaders Anne Douglass NASA GSFC Bernd Kärcher IPA/DLR, Germany Wei-Chyung Wang SUNY Albany

Other Invited Participants Steve Baughcum Boeing Co. Robert P. d'Entremont AER, Inc. Andy Dessler Texas A & M Univ. Paul Ginoux GFDL NOAA Redina Herman Western Illinois Univ. Andy Heymsfield MMM/ESSL NCAR Ivar Isaksen Univ. Oslo, Norway Daniel Jacob Harvard Univ. Sanjiva Lele Stanford Univ. Jennifer Logan Harvard Univ. J. McConnell York Univ., Canada Rick Miake-Lye Aerodyne Res. Inc. Pat Minnis LaRC NASA Dave Mitchell DRI Dan Murphy NOAA CSD/ESRL Laura Pan NCAR Joyce Penner Univ. Michigan Michael Prather Univ. California-Irvine Jose Rodriguez NASA GSFC Karen Rosenlof ESRL NOAA Ken Sassen Univ. Alaska-Fairbanks Robert Sausen IPA/DLR, Germany Keith Shine Univ. Reading, UK Azadeh Tabazadeh Stanford Univ. Ian Waitz MIT Ping Yang Texas A & M University Fangqun Yu SUNY-Albany

*Appendix 2 lists participants by subgroups and Appendix 3 lists other attendees

Acknowledgement I first want to acknowledge the effort of the Next Generation Air Transportation Sys- tem/Joint Planning and Development Office (NGATS/JPDO) Environmental Integrated Product Team (EIPT) and the Partnership for AiR Transportation Noise and Emissions Reduction (PARTNER) in getting the Workshop established and to thank the FAA Office of Environment and Energy and NASA Science Mission Directorate’s Applied Science Program for providing the support for setting up the actual Workshop. Given that there were no special travel funds provided for attending the Workshop, I also thank the many organizations that contributed this to the participants. The Subgroup leaders, Anne Douglass, Bernd Kärcher, and Wei-Chyung Wang, did a wonderful job of chairing their sessions and organizing the writing for the Workshop report. A special thank you to Mohan Gupta and Malcolm Ko for their extensive help with the meeting preparations and in reviewing this report. I also greatly appreciate the additional reviews of the report we received from David Fahey, Mark Jacobson, and Brian Toon. Finally, I want to also express a big thank you to all of the speakers and participants for their time and efforts. The community interest and participation in this Workshop was exemplary, particularly given the short-lead time for the Workshop. It is this community interest and involvement in establishing the relevance of the issues and in laying the groundwork for the research associated with aviation and climate change that will lead to the successful new research efforts we all hope will result from this Workshop.

Workshop Chair Don Wuebbles

correctly capture the climate impacts of Executive Summary aviation emissions. The effects of aircraft emissions on the As a first step in response to the need to current and projected climate of our address these issues, the Next Genera- planet may be the most serious long- tion Air Transportation System/Joint term environmental issue facing the Planning and Development Office aviation industry [IPCC, 1999; Aviation (NGATS/JPDO) Environmental Inte- and the Environment – Report to the grated Product Team (EIPT)1 and the United States Congress, 2004]. How- Partnership for AiR Transportation ever, there are large uncertainties in our Noise and Emissions Reduction present understanding of the magnitude (PARTNER)2 convened a panel of ex- of climate impacts due to aviation emis- perts from around the world to partici- sions. With extensive growth in demand pate in a “Workshop on the Impacts of expected in aviation over the next few Aviation on Climate Change” during decades, it is imperative that timely ac- June 7-9, 2006 in Boston, MA. The tion is taken to understand and quantify stated goals of the workshop were to the potential impacts of aviation emis- assess and document the present state sions to help policymakers address cli- of knowledge of climatic impacts of avia- mate and other potential environmental tion; to identify the key underlying uncer- impacts associated with aviation. tainties and gaps in scientific knowl- The climatic impacts of aviation emis- edge; to identify ongoing and further re- sions include the direct climate effects search needed and to make prioritized from carbon dioxide (CO2) and water recommendations as to how additional vapor emissions, the indirect forcing on funding may be leveraged to take ad- climate resulting from changes in the vantage of other on-going funded re- distributions and concentrations of search programs; to explore the devel- ozone and methane as a consequence opment of metrics for aviation climate- of aircraft nitrogen oxide (NOx) emis- related trade-off issues; and to help fo- sions, the direct effects (and indirect ef- cus the scientific community on the avia- fects on clouds) from emitted aerosols tion-climate change research needs. and aerosol precursors, and the climate effects associated with contrails and cirrus cloud formation. To enable the de- 1 The EIPT is one of eight integrated product velopment of the best strategy to miti- teams of the NGATS/JPDO charged with the gate these climatic impacts scientists role of formulating a strategy that allows aviation must quantify these impacts and reduce growth consistent with the environmental-related goals of NGATS. The JPDO, jointly managed by current uncertainties to enable appropri- FAA and NASA, serves as a focal point for co- ate action. The only way to ensure that ordinating the research related to air transporta- policymakers fully understand trade-offs tion for all of the participating agencies (Federal from actions resulting from implement- Aviation Administration, NASA, the Departments ing engine and fuel technological ad- of Commerce, Defense, Homeland Security, Transportation, and the White House Office of vances, airspace operational manage- Science and Technology Policy). ment practices, and policy actions im- 2 posed by national and international bod- PARTNER is a Center of Excellence supported by the Federal Aviation Administration, the Na- ies is to provide them with metrics that tional Aeronautics and Space Administration, and Transport Canada.

This report documents the findings and sources of carbon dioxide, the expecta- recommendations of the Workshop. tion is that this percentage will increase While the report does not represent a because of projected increase in avia- peer reviewed assessment, the intention tion and the likely decrease in other is to use the results to provide guidance combustion sources as the world moves for the agencies and private sector away from fossil fuels towards renew- stakeholders participating in the EIPT to able energy sources. In addition, air- develop and implement an aviation cli- craft nitrogen oxides released in the up- mate research plan to reduce uncertain- per troposphere and lower stratosphere ties to “levels that enable appropriate generally have a larger climate impact actions” [Next Generation Air Transpor- than those emitted at the surface, al- tation System-Integrated Plan, 2004]. though some of the much larger surface A major internationally coordinated effort emissions from energy and transporta- to assess the impacts of aviation on the tion sources also reach the upper tropo- global atmosphere was sponsored by sphere. the U.N.’s Intergovernmental Panel on The workshop participants acknowl- Climate Change (IPCC, 1999). Since edged the need for focused research then, while new information has become efforts in the United States specifically available, there has been no compre- to address the uncertainties and gaps in hensive attempt to update the assess- our understanding of current and pro- ment of the science and the associated jected impacts of aviation on climate uncertainties. During this time period, and to develop metrics to characterize there have also been discussions of these impacts. This could be done how to define the best metrics to ac- through co-ordination and/or expansion count for the wide range of spatial and of existing and planned climate research temporal scales of aviation-induced cli- programs or may entail new activities. mate impacts. This workshop agreed Such efforts should include strong and with IPCC (1999) that the three most continuing interactions between the sci- important ways that aviation affects cli- ence community, aviation system opera- mate are (1) direct emissions of green- tors and policy developers to ensure house gases including CO2 and water that the most up-to-date science is read- vapor, (2) emissions of NOx that interact ily available for policy considerations with ozone, methane, and other green- and that scientists are aware of the house gases, and (3) persistent con- questions and challenges faced by op- trails and their effects on cirrus cloud erators and policy developers. Partici- distribution. For this workshop, our focus pants were asked to provide both short- was on the impacts of subsonic aviation term (three to five years) and long-term emissions at cruise altitudes in the up- prioritized research needs with achiev- per troposphere and lower stratosphere able goals and objectives that can help (UT/LS) and on the potential response to address the uncertainties and gaps of the climate system to these emis- associated with key questions on avia- sions. However, it is also recognized tion induced climate impacts. These re- that all aircraft emissions must be con- search priorities also include the need to sidered. identify, develop and evaluate metrics to characterize the climate impacts of avia- Although current fuel use from aviation tion to aid the policy-making process. is only a few percent of all combustion

Research into the environmental effects background state, emphasizing the of aircraft emissions cannot be decou- composition of the upper troposphere pled from basic atmospheric research, and the transition from the troposphere such as fundamental understanding of to the lowermost stratosphere. Im- the physics and chemistry in the UT/LS provements to the representation of the region and aerosol-cloud interactions in atmospheric circulation have resulted in the context of global climate modeling. better models for the composition and The overall science focus of this work- fluxes of ozone and other species in this shop was divided into three particular region. The magnitude of this flux and Subgroups: (1) Emissions in the UT/LS its spatial and seasonal variability are and resulting chemistry effects, (2) Ef- similar to quantities derived from obser- fects of water and particle emissions on vations in some models, and uncertain- contrails and on cirrus clouds, and (3) ties have been reduced. These im- Determining the resulting impacts on provements should help to better assess climate from aircraft emissions and de- the aviation impacts. There are continu- fining metrics to measure these impacts. ing efforts that compare simulations with observations that will likely lead to development of observation-based performance metrics for atmospheric mod- Key Findings els.

A. Emissions in the UT/LS and Uncertainties and Gaps Resulting Chemistry Effects One family of uncertainties in evaluating State of the Science aviation effects on climate derives from specific model formulation errors and The potential importance of aircraft can be addressed by laboratory, nu- emissions of nitrogen oxides (NOx) (and merical, or atmospheric studies that fo- also from hydrogen oxides (HOx) from cus on the specific process. water vapor emissions into the strato- • Aircraft emissions of gases and parti- sphere) on tropospheric and strato- cles. Remaining uncertainties in the spheric ozone has been recognized for emissions, both from uncertainties in several decades. The effects on ozone how much is emitted (e.g., soot, sul- (O3) can also affect the production of fates) from each aircraft and from un- tropospheric OH and thus affect levels certainties in the global distribution of of methane (CH4). There have been these emissions with altitude, latitude, substantial improvements in the chemis- and longitude, need further considera- try-transport modeling tools used to tion. evaluate the impacts of aviation NOx The fundamental NOx and HOx chem- emissions on O3 and CH4 since the • 1999 IPCC assessment. The database istry of the upper troposphere. Large of observations in the UT/LS has been disagreements between the modeled greatly expanded, and data from the and measured abundances of HOx European MOZAIC program (instru- and NOx gases in the upper tropo- ments on commercial aircraft) and fo- sphere point to errors in either the cused field campaigns are commonly measurements or in the tropospheric being used to evaluate the global model chemical mechanisms and rates. This

discrepancy is not found in the lower tion or other breakdowns in atmos- stratosphere. pheric stability are a major uncertainty in evaluating aviation impacts today. • Lightning NOx. Better understanding of lightning NOx in terms of sources Research Recommendations and and their relationship to convection is Priorities needed to evaluate the aircraft perturbation, especially for future climates. • Models and Measurements Intercom- parison. A Models and Measurements • Plume processing of aircraft NOx in Intercomparison, emphasizing the the first 24 hours. Better understand- UT/LS and free troposphere, should ing of the possible conversion of NOx be conducted. This process should to nitric acid (HNO3) in the aircraft lead to model improvements and re- plume needs to be attained. duction of uncertainty in model predic- A second type of uncertainty involves tions. coupling across different Earth system • Vertical transport processes between components, and possible non-linear 2 and 10 km. Additional measure- responses to perturbations and/or feed- ments and data analyses, along with backs within the chemical system. modeling analyses, are needed to re- • The coupling and feedbacks of tropo- duce uncertainties in treatment of con- spheric CH4-CO-OH-O3. There is no vection and other transport processes, single test (based on observations) and in the treatment of lightning ef- that gives confidence that any model fects. accurately responds to a NOx pertur- • Data analysis and modeling. Expand bation. the analysis of the wealth of data be- • Climate change. Analyses of future ing obtained in the UT/LS by different aircraft fleets need to be considered aircraft and satellite platforms to fur- relative to the climate expected. Most ther constrain the magnitude and sea- model studies and all observations of sonality of turnover rates and mixing the meteorology and background processes. chemistry are derived from today's • Re-examine the impacts of aviation in climate. the UT/LS using several of the im- • Scavenging. The process whereby proved models. gaseous HNO3 (and thus NOx) is re- In the longer-term, field campaigns may moved from the atmosphere involves be needed to address issues with HOx- large-scale transport, convection, NOx chemistry in the UT and to better cloud processes and precipitation. understand background processes. Fur- This coupling is a major uncertainty in ther improvements are needed in labo- current chemistry-transport models. ratory studies of heterogeneous proc- • Transport and Mixing. Aircraft emis- esses and low-temperature kinetics. sions will accumulate in atmospheric regions of relatively slow, or stagnant mixing. The seasonality of these re- B. Contrails and Cirrus gions, the apparent barriers to mixing, and the rapid mixing through convec- State of the Science

Contrails form if ambient air along the to modify high cloudiness in the ab- flight track is colder and moister than a sence of contrails is affected by the fre- threshold based on known thermody- quent observation of high supersatura- namic parameters. Contrails initially tions with respect to the ice phase, the contain more but smaller ice crystals relatively small number of heterogene- than most cirrus clouds. Early contrail ous ice nuclei (IN) in cirrus conditions, evolution depends, in not well under- and the ever-presence of mesoscale stood ways, on aircraft and engine temperature fluctuations inducing large emission parameters. At times contrails cooling rates and setting the stage for organize themselves in long-lived, re- cirrus formation. gional-scale clusters in ice supersaturated air masses. The radiative effect of Uncertainties and Gaps contrails is different during the day than A number of uncertainties and gaps at night. Aircraft-induced contrail-cirrus were identified in contrail-cirrus and add significantly to the natural high other aircraft-induced effects on cirrus cloud cover and have the potential, al- clouds: beit with large uncertainties, for a rela- Plume particle processing. It is not well tively large positive radiative forcing (di- • understood how properties (number rect effect). Line-shaped contrails are concentration, surface area, composi- only a portion of the total climate impact tion, and mixing state) of ambient of aviation on the cloudiness. aerosols are perturbed in the presence Recent correlation analyses between of jet engine emissions under various real-time regional-scale air traffic atmospheric conditions and aircraft movements and the occurrence of con- configurations. Detailed investigations trail structures detectable with satellites, of the microphysical and chemical suggest the global coverage of persis- processes governing the evolution of tent, spreading contrails (contrail-cirrus) aviation aerosols in the time scale of and inferred radiative forcing might be days after emission is required. underestimated by an order of magni- Optical properties of contrails, contrail- tude or more, but large uncertainties • cirrus, and cirrus. Factors controlling remain. the radiative properties of cirrus clouds Homogeneous freezing of supercooled and contrail-cirrus (ice crystal habit, aqueous solution droplets initiated by vertical profiles of ice water content, rapid mesoscale temperature fluctua- effective radius) are poorly constrained tions is a ubiquitous pathway to form cir- by observations. The balance between rus clouds in-situ globally. A global im- cooling from reflection of sunlight and pact of aircraft soot particles processed warming from trapping of heat radia- in dispersing plumes on cirrus (indirect tion is also poorly understood. effect) cannot be excluded. By number, Detection and prediction of ice super- aviation might double the background • saturation. Contrails and the expan- black carbon loading in the UT/LS. The sion of contrails into cirrus clouds oc- indirect effect depends, along with de- cur in an ice supersaturated environ- tails of plume processing, on the ability ment. However, the global distribution of background aerosol particles to act as of supersaturation in the upper tropo- ice-forming nuclei. The potential of soot sphere is not adequately known. particles emitted by aircraft jet engines

• In-situ measurements of aerosol a single class of clouds in terms of chemistry and small ice crystals. Cir- their radiative properties; thus they are rus ice crystals can range from a few not capable of predicting contrail-cirrus to hundreds of !m or more in size. cloud fraction from first principles. Measuring this large range requires • Representation of aerosols and con- several instruments and improved trails in global atmospheric models. agreement between instruments. Ex- Both models and satellite datasets isting instrumentation also cannot eas- lack the horizontal and vertical resolu- ily measure the shape of the numer- tion to address many contrail issues. ous very small crystals that have been Data useful for validation of aerosol found in contrails and cirrus clouds. modules is lacking, especially for car- • Properties of heterogeneous ice nuclei bonaceous aerosol. Global models from natural and anthropogenic and contrail/cirrus studies need to es- sources. The atmospheric effects of tablish the essential parameters for ice formation from aviation particles properly incorporating aviation aero- depend on the ice nucleation proper- sols and their effects into atmospheric ties of particles from other anthropo- calculations. genic and natural sources. However, • Long-term trends in contrail-cirrus and concentration and chemical composi- cirrus. Long-term trends can only be tion of IN in the upper troposphere are ascertained from consistent measure- not well known and are difficult to pre- ments over extensive periods, but the dict with models. current satellite record has many un- • Interactions between heterogeneous certainties, limiting the ability to exam- ice nuclei and cirrus clouds. Ice nu- ine past trends. Special care is needed cleation processes occur within short to ensure homogeneous datasets for time scales and are rather localized. estimation of future trends in cirrus Hence it is difficult to determine their measurements. relative importance using in-situ measurements. Ice nucleation path- Research Recommendation and Pri- ways can be isolated in the laboratory, orities but the question arises whether the • In-situ probing and remote sensing of employed IN particles are representa- aging contrail-cirrus and aircraft tive. This issue is particularly important plumes. A series of coordinated re- for aircraft because real engine soot gional-scale campaigns should be de- and its processing cannot easily be signed and executed to measure the represented in laboratory measure- appropriate variables using in-situ and ments. remote sensing measurements with • Incorporation of effects of aviation- the aim to characterize the growth, de- induced particles and cirrus into global cay and trajectories of contrail ice par- models. Accurate knowledge of ice ticle populations. Such measurements supersaturation is crucial for quantify- are also needed to define the abun- ing direct and indirect effects of avia- dance and properties of ambient aero- tion on cirrus cloudiness. Most current sols as well as gaseous aerosol pre- models do not adequately represent cursor concentrations in the tropo- ice supersaturation, and treat cirrus as sphere.

