tracer transport for realistic aircraft emission scenarios

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 100, NO. D3, PAGES 5203-5214, MARCH 20, 1995

Tracer transport for realistic aircraft emission scenarios calculated using a three-dimensional model

Clark J. Weaver

Applied Research Corporation, Landover, Maryland

Anne R. Douglass and Richard B. Rood Laboratory for Atmospheres, Goddard Space Flight Center, Greenbelt, Maryland

Abstract. A three-dimensional transport model, which uses winds from a stratospheric data assimilation system, is used to study the transport of supersonic aircraft exhaust in the lower stratosphere. A passive tracer is continuously injected into the transport model. The tracer source distribution is based on realistic scenarios for the daily emission rate of reactive nitrogen species for all forecasted flight routes. Winds are from northern hemisphere winter/spring months for 1979 and 1989; there are minimal differences between the tracer integrations for the 2 years. During the integration, peak tracer mixing ratios in the flight corridors are compared with the zonal mean and found to be greater by a factor of 2 or less. This implies that the zonal mean assumption used in two dimensional models is reasonable during winter and spring. There is a preference for pollutant buildup in the heavily traveled North Pacific and North Atlantic flight corridors. Pollutant concentration in the corridors depends on the position of the Aleutian anticyclone and the northern hemisphere polar vortex edge.

1. Introduction

Evaluating the environmental impact of a proposed fleet of supersonic aircraft flying in the stratosphere requires a variety of modeling approaches. While zonally averaged two- dimensional (2-D) models are able to perform long-term, multi- year assessment calculations that account for a large number of chemical reactions, the zonal mean assumption means they cannot simulate the zonally asymmetric emission of aircraft exhaust. It is not yet computationally feasible to run a three- dimensional (3-D) chemistry transport model for the long integrations necessary to evaluate both the full environmental impact and the sensitivity of calculated results to model inputs and parameters. Use of both 2-D and 3-D transport models can ease the limitations inherent in both approaches. A 3-D transport model with a passive tracer is used here to test the validity of the 2-D zonal mean assumption. This question has also been addressed by Douglass et al. [1993] and Rasch et al. [1994]. The current study builds on this previous work by considering more realistic aircraft emissions scenarios that account for the cumulative effect of all flight routes instead of a single route.

Both the study reported byDouglass et al. [1993] and the present study use a 3-D transport model with winds from a data assimilation system. The advantage of this approach is the realistic simulation of the actual atmospheric circulation. The assimilated data sets available at the beginning of this study spanned January-April for 2 different years. This time duration is long enough to study the effect of zonally asymmetric emissions on a seasonal timescale.

Copyright 1995 by the American Geophysical Union.

Paner number 9411303320.

0148-0227195/94JD-03320505.00

The tracer transport in 2-D and 3-D models was compared by Douglass et al. [1993] by integrating both models with a zonally symmetric passive source. In that earlier study, 4- month calculations using a 2-D model with tracer material injected continuously between 40 ø and 50 ø north and south latitudes at 50 mbar were compared with similar 3-D calculations from the same period. The seasonal behavior of the tracer fields in both models was similar. The zonal mean of

the 3-D tracer distribution was similar to the 2-D distribution when the 2-D calculation used a residual circulation derived

from the wind fields used in the 3-D calculation. At the level of

injection for a given latitude, the standard deviation of the 3-D tracer distribution was about 25% of the zonal mean and

represented the zonal asymmetry due to dynamical transport. Also presented were results of 3-D transport model runs where exhaust was continually emitted into three single flight corridors: Boston-London, Los Angeles-Tokyo and Los Angeles-Sydney. The tracer distribution from all three integrations were zonally asymmetric. The North Atlantic route showed that peak mixing ratios at the flight level were usually 3 or 4 times the zonal mean for the year simulated. A significant portion of tracer from the Los Angeles-Tokyo route remained trapped in the Aleutian anticyclone and tracer from the Boston-London route was excluded. The single corridor study focussed on individual dispersion characteristics of a single route and used dimensionless injection rates. The present work is a more realistic assessment, since it includes the relative contribution of all proposed flight corridors and the tracer injection rate is the amount of nitrogen oxides expected from the projected aircraft fleet.

