Global NOx Production by ... - Springer Link

Journal of Atmospheric Chemistry 43: 61–74, 2002. © 2002 Kluwer Academic Publishers. Printed in the Netherlands.

61

Global NOx Production by Lightning XUEXI TIE 1 , RENYI ZHANG 2, GUY BRASSEUR 3 and WENFANG LEI 2 1 National Center for Atmospheric Research, Boulder, Colorado, U.S.A. 2 Department of Atmospheric Sciences, Texas A&M University, College Station, Texas, U.S.A. 3 Max-Planck Institute for Meteorology, Hamburg, Germany

(Received: 21 May 2001; in final form: 30 November 2001) Abstract. Lightning is thought to represent an important source of tropospheric reactive nitrogen species NOx (NO + NO2 ), but estimates of global production of NOx by lightning vary considerably. We evaluate the production of NOx by lightning using a global chemical/transport model, satellite lightning observations, and airborne NOx measurements. Various model calculations are conducted to assess the global NOx production rate of lightning by comparing the model calculations with airborne measurements. The results show that the simulated NOx in the tropical middle and upper troposphere are very sensitive to the amount and altitude of the lightning NOx used in the model. A global lightning NOx production of 7 Tg N yr−1 uniformly distributed in convective clouds or 3.5 Tg N yr−1 distributed in the upper cloud regions produces good agreement between calculated and measured NOx concentrations in the tropics. Key words: lightning, NOx .

1. Introduction Reactive nitrogen species NOx (NO + NO2 ) play an important role in influencing tropospheric ozone concentrations. Crutzen (1973) and Chameides and Walker (1973) suggest that oxidation of CH4 , CO, and other non-methane hydrocarbons (NMHCs) in presence of NOx leads to a significant amount of ozone production in the troposphere. NOx is emitted into the atmosphere from various sources, including fossil fuel combustion, biomass burning, aircraft emission, oxidation of ammonia, and lightning (e.g., Brasseur et al., 1996; Seinfeld and Pandis, 1998). Model studies suggest that lightning can significantly affect the NOx budget in the tropical troposphere (Lawrence et al., 1995; Lamarque et al., 1996; Tie et al., 2001a). In particular, the recent study by Tie et al. (2001a) reveals that lightning not only affects NOx concentrations, but also enhances nitrogen reservoirs (HNO3 , PAN, etc.). Among the nitrogen species, HNO3 is mostly influenced, comprising approximately 60–80% of the total increase of the nitrogen species. Furthermore, the enhancement of the nitrogen reservoir species by lightning has been shown to be responsible for NOx increases over remote tropical regions due to recycling of the nitrogen reservoir species. Atmospheric measurements also reveal a considerable enhancement of NOx by lightning near convective storms in the middle

62

XUEXI TIE ET AL.

and upper troposphere (Ridley et al., 1996; Huntrieser et al., 1998; Stith et al., 1998; Zhang et al., 2000). A large-scale ozone enhancement due to lightning has been inferred from a multiple-year record of tropical troposphere ozone columns (TTOCs) (Martin et al., 2000). In addition, lightning is also believed to represent an important source of NOx in the regional atmosphere (Bond et al., 2001). The global production of NOx by lightning has been a subject of numerous studies (e.g., Liaw et al., 1990; Price et al., 1997; Nesbitt et al., 2000). Estimates of the global production of NOx by lightning requires understanding of the lightning physical characteristics, the chemical mechanism of NOx formation by lightning, and the global and seasonal lightning distributions, all of which are not well characterized presently. In particular, there is considerable uncertainty about the NOx production rate by individual lightning flashes, with reported values from 0.23 × 1026 to 21 × 1026 NO molecules for each cloud-to-ground (CG) flash (e.g., Price et al., 1997). Inevitably, such uncertainty leads to an estimate of global NOx production rate by lightning that varies considerably from previous studies (Nesbitt et al., 2000). Hence lightning represents the largest uncertainty among the various NOx sources (Price et al., 1997; Nesbitt et al., 2000). In this study, we have used a global three-dimensional chemical/transport model to investigate the magnitude, geographical distribution, and vertical distribution of the global lightning NOx emission. Two sources of the global lightning NOx production were considered. A parameterization based on the convective cloud height was used to calculate lightning-produced NOx according to Price et al. (1997). We also used the global NOx production estimated from satellite lightning observations of the Optical Transient Detector (OTD) (Nesbitt et al., 2000). The model results are then compared to airborne measurements from PEM-West-A, PEM-Tropics, TRACE-A, CITE-2, CITE-3, ECHEM, and ABLE to constrain the global production of NOx due to lightning.

