Contrails: Research, texts, comments and links

Contrails and aviation-cirrus

European Workshop A2C3

November 27, 2000
Report on the European Workshop "Aviation, Aerosols, Contrails and Cirrus Clouds" (A2C3),
Seeheim near Frankfurt/Main, July 10-12, 2000

Aircraft engine emissions cause the formation of small aerosol particles and contrails in the upper troposphere and lower stratosphere, which - together with other aerosols - may influence cirrus cloud formation (Figure 1, 2). The special report by the Intergovernmental Panel on Climate Change (IPCC) on "Aviation and the Global Atmosphere" (IPCC, 1999) concluded that the largest uncertainties limiting our ability to project aviation impacts on climate and ozone comes from the influence of contrails and aerosols on cloud changes (Figure 3). 

Under guidance of the World Meteorological Organisation (WMO), the "Scientific Assessment of Ozone Depletion: 1998" was prepared (WMO, 1999) which notes incomplete understanding of the influence of aerosol and cloud particles on ozone in the lower stratosphere and the upper troposphere. 

The Second (IPCC, 1996) and the forthcoming Third Assessment Report of IPCC on "Climate Change" note large uncertainties in quantifying the direct and indirect effects of aerosols and changes in cloudiness on climate. At present, less is known about the impact of aerosols and contrail-induced cirrus changes than about the impact of gaseous emissions. Many of the emissions of aircraft engines may be reduced with more advanced technology. But what can be done against contrails and cirrus changes? Therefore  this workshop concentrate on the aerosol and contrail issues.

The three-days European workshop "Aviation, Aerosols, Contrails and Cirrus Clouds" (A2C3) took place in Seeheim, near Frankfurt/Main, Germany, July 10-12, 2000. The workshop started with two days with four scientific sessions followed by one day with working group meetings and discussions. 26 invited oral papers and 55 poster contributions were presented and discussed by the 96 participants from Europe, Russia, and the USA, with representatives from atmospheric sciences, combustion sciences, industry, and various agencies. The first session provided the frame of reference for the following discussions by reviewing results of previous assessments and by presenting results from scientific work in related fields. The following three sessions were devoted to the specific workshop topics: 1) particle formation and properties; 2) contrails and cloud changes; and 3) parameters effecting aviation impact.

The third day was dedicated to discussions at the posters, working groups and a closing panel session. The proceedings (Schumann and Amanatidis, 2000) are in press and will appear in the series of Air Pollution Research Reports of the European Commission. Main results of the papers presented in the proceedings and in posters during the workshop are summarised briefly below.

Black-Carbon and Nonvolatile Soot Particles
Aircraft soot particles (composed of black carbon and other non-volatile materilas) may act as condensation and freezing nucleus for ice particle formation. Soot emissions have been measured in altitude test facilities (A. D�pelheuer and C. Wahl, D.
E. Hagen et al.) and in exhaust plumes and contrails behind cruising aircraft (A. Petzold et al.). Such measurements have been used to validate the soot calculation procedures which are necessary to set up a 3d global aircraft-generated soot inventory (Figure 4). Aircraft emit on average 0.04 g black carbon per kg of burnt fuel; 40% of all soot gets emitted above 10 km altitude, mainly over North America and Europe (A. D�pelheuer and C. Wahl). The number and mass of soot particles emitted per kg of burnt fuel by modern engines is lower (typically 0.01 g/kg, 3�1014 particles/kg fuel) than in older ones (typically 0.1 g/kg and 2�1015/kg). The November 27, 2002 soot emission mass and number is independent of fuel sulphur content. Primary soot particles are typically 25 to 60 nm large in diameter, soot agglomerates reach larger than 150 nm in size (A. Petzold et al.).

Aircraft engine soot may not be as hydrophobic as assumed in the past. High initial water uptake has been measured on aviation soot. Such soot particles acquire a substantial fraction of a water monolayer already in the very young exhaust plume, which would explain contrail formation independent of fuel sulphur content (O.Popovitcheva et al.). 

The activation of soot particles from a graphite spark gap generator with mobility equivalent diameters of 50 to 200 nm has been investigated in a new large aerosol chamber (AIDA in Karlsruhe) at 255 and 239 K. For 255 K, the soot was observed to become activated to form droplets when reaching liquid saturation; for 239 K, the droplets freeze and evaporate quickly when the humidity drops below ice saturation (A. Nink, O. M�hler et al.).