Isentropic trajectory calculations can also be used to study the dispersal to aircraft exhaust. Sparling et al. [1995] simulate the emission of exhaust by the daily initialization of air parcels along the New York-London and Los Angeles-Tokyo

5203

5204 WEAVER ET AL.: AIRCRAFT F3dISSION TRANSPORT

flight paths on the 500 K isentropic surface. Integrations were for 20 days in January for each of the years 1980-1994. In agreement with the single corridor study using the 3-D transport model, parcels released during winter along the Los Angeles-Tokyo corridor often get trapped in the Aleutian anticyclone while the New York-London emissions are excluded from it. To study the seasonal variability, 20-day integrations were done for January, March, July, and November 1992 for only the New York-London corridor. The buildup of emissions was strongest for July and weakest for January and November.

2. Model Description

Since this is a sequel to the single corridor study, the model as well as the winds used are identical to those used by Douglass et al. [ 1993]. The transport model is the Goddard 3-D chemical transport model (CTM) configured to run a passive tracer. An upwind flux limited scheme [van Leer, 1974] is used to calculate horizontal transport at 2 latitude x 2.5 longitude resolution; further details concerning the implementation of this scheme are given by Allen et al. [1991]. The CTM has 8 tropospheric sigma levels and 11 stratospheric pressure levels in the vertical with an interface at 120 hPA. This is about a 2-

km resolution in the troposphere and 3.5-kin resolution in the stratosphere. The nondiffusive transport scheme developed by Prather [1986] is used for vertical advection to compensate somewhat for the coarse vertical resolution.

Winds for the transport are taken from the STRATAN data assimilation procedure as described by Rood et al. [1989]. Previous studies have shown that the assimilation dynamics faithfully represent planetary- and synoptic-scale features in the transport model. Ozone simulations with the 3-D CTM, which use parameterized chemical production and loss rates from a 2-D model, compare favorably with observations from sondes, aircraft and satellite from the Total Ozone Mapping Spectrometer (TOMS) and Limb Infrared Monitor of the Stratosphere (LIMS) [Rood et al., 1991, 1992]. While synoptic-scale ozone features are accurately reproduced, on longer timescales there is evidence that an excessive residual circulation in the CTM leads to an unrealistic buildup of ozone in the upper troposphere and lower stratosphere at middle latitudes. This shortcoming is due to a large residual in the expected thermodynamic balance in the assimilated winds and is more thoroughly discussed by Weaver eta/. [1993]. A more recent assimilation procedure which gradually updates the incremental analyses [Schubert et al., 1993] is in closer thermodynamic balance and has a more reasonable residual circulation. This study is focussed on the horizontal distribution of exhaust at the level of injection and is insensitive to this aspect of the wind fields.

The rate of injection of passive tracer into the transport model is based on the 1993 Interim Assessment Scenarios

[Wuebbles et al., 1993]. The amount of nitrogen oxides, hydrocarbons, and carbon monoxide emitted as functions of latitude, longitude, and altitude are specified in these scenarios. The emission rates depend on the aircraft performance and engine characteristics. These are summarized by an emission index, which gives the amount of a substance released per kilogram of fuel burned. The latitude and longitude of emission depend on the aircraft routes, which are based on marketing forecasts. The altitude of emission of the aircraft is higher for higher Mach number (Mach number M = speed of aircraft /

speed of sound). Our study considers only NOx (NO+NO2) emissions from scenario C (Mach number 1.6 (M1.6), emission index EI=5) and emission scenario E (M2.4, EI=5) which are reported in molecules per year on a 1 ø latitude x 1 ø longitude x 22 level grid from 0 to 22 krn. To prepare the scenarios for the transport model, these values are converted to parts per million by volume (ppmv) and interpolated onto the CTM grid with care taken to conserve mass. Both of these scenarios include the emissions from subsonic aircraft.

Wuebbles et al. [1993] provide emissions for a reference fleet for year 2015 that includes only subsonic aircraft assuming the existence of supersonic aircraft (scenario J); this subsonic scenario is subtracted from each of the above scenarios.

Inclusion of subsonic aircraft emissions would only have a small effect on the results in the stratosphere. Single corridor tracer studies of emissions from subsonic aircraft show that

only a small amount of tropospheric emissions is vertically transported to the stratosphere (R. Rood, personal communication, 1994). Tropospheric emissions are associated with takeoff and landing of supersonic aircraft. To ensure that passive tracer found in the troposphere is from vertical transport rather than from aircraft takeoff or landing, tropospheric emissions are not included in these simulations. Figure l a shows a vertical profile of the emission rates for the two scenarios. The M2.4 scenario has peak emission at the 53 mbar model pressure level, and the M1.6 scenario peaks at the 91 mbar model level. Figure lb shows the source locations at the peak emission levels for M2.4. The North Atlantic and North Pacific flight corridors have the highest emission rates; tracks to Hawaii and those in the Indian Ocean are moderate by comparison. Emission rates in the southern hemisphere (not shown) are minor. The initial mixing ratio field is set to zero and at each time step (450 s) the mixing ratio is augmented using these emission rates.