• Regional studies of supersaturation approaches and methods of IN meas- and contrails using weather forecast urements. models have the potential to include Long-term research needs include: (1) our best physics and the high resolu- development or improvement of instru- tion needed to more accurately predict ments that help establish background supersaturation. Development of these concentrations and characteristics of models and associated observational heterogeneous ice nuclei and measure datasets may be the best approach for supersaturation accurately, and (2) the developing our knowledge in the near development and implementation of new term, supporting development of more concepts for ice phase-related micro- accurate tools for use in global mod- physics, supersaturation, radiation, and els. cloud fraction in climate models ena- • Global model studies addressing direct bling a consistent treatment of global and indirect effects. Enhance the aviation effects. treatment of relevant processes, in- The current suite of satellite instruments cluding appropriate parameterizations, is inadequate for evaluating supersatu- in global climate models (GCMs) to ration. Higher resolution (both horizon- improve analyses of contrails and cir- tally and vertically) is required for the rus associated with aircraft. observations. Very accurate measure- • Use of existing or upcoming informa- ments of temperatures as well as water tion from space-borne sensors. Inves- vapour mixing ratio are required to de- tigate the optical and microphysical rive high quality fields of relative humid- properties of contrail and contrail- ity. cirrus, e.g., optical thickness and ef-

fective particle sizes (parameters that are essential to the study of the radia- C. Climate Impacts and Climate tive forcing of these clouds) from Metrics space-borne sensors. • Process studies of plume and contrail State of the Science development. Studies that explore the In assessing the overall impact of avia- role of emitted aerosol particles, and tion on climate, and to quantify the po- how volatile aerosols interact with tential trade-offs on the climate impact each other and with background aero- of changes in aircraft technology or op- sols, are required to understand the erations, metrics for climate change are indirect effect. Studies that investigate needed to place these different climate contrail development as a function of forcings on some kind of common scale. emissions and aircraft design and how Radiative forcing (RF) has been used as contrails evolve into cirrus-like clouds a proxy for climate impact for well-mixed would better quantify the direct effect. greenhouse gases. However, recent • Laboratory measurements of ice nu- analyses have demonstrated that a unit cleation. Laboratory data are urgently radiative forcing from different climate needed to develop aerosol-related pa- change mechanisms does not necessar- rameterisations of heterogeneous ice ily lead to the same global mean tem- nucleation for use in models. It is also perature change (or to the same re- recommended to compare different gional climate impacts). The concept of

11 efficacy (E) has been introduced to ac- exception of the radiative forcing from count for this (i.e., E depends on the the CO2 emissions. The ozone and specific perturbation to the climate sys- methane RFs from NOx emissions are tem, such as changes in ozone or aero- opposite in sign, so the extent to which sol distributions related to aircraft emis- they offset each other is an important sions). Hence, it is the product of E and uncertainty. Estimates for contrails and RF that should be evaluated and inter- cirrus are particularly highly uncertain. compared for the various climate im- • Optical properties of contrails, contrail- pacts from aviation. However, RF is not cirrus, and cirrus. As discussed in pre- an emissions metric capable of compar- vious section. ing the future impact of different aviation emissions. The applicability of emission • Defining metrics for trade-offs. The metrics, such as Global Warming Poten- scientific community may be able to tials (GWPs), have not been adequately define useful metrics for the climate tested and evaluated. In addition, change and climate impacts associ- changes in precipitation and other cli- ated with aviation, but further study mate variables besides temperature are and consensus building is needed. of interest. Climate metrics for aviation need to be done in the context of cli- Recommended Research and Priori- mate metrics for other short-lived per- ties turbations from other sectors. • Radiative effects on climate from contrails and cirrus. In addition to previ- An update of the IPCC (1999) radiative ously mentioned studies, specific stud- forcing (RF) from aviation for the “cur- ies aimed at better understanding the rent” time period finds that, with one ex- climate impacts from contrails and cir- ception, the IPCC findings have not rus. Intercomparisons (model to significantly changed, apart from the in- model) and evaluations (compare crease in air traffic from 1992 to 2000 model to observations) of climate and (Sausen et al., 2005). The exception is radiative transfer models. RF from linear contrails, which appear to be at least a factor of three smaller. • Systematic model intercomparison of There is still no reliable estimate of RF efficacy studies. Evaluate inhomoge- from aviation-induced cirrus clouds. neous vertically and horizontally dis- Based on recent correlation analyses tributed forcing agents. Analyze cirrus some authors suggest that this RF might changes, ozone changes, CH4, and di- be dominating all other aircraft effects. rect particle effects, and effects of It is critical that appropriate metrics be changes in climate state. established before assuming relative • Identify, develop and evaluate metrics climate impacts for various contributions for climate impact assessment and based on potentially inappropriate met- examine their scientific basis. rics. • Quantify the uncertainty in proposed Uncertainties and Gaps metrics and how it propagates (both • Climate impacts are highly uncertain. parametric input uncertainties and There remain significant uncertainties model uncertainties). on almost all aspects of aircraft environmental effects on climate, with the

D. Studies for Trade-offs Amongst There is currently no study in the peer Aviation Emissions Impacting reviewed literature that can be cited to Climate justify, based on the scientific understanding of the impact of aviation emis- Along with furthering scientific under- sions, the possible choices of metrics standing, there are policy-related needs suitable for trade-off application. Re- for sensitivity analyses of the net effects search is needed to examine the effect of trade-offs between various interven- of different metrics, and the choices tions in aircraft operations and emis- within each metric (e.g. time horizon), sions including: on evaluating the relative importance of • NOx reduction technology versus fuel different aviation emissions. Such stud- efficiency (i.e., CO2 emissions) ies would need to explore the potential • flight altitude effects (e.g., effects on of existing metrics and the possibility of ozone and contrail formation) designing new metrics. It must be • changing future geographical distribu- stressed that even if there is a philoso- tions of the fleet phical agreement on an acceptable met- • differential impact of day/night opera- ric, current atmospheric models may not tions be able to calculate these metrics with • routings to avoid certain regions with acceptable accuracy. specialized chemistry (e.g. supersaturated air, cirrus, or polar) • studies of the co-dependence of physical impacts, e.g. how future climate change may alter the ozone response

Climate change metrics are expected to play an important role in these analyses. The IPCC report provided instantaneous forcing due to the cumulative impacts of aviation. While this is a measure of how the atmosphere has changed due to historical aviation activities, such estimates of radiative forcing for aviation may not provide an appropriate basis for making policy or operational decisions (time- dependent effects likely necessary), nor are they an appropriate basis for fully evaluating the relative impacts of various aviation effects. This argument is not new; indeed, it is widely understood and accepted, and a variety of alternative integral measures have been pursued.

military operations.” One of the chal- 1. Background lenges posed by the vision is achieving Aviation has become an integral seg- growth while reducing environmental ment in global economic and transporta- impacts. tion systems, and thus is an indispensa- ble part of modern society, affecting the The potential impact of aircraft emis- lives of millions of people directly or indi- sions on the current and projected cli- rectly. Environmental concerns from mate of our planet is one of the more aviation include effects on noise levels important environmental issues facing near airports, and effects of aircraft the aviation industry. For example, the emissions on local air quality, regional recent Report to the U.S. Congress on and global scale atmospheric chemical Aviation and the Environment (Waitz, et composition, and on climate change, al., 2004) states that for the aviation in- and resulting impacts of these effects on dustry, “The (environmental) topic of human and ecosystem health. However, greatest uncertainty and contention is the absolute magnitude and relative im- the climate impact of aircraft.” portance of these environmental effects Although current fuel use from aviation are not fully understood. is only a few percent of all combustion The U.S. projects demand for air trans- sources of carbon dioxide, the expecta- portation services to grow three fold by tion is that this percentage will increase 2025 (e.g., Next Generation Air Trans- because of projected increase in avia- portation System, 2004). It is a daunting tion and the likely decrease in other challenge for both the scientific and combustion sources as the world moves technological communities to satisfy this away from fossil fuels towards renew- increasing demand, while still protecting able energy sources. In addition, air- our environment. With extensive growth craft nitrogen oxides released in the up- demand expected in aviation over the per troposphere and lower stratosphere next few decades, it is imperative that generally has a larger climate impact vigorous action be taken to understand than those emitted at the surface, al- the potential impacts of aviation emis- though some of the much larger surface sions to help policymakers address cli- emissions from energy and transporta- mate and other potential environmental tion sources also reach the upper tropo- impacts associated with aviation. To sphere. meet the challenges presented by this The integrated national plan for imple- growth, the President of the United mentation of the NGATS initiative is car- States signed ‘Vision 100 – Century of ried out by a Joint Planning and Devel- Aviation Reauthorization Act” in 2003 opment Office (JPDO). The JPDO is and created a multi-agency integrated comprised of a number of U.S. agen- plan for the development of a Next cies: National Aeronautics and Space Generation Air Transportation system Administration (NASA), Federal Aviation (NGATS). The vision of the NGATS is Administration (FAA), Department of “A transformed aviation system that al- Transportation (DOT), Department of lows all communities to participate in the Homeland Security (DHS), Department global market-place, provides services of Commerce (DOC) and the White- tailored to individual customer needs, house Office of Science and Technology and accommodates seamless civil and Policy (OSTP). The Environmental Inte-

14 grated Product Team (EIPT) of JPDO 9, 2006 in Boston, MA. This report has been tasked with incorporating envi- documents the findings and recommen- ronmental impact planning into the dations of the Workshop. This report is NGATS. To fulfill this strategy, we must also aimed at providing guidance for the quantify the climatic impacts of aviation agencies and private sector emissions (e.g., the direct climate ef- stakeholders participating in the EIPT to fects from CO2 and water vapor emis- develop and implement an aviation cli- sions, the more indirect forcing on climate research plan to reduce uncertain- mate resulting from changes in the dis- ties to “levels that enable appropriate tributions and concentrations of ozone actions”. and methane associated with aircraft In 1999, a major international coordi- NOx emissions, the direct effects (and nated effort to assess the impacts of indirect effects on clouds) from emitted aviation on the global atmosphere was aerosols and aerosol precursors, and sponsored by the U.N.’s Intergovern- the climate effects associated with con- mental Panel on Climate Change (IPCC, trails and cirrus cloud formation and 1999). Since then, while new informa- modification) to enable appropriate pol- tion has become available, there has icy considerations and actions. Under- been no comprehensive U.S. or interna- standing aviation’s climate impact is tional attempt to update the assessment also critical to informing the United of the science and the associated uncer- States in the best considerations and tainties. Also, during this time, there trade-offs for setting standards in engine have also been discussions of how to emissions, special flight operations, or define the best metrics to account for other potential policy actions through the the wide range of spatial and temporal International Civil Aviation Organization. scales of aviation-induced climate im- This cannot be adequately done until pacts. This workshop agreed with IPCC the policymakers can correctly capture (1999) that the three most important the environmental effects of aviation ways that aviation affects climate are (1) emissions, including climate impacts. direct emissions of greenhouse gases The extensive investment of new aircraft including CO2 and water vapor, (2) in the marketplace, with their long serv- emissions of nitrogen oxides that inter- ice lifetime (25-30 years or longer), em- act with ozone, methane, and other phasizes the urgent need for improving greenhouse gases, and (3) persistent our current understanding of the effects contrails along with the effects on cirrus of aviation on climate. clouds from contrails and from particles In response to the need to address emitted from aircraft. For this workshop, these issues, the Science/Metrics Panel our focus was on the impacts of sub- of the EIPT and the Partnership for AiR sonic aviation emissions in the upper Transportation Noise and Emissions troposphere and lower stratosphere Reduction (PARTNER), the (UT/LS) and on the potential response FAA/NASA/Transport Canada spon- of the climate system to these emis- sored Center of Excellence convened a sions. panel of science experts from around In addition to studies on aviation, the the world to participate in a two and half Workshop also drew upon results from day “Workshop on the Impacts of Avia- studies associated with other activities tion on Climate Change” during June 7- (e.g., intercontinental (long-range)

15 transport of air pollutants and their ef- uncertainty in understanding aircraft im- fects on ozone in the upper troposphere, pacts on climate. As shown in the particles and associated microphysical agenda and participants lists presented properties of cirrus clouds) so that the in the Appendices, the Workshop was aviation impacts could be put in context based around three Subgroups repre- with other anthropogenic activities. Re- senting these three areas of interest. search into the environmental effects of Plenary sessions attempted to deal with aircraft cannot be decoupled from re- the extensive amount of overlap among search on other atmospheric science these three areas. The discussion below issues, such as the basic understanding centers around each of these areas and of the physics and chemistry in the the findings of the designated Sub- UT/LS and aerosol-cloud interactions in groups at the Workshop. Note that, the context of global climate modeling. there is no attempt at completeness in The stated goals of the Workshop were referencing the discussion that follows. to assess and document the present 1.1 An Introduction to the Issues state of knowledge of climatic impacts of aviation; to identify the key underlying The types of emissions produced by air- uncertainties and the gaps in scientific craft engines are similar in many ways knowledge; to identify ongoing and fur- to the emissions resulting from other ther research needed and to make pri- forms of fossil fuel combustion. Figure 1, oritized research recommendations; to adapted from the IPCC (1999) report explore the development of metrics for shows the real combustion products climate-aviation trade-off issues; to ex- from aircraft engines as compared with amine how to leverage ongoing re- ideal combustion process and also illus- search programs to help address avia- trates the scale of the combustion prod- tion-related needs; and to help focus the ucts. Aircraft emissions are distinct from scientific community on the aviation- other sources burning fossil fuel since climate change research needs of the most emissions occur at cruise altitudes. NGATS. The workshop was designed The most important chemical species as an effort to gather guidance on how emitted by aircraft engines are carbon to address scientific uncertainties and dioxide (CO2), water (H2O), nitrogen ox- gaps associated with climate impacts of ides (NO and NO2, collectively termed aviation. The overall science focus of as NOx), and sulfur oxides (SO2 and sul- this workshop was divided into three fate, collectively known as SOx). The particular areas of interest: (1) Emis- small amount of soot emitted may also sions in the upper troposphere / lower have chemical and radiative effects. stratosphere (UT/LS) and resulting Previous studies have shown that the chemistry effects, (2) Effects of water small amount of emissions of hydrocar- and particle emissions on contrails and bons and carbon monoxide are on cirrus clouds, and (3) Determining the impacts on climate resulting from aircraft emissions and defining metrics for evaluating these impacts. These focal areas were chosen because they are generally recognized by the science community as being the key areas of

Figure 1. Schematic of ideal combustion products and all existing combustion products (adapted from IPCC, 1999). UHC is unburned hydrocarbons.

not important in evaluating effects of air- • Under the right meteorological condi- craft emissions on the global environ- tions, aircraft emissions of water vapor ment. (and aerosols) can lead to formation of Aircraft emissions can alter the radiative contrails and possibly result in effects budget of the Earth and contribute to on upper tropospheric cirrus clouds – human-induced climate change through these effects may exert spatially inho- several different ways: mogeneous radiative impacts on climate. • Aircraft engines emit CO2 and water vapor, important greenhouse gases, Most aircraft NOx emissions are re- that directly affect climate through their leased directly into the chemically com- absorption and reemission of infrared plex and radiatively sensitive UT/LS be- radiation; tween 8-13 km. At the time of the IPCC assessment, there was concern that • Aircraft emit NOx (and HOx produced heterogeneous chemistry following im- from water vapor emissions into the mediate conversion of sulfur to aerosols stratosphere) that modifies atmos- from the aircraft engines could affect the pheric ozone concentrations. Ozone impact on ozone from the NOx emis- affects the radiative balance of the sions. However, recent measurements climate system through both its short- suggest that this immediate conversion wave and infrared (greenhouse effect) is small, such that it will have only a mi- absorption; nor effect on the expected changes in • Through its resulting net production of ozone. upper tropospheric and lower strato- The effect of aircraft emissions on at- spheric ozone, NOx emissions from mospheric ozone concentration de- subsonic aircraft also reduce the at- pends on the altitude at which the emis- mospheric abundance of CH4, another sions are injected. The importance of important greenhouse gas, through ozone production cycles from the NOx feedback effects on concentrations of emissions through the oxidation of tropospheric hydroxyl radicals (OH), methane and hydrocarbons become the primary reactant for destruction of less effective with altitude while the methane; catalytic ozone loss cycles become • Aircraft emit aerosols in the form of more efficient. Any uncertainties in how liquid particles containing sulfate and well we understand the atmospheric organics, and soot particles. Emis- chemical and physical processes in the sions of sulfur dioxide also increase UT/LS affect our ability to understand the aerosol mass in aging plumes. the magnitude of the aviation effects on These aerosols can be radiatively ac- ozone and methane. tive themselves, either by scattering As mentioned above, aircraft influence, (sulfates) or absorbing (soot) solar ra- under the right meteorological condi- diation or can indirectly affect climate tions, high clouds directly by producing through triggering the formation of line-shaped contrails that persist and persistent condensation trails or alter- spread in ice-supersaturated air. The ing the natural cloudiness; resulting clusters of contrail-cirrus are observed on regional scales, sometimes

18 also in regions without significant air is defined by IPCC as the change in net traffic, because they are advected with (down minus up) irradiance (solar plus the wind field over large distances. long-wave; in Wm-2) at the tropopause There are also indications that aircraft after allowing for stratospheric tempera- can influence high clouds indirectly by tures to readjust to radiative equilibrium, injecting aerosol particles that act as but with surface and tropospheric tem- heterogeneous ice nuclei at some point peratures and state held fixed at the un- after emission, without contrail-cirrus perturbed values. The radiative forcing being involved. In the absence of aircraft (RF) concept provides a first order esti- emissions, a cirrus cloud might not have mate of climate change without the need formed or the resulting cirrus might have to actually conduct time consuming and different optical properties. computationally expensive climate It is important to recognize that a cou- model simulations. However, as dis- pling may exist between the pure direct cussed later, this concept has significant and indirect effects. It is conceivable limitations for spatially inhomogeneous that the indirect effect occurs along with perturbations to the climate system and contrail-cirrus, because ice supersatura- can be a poor predictor of the global tion required for persistence facilitates mean climate response. As a result, al- ice formation at low supersaturations. If ternative definitions have been devel- that occurs, contrail-cirrus can exert an oped (as will be discussed later). indirect effect on their own. Cirrus may Figure 2, adapted from IPCC 1999 have different properties because they (Prather et. al., 1999), portrays a causal nucleate in regions with preexisting ice chain whereby the direct emissions of and share the available water. The aircraft accumulate in the atmosphere, situation is further complicated by the change the chemistry and the micro- fact that background cirrus itself might physics, and alter radiatively active sub- be affected by aircraft in heavily traveled stances in the atmosphere, which flight corridors. change radiative forcing and hence the The climate effects from current aircraft climate. emissions are only a small fraction of It is important to consider the relative the total effects of human activities on time scales associated with the effects climate (e.g., emissions of carbon diox- on climate from these various emis- ide from aviation are a few percent of sions. After emission, effects of the NOx the total emissions from fossil fuel burn- emissions from aircraft on ozone will last ing and changes in land use). As a re- for weeks to months (perhaps a little sult, it is very difficult to use a climate longer for stratospheric emissions), model to directly evaluate the climate while affects on methane, because of its effects resulting from aviation. Radiative long atmospheric lifetime, last for years. forcing has been widely used as a met- Effects from emissions of particles and ric of climate change to measure the from contrail formation should last much relative efficacy of climate change shorter, generally not more than days to mechanisms. The radiative forcing of a few weeks. On the other hand, emis- the surface-troposphere system due to sion of CO2 will affect climate for centu- the perturbation of an agent (say, a ries or longer. After the initial uptake of change in greenhouse gas concentra- about half of emitted CO2 by vegetation tions or a change in the solar constant), and the surface ocean, some stays in

19 the atmosphere with a decay time of in the atmosphere for many thousands 100 or more years, and up to a quarter of years. of the emitted CO2 is expected to remain

Figure 2. Schematic of possible mechanisms whereby aircraft emissions impact climate. Climate impact is represented by changes in global mean surface temperature (T s) and global mean sea level rise ( msl). Adopted from IPCC (1999).