2. Method 2.1. MODEL DESCRIPTION The numerical model used in the present study is a global three-dimensional chemical/transport model (MOZART) developed at the National Center for Atmospheric Research (NCAR) (Brasseur et al., 1998). The model is a comprehensive tropospheric chemical/transport model, calculating the global distribution of 56 gas-phase chemical species. Version 1 of the model is configured with a T42 (2.8◦ × 2.8◦ ) horizontal resolution and 25 hybrid vertical levels ranging from the surface to 4 mb. Meteorological information (i.e., winds, temperature, etc.) is provided by the NCAR Community Climate Model (CCM-2) every 3 hours on the basis of pre-calculated results (off-line). The model time step for chemistry and transport is 20 minutes. Chemical species are transported by advection (Rasch and Williamson, 1991), vertical diffusion (Holtslag and Boville, 1993), and convection

GLOBAL NOx PRODUCTION BY LIGHTNING

63

(Hack, 1994). Details on the treatment of the gas-phase chemical mechanisms and transport processes are described by Brasseur et al. (1998). The nitrogen species in the model include NOx (NO + NO2 ), NO3 , PAN (CH3 CO3 NO2 ), and HNO3 , N2 O5 , HNO4 , and NOy (NOx + NO3 + PAN + HNO3 + 2N2 O5 + HNO4 ). Heterogeneous reaction of N2 O5 on sulfate aerosols is also included; this reaction provides an important pathway converting NO2 to HNO3 at high latitudes in the Northern Hemisphere during winter (Dentener and Crutzen, 1993; Tie et al., 2001b). An important daytime loss process of NOx is the reaction of NO2 reaction with OH to form HNO3 which is ultimately removed from the atmosphere by wet deposition to the ground. The global OH budget is obtained by calculating the CH4 chemical lifetime. The model estimated lifetime of CH4 is about 10 years which is comparable or slightly higher than previous estimates (Prinn et al., 1987; Tie et al., 1992). Lightning occurs only in well developed convective clouds, and thus the convective transport is an important process to re-distribute NOx released from lightning (Pickering et al., 1998). In MOZART (version 1), the convective transport scheme is developed by Hack (1994), which is a step-wise vertical diffusion scheme according to atmospheric instability. This method has been suggested to produce a slower mixing in the cloud compared to the deep convection scheme developed by Zhang and McFarlance (1995). However, both schemes contain uncertainties and need to be validated in the future. In the present study, we conduct two sets of model simulations. First, NOx produced by lightning is uniformly distributed in the clouds as a constant mass throughout the total column, and is re-distributed by convective transport and other processes. Second, lightning-produced NOx is released at the top four levels of the clouds, which can be considered to resemble the fast convective transport, such as the scheme suggested by Zhang and McFarlance (1995). As we will illustrate in Section 3, an increase in convective transport (simulated by direct NOx release at the top of the convective clouds) can lead to a difference of a factor of two in the estimated global budget of lightning-produced NOx . The surface emission of NOx is based on Muller (1992) with some adjustments to account for more recent data. The model contains all possible surface NOx emission sources, including fossil fuel combustion, biomass burning, and soil release. The total surface emission strength of NOx is 32.5 Tg N yr−1 which is within the range estimated by WMO (1995).

2.2. FIELD NOx MEASUREMENTS The field NOx data considered in this study are complied by Emmons et al. (1997), based on campaigns from PEM West-A, PEM Tropics, CITE-2, CITE-3, ECHEM, ABLE, and TRACE-A. Because the chemical lifetime of NOx is relatively short, we have chosen geographical sites that are located at or near active lightning regions in the tropics and in the United States. In the tropical troposphere NOx concentrations are significantly enhanced by lightning emission (Tie et al., 2001a),

64

XUEXI TIE ET AL.

Table I. Locations and date of the NOx measurements for comparison with model results No.

Campaign

Date

Region

Lat., lon.