The published emissions rates are given in terms of NO• perturbation due to aircraft exhaust because NO x is the major odd nitrogen species at the tail pipe. However, once in the atmosphere, some of the NO x is converted to less reactive nitrogen species such as HNO 3 and N20 5. Therefore the total

odd nitrogen (NOy) perturbation due to aircraft exhaust is a more realistic label for the tracer mixing ratio and will be used hereinafter.

10 Emission rates

-- Scenario M2.4 .... Scenario M1.6

100

1000

0 1.0,1c 2.0-10 '8 3.0-10 '8 ppbv / sec

Figure la. Aircraft emission rates for selected scenarios after interpolation to the transport model grid. Shown are vertical profiles of the global averages for scenario M2.4 and scenario M 1.6.

WEAVER ET AL.: AIRCRAFT EMISSION TRANSPORT 5205

Scenario M2.4 Injection rate at 53 hPa

(

270

stratospheric aircraft. For example, high local concentrations of exhaust tracer may lead to nonlinear chemical effects, particularly if the concentration is large enough to trigger a threshold process such as polar stratospheric cloud (PSC) formation. Using a PSC parameterization in a 2-D chemical model, Considine et al. [1994] consider ozone changes from increased PSC formation due to aircraft exhaust perturbation. Although this parameterization accounts for longitudinal variation in the temperature field, it considers a zonal mean increase in HNO 3 and H20 when calculating the probability of PSC formation. Even with the zonal mean assumption for the constituents, increased PSC formation causes decreases in ozone in the southern hemisphere. Knowledge of maximum NO• values along with minimum temperatures could improve the prediction of PSC formation and determine the likelihood of PSC formation.

4. Results

The time series of the average of the grid boxes which contain 1% and 10% of the highest values are shown in Figure 2. These time series are not as noisy as time series of the daily maximum values and are more representative of the modeled corridor buildup. If the tracer had been injected in the model and

o -• (30 N) o3 o o o o o o

ppbv, l e6 / sec

Figure lb. Aircraft emission rates for scenario M2.4 at the 53 mbar level for the northern hemisphere.

3. Background The tracer evolution is studied for 1979 and 1989 during

winter and spring months in the northern hemisphere. The integration periods are from December 28, 1978 to April 20, 1979 and from December 28, 1988 to March 31, 1989. Meteorological conditions control the evolution of the tracer distribution. Both winters have been studied extensively by Smith et al. [1984] and Leovy et al. [1985] for 1979 and by Newman et al. [1989] for 1989. The 1978-1979 winter had a minor wavenumber 1 warming in the lower stratosphere beginning January 18 as the vortex slid away from the pole toward Europe. Between February 17 and 24 there was a major wavenumber 2 warming. The 1989 winter had a pronounced wavenumber 1 feature around January 21 leading to a slight stratospheric warming and a major wavenumber 2 warming from February 15 to 19.

In the simulations, stagnant wind regimes favor a large buildup of pollutant in the corridor. There is a "corridor effect" because the highest concentrations are typically found in the areas of largest emissions, i.e., in flight corridors under conditions of weak winds. The ratio of the maximum value to the zonal mean may be large enough to invalidate the assumption that chemical effects may be calculated from the zonal mean perturbation. On the other hand, rapid evacuation of the pollutant, perhaps due to high wind velocities, results in a weak corridor effect [Douglass et al., 1993; Sparling et al., 1995]. This study evaluates the corridor effect by identifying the local maximum values of exhaust tracer and comparing these values with a global or zonal mean. Use of the zonal mean assumption when there is a strong corridor effect could potentially lead to an unrealistic calculation of the effect of the

Peak mixing ratios 91 hPa Scenario M1.6

f ' ' ' I I ' ' ' I I 0.35 Top 1% 1979 Top 1% 1989

0.30 -- Top 10 % 1979 ' - - - Top 10 % 1989 :"'

0.10

20 40 60 80 100 120 Model day

Peak mixing ratios 53 hPa Scenario M2.4

0.25

"" 0.20

0.15

0.50

0.40

• Top 1% 1979 .... Top 1% 1989 -- Top 10 % 1979

>

.o 0.30

0.20

- - - Top 1/•.,,•/I.]•11 "'"" '""" ;• '

0.101: --I.. --, .... i ...... i.' ....