2. State of the Science 2.1 Emissions in the UT/LS and Resulting Chemistry Effects The IPCC (1999) assessment served as the starting place for the Workshop. There have been substantial improve- Evaluation of the current state of under- ments in the chemistry-transport model- standing then grew out of the scientific ing tools used to evaluate the impacts of papers and other publications that have aviation NOx emissions on O3 and CH4, been published since then. Each Sub- since the 1999 IPCC assessment, but group provided their insights into key as noted in the 2004 report to the United aspects of the current understanding of States Congress Aviation and the Envi- the science relative to determining air- ronment, “there are currently no major craft effects on climate. research programs in the United States to evaluate the unique climatic impact of aviation.” Although there has been no recent assessment, there have been substantial improvements in the models that would be used to produce such an

20 assessment (and a few sensitivity stud- sessment of the effects of aviation using ies of aircraft effects using European state-of-the-art models. models). In the U.S., coupled tropo- There are continuing efforts that con- sphere-stratosphere chemistry-transport front simulations with observations, and models include those developed by the these efforts may lead to the develop- NASA Global Modeling Initiative (GMI) ment of observation-based performance and the National Center for Atmospheric metrics for models. Such efforts serve Research (NCAR) MOZART. An as- several purposes. The first is to identify sessment of aviation effects using state- and characterize model deficiencies; of-the-art models would decrease the such activities lead to model improve- uncertainty in the climate response due ments. The second is to identify the ex- to chemical changes associated with isting models for which transport and aircraft emissions because of improved chemical processes combine to produce representation of atmospheric transport the most realistic simulated background in the upper troposphere and lower atmosphere with which simulated air- stratosphere UT/LS. However, signifi- craft exhaust will interact. The third is to cant uncertainties remain as discussed develop a strategy for understanding the below. different responses among models used The data base of observations in the in aircraft assessment efforts. UT/LS has been greatly expanded since 1999, and data from the European MO- ZAIC program (in-situ ozone and water 2.2 Contrails and Cirrus instruments on commercial aircraft) and field campaigns are commonly being Because the IPCC identified contrails, used to evaluate the global model back- contrail-cirrus, and modifications of cir- ground state, emphasizing the composi- rus by aircraft exhaust as the most un- tion of the upper troposphere and the certain components of the aviation im- transition from the troposphere to the pact on climate, the majority of recent lowermost stratosphere. However, the studies have focussed on cloud proc- amount of data available is still insuffi- esses, while a limited number of studies cient to fully establish a climatology for also addressed chemical effects. More NOx, HOx and many other constituents recently, ground-based experimental of the UT/LS. studies on jet engine emissions have been carried out. Most of this research Improvements to model representations is summarized in a few review papers of the stratospheric circulation have re- (Kärcher et al., 2004; Schumann, 2005; sulted in better simulations of the com- Miake-Lye, 2005), from which the follow- position of the lower stratosphere and in ing key results have been drawn. more realistic fluxes of ozone and other stratospheric species from the lower 1. Aircraft emissions cause significant, stratosphere to the upper troposphere. detectable enhancements of NOx, vola- The magnitude of these fluxes and their tile particles, and soot (and possibly H2O spatial and seasonal variability are simi- in the lowermost stratosphere) on the lar to quantities derived from observa- plume scale at cruise. Signals at re- tions in some models, and uncertainties gional scales can only be inferred from have been reduced. These improve- models, the magnitudes of which are ments would have an impact on an as- determined by emission rates and at-

21 mospheric lifetimes. The contribution The following findings affect the poten- relative to ambient levels in the global tial of aircraft exhaust to modify high UT/LS is variable owing to the large cloudiness in the absence of contrails variability of particles and most trace (indirect effect). (i) The frequent obser- gases. vation of high relative humidities with 2. Contrail formation relies on known respect to the ice phase up to the level thermodynamic parameters and they where homogeneous freezing com- develop if the air is colder and moister mences (>150%). (ii) The relative pau- than a threshold. They initially contain city (low number) of heterogeneous ice more but smaller ice crystals than most nuclei (IN) – insoluble particles that cirrus clouds. Early contrail evolution might facilitate the formation of the ice depends in poorly understood ways on phase – in cirrus conditions. (iii) The aircraft and engine emission parame- ever-presence of mesoscale tempera- ters. ture fluctuations inducing large cooling rates (~10 K/h) on short time scales (10- 3. Aircraft-induced contrail-cirrus add 20 minutes) and setting the dynamical significantly to the natural high cloud stage for cirrus formation. These obser- cover and modulate the radiative bal- vations place bounds on the effect soot ance, with effects that are particularly particles emitted by aircraft jet engines important during the daytime. Line- can exert in the atmosphere: it seems shaped contrails also contribute to a unlikely that exhaust soot particles portion of the total climate impact from would be better IN (i.e., cause ice parti- aviation effects on cloudiness. The di- cle formation close to ice saturation) rect contrail impact is affected by wind than mineral dust grains or other natural shear and the duration and size of ice IN, and soot has to compete with IN supersaturated regions controlling the from these other sources (particularly contrail-to-cirrus transition and contrail- with mineral dust aerosol) and with liq- cirrus properties. uid particles that freeze homogeneously 4. The upper troposphere is frequently at times. (the lowermost stratosphere occasion- Another major research result concerns ally) supersaturated with respect to ice. the direct effect of persistent, spreading Homogeneous freezing of supercooled contrails (contrail-cirrus). According to aqueous solution droplets initiated by recent, yet unpublished, correlation rapid mesoscale temperature fluctua- analyses between real-time regional- tions is a ubiquitous pathway to form cir- scale air traffic movements and the oc- rus clouds in-situ globally. currence of contrail structures that are 5. The global impact of the indirect ef- detectable with satellites, it seems plau- fect of aircraft soot particles processed sible that the global persistent contrail in dispersing plumes on cirrus is un- coverage (and inferred radiative forcing) known, and depends, in addition to de- might exceed the IPCC and more recent tails in the plume processing, on the European TRADEOFF estimates by an ability of background aerosol particles to order of magnitude or more (but error act as ice-forming nuclei. By number, bars in these studies are also very aviation might double the background large). As satellites are more capable of black carbon loading in the UT/LS detecting aged, spread out contrail- (Hendricks et al., 2004). cirrus rather than young, very narrow

contrails, it is difficult to arrive at more traffic from 1992 to 2000. The O3 and accurate estimates of the direct effect CH4 effects are changed due to more without advances in observational tools recent analyses from European chemi- or the help of global models. To date, cal-transport models. The other major however, global models are not yet able change is found for the direct global RF to simulate contrail-cirrus properly due from (linear) contrails; the new value is mainly to our limited understanding of roughly a factor of 3 smaller than IPCC the processes affecting them and, as a (1999) based on results from Marquart result, our lack of suitable parameteriza- et al. (2003) and Myhre and Stordal tions for their subgrid-scale nature. (2001), which were scaled (by fuel burn) 2 to the year 2000 resulting in 6 mW/m and 15 mW/m2, respectively. However, 2.3 Climate Impacts and Climate another, yet unpublished, study (Penner Metrics and Chen, 2006) suggests that the radiative forcing from cirrus, i.e., cloud The discussion starts with an update changes beyond linear contrails, could based on recent derivations of the radia- be significantly larger than those in Fig- tive forcing from aircraft emissions rela- ure 3 or Table 1. The overall conclusion tive to the findings in IPCC (1999). This from these analyses is that significant discussion is then followed by an analy- uncertainties still remain in quantifying sis of the concept of climate metrics the impacts of aviation emissions on such as radiative forcing. climate. Note that the RF for direct soot in Figure 3 and Table 1 are based on 2.3.1 Radiative Forcing Update the atmospheric soot concentrations, An update of the IPCC (1999) radiative and does not include the soot incorpo- forcing (RF) from aviation for the “cur- rated into clouds or long-term deposition rent” time period has been presented by to the ground. Sausen et al. (2005). Figure 3 and Table 1 summarize their results as well as the In IPCC (1999), similar analyses for pro- findings from IPCC (1999). In view of jections of the commercial aircraft fleet the large error bars of IPCC (1999), the to 2050 showed much higher radiative forcings from aircraft for 2050 compared RF from CO2, H2O and direct effect of sulfate aerosols have not changed sig- to 1992, but with even larger uncertain- nificantly, apart from the increase in air ties on the individual effects.

Figure 3: Global radiative forcing (RF) [mW/m2] from aviation for 1992 and 2000, based on IPCC (1999) and TRADEOFF results. The whiskers denote the 2/3 confidence intervals of the IPCC (1999) values. The lines with the circles at the end display different estimates for the possible range of RF from aviation induced cirrus clouds. In addition the dashed line with the crosses at the end denotes an estimate of the range for RF from aviation induced cirrus. The total does not include the contribution from cirrus clouds (Sausen et al., 2005).

Table 1: Radiative forcing (RF) derived for various emissions from aviation [mW/m2]. The best estimates for 1992 by IPCC (1999) and two estimates for 2000 are given: one is derived from IPCC (1999) by linear interpolation; the second is based on the mean values resulting from the TRADEOFF project. As in IPCC (1999), the European TRADEOFF RFs for CO2, O3 and CH4 were scaled by a factor of 1.15. The RFs from O3 and CH4 are both a result of aircraft NOx emissions (Sausen et al., 2005). Year Study RF [mW/m2]

o s e / l t

i )

w a a t t ( s

l c c t t

u p a 2 4 e e r n l o m t

O r r r 3 O H 2 i i o i u o u o C O C H D S D S C T S C 1992 IPCC (1999) 18.0 23.0 - 14.0 1.5 - 3.0 3.0 20.0 48.5 2000 IPCC (1999)* 25.0 28.9 -18.5 2.0 - 4.0 4.0 33.9 71.3 2000 TRADEOFF 25.3 21.9 -10.4 2.0 - 3.5 2.5 10.0 47.8 * scaled to year 2000

complex models and be relatively easily 2.3.2 Climate metrics for aviation applied. Hence transparency and sim- Emissions by aviation are responsible plicity of application are necessary at- for a range of atmospheric changes that tributes of metrics. Clearly such trans- perturb the radiation budget and hence parency and simplicity come at a cost in force climate change. In assessing the terms of compromising the sophistica- overall impact of aviation on climate, tion with which climate effects can be and to quantify the potential trade-offs in properly represented. The metrics allow the climate impact of changes in aircraft the user to explore, for example, the technology, operations, or even the trade-offs in terms of climate impacts of amount of aircraft traffic, it is important changes in aircraft technology or opera- to place these different climate forcings tions. By providing ranges in the values on some kind of common scale. We re- of metrics, they can also be used to fer to methods that attempt to achieve communicate the level of scientific un- this as “metrics”. Although their existing certainty, and the impact of that uncer- application to aircraft issues is much tainty, on possible policy options. more limited, the general usefulness and uncertainties associated with met- Radiative Forcing rics for climate change has been the The most straightforward metric is the subject of many published research traditional one, namely radiative forcing studies. (RF) at some given time due to the cu- There are many difficulties in developing mulative impact (both direct and indi- such metrics, which while not unique to rect) of aviation emissions during some aviation, are certainly exacerbated by prior time period. The global mean RF is the nature of aviation’s impacts on cli- linearly related to the expected equilib- mate. Aviation directly leads to in- rium global mean surface temperature creases in carbon dioxide and indirectly change via a model dependent climate to decreases in methane, both of which, sensitivity parameter (e.g. IPCC, 2001). as a result of their long atmospheric RF allows a first order estimate of the residence times, are spread relatively overall impact of aviation on climate, homogeneously across the globe. Avia- and also the relative size of the different tion also leads to increases in ozone contributors to aviation-induced climate and contrails, and other cloudiness change. changes, which, partly because they are It was originally thought that a unit radia- shorter-lived, depend sensitively on the tive forcing from different climate height, time and geographical location of change mechanisms in the same posi- the emissions. tive or negative direction (e.g., due to The quantification of climate metrics changes in amounts of greenhouse should, wherever possible, use input gases or changes in the solar output) from state-of-the-art models; but a prime would lead to the same global mean aim of the development of metrics in this temperature change. However, recent context is that they should then facilitate research has indicated that this does not at least a first order assessment of im- appear to be generally the case, and the pacts without compromising the science concept of efficacy E has been intro- and without the further use of these duced into the radiative forcing concept

25 to account for this (i.e., E depends on the large contrasts in the residence time the specific perturbation to the climate of the different contributors to RF (as system, such as changes in ozone or discussed earlier in this report). aerosol distributions related to aircraft IPCC (1999) also introduced the con- emissions). Hence, it is the product of E cept of the radiative forcing index (RFI), and RF that should be evaluated and which is simply the ratio of the total forc- intercompared relative to the environ- ing at a given time to the forcing from mental impacts. In addition, the equal carbon dioxide at the same time. Unfor- global-mean RFs do not necessarily tunately, the RFI has been misapplied in lead to the same distribution of surface some quarters (as discussed by Forster temperature response; broadly the re- et al. 2006) as a way of crudely account- sponse is largely concentrated in the ing for the future non-CO climate hemisphere in which it occurs, but be- 2 change impacts of aviation, by simply yond this there is little relationship be- multiplying the CO2 emission scenarios tween the spatial patterns of forcing and by the RFI. This example points to the response (e.g., Joshi et al. 2003; Han- need for scientific oversight of applica- sen et al. 2005). This is particularly im- tions of climate metrics in policy consid- portant in the context of NOx driven erations. changes in ozone and methane; the ozone changes, and the resulting posi- Emission Metrics tive RF are largely in the hemisphere of For comparison of the climate impact of emission while the negative RF associ- emissions, a whole class of metrics has ated with methane decrease is more been proposed (see e.g., Fuglestvedt et globally distributed. Hence offsets in the al., 2003). These aim to provide an ex- global mean temperature response are change rate, so that each emission can not necessarily reflected in the local be given a CO2-equivalence. temperature change (e.g. Berntsen et al. 2005). The Kyoto Protocol to the UN Frame- work Convention on Climate Change IPCC (1999) calculated the change in has adopted the Global Warming Poten- radiative forcing due to cumulated emis- tial (GWP) concept as developed for the sions from the historical aircraft fleet. It IPCC climate assessments to provide was a way to address the question of this equivalence. The GWP is the time- how the forcing will be different if no air- integrated radiative forcing following a craft have been operating or a specific pulse emission of a gas, relative to the emission is omitted. It was NEVER in- same quantity for a pulse emission of tended that those radiative forcing num- carbon dioxide. Hence it accounts for bers should be used directly in policy both the radiative strength of the climate considerations. The convention of calcu- change agent and its persistence in the lating RF in this way, as an absolute atmosphere; the Kyoto Protocol has change in concentrations over a given chosen the time horizon of 100 years. period of time, is a “backward looking” The GWP concept has been the subject approach that is not a suitable metric for of considerable scrutiny and criticism comparing the impact of emissions (as (e.g. Smith and Wigley, 2000a,b; compared to concentration changes) O’Neill, 2000). Nevertheless, their rou- during some specified period on the tine use is now deeply embedded in the subsequent climate change, because of policymaking community, possibly be-

26 cause of their ease of application, trans- transport models and radiative transfer parency and the absence of an obvious schemes and hence should be regarded replacement (Fuglestvedt et al. 2003). as somewhat crude and preliminary. Modifications to the GWP have been There are many potential choices in proposed – for example, the GWP can metric development (e.g. Fuglestvedt et be multiplied by the efficacy (Berntsen al. 2003; Shine et al 2005). Should et al., 2005), and a version of the GWP pulse or sustained emissions be cho- for sustained (step-change) emissions, sen? What is the dependence on the rather than pulse emissions, has been background atmosphere? Should the proposed and used (e.g. Johnson and metric integrate the “climate effect” over Derwent, 1996; Berntsen et al., 2005). time (as the GWP does, integrating the The gases controlled by the Kyoto Pro- radiative forcing)? Or should it be the tocol are long-lived greenhouse gases, climate effect (e.g. temperature change) and each gas can be represented by a at some particular time? In either case, single value for a given time horizon. a choice has to be made of the time of The situation is much more difficult for the end-point. What climate effect should be used? The most basic is ra- emissions of short-lived gases, where diative forcing, but in principal, it could the GWP may depend significantly on be temperature change, precipitation the location and time of emission, and in change, frequency of extreme events, the case of aircraft, the altitude of emis- sea-level rise, or even the monetary sion. This complexity led IPCC (1999) to value of the “damage” due to the climate reject the possibility of applying GWPs change. In general, the more a metric for aviation, although they did not rec- reflects the actual impact of climate ommend any alternatives. Others (Klug change, the more uncertain is its calcu- and Ponater, 2000; Svensson et al., lation, and a balance between practica- 2004) have taken a more pragmatic bility and relevance has to be struck view and attempted to develop height- (e.g. Fuglestvedt et al. 2003 – see Fig- dependent metrics for aviation emis- ure 4 below). sions; however these efforts did not use input from state-of-the-art chemical