1 2 3 4 5 6 7 8 9 10 11 12 13 14

PEM-West-A PEM-West-A PEM-West-A PEM-West-A PEM-Tropics TRACE-A TRACE-A TRACE-A CITE-3 TRACE-A TRACE-A ECHEM CITE-2 ABLE

Sep 16–Oct 21, 1991 Sep 16–Oct 21, 1991 Sep 16–Oct 21, 1991 Sep 16–Oct 21, 1991 Aug 30–Oct 5, 1996 Sep 21–Oct 26, 1992 Sep 21–Oct 26, 1992 Sep 21–Oct 26, 1992 Aug 22–Sep 29, 1989 Sep 21–Oct 26, 1992 Sep 21–Oct 26, 1992 Jul 07–Aug 22, 1989 Aug 11–Sep 05, 1986 Jul 06–Aug 15, 1990

China_Coast_E Japan_Coast_E Philippine sea Pacific_Tropics_W Pacific Africa_S Africa_Coast_W Atlantic_S Natal Brazil_E Brazil_Coast U.S. U.S. U.S.

20–30◦ N, 115–130◦ E 25–40◦ N, 135–150◦ E 5–20◦ N, 135–150◦ E 5◦ S–15◦ N, 155–165◦ E 30–10◦ S 170–190◦ E 25–5◦ S, 15–35◦ E 25–5◦ S, 0–10◦ E 20◦ S–0, 340–350◦ E 5◦ S–5◦ N, 325–335◦ E 15–5◦ S, 310–320◦ E 35–25◦ S, 310–320◦ E 30–35◦ N, 250–255◦ E 35–45◦ N, 235–250◦ E 35–45◦ N, 280–290◦ E

but are less impacted by industrial emission, stratospheric intrusion, and aircraft emission (Lamarque et al., 1996). As a result, the tropical region is suitable for evaluating the impact of lightning on NOx by comparing calculations with observations. In the continental United States, the lightning activities are quite strong in the summer (Bond et al., 2001). Thus, the NOx observation in the upper troposphere during the summer also provides valuable information on lightning production of NOx . The measurements of each campaign have been categorized by the geographical location and sampling date, as summarized in Table I. From the locations of the observation sites (see Table I) and the distribution of lightning activities (see Figure 1), we can characterize the observation sites that whether they are close or far away from active thunderstorm. Sites 1 and 2 are near moderate lightning activities from Japan and China. Sites 3, 4 and 5 are over the Pacific, and are far away from lightning activities, and surface emissions. Site 6 is in the south of Africa, and is near strong lightning activities, and close to strong surface emissions. Site 7 is on the coast of the south of the continent of Africa, and is not too far from strong lightning activities. Site 8 is over the Atlantic Ocean, and has relatively weak lightning activities. Sites 9 and 11 are on coast of South America, and is not too far from strong lightning activities. Site 10 is inland in South America, and is close to both strong lightning activities and surface emissions. Sites 12, 13, and 14 are located in the continental United States, and are close to strong lightning activities during the summer.


65

Figure 1. Global horizontal distribution of lightning NOx emission averaged for August, September, and October: (a) parameterization according to Price et al. (1997); (b) data estimated from OTD satellite data (Nesbitt et al., 2000).

66

XUEXI TIE ET AL.

2.3. GLOBAL NOx EMISSION BY LIGHTNING Two sources of global lightning NOx emissions were considered in this work. One is the model calculated global lightning NOx emission according to the method developed by Price et al. (1997). In their study, the authors parameterized the global and seasonal distributions of lightning-produced NOx , based on the observed distribution of electrical storms, the cloud heights, and the physical properties of lightning strokes. In the MOZART model the parameterization by Price et al. (1997), based on convective cloud height, was used to calculate lightning NOx production. The lightning NOx production is expressed as P = A × H exp(B), where A is 3.44 × 10−5 for continental thunderstorms and 6.40 × 10−4 for marine thunderstorms. H is the height of the convective clouds, and B is 4.92 and 1.73 for continental and marine thunderstorm, respectively. As mentioned above, we have made several assumptions regarding the vertical distribution of the NOx production within the clouds (uniformly distributed or distributed at the top four levels of the clouds). We also employed the global NOx emission data estimated from satellite observations of the Optical Transient Detector (OTD) (Nesbitt et al., 2000). OTD is the first long-term observer of lightning from space on the MicroLab-1 satellite, which was launched on 3 April 1995 into an orbital altitude of 710 kilometers and an inclination of 70 degrees (Christian and Latham, 1998). The satellite orbits the earth approximately once every 100 minutes. The OTD sensor has a 100-degree viewing angle that facilitates a 1300 × 1300 km viewing footprint. This allows viewing of approximately 1/300 of the total surface area of the earth at a particular moment. The instrument has a spatial resolution of 10 km in the horizontal and a temporal resolution of 2 ms. The OTD detects lightning flashes with an efficiency between 40% and 65%, depending on viewing conditions such as sun glint and radiation (Boccippio et al., 2001). The use of the OTD data to obtain the global NOx production by lightning requires extrapolation because of the relatively small viewing area of the OTD satellite (Nesbitt et al., 2000). 3. Results and Discussions The global distribution of lightning NOx emission derived from the satellite OTD data (Nesbitt et al., 2000) is compared to the parameterization by Price et al. (1997) in Figure 1. To examine the difference in geographical distributions between the two results, we scaled both data sets by an annual NOx emission rate of 7 Tg N yr−1 . Figure 1 indicates that there is a similarity between the parameterized lightning NOx emission (Price et al., 1997) and the OTD data, showing that the most lightning NOx emission is located in the Amazon, central Africa, and South Asia. However, there is a noticeable difference in the lightning activity between the modeled and OTD data, especially in South Asia where the modeled lightning emission is about twice as high as the OTD values. In addition, the OTD data reveals negligible lightning-NOx production over the Sahara desert, which is in