20 40 60 80 100 120 Model day

Figure 2. Time series of peak mixing ratio values for years 1979 and 1989 for (a) scenario M1.6 at 91 hPa and (b) scenario M2.4 at 53 hPa. For a given day, mixing ratios from 3312 grid boxes at the stated level are sorted by magnitude. The average mixing ratio for the grid boxes which contain values which are in the top 1 and 10% of the distribution are plotted. The 10% trace averages 331 of the highest values; the 1% trace averages 33 of the highest values.

5206 WEAVER ET AL.: AIRCRAFT EMISSION TRANSPORT

transport turned off, the traces would increase linearly and steeply. By day 20, peak values would be larger than any seen in Figure 2. The weaker upward trend of both traces in Figure 2 demonstrates the dispersal of the pollutant by the dynamics. The smaller sample size of the 1% trace makes it more sensitive to local dynamic events. Tracer buildup from a single event will have less influence on the top 10% trace, producing a more gradual trend. At the end of the integration, the peak concentrations as shown in Figure 2 are still increasing, so there is no estimation of the peak concentration of pollutant in the real world for the spring/winter season. Isentropic trajectory calculations of Sparling et al. [1995] show that pollutant injected in the summer months will likely build up in the corridor.

To examine the spatial variability of tracer, normalized mixing ratios (NMR) are calculated by scaling the local mixing ratio by the global amount of tracer in the model. This partially removes the effect of the continual injection of tracer and allows comparison of time averages as well. The peak values can also be compared with the zonal mean to indicate the validity of the zonal mean assumption in the 2-D models. Throughout the integration the highest concentrations of exhaust are found at the vertical level of peak emissions: 91 mbar and 53 mbar for M1.6 and M2.4, respectively.

Figure 3 shows a time average of the NMRs during January 19 to February 5 (model integration days 22-39) at the peak emission level (53 mbar) for scenario M2.4 for both years. Since this is the early stage ef the integration, the NMR are maintained by local emission of the aircraft. For both 1979 and 1989 the highest values are seen in regions of heavy air traffic: (1) North Atlantic corridor, (2) U.S. west coast to Japan corridor, and (3) U.S. west coast to Hawaii corridor. Figures 2a and 2b show that these peak NMR values are associated with NOy mixing ratios from aircraft perturbation between 0.2 and 0.3 ppbv. This is at most twice as large as the zonal mean perturbation mixing ratios at the flight corridor latitude (0.15 ppbv) and 4-6 times the global mean of the perturbation (0.05 ppbv) for both years. The pollutant source in the northern hemisphere polar vortex consists of two lightly traveled routes. This weak source in concert with the restricted

meridional transport across the vortex edge keeps the vortex NMR less than 1. Transport simulations using a lower mean flight altitude, scenario M1.6 (not shown), are qualitatively similar to those described in scenario M2.4. For both

scenarios the same three heavy traffic regions are found, the peak NMR values are about 4 times the global mean, and the vortex remains largely free of pollutant.

Scenario M2.4 normalized mixing ratio 53 hPa Time mean 790119 - 790205

180

270

o .... Fo

• o • o • o

Normalized mixing raUo

Figure 3a. Time average of the normalized mixing ratio values for scenario M2.4 at 53 mbar from January 19 to February 5 for 1979.

WEAVER ET AL.: AIRC• EMISSION TRANSPORT 5207


180

270

::: :::::::::::::::::::::::: ::::::::::::::::::::: ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: - ': ".'.;i ;.i.; ;;':.;;;::.i.::-i.:: 5 ::.::.... 5 5.:..:.i.:.:: ........ :..: :: ....... :::i:i ....... •ii:!,:,!:!..! !: :•:i:!:.:,:-...'<:.-':•:•:•,-•i.'.'-'-'.'-'-'•--'.•:,..-•... ,;•':..'..•..'.:•.- I I I I I I I I I

? o o ? .... o k• L. -4 o • u• -4 o

O• 0 O• 0 • 0 O• 0

Normolized mixing r0tio

Figure 3b. Same as Figure 3a but for 1989.

Cleansing of North Atlantic Corridor Normalized mixing ratios for 6-23 February (model

integration days 40-57) using scenario M2.4 are shown in Figure 4. For both years the polar vortex remains isolated, and NMR values greater than 4 occur in the eastern Pacific (45'N, 210'E). For 1979 the North Atlantic corridor (along 45'N between 300'E and 360'E) is clearly defined with NMR values above 4. However, for 1989 the location of buildup is not in the relatively clean corridor but instead over Western Europe around 45'N, 0'E. For both years, Figure 2 shows that the M1.6 scenario has peak mixing ratios of about 0.3 ppbv, which is a factor of 2 greater than the zonal mean (0.15 ppbv) and 4 times the global background mean.