Figure 4. A simplified representation of the cause-effect chain from emissions to climate change and damages, as presented by Fuglestvedt et al. (2003).

nisms and rates. The discrepancy is 3. Uncertainties and Research not found in the lower stratosphere. Gaps This uncertainty is fundamental to the Each Subgroup identified the key uncer- calculated increase in O3 and de- tainties and research gaps associated crease in CH4 from aircraft NOx. with their topic. The analyses of the • Lightning NOx. Lightning-generated state of the science led directly into the NO provides much of the background discussion on remaining uncertainties NOx in the upper troposphere. Be- and gaps. cause the response to NOx perturbations is non-linear and depends on this background, it is essential to deter- 3.1 Emissions in the UT/LS and mine what NOx levels would be with- Resulting Chemistry Effects out aircraft. An understanding of lightning NOx in terms of sources and their There are two types of scientific uncer- relationship to convection is needed to tainty that greatly impact the evaluation evaluate the aircraft perturbation, es- of aviation's impact from emissions of pecially for future climates. NOx on climate and global air quality. One family of uncertainties derives from • Plume processing of aircraft NOx in specific model formulation errors and the first 24 hours. The high concentra- can be addressed by laboratory, nu- tions of NOx in the exhaust plume merical, or atmospheric studies that fo- along with other short-lived exhaust cus on the specific process. In terms of products and ice particles can lead to aviation impacts, some key process un- rapid conversion of NOx to HNO3, certainties are: which effectively cuts off the aircraft perturbation to O and CH . Different • Aircraft emissions of gases and parti- 3 4 implementations of the plume process- cles. Remaining uncertainties in the ing in global models still shows a wide emissions, both from uncertainties in range (e.g., from 2 to 10% of the NO how much is emitted (e.g., soot, sul- x removed) and needs to be resolved. fates) from each aircraft and from un- Zero plume-processing produces a certainties in the global distribution of maximum change in ozone, and thus these emissions with altitude, latitude, limits can be attached to this effect. and longitude, need further consideration. A second type of uncertainty involves coupling across different Earth system • The fundamental NOx and HOx chem- components, and possible non-linear istry of the upper troposphere. Large responses to perturbations and/or feed- disagreements between the modeled backs within the chemical system. and measured abundances of OH, HO2, NO and NO2 in the upper tropo- • The coupling and feedbacks of troposphere have been found in INTEX and spheric CH4-CO-OH-O3. These feed- other airborne campaigns. These dis- backs determine the balance of the crepancies point to errors in either the aviation NOx climate impacts through measurements or in understanding of greenhouse gases: i.e., between O3 surface emissions, the effects of long- increases and CH4 decreases. There range transport of pollutants, and/or is no single test (based on observa- the tropospheric chemical mecha- tions) that gives confidence that a

model accurately responds to a NOx • Correlation between operational flight perturbation since the impact on CH4 routing and meteorology. Emissions is a remote response that involves CO inventories have generally been and transport away from the region of evaluated as a mean set of emissions maximum NOx perturbation. A related operating on a varying meteorology uncertainty involves the cross-over (and hence transport of those emis- point, i.e., the altitude (as a function of sions). The correlation of emissions latitude) in the lower stratosphere at with the meteorology of the day could which the addition of aircraft NOx lead to biases in the location of emis- changes from an O3 increase (low alti- sions relative to the tropopause and tudes) to a decrease. This cross-over also in preferentially upward moving point depends not only on the chemi- air masses (leading to greater accu- cal coupling, but also the transport in mulation of emissions) or downward the lower stratosphere. moving ones (less accumulation and impact). • Climate change. Our models and observations of the meteorology and • Transport and Mixing. Aircraft emis- background chemistry are derived sions will accumulate in atmospheric from today's climate. It is clear that regions of relatively slow, or stagnant climate is changing and that the future mixing. The seasonality of these re- aircraft fleet will operate in a different gions, the apparent barriers to mixing, atmosphere in terms of both chemistry and the rapid mixing through convec- and transport. Key changes affecting tion or other breakdowns (e.g., tro- aviation impacts are those to upper popause folds) in atmospheric stability troposphere temperature and relative are readily observed in the strato- humidity (for contrails and cloud proc- sphere and upper troposphere. It is not essing, background levels of O3 and easy to quantify this transport, and NOx and aerosols, the tropopause hence the accumulation of emissions boundary (e.g., it has moved up 200 m in these regions. This is a key uncer- over the last 2 decades), and the tainty in evaluating aviation impacts overall rates of vertical transport in the today. For example, the transport UT/LS. through the lower stratosphere and the transition regions in the upper tropo- • Scavenging. The removal of aircraft sphere may be slow enough to accu- NO from the atmosphere typically in- x mulate and enhance the background volves chemical conversion to HNO 3 levels of H O in these regions, leading followed by scavenging in aerosols or 2 to a direct greenhouse effect as well hydrometeors that deposit it on the as perturbing the HO ground. The process whereby gase- x-NOx chemistry. ous HNO3 is scavenged thus involves large-scale transport, convection, cloud processes and precipitation. 3.2 Contrails and Cirrus This coupling is a major uncertainty in A number of uncertainties and gaps current CTMs; however, since most of were identified relative to understanding the HNO3 will not be chemically re- the effects of contrails and aircraft- activated to NOx, this introduces only a induced effects on cirrus clouds: small uncertainty in aviation impacts.

• Detection and prediction of ice super- derstood. For example, the behavior of saturation. Contrails and the expan- instruments that measure in-situ con- sion of contrails into cirrus clouds oc- centrations and sizes of ice particles cur in a supersaturated environment. may differ dramatically. Some instru- However, global distributions of super- ments based on particle impaction do saturation in the upper troposphere, not measure high concentrations of where aviation-produced cirrus is likely small crystals (diameter <70 !m), to occur, are not well known. Satellites while optical probes may measure are currently on line that may be ca- small crystal concentrations orders of pable of improved detection of super- magnitude greater than the impaction saturation in the upper troposphere, instruments. The radiative properties including the MLS (Microwave Limb of contrails also depends on the Sounder), AIRS (Atmospheric Infrared shapes of their crystals, and due to the Sounder) and TOVS (TIROS Opera- small crystal sizes and the spatial tional Vertical Sounder) instruments. resolution of the imaging probes, the Required is a remote sounding instru- shapes of these small crystals are un- ment that measures both temperature certain. The radiative properties of cir- and humidity with good vertical resolu- rus clouds and contrail-cirrus also de- tion and/or can detect relative humidity pend on the vertical distribution of their directly. However, current retrievals microphysical properties. An examina- adequate for water vapor and tem- tion of those radiative properties using perature measurements under sub- remote sensing instruments (e.g., li- saturated conditions may not be suffi- dar, radar, radiometers, and interfer- cient for supersaturated cases. Exist- ometers) is often confounded by the ing satellites do not have the horizon- presence of lower level clouds. tal or vertical resolution to accurately • Interactions between heterogeneous define the frequency and extent of su- ice nuclei and cirrus clouds. For cirrus persaturated regions. clouds, nucleation of ice particles can • Optical properties of contrails, contrail- occur via the heterogeneous action of cirrus, and cirrus. Existing research on insoluble ice nuclei (IN). At tempera- contrails and contrail-cirrus indicates tures below about –38°C, nucleation that ice particle sizes are smaller on occurs by the homogeneous freezing average than in natural cirrus (Baum- of liquid solution droplets. For cirrus gardner and Gandrud, 1998; Minnis et forming in-situ, homogeneous freezing al., 1998; Sassen and Hsueh, 1998; occurs in increasingly concentrated so- Schröder et al., 2000). Because the lution droplets as temperature de- radiative effects depend contrail dy- creases below –36°C. Heterogeneous namics and spatial structure and on ice nucleation in particles that are par- many environmental factors including tially or fully soluble can potentially the surface temperature and presence cause cirrus formation at warmer tem- of low level clouds, it is difficult to de- peratures, and, for temperatures below termine the overall net warming effect –38°C, lower relative humidities. Het- on climate for contrails and contrail- erogeneous ice-nucleation mecha- cirrus. nisms (modes) most relevant for UTLS

The extent of feedbacks between mi- conditions include immersion freezing crophysics and radiation is poorly un- (ice nucleation induced by an IN previ-

31 ously immersed in a liquid aerosol sources. Laboratory studies using sur- droplet), contact freezing “inside-out” rogates for airborne crustal and min- (freezing initiated when a solid IN im- eral dust particles predict the strong mersed in a liquid aerosol particle col- ice-nucleation efficiency of such parti- lides with the drop surface from the in- cles throughout cirrus cloud forming side), and deposition nucleation (direct temperatures. nucleation of ice from vapor at a solid Cirrus cloud ice crystal concentrations particle surface) which may occur in for sizes above 100 !m and for tem- rare cases when IN reach the upper peratures both above and below – troposphere in a dry state. Very recent –3 38°C are typically of order 0.01 cm , research demonstrates the relevance similar to observed IN concentrations. of organics in heterogeneous ice nu- From this, it is reasonable to conclude cleation (Zuberi et al., 2001; Zobrist et that the first ice particles in cirrus are al., 2006; Shilling et al., 2006). Not all formed through heterogeneous ice nu- of these modes have been shown to cleation, if IN are available. operate efficiently at cold temperatures. The concentration and nuclea- Aircraft soot particles have to compete tion relative humidity of IN mainly de- with efficient IN in dust layers, but dust termine their impact on cirrus cloud aerosol is geographically and season- properties (DeMott et al., 1997; Kär- ally highly variable and it remains un- cher and Lohmann, 2003). certain how many dust particles are actually present in aircraft flight corri- The basic uncertainty associated with dors. In the absence of dust, meas- ice nucleation processes is that they urements in aging aircraft plumes face occur within short time scales (often the difficulty of distinguishing between only within seconds) and are rather lo- soot particles from aircraft exhaust and calized (in sufficiently supersaturated those from other sources (biomass patches of air). For this reason, it is burning, forest fires) present in the extremely difficult to determine their background atmosphere. Recent find- relative importance in in-situ meas- ings that certain organics might cause urements, or to even determine the precipitation of ice-nucleating crystal- basic nucleation mode. It is possible to line solids in liquid particles render ef- isolate different ice nucleation path- forts to disentangle the roles of various ways in the laboratory, but the ques- particle types in ice formation even tion arises whether the employed IN more complicated. particles are representative of atmospheric particles, an issue particularly • In-situ measurements of aerosol com- important for aircraft because real en- position and small ice crystals. It is dif- gine soot and its processing cannot ficult to make in-situ measurements of easily be represented in laboratory both the aerosol composition and measurements. small ice crystals. The low mass loading of particles provides a challenge to Past studies of IN compositions have instrumentation even if they can ade- identified clay particles and mineral quately measure particle composition dust as important atmospheric IN. Li- at lower altitudes. Mass spectrometric dar studies have documented the data indicate that, at least in some re- strong cloud-glaciating effect of dust gions, a majority of the particle mass particles from both Asian and Saharan

32 in the upper troposphere is carbona- • Representation of emitted aerosols ceous (e.g., include organic and ele- and contrails in global atmospheric mental carbon). These data do not ex- models. Assessment of aviation ef- tend to the most numerous small parti- fects on cirrus cloudiness in global cles below ~100 nm in diameter, nor is models requires inputs of the emitted there information on the type of or- aerosol quantities and their relevant ganic molecules. These organics could properties. Currently no detailed parti- significantly change the freezing be- cle emissions inventories exist, as havior of particles, affecting the evolu- have been used for assessing gase- tion of contrail-cirrus and cirrus. ous aviation emissions like NOx and CO as a function of fleet technology There are significant problems with the and operation under various scenar- measurement of small ice particles. ios. Such detailed inventories of parti- Cirrus ice crystals can range from a cles including their gaseous precur- few to hundreds of !m or more in di- sors SO and hydrocarbons are ameter. Measuring this range requires 2 needed to account for the variations in several instruments and the agree- non-volatile and volatile particle num- ment between instruments in the size ber, size, and character across the range has not always been good. range of engine technologies currently Although in cirrus clouds the basic size in use or expected in the future. Be- modes are not in as much question as cause the changes in emission inven- for mixed phase clouds, it is question- tories have not been addressed, there able whether any of the existing have only been a few attempts to probes can obtain accurate sizes of study the climate impact of persistent non-spherical ice crystals. Further- contrails. more, not only the size but also the Process studies of the evolution of crystal habit is important for the radia- particle properties and their role as IN tive properties of the cloud. Existing will be required to generate effective instrumentation cannot easily measure emission indices for input to the global the shape of the numerous very small models, since the characteristics of the (diameter <20 !m) crystals that have emitted particles may have changed often been found in contrails and cirrus significantly by the time the emissions clouds. Ice can form much more easily have mixed to global model grid on some particles in the atmosphere. scales. In that regard, global models The measurement of these ice nuclei and contrail/cirrus studies need to es- in both contrail-cirrus and the back- tablish the essential parameters for ground atmosphere is crucial to un- properly incorporating aviation aero- derstanding how particles emitted from sols and their effects into atmospheric aircraft compete with background par- calculations. As these inputs are de- ticles in the formation of new or in the termined, the global atmospheric modification of existing cirrus. There models can account for the effects are many fewer IN than other particles emitting particles as a function of the and improved instrumentation to frequency and extent of supersatura- measure their number and properties tion with respect to ice, the optical is needed. properties of the generated aerosol, and the radiative implications of the

resulting high cloud cover generated of particles of different sizes as a func- due to aviation operations. tion of time in dispersing aircraft plumes. The aircraft-generated parti- • Plume particle processing. Current cles interact with background aerosols global models treat aircraft emissions through coagulation and mixing, and as well mixed within the grid box, ig- will eventually become part of ambient noring plume processing of the emis- aerosols. In addition, photochemistry sions. However, the effects of non- will provide additional condensable linear plume processes (both chemical material (e.g., sulfuric acid from emit- and microphysical) have not been ted SO ). evaluated in depth in the context of 2 global chemical transport models. Research is lacking on how the prop- None of the currently available emis- erties (number concentration, surface sion inventories considers the effect of area, composition, and mixing state) of plume processing on species or parti- ambient aerosols are perturbed at the cle mixing ratios (e.g., NOx to NOy re- presence of jet engine emissions un- partitioning, volatile and soot aerosol der various conditions. In this regard, a number concentrations and size) that detailed investigation of the micro- eventually enter global simulations. physical (condensation and coagulation) and chemical processes (oxida- Two types of aerosols are known to tion of precursor gases) governing the exist in aircraft plumes: the first is as- evolution of aviation aerosols in the sociated with soot particle emission time scale of days to weeks after and has number emission index of 14 15 emission is required. Research is also ~10 -10 /kg-fuel; the second is due needed to define the abundance and to the formation of volatile particles in- properties of ambient aerosols as well duced by chemi-ions (e.g., Eichkorn et as gaseous aerosol precursor concen- al., 2002) and has number emission 16 17 trations in the troposphere. Both theo- index of ~10 -10 /kg-fuel. The avia- retical modeling and in-situ measure- tion-generated particles may perturb ments are needed to advance our the abundance and properties of cli- knowledge about the perturbation of mate-relevant particles in the upper the climate-relevant particles in the troposphere. To properly assess this upper troposphere by aviation emis- perturbation and associated climatic sions. effect, further research is needed to understand the properties, transforma- • Properties of heterogeneous ice nuclei tion, and fate of aircraft-generated par- from natural and anthropogenic ticles. sources. The formation of cirrus clouds is characterized by a competition be- Aviation aerosols are composed of wa- tween freezing particles for the avail- ter, sulfuric acid, organics, and soot. able water vapor. Because of this Particle composition may affect their competition, the ice-nucleating behav- potential to act as IN. In this regard, it ior of particles from aviation depends is necessary to characterize the de- on the ice nucleation properties of par- pendence of particle composition on ticles from other anthropogenic and engine operation conditions and fuel natural sources. The chemical compo- properties and the relative contribution sition of IN in the free troposphere is of organics versus sulfur to the mass important to understand both the de-

