JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 117, D11201, doi:10.1029/2011JD017019, 2012

Contrails developed under frontal influences of the North Atlantic B. A. Laken,1,2 E. Pallé,1,2 D. R. Kniveton,3 C. J. R. Williams,4 and D. A. Kilham5 Received 14 October 2011; revised 24 April 2012; accepted 24 April 2012; published 1 June 2012.

[1] Satellite imagery reveals a visually striking pattern of persistent line-shaped located to the Northwest of the British Isles on 1 September 2007, just before sunrise. These contrails formed over the heavily trafficked eastbound North Atlantic Track (NAT) flight paths, as they intersected an area of prefrontal mixing. The high relative humidity with respect to within the prefrontal mixing zone allowed the contrails to persist, while the strictly regulated flight paths over the region account for their remarkable shapes. The positioning of the NAT flight paths to take advantage of the jet stream likely maximized regional formation. An estimation of the outgoing top of the atmosphere longwave (LW) flux from the CERES instrument shows that the contrails reduced the local instantaneous LW emissivity by 20.96 (Æ0.26) W/m2. This example demonstrates that for middle latitude regions, prefrontal mixing is an important factor governing the radiative influence of contrails. However, a full estimation of the radiative impact is not possible, as we cannot specify the amount of shortwave forcing caused by the persistence of the contrails into daytime; moreover, several hours after formation the line-shaped contrails spread and overlapped with pre-existing structures. Citation: Laken, B. A., E. Pallé, D. R. Kniveton, C. J. R. Williams, and D. A. Kilham (2012), Contrails developed under frontal influences of the North Atlantic, J. Geophys. Res., 117, D11201, doi:10.1029/2011JD017019.

1. Introduction conditions below the ice saturation level are short-lived, as their constituent ice crystals sublimate rapidly after forma- [2] Condensation trails (contrails) are aviation-induced tion. Alternatively, if the RHI is greater than ice saturation ice-phase , usually formed at pressures below 300 mb, levels they may persist as nearly linear structures for an at air temperatures less than À30C. They can persist in air extended period of time (termed persistent contrails), or may that is saturated with respect to ice [Jensen et al., 1998]. even develop into a cirrostratus cloud layer (referred to as They primarily form due to the mixing of the warm, water spreading contrails) [Jensen et al., 1998; Haywood et al., vapor (and aerosol) rich air of the aircraft exhaust plumes, 2009]. Additionally, the injection of aerosols into the upper with the cold and less humid ambient air [Appleman, 1953; troposphere from aircraft exhaust plumes also provides a Scorer, 1955]. Like all clouds, contrails both cool and warm source of ice forming nuclei which may subsequently influ- the climate by influencing Earth’s radiative balance: they ence cirrus properties via heterogeneous nucleation reflect incident shortwave (SW) radiation back to space, mechanisms (the aerosol indirect effect) [Lee et al., 2009]. resulting in a cooling effect, while they simultaneously trap [4] Contrails lead to an increase in cloud coverage due to outgoing terrestrial longwave (LW) radiation, resulting in a the fact that during the mixing of aircraft exhaust plume air warming effect. The net influence of contrails is believed to with the surrounding air, ice crystals are formed even though be a small magnitude warming effect [Forster et al., 2007]. the conditions for the formation of natural clouds may not [3] The lifetime of contrails depends on the ambient yet have been reached. Observational studies suggest that atmospheric conditions, specifically the relative humidity the total coverage of line-shaped contrails over regions with with respect to ice (RHI). Contrails formed under RHI heavy air traffic may be around several percent [Bakan et al., 1994; Meyer et al., 2002]. Observational estimates of the global coverage of aircraft-related cirrus clouds have been 1 Instituto de Astrofísica de Canarias, San Cristóbal de La Laguna, made by Stordal et al. [2005] from extrapolations of regional Spain. 2Department of Astrophysics, Faculty of Physics, Universidad de La correlations between cirrus detected by satellite data and Laguna, San Cristóbal de La Laguna, Spain. modeled air traffic simulations. These results suggest that in 3Department of Geography, University of Sussex, Falmer, UK. the year 2000 global aircraft-related cirrus cover was approx- 4NCAS-Climate, University of Reading, Reading, UK. – 5 imately 0.1 0.38%; for a detailed comparison of regional Asgard Consulting, Basingstoke, UK. and global contrail coverage estimates, see Rap et al. [2010] Corresponding author: B. A. Laken, Instituto de Astrofísica de and references therein. In addition to increasing cloud cover, Canarias, Via Lactea s/n, E-38205 La Laguna, Tenerife, Spain. it has also been noted that contrail cirrus may reduce natural ([email protected]) formation. Model studies suggest that these Copyright 2012 by the American Geophysical Union. reductions may be around 1–2% regionally and may partially 0148-0227/12/2011JD017019