To understand why there is build up of pollutant in the Pacific corridor in 1989 and at the same time a cleansing of the North Atlantic corridor, we use isentropic back trajectory analysis. This will tell the past history of air parcels in the corridor. For both the Ariantic and Pacific corridor, 50 points ending along the aircraft flight routes were used to initialize the Goddard Isentropic Trajectory Model [Schoeberl and Sparling, 1995]. Figure 5 overlays a subset of trajectories on an NMR field for model day 56. The trajectory model keeps parcels on the 500 K potential temperature surface and calculates a position about every 15 min. It is conveniently set

up to use NMC balanced winds [Newman et al., 1989] which are interpolated in time and space to the parcel location. Although use of the assimilated winds (used in the transport) is more appropriate for this comparison, the qualitative differences are minimal [Sparling et al., 1995]. Kinks in the trajectories indicate parcel positions at 0000 UT for the previous 6 days. The most polluted region in the eastern Pacific (50øN, 210øE) is collocated with highly anticyclonic trajectories. Exhaust emitted in this region during the last 6 days remains trapped in this anticyclonic circulation. The western end of the corridor (45øN, 150øE) is much cleaner; parcels ending closest to Japan do not encounter any flight corridor as they travel from China. Similarly, some parcels ending along the clean Atlantic corridor have been circulating around the polar vortex (two are shown), encountering only the lightly traveled routes across the vortex. The other remaining Atlantic trajectories (one is shown) have origins in the clean subtropics. Geopotential height fields at 250 mbar show a high-pressure system centered at 30øN, 310øE that is associated with this northward advection of subtropical air. Since the 18-day time mean (Figure 4b) also shows a tongue of cleaner air around 310øE that extends from the subtropics to the midlatitudes, this advection pattern is a stationary feature. For this case, buildup occurs when pollutant is introduced in anticyclonic circulation patterns; pollutant is

5208 WE.AVER ET AL.: AIRCRAFF EMISSION TRANSPORT


180

270

.:::::::::::::::::: ........ ::::::::::::::::::::::::::::::::::::::::::: ........ :::::::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: ....................................................................................... :::::::::::::::::::::: ::::::::::::::::::::::::::: :5: •:: 5::•:½ :::::::::::::::::::::::::::: ::!•::.-'•!•ii•i•!•i½•i•' •' "":- ...:. ß, '•-.... ,;.... •'•.'.j• ..'•

o o o o .... .• b f,o L• :,4 o • o• -4 o

On 0 On 0 On 0 On 0

Normalized mixing ratio

Figure 4a. Time average of the normalized mixing ratio values for scenario M2.4 at 53 mbar from February 6 - 23 for 1979.

dispersed when it is introduced at the edge of the vortex or when is it is mixed with clean intrusions of subtropical air.

The North Atlantic corridor is more polluted in 1979 than 1989 because of the differences in the orientation of the

vortex and position of the planetary waves. Figure 6 shows geopotential height fields typical of this 18-day period for both years. Both years have major wavenumber 2 warming events during this period of the integration. The wavenumber 2 feature is pronounced on February 22, 1979; the vortex is split with one lobe over North America and the other over central

Asia. In concert with this are anticyclonic regions over Western Europe/eastern North Atlantic and the North Pacific, both of which are collocated with heavily traveled aircraft routes. The pollutants are trapped, and the corridors remain def'med for 1979. On February 22 the wavenumber 2 feature is shifted slightly eastward compared with 1989. Now the vortex edge is over the North Atlantic corridor where the pollutant can be quickly dispersed and the anticyclone is over Eastern Europe away from the corridor sources.

The M1.6 scenario for February 6-23 (model integration days 40-•7) has locations of peak NMR for both years in almost the same positions as the M2.4 scenario (Figures 4a

and 4b). The relatively clean Atlantic corridor is apparent in 1989 for this integration also.

Vortex Breakup

For the period February 24 - March 13 (model integration days 58-75), values of NMR above 4 are still maintained in the 1979 and 1989 integrations for the M2.4 scenario. For 1989 the center of the pollutant maximum in the Pacific shifts westward from the previous 18-day period (Figure 7). For 1989 this coincides with a weakening polar vortex centered over Siberia. The M1.6 scenario (not shown) also shows this shift in the center of the pollutant maximum.