34 tails of their freezing behavior and their • Contrail-cirrus development. The de- sources. velopment of cirrus clouds from con- A special case is elemental or soot-like trails and the resulting radiative effects carbon. The available data indicate are poorly characterized in current cli- that their ice nucleating behavior de- mate models and have only been stud- pends on their source and processing ied on a limited basis from satellites in the atmosphere such as addition of and in detailed-cloud scale models. sulfate or organics. For example, ele- The role of wake dynamics in deter- mental carbon from biomass burning mining immediate contrail ice particle probably has different ice-nucleating concentrations has not been fully ex- properties than aviation soot. None of plored. For example, the interactions the laboratory studies of ice nucleation between the wakes of four versus two has used authentic aviation soot. Sul- engines could, perhaps, dissipate fates and organics have been shown young contrails through induced sub- to affect ice nucleation ability, but the sidence, even in nominally supersatu- role of organics that condense in the rated conditions. Otherwise, once a plume behind a jet engine has not contrail forms in supersaturated condi- been studied in cruise conditions. tions, it will continue growing and spreading. However, knowledge of the It is known from measurements carried impact of the type and numbers of out in wave clouds at temperatures primary and secondary emission parti- close to –38°C that certain aerosol cles on the number of ice crystals and particles nucleate ice at lower super- hence the potential for particle growth saturations once they have nucleated and precipitation is inadequate. Such ice before (Field et al., 2001). This pre- factors, along with wind shear, local conditioning effect is neither theoreti- vertical humidity profiles, and wake cally understood nor well explored ex- turbulence, will determine how the perimentally. Short-lived contrails that contrails grow vertically and horizon- form in subsaturated air could lead to tally and whether they dissipate in a pre-conditioning of exhaust soot parti- few minutes or in many hours. The re- cles, as it is known that contrail ice sulting vertical distribution of the parti- crystals mainly form on emitted soot cles and their sizes determine the con- particles. After sublimation, these trail-cirrus optical depth and effective modified soot particles could facilitate particle size that govern the radiative ice formation in the atmosphere, in- effects of the cloud. creasing the relative importance of the indirect effect. Conversely, if the soot A few modeling studies have exam- is not so transformed by this condition- ined the transformation of young con- ing, it may never be as effective as trails to cirrus clouds and its sensitivity ambient IN, even though these soot to the number of nucleating particles particles previously served as nuclei and wind shear, but the effects of real- for contrail particles. Whether or not istic emission particle distributions, in- soot particles are effective IN and to duced turbulence, radiation interac- what extent contrail processing tions, and the mesoscale environment changes their properties is a central have not yet been examined in a question. meaningful way. Such modeling studies are clearly needed but will remain

theoretical exercises until the relevant models. Current operational weather variables can be measured simultane- forecast models [with the exception of ously. While early field campaigns an upcoming release of the European were conducted to accomplish that Centre for Medium Range Weather goal, the amount of useful data is in- Forecasts (ECMWF) Integrated Fore- sufficient to confidently model and un- cast System] and climate models [with derstand the processes determining the exception of a recent test version the microphysical and optical proper- of the German climate model ECHAM] ties of contrails and contrail-cirrus in a are not capable of predicting super- wide range of atmospheric conditions. saturation in the UTLS region. Instead, Cloud-to-regional scale measurements saturation adjustment schemes are and modeling are necessary steps in employed that move any excess of building a dependable set of tools for water vapor above saturation obtained determining contrail-cirrus impacts on within one time step due to cooling or climate. They form the basis for mod- transport of ice water. The saturation eling the effects on a global scale. adjustment dates back to the early However, knowledge of the global dis- days of climate modeling in the late tribution of contrail-cirrus optical prop- 1960s. It is well justified for warm (liq- erties and coverage still remains un- uid phase) clouds, because supersatu- certain. To date, satellite-based esti- ration with respect to supercooled wa- mates of contrail particle sizes, optical ter in the atmosphere is very small (up depths, altitudes, and coverage are to a few percent at most). However, it confined to only a few regions, sea- is known that cirrus clouds require sons, or years. The most studied area tens of percent of ice supersaturation is Western Europe followed by the to form. Accurate knowledge of ice su- United States. A more comprehensive persaturation is crucial for quantifying climatology of aircraft-induced cirrus both, direct and indirect effects of avia- properties and radiative effects is tion on cirrus cloudiness. High- needed, at least, for those areas resolution regional models may be where air traffic is significant or will needed to do this well. become significant in the near future, Many cloud schemes in general circu- e.g., eastern and southern Asia. The lation models (GCMs) diagnose cloud climatology should include several fraction, mostly as a function of the years that differ in upper tropospheric grid mean relative humidity, which is humidity in order to determine the appropriate for short-lived low levels variability over the actual range of clouds. Stratiform cirrus clouds in par- conditions that occur over time scales ticular are known to be long-lived and greater than a decade. A database of can be transported over many grid this type will serve as the basis for un- boxes of a large-scale model during derstanding the direct impact of con- their lifetime. For cirrus, a prognostic trails and contrail-cirrus and for guiding description of cloud cover would be and validating global climate models more appropriate. The same holds for that include this new class of ice long-lived contrail-cirrus that add to clouds. high cloudiness. In many cases, only ice water mass is prognosed in global • Incorporation of effects of aviation- models. This hampers the introduction induced particles and cirrus into global

36 of physically-based links to the nuclea- components related to cirrus in global tion of ice crystals in cirrus clouds, an models is a highly interdisciplinary re- issue clearly required to distinguish search issue requiring a long-term between the many different types of commitment in well-coordinated scien- cirrus and to track the indirect effect of tific projects. This research has to be aircraft-produced aerosols. carried out in concert with laboratory, In most global models cirrus is treated in-situ, and remote sensing analyses as one class of clouds in terms of their that may provide guidance in develop- radiative properties, although it is well ing parameterization schemes of sub- known that anvil cirrus originating from grid-scale processes used in GCMs. deep convective outflow, stratiform cir- • Global distribution and properties of rus associated with synoptic weather supersaturation, aerosols, and thin cir- systems, or wave clouds triggered by rus. Even if the degree and frequency strong orographic forcing exhibit dis- of occurrence of supersaturation and tinct size distributions and geometrical the composition and size distributions properties that eventually control their of aerosol and cirrus cloud particles radiative forcing. Contrail-cirrus form a were calculated by GCMs, it would be class of high ice clouds on their own, a difficult task to actually validate as they are composed of more numer- them. The problems with detecting and ous small (diameter ~10-30 !m) ice verifying ice supersaturation meas- crystals than natural cirrus and likely urements from satellites have been lack a large (diameter >100 !m) ice addressed before. There is no global crystal mode. Also, known feedbacks inventory available that would yield between radiation and internal cloud quantitative information about the motion remain largely unconsidered in aerosol budget in the UTLS region, in GCMs, because such processes can- particular for soot-containing aerosol. not easily be parameterized owing to Current global model aerosol valida- their small-scale nature. According to tion exercises (e.g., the AEROCOM cloud-scale numerical simulations, initiative) strongly focus on lower tro- such feedbacks can also occur in con- pospheric aerosols. Cloud climatolo- trail-cirrus. gies such as contained in the Interna- tional Satellite Cloud Climatology Pro- Before at least some of the above de- ject (ISCCP) data set do not include scribed first order problems have been high clouds with optical depths below solved, it will be very difficult to provide about 0.2, but such thin cirrus are reliable global real-time forecasts of common in regions where aircraft contrail-cirrus in support of potential cruise. Stratospheric aerosol and sub- contrail avoidance strategies, or pos- visual cloud climatologies [especially sible to accurately predict the impact those from the Stratospheric Aerosol of aviation-induced cloudiness in fu- and Gas Experiment (SAGE) and the ture climate change scenarios. While Halogen Occultation Experiment (HA- some progress has been made in the LOE)] have limited value in the tro- past 5 years to deal with ice super- popause region and upper tropo- saturation and to parameterize the nu- sphere. Aircraft measurements are cleation of ice crystals in cirrus clouds, useful in checking whether global harmonizing the complete suite of dy- model predictions capture the atmos- namical, microphysical, and radiative

pheric variability, but do not provide a tudes of the trends. The humidity issue global coverage. has been skirted, to a degree, by focusing on the relationships between • Long-term trends in contrail-cirrus and cirrus trends or amounts and upper cirrus. Based on current knowledge, it tropospheric air traffic. Such studies is expected that cirrus coverage will generally agree that cirrus coverage is increase as a result of contrail-cirrus greater in areas where air traffic oc- formation that taps hitherto cloud-free curs, but they do not answer the ques- supersaturated air. Whether contrails tion regarding the suppression of cir- cause additional cirrus coverage or rus in other areas. Solid answers to cause a decrease in cirrus in other lo- those questions appear to be depend- cations resulting a zero net change ent on understanding humidity variabil- can only be answered by studies of ity in both areas, with special empha- long-term variability. Long-term trends sis on ice supersaturated regions. can only be ascertained from consistent measurements over extensive pe- While it is not possible to return to the riods. Subjective, but consistently past and reconstruct a more accurate guided, surface observations of cloud UTH record, it is recommended that types and coverage have been taken improved methods for measuring UTH over much of the world since the be- and supersaturation be standardized ginning of the jet age and over many and applied consistently on a global locations in earlier years. Satellite ob- basis in the future. Development of in- servations, primarily in the form of data novative methods to unscramble the from the ISCCP, have been analyzed natural and anthropogenic effects since 1983. Each type of observation should be continued. Furthermore, has its uncertainties and either one both surface and satellite observations alone is insufficient to confidently de- should be sustained in order to detect termine the overall impact of aircraft cirrus changes as air traffic patterns on cirrus. evolve over the coming years. Recent studies show that for common periods of record, the surface and satellite data agree in the general direc- 3.3 Climate Impacts and Climate tions of the trends but not in magni- Metrics tude. Ensuring that the trends are due There are only limited analyses of the to air traffic requires some knowledge uncertainties associated with current of the concomitant trends in upper tro- evaluations of radiative forcing resulting pospheric humidity (UTH), a parame- from aircraft emissions. Sausen et al. ter that has not been measured either (2005) did not provide new uncertainty adequately or consistently for any estimates (beyond IPCC, 1999), as the length of time. While recent efforts to sample size of independent experiments separate the natural humidity effects was too small to do this on a sound ba- and anthropogenic impact have had sis. Subjective estimates indicate that limited success, confidence in the re- the smallest uncertainties are associ- sults remains tepid because of the un- ated with RF from CO2, while the largest certainties in the humidity record and uncertainties are from cloud effects be- the differences between surface and yond linear contrails. More specifically, satellite observations in the magni-

38 the following main uncertainties with sponding heating rates are horizon- relevance remain: tally and vertically inhomogeneously distributed, the regional pattern of • RF from O3 and CH4: Here the main climate change will be different for dif- uncertainty arises from the change in ferent aircraft effects, e.g., the climate abundance of the species as stated in response to aircraft-induced ozone Section 3.1. The ozone and methane increase and methane decrease have RFs from NOx emissions are oppo- a different structure and will not com- site in sign, so the extent to which pensate. The actual pattern needs to they offset each other is an important be explored. uncertainty. • Metrics for aviation induced climate • Impact from cirrus clouds: While the change: It is still not clear which cli- knowledge on linear contrails has in- mate metrics are most appropriate for creased since IPCC (1999), still at analyzing present-day and future avi- least a factor of 2 remains as uncer- ation activity on climate. There is also tainty. Like IPCC (1999), the Sausen a need to use metrics to compare et al. (2005) paper does not provide a climate impacts due to various trade- meaningful estimate of the radiative off options e.g., between CO forcing associated with aviation- 2 and and induced cirrus changes beyond linear NOx emission, or between CO2 contrail plus contrail cirrus. contrails. Sausen et al. rather report of an estimate by Stordal et al. (2005) based on studying the correlation between air traffic and observed cirrus 4. Use of Climate Metrics in cloud changes: Their "mean" esti- Trade-off Studies mate is 30 mW/m2 with very large un- There are a variety of potentially impor- certainty (an upper bound at 80 2 tant trade-offs for aviation and climate mW/m ). More recently, a yet unpub- change. These include trade-offs among lished study (Penner and Chen, aircraft and fuel technology, aviation op- 2006) suggests much larger effects erations, and policy options. The selec- from indirect cirrus. However, other tion of one option over another may po- studies suggest, that even the sign of tentially result in different magnitudes of this effect is uncertain (Hendricks et climate impacts. More broadly, there are al., 2005). also interdependency oriented trade-offs • Efficacy of aircraft-induced radiative with other environmental impacts such forcings: There are some indications as local air quality and community noise. that the efficacy of contrails (and po- These interdependencies are covered in tentially contrail cirrus) is smaller than section 4.3. 1, while the efficacy of aircraft in- The focus of this workshop was on duced ozone changes appears to be those science issues that affect climate larger than 1. These findings need to impact related trade-offs issues and a be confirmed by more modeling stud- need to assess these climate impacts in ies. terms of suitable metrics. For example, • Aircraft induced climate change: As it has been suggested that reducing aircraft-induced perturbations in the cruise altitude would reduce the number radiative active species and corre- of contrails (it is unclear what the effects

39 would be on cirrus) and reduce the NOx ment may take 10 years, followed by a effects on ozone, while affecting fuel 20 year production period, followed by burn and the amount of CO2 emissions. as long as 30 years in service. There- Flying higher, on the other hand, would fore, airplane technology under devel- reduce contrail production, but would opment today may still be flying 50 enhance residence time of pollutants, years from now. with resulting chemistry implications. In- creasing the engine bypass ratio could 4.1 Issues with Metrics for Assess- reduce CO2 emissions and produce less ing Aviation Climate Impacts noise, but result in increased NOx emis- Estimates of radiative forcing for avia- sions. Some studies have suggested tion, such as those produced in IPCC that reducing the amount of high latitude (1999) or Sausen et al. (2005), repre- routing might reduce effects on strato- sent the radiative forcing at a given time spheric ozone, but this would require due to all prior and current aviation ac- more fuel burn and increased CO 2 tivity (e.g., effects of accumulated CO2 emissions. Recently, because of the emissions, plus present day, short-lived competing effects of solar versus infra- impacts like contrails). Since some of red radiative effects from contrails, it has the effects have very different times- been suggested that reducing the num- scales, such radiative forcing estimates ber of night flights would generally be for either current or projected fleets of expected to reduce the climatic effects aircraft do not capture the relative im- of the contrails produced [e.g., Stuber et portance of short-lived and long-lived al., 2006]. However, to some extent, the effects. Further, as discussed earlier, benefit of such reduction in nighttime different sources for radiative forcing flights would be offset by an increment (e.g., contrails, CO2, NOx through in climate impacts due to enhanced day- ozone, and NOx through methane) pro- time aviation activities. duce different levels of temperature There are two primary motivations for change at the surface of the Earth per understanding such trade-offs. First, in unit change in radiative forcing. As men- order to select environmental policies tioned earlier, a limitation of considering that balance society’s economic and en- relative levels of radiative forcing is that vironmental needs, national and inter- the different impacts on temperature are governmental agencies must have the not fully accounted for, although the use ability to assess the complete impact of of efficacy factors may improve this each suggested option, accounting for evaluation. potential interdependencies. Second, it Thus, climate metrics that integrate over is important to provide guidance to past aviation activities, in of themselves, manufacturers and airlines as they seek may not be an appropriate basis for to balance a variety of environmental, making policy decisions, nor are they an safety and performance objectives. This appropriate basis for fully evaluating the is particularly important because of the relative impacts of various aviation ef- capital-intensive nature of the industry – fects. This argument is not new; indeed, roughly $10B for a new airplane devel- it is widely understood and accepted, opment effort -- and because of the long and a variety of alternative integral time-scales for development and use. measures have been pursued [see e.g., Technology development and deploy- Forster et al., 2006].

4.2 Key Challenges and Gaps in ferences in emissions, flight altitude and addressing Trade-offs among the latitude distribution of emissions, we Aviation Climate Impacts are unsure whether current models could represent the statistical signifi- Extensive discussion at the Workshop cance of the difference. The forcings highlighted the current uncertainties and from changes in O and CH are ex- difficulties in addressing trade-offs 3 4 pected to have opposite signs, cancel- among climate impacts like the exam- ing effects would broaden the uncer- ples mentioned above. Based on the tainty, and even the sign of the net im- information from previous assessments pact might be uncertain. Finally, note and from more recent published investi- that implications for the trade-offs de- gations, it was concluded that the scien- veloped based on the present fleet and tific community would not currently be atmospheric conditions may be poten- able to reach consensus in quantifying tially different for the future projections. the climate impacts, and the associated Uncertainty about the changes in the metrics, associated with the NO -O - x 3 future background atmosphere relative CH impacts or with the contrail and cir- 4 to today adds to the uncertainties de- rus impacts for the purpose of trade- scribed earlier in the Uncertainties and offs. Research Gaps section of this docu- Present models are far more capable ment. than those used in the 1999 IPCC as- In a more general sense, there are no sessment, and could be used with the published studies examining the possi- goal of developing a limited set of trade- ble choices, dependencies and prob- off metrics. Simulations with present lems for metrics suitable for application models would still be subject to uncer- to aviation, that adequately utilize our tainties as outlined above, but could be understanding of the impact of aviation used to develop metrics for aspects emissions on atmospheric composition, such as cruise altitude and routing. For as reflected in state-of-the-art atmos- a given fleet and distribution of emis- pheric models. Research is needed to sions, it is possible to quantify the re- examine the effect of different metrics, sponse for a change in a single aspect and the choices within each metric (e.g. of the fleet, such as a reduction of NO x time horizon), on evaluating the relative emissions. Significant effort would be importance of different aviation emis- required to produce and analyze the sions. Such studies would need to ex- necessary simulations. It is important plore the potential of existing metrics that the response to the aircraft emis- and the possibility of designing new sions and its uncertainty are presented metrics. in a manner that takes into account their interdependence. Uncertainty should not It was noted that in discussing trade-offs be interpreted as representing a normal involving contrail-cirrus effects that distribution. For example, the 1999 these effects will depend on latitude, IPCC assessment did not assume a with pronounced regional differences normal distribution for uncertainties. The expected. More robust estimates de- quoted uncertainty represented a 2/3 pend on progress in understanding su- likelihood interval. If present models and persaturation and cloud modeling, but tools were used to compare the effects also on scenarios of how future flight of two “future” fleets, with combined dif- routes will be distributed. We cannot tell

41 with confidence what these tradeoffs will ICAO estimated the cost of this increase look like in a future climate, and it will be in stringency to be approximately $5 bil- difficult to develop trade-off metrics for lion. Aircraft and engines designed to contrail-cirrus and cirrus changes based meet this standard are likely to have to on existing information. make compromises on fuel burn and Although the complexities and uncer- weight, and thus, on climate and noise tainties surrounding trade-off alternative impacts. Policymakers have also ex- can be perceived as barriers to policy- pressed an interest in comparing avia- making, it is important to bear in mind tion to other industries (e.g., other that aircraft manufacturers continue to transportation systems). The European bring forward new products and the pol- Union is considering options for includ- icy-making community must continue to ing aviation in emissions trading to miti- consider environmental policy decisions gate climate impacts. These attempts related to aviation. Climate change met- raised many policy questions. The atten- rics are expected to play an important tion of workshop participants was drawn role in this decision-making process. to these interdependencies and trade- offs. However, given the science focus Research is needed to examine different on the aviation-induced climate impacts, metrics, and the choices within each there was no further discussion on these metric (e.g. time horizon), on assessing issues. the relative climate impacts of different aviation emissions. It must be stressed that even if agreement on an acceptable metric design was reached, it may be 5. Research Needs and Priori- that the current atmospheric models are tized Recommendations unable to calculate the input parameters The workshop identified the need for fo- for these metrics with acceptable accu- cused research efforts in the United racy. Hence, studies would need to ex- States to address the uncertainties and amine the impact of both model and pa- gaps in our understanding of current rametric uncertainties on trade-off is- and projected impacts of aviation on sues. climate and to develop metrics to characterize these impacts. This could be 4.3 Extending the Trade-off Space done through coordination and/or ex- The trade-off concept can be expanded pansion of existing and planned climate if one attempts to evaluate interdepend- research programs or it may entail new encies among local air quality, noise activities. Such efforts should include and climate impacts due to provisions of strong and continuing interactions be- addressing the one relative to the other. tween the science community, aviation In 2004 the International Civil Aviation system operators and policy developers Organization adopted new certification to ensure that the most up-to-date sci- standards for aircraft engine NOx emis- ence is readily available to policy con- sions. The new standards represent a siderations and that scientists are aware 12 percent increase in stringency of the questions and challenges faced (tighter standards) to be introduced in by operators and policy developers. 2008 and are designed to mitigate the These efforts will also require leveraging local air quality impacts of aviation. of ongoing research programs of the climate system.