D11201 1of7 D11201 LAKEN ET AL.: CONTRAILS UNDER FRONTAL INFLUENCES D11201

Figure 1. Contrail development from the AVHRR. A series of AVHRR Channel 5 (11.5–12.5 mm) images (1.1 km2 resolution) between 31/08/2007 and 01/09/2007 showing the precession of the frontal cloud system and development of the contrails over an area of the North Atlantic.

offset the impact of contrails on climate change [Burkhardt [6] The radiative impacts of contrails are primarily con- and Kärcher, 2011]. trolled by their extent, optical depth and lifetime. Persistent [5] The radiative forcing (RF) resulting from aviation- RHI supersaturations are often observed in the upper tro- induced clouds may alter Earth’s climate, and the magnitude posphere of the tropics and the lower troposphere of polar of this effect is expected to increase over time as the volume regions, while middle latitude (storm track) regions fre- of air-traffic increases [Marquart et al., 2003]. The IPCC quently show large variability in saturation levels [Kahn fourth assessment report rates the level of understanding et al., 2009]. Cirrus clouds may form as a result of hetero- of the global radiative forcing of contrails to be low (due geneous or homogenous nucleation processes: homogenous largely to the difficulties in identifying aircraft-induced ice nucleation requires supersaturations of >45% [Koop cirrus) [Forster et al., 2007]. Estimates suggest that for et al., 2000] and may be the dominant process of natural persistent line-shaped contrails the RF in the year 2000 may cirrus formation, whereas heterogeneous ice nucleation have been approximately +0.01 W/m2 [Sausen et al., 2005]. mechanisms operate at lower supersaturations [DeMott et al., However, considering the additional effects of contrails 2003; Möhler et al., 2006], and are an important nucle- cirrus, model results suggest that the global RF (of line- ation process influencing contrail formation and properties shaped and contrail cirrus combined) may be much higher, [Kärcher and Yu, 2009]. Consequently, there are many namely 0.0375 W/m2 for the year 2002 [Burkhardt and cloud free regions where supersaturation conditions exist Kärcher, 2011]. at levels too low to allow for homogenous nucleation to

2of7 D11201 LAKEN ET AL.: CONTRAILS UNDER FRONTAL INFLUENCES D11201

Figure 2. Pressure and wind conditions. (a) Synoptic surface level pressure conditions (mb) from the UK Met Office at 00:00 UT 01/09/2007, and (b) vector winds at 225 mb over a polar stereographic projection of the Northern hemisphere, from the NCEP Climate Forecast System Reanalysis (CFSR), as a mean of 00:00–06:00 UT on 01/09/2007. occur, but which are susceptible to contrail formation via with the suggestion of Burkhardt and Kärcher [2009] that condensation freezing of either aircraft or pre-existing ambient contrail-induced cirrus coverage globally may be dominated aerosols [Gierens et al., 1999; Kärcher et al., 2007]. by a few major events, and scale with supersaturation [7] The heterogeneous freezing of aging soot particles rather than the frequency of contrail formation. Conse- from the aircraft exhaust plume may cause an additional quently, the uncertainty regarding future changes in air increase in cloud cover. However, the magnitude of this temperature and RHI at upper tropospheric levels may have effect is not yet known, as it is currently unclear if soot par- significant consequences for the RF of aviation induced ticles are efficient ice forming nuclei in atmospheric condi- cloud [Forster et al., 2007]. In this work, we present a case tions. Furthermore, an increase in cloud cover resulting from study of a persistent contrail event formed under the influ- this process may not happen within the plume phase, making ence of frontal mixing in the North Atlantic region, including it problematic to observe resulting cloud cover changes an estimate of the line-shaped contrails LW radiative impact, [Jensen and Toon, 1997]. with the intention of increasing the sparse number of obser- [8] Once formed, the radiative impacts of persistent and vational studies available for evaluating model estimates. spreading contrails may be locally intense. Using a case study of the radiative effects of a spreading contrail devel- 2. Satellite Observations of North Atlantic oping in to a cirrostratus cloud over England, Haywood et al. Contrails on 1 September 2007 [2009] estimated a local net radiative forcing of +10 to +30 W/m2 for daytime and nighttime respectively. This case [9] Satellite telemetry reveals the presence of several study showed a single localized event generating a global- striking high level cloud formations to the northwest of mean RF equivalent to 7% of the estimated diurnally aver- England on 1 September 2007 prior to sunrise. Figure 1 aged linear contrail RF globally. This result is in agreement shows the development of these features as recorded by the