For the period March 14-31 (model integration day 76-93), Figure 8 shows that the contrast between the clean polar vortex and the dirtier midlatitudes lessens for scenario M2.4. Not

shown is the 1979 field, which exhibits the same features. Peak values of NMR decline, but the corridors are clearly evident in the midlatitudes. Figure 2 shows that the M2.4 scenario has peak mixing ratios of 0.35 ppbv, which is about 1.75 times the perturbation zonal mean (0.2 ppbv). The lessened contrast between the polar and midlatitude regions

WEAVER ET AL.: AIRCRAFT F.34ISSION TRANSPORT 52o9


180

270

o o ....

o c• o c• o

Normalized mixing ratio

Figure 4b. Same as Figure 4a but for 1989.

results from the breakup of the polar vortex. During the winter months leading up to the beginning of this period, fields of Ertel potential vorticity (Epv) for both years show a strong gradient between the high Epv values of the vortex and lower values of the midlatitudes. Since Epv is conserved for short time periods, this strong gradient is indicative of a barrier to meridional transport and is consistent with the meridional gradient in the tracer distribution. However, during this period the vortex Epv values decline enough to allow mixing between the polar region and the midlatitudes. Gradients in temperature and tracer that were maintained during the winter decline as the final polar stratospheric warming occurs and pollutant concentrations rise at the pole. Another casualty during this period is the Aleutian anticyclone, which disappears by model day 78 (March 16) for 1979 and by day 85 (March 23) for 1989. For scenario M1.6 the breakup of the vortex is less evident during 1989.

5. Discussion

Near the end of the integration during March (model days 63-93), the peak NO y mixing ratios from aircraft perturbation, 0.35 ppbv (see Figure 2), and the global mean, 0.2 ppbv, are

small compared with the accepted estimates of NOy global mean background levels. March 50 mbar global mean nighttime values of HNO 3 and NO 2 from LIMS are about 3

ppbv. Total NO y estimated from 2-D model runs is about 1.5 ppbv [Douglass et al., 1992]. An EI of 15 g NOx/kg fuel instead of 5 g NO•/kg fuel as assumed here would yield peak values of 0.6-0.9 ppbv; these are much closer to the background values. Pollution levels using an EI of 5 will probably become significant with respect to the background values within a year of integration. To extend the integration the loss mechanism due to tropospheric rain out, which has been neglected in these calculations, needs to be included.

Time series of peak NMR values for 1979 and 1989 (not shown) are qualitatively similar to the time series for peak mixing ratios shown in Figure 2. These traces (not shown) are similar to Figure 2 except that mixing ratios are normalized with the global mean. Both years show values greater than 9 during first 20 days of the integration for the 1% average; thereafter, peak values decline and then stabilize at around 4 during the remainder of the run. The global mean value of tracer is initially zero. Therefore any injection of tracer is a significant change. As the integration proceeds, the global mean tracer values increase, so the relative impact of the injection is less.

WEAVER ET AL.: AIRCRAFT EMISSION TRANSPORT

Scenario M2.4 normalized mixing ratio 53 hPa 890222

180

27O

.o o o .o .... o k• L• -4 o

rd• 0 rd• 0

Norm01ized mixing

Figure 5. Normalized mixing ratio values for scenario M2.4 at 53 mbar on February 22, 1989. Seven backward isentropic trajectories ending on this date are shown in white. See text for details.

The distribution of pollutants at other levels can be analyzed qualitatively. For both years the NMR distribution at 250 mbar favors high values in the midlatitudes and a relatively clean tropical region. The distribution in the midlatitudes is zonally asymmetric; it is random over time with no one location showing a significant buildup of pollutant. The previous corridor study of Douglass et al. [1993] showed a good correlation between the distribution of tracer in the upper troposphere to 500-hPa geopotential height fields which suggests that the stratosphere troposphere exchange in the model is produced by upper tropospheric events in the midlatitudes.

The previous corridor study also showed that the Los Angeles to Sydney corridor has a high degree of zonal asymmetry in the tropical latitudes. This is due to relatively weak winds that are often highly variable with longitude and not fast enough to evacuate the corridor. The multicorridor emission scenario shown in Figure lb shows that the Los Angeles-Sydney corridor is one of many that cross the central and western Pacific. So it is not surprising that an equatorial projection of NMR for January 19 to February 23, 1989, does not show a pollutant asymmetry for this particular corridor.

