SEARCH

SEARCH BY CITATION

Keywords:

  • source-receptor relationships;
  • retroplumes;
  • trajectories;
  • aircraft measurements

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Aircraft Instrumentation
  5. 3. Model Description
  6. 4. Results
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[1] In this paper we present simulations with a Lagrangian particle dispersion model to study the intercontinental transport of pollution from North America during an aircraft measurement campaign over Europe. The model was used for both the flight planning and a detailed source analysis after the campaign, which is described here with examples from two episodes. Forward calculations of emission tracers from North America, Europe, and Asia were made in order to understand the transport processes. Both episodes were preceded by stagnant conditions over North America, leading to the accumulation of pollutants in the North American boundary layer. Both anthropogenic sources and, to a lesser extent, forest fire emissions contributed to this pollution, which was then exported by warm conveyor belts to the middle and upper troposphere, where it was transported rapidly to Europe. Concentrations of many trace gases (CO, NOy, CO2, acetone, and several volatile organic compounds; O3 in one case) and of ambient atmospheric ions measured aboard the research aircraft were clearly enhanced in the pollution plumes compared to the conditions outside the plumes. Backward simulations with the particle model were introduced as an indispensable tool for a more detailed analysis of the plume's source region. They make trajectory analyses (which, to date, were mainly used to interpret aircraft measurement data) obsolete. Using an emission inventory, we could decompose the tracer mixing ratios at the receptors (i.e., along the flight tracks) into contributions from every grid cell of the inventory. For both plumes we found that emission sources contributing to the tracer concentrations over Europe were distributed over large areas in North America. In one case, sources in California, Texas, and Florida contributed almost equally, and smaller contributions were also made by other sources located between the Yucatan Peninsula and Canada. In the other case, sources in eastern North America, including moderate contributions from forest fires, were most important. The plume's maximum was mainly caused by anthropogenic emissions from the New York area. To our knowledge, this is the first case reported where a pollution plume from a megacity was reliably detected over another continent.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Aircraft Instrumentation
  5. 3. Model Description
  6. 4. Results
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[2] For a long time it was thought that intercontinental transport (ICT) of air pollution has only a minor effect on the chemical composition of the atmosphere over Europe. At Mace Head, Ireland, for instance, pollution episodes caused by transport from North America are rare events, and corresponding ozone (O3) enhancements are marginal [Jennings et al., 1996; Ryall et al., 1998], prompting Derwent et al. [1998] to conclude “that either the North American O3 and CO plume does not intersect the European coastline over Mace Head or that this plume had become merged into the Northern Hemisphere background in transit” (p. 153). Similar difficulties in the definite detection of air pollution plumes from North America have been faced at other European stations, for example, at Porspoder on the French Atlantic coast [Dutot et al., 1997; Fenneteaux et al., 1999] or at Izaña on Tenerife to the west of Africa [Schmitt, 1994]. On the other hand, strong pollution signals were measured over the western North Atlantic [Parrish et al., 1993] under conditions of continental outflow from North America, and smaller signals were also found in the central North Atlantic [Parrish et al., 1998]. In the meantime, these contradictions are resolved, as it is now known that ICT takes place mainly in the upper troposphere [Stohl and Trickl, 1999; Yienger et al., 2000; Stohl et al., 2002a] after surface emissions have been transported upward in warm conveyor belts (WCBs) [Stohl, 2001] and/or by deep convection [Prados et al., 1999]. This is true for ICT of both Asian and North American emissions, but we focus here on ICT from North America to Europe. Note that European emissions behave differently, as they tend to remain in the lower troposphere for a longer time [Stohl et al., 2002a].

[3] While the major impact of ICT is certainly in the upper troposphere, a recent modeling study [Li et al., 2002] suggests that surface O3 concentrations in Europe are enhanced, too, by 2–4 ppbv on average in summer through photochemical O3 formation from North American emissions. These authors estimate that 20% of the violations of the European Council O3 standard (55 pbbv, 8-hour average) would not have occurred in the absence of anthropogenic emissions from North America. Similar results were obtained by Wild and Akimoto [2001]. However, in order to be observable at the surface, a pollution plume must either travel all the way at low altitudes or it must descend from higher altitudes. Both processes are relatively slow compared to direct transport at high altitudes, and the chemical environment in the lower troposphere, where the lifetimes of many substances are shorter than they would be higher in the troposphere, is less favorable for long-range transport [Wild et al., 1996; Schultz et al., 1998]. Furthermore, mixing with surrounding air masses is stronger in the atmospheric boundary layer (ABL) than in the free troposphere. All these circumstances make it difficult to observe the effects of ICT at the surface.

[4] While the meteorological conditions favoring fast ICT are relatively well understood [Stohl and Trickl, 1999; Stohl, 2001], the chemical changes occurring in the process are much less well characterized. One reason for this is that transport to the upper troposphere takes place mostly in clouds, leading to washout of water-soluble gases and aerosols, a process which is still poorly parameterized in chemistry transport models (CTMs). Thus simulations of ICT with these models are difficult. Another reason is the lack of suitable measurement data to constrain and validate the models. Because ICT can be observed best in the middle and upper troposphere, aircraft measurements at high altitudes are required (except for O3, for which lidar and sonde measurements are available), which are quite expensive. Several aircraft campaigns have studied continental pollution outflow, for example, the Southern African Regional Science Initiative (SAFARI) and the Transport and Atmospheric Chemistry Near the Equator-Atlantic (TRACE-A) campaign [Andreae et al., 1996], the Transport and Atmospheric Chemistry Near the Equator-Pacific (TRACE-P) campaign, the Indian Ocean Experiment (INDOEX) [Lelieveld et al., 2001], the Biomass Burning and Lightning Experiment (BIBLE) [Kondo, 2000], and the North Atlantic Regional Experiment (NARE) [Fehsenfeld et al., 1996]. However, in the past, aircraft measurement campaigns were not designed specifically for studying ICT, and thus observations of pollution plumes from upwind continents were coincidental and quite rare and sometimes possibly not recognized as such. A notable exception are the airborne measurements by Arnold et al. [1997] north of Ireland at an altitude of 9000 m in a plume originating from North America. They found concentrations of both sulfur dioxide and acetone of up to 3 ppb along with enhanced levels of condensation nuclei and carbon dioxide. At such high concentrations, acetone, whose chemical budget is still poorly constrained [Jacob et al., 2002], is probably the most important source of HOx radicals in the upper troposphere, while sulfur dioxide reacts with OH, leading to the condensable trace gas sulfuric acid. The latter promotes aerosol particle formation and growth.

[5] Convective Transport of Trace Gases into the Upper Troposphere Over Europe: Budget and Impact on Chemistry (CONTRACE) is a project that investigates the role of different convective transport mechanisms for the trace gas budget in the upper troposphere. One aim of CONTRACE is to study the outflow of WCBs, which sometimes transport polluted air from the North American ABL to the free troposphere over Europe. In November 2001 the first CONTRACE field experiment was carried out from the Deutsches Zentrum für Luft- und Raumfahrt (DLR) operation site in Oberpfaffenhofen (48°N, 11°E in southern Germany). The DLR research aircraft Falcon was used as a platform for the in situ measurements of a large variety of species and was directed into pollution plumes from North America. Seven flights were undertaken, four of which intercepted pollution plumes from North America while the other three studied European air pollution. This paper describes the modeling efforts undertaken during CONTRACE to trace the plumes back to their sources. Backward modeling with a Lagrangian particle dispersion model is introduced as a tool for this purpose, making back trajectory analyses (which, to date, were mostly used to interpret aircraft measurements) obsolete. In section 2, we briefly describe the aircraft's instrumentation. In section 3, we describe the model simulations and the backward modeling approach. Results for three flights, during which pollution clouds originating from North America were observed, are presented in section 4. Finally, conclusions are drawn in section 5. A companion paper (H. Huntrieser et al., manuscript in preparation, 2003) presents a detailed characterization of the chemical measurements taken within the North American pollution plumes.