3of7 D11201 LAKEN ET AL.: CONTRAILS UNDER FRONTAL INFLUENCES D11201

the persistent line-shaped contrails are no longer visible, although less apparent contributions to the high-level cloud coverage may still remain (Figure 1). [11] The remarkable shapes and high densities of the line- shaped contrail clusters observed are attributed to the east- bound (nighttime) North Atlantic Track (NAT) flight paths of transatlantic routes. The North Atlantic region has a high volume of heavily regulated air traffic. On the date of 01/09/2007 between the hours of 00:00 – 05:00 UT approx- imately 600 aircraft were flying along 5 NAT paths. The locations of the NATs are primarily determined by the meteorological conditions: the NATs are planned so that the eastbound aircraft fly with the aid of the polar jet stream. On 01/09/2007 the jet stream was concentrated over a small region, indicated by the frontal cloud system (the jet stream conditions are described in detail in the following section), as a result the NATs are located closely together (the mini- mum legal distance between individual tracks is 1). An interesting feature of the contrail groups is the apparent convergence of each track to a single point. This reflects a convergence of air traffic along the transatlantic flight corridor, which appears relatively dispersed toward its American side but gets narrower as aircraft approach the heavily trafficked and tightly regulated West European air- space. As a result of the aircraft converging on the NAT paths and traveling in great circles the observed pattern of inter- secting points is produced (clearly seen at 05:11 UT in Figure 1).

3. Analysis of Atmospheric Conditions

[12] Synoptic surface level pressure conditions on 1 September 2007 at 00:00 UT from the UK Met Office are presented in Figure 2a. The figure shows cold fronts advecting across the North Atlantic toward a region of high pressure, while occluded fronts advance northeastward across the UK. The position of the Polar jet stream is apparent from the 225 mb vector winds, shown in Figure 2b with data from the NCEP Climate Forecast Reanalysis System [Saha et al., Figure 3. Relative humidity and contrails. (a) Contrail 2010]; from a consideration of the synoptic conditions it is and high level cloud features from AVHRR (channel 5) clear that the cloud system is formed under the influence of 05:11 UT overlaid on to NCEP CFSR relative humidity with a polar low pressure front. respect to ice (%) data at 230 mb, which has been interpolated [13] The form and location of the contrails themselves from a 0.5 Â 0.5 grid. Reanalysis data is a six-hour average give vital clues regarding the local atmospheric conditions. between 00:00 to 06:00 UT. The edge of the AVHRR sensor Specifically, the contrails all terminate abruptly at roughly swath is denoted by the dotted line, and intersects the two the same distance from the frontal cloud system. This indi- most northerly contrail clusters. (b) Meteosat SEVIRI IR/ cates a region of ice supersaturation, which allows the line- water vapor channel (5.35–7.15 mm) retrieved at 06:00 UT. shaped contrails to persist. This notion is confirmed from an examination of the RHI and water vapor sensitive satellite Advanced Very High Resolution Radiometer (AVHRR) channels presented in Figure 3. Figure 3a shows the high- instrument in channel 5 (11.5–12.5 mm) over a 14-h period. level cloud retrieved from the 05:11 UT AVHRR satellite pass overlaid on to a 6-h average of RHI data (between During this time an extensive, high-level cloud system is – present to the northwest of the British Isles, extending 00:00 06:00 UT) at 230 mb from the NCEP CFSR project, roughly from the southern tip of Greenland to the west of while panel 3B shows data from the Meteosat Spinning Enhanced Visible and Infrared Imager (SEVIRI) IR/water Norway, moving in a southeasterly direction with the typical – m characteristics of a frontal cloud system. vapor channel (5.35 7.15 m) taken at 06:00 UT. [14] Although the cloud and RHI data are not simulta- [10] Figure 1 shows that a series of persistent line-shaped contrails clustered in groups are visible within the cloud neous, panel A does provide a way to approximately assess system, extending beyond its southern edge at 03:32 UT. the relationship between the contrails and RHI. As a result of At 05:11 UT the density and extent of the contrail clusters the southeasterly motion of the cloud system, the 6-h aver- has greatly increased, and they have developed into a aging means that the cloud/contrail mask is likely offset from remarkable distribution pattern. By the mid-day satellite pass the RHI. However, from the figure we see that the contrails