The main processes leading to soot particles are now known and can be reasonably approximated in models at least for laminar flame situations. However, a reliable technique for quantitative calculation of soot emissions from engine combustors has still to be developed. The most important contribution to soot comes from surface growth. Particle interception and coagulation dominate the particle dynamics. Large amounts of soot forms in the early combustion process but most of it gets oxidised before leaving the combustion chamber. Soot prediction within aero-gas turbine combustion chambers require soot models coupled to three-dimensional (3d) fluid dynamics and radiation transfer calculations (H.Bockhorn, H. Brocklehurst, and D. Hu, P. Frank et al.).

Chemiions (CI) and Volatile particles
Large numbers (about 2.4�1017 per kg of burnt fuel) of volatile particles form behind aircraft at cruise. The volatile particles may contribute to contrail and cirrus particle formation (Figure 5). The size spectrum of particles from 3 nm to 100 nm has been measured with a battery of condensation nucleus counters of different size sensitivities; size spectra of larger particles are measured with optical probes (A. Petzold et al.; F. Schr�der). Smaller particles of different masses are detectable as chemiions (CI).

Engines emit large numbers of positive and negative chemiions. Chemiions are produced by chemistry in the combustors. They may enhance the growth of volatile particles, and hence influence the formation of ice particles. Chemiions may have large sizes, possibly more than 8500 atomic mass units, and hence may reach up to the smallest condensation nuclei measurable today (about 3 nm) (F. Arnold et al., K.-H. Wohlfrom et al.). Very massive CI were observed even for nearly sulphur free fuels. However the number of negative CI grows with fuel sulphur content (FSC). Negative CI are larger in mass than positive CI (K.-H. Wohlfrom et al.). Negative CI contain HSO4-H2SO4 cluster ions. Positive CI have been identified as at least partly oxygenated hydrocarbons (A. Kiendler et al.). Models explain the measured number of chemiions in hot aircraft exhaust plumes by ion-ion-recombination and initial ion concentrations of the order of 108 cm-3 (A. Sorokin et al.; F. Arnold et al.).

The concentration of volatile particles strongly increases as diameter decreases toward the sizes of the large molecular clusters. The number of volatile particles larger than 3 nm detectable in a young (order 0.5 to 10 s) plume increases with FSC for FSC in the range from 60 to 1000 �g/g (Figure 6). Particle formation does not change significantly when the FSC gets reduced below 60 �g/g, apparently because of condensable hydrocarbons contributing to the formation of volatile aerosols in the exhaust plume (Petzold et al.). 

The variation in the number of measured volatile particles behind several aircraft can be explained with a model in terms of variations in plume age, size detection limit of the particle counters, FSC, the amount of condensable organic material, and, most important, with the number of CI emitted (B. Karcher et al.).

Sulphuric acid
Part of the sulphur contained in the aircraft fuel may get converted to sulphuric acid which affects the formation of volatile particles and the hydration properties of soot. The conversion fraction of fuel sulphur to S(VI) (SO3, H2SO4) depends on pressure and temperature at combustor exit and on the residence times. The conversion is largest at a temperature of about 1000 K and high pressures. Cooling of turbine blades in the high pressure turbine enhances the conversion (R. Miake-Lye et al.). In flight measurements indicate that around 3 % of FSC
gets converted to sulphuric acid, less than believed a few years ago (F. Arnold et al., K.-H.Wohlfrom et al.).

Nitric acid and other aerosol consitutents
H2O-H2SO4 aerosols take up HNO3 only at small mole fractions (<1%) and only for temperatures below 222 K and high ambient humidity (G. Gleitsmann and R. Zellner). The uptake coefficient of HNO3 (order 10-3 mg-1) and H2O (5�10-4 mg-1) on soot have been measured. Uptake of HNO3 seems to be of little importance for volatile aerosols, soot activation, and contrail formation in aircraft exhaust plumes (T. Keil et al.). The uptake of NOy on cirrus cloud particles was inferred from measurements of gas- and particle phase NOy,
H2O, HNO3, and aerosols for temperatures near 200 K. About 2 to 10% of the NOy gets removed by a coverage of ice particles with about 0.1 monolayers of NOy (H. Schlager et al.).

A thermodynamic model of hydrogen-ammonium-sulphate-nitrate-water aerosols has been provided for upper tropospheric temperatures (S. Clegg and P. Brimblecombe). The nonreactive uptake of HCl on frozen film ice surfaces was determined for temperatures from 205 - 235 K. The uptake coefficient decreases substantially when T increases from 205 to 215 K (R. G. Hynes et al.).