2. Aircraft Instrumentation

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Aircraft Instrumentation
  5. 3. Model Description
  6. 4. Results
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[6] The airborne measurements during CONTRACE were performed with the DLR research aircraft Falcon, operating mainly in the middle and upper troposphere (4–10 km). A variety of chemical tracers for polluted air masses like CO, CO2, O3, NO, NOy, and ions were measured with high temporal resolution (1–4 s). The chemical instrumentation (technique, detection limit, sampling rate) onboard the Falcon is presented in Table 1. For a more precise description of the instruments, which have been used in several previous field experiments, see Schlager et al. [1997], Schulte et al. [1997], Feigl [1998], Gerbig et al. [1996, 1999], Huntrieser et al. [1998], Schröder et al. [1998], Ziereis et al. [1999], Arnold et al. [1999], and Eichkorn et al. [2002]. In addition, two J(NO2) radiometers (upward-looking and downward-looking sensors) were included on the Falcon. The radiometers measure the solar actinic flux in the 300–380 nm spectral band. Corrections applied to the J(NO2) signals (due to spectral and angular response, altitude, solar zenith angle, and temperature) are described by Volz-Thomas et al. [1996], Beine et al. [1999], and Shetter and Müller [1999]. Position, altitude, temperature, humidity, pressure, and wind (u, v, w) were measured with the standard Falcon meteorological measurement system [see Schumann et al., 1995, 2000]. Other measurements (e.g., of acetone, sulfur dioxide, and volatile organic compounds) were performed, too, but will be discussed elsewhere (H. Huntrieser et al., manuscript in preparation, 2003).

Table 1. Overview of the Falcon Instrumentation
SpeciesTechniqueaDetection LimitSampling Rate, sAccuracy, %Reference
  • a

    CL, chemiluminescence; NDIR, nondispersive infrared.

NOCL, ppt518Schlager et al. [1997], Feigl [1998], Ziereis et al. [1999]
NOyAu-converter plus CL, ppt15112Feigl [1998]
O3UV absorption, ppb145Schlager et al. [1997], Huntrieser et al. [1998]
COVUV fluorescence, ppb2110Gerbig et al. [1996, 1999]
CO2NDIR, ppm0.110.1Schulte et al. [1997]
J(NO2)filter radiometry, s−11E-4117Volz-Thomas et al. [1996]
Ionsmass spectrometer, ppq1E-17 (0.01)315Arnold et al. [1999], Eichkorn et al. [2002]

3. Model Description

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Aircraft Instrumentation
  5. 3. Model Description
  6. 4. Results
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[7] To simulate transport processes, we employed the Lagrangian particle dispersion model FLEXPART (version 4.4) [Stohl et al., 1998; Stohl and Thomson, 1999] (information on FLEXPART can be obtained via the internet at http://www.forst.tu-muenchen.de/EXT/LST/METEO/stohl/). FLEXPART was validated with data from three large-scale tracer experiments in North America and Europe [Stohl et al., 1998], and it was used previously in case studies of ICT [Stohl and Trickl, 1999; Forster et al., 2001; Spichtinger et al., 2001] and to simulate intrusions of stratospheric air into the lower troposphere [Stohl et al., 2000]. It was also employed to develop a 1-year “climatology” of ICT [Stohl et al., 2002a].

[8] FLEXPART treats advection and turbulent diffusion by calculating the trajectories of a multitude of particles. Stochastic fluctuations, obtained by solving Langevin equations [Stohl and Thomson, 1999], are superimposed on the grid-scale winds from global meteorological data sets to represent transport by turbulent eddies, which are not resolved. Global data sets also do not resolve individual convective cells, although they reproduce the large-scale effects of convection (e.g., the strong ascent within WCBs). To account for sub-grid-scale convective transport, FLEXPART was recently equipped with the convection scheme developed by Emanuel and ŽivkovićRothman [1999]. Its implementation is described by Seibert et al. [2001].

[9] FLEXPART can be driven either with model-level data from the European Centre for Medium-Range Weather Forecasts (ECMWF) [European Centre for Medium-Range Weather Forecasts, 1995] or with pressure-level data from the Aviation Model (AVN) of the National Center for Environmental Prediction. During the CONTRACE campaign we produced tracer forecasts using AVN data (resolution 1°, 26 vertical levels) to support the flight planning. Furthermore, trajectory-based WCB forecasts using the FLEXTRA model [Stohl et al., 1995], meteorological forecasts (e.g., maps of geopotential, tropopause heights, etc.), and satellite imagery were provided that were all organized on a common web page. All our forecast products were updated every 6 hours and were available within <12 hours from the nominal start of the AVN forecast, including the time required for the AVN forecast itself, the transfer of the AVN data to our institute, the FLEXTRA and FLEXPART computations, and the production of the graphical output. The airborne in situ measurements show that the forecasts very successfully guided the Falcon into the polluted layers (H. Huntrieser et al., manuscript in preparation, 2003), although the convection scheme was not used for the forecast runs. Our forecast system also supported flight planning during the recent Intercontinental Transport and Chemical Transformation 2002 (ITCT 2k2) campaign, which took place over the eastern Pacific Ocean. A detailed description and evaluation of our forecasts during this experiment was prepared by C. Forster et al. (Lagrangian transport model forecasts and a transport climatology for the Intercontinental Transport and Chemical Transformation 2002 (ITCT 2k2) measurement campaign, submitted to Journal of Geophysical Research, 2003). During CONTRACE, other institutes provided forecasts with a global CTM [Lawrence et al., 2003] and with a mesoscale CTM [Jakobs et al., 1995], too.

[10] In this paper we present a postanalysis of the CONTRACE flights; forecasts are not considered here. For this we used global ECMWF data with a horizontal resolution of 1°, 60 vertical levels, and a time resolution of 3 h (analyses at 0000, 0600, 1200, and 1800 UTC; 3-hour forecasts at 0300, 0900, 1500, and 2100 UTC). Data with 0.5° resolution covering the domain 120°W–30°E and 18°N–66°N were nested into the global data in order to achieve higher resolution over the region of main interest, i.e., North America, the North Atlantic, and Europe.

3.1. Forward Simulations

[11] Forward simulations are useful for visualizing the dispersion of emission tracers and for understanding its relation to the synoptic situation. We calculated the transport of six tracers, representing anthropogenic emissions of two species (carbon monoxide (CO) and nitrogen oxides (NOx)) from three continents (North America, Europe, and Asia). We employed two emission data sets: For Europe we used the inventory of the Co-operative Programme for Monitoring and Evaluation of the Long-Range Transmission of Air Pollutants in Europe (base year 1999, resolution 50 km, see http://www.emep.int), and for outside Europe we used the EDGAR Version 3.2 inventory [Olivier and Berdowski, 2001] (base year 1995, 1° resolution). All six tracers behave purely passively in the model; that is, no chemical processes are simulated. The simulations started on 28 October, allowing 17 days spin-up until the first CONTRACE flight on 14 November, and ended on 28 November 2001. During this period, particles were released between the surface and 150 m above ground at a constant rate according to the emission distribution. The total number of particles used was chosen as 25, 15, and 8.5 million for North America, Europe, and Asia, respectively. Relatively few particles were used for Asian emissions, as these are least important for CONTRACE. More particles were required for North American emissions than for European ones because North American tracer plumes can be relatively diluted when they reach Europe. The number of particles released in a particular grid cell was determined by scaling the emissions in that cell relative to the total emissions from the respective continent. Each particle carried the emitted masses of both CO and NOx and was flagged with its release time so that age spectra of the tracers could be created. Particles were dropped from the simulation after 20 days. Since the lifetime of CO is on the order of a month or longer, our simulations do not account for background CO concentrations but only for enhancements over the CO background due to emissions during the last 20 days.

[12] A seventh tracer run was made using a special version of FLEXPART as described by Stohl et al. [2000]. O3 was initialized in the stratosphere according to a statistical relationship with potential vorticity at the start of the model run and at the inflowing boundary of the simulation domain that, in this case, was limited to 80°W–40°E and 25°N–88°N. O3 was carried as a passive tracer in order to identify portions of the flight tracks that were influenced by transport from the stratosphere. Tracer concentrations were determined every hour on a three-dimensional grid with 500 m vertical resolution and with horizontal resolutions of 0.5° for the Europe tracers, 1° for the North America tracers, and 2° for the Asia and stratospheric O3 tracers.