4of7 D11201 LAKEN ET AL.: CONTRAILS UNDER FRONTAL INFLUENCES D11201

Figure 4. CERES instantaneous LW flux. Top of the atmosphere upward LW flux (5–100 mm) from the CERES instrument (FM3) cross-track scanner on board NASA’s AQUA satellite from the CERES-CES data set for 01/09/2007 (04:00–05:00 UT). The data have been converted to a 0.25  0.25 grid. appear to terminate at RHI saturations of approximately 85%. pixels located between the contrails (hereafter referred to as To compliment this data, panel B shows that the position of the intracontrail pixels). For that we use the raw (uninterpolated) contrails is strongly constrained by the presence of high water CERES FM3 retrievals. From these data we found the vapor concentrations, resulting from pre-frontal mixing. intracontrail pixels to have a mean (error given is the 1s 2 [15] The contrails do not resume after the reduction in RHI standard error of the mean) of 247.08(Æ0.14) W/m , while south of the frontal cloud system despite the presence of the contrail pixels had a value of 226.12 (Æ0.26) W/m2, additional regions of high RHI (Figure 3). This discontinuity thus the contrails are estimated to have a local instantaneous may be explained by the flight paths of the air traffic, which, LW impact of 20.96 (Æ0.26) W/m2. A paired (two-tailed) rather than crossing through the additional regions of high Student’s T-test performed on ensembles of the uninterpolated RHI instead pass through a corridor of lower RHI extending contrail/intracontrail pixels shows the difference between the toward England. Alternatively, the contrails may not be seen contrail and non-contrail sites to be statistically significant in additional areas of supersaturation as the air temperatures above the 0.99 probability level, confirming that the contrails may be too high for formation to occur. have significantly reduced the local outgoing instantaneous LW radiation flux. 4. Estimated Radiative Changes due [19] During daylight hours, a large portion of this LW to the Presence of Persistent Contrails forcing would be offset as a result of an opposite sign SW forcing. It should also be noted that our estimation of [16] Figure 4 shows an estimate of the change in outgoing LW forcing might be overly simplistic, as the radiative – m LW radiation (5 100 m) from the top of the atmosphere as forcing of a contrail is the flux change due to the presence of measured by the Clouds and Earth Radiant Energy System the contrail with all other parameters held constant. However, (CERES) instrument, on board the AQUA satellite platform we make no attempt to account for any variations in tem- [Wielicki et al., 1996]. The figure shows the field of view perature or humidity between the contrail/intracontrail pixels (FOV) of the CERES FM3 (cross track scan mode) sensor in our calculation. Although we have estimated the impact (available from the CERES CES, Cloud and Radiation of the line-shaped contrails on the LW radiation balance, – Swath data set), between 04:00 05:00 UT on 01/09/2007 the full extent of the contrails impact on the local net   presented on a 0.25 0.25 resolution stereographic plot. radiation balance is difficult to determine, as the contrails [17] The frontal cloud is evident as the blue (low value) persist into the daytime and undergo spreading, developing pixels, as a result of the cloud absorbing large parts of the in to cirrostratus clouding complicating the consideration of outgoing surface LW radiation, while regions of clear sky/ their forcing considerably. low cloud are indicated by red (high value) pixels. The [20] Figure 5 shows a series of 10.8 micron infrared contrails can be clearly seen extending from the cloud as retrievals taken at hourly intervals between 08:00 and several distinct light colored protrusions. It is important to 11:00 UT from the Meteosat Second Generation (MSG) 2 note that the intensity of these regions has likely been Spinning Enhanced Visible and InfraRed Imager (SEVIRI) diminished by the interpolation that we applied to the data. instrument. These images show the spreading of the line- [18] To estimate the local instantaneous outgoing top of shaped contrails into the high level cloud of the frontal the atmosphere (TOA) LW radiative forcing of the contrails system, gradually becoming indistinguishable for both cloud we objectively identify contrail pixels, and compare them to free areas and pre-existing frontal clouds. The ambiguity