Global aerosol distribution
A global stratospheric and tropospheric sulphate aerosol model has been developed, compared to observations from the SAGE satellite instrument in terms of extinction, optical depth, mass and surface area density, and applied to study future changes of stratospheric aerosols in relation to projected changes of anthropogenic sulphur emissions. Subsonic aircraft were found to contribute about 3% to the total aerosols mass in the stratosphere and 20 % of the surface area density at 14 km altitude north of 45�N (G. Pitari and E. Mancini).

Contrails trigger cirrus cloud formation. In air masses with ice-supersaturation of a few percent no cirrus cloud would have formed otherwise, see Figure 2 (U. Schumann). Aerosols from aircraft may influence cirrus formation, possibly long after the time of emission. The crucial contrail parameters are their cover and optical properties. Contrails may warm the Earth surface in particular over warm and bright surfaces. During day, the enhanced Earth's albedo caused by contrails may cause a cooling. Around 17% of the Earth is covered with air masses which are ice-supersaturated and cold enough so that persistent contrails form when aircraft fly in these region. The area coverage by such regions susceptible to contrails is 21% when aircraft would burn liquid hydrogen instead of kerosene (K. Gierens and S. Marquart).

Several recent studies show that supersaturation with respect to ice is very common in the upper troposphere (K. Gierens, E. Jensen, J. Ovarlez et al.), see Figure 7. In the Arctic winter, supersaturation occurs even in the lowermost stratosphere due to the extremely low temperatures in this region. Also the tropical upper troposphere is often ice supersaturated.

Cirrus clouds rarely extend more than 1 km above the tropopause (E. Jensen et al.). Large ice supersaturation also occurs within cirrus clouds (J. Ovarlez et al.). Short lived contrails of 2-engined aircraft were observed to evaporate earlier than contrails from 4-engines aircraft, consistent with model calulations (K. Gierens and R. Sussmann). The microphysical properties of contrails and cirrus clouds have been measured during more than 20 airborne missions over Central Europe at temperatures below -50�C. Contrails contain large concentrations of nearly spherical ice crystals, initially up to 105 cm-3, later diluted to typically a few 100 cm-3, while the mean diameter range from 1-10 �m. Young cirrus clouds contain mostly regularly shaped ice crystals in the size range of 10-20 �m at typical concentrations of 2-5 cm-3. Contrail growth is affected only weakly by pre-existing cirrus clouds (F. Schr�der).

In contrails and cirrus clouds, the scattering phase function has been measured in situ with a Polar Nephelometer. The scattering by cold ice clouds differs strongly from that of spherical particles, implying 40% higher albedo, i.e. a stronger shortwave cooling by contrails than computed for spherical particles (F. Auriol et al.). 

Particle measurement systems may be strongly disturbed by electromagnetic interferences. Older probes with slower electronics may have difficulties in registering crystal size and number density accurately (E. Raschke et al.).

Contrails have been estimated to cover currently about 0.1% of the global sky (K. Gierens and S. Marquart). Over the USA, the mean contrail coverage has been determined from 4 months of satellite data to be 1.8%. For 55 contrails over different backgrounds the mean optical depth and the effective particle diameter have been determined and found to be 0.46 and 36 �m, respectively, see Figure 10 (P. Minnis et al.). Over central Europe, the annual daytime average contrail has been determined from 2 years of NOAA-14 satellite data with an automated contrail detection algorithm. The annual day time average cover is 0.7% (Figure 9) with a strong annual cycle (1% in winter, 0.4% in summer). The mean visible optical depth in this region is determined from the same data to be 0.11. During night the cover is 3 time smaller. The resultant radiative forcing is found to be one order of magnitude below the results of previous assessments. (R. Meyer et al.). Data from traffic management systems indicate strong diurnal variations in air traffic, which is important in assessing the mean radiative forcing by contrails; contrails warm during night but may cool during day time (S. Baughcum et al.).

A case study of a contrail has been reported which has been measured from ground with a Lidar, and in situ with particle size spectrometers. Pyranometers and pyrgeometers were used to measure the radiative shortwave and longwave vertical fluxes near the contrail. The solar optical depth of the contrail varies from 0.05 to 0.23. The measured radiation fluxes indicate smaller radiative forcing than what is computed for the measured size spectrum assuming spherical ice particles (P. Wendling et al.).