67


Table II. Description of the model calculations No.

Symbol

Source

Magnitude (Tg N yr−1 )

1 2 3 4 5 6

LT0 LT1 LT2 LT3 LT4 LT5

No lightning-NOx Parameterization a OTD b Parameterization a Parameterization a Parameterization a

0.0 7.0 7.0 7.0 1.0 3.5

Altitude

Uniformed in cloud Uniformed in cloud In upper cloud Uniformed in cloud In upper cloud

a Lightning NO according to Price et al. (1997). x b Lightning NO estimated by Nesbitt et al. (2000). x

contrast to the parameterization according to Price et al. (1997). Because the distribution of lightning NOx production is based on the distribution of cloud height, it appears that the convective clouds are overestimated near the Sahara in the model. In general, lightning activity depends on the incoming solar heating, and on the geographical distribution of land/ocean surfaces. In addition, moisture, and temperature gradient are also important factors affecting lightning activity. The most intensive lightning activity occurs over land where the surface heating is maximal. Hence the seasonal variation of lightning activity is significant. For example, in June, intensive lightning occurs at low latitudes in the Northern Hemisphere, while in December the lightning activity occurs primarily at low latitudes in the Southern Hemisphere. Marine convection, although occurring throughout the year, yields little lightning over the central Atlantic and Pacific (Nesbitt et al., 2000). To constrain the global lightning NOx emission and to reduce its uncertainty, various model simulations were conducted, as summarized in Table II. Simulation 1 (LT0) represents a model calculation without lightning NOx emission. Simulation 2 (LT1) is a model calculation according to the parameterized lightning NOx emission of 7 Tg N yr−1 (Price et al., 1997) uniformly distributed with height below the top of convective clouds and covers the entire grid-cell in which clouds are present. Simulation 3 (LT2) is the same as LT1, except that the OTD lightning distribution is implemented by multiplying a scaling factor of O(x, y)/M(x, y), where O(x, y) is the monthly mean lightning NOx distribution from OTD and M(x, y) is the monthly mean lightning NOx distribution from the model. Simulations 4 (LT3) and 5 (LT4) are also similar to LT1, except that the NOx production within clouds is distributed only in the top four levels of the convective clouds in LT3 and that the total NOx production from lightning is reduced to 1 Tg N yr−1 in LT4, respectively. Simulation 6 (LT5) is the same as LT3, except that the total NOx production from lightning is reduced to 3.5 Tg yr−1 . Figure 2 shows the calculated zonally averaged changes in the NOx concentration due to lightning from different simulations, i.e., the difference between

68

XUEXI TIE ET AL.

Figure 2. Zonally averaged changes of NOx due to lightning from different simulations in September. The simulation information is given in Table II.