5of7 D11201 LAKEN ET AL.: CONTRAILS UNDER FRONTAL INFLUENCES D11201

Figure 5. Development of spreading contrails. Meteosat Second Generation (MSG) 2 Spinning Enhanced Visible and InfraRed Imager (SEVIRI) 10.8 micron infrared retrievals (1 km2 resolution), at hourly intervals from 08:00–11:00 UT 01/09/2007. The spreading of the line-shaped contrails in to the frontal cloud system is clearly visible, indicating that the aviation-induced cloud has enhanced the natural cloud coverage, making the true extent of the radiative impacts of the contrails difficult to determine.

between the aviation-induced cloud and natural cloud is While such a strategy reduces CO2 emissions, it may also clear; this example highlights the difficulties with assessing maximize the potential regional contrail radiative impact. the full impacts of aviation-induced cloud on the climate system. 5. Concluding Remarks [21] The contrail development identified in this study illustrates the importance of frontal processes for the gener- [23] We have presented a case study of persistent con- ation of persistent contrails, by the creation of areas of clear trails, developing in an area of prefrontal mixing over the sky, with high relative humidity levels. These results are North Atlantic region. These contrails are remarkable due to consistent with the findings that contrail development (and their unusual appearance; this is attributed to a clustering of outbreaks) are found to frequently occur over regions asso- individual contrails around NAT flight paths, resulting in the ciated with frontal activity, cyclonic waves, and jet stream appearance of five distinct contrail groups with intersecting activity, where the upper tropospheric areas are associated points. We have analyzed data regarding the evolution and with baroclinic ridging, and the conditions become higher properties of these contrails from a variety of satellite and colder than average [Kästner et al., 1999; DeGrand instruments including a CERES sensor pass. We estimate et al., 2000; Carleton et al., 2008]. that the contrail groups significantly reduced the instanta- neous outgoing LW radiation flux by 20.96 (Æ0.26) W/m2. [22] For regions such as the North Atlantic, where there is This value is comparable in magnitude to the (nighttime) both a high volume of air traffic and where the favorable 2 conditions for contrail formation and persistence are pro- value obtained by Haywood et al. [2009] of 30 W/m . vided by mixing processes along frontal zones, the radiative Although, the aforementioned study included the effects of impact of contrails in the future may vary depending on how contrail cirrus in addition to line-shaped contrails, while this climate change influences frontal activity over the region. study only estimates the radiative impact of line-shaped Furthermore, it is significant to note that eastbound (night- contrails. A chief limitation of estimating the effects of con- time) NAT paths are specifically designed to maximize the trails on climate from observations lies in the ambiguity of volume of air traffic flying within the polar jet stream, in distinguishing aircraft initiated cloud from natural cirrus efforts to minimize fuel consumption and journey time. formations. While for persistent line-shaped contrails this is