3.2. Backward Simulations

[13] As long as we consider only linear processes, it is possible to reverse a dispersion problem simply by running a Lagrangian particle dispersion model backward in time (see Flesch et al. [1995] for a proof). In analogy to forward simulations we may call the particle cloud produced by a backward simulation a “retroplume.” The theory of backward simulations, as we apply it here, was described by Seibert and Stohl [1999], Seibert [2001], and P. Seibert and A. Frank (Source-receptor matrix calculation with a Lagrangian particle dispersion model in backward mode, submitted to Atmospheric Chemistry and Physics, 2003, hereinafter referred to as Seibert and Frank, submitted manuscript, 2003). Seibert and Frank (submitted manuscript, 2003) present a detailed validation of the method using both idealized and realistic test cases, and they compare results from forward and backward simulations. Essentially, a large number of particles is released at the receptor (i.e., at a point located at the flight track in our case) and transported backward in time. Then the residence time of all particles, normalized by the total number of released particles, is determined on a uniform grid. The residence time in a particular grid cell is proportional to the contribution a source with unit strength in this cell would make to the mixing ratio at the receptor. By multiplying the residence time with a tracer's actual source strength (in ppbm s−1) in the respective grid cell (assuming instantaneous mixing of the emitted tracer within that cell), we obtain the actual contribution to the mixing ratio at the receptor (in ppbm) from this grid cell. Summing up the contributions from all grid cells finally gives the total mixing ratio at the receptor. A flowchart comparing forward and backward simulations with FLEXPART is presented in Figure 1. Note that the last step, the multiplication with actual source strengths, is done a posteriori, allowing different emission scenarios to be considered without having to rerun FLEXPART.

image

Figure 1. Flowchart comparing the basic steps in the forward and backward model simulations.

Download figure to PowerPoint

[14] As we are concerned here with surface emissions only, we concentrate on a shallow layer adjacent to the ground (for practical reasons we use a layer of 300 m thickness, which is within the ABL most of the time but which is still high enough to allow a sufficiently large number of particles to be sampled), the residence times in which we call the footprint. We take emission source strengths from the EDGAR emission inventory and assume that the emissions mix instantaneously into the footprint layer. Our footprint concept is, except for the numerical implementation, equivalent to that of Flesch et al. [1995], where particles take up emissions upon touching the ground. Note that it would be possible to also consider sinks (Seibert and Frank, submitted manuscript, 2003), but we focus here on CO, which has a lifetime longer than the transport times considered. Thus CO is simulated as a passive tracer like in the forward simulations.

[15] The residence time fields could be used conveniently for inverse modeling using the measurement data, where the strengths and the distributions of the sources can be quantitatively (albeit possibly with large uncertainties) determined without prior knowledge of the emissions [Seibert, 2001]. However, in this paper we do not go that far because anthropogenic CO emissions are known relatively well. Improving the emission inventory is thus not an aim of this study. Backward simulations, nevertheless, have four major advantages over the forward ones: First, higher resolution can be achieved. In the forward case we have to “fly” through the model output and sample the tracer along the flight tracks. Unfortunately, the resolution of the model output is limited by both computation time (more particles are needed for higher resolution) and hard disk storage space. In the backward case we released 20,000 particles from each of a number of small boxes that were created along the flight track whenever the aircraft had covered a distance of >0.1° or changed its altitude by >200 m. Depending on the length and geometry of a flight, between 85 and 271 boxes were produced that way, yielding a much higher space and time resolution than did the corresponding forward simulation.

[16] The second advantage is that our backward simulations can replace traditional backward trajectory calculations. Unlike trajectories, which account for advection only, our simulations also consider turbulence, convection, and the filamentation of the initial sampling volume by the large-scale advection. Stohl et al. [2002b] have shown recently that neglecting these processes can lead to errors in source-receptor relationships that are larger than the trajectory errors (i.e., errors resulting from errors in the wind fields, interpolation, and truncation) themselves. Maps of the residence times at various altitudes and, especially, of the footprint can replace trajectory plots for identifying potential source regions.

[17] Third, if the emission distribution is already known, maps of the source contribution to the total mixing ratio at the receptor can be constructed. To achieve this in a forward simulation, a large number of tracers (one per grid cell) would be required, which is not practical. Furthermore, alternative emission inventories can be used and different geographical definitions for tracers can be applied a posteriori without requiring a rerun of the dispersion model.

[18] Fourth, the backward runs are computationally more efficient than the forward simulations.

4. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Aircraft Instrumentation
  5. 3. Model Description
  6. 4. Results
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

4.1. Flights on 19 November 2001

[19] A pollution plume from North America was forecast to arrive over Europe on 19 November 2001. The forecast suggested that it would first reach southern Scandinavia and then travel toward central Europe. Therefore the Falcon flew out of its base in Oberpfaffenhofen heading north along 10°E to 60°N (approximately the location of Oslo) and then along 60°N eastward to Stockholm (18°E), where it was refueled. It returned to Oberpfaffenhofen on a more easterly route (see Figure 3j). Figure 2 shows some of the measurements taken during these two flights along with the age spectra of the three CO tracers and the stratospheric O3 tracer, obtained by “flying” through the forward model output. FLEXPART suggested relatively strong and recent (age of 0–2 days) stratospheric influence close to Oberpfaffenhofen at ∼10 km altitude, which is partly confirmed by the measurements (Figure 2b). Relative humidity was <20% in these flight sections, and at 1020 UTC an O3 peak of 94 ppb was observed (Figure 2a). However, the O3 mixing ratios were much lower than predicted by the model, which is likely due to the inaccurate positioning of the tropopause in the coarse-resolution model output.

image

Figure 2. Age spectra of three CO tracers (from Asia, Europe, and North America) and a stratospheric O3 tracer obtained from forward model simulations (colored bars) and aircraft measurement data (lines) for the flights on 19 November 2001: (a) measured O3 (black line) and relative humidity (red line); (b) stratospheric O3 tracer and aircraft altitude (black line); (c) Asia CO tracer, measured NOy (black line), and measured NO (red line); (d) Europe CO tracer and measured CO2 (black line); and (e) North America CO tracer and measured CO (black line).

Download figure to PowerPoint

[20] The FLEXPART Asia CO tracer (Figure 2c) shows mixing ratios below 8 ppb at ages of 10 days and older, suggesting only a weak influence of Asian emissions, except possibly for the “background” CO concentrations. The Europe CO tracer (Figure 2d) is zero all the time, except for short periods during takeoff and landing. In contrast, North America CO tracer shows four peaks at ∼1200, 1300, 1430, and 1615 UTC (Figure 2e), with an age of 3–10 days. At all these times, measured CO increased, too, from background levels of ∼80 ppb to 110 ppb. A spike of 170 ppb CO was observed at ∼1150 UTC, when, also, the FLEXPART CO tracer shows a maximum of 110 ppb. Assuming a background of 80 ppb, FLEXPART overestimates the CO increases in the plume somewhat, but the plume locations are well predicted. CO2 measurements suffered from some contamination with cabin air during descents of the aircraft and thus must be interpreted with caution. They nevertheless show a clear correlation with CO in the plume sections of the flights (Figure 2d). NOy (Figure 2c) shows values of a few hundred ppt in the plume and a spike of >1 ppb at the time of the CO spike. NO remains at very low levels, implying that NOx has been oxidized already to other species of the NOy family and thus a plume age of a few days (the lifetime of NOx in the free troposphere) at least. Other NOy spikes of very short duration are due to occasional encounters of fresh aircraft exhaust, especially at higher altitudes. Note that these spikes are uncorrelated with CO but normally go along with NO enhancements.

[21] Acetone (not shown) has maxima of almost 5 ppb in the plume, compared to values of ∼1.3 ppb outside the plume, and concentrations of sulfur dioxide (not shown) are enhanced, too. Analyses of flask air samples taken during the flights also reveal maxima of several volatile organic compounds (e.g., benzene and toluene) in the plume (not shown). O3 is 5–10 ppb higher within the plume than outside and is positively correlated with CO in the plume, which suggests photochemical O3 formation. Relative humidity is variable in the plume with maxima up to 80% (e.g., at 1300 UTC), indicative of recent uplifting of the air mass.