Gas phase chemistry and heterogeneous reactions on contrail and cirrus particles 
Aerosols and ice particles contribute to changes in ozone and other air constituents. NOx emissions by subsonic aircraft lead to net ozone production in the tropopause region. However, within contrails, halogen components may get activated and NOx and hydroxyl radicals may get reduced, causing less ozone production, and ozone depletion is possible at least locally within contrails. In the lowermost stratosphere, ozone destruction is caused by heterogeneous chlorine activation on contrail ice particles, whereas in the upper troposphere ozone production is compensated by heterogeneous reduction of OH radicals (S. Meilinger et al.). A chemistry box model study shows a large sensitivity of upper tropospheric chemistry to heterogeneous processes in cirrus clouds. Ice particle sedimentation may reduce upper tropospheric HNO3, NOx, and OH, resulting in significant ozone loss (A. Meier and J. Hendricks). A 3d chemical transport model (CTM) has been used to investigate the role of subvisible ice clouds in chemical ozone loss near the tropopause. The calculated zonal mean ozone loss reaches a maximum in the mid- and high latitude tropopause region of about 1% in winter and 3% in summer (B. Bregman et al.).

Seven years of ozone, humidity and temperature profile observations at 21�S, 55�E reveal a systematic diminution of ozone abundance in tropical cirrus clouds, possibly due to ice cloud induced denoxification (S. Roumeau et al).

Only a few attempts have been made to model the formation and microphysics of cirrus clouds (W. Wobrock and A. Flossmann; T. Chourlaton et al.). From measurements and models, it was found that aerosols which passed through cirrus behave differently from aerosol that was not yet contained in frozen particles (T. Choularton). Large aerosol may act as ice nuclei by homogeneous freezing, small aerosols may enhance freezing by acting as heterogeneous ice nucleus (W. Wobrock). Bulk and size-resolved cirrus process-models show large differences even when applied to idealised cases with prescribed ambient conditions; particle sedimentation rate is one of the crucial model parameters (GEWEX-cirrus modelling initiative, poster of D. O'C. Starr).

Volatile aerosol freezes for temperatures, T, in the range 233.5 K < T < 273.15 K, when the ambient air reaches liquid saturation. Below 233 K, homogeneous ice nucleation occurs in particles of 0.2 �m radius within one minute when the relative humidity over ice (RHI) exceeds RHI = 238.7 - 0.398 T (RHI in %, T in K; Koop et al.).

The ECHAM4 global circulation model (GCM) has been used to study the sensitivity of ice clouds to different assumptions about ice nuclei. Ice nuclei are assumed to be either just a function of temperature and altitude, insoluble carbonaceous aerosols or dust aerosols. The different models cause variations in the longwave cloud forcing by 4 W m-2 (U. Lohmann). A first GCM study with parameterised contrails was presented, showing variations in the cover and optical depth of contrails with latitude, altitude, and season; the optical depth of contrails appears to be larger over the USA than over Europe because of different temperature and humidity conditions (M. Ponater et al.).

Lidar observations in Southern France (44�N, 6�E) have been used to determine the occurrence frequency and altitude range of thin cirrus at various optical depths. Subvisible cirrus (532 nm optical depth < 0.03) constitute a substantial portion (40%) of the calculated mean optical depth. Fewer than 5% of the cirrus occurrences are (at most <1 km) above the tropopause (L. Goldfarb et al.).

For the first time, in situ observations of cirrus clouds have been performed in the Southern hemisphere during the Interhemispheric Differences in Cirrus Properties From Anthropogenic Emissions (INCA) experiment in March/April 2000. An extensive scientific payload to characterise aerosol-cloud interactions, cloud microphysical properties, and trace gases was deployed on the DLR research aircraft Falcon. A second experiment in the Northern Hemisphere midlatitudes has been performed in Sept./Oct. 2000 to compare clean and polluted air masses in order to study whether anthropogenic emissions have a measurable effect on cirrus cloud properties. First results were presented showing clear interhemispheric differences in various aerosol components (volatile and non-volatile particles of various sizes; crystal residuals and interstitial aerosols; see Figure 8) and trace gases (CO, O3, NO, and NOy) (J. Str�m et al.). The observations are accompanied by ground based Lidar (Immler et al.) and satellite observations (P. Minnis et al.). At Punta Arenas the mean cirrus cover was 45%. On the way from Punta Arenas to Bremerhaven, the Lidar was mounted on the POLARSTERN and observed extended cirrus clouds in the tropics (F. Immler et al.).

Climate Impact
Little progress has been made beyond that described in the IPCC report in understanding which aviation parameter is most important in controlling the climate impact by contrails and cirrus. The total aviation climate impact (in terms of radiative forcing, and without cirrus changes) is about three times as large as the radiative forcing from aircraft-induced carbon dioxide alone (R. Sausen; U. Schumann), see Figure 3 (IPCC,1999).