Substantial progress across the issues will require a long-running science program to be established. The Atmos- 5.1 Emissions in the UT/LS and pheric Effects of Aviation Project Resulting Chemistry Effects (AEAP), the NASA led research program during the 1990s, with its numer- On-going Research ous European collaborations along with Aircraft research cannot be decoupled its annual meetings, provides an excel- from on-going research in atmospheric lent model. The effect of low-cost, quick- chemistry and climate. New efforts that and-dirty approaches to improve scien- address aircraft issues will benefit from tific understanding is questionable. and be enhanced by on-going efforts in However, because there are potential climate research. needs for near term policy considerations, the new research efforts need to • Aircraft impact studies. consider short-term as well as long-term There is significant atmospheric re- research objectives. Each of the three search funded through EU Framework Subgroups at the Workshop addressed programs focusing on emissions from both the short-term and long-term re- aircraft and their impact on atmospheric search needs and developed a priori- composition and climate. Studies focus tized list of recommendations that on impact in the UT/LS. These studies should be considered for an aviation- include the chemical compounds O3, focused research program. The short- NOx, OH, CH4 and water vapor. Recent term research priorities assumed that an EU funded research projects have fo- interim assessment of the effects of cused on current and future emissions aviation on climate was needed within from aircraft and the impact on atmos- the next three (to five) years to meet pheric composition and climate. The NGATS objectives. TRADEOFF [Sausen et al, 2005] and In establishing the objectives of the fo- CRYOPLANE [Gauss et al., 2004] pro- cused research efforts, it will be impor- jects included studies of tradeoff options tant to consider existing, ongoing re- like change in routings (TRADEOFF) search where value-added contributions and the use of hydrogen as fuel can be made by additional consideration (CRYOPLANE). The SCENIC project of the issues associated with the emis- investigated impact on the atmospheric sions from aviation. To the degree pos- composition from introducing a fleet of sible, the Subgroups also addressed the supersonic aircraft. Current and future ongoing research that should be con- aircraft emissions have been updated in sidered in meeting program objectives. connection with the projects. These It is also recognized that the short-term studies are followed by ongoing re- and long-term research efforts dis- search in the Integrated Project (IP) cussed below will require new or addi- QUANTIFY (2005 – 2010, http://ip- tional research funding. quantify.eu). The goals of QUANTIFY have been expanded over prior pro- In the discussion below, the highest pri- grams to include the impact on atmos- ority items for short-term or long-term pheric composition and climate from the research are specifically noted. All other transport sector (aircraft, ship, land- items should be regarded as being of based transport). Studies include im- medium to high priority.

43 provements in model performance launched to address important ques- (transport processes, plume parameter- tions regarding global air quality, climate izations, chemistry), and current and fu- and the stratospheric ozone layer (Aura, ture contributions from the different SCISAT-1 ACE, ENVISAT etc.). Satel- transport sectors to atmospheric com- lite observations and data obtained position changes and climate impact through their validation programs (e.g. (through metrics). Further aircraft impact AVE) have contributed to a substantial studies are being performed through the database of tropospheric, UT/LS, and EU-led European Compatible Air- stratospheric measurements. Instru- Transport System (ECATS, ments such as MLS on Aura measure http://www.pa.op.dlr.de/ecats/) Network CO and O3 in this region. ACE meas- of Excellence, where the focus is on ures 10-20 species down to 5 km. In technological aspects and the relation to addition, measurements of important environmental impact. One of the objec- species in the free troposphere, such tives is to improve modeling capability to as, for example, CO from MOPITT reduce uncertainties and develop com- (EOS-Terra), MLS and TES (both on mon modeling tools for large-scale stud- EOS-Aura) can be used to address is- ies such as climate impact. sues of tropospheric transport. • Climate Research. NSF UT/LS-HIAPER. The NSF commu- SCOUT-O3. This international program nity is making a concerted effort to in- contributes to the understanding of vestigate the UT/LS, using the new high processes in the UT/LS, where the fo- altitude research aircraft, Gulfstream V, cus is on ozone changes in the LS re- known as High-performance Instru- gion, including stratosphere troposphere mented Airborne Platform for Environ- exchange, convective transport to the mental Research (HIAPER). Field cam- UT, chemical processes in the UT/LS paigns planned for the next 5-10 years and climate-chemistry interactions. have specific foci for UT/LS chemistry, dynamics, and microphysics, including U.S. agency research programs. Sev- the mixing of chemical constituents in eral U.S. agencies (e.g., NASA, NOAA, the tropopause region, the role of con- DOE, NSF, and EPA) sponsor individual vective transport in coupling the bound- investigator research related to climate ary layer and the UT/LS, connections of concerns. For example, DOE has the chemistry and microphysics, lightning Atmospheric Radiation Measurement production of NOx. The data obtained in (ARM) program and NASA sponsors this effort will be interpreted in conjunc- climate research within a number of its tion with satellite observations using programs. NASA also sponsors the de- multi-scale models. The set of questions velopment of global three-dimensional and tools in development for these mis- models for studying the chemistry and sions are all highly relevant to aviation physics of the troposphere and strato- specific research. The NASA DC-8, WB- sphere though the Global Modeling Ini- 57, and the NOAA P-3 also have key tiative (GMI). roles in understanding atmospheric Satellite Measurement Programs includ- chemistry issues relevant to the climate ing validation campaigns (Aura, AVE, impacts of aviation. ACE, ENVISAT, etc.) In the last 5 to 6 SPARC CCMval. The SPARC commu- years many satellites have been nity has developed a strategy for proc-

44 ess-oriented validation of coupled chem- model predictions. Studies of the istry-climate models (CCMs) termed UT/LS also need to investigate the ef- CCM-Val (Chemical Climate Model fects of higher resolution on represen- Evaluation) tation of constituents and relevant [http://www.atmosp.physics.utoronto.ca/ processes in that region. SPARC; Eyring et al., 2005/]. The focus • Use of state-of-the-art models to pro- has been to develop analysis vide parametric relationships between tools/methods in the areas of radiation, aviation emissions and environmental dynamics, transport, and chemistry and impact. Since the aircraft fly in the microphysics better understanding UT/LS, models with detailed represen- model processes. The performance cri- tation of both troposphere and strato- teria and diagnostics developed within sphere are needed. Such models exist this effort will also be applicable to the (e.g., GMI, MOZART), and continue to chemical transport models (CTMs) and be tested with different measurements. serve to assess how well the new gen- A parameter space for aircraft emis- eration of models can be used to repre- sions needs to be defined, with appro- sent aviation related processes. priate sampling to derive meaningful Short-Term Research Needs approximations. This activity needs to be coordinated with the aircraft engi- The activities below are deemed feasi- neering community. This activity would ble in the next three years. As pointed be most useful if conducted following out below, some of these efforts will ex- the Models and Measurements Inter- pand ongoing research in the US and comparison. other countries, or will involve collaboration with such research. • Vertical transport processes between 2 and 10 km. Vertical transport proc- • [High Priority] Models and Measure- esses bring NOx and other ozone pre- ments Intercomparison. A Models and cursors to flight altitudes (e.g., convec- Measurements Intercomparison, em- tion), and transport aircraft exhaust phasizing the UT/LS and free tropo- downwards (e.g., downdrafts and sphere, could be conducted following large-scale downwelling). Lightning is similar efforts previously carried out for also associated with deep convective the stratosphere (e.g., Park et al., regions. Various representations of 1999), as well as ongoing activities in these processes are used in current Europe (ACCENT, SCOUT-O3). It models, and all are highly uncertain. would require: a) detailed definition of Activities in this area should profit from modeling experiments testing different collaboration with ongoing model in- model aspects (transport, chemistry), tercomparison and evaluation by the and diagnosing model differences European ACCENT program. through process oriented model intercomparison; b) compilation of a data- • Data analysis and modeling. A wealth base of measurements from different of data has been obtained in the low- platforms, to be used in evaluating ermost stratosphere by different air- model performance; c) identification of craft and satellite platforms. A wealth model deficiencies through the above of data has been obtained from past process, leading to model improve- measurement campaigns such as the ments and reduction of uncertainty in NASA supported SONEX measure-

ments that have not been fully ana- from aircraft effluents. Such a cam- lyzed for the possible link to aviation paign can also include instruments to activities in the region of the observa- examine microphysical processes for tions. Also, the NASA A-train constel- contrails and cirrus. lation of satellites, and the European • Background NOx and its control via satellites ENVISAT and SCISAT-1 are lightning and convection. Because of providing a wealth of data in this re- non-linearities in the atmosphere, the gion. Different data analysis and mod- ozone response from the same eling groups are currently using these amount of NOx deposited in the upper data to constrain the magnitude and troposphere by aviation can vary de- seasonality of turnover rates and mix- pending on the background concentra- ing processes in the lowermost strato- tion of NO from non-aviation sources. sphere. Results from these studies will x It is important to derive the back- further reduce uncertainties in the dis- ground concentrations in the upper persion of aircraft exhaust in this re- troposphere from observations, and to gion. understand the relative importance of • Model analyses of chemical and cli- the lightning source and transport of mate impacts. Re-examine the im- NOx emissions from the boundary pacts of aviation in the UT/LS using layer in understanding these back- several of the improved models. ground concentrations. A similar argument could be made for HOx pre- Long-term Research Needs cursors. • [High Priority] Field Campaign(s) to • Relationship between surface reaction address issues with HOx-NOx chemis- efficiencies of heterogeneous chemistry in the UT. As discussed in the “Un- try and upper tropospheric aerosol certainty” section, there are significant composition. Heterogeneous reactions discrepancies between measured HO2 on aerosol surfaces transform NOx concentrations and constrained box species to HNO3 that is removed effi- model calculated values in the upper ciently by washout processes. Thus, troposphere. If the measured HO2 val- the reaction efficiency plays an impor- ues are accurate, this indicates either tant role in determine the residence missing mechanisms in the photo- time of NOx in the troposphere. The ef- chemical scheme, or wrong kinetic pa- ficiency depends on the chemical rameters. Redundant measurements composition of the aerosol. It is impor- of OH, HO2, NO, NO2 species are tant to make in situ measurements of needed to resolve this issue. This is aerosol composition and use meas- classified as a long-term goal because ured concentrations of trace species to it requires development and adapta- derive reaction efficiency factors. tion of instruments to operate on the same aircraft platform. To be useful, • Status of low-temperature kinetic the campaign must cover a range of rates. Many reaction rate constants environments – pristine, polluted, are measured at higher temperatures moderately polluted, to test non- and the measured results are extrapo- linearities in the system. It would be lated to lower temperature appropriate important also to sample areas known for the upper troposphere. Unfortu- to have elevated NOx concentrations nately, it is not immediately obvious

which reaction(s) is/are responsible for microphysical calculations carried out discrepancies between modeled and along representative dilution trajecto- measured concentrations. Continued ries. Both types of studies are pro- efforts are needed to identify those re- posed for longer time periods and fur- actions followed by direct measure- ther contrail evolution in the future. ments of the rate for the appropriate • APEX ground-based particle emission temperature range. studies (2004-2006). A series of field • Access to high-performance comput- studies have been carried out to ing. Computational resources continue measure particle precursors and parti- to be a limiting factor in performing cles, both volatile and non-volatile, us- simulations with the ever finer model ing commercial high bypass ratio tur- resolution required to represent at- bine engines on ground-based air- mospheric processes. There is a need planes. This series of Aircraft Particle to continuous upgrading of the com- Emissions Experiments (APEX, puter hardware to take full advantage JETS/APEX2, APEX3, UNA-UNA), of new developments for maximum sponsored by a number of U.S. agen- throughput. This strategy also turns cies and with broad measurement out to be cost effective, as obsolete team representation, has generated equipment inevitably requires higher results both from dedicated engines maintenance cost. tests and from advected plume studies during routine operations at airports.

While the tests were carried out at 5.2 Contrails and Cirrus ground level, intermediate engine powers were explored during dedi- Asterisks indicate areas of research that cated engine tests, so that cruise op- are being addressed to some degree in erations might be estimated in the fu- current programs. ture after altitude chamber measurements are made to develop algorithms On-going Research Projects (and for extrapolating static sea level Some Recently Proposed but not yet measurements to cruise conditions. Supported) Projects marked with the # symbol are • NASA LaRC CMAI (2006-2008). A planned or proposed and not yet study, sponsored by the NASA Cloud funded. Modeling and Analysis Initiative, is un- derway at NASA Langley Research • PARTNER contrail studies (2005- Center (LaRC) to examine the poten- 2007). This project being undertaken tial for modeling persistent contrails in PARTNER aims to find relationships using high temporal and spatial resolu- between aircraft parameters and the tion numerical weather analyses properties of contrails generated by (NWA). Contrail formation, properties, the aircraft under a variety of atmos- and spreading parameterizations used pheric conditions. Large-Eddy Simula- in the model will be determined em- tion studies are being conducted, ini- pirically based on satellite observa- tially with simple ice-microphysics and tions of contrails in the absence of focusing on contrail evolution for the natural cirrus in conjunction with the first few minutes. These are comple- atmospheric conditions in the NWA. mented by parallel, detailed particle Aircraft will be flown through the model

to form contrails; the results will be ticular the effects of emissions of used to compute radiative impacts and ozone precursors and particles, as will be validated using satellite data. well as of contrails and ship tracks. The project goal includes the provision • DLR-IPA PAZI (2004-2007). The pro- of forecasts and other policy-relevant ject conducted at the Institute of At- advice, which will be supplied to gov- mospheric Physics (IPA) and spon- ernments and to international assess- sored by DLR concentrates on the ef- ments of climate change and ozone fects of particles from aviation and depletion, such as the IPCC climate other anthropogenic and natural assessments and the WMO-UNEP sources on aerosols, cirrus clouds, ozone assessments. Using signifi- and climate. Besides IPA, other Re- cantly improved transport emission in- search Centers from the Helmholtz- ventories, better evaluated and hence Society in Jülich, Karlsruhe, and more reliable models, these new fore- Bremerhaven, the institutes of Meteor- casts in QUANTIFY will represent a ology and Chemistry from the Max- considerable improvement of current Planck-Society and various universi- predictions. QUANTIFY work pack- ties, and partners from France, Swe- ages include focused field measure- den, Switzerland, England, and Can- ments, exploitation of existing data, a ada all contribute to this project. Its range of numerical models, and new first funding phase ended in 2003, the policy-relevant metrics of climate results of which have been summa- change. rized in the Proceedings of the Euro- (http://www.pa.op.dlr.de/quantify/) pean Conference on Aviation, Atmos- phere and Climate (AAC). Much of the • NASA TC-4 / Costa Rica (2007-8)#. work carried out in PAZIs second fund- The Tropical Composition, Cloud, and ing phase is centered on the direct and Climate Coupling (TC-4) mission is a indirect effects of aircraft emissions on proposed NASA Earth Science Enter- high cloudiness and their associated prise (ESE) investigation. The primary climate impact. The project homepage goals of TC-4 are to gain a better un- provides more detailed information (in- derstanding of chemical, dynamical, cluding all references that have and physical processes occurring in emerged from PAZI-related studies), the tropical upper troposphere and in and also the first phase summary re- the layer surrounding the tropical tro- port can be downloaded from this site popause as well as the roles that the (http://www.pa.op.dlr.de/pazi/). anvils of deep convective clouds and tropical cirrus play in humidifying the • EU Integrated Project QUANTIFY (2005-10). The main goal of QUAN- upper troposphere and lower strato- TIFY is to quantify the climate impact sphere. These issues are relevant to of global and European transport sys- studies involving global climate tems for the present situation and for change, stratospheric ozone depletion, several scenarios of future develop- global tropospheric chemistry, and the ment. The climate impact of various Earth's radiation balance. While not transport modes (land surface, ship- connected to aviation research, results ping, aviation) will be assessed, in- from TC-4 could help improve under- cluding those of long-lived greenhouse standing and detection of supersaturation in cirrus levels. gases like CO2 and N2O, and in par-