However, the western equatorial Pacific and the Indian Ocean are favored regions of buildup compared with the equatorial Atlantic, consistent with the larger emission rates in these regions as shown in Figure lb.

6. Conclusion

Transport simulations of the evolution of aircraft exhaust for the 2 years examined are qualitatively similar. As expected, the tracer distributions at the aircraft flight altitude using all proposed flight corridors are more zonally symmetric than the single corridor simulations discussed by Douglass et al. [1993]. Peak perturbation mixing ratios are usually no more than a factor of 2 greater than the zonal mean and usually less than 5 times the global mean values. Since the maxima from these 3-D model results deviate from the zonal mean by a factor of 2 or less, the zonal mean assumption used in 2-D chemical models is reasonable.

NMR fields show a preference for pollutant buildup in the heavily traveled North Pacific and North Atlantic flight

WEAVER ET AL.: AIRCRAFF EMBSION TRANSPORT 5211

Z 50 hPa 790222

180

270

O0 r42 r42 r42 0 0

0 0'• I"',0 O0 • 0

km

Figure 6a. NMC geopotential heights at 50 mbar for February 22, 1979.

Z 50 hPa 890222

180

27O

b b L• b o o• I',O c•

km

Figure 6b. Same as Figure 6a but for February 22, 1989.

5212 WEAVER ET AL.: AIRCRAFF EMISSION TRANSPORT

Scenario M2.4 normalized mixing ratio 53 hPa Time mean 890224 -890313

180

27O

................................................ ,.• ....... :• ...... •. :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::

0 0 0 0 --•

b i,o b• ¬ o r• 0 • 0

Normolized mixing rotio

Figure 7. Time average of the normalized mixing ratio values for 1989 integration for scenario M2.4 at 53 mbar from February 24 - March 13.

corridors. However, neither corridor shows a strong preference for pollutant buildup over another. For the two winters considered, pollutants from the North Pacific routes are effectively trapped when emitted within the Aleutian anticyclone. Time averages from the integrations show a local maximum somewhere in the North Pacific. This is not

surprising, since the Aleutian anticyclone is a persistent pattern. We would expect similar behavior during other northern hemisphere winters. Trajectory calculations of parcels released in the Los Angeles-Tokyo corridor show pollutant buildup for 1980-1994 [Sparling et al., 1995]. The North Atlantic corridor is also defined by a local exhaust maximum, but in one instance was cleansed during a wavenumber 2 pat 'f I that placed a vortex lobe just north of the corridor. The t, ,pical flight corridors do not develop the same high pollutant levels found in the northern hemisphere

midlatitudes, on account of lower emissions. Because of the

moderate number of tropical corridors spread out over the Pacific, the slight zonal asymmetry in pollutant in the tropics is not due to any particular corridor but instead due to more air traffic in the tropical Pacific as compared to the Atlantic.

Future integrations of more than 1 year using assimilated winds with a realistic residual circulation will allow us to

examine exhaust behavior in summer, when 2-D models predict the greatest effect of aircraft emissions on stratospheric ozone and when trajectory calculations show the greatest potential for pollutant buildup [Sparling et al., 1995]. These integrations will also allow us to consider the long-term buildup at flight level and the importance of vertical transport to upper levels. These will provide further insight concerning the question of the applicability of 2-D models to evaluate the chemical effects of aircraft.

WEAVER ET AL.: AIRCRAFT EMISSION TRANSPORT 5213


180

27O

I I I

0 (Ha 0

Normolized mixing rotio

Figure 8. Time average of the normalized mixing ratio values for 1989 integration for scenario M2.4 at 53 mbar from March 14 to March 29.

Acknowledgments. We thank Mark R. Schoeberl for constructive comments on this manuscript. The authors acknowledge NASA Headquarters High Speed Research Program for support. This is SGCCP contribution #76.

References

Allen, D. J'., A. R. Douglass, and R. B. Rood, Application of a monotonic upstream transport scheme to three-dimensional constituent transport calculations, Mon. Weather Rev., 119, 2456-2464, 1991.

Considine, D. B., A. R. Douglass, and C. H. Jackman, Effects of a polar stratospheric cloud parameterization on ozone depletion due to stratospheric aircraft in a two-dimensional model, J. Geophys. Res., 99, 18,879-18,894, 1994.

Douglass, A. R., C. H. Jackman, R. B. Rood, A. C. Aikin, R. S. Stolarski, M.P. McCormick, and D. W. Fahey, Natural cycles, Gases, in The Atmospheric Effects of S tratospheric Aircraft: A First Program Report, edited by M. J. Prather and H. L. Wesoky, chap. 3a, 33-62, NASA Ref. Publ. 1272, 1992.