[22] A detailed chemical characterization of the plume is presented elsewhere (H. Huntrieser et al., manuscript in preparation, 2003). Here we concentrate on transport aspects and on determining the plume's source region. On 8 November a large anticyclone with a core sea level pressure of ∼1035 hPa formed over the central United States, leading to the accumulation of pollutants underneath its subsidence inversion. During the following days the anticyclone moved southward, opening a transport route from the West Coast to the east along its northern edge. At the same time, pollutants from the southern United States (Houston, Texas, for instance) could accumulate in the center of the anticyclone. On 12 November the anticyclone moved to the Great Lakes region, and on 13 November it pushed farther east as shown in Figure 3a, which presents a sea level pressure map and frontal analysis combined with a GOES-East infrared satellite image. On the anticyclone's backside the pollution that had accumulated during the previous days travelled northward to the Great Lakes region (see the “river” of CO tracer stretching from Houston to the Great Lakes in Figure 3f). This occurred largely under cloud-free conditions. On 15 November the anticyclone's center pulled offshore while a cyclone formed over Hudson Bay and Newfoundland (Figure 3b), channelling the flow in between the two features toward the northeast. This allowed the air to take up more emissions while travelling along the eastern seaboard (Figure 3g). O3 measurements at Horton station in the Appalachian Mountains clearly confirm the arrival of this pollution front by an increase in O3 from ∼40 ppb to 60 ppb during the night of 14–15 November (H. Huntrieser et al., manuscript in preparation, 2003). At that time the plume was still located in a mostly cloud-free environment, and thus the simulated CO remained in the lower troposphere. As the cyclone intensified over Newfoundland and moved offshore on 17 November (Figure 3c), its cold front pushed the (formerly) warm sector air out into the North Atlantic (Figure 3h). The CO tracer was contained in the cold frontal cloud band (the WCB) [Browning, 1990], which transported it into the middle and upper troposphere. This is exactly the type of continental outflow situation studied extensively during the NARE campaigns [e.g., Fehsenfeld et al., 1996; Cooper et al., 2001, 2002].

image

Figure 3. (left) Infrared satellite images from GOES-East plus isolines of sea level pressure from Aviation Model analyses (black) and surface fronts for (a) 13 November 2001 0000 UTC, (b) 15 November 2001 0000 UTC, and (c) 17 November 2001 0000 UTC and combined GOES-East/Meteosat infrared images plus geopotential at 500 hPa for (d) 18 November 2001 1200 UTC and (e) 19 November 2001 1200 UTC. No satellite data are available for white areas in the northern parts of the images. (right) Total columns of the North America CO tracer on (f) 13 November 2001 0000 UTC, (g) 15 November 2001 0000 UTC, (h) 17 November 2001 0000 UTC, (i) 18 November 2001 1200 UTC, and (j) 19 November 2001 1200 UTC. The aircraft's flight legs are shown in Figure 3j.

Download figure to PowerPoint

[23] To visualize the WCB, we started 5-day forward trajectories (based on ECMWF data) from a regular grid with 1° resolution in the domain 120°W–60°W, 30°N–60°N at an altitude of 500 m above the ground on 14 November at 1200 UTC. Trajectories ascending >5000 m during the last 4 days are plotted in Figure 4. Such a criterion isolating ascending trajectories is effective in displaying WCB-like flow features [Wernli and Davies, 1997; Stohl, 2001]. The trajectories first travelled at low altitudes over high-emission regions (compare with Figure 3) and then, upon export from North America, ascended in the WCB to altitudes of ∼6–8 km. Many of the trajectories arrived in the middle troposphere over Scandinavia and northern Germany. The WCB can also be seen nicely in the combined GOES-East/Meteosat infrared satellite image on 18 November at 1200 UTC (Figure 3d). At that time the WCB cloud band was located over and south of Greenland. Note the similarity of the structures in the satellite image and the trajectory plot despite the fact that the trajectories integrate over a 5-day period, whereas the satellite image is instantaneous.

image

Figure 4. Five-day forward trajectories starting from a regular grid (120°W–60°W, 30°N–60°N) with 1° resolution at an altitude of 500 m above ground level on 14 November 1200 UTC. Only trajectories that ascend >5000 m during the last 4 days are plotted. The trajectories' colors, coded as in the label bar, refer to their actual altitude.

Download figure to PowerPoint

[24] A comparison of the CO tracer plume (Figure 3i) with the satellite image (Figure 3d) on 18 November shows that most of the tracer was contained in the northernmost section of the WCB, in the cyclone's occluded part. The cold conveyor belt [cf. Cooper et al., 2002] at the west coast of Greenland transported tracer, as well. South of Greenland the WCB contained only moderate amounts of CO tracer. The reason for this is that the tracer that had accumulated over North America over several days was swept out to the North Atlantic as soon as the outflow began (compare with Figures 3c and 3h). Outflow conditions prevailed for some time, and thus tracer emitted later into the outflowing air was contained in the southern part of the WCB but at lower concentrations, as it had less time to accumulate over the source region. One day later, the pollution plume reached Europe (Figure 3j) at the forward flank of a ridge centered over Great Britain. At that time the plume's center was located already in the outflow of the WCB, ahead of the clouds (Figure 3e), where remaining cloud droplets evaporated as the air descended. Thus the “WCB processing” was finished before the pollution plume arrived over Europe.

[25] Figure 5 shows vertical sections through the FLEXPART output, oriented in approximately meridional direction along the flight tracks, at about the times of the two flights. On the northbound flight the pollution plume was missed upon ascent from Oberpfaffenhofen (at the time of departure the plume was even farther to the north than shown in Figure 5, which is valid for 1200–1300 UTC). The aircraft then flew above the plume, but it passed through the plume's maximum during the first stepwise descent/ascent close to Oslo at ∼1200 UTC. It then climbed out of the plume but crossed it once more when it descended into Stockholm at 1300 UTC.

image

Figure 5. Vertical cross sections cutting through the North America CO tracer field (a) along 10°E at 1200–1300 UTC and (b) along a straight line connecting Stockholm and Oberpfaffenhofen at 1400–1500 UTC on 19 November 2001. The bold lines are the vertical profiles of the Falcon on the flights from Oberpfaffenhofen to Stockholm and on the return flight, respectively.

Download figure to PowerPoint

[26] On the return flight the pollution plume was encountered twice: first, during the stepwise ascent from Stockholm at 1430 UTC and second, upon the descent into Oberpfaffenhofen at 1615 UTC. Note the plume's southward shift during the return flight (Figure 5b), although Figures 5a and 5b are separated by only 2 hours. Between 1000 and 1700 UTC the plume's tip actually travelled >3° of latitude southward, explaining why the plume was encountered close to Oberpfaffenhofen during the return flight only.

[27] Now, from where in North America did the tracer plume originate? To answer this question, we made a detailed source analysis with backward simulations starting from along the flight tracks. We start with a single event, namely, the tracer maximum at 1153 UTC (see Figure 2). Twenty thousand particles were released from a small box (0.005° × 0.035° × 200 m) around the aircraft position within a 27-s time interval. Figure 6 shows the particle residence times for this event both in the whole atmospheric column and in the footprint layer. Initially, the retroplume remains compact, but over Greenland the particle cloud starts to disperse (backward in time), and once it reaches the North American east coast, it grows rapidly. This is where the WCB's ascent began (compare with Figure 4). Note that it is typical that WCBs suck up air from a large region because convergence in the ABL is required to feed their inflow. The retroplume covers large parts of North America, but its footprint residence times are longest over the eastern United States, where, also, the anthropogenic emissions are greatest. The scattered small maxima over Europe and Africa are due to the retroplume circling the globe within the 20 days considered.

image

Figure 6. Residence time distribution for the particles arriving in a small receptor box (0.005° × 0.035° ×200 m) on 19 November at 1153 UTC within a 27-s time interval (a) in the whole atmospheric column and (b) in the footprint layer. Values are given as percentages of the maximum residence time, indicated below each panel. Residence times were calculated from arrival until 20 days back.

Download figure to PowerPoint

[28] Multiplying the footprint residence times with the CO emission strengths from the EDGAR inventory yields the actual source contribution per 0.5° grid box (Figure 7) (note that the emissions are available only with 1° resolution) to the total CO tracer mixing ratio of 130 ppb (compare with Figure 8). Emissions in large regions contributed to the simulated peak in CO tracer mixing ratio, but the largest contributions came from a rather small region on the east coast, around New York.

image

Figure 7. The source contribution per 0.5° grid box to the total CO tracer mixing ratio for a small receptor box (0.005° × 0.035° × 200 m) along the flight leg and a 27-s time interval ending on 19 November at 1153 UTC.