• NASA MODIS (2006-)#. The MODIS search aircraft would lay out a contrail. cloud products contain rich information It would soon thereafter make a set of about the properties of ice clouds. The measurements in the contrail with ice level-2 MODIS cloud properties particle probes to determine the initial (MOD06_L2 for Terra satellite and state (number concentrations and MYD06_L2 for Aqua satellite) are re- sizes). Thereafter, the aircraft would trieved on a granule scale, which con- climb above the cloud. With a cloud sist of cloud top physical parameters Doppler radar, lidar, and radiometers, (temperature, pressure, effective it would map out the evolution of parti- emissivity), cloud phase, cloud optical cles in the vertical using a combination parameters (optical depth, effective of radar/lidar (to obtain mean effective particle radius, water path), cirrus re- radius and ice water content), li- flectance in the visible band, contrail dar/radiometers to retrieve mean ef- flag. Of the MODIS ice cloud products, fective size through the depth of the the cirrus reflectance derived from the contrail, and Doppler fall velocity which visible and 1.375-!m bands is quite can be linked to the mean sedimenta- unique as it is available even for very tion velocity of the population. Occa- thin cirrus clouds. The MODIS level-2 sional ascents and descents through products are further integrated into the the contrail can be used to provide MODIS level-3 products that have a evaluation/validation of the retrieved one-degree resolution in both latitude properties, to determine whether addi- and longitude. The MODIS level-3 tional ice nucleation is occurring in cloud products can be directly applied contrail-cirrus, and to measure vertical to climate simulations. motions within the cloud column. • NASA CALIPSO (2006-). One of the A similar measurement strategy would science objectives of the NASA CA- be applicable for airborne observations LIPSO satellite mission is to provide of aging exhaust plumes in the ab- comprehensive measurements of sence of contrails, although it will be cloud vertical structure. CALIPSO much more difficult to track individual products will include: cloud height (for plumes for an extended period of time. layers with optical depths" > 0.01), Such measurements additionally re- cloud thickness (for layers with ! < 5) , quire the detection of relative humidity, extinction profile, ice/water phase, ice SO2, NOx and NOy (which includes cloud emissivity, and ice particle size. the longer lived reservoirs gases like ! HNO3, to study chemical conversion of Short-Term Research Needs nitrogen oxides), CO2 (to infer plume • [High Priority] In-situ probing and re- age), and organics in condensed mote sensing of aging contrail-cirrus phase. Additional instruments include and aircraft plumes. A combination of an aerosol mass spectrometer, aero- in-situ and remote sensing measure- sol probes to quantify interstitial aero- ments can be used to characterize the sol properties, and condensation nu- growth, decay and trajectories of con- clei and ice nuclei counter. To this trail ice particle populations in the mid- end, currently available aerosol mass latitude upper troposphere. In pristine spectrometers need enhanced sensi- humid air, devoid of ice particles and tivity to measure small concentrations outside of aircraft flight corridors, a re- and novel IN counter that are under

development need to operate in-flight persaturation. Development of these in cirrus conditions. The deployment of models and associated observational instruments detecting single black car- datasets may be the best approach for bon particles is highly desirable. Such developing our knowledge in the near instruments are capable of quantifying term before attempting to use global number size distributions as well as models. the mixing state of soot particles (i.e., • [High Priority] Global model studies the degree of coating with soluble mat- addressing direct and indirect effects*. ter). Along with accurate supersatura- In global climate models (GCMs) that tion measurements, the in-situ detec- enable the prediction of number and tion of single soot particles might pro- mass of ice crystals in cirrus clouds, it vide a clue for unravelling the indirect is possible to study the effect of soot effect. aerosols on cirrus formation. These To increase our understanding of con- models need some predictive capabil- trail-cirrus development, a set of coor- ity for soot particle abundance, both dinated regional-scale campaigns from aircraft and other sources, along should be designed and executed to with ice nucleation parameterizations measure the appropriate variables that include realistic background cirrus both with in-situ and passive and ac- formation scenarios. By their very na- tive remote sensing instruments. The ture, GCM studies of the indirect effect measurements should cover the cloud- remain parametric until the freezing free air prior to contrail initiation properties of aircraft soot and associ- through the entire life of the contrail- ated plume processing effects are bet- cirrus until it becomes part of larger ter characterized. natural system or dissipates. In the lat- Provided appropriate parameteriza- ter case, the cleared air should be tions can be developed that are able to sampled to determine the particulates track contrail-cirrus as a distinct class that remain after the episode is com- of cirrus clouds consistent with the plete. Because contrail-cirrus persis- physics already employed in GCM tence is highly variable, lasting from a cloud modules, the direct effect of per- half hour to more than 14 hours, a sistent, spreading contrails on cloud large number of events should be cover and radiative forcing can be sampled to gain the needed informa- studied even in the framework of con- tion. Such regional-scale experiments ventional climate models. Even without should ideally be combined with high- explicit calculation of supersaturation, resolution space-borne remote sens- such studies will be more realistic than ing, for instance, to address the ques- existing GCM estimates and lead to tion of fractional coverage, with obvi- improved prediction of the hitherto ous benefits for global model valida- poorly quantified global contrail- tion. climate impact. [High Priority] Initiate regional model- • Both types of studies will yield sepa- ing studies of supersaturation and con- rate estimates for the direct effect of trails using weather forecast models contrail-cirrus alone and the indirect that have the potential to include our effect of soot emissions on cirrus best physics and the high resolution properties alone. Interpretation of the needed to more accurately predict su-

resulting global radiative forcing is Aerosol Lidar with Orthogonal Polari- complicated by the fact that the refer- zation (CALIOP) on CALIPSO ence climate changes due to the avia- launched on April 28, 2006 provides tion impact (i.e., the cirrus cloud prop- new capabilities to study the vertical erties change). A combined treatment structures of contrails and contrail- of direct and indirect effects is not fea- cirrus clouds. The polarization capabil- sible in the short term. ity of the CALIOP at 532 nm will allow fundamental advances in our under- • [High Priority] Use of existing or up- standing of the particle morphologies coming information from space-borne of ice crystals in contrails and cirrus. sensors. A high priority will be the investigation of the optical and micro- The research goal of this priority in the physical properties of contrail and con- next 3-5 years will be the use of syn- trail-cirrus, e.g., the optical thickness ergetic data from the existing passive and effective particle sizes (two impor- and active sensors to systematically tant parameters that are essential to analyze the global distributions and the study of the radiative forcing of characteristics of contrail/cirrus clouds. these clouds) from space-borne sen- This effort will include the development sors. The satellite-based radiometric of robust new algorithms to use the measurements facilitate a unique ap- synergetic data sets to detect contrails proach to understand the characteris- and contrail-cirrus clouds and to re- tics of contrails and contrail-cirrus on a trieve their properties. global scale. Especially, the passive • [High Priority] Evaluation of capability [e.g., the Moderate Resolution Imaging for supersaturation from satellite ob- Spectroradiometer (MODIS)] and ac- servations. As an initial high priority, tive [e.g., Cloud-Aerosol Lidar and In- short-term study, a careful analysis frared Pathfinder Satellite Observation and comparison of all available cor- (CALIPSO)] sensors aboard the A- relative data on supersaturation at air- train platforms provide unprecedented craft flight levels are needed. Also on a opportunity to study aviation-induced short-term basis, systematic studies and modified high clouds. should be undertaken that compare The MODIS sensors on both Terra locations of observed contrail-cirrus as and Aqua platforms have a strong wa- detected from satellite or from ground ter vapor absorption channel centered based observations with regions of at 1.375 !m. This channel is quite ef- supersaturation detected from satellite. fective for detecting contrails and cir- Such comparisons would be a reason- rus clouds because the reflection of able first step for assessing the viabil- sunlight by the surface and lower at- ity of current satellite measurements to mosphere including low-level water accurately detect upper tropospheric clouds is absorbed by water vapor be- supersaturation. On a longer-term ba- low the clouds. Under contrail and sis, studies to improve retrievals for contrail-cirrus cloudy conditions, the existing instruments and develop new radiometric measurements acquired instruments for measuring supersatu- by the AIRS provide the hyperspectral ration at aircraft flight levels are signatures of these clouds, which can needed. be used to effectively infer the properties of these clouds. The Cloud-

• Correlations between supersaturation saturation with respect to ice is likely. fields and contrail-cirrus. Satellites Quantifying the frequencies of occur- provide global observations of the at- rence of contrail-cirrus with respect to mosphere and clouds. Enhancing supersaturated air parcels provides techniques for automated detection of critical guidance for parameterizing supersaturated air parcels and persis- contrail processes at the sub-grid tent contrails from space-borne pas- scales, both spatial and temporal, re- sive sensors will lend insight to the quired by a GCM. global nature of their presence and in- With the advent of active space-based fluence on radiative forcing. Many platforms it will be possible to supple- techniques rely on reflected-solar radi- ment passive-radiance contrail-cirrus ances, which has a limiting influence retrievals with active radar and lidar on the quality of the contrail detection observations, in part to compare and and analysis because: (i) radiance in part to further constrain the retriev- calibration is critical; (ii) the contrails als, albeit for only a portion of the total and contrail-cirrus often exhibit self- satellite swath. Active lidar/radar and shadowing attributes that confound the passive visible/infrared radiometer ob- microphysical and radiative-property servations from aircraft augment the retrievals; (iii) these retrievals are lim- spaceborne set, and are vital for as- ited to daytime only. Retrieval tech- sessing the atmospheric and contrail niques are emerging that exploit the state in detail. Aircraft radiance meas- influence of ice-particle shape and size urements from selected spectral bands on passive multispectral infrared sig- can be used to retrieve contrail radia- natures. It is important to place in- tive and microphysical properties that creased emphasis on coupled visible- in turn drive radiative flux mod- infrared retrieval techniques that els. These "retrieved" shortwave and gracefully degrade to infrared-only re- longwave fluxes can then be com- trievals at night. pared with coincident in-situ Developing a truly global day-night broadband flux measurements to en- contrail detection and analysis capabil- sure that the detailed ice-particle re- ity affords the construction of a con- trievals and associated microphysical trail-cirrus climatology that provides models are radiatively consistent with not only frequency of occurrence but observation. also trends in the diurnal cycle and • Process studies of plume and contrail statistics on radiative forcing changes. development*. A complete relationship Space-based detection of supersatu- has not yet been established between rated air parcels will enhance this cli- aircraft technology, neither engine matology. Detecting supersaturated air characteristics nor airframe configura- parcels from infrared and microwave tion, and the resulting aerosol proper- sounders is limited to some degree by ties of aviation-induced cloud particles. the vertical and horizontal spatial reso- Process studies that explore the role lutions of both the parcels and the sat- of emitted non-volatile (soot) particles, ellite observations themselves. How- nucleated volatile particles, and how ever it is possible even in limiting con- these aerosols interact with each other ditions to identify atmospheric water and with background ambient aerosols vapor conditions under which super- is required to understand how contrails

evolve into cirrus-like clouds and how approaches and methods of IN meas- aviation particle emissions perturb the urements also need to be compared. background aerosol in the absence of contrail formation. The role of aircraft Long-Term Research Needs configuration (engine placement, wing • The development or improvement of loading, etc.) may also have impact on instruments that help establish back- the nature and properties of the result- ground concentrations and character- ing ice clouds that mediate the radia- istics of heterogeneous ice nuclei and tive impact that aviation imposes. Re- measure supersaturation accurately, in liable estimates of the amounts and particular at low water vapor volume properties of aerosols and ice clouds mixing ratios (at the 10-50 ppm level at generated by aviation requires making tropopause temperatures and pres- a causal connection between the sures). This includes verification of su- emissions left behind the airplane and persaturation measurements from sat- those deposited in the atmosphere at ellites. More accurate temperature the global grid scale. measurements are also needed. • Laboratory measurements of ice nu- • The development and implementation cleation. Assessing the impact of con- of new concepts for ice phase-related trail-cirrus interaction on climate microphysics, supersaturation, radia- change requires, in addition to field tion, and cloud fraction, enabling a and modeling studies, careful labora- consistent treatment of aviation effects tory measurements of respective in the framework of climate and opera- aerosol-induced processes. Needed tional weather forecast models. This are experiments that are sensitive not includes the creation and validation of only to the onset threshold of ice nu- a number of parameterization cleation but also to the number frac- schemes, as the crucial processes to tion of aerosol particles acting as het- be treated will remain unresolved, both erogeneous IN as function of tempera- spatially and temporally, by global at- ture and ice supersaturation. Such mospheric models in the foreseeable data is urgently needed to develop future. aerosol-related parameterisations of

heterogeneous ice nucleation for use in models. Future experiments should focus on the ice nucleation efficiency 5.3 Climate Impacts and Climate of representative aircraft-emitted Metrics (soot) and natural (e.g., dust) aerosols. Important are systematic varia- On-going Research tions of soot parameters like the or- Limited efforts are continuing to address ganic carbon content, which are al- the adequacy of the radiative forcing ready known from previous studies to concept as a metric in a general sense influence the ice nucleation efficiency for climate change and the use of effi- of soot. Another important issue is the cacy within this concept, but these ef- impact of particle aging and mixing forts are not focused on aviation. due to coagulation growth and con- Short-Term Research Needs densation of less volatile substances like sulfuric acid or organics. Different

Cloud changes beyond linear contrails. well as their estimates of radiative One of the major uncertain forcings is forcing. The radiative forcing and re- that from cloud changes. Changes may sponse of the models need to be com- be caused by spreading contrails (and pared to observations as much as so they are initiated by the formation of possible. contrails). Changes are also possible in • [High Priority] Systematic model in- the ice number concentration associated tercomparison of efficacy studies. In- with changes to the ice number concen- homogeneous vertically and horizon- tration from ice nuclei associated with tally distributed forcing agents. Cirrus aircraft emissions. The following sub- changes, ozone changes, CH4, soot. projects are recommended. Effect of changes in climate state. [High Priority] Radiative forcing from • More research is needed to examine all cirrus and contrails. In addition to the effect of different metrics on evalu- previously mentioned studies, specific ating the relative importance of avia- studies are needed that are aimed at tion effects from different emissions. better understanding the radiative forc- Different metrics can change the rela- ing from contrails and cirrus. Inter- tive importance of the different effects comparisons (model to model) and from aviation emissions. A metric is evaluations (compare model to obser- needed that faithfully captures the cli- vations) of radiative transfer models mate impact of different components are also needed. Also, studies are of aviation emissions. needed to analyze the dependence on radiative transfer scheme and on rep- • [High Priority] Explore different exist- resentation of size and shape of ice ing and possible new metrics (both in- clouds. tegrative metrics and endpoint metrics). For example, RF, global average Develop cloud scale models with • surface temperature change, regional enough physics to predict macroscopic temperature changes, precipitation cloud changes. changes, extremes. • Chemistry-climate interaction relevant • [High Priority] Quantify the reliability to aviation (future climate). Chemistry- in determining various metrics and climate interactions from aircraft are how uncertainties propagate (both pa- currently evaluated as being relatively rametric input uncertainties and model small. However, this may not remain uncertainties). true in future climates. Further model studies of this aspect are recom- Long-Term Research Needs mended. • Experiments designed to not form con- More research is needed to determine trails -- attempt to detect the signature the efficacy of aviation related climate of aerosol emissions along the aircraft forcings. Current studies with limited track in formation of cirrus. number of models have shown that Evaluate the change of water vapor while climate model sensitivity to forc- • within the UT/LS associated with air- ing varies substantially, that of efficacy craft. Although this forcing is relatively (the relative change in temperature to small at present, it may become larger that caused by CO2) is not as variable. in the future. This issue also involves More models need to be examined as

the turnover time from the lower offs. Trade-offs need to be evaluated stratosphere to the troposphere and is using the various metrics established in connected with issues affecting the ef- research discussed above. Some of the fects of NOx. A coordinated program possible trade-offs to consider for addi- of model studies in comparison with tional evaluation include: measurements that includes this as- • Engine NOx reduction technology ver- pect is recommended. sus fuel efficiency (i.e., CO2 emis- • Trend studies of cirrus clouds: what sions) fraction is due to aviation? • noise abatement versus fuel efficiency GCMs have not been developed with • Changing cruise flight altitudes (which any capability to predict supersatura- is also affected by aircraft design) tion. In addition, cloud fractions within these models are generally diagnostic, • Changing future geographical distribu- not prognostic. Further model devel- tions of fleet (e.g., more flights to Asia) opment is needed to correct these in- • Flight re-routings (daytime vs. night- adequacies for treating the effects of time flights; avoiding regions of super- aircraft on clouds. saturation; avoiding certain regions, • [High Priority] Development of fore- e.g. polar routing, with specialized casting methods for supersaturation chemistry or within or outside existing (possibly based on commercial aircraft cirrus) measurements) • Calculations of the long-term O3 and • [High Priority] Development of meth- CH4 changes due to aviation. These ods to prognostically calculate cloud calculations are needed for metrics fraction within GCMs. studies, and are related to studies recommended above.