69

the simulation without lightning-NOx and the one including lightning contribution with an assumed strength and height for the lightning-NOx emission. The results suggest that lightning has a major impact on NOx in the tropical upper troposphere, increasing NOx concentrations by 20 to 200 pptv. This corresponds to about 40 to 90% increases in the NOx concentrations simulated with and without the lightning-produced NOx for the different cases. As suggested by Lamarque et al. (1996), with 5 Tg N yr−1 lightning production of NO, the NOx concentrations in the tropics increase by about 70%. However, the percentage can be as large as 600% if the changes in NOx are relevant to the lightning-free case in which the NOx concentrations in the tropics are typically very small (Tie et al., 2001a). With the smallest modeled NOx production by lightning (i.e., 1 Tg N yr−1 in LT4), the NOx concentrations increase by 20 to 40 pptv in the tropical upper troposphere. The LT3 simulation produces the largest increase in NOx concentrations (up to 200 pptv), indicating that lightning NOx production directly emitted in the upper troposphere plays a significant role in controlling the NOx concentration in this region. The results of the LT5 simulation provide further evidence for the importance of the upper tropospheric lightning NOx release on the NOx concentration. Figure 2 shows that the results from LT1 and LT5 are quite similar in the upper troposphere. However, the different amounts of lightning NOx emission of 7 Tg N yr−1 in LT1 and 3.5 Tg N yr−1 in LT5 suggest that the direct emission of NOx in the lower troposphere plays an insignificant role in controlling the NOx concentration. As discussed by Tie et al. (2001a), the chemical lifetime of NOx increases from a few hours in the planetary boundary layer to a few days in the upper troposphere. Hence the chemical conversion of NOx to HNO3 occurs rapidly in the lower tropical troposphere, and the major enhancement of NOx from direct lightning emission is confined mainly to the upper troposphere. Figure 3 compares the NOx concentrations between the model calculations and observations made during several aircraft field campaigns. The various sites were selected to represent the different levels of lightning activity. Sites 1 and 2 are located in China and Japan, with relatively active lightning activity. The results show that without lightning NOx emission the calculated NOx concentrations are significantly lower than the observations. LT1 and LT5 produce reasonable results compared with the observed values at site 1. At site 2, LT1 and LT5 tend to slightly underestimate and LT3 tends to slightly overestimate the NOx concentration. Sites 3 and 4 are located on the East Coast of Asia where NOx is not directly emitted by lightning. As a result, enhancement of NOx concentrations due to lightning is visible, but is much smaller than that at sites 1 and 2. The simulations of LT1 and LT3 have a better representation of the NOx concentrations compared to the observations in those locations. Site 5 is located over the Pacific Ocean and is distant from Australia. All model simulations underestimate the NOx concentration from the surface to 12 km at this site, indicating an underestimation of NOx emission from the nearby continental regions or the recycling of NOx from nitrogen reservoirs (HNO3 and PAN) at this site (Tie et al., 2001a). Similarly, at the oceanic sites 7,

70

XUEXI TIE ET AL.

Figure 3. Comparison of NOx between model simulations (lines) and observations (squares) in eleven locations. The observation sites are summarized in Table I.

8, and 9, all simulations except LT3 underestimate the NOx concentrations. Sites 6, 10, and 11 are located in regions of high lightning activity (Africa and South America), and they are also located in regions of high biomass burning activities. At these sites, lightning shows a strong impact on the NOx concentration, with both LT1 and LT5 producing better agreement with the observations. Sites 12, 13, and 14 are located in the United States. During the summer months (i.e., June, July, and August), strong lightning activities occur over the U.S. (Bond et al., 2001). Thus,