Download figure to PowerPoint

image

Figure 8. Age spectra of North America CO and NOx tracers obtained from backward model simulations (colored bars) and superimposed aircraft measurement data (lines) for the flights on 19 November 2001: (a) North America NOx tracer, measured NOy (black line), and NO (blue line); and (b) North America CO tracer, measured CO (black line), and aircraft altitude (blue line, relative units, maximum 9400 m).

Download figure to PowerPoint

[29] The patterns of both the footprint and the source contributions are quite similar for the different plume encounters (not shown) but with smaller absolute values than shown in Figures 6 and 7. The footprint residence times over New York especially are much smaller than for the tracer maximum, confirming that it is this region that caused the tracer maximum within the plume. This may be the first case reported where a megacity plume was detected over another continent.

[30] The source contribution for the last plume encounter close to Oberpfaffenhofen, where the aircraft crossed the plume's leading edge (see Figure 5), is somewhat different from the others. Here the contribution from sources farther west in North America is significantly larger, and sources around the Great Lakes, Houston, and Los Angeles, California, contribute almost as much as those on the east coast.

[31] Figure 8 shows the North America CO and NOx tracers obtained from all the backward simulations for the flights on 19 November. They show many fine-scale details that are missing in the corresponding forward simulation (compare with Figure 2d). Note, for instance, the CO tracer spikes of 130 ppb at 1153 UTC and 65 ppb at 1615 UTC, which are both almost hidden in Figure 8b by simultaneous spikes in measured CO.

[32] The NOy data are also in very good agreement with the NOx tracer (Figure 8a), except that they are more than 1 order of magnitude smaller. This suggests that >90% of the NOy has been removed from the atmosphere during the transport, most likely due to washout. Using a correlation analysis of NOy and CO measurements within the plume (H. Huntrieser et al., manuscript in preparation, 2003), we obtain almost the same fraction of NOy removal. An analysis by Stohl et al. [2002c], based on aircraft data from the western Atlantic, has revealed similar fractions of NOy removal, which occurred within, or upon export from, the North American ABL. The present data show that the remaining NOy can be transported over intercontinental distances.

[33] FLEXPART underestimates the duration of the plume encounter at ∼1200 UTC and possibly also at 1600 UTC, and there is a time period at ∼1500 UTC with moderately enhanced CO that does not correspond to any CO tracer. The first underestimate can be due to the simulated plume being positioned too low in the model by as little as ∼100 m at the beginning and 300 m at the end of the plume encounter. These very small errors may, in fact, even be explained by inaccuracies in aircraft altitude and/or model topography. However, an alternative explanation is that the additional CO enhancements were caused by forest fire emissions.

[34] Extensive forest fires burned in eastern North America, especially in South Carolina, Tennessee, and Georgia, during the stagnant period preceding the pollution export from North America. This high fire activity is unusual for the season and was caused by the prevailing extremely dry conditions. Following our previous work [Forster et al., 2001; Spichtinger et al., 2001], we used the fire reports from the National Interagency Fire Center (http://www.cidi.org/wildfire/index.html) to create a daily inventory of fire CO emissions for November 2001. Figure 9 shows the forest fire emission boxes that we used for the FLEXPART simulation. For each of these boxes the area burned is known on a daily basis. Total fire activity peaked from 8 to 13 November (see Table 2) during the period when pollution could accumulate over eastern North America. Burned areas were converted to CO emissions using an emission factor of 4500 kg CO hectare−1 [Cofer et al., 1998; Andreae and Merlet, 2001].

image

Figure 9. Forest fire emission boxes used for the FLEXPART simulations, shown on a map of the eastern United States.

Download figure to PowerPoint

Table 2. Areas Burned by Forest Fires in the Eastern United States in November 2001a
DayArea, ha
13625
26790
31728
41934
57100
61974
72279
811,098
910,728
107533
112153
121048
1313,532
146438
151058
163565
176210
181047
196983
204106
211099
22100
23103
24132

[35] One difficulty in simulating the transport of forest fire emissions is that the effective source height of the emissions can vary considerably depending on the burning phase (e.g., crown fires versus smoldering fires) [Lavoué et al., 2000]. Therefore we combined the forest fire emission inventory with the results of the backward simulations and tested release heights of 0–1 km, 1–2 km, 2–3 km, and 3–5 km. The forest fire CO tracer mixing ratios obtained are all rather low, on the order of a few ppb, irrespective of the release height. With a 0- to 1-km source height the maxima of the forest fire CO tracer (∼10 ppb) tend to be colocated with the maxima of the anthropogenic North America CO tracer along the flight tracks, but the plumes simulated along the flight track are broader (mostly because the forest fire emissions are found in layers extending higher up than the anthropogenic emissions). With source heights >2 km the forest fire CO tracer shows even broader plumes along the flight tracks, and its mixing ratios are very low (maxima ∼4–6 ppb). Figure 10 thus shows the results obtained with a 1- to 2-km source height, which is intermediate between these two extremes and which explains best the missing CO. Given this source height, the forest fire CO tracer arrives over Europe in a layer above the anthropogenic North America CO tracer and thus can explain some of the missing CO before 1200 UTC, at 1230 UTC, and around 1500 UTC. Simulated mixing ratios are very low, but they could be higher if the emission factor was underestimated or if the daily burned area information is inaccurate. In summary, North American forest fire emissions seem to have had a moderate, but not entirely negligible, influence on the CONTRACE measurements. Forest fire emissions have been mixed into the anthropogenic plumes, but over Europe their contribution was most likely largest in layers immediately above the anthropogenic pollution.

image

Figure 10. Age spectra of the forest fire CO tracer obtained from backward model simulations (colored bars) and superimposed aircraft measurement data of CO (black line) for the flights on 19 November 2001.

Download figure to PowerPoint

4.2. Flight on 27 November 2001

[36] On 21 November a cyclone started to form over central North America, terminating a period of stagnant, anticyclonic weather conditions. The cyclone, whose center moved to the Great Lakes region during the next days, rapidly intensified at the surface, and the trough's base at 500 hPa extended to the Gulf of Mexico. The cloud bands associated with the cyclone's fronts were prominent features in the satellite imagery (not shown). The system caused severe weather, including the formation of a series of tornadoes in Arkansas, Mississippi, and Alabama on 24 November.

[37] The WCB associated with the cold frontal cloud band pushed pollution that had accumulated in the North American ABL out over the Atlantic Ocean and into the upper troposphere on 24–26 November. The plume reached western Europe on 27 November, as seen in an image sequence of the North America CO tracer in Figure 11. The CO tracer patterns were correlated with the cloud structures in the satellite images and thus the pollution was likely contained in clouds all the time. It was also detected in the cloudy parts of the Falcon flight on 27 November. On this day the Falcon flew from Oberpfaffenhofen eastward to Vienna then westward to Brussels, from where it returned to its base (the flight leg is shown in Figure 11d). The first part of the flight was scheduled to explore pollution of European origin (which was only partly successful), while the second part was aimed at the North American pollution plume. Figure 12 shows a vertical section through the CO tracer at approximately the latitude where the aircraft flew through the plume. The plume, tilted forward as is typical for WCB transport, was intersected by the aircraft only at its easternmost end. Figure 13 shows North America NOx and CO tracers from the backward simulations along the flight path. North America tracers were low until 1405 UTC, when they suddenly increased to ∼3–5 and ∼20–40 ppb (age 3–10 days), respectively. Tracer mixing ratios decreased when the aircraft ascended to the top of the plume (compare with Figure 12), increased again when the aircraft descended, and finally dropped to almost zero as the aircraft headed back to Oberpfaffenhofen.

image

Figure 11. Total columns of the North America CO tracer on (a) 24 November 2001 1200 UTC, (b) 25 November 2001 1200 UTC, (c) 26 November 2001 1200 UTC, and (d) 27 November 2001 1500 UTC. The aircraft's flight leg is shown in Figure 11d.