6of7 D11201 LAKEN ET AL.: CONTRAILS UNDER FRONTAL INFLUENCES D11201 relatively straightforward due to their unnatural line-shaped lower stratosphere derived from three years of MOZAIC measurements, structure, when considering spreading contrails developing Ann. Geophys., 17, 1218–1226, doi:10.1007/s00585-999-1218-7. Haywood, J., et al. (2009), A case study of the radiative forcing of persistent in to cirrostratus cloud it becomes far more difficult [Forster contrails evolving in to contrail-induced cirrus, J. Geophys. Res., 114, et al., 2007]. From Figure 5, it is clear that the contrails D24201, doi:10.1029/2009JD012650. studied here underwent spreading. As a result, the contrails Jensen, E., and O. Toon (1997), The potential impact of soot particles from – presence certainly lead to further radiative effects in addition aircraft exhaust on cirrus clouds, Geophys. Res. Lett., 24(3), 249 252, doi:10.1029/96GL03235. to secondary impacts on the dynamics of the frontal system, Jensen, E., O. Toon, S. Kinne, G. Sachse, B. Anderson, K. Chan, C. Twohy, which are unaccounted for in our study. Further, it is unclear B. Gandrud, A. Heymsfield, and R. Miake-Lye (1998), Environmental if natural clouds would have formed in the absence of the conditions required for contrail formation and persistence, J. Geophys. Res., 103(D4), 3929–3936, doi:10.1029/97JD02808. contrail-induced cirrus. Kahn, B., A. Gettelman, E. Fetzer, A. Eldering, and C. Liang (2009), Cloud and clear-sky relative humidity in the upper troposphere observed by the [24] Acknowledgments. We kindly thank Steve Ghan (Pacific A-train, J. Geophys. Res., 114, D00H02, doi:10.1029/2009JD011738. Northwest National Laboratory), and three anonymous reviewers for their Kärcher, B., and F. Yu (2009), Role of aircraft soot emissions in contrail comments. The CFSR data were developed by NOAA’s National Centers formation, Geophys. Res. Lett., 36, L01804, doi:10.1029/2008GL036649. for Environmental Prediction (NCEP). Reanalysis data for this study are Kärcher, B., O. Möhler, P. DeMott, S. Pechtl, and F. Yu (2007), Insights from NOAA’s National Operational Model Archive and Distribution System into the role of soot aerosols in cirrus cloud formation, Atmos. Chem. (NOMADS), maintained at NOAA’s National Climate Data Center (NCDC). Phys., 7(16), 4203–4227, doi:10.5194/acp-7-4203-2007. The CERES and CALIPSO data were obtained from the NASA Langley Kästner, M., R. Meyer, and P. Wendling (1999), Influence of weather con- Research Center Atmospheric Sciences Data Center (ASDC). Meteosat ditions on the distribution of persistent contrails, Meteorol. Appl., 6(3), and AVHRR imagery were provided by NEODAAS/University of Dundee. 261–271, doi:10.1017/S1350482799001231. Synoptic surface pressure plot was created by the UK Met Office. Special Koop, T., B. Luo, A. Tsias, and T. Peter (2000), Water activity as the deter- thanks go to Neil Lonie (University of Dundee), Gill Compo (University minant for homogenous ice nucleation in aqueous solution, Nature, 406, of Colorado), Keith Butt (ICAO supervisor, Gander ACC) and Gabriel Pérez 611–614, doi:10.1038/35020537. (Instituto de Astrofísica de Canarias). E. Pallé and B. Laken acknowledge Lee, D., D. Fahey, P. Forster, P. Newton, R. Wit, L. Lim, B. Owen, support from the Spanish MICIIN grant CGL2009–10641. and R. Sausen (2009), Aviation and global climate change in the 21st century, Atmos. Environ., 43(22–23), 3520–3537, doi:10.1016/ j.atmosenv.2009.04.024. References Marquart, S., M. Ponater, F. Mager, and