Douglass, A. R., R. B. Rood, C. J'. Weaver, M. C. Cerniglia, and K. F. Brueske, Implications of three-dimensional tracer

studies for two-dimensional assessments of the impact of supersonic aircraft on stratospheric ozone, J. Geophys. Res., 98, 8949-8963, 1993.

Leovy, C. B., C.-R. Sun, M. H. Hitchman, E. E. Remsberg, I. J. M. Russell, L. L. Gordley, J. C. Gille, and L. V. Lyjak, Transport of ozone in the middle stratosphere: Evidence for planetary wave breaking, J. Atmos. Sci., 42,230-244, 1985.

Newman, P. A., L. R. Lait, M. R. Schoeberl, R. M. Nagatani and A. J. Krueger, Meteorological atlas of the northern hemisphere lower stratosphere for January and February 1989 during the Airborne Arctic Stratospheric Expedition, NASA Tech. Memo., 4145, pp. 185, 1989.

Prather, M. J., Numerical advection by conservation of second- order moments, J. Geophys. Res., 91, 6671-6681, 1986.

Rasch, P. J., X. Tie, B. A. Boville, and D. L. Williamson, A three-dimensional transport model for the middle atmosphere, J. Geophys. Res., 99, 999-1017, 1994.

Rood, R. B., D. J. Allen, W. E. Baker, D. J. Lamich, and J. A. Kaye, The use of assimilated stratospheric data in constituent transport calculations, J. Atmos. Sci., 46, 687- 701, 1989.

Rood, R. B., A. R. Douglass, J. A. Kaye, M. A. Geller, Chi Yuechen, D. J. Allen, E. M. Larson, E. R. Nash, and J. E.

5214 WEAVER ET AL.: AIRCRAFT EMISSION TRANSPORT

Nielsen, Three-dimensional simulations of wintertime ozone variability in the lower stratosphere, J. Geophys. Res., 96, 5055-5071, 1991.

Rood, R. B., J. E. Nielsen, R. S. Stolarski, A. R. Douglass, J. A. Kaye, and D. J. Allen, Episodic total ozone minima associated effects on heterogeneous chemistry and lower startospheric transport, J. Geophys. Res., 97, 7979-7996, 1992.

Schoeberl, M. R. and L. C. Sparling, Trajectory modeling, in Diagnostic Tools in Atmospheric Physics, Proceedings of the International School of Physics, Course CXIXl, edited by 13. Fiocco and 13. Visconti, North Holland, Amsterdam, 1995.

Schubert, S. D., R. B. Rood, and J. Pfaendtner, An assimilated dataset for Earth science applications, Bull. Am. Meteorol. Soc., 74, 2331-2342, 1993.

Smith, A. K., J. t21. Gille, and L. V. Lyjak, Wave-wave interaction in the stratosphere: Observations during quiet and active wintertime periods, J. Atmos. Sci., 41,363-373, 1984.

Sparling, L. C., M. R. Schoeberl, A. R. Douglass, C. J. Weaver, P. A. Newman, and L. R. Lair, Trajectory modeling of emissions from lower stratospheric aircraft, J. Geophys. Res., 100, 1427-1438, 1995.

van Leer, B., Towards the ultimate conservative difference

scheme, H, Monotonicity and conservation combined in a second order scheme, J. Cornput. Phys., 14,361-370, 1974.

Weaver, C. J., A. R. Douglass, and R. B. Rood, Thermodynamic balance of three-dimensional stratospheric winds derived from a data assimilation procedure, J. Atmos. Sci., 50, 2987-2993, 1993.

Wuebbles, D. J., D. Maiden, R. K. Seals Jr., S. L. Baughcum, M. Metwally, and A. Mortlock, Emissions scenarios development: Report of the Emissions Scenarios Committee, The Atmospheric Effects of $tratospheric

Aircraft: A Third Program Report.,edited by R. S. Stolarski and H. L. Wesoky, NASA Ref. Publ. 1313, chap. 3, 63-208, 1993.

A. R. Douglass and R. B. Rood, Laboratory for Atmospheres, NASA Goddard Space Flight Center, Greenbelt, MD 20771.

C. J. Weaver, Code 916, NASA Goddard Space Flight Center, Greenbelt, MD 20771. (e-mail: [email protected])

(Received August 22, 1994; revised December 2, 1994; accepted December 8, 1994.)

tracer transport for realistic aircraft emission scenarios

Documents