• Studies of the co-dependence of 5.4 Research for Trade-off Studies physical impacts (e.g., ozone and Trade-off studies need to be placed in methane effects are inter-related, not context of the research priorities above. independent) and resulting metrics All of these studies will have a high pri- and methods used in trade-offs con- ority for policy considerations; however, siderations. research first needs to focus on of the Climate change metrics are expected to identification, development and evalua- play an important role in these analyses. tion of metrics that can relate various There is currently no study in the peer climate impacts due to aviation. reviewed literature that can be cited to Besides the scientific research interest justify, based on the scientific under- in analyzing the net effects of potential standing of the impact of aviation emis- trade-offs, there is also a policy need to sions, the possible choices of metrics understand the benefits of emission re- suitable for trade-off application. Re- ductions due to various interventions as search is needed to examine the effect a result of air technology, operations of different metrics, and the choices and policy options. Model intercompari- within each metric (e.g. time horizon), son studies can be used to establish on evaluating the relative importance of how well we can evaluate various trade- different aviation emissions. Such stud-

55 ies would need to explore the potential with "business as usual" research in- of existing metrics and the possibility of vestment. designing new metrics. It must be • Identifying investment required to stressed that even if there is a philoso- achieve meaningful progress toward phical agreement on an acceptable met- addressing key questions based on ric, current atmospheric models may not specific timelines (e.g., 3, 5 and 10 be able to calculate these metrics with years). acceptable accuracy. • Creating a roadmap with identified

roles and responsibilities of various participating agencies and 6. Next steps stakeholders. Following this Workshop, there are a range of possible next steps that could • Establishing a peer review process to be taken. Most participants felt strongly gather input on analyzing the climate that a solid, quantitative evaluation of impacts of aviation within the perspec- climate change impacts for a range of tive of other environmental impacts aviation technological, fuel, operational such as noise and air quality. and policy related options would be desirable. For the sponsors, there are additional steps that could be taken to make the results more useful for decision-making. For example, a smaller group of science and other aviation experts could be brought together to take the outcome of this Workshop as a starting point towards addressing issues such as: • Ranking and prioritizing research needs across all three subgroups based on the ability to contribute toward answering the practical needs faced by aviation community (operators, manufactures, regulators and policymakers). • Establishing a forum for dialogue between the science and aviation (manufacturers, operators, and regulators) communities to provide input to the prioritization process. • Identifying what existing research activities can be leveraged toward meeting practical needs. • Identifying how long it would take to achieve the identified practical needs

CO2 effects of aviation in emission trading References schemes. Atmos. Environment, 40, 1117- Baumgardner, D., and B. E. Gandrud, 1998: 1121. doi:10.1016/j.atmosenv.2005.11.005. A comparison of the microphysical and opti- Fuglestvedt, J. S., T. K. Bernsten, O. Godal, cal properties of particles in an aircraft con- R. Sausen, K. P. Shine, and T. Skodvin, trail and mountain wave cloud. Geophys. Metrics of climate change: assessing radia- Res. Lett., 25, 1129-1132. tive forcing and emission indices. Climatic Berntsen, T. K., J. S. Fuglestvedt, M.M. Change, 58, 267-331, 2003 Joshi, K. P. Shine, N. Stuber, M. Ponater, Gauss, M., I. Isaksen, V. Grewe, M. Köhler, R. Sausen, D. A. Hauglustaine, and L. Li, D. Hauglustaine, D. Lee, 2004: Impact of 2005: Response of climate to regional Aircraft NOx Emissions: Effects of Changing emissions of ozone precursors: sensitivities the Flight Altitude. In R. Sausen, C. Fichter and warming potentials. Tellus, 57B, 283- and G. Amanatidis (eds.): European Con- 304. ference on Aviation, Atmosphere and Cli- DeMott, P.J., D.C. Rogers, and S.M. Krei- mate (AAC). Proceedings of an International denweis, 1997: The susceptibility of ice for- Conference, Friedrichshafen, Germany, 30 mation in the upper tropospheric clouds to June to 3 July 2003, EUR 21051, 122-127. insoluble aerosol components. J. Geophys. Hansen, J., Mki. Sato, R. Ruedy, L. Naza- Res., 102, 19575-19584. renko, A. Lacis, G.A. Schmidt, G. Russell, I. Eichkorn, S., K.-H. Wohlfrom, and F. Arnold: Aleinov, M. Bauer, S. Bauer, N. Bell, B. 2002. Massive positive and negative chemi- Cairns, V. Canuto, M. Chandler, Y. Cheng, ions in the exhaust of an aircraft jet engine A. Del Genio, G. Faluvegi, E. Fleming, A. at ground level: Mass distribution measure- Friend, T. Hall, C. Jackman, M. Kelley, N. ments and implications of aerosol formation. Kiang, D. Koch, J. Lean, J. Lerner, K. Lo, S. Atmospheric Environment, 36, 11, 1821- Menon, R. Miller, P. Minnis, T. Novakov, V. 1825. Oinas, Ja. Perlwitz, Ju. Perlwitz, D. Rind, A. Romanou, D. Shindell, P. Stone, S. Sun, N. Eyring, V., Harris, N. R. P., Rex, M., Shep- Tausnev, D. Thresher, B. Wielicki, T. Wong, herd, T. G., Fahey, D. W., Amanatidis, G. M. Yao, and S. Zhang, 2005: Efficacy of T., Austin, J., Chipperfield, M. P., Dameris, climate forcings. J. Geophys. Res. 110, M., Forster, P. M. De F., Gettelman, A., D18104, doi:10.1029/2005JD005776. Graf, H. F., Nagashima, T., Newman, P. A., Pawson, S., Prather, M. J., Pyle, J. A., Hendricks, J., B. Kärcher, A. Döpelheuer, J. Salawitch, R. J., Santer, B. D., Waugh, D. Feichter, U. Lohmann, and D. Baum- W., A strategy for process-oriented valida- gardner, Simulating the global atmospheric tion of coupled chemistry–climate models. black carbon cycle: A revisit to the contribu- Bull. Amer. Met. Soc., 86: 1117-1133, 2005. tion of aircraft emissions. Atmos. Chem. Phys. 4, 2521-2541, 2004. Field, P. R., R. J. Cotton, K. Noone, P. Glantz, P. H. Kaye, E. Hirst, R. S. Gre- Hendricks, J., B. Kärcher, U. Lohmann and enaway and C. Jost, R. Gabriel, T. Reiner, M. Ponater, Do aircraft black carbon emis- M. Andreae, C. P. R. Saunders, A. Archer, sions affect cirrus clouds on the global T. Choularton, M. Smith, B. Brooks, C. scale? Geophys. Res. Lett. 32, L12814, Hoell, B. Bandy, D. Johnson and A. Heyms- doi:10.1029/2005GL022740, 2005. field, 2001: Ice nucleation in orographic Intergovernmental Panel on Climate wave clouds: Measurements made during Change, Aviation and the Global Atmos- INTACC. Q. J. R. Meteorol. Soc, 127, phere. J. E. Penner, D. H. Lister, D. J. 1493—1512. Griggs, D. J. Dokken, and M. McFarland Forster, P. M. de F., K. P. Shine, and N. (eds.), Cambridge University Press. Cam- Stuber, 2006: It is premature to include non- bridge, UK, 1999.

Intergovernmental Panel on Climate Marquart, S., M. Ponater, F. Mager and R. Change, 2001: Climate Change 2001: The Sausen, 2003: Future development of con- Scientific Basis. Contribution of Working trail cover, optical depth and radiative forc- Group I to the Third Asssessment Report of ing: Impacts of increasing air traffic and cli- the Intergovernmental Panel on Climate mate change. J. Climate 16, 2890-2904. Change [Houghton, J.T., Y. Ding, D.J. Miake-Lye, R. C., 2005: Advancing the Un- Griggs, M. Noguer, P.J. van der Linden, X. derstanding of Aviation’s Global Impact. Dai, K. Maskell, and C.A. Johnson (eds.)]. Partnership for Air Transportation Noise and Cambridge University Press, United King- Emissions Reduction. Report No. PART- dom and New York, NY, USA, 881 pp. NER-COE-2005-003, Massachusetts Insti- Johnson, C.E. and R.G. Derwent, 1996: tute of Technology, Cambridge, MA. Relative radiative forcing consequences of Minnis, P., and D. F. Young and D. P. Gar- global emissions of hydrocarbons, carbon ber and L. Nguyen, W. L. Smith Jr. and R. monoxide, and NO from human activities x Palikonda, 1998: Transformation of contrails estimated with a zonally-averaged two- into cirrus during SUCCESS. Geophys. dimensional model. Clim. Change, 34, 439- Res. Lett., 25, 1157-1160. 462. Myhre, G., and F. Stordal, 2001: On the Joshi, M., K. Shine, M. Ponater, N. Stuber, tradeoff of the solar and thermal infrared R. Sausen, and L. Li, A comparison of cli- radiative impact of contrails. Geophys. Res. mate responses to different radiative forc- Lett. 28, 3119-3122. ings in three general circulation models: towards an improved metric of climate Next Generation Air Transportation System, change. Climate Dynamics, 20, 843-854, Federal Aviation Administration report to the 2003. U.S. Congress, 2004. Kärcher, B. and U. Lohmann, 2003: A Pa- O’Neill, B. C., 2000: The jury is still out on rameterization of cirrus cloud formation: Global Warming Potentials. Clim. Change, Heterogeneous freezing. J. Geophys. Res., 44, 427-443. 108 (D14), 4402, Park, J., et al., 1999: The atmosphere Ef- doi:10.1029/2002JD003220. fects of Stratospheric Aircraft: Reports of Kärcher, B., et al., 2004: Particles and cirrus the 1998 Models and Measurements II clouds (PAZI): Overview of results 2000- workshop, NASA/TM-1991-209554. 2003. In R. Sausen, C. Fichter and G. Ama- Penner, J. and Y. Chen, 2006: Influence of natidis (eds.): European Conference on anthropogenic aerosols on cirrus clouds and Aviation, Atmosphere and Climate (AAC). global radiation. In preparation. Proceedings of an International Conference, Friedrichshafen, Germany, 30 June to 3 Prather, M., R. Sausen, A. S. Grossman, J. July 2003, EUR 21051, 197-206. M. Haywood, D. Rind, B. H. Subbaraya, P. Forster, A. Jain, M. Poater, U. Schumann, Klug, H. G., and Ponater, M., 2000: Impact W.-C. Wang, T. M. L. Wigley, and D. J. of hydrogen fuels on contrails and radiative Wuebbles, 1999: Potential Climate Change forcing. In Aviation, Aerosols, Contrails and from Aviation. Chapter 6 in Intergovernmen- Cirrus Clouds (A2C3). U. Schumann, editor, tal Panel on Climate Change, Aviation and Proceedings of a European Workshop, the Global Atmosphere. J. E. Penner, D. H. Seeheim, Germany, July 10-12, 2000 Lister, D. J. Griggs, D. J. Dokken, and M. Mannstein, H. and U. Schumann, 2005: Air- McFarland (eds.), Cambridge University craft induced contrail cirrus over Europe. Press. Cambridge, UK. Meteorol. Z. 14, 549-554. Sassen, K., and C.-Y. Hsueh, 1998: Contrail properties derived from high-resolution po-

58 larization lidar studies during SUCCESS. al., 2004: Atmos. Chem. Phys. Discuss., 4, Geophys. Res. Lett., 25, 1165-1168. 6473-6501. Sausen, R., I. Isaksen, V. Grewe, D. Stuber, N., Piers Forster, Gaby Rädel and Hauglustaine, D.S. Lee, G. Myhre, M.O. Keith Shine, 2006: the importance of the Köhler, G. Pitari, U. Schumann, F. Stordal diurnal and annual cycle of air traffic for and C. Zerefos, 2005: Aviation radiative contrail radiative forcing. Nature, 441, 864- forcing in 2000: An update on IPCC (1999). 867 (15 June 2006) | Meteorol. Z. 14, 555-561. doi:10.1038/nature04877. Schröder, F.P., B. Kärcher, C. Duroure, J. Svensson, F., A. Hasselrot and J. Molda- Ström, A. Petzold, J.-F. Gayet, B. Strauss, nova, 2004: Reduced environmental impact P. Wendling, and S. Borrmann, 2000: The by lowered cruise altitude for liquid hydro- transition of contrails into cirrus clouds. J. gen-fuelled aircraft. Aerospace Science and Atmos. Sci., 57, 464-480. Technology, 8, 307-320. Schumann, U., 2005: Formation, properties, Waitz, I., J. Townsend, J. Cutcher- and climatic effects of contrails. Comptes Gershenfeld, E. Greitzer, and J. Kerrebrock, Rendus Physique, 6, 549-565. Report to the United States Congress: Avia- tion and the Environment, A National Vision, Shilling, J.E., T.J. Fortin, M.A. Tolbert, 2006: Framework for Goals and Recommended Depositional ice nucleation on crystalline Actions. Partnership for AiR Transportation organic and inorganic solids. J. Geophys. Noise and Emissions Reduction, MIT, Cam- Res., 111, D12204, bridge, MA, 2004. doi:10.1029/2005JD006664. Zobrist, B. , Marcolli, C., Koop, T., Luo, B. Smith, S. J., and T. M. L. Wigley, 2000a: P., Murphy, D. M., Lohmann, U., Zardini, A., Global Warming Potentials: 1. Climatic im- Krieger, U. K., Corti, T., Cziczo, D. J., Fueg- plications of emissions reductions. Clim. listaler, S., Hudson, P. K., Thomson, D. S. Change, 44, 445-457. and Peter, T., 2006: Oxalic acid as a het- Smith, S. J., and T. M. L. Wigley, 2000b: erogeneous ice nucleus in the upper tropo- Global Warming Potentials: 2. Accuracy. sphere and its indirect aerosol effect. At- Clim. Change, 44, 459-469. mos. Chem. and Phys. Disc.6, 3571-3609. Stordal, F., G. Myhre, D.W. Arlander, T. Zuberi, B., A.K. Betram, T. Koop, L.T. Svendby, E.J.G. Stordal, W.B. Rossow, and Molina, and M.J. Molina, 2001: Heteroge- D.S. Lee, 2005: Is there a trend in cirrus neous freezing of aqueous particles induced cloud cover due to aircraft traffic? Submitted by crystallized (NH4)2SO4, ice, and letovic- to Atmos. Chem. Phys.. See also Stordal et ite. J. Phys. Chem. A, 105, 6458-6464.

Appendices

Appendix 1

Workshop Agenda

Workshop on the Impacts of Aviation on Climate Change

June 7-9, 2006 Boston, MA June 7 8:00 - 9:00 a.m. Registration and Breakfast* 9:00 - 9:10 a.m. Mohan Gupta – Welcome and logistics 9:10 - 9:25 a.m. Lourdes Maurice – Motivation and Vision for the Workshop: JPDO/NGATS/FAA Context 9:25 - 9:40 a.m. Malcolm Ko – The EIPT-S/M Panel charge to the workshop 9:40 - 10:15 a.m. Don Wuebbles – Science overview, workshop goals & objectives, expected outcomes and format 10:15 - 10:30 a.m. Break 10:30 - Noon Charge to subgroup leaders: Key questions by the subgroup leaders and general discussion Don Wuebbles – Discussion leader 10:30 - 11:00 a.m. Anne Douglass – UT/LS and Chemistry Effects 11:00 - 11:30 a.m. Bernd Kärcher – Contrails and Cirrus Clouds 11:30 - Noon W.-C. Wang – Climate Impacts and Climate Metrics Noon - 1:00 p.m. Lunch 1:00 - 3:00 p.m. Parallel subgroup meetings Focus: Review the present state of scientific knowledge and key underlying uncertainties 3:00 - 3:15 p.m. Break 3:15 - 5:15 p.m. Plenary session: Updates from each subgroup with discussions, (approx. 40 minutes each) Don Wuebbles – Discussion leader 5:15 - 5:30 p.m. Don Wuebbles – Summary and charge for the next session

June 8 - 8:00 a.m. Breakfast* 8:00 - 10:00 a.m. Parallel subgroup meetings Focus: Explore the metrics of aviation emissions relative to other emissions. Discuss the aviation-related trade-off issues. Identify the gaps in aviation-related research needs. 10:00 - 10:15 a.m. Break 10:15 - 11:45 a.m. Plenary session: Updates from each subgroup with discussions, (approx. 30 minutes each) Don Wuebbles – Discussion leader 11:45 - Noon Don Wuebbles – Summary and charge for the next session Noon - 1:00 p.m. Lunch 1:00 - 3:00 p.m. Parallel subgroup meetings Focus: Identify short- and long-term priorities of aviation needs. Identify the ongoing and already planned future research programs and make recommendations on how to leverage upon them to meet aviation needs. 3.00 - 3.15 p.m. Break 3:15 - 4:45 p.m. Plenary session: Updates from each subgroup with discussions, (approx. 30 minutes each) Don Wuebbles – Discussion leader 4:45 - 5:00 p.m. Don Wuebbles – Summary and charge for the next session

June 9 - 8.00 a.m. Breakfast* 8:00 - 10:00 a.m. Parallel subgroup meetings Focus: Develop final consensus on all issues. 10:00 - 10:15 a.m. Break 10:15 - 12:30 p.m. Plenary session: Updates from each subgroup with discussions, (approx. 45 minutes each) Don Wuebbles – Discussion leader 12:30 - 12:55 p.m. Discussion on publication of the final report, involvement of attendees, review and overall schedule. Don Wuebbles – Discussion leader 12:55 - 1:00 p.m. Concluding remarks – Mohan Gupta, Malcolm Ko, Don Wuebbles 1:00 p.m. Adjourn

Appendix 2

Workshop Participants/Authors for Each Subgroup

Workshop Chair: Don Wuebbles, Univ. Illinois - Urbana Champaign

Subgroup 1: Emissions in the UT/LS and Resulting Chemistry Effects Anne Douglass, GSFC NASA : Subgroup Leader Ivar Isaksen, Univ. Oslo, Norway Daniel Jacob, Harvard Univ. Jennifer Logan, Harvard Univ. J. McConnell, York Univ., Canada Dan Murphy, CSD/ESRL NOAA Laura Pan, NCAR Michael Prather, Univ. California-Irvine Jose Rodriguez, GSFC NASA

Subgroup 2: Contrails and Cirrus B. Kärcher, IPA/DLR, Germany : Subgroup Leader Steve Baughcum, Boeing Co. Robert P. d'Entremont, AER, Inc. Andy Dessler, Texas A & M Univ. Paul Ginoux, GFDL NOAA Andy Heymsfield, MMM/ESSL NCAR Sanjiva Lele, Stanford Univ. Rick Miake-Lye, Aerodyne Res. Inc. Pat Minnis, LaRC NASA Dave Mitchell, DRI Karen Rosenlof, ESRL NOAA Ken Sassen, Univ. Alaska Azadeh Tabazadeh, Stanford Univ. Ping Yang, Texas A & M Univ. Fangqun Yu, SUNY-Albany

Subgroup 3: Climate Impacts and Climate Metrics Wei-Chyung Wang, SUNY-Albany : Subgroup Leader Redina Herman, Western Illinois Univ. Joyce Penner, Univ. Michigan Robert Sausen, IPA/DLR, Germany Keith Shine, Univ. Reading, UK Ian Waitz, MIT Don Wuebbles, Univ. Illinois - Urbana Champaign 63

Appendix 3

Other Attendees at the Workshop

Nathan Brown, FAA Mohan Gupta, FAA Curtis Holsclaw, FAA Brian Kim, Volpe Center DOT Malcolm Ko, LaRC NASA Joel Levy, NOAA Karen Marais, MIT Lourdes Maurice, FAA Chris Roof, Volpe Center DOT Saleem Sattar, Transport Canada