Download figure to PowerPoint

image

Figure 12. Vertical cross section cutting through the North America CO tracer field from 10°W, 57°N to 15°E, 47°N on 27 November 1400–1500 UTC. The bold line shows the vertical profile of the Falcon flight.

Download figure to PowerPoint

image

Figure 13. Age spectra of North America CO and NOx tracers obtained from backward model simulations (colored bars) and superimposed aircraft measurement data (lines) for the flight on 27 November 2001: (a) North America NOx tracer, measured NOy (black line), and NO (blue line); and (b) North America CO tracer, measured CO (black line), and aircraft altitude (blue line, relative units, maximum 9900 m).

Download figure to PowerPoint

[38] Again, the measurements presented in Figure 13 are in very good agreement with the tracer calculations. CO increased from a baseline level of ∼80 ppb to 100 ppb (with a spike of 110 ppb) in the plume, and NOy increased from ∼100 ppt to 500 ppt. The NOy increase, again, is ∼10% of the NOx tracer, suggesting that a similar fraction of NOy was removed as was in the first case study. The spikes in the NOy data (mostly correlated with NO) are due to relatively frequent encounters of fresh aircraft exhaust, as the flight was within a major European flight corridor. Nevertheless, the measurements strongly support the presence of the North America plume suggested by the model calculations. CO2 (not shown) had its maximum in the plume, too.

[39] Figure 14 shows ion data during the Falcon flight of 27 November 2001 obtained by the large ion mass spectrometer (LIOMAS) of the Max-Planck-Institut für Kernphysik, Heidelberg, Germany. Plotted are fractional abundances f2, f4, and f6 of positive ions with mass numbers m >200, 400, and 600. During interception of the plume the large ion abundances increased very markedly. For example, f4 increased by a factor of about 10 and is closely correlated with the pollutant tracers CO and NOx. The observed large ions probably have been formed by the growth of small ions (m < 200) resulting from galactic cosmic ray ionization. The ion growth proceeds via clustering to ions of condensable gases X. These condensable gases must be short-lived (∼1 hour), as they condense not only on ions but also on aerosols. Therefore these condensable gases X must be formed locally in the upper troposphere. One candidate for X is sulfuric acid, which is formed from SO2 [see Eichkorn et al., 2002]. The presence of large ions in the plume has three important implications (see above): aerosol formation by ion-induced nucleation was operative in the plume, at least one condensable gas X with elevated concentrations must have been present, and efficient growth of fresh aerosols by X condensation was operative.

image

Figure 14. Positive ion data obtained by the large ion mass spectrometer during the Falcon flight of 27 November 2001. Plotted are fractional abundances of positive ions with mass numbers >200, 400, and 600 atomic mass units. Also given is the flight altitude (black line). For details, see the text.

Download figure to PowerPoint

[40] Figure 15 shows the residence time distribution both in the whole atmospheric column and in the footprint layer averaged over all backward calculations (43 in total) from boxes along the plume section of the flight (1410–1445 UTC). Again, the retroplume was sharply focused over the North Atlantic (see Figure 15a) but broadened substantially once it reached Hudson Bay (2 days back) upon (retro) outflow from the WCB. One branch of the retroplume turned to the south (reaching the Caribbean after 5 days), while another one travelled to the North American west coast (reaching it after 6 days). Both branches of the retroplume spent several days in these regions, respectively, due to slow low-level transport. The additional retroplume structures over the Atlantic and Europe with residence times <0.4% of the maximum are due to the retroplume circling the globe and are older than 10 days.

image

Figure 15. Average residence time distribution for the particles arriving in 43 small boxes along the flight track on 27 November between 1410 and 1445 UTC (a) in the whole atmospheric column and (b) in the footprint layer. Values are given as percentages of the maximum residence time, indicated below each panel. Residence times were calculated from arrival until 20 days back.

Download figure to PowerPoint

[41] The retroplume touched down into the footprint layer (Figure 15b) just over North America. Highest footprint residence times are therefore located in the southern United States, Mexico, and the Caribbean, encompassing both very clean and very polluted regions. Note the similarity of the footprint patterns with the climatologically averaged WCB source region in the North Atlantic, whose occurrence is favored by the high latent heat fluxes in the Caribbean [Stohl, 2001]. The source contribution map of CO (Figure 16) is very different from the first case. In this case the largest single contribution to the CO tracer came from the Los Angeles region, followed by major contributions from Florida, Houston, and Atlanta, Georgia.

image

Figure 16. Average source contribution per 0.5° grid box to the total CO tracer mixing ratio for 43 small boxes along the flight track on 27 November between 1410 and 1445 UTC.

Download figure to PowerPoint

5. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Aircraft Instrumentation
  5. 3. Model Description
  6. 4. Results
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[42] In this paper we have studied two episodes of ICT of pollutants from North America to Europe using aircraft measurements and Lagrangian particle dispersion model calculations. We can draw the following conclusions:

[43] 1. During CONTRACE the research aircraft Falcon could be successfully guided into pollution plumes over Europe that originated from sources in North America by using Lagrangian tracer model forecasts with relatively high resolution for the flight planning.

[44] 2. Comparisons of the model output with data measured on board the aircraft showed that the ICT plume locations were accurately captured by FLEXPART, both in the model predictions during the campaign and, even more so, in the postanalysis simulations.

[45] 3. Both episodes were preceded by stagnant conditions over North America, which led to the accumulation of pollutants in the North American ABL. WCBs were crucial for the export of this pollution into the middle and upper troposphere and its subsequent rapid transport to Europe. The plumes were often colocated with cloud structures in the satellite images, allowing fine tuning of the flight routes using real-time meteorological satellite products.

[46] 4. Concentrations of many chemical trace species (CO, NOy, CO2, acetone, and volatile organic compounds; O3 in one case) measured aboard the Falcon were clearly enhanced within the plumes as compared to outside.

[47] 5. In the plume observed on 27 November the LIOMAS instrument on board the research aircraft detected large ions. These were closely correlated with the pollution tracers CO and NOx. The presence of large ions has three important implications: aerosol particle formation by ion-induced nucleation must have been operative, at least one condensable gas X with elevated concentrations must have been present, and accelerated condensational growth of new aerosol particles must have been operative.

[48] 6. North American forest fire emissions made a moderate additional contribution to the plume observed on 19 November.

[49] 7. The anthropogenic CO tracer simulated with FLEXPART was quantitatively comparable to the CO enhancements in the measurement data, while NOy concentrations were substantially overestimated by the model NOx tracer. This suggests that in the plumes, ∼90% of the NOy was removed from the air.

[50] 8. Backward simulations with FLEXPART proved to be extremely valuable in the interpretation of the measurement data. Higher resolution than in the forward model runs could be achieved for calculating tracer mixing ratios along the flight tracks.

[51] 9. Residence time maps obtained from the backward simulations give qualitatively similar information as backward trajectories, but in contrast to a trajectory model the particle dispersion model also accounts for the growth of the retroplume. This is especially important for ICT, as retroplumes grow very fast upon (retro) outflow into the WCB source regions due to the low-level convergence feeding the WCB.

[52] 10. Owing to the linearity of the tracer dispersion problem, a detailed source contribution analysis can be made with Lagrangian backward model simulations. Given an emission inventory, the tracer mixing ratio at the receptor location can be decomposed into contributions from any number of sources, which can be plotted on a map. For both North America plumes studied in this paper we found that tracer sources contributing to the tracer concentrations over Europe were distributed over large areas in North America. For the plume observed on 27 November, sources in California, Texas, and Florida contributed almost equally, and smaller contributions were made by sources reaching from the Yucatan Peninsula to Canada.