R. Sausen (2003), Future develop- Appleman, H. (1953), The formation of exhaust contrails by jet aircraft, ment of contrail cover, optical depth and radiative forcing: Impacts of Bull. Am. Metereorol. Soc., 34,14–20. increasing air traffic and climate change, J. Clim., 16, 2890–2904, Bakan, S., M. Betancor, V. Gayler, and H. Graßl (1994), Contrail frequency doi:10.1175/1520-0442(2003)016<2890:FDOCCO>2.0.CO;2. over Europe from NOAA-satellite images, Ann. Geophys., 12(10–11), Meyer, R., H. Mannstein, R. Meerkötter, U. Schumann, and P. Wendling 962–968, doi:10.1007/s00585-944-0962-y. (2002), Regional radiative forcing by line-shaped contrails derived from Burkhardt, U., and B. Kärcher (2009), Process-based simulation of contrail satellite data, J. Geophys. Res., 107(D10), 4104, doi:10.1029/ cirrus in a global climate model, J. Geophys. Res., 114, D16201, 2001JD000426. doi:10.1029/2008JD011491. Möhler, O., et al. (2006), Efficiency of the deposition mode ice nucleation Burkhardt, U., and B. Kärcher (2011), Global radiative forcing from con- on mineral dust particles, Atmos. Chem. Phys., 6(10), 3007–3021, trail cirrus, Nat. Clim. Change, 1,54–58, doi:10.1038/nclimate1068. doi:10.5194/acp-6-3007-2006. Carleton, A., D. Travis, K. Master, and S. Vezhapparambu (2008), Com- Rap, A., P. Forster, A. Jones, O. Boucher, J. Haywood, N. Bellouin, and posite atmospheric environments of jet contrail outbreaks for the United R. De Leon (2010), Parameterization of contrails in the UK Met Office Sates, J. Appl. Meteorol., 47(2), 641–667, doi:10.1175/2007JAMC1481.1. climate model, J. Geophys. Res., 115, D10205, doi:10.1029/2009JD012443. DeGrand, J., A. Carleton, D. Travis, and P. Lamb (2000), A satellite-based Saha, S., et al. (2010), The NCEP Climate Forecast System Reanalysis, Bull. climatic description of jet aircraft contrails and associations atmospheric Am. Meteorol. Soc., 91(8), 1015–1057, doi:10.1175/2010BAMS3001.1. conditions, 1977–79, J. Appl. Meteorol., 39(9), 1434–1459, doi:10.1175/ Sausen, R., et al. (2005), Aviation radiative forcing in 2000: An update on 1520-0450(2000)039<1434:ASBCDO>2.0.CO;2. IPCC (1999), Meteorol. Z., 14(4), 555–561, doi:10.1127/0941-2948/ DeMott, P., D. Cziczo, A. Prenni, D. Murphy, S. Kreidenweis, D. Thomson, 2005/0049. R. Borys, and D. Rogers (2003), Measurements of the concentration and Scorer, R. (1955), Condensation trails, Weather, 10, 281–287. composition of nuclei for cirrus formation, Proc. Natl. Acad. Sci. U. S. A., Stordal, F., G. Myhre, E. Stordal, W. Rossow, D. Lee, D. Arlander, and 100(25), 14,655–14,660, doi:10.1073/pnas.2532677100. T. Svendby (2005), Is there a trend in cirrus cloud cover due to aircraft Forster, P. et al. (2007), Changes in atmospheric constituents and in radia- traffic?, Atmos. Chem. Phys., 5, 2155–2162, doi:10.5194/acp-5-2155- tive forcing, in Climate Change 2007: The physical science basis. Contri- 2005. bution of Working Group 1 to the Fourth Assessment Report of the Wielicki, B., B. Barkstrom, E. Harrison, R. Lee, G. Smith, and J. Cooper Intergovernmental Panel on Climate Change, edited by S. Solomon (1996), Clouds and the Earth’s Radiant Energy System (CERES)—An et al., pp. 129–234, Cambridge Univ. Press, Cambridge, U. K. Earth Observing System Experiment, Bull. Am. Meteorol. Soc., 77(5), Gierens, K., U. Schumann, M. Helten, H. Smit, and A. Marenco (1999), 853–868, doi:10.1175/1520-0477(1996)077<0853:CATERE>2.0.CO;2. A distribution law for relative humidity in the upper troposphere and

7of7