The climate response is more sensitive to aircraft induced ozone than to a CO2 perturbation with the same radiative forcing. Global warming leads to a smaller increase in contrail coverage than would be expected from growing air traffic in an unchanged atmosphere (R. Sausen). Some observations indicate long-term cirrus cover trends (P. Minnis et al.). In general, long term observations are needed to identify any systematic cloud or climate change due to aviation, if present, because of the large variability of cloud systems (V. Ramaswamy).

Mitigation Potentials
Contrails form for thermodynamic reasons and, hence, can be avoided only by flying in warmer air either below or above the altitude range (around the tropopause at midlatitudes) which is cool enough to let contrails form. Recent experiments have provided evidence that contrails form at lower altitude and hence more frequently when using more efficient engines. However, more efficient engines also emit less carbon dioxides. The trade-off between both effects depends on the relative sensitivity of climate to contrails and carbon dioxide changes and on the scenario of future aviation (U. Schumann).

Fuel properties have some impact on the formation of soot, volatile particles, and the carbon dioxide and water vapour emissions. The impact of variations within various jet fuels for soot formation is small compared to the variability between various combustor concepts (poster by P. Madden). Elimination of fuel sulphur would reduce the size, and to minor degree also the number of volatile aerosol particles formed. However, volatile particles would form even for a very low fuel sulphur content (B. K�rcher et al.).

Hydrogen fuelled aircraft have been proposed in order to reduce carbon dioxide emissions by aircraft. Very low NOx emissions from such aircraft have been demonstrated. The radiative forcing by the emitted water vapour is small, but the impact on contrails and on cirrus clouds has still to be determined (H. Klug and M. Ponater). Replacing the fleet of kerosene driven aircraft by a fleet of liquid hydrogen driven aircraft would pay off in terms of reduced radiative forcing by carbon dioxides only in the long term, after several decades (S.Marquart et al.). Contrails behind liquid hydrogen driven engines may have smaller climatic impact because of different particle properties (larger in size but less in number) compared to contrails behind kerosene burning engines.

It is not yet known whether the climatic impact depends just on fuel consumption of the aircraft flying in sufficiently cold and humid air masses, or whether other parameters are more or equally important, such as the number of aircraft in operation, the number and size of soot particles emitted, the sulphur content in the fuel used, or just the flight route. In the absence of a reliable estimate of the climatic impact of contrails, it is not yet possible to assess various flight routing and operational options (U. Schumann).

Ongoing and future research programmes 
Several research programmes were described: Atmospheric research in Germany supported by the Bundesministerium f�r Bildung und Forschung (G. Hahn); laboratory studies of heterogeneous processes on ice (CUT-ICE; J. N. Crowley et al.); aerosol/particulate research activities supported by the NASA ultra-efficient engine technology (UEET) program (C. C. Wey); measurements of aged aircraft exhaust in the ACCENT mission (R. Friedl et al.); environmental considerations in aircraft design (C. Hume); the CRYOPLANE project considering the potential of liquid hydrogen as aircraft fuel (H. Klug); perspective of the U.S. federal aviation administration (H. Wesoky); overview on NASA's plans to address the concerns identified in the IPCC special report (D. Anderson and R. Friedl); programmatic perspectives of the European research on aviation's atmospheric effects (G. Amanatidis); and industry and agencies perspectives were presented in opening remarks and panel discussions (J. Raps, F. Beul�, P. Newton, F. Walle). A series of SULFUR experiments (1-7) was performed by the Deutsches Zentrum f�r Luft- und Raumfahrt (DLR) in Germany; DLR and partners of other major national research centres (within the Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren, HGF) perform a 3-years project Particles from Aircraft Engines and Their Impact on Contrails, Cirrus Clouds, and Climate (PAZI).

Important open questions addressed in the working groups include: The prediction of the number of soot particles emitted from the fleet of aircraft; the hydration and ice nucleation properties of soot and other aerosol components; the global cover and optical properties of contrails and aircraft induced cirrus, and their long-term trends; the heterogeneous chemistry on cirrus and contrail particles and their effects on air composition; the statistics of vertical wind, humidity, and cirrus properties in the tropopause region; the identification of the most relevant aerosols component which should get reduced for an environmentally sustainable air traffic development.

Aerosols are key issue for understanding of aircraft impact, ozone chemistry and climate changes. Aviation emissions are contributing a small but growing fraction of all anthropogenic disturbances to the climate system. In view of the rather well defined disturbance, understanding of aviation effects helps in understanding the more general aerosol-cloud-climate-chemistry interactions from all sources.

November 27, 2000

Contrails are Bad News