[53] 11. For the plume observed on 19 November, sources at the North American east coast were most important, and the tracer maximum was caused by emissions from the New York region. To our knowledge, this is the first case reported in the literature where a pollution plume from a megacity was reliably detected over another continent.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Aircraft Instrumentation
  5. 3. Model Description
  6. 4. Results
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[54] This study is part of the project CONTRACE, funded by the German Federal Ministry for Education and Research within the Atmospheric Research Program 2000 (AFO 2000). ECMWF and the German Weather Service are acknowledged for permitting access to the ECMWF archives. The GOES-East infrared images were made available through the UNIDATA McIDAS data stream, and the Meteosat images were released by EUMETSAT and were made available through the NASA Marshall Space Flight Center. Discussions with P. Seibert on the theory of inverse modeling and her comments on the manuscript were very helpful. VOC data were made available by B. Rappenglück from the Research Center Karlsruhe.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Aircraft Instrumentation
  5. 3. Model Description
  6. 4. Results
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information
  • Andreae, M. O., and P. Merlet, Emission of trace gases and aerosols from biomass burning, Global Biogeochem. Cycles, 15, 955966, 2001.
  • Andreae, M. O., J. Fishman, and J. Lindesay, The Southern Tropical Atlantic Region Experiment (STARE): Transport and Atmospheric Chemistry near the Equator-Atlantic (TRACE-A) and Southern African Fire-Atmosphere Research Initiative (SAFARI): An introduction, J. Geophys. Res., 101, 23,51923,520, 1996.
  • Arnold, F., J. Schneider, K. Gollinger, H. Schlager, P. Schulte, D. E. Hagen, P. D. Whitefield, and P. van Velthoven, Observation of upper tropospheric sulfur dioxide and acetone pollution: Potential implications for hydroxyl radical and aerosol formation, Geophys. Res. Lett., 24, 5760, 1997.
  • Arnold, F., J. Curtius, B. Sierau, V. Bürger, R. Busen, and U. Schumann, Detection of massive negative chemiions in the exhaust plume of a jet aircraft in flight, Geophys. Res. Lett., 26, 15771580, 1999.
  • Beine, H. J., A. Dahlback, and J. B. Ørboek, Measurements of J(NO2) at Ny-lesund, Svalbard, J. Geophys. Res., 104, 16,00916,019, 1999.
  • Browning, K. A., Organization of clouds and precipitation in extratropical cyclones, in Extratropical Cyclones: The Erik H. Palmen Memorial Volume, edited by C. Newton, and E. Holopainen, pp. 129152, Am. Meteorol. Soc., Boston, Mass., 1990.
  • Cofer, W. R.III, E. L. Winstead, B. J. Stocks, J. G. Goldammer, and D. R. Cahoon, Crown fire emissions of CO2, CO, H2, CH4, and TNMHC from a dense jack pine boreal forest fire, Geophys. Res. Lett., 25, 39193922, 1998.
  • Cooper, O. R., et al., Trace gas signatures of airstreams within North Atlantic cyclones: Case studies from the North Atlantic Regional Experiment (NARE '97) aircraft intensive, J. Geophys. Res., 106, 54375456, 2001.
  • Cooper, O. R., J. L. Moody, D. D. Parrish, M. Trainer, T. B. Ryerson, J. S. Holloway, G. Hübler, F. C. Fehsenfeld, and M. J. Evans, Trace gas composition of midlatitude cyclones over the western North Atlantic Ocean: A conceptual model, J. Geophys. Res., 107(D7), 4056, doi:10.1029/2001JD000901, 2002.
  • Derwent, R. G., P. G. Simmonds, S. Seuring, and C. Dimmer, Observation and interpretation of the seasonal cycles in the surface concentrations of ozone and carbon monoxide at Mace Head, Ireland from 1990 to 1994, Atmos. Environ., 32, 145157, 1998.
  • Dutot, A. L., et al., Tropospheric ozone and precursors at the Porspoder station, France, in Tropospheric Ozone Research, edited by. Ø. Hov, pp. 216221, Springer-Verlag, New York, 1997.
  • Eichkorn, S., S. Wilhelm, H. Aufmhoff, K. H. Wohlfrom, and F. Arnold, Cosmic ray-induced aerosol-formation: First observational evidence from aircraft-based ion mass spectrometer measurements in the upper troposphere, Geophys. Res. Lett., 29(14), 1698, doi:10.1029/2002GL015044, 2002.
  • Emanuel, K. A., and M. Živković-Rothman, Development and evaluation of a convection scheme for use in climate models, J. Atmos. Sci., 56, 17661782, 1999.
  • European Centre for Medium-Range Weather Forecasts, User Guide to ECMWF Products 2.1, Meteorol. Bull., M3.2, Reading, England, 1995.
  • Fehsenfeld, F. C., M. Trainer, D. D. Parrish, A. Volz-Thomas, and S. Penkett, North Atlantic Regional Experiment 1993 summer intensive: Foreword, J. Geophys. Res., 101, 28,86928,875, 1996.
  • Feigl, C., Aufbau und Charakterisierung eines Messsystems für NO, NO2 und NOy: Laboruntersuchungen und Einsatz in der unteren arktischen Stratosphäre, Ph. D. thesis, DLR-FB-98-21, Dtsch. Zentrum für Luft- und Raumfahrt-Oberpfaffenhofen, Wessling, Germany, 1998.
  • Fenneteaux, I., et al., Influence of continental sources on oceanic air composition at the eastern edge of the North Atlantic ocean, TOR 1992–1995, J. Atmos. Chem., 32, 233280, 1999.
  • Flesch, T. K., J. D. Wilson, and E. Yee, Backward-time Lagrangian stochastic dispersion models and their application to estimate gaseous emissions, J. Appl. Meteorol., 34, 13201332, 1995.
  • Forster, C., et al., Transport of boreal forest fire emissions from Canada to Europe, J. Geophys. Res., 106, 22,88722,906, 2001.
  • Gerbig, C., D. Kley, A. Volz-Thomas, J. Kent, K. Dewey, and D. S. McKenna, Fast response resonance fluorescence CO measurements aboard the C-130: Instrument characterization and measurements made during North Atlantic Regional Experiment 1993, J. Geophys. Res., 101, 29,22929,238, 1996.
  • Gerbig, C., S. Schmitgen, D. Kley, A. Volz-Thomas, K. Dewey, and D. Haaks, An improved fast-response vacuum-UV resonance fluorescence CO instrument, J. Geophys. Res., 104, 16991704, 1999.
  • Huntrieser, H., H. Schlager, C. Feigl, and H. Höller, Transport and production of NOx in electrified thunderstorms: Survey of previous studies and new observations at midlatitudes, J. Geophys. Res., 103, 28,24728,264, 1998.
  • Jacob, D. J., B. D. Field, E. M. Jin, I. Bey, Q. Li, J. A. Logan, R. M. Yantosca, and H. B. Singh, Atmospheric budget of acetone, J. Geophys. Res., 107(D10), 4100, doi:10.1029/2001JD000694, 2002.
  • Jakobs, H. J., H. Feldmann, H. Hass, and M. Memmesheimer, The use of nested models for air pollution studies: An application of the EURAD model to a SANA episode, J. Appl. Meteorol., 34, 13011319, 1995.
  • Jennings, S. G., T. G. Spain, B. G. Doddridge, H. Maring, B. P. Kelly, and A. D. A. Hansen, Concurrent measurements of black carbon aerosol and carbon monoxide at Mace Head, J. Geophys. Res., 101, 19,44719,454, 1996.
  • Kondo, Y., The Biomass Burning and Lightning Experiment, IGACtivities Newsl., 20, 47, 2000.
  • Lavoué, D., C. Liousse, H. Cahchier, B. J. Stocks, and J. G. Goldammer, Modeling of carbonaceous particles emitted by boreal and temperate wildfires at northern latitudes, J. Geophys. Res., 105, 26,87126,890, 2000.
  • Lawrence, M., et al., An operational global chemical weather forecast system for field campaign support: Predictions and observations of large-scale features during INDOEX, MINOS, and CONTRACE, Atmos. Chem. Phys. Discuss., 2, 15451597, 2003.
  • Lelieveld, J., et al., The Indian Ocean Experiment: Widespread air pollution from south and southeast Asia, Science, 291, 10311036, 2001.
  • Li, Q., et al., Transatlantic transport of pollution and its effects on surface ozone in Europe and North America, J. Geophys. Res., 107(D13), 4166, doi:10.1029/2001JD001422, 2002.
  • Olivier, J. G. J., and J. J. M. Berdowski, Global emissions sources and sinks, in The Climate System, edited by J. Berdowski, R. Guicherit, and B. J. Heij, pp. 3378, A. A. Balkema, Brookfield, Vt., 2001.
  • Parrish, D. D., J. S. Holloway, M. Trainer, P. C. Murphy, G. L. Forbes, and F.C. Fehsenfeld, Export of North American ozone pollution to the North Atlantic Ocean, Science, 259, 14361439, 1993.
  • Parrish, D. D., M. Trainer, J. S. Holloway, J. E. Yee, M. S. Warshawsky, F. C. Fehsenfeld, G. L. Forbes, and J. L. Moody, Relationships between ozone and carbon monoxide at surface sites in the North Atlantic region, J. Geophys. Res., 103, 13,35713,376, 1998.
  • Prados, A. I., R. R. Dickerson, B. G. Doddridge, P. A. Milne, J. L. Moody, and J. T. Merrill, Transport of ozone and pollutants from North America to the North Atlantic Ocean during the 1996 Atmosphere/Ocean Chemistry Experiment (AEROCE) intensive, J. Geophys. Res., 104, 26,21926,233, 1999.
  • Ryall, D. B., R. H. Maryon, R. G. Derwent, and P. G. Simmonds, Modelling long-range transport of CFCs to Mace Head, Ireland, Q.J.R. Meteorol. Soc., 124, 417446, 1998.
  • Schlager, H., P. Konopka, P. Schulte, U. Schumann, H. Ziereis, F. Arnold, M. Klemm, D.E. Hagen, P.D. Whitefield, and J. Ovarlez, In situ observations of air traffic emission signatures in the North Atlantic flight corridor, J. Geophys. Res., 102, 10,73910,750, 1997.
  • Schmitt, R., Simultaneous measurements of carbon monoxide, ozone, PAN, NMHC and aerosols in the free troposphere at Izaña, Canary Islands, in The Proceedings of EUROTRAC Symposium '94, edited by P. M. Borrell, pp. 313316, SPB Acad., The Hague, 1994.
  • Schröder, F. P., B. Kärcher, A. Petzold, R. Baumann, R. Busen, C. Hoell, and U. Schumann, Ultrafine aerosol particles in aircraft plumes: In situ observations, Geophys. Res. Lett., 25, 27892792, 1998.
  • Schulte, P., H. Schlager, H. Ziereis, U. Schumann, S. L. Baughcum, and F. Deidewig, NOx emissions indices of subsonic long-range jet aircraft at cruise altitude: In situ measurements and predictions, J. Geophys. Res., 102, 2143121442, 1997.
  • Schultz, M., R. Schmitt, K. Thomas, and A. Volz-Thomas, Photochemical box modeling of long-range transport from North America to Tenerife during the North Atlantic Regional Experiment (NARE) 1993, J. Geophys. Res., 103, 13,47713,488, 1998.
  • Schumann, U., P. Konopka, R. Baumann, R. Busen, T. Gerz, H. Schlager, P. Schulte, and H. Volkert, Estimate of diffusion parameters of aircraft exhaust plumes near the tropopause from nitric oxide and turbulence measurements, J. Geophys. Res., 100, 14,14714,162, 1995.
  • Schumann, U., H. Schlager, F. Arnold, J. Ovarlez, H. Kelder, Ø. Hov, G. Hayman, I. S. A. Isaksen, J. Stähelin, and P. D. Whitefield, Pollution from aircraft emissions in the North Atlantic flight corridor: Overview on the POLINAT projects, J. Geophys. Res., 105, 36053631, 2000.
  • Seibert, P., Inverse modelling with a Lagrangian particle dispersion model: Application to point releases over limited time intervals, in Air Pollution Modeling and its Application XIV: Proceedings of ITM Boulder, edited by S. E. Gryning, and F. A. Schiermeier, pp. 381389, Plenum, New York, 2001.
  • Seibert, P., and A. Stohl, Inverse modelling of the ETEX-1 release with a Lagrangian particle model, paper presented at the 3rd GLOREAM Workshop, Cent. for Res. on Parallel Comput. and Supercomput. of Naples, Ischia, Italy, 22 – 24 Sept. 1999.
  • Seibert, P., B. Krüger, and A. Frank, Parametrisation of convective mixing in a Lagrangian particle dispersion model, paper presented at the 5th Global and Regional Atmospheric Modelling Project Workshop, Paul Scherrer Inst. Lab. of Atmos. Chem., Wengen, Switzerland, 24 – 26 Sept. 2001.
  • Shetter, R. E., and M. Müller, Photolysis frequency measurements using actinic flux spectroradiometry during the PEM-Tropics mission: Instrumentation description and some results, J. Geophys. Res., 104, 56475661, 1999.
  • Spichtinger, N., M. Wenig, P. James, T. Wagner, U. Platt, and A. Stohl, Satellite detection of a continental-scale plume of nitrogen oxides from boreal forest fires, Geophys. Res. Lett., 28, 45794582, 2001.
  • Stohl, A., A 1-year Lagrangian “climatology” of airstreams in the Northern Hemisphere troposphere and lowermost stratosphere, J. Geophys. Res., 106, 72637279, 2001.
  • Stohl, A., and D. J. Thomson, A density correction for Lagrangian particle dispersion models, Boundary Layer Meteorol., 90, 155167, 1999.
  • Stohl, A., and T. Trickl, A textbook example of long-range transport: Simultaneous observation of ozone maxima of stratospheric and North American origin in the free troposphere over Europe, J. Geophys. Res., 104, 30,44530,462, 1999.
  • Stohl, A., G. Wotawa, P. Seibert, and H. Kromp-Kolb, Interpolation errors in wind fields as a function of spatial and temporal resolution and their impact on different types of kinematic trajectories, J. Appl. Meteorol., 34, 21492165, 1995.
  • Stohl, A., M. Hittenberger, and G. Wotawa, Validation of the Lagrangian particle dispersion model FLEXPART against large scale tracer experiment data, Atmos. Environ., 32, 42454264, 1998.
  • Stohl, A., et al., The influence of stratospheric intrusions on alpine ozone concentrations, Atmos. Environ., 34, 13231354, 2000.
  • Stohl, A., S. Eckhardt, C. Forster, P. James, and N. Spichtinger, On the pathways and timescales of intercontinental air pollution transport, J. Geophys. Res., 107(D23), 4684, doi:10.1029/2001JD001396, 2002a.
  • Stohl, A., S. Eckhardt, C. Forster, P. James, N. Spichtinger, and P. Seibert, A replacement for simple back trajectory calculations in the interpretation of atmospheric trace substance measurements, Atmos. Environ., 36, 46354648, 2002b.
  • Stohl, A., M. Trainer, T. B. Ryerson, J. S. Holloway, and D. D. Parrish, Export of NOy from the North American boundary layer during 996 and 1997 North Atlantic Regional Experiments, J. Geophys. Res., 107(D11), 4131, doi:10.1029/2001JD000519, 2002c.
  • Volz-Thomas, A., A. Lerner, H.-W. Pätz, M. Schultz, D.S. McKenna, R. Schmitt, S. Madronich, and E. P. Röth, Airborne measurements of the photolysis frequency of NO2, J. Geophys. Res., 101, 18,61318,627, 1996.
  • Wernli, H., and H. C. Davies, A Lagrangian-based analysis of extratropical cyclones. I: The method and some applications, Q.J.R. Meteorol. Soc., 123, 467489, 1997.
  • Wild, O., and H. Akimoto, Intercontinental transport of ozone and its precursors in a three-dimensional global CTM, J. Geophys. Res., 106, 27,72927,744, 2001.
  • Wild, O., K. S. Law, D. S. McKenna, B. J. Bandy, S. A. Penkett, and J. A. Pyle, Photochemical trajectory modeling studies of the North Atlantic region during August 1993, J. Geophys. Res., 101, 29,26929,288, 1996.
  • Yienger, J. J., M. Galanter, T. A. Holloway, M. J. Phadnis, S. K. Guttikunda, G. R. Carmichael, W. J. Moxim, and H. Levy II, The episodic nature of air pollution transport from Asia to North America, J. Geophys. Res., 105, 26,93126,945, 2000.
  • Ziereis, H., H. Schlager, P. Schulte, I. Köhler, R. Marquardt, and C. Feigl, In situ measurements of the NOx distribution and variability over the eastern North Atlantic, J. Geophys. Res., 104, 16,02116,032, 1999.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Aircraft Instrumentation
  5. 3. Model Description
  6. 4. Results
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.