71

lightning shows a major impact on the NOx concentrations during July and August, especially at site 12 (at sites 13 and 14, NOx observations are limited below 6 km), with LT5 producing the best agreement with the observations. Hence the model calculations indicate that in the tropical troposphere NOx production from lightning plays an important role in controlling the NOx concentrations, resulting in more than 100 pptv enhancement. Over oceanic locations, the model simulations often underestimate the NOx concentrations at all altitudes. This may occur as a result of inadequate representation of the NOx emission from the nearby continental regions or the recycling of NOx from nitrogen reservoirs (HNO3 and PAN) in the model. The geographical difference in the lightning activity between the parameterization by Price et al. (1997) and satellite OTD data (Nesbitt et al., 2000) produced some noticeable differences in the calculated NOx concentrations, especially in South Asia. The vertical distribution of NOx release by lightning is also important in controlling the NOx concentrations. The model simulation using a global NOx production of 3.5 Tg N yr−1 emitted in the upper levels of clouds (LT5) or 7 Tg N yr−1 emitted uniformly in clouds (LT1) produces similar NOx concentrations in the upper troposphere. As mentioned before, the direct release of lightning-produced NOx at the top of convective clouds can be considered to represent the case of fast convective transport of NOx from the lower part to the upper part of the clouds. Hence the uncertainty in estimating the global NOx production by lightning due to convective transport is approximately a factor of two. Figure 3 also suggests that the most significant influence on NOx concentrations due to lighting NOx production is located above 5 km. Table III summarizes the calculated NOx concentrations averaged above 5 km for various lightning NOx source strengths and vertical releases and the comparison to the observed NOx values. The NOx values are averaged from 14 observation sites indicated Table I. The percentage differences between the observations and calculations are given in the third column. Table I indicates that on the average LT0 underestimates NOx by 87%; LT1 underestimates NOx by 17%; LT2 overestimates NOx by 15%; LT3 overestimates NOx by 63%; LT4 underestimates NOx by 67%; and LT5 underestimates NOx by 22%. The simulations with LT1, LT2, and LT5 produce better agreements between observed and calculated NOx concentrations in the upper troposphere. Hence, a global lightning NOx production of 7 Tg N yr−1 uniformly distributed in clouds or 3.5 Tg N yr−1 distributed in the upper cloud regions produces less disagreement between calculated and measured NOx concentrations in the tropics. 4. Conclusions We have assessed the global production of NOx by lightning using a global chemical/transport model, satellite lightning observations, and airborne NOx measurements. From sensitivity studies with respect to the geographical distributions, magnitudes, and vertical positions of NOx emission from lightning, the model

72

XUEXI TIE ET AL.

Table III. Calculated NOx concentrations averaged above 5 km for different vertical NOx releases by lightning and comparison to observed NOx values. The NOx values are averaged from 14 observation sites indicated Table I Symbol

NOx (pptv)

NOx (%) (Obs–Cal)/Obs

Obs LT0 LT1 LT2 LT3 LT4 LT5

109 14 90 126 178 35 84

–87 –17 +15 +63 –67 –22

simulations suggest that NOx production from lightning plays an important role in controlling the NOx concentrations, producing more than 500% percent increase in the tropical troposphere. The most important impact of lightning on NOx is in the upper troposphere and over land areas. The simulations with a global lightning NOx emission of 7 Tg N yr−1 uniformly distributed in clouds or 3.5 Tg N yr−1 distributed in the upper regions of clouds produce the best agreement between calculated and measured NOx concentrations over continents.

Acknowledgements The work of X. Tie was partially supported by the DOE Atmospheric Chemistry Program under contract DE-AI05-98ER62579. Renyi Zhang and Wenfang Lei were supported by the NASA New Investigator Program in Earth Science and Texas Air Research Center (TARC). The National Center for Atmospheric Research (NCAR) is sponsored by the National Science Foundation. The referees provided valuable comments for improving this manuscript.

References Brasseur, G. P., Muller, J. F., and Granier, C., 1996: Atmospheric impact of NOx emission by subsonic aircraft: A three-dimensional model study, J. Geophys. Res. 101, 1423–1428. Boccippio, D. J., Cummins, K. L., Christian, H. J., and Goodman, S. J., 2001: Combined satellite and surface-based estimation of intracloud/cloud-to-ground lightning ratio over the continental United States, Mon. Wea. Rev. 129, 108–122. Bond, D. W., Zhang, R., Tie, X., Brasseur, G., Huffines, G., Orville, R. E., and Boccippio, D. J., 2001: NOx production by lightning over the continental United States, J. Geophys. Res. 106, 27701–27709.


73

Brasseur, G. P., Hauglustaine, D. A., Walters, S., Rasch, P. J., Müller, J.-F., Granier, C., and Tie, X. X., 1998: MOZART, a global chemical transport model for ozone and related chemical tracers. 1. Model description, J. Geophys. Res. 103, 28,265–28,289. Chameides, W. L. and Walker, J. C. G., 1973: A photochemical theory of tropospheric ozone, J. Geophys. Res. 78, 8751–8760. Christian, H. J., Blakeslee, R. J., and Goodman, S. J., 1989: The detection of lightning from geostationary orbit, J. Geophys. Res. 94, 13,329–13,337. Christian, H. J. and Latham, J., 1998: Satellite measurements of global lightning, Quart. J. Roy. Meteorol. Soc. 124, 1771–1773. Crutzen, P. J., 1973: A discussion of the chemistry of some minor constituents in the stratosphere and troposphere, Pure. Appl. Geophys. 106, 1385–1399. Dentener, F. J. and Crutzen, P. J., 1993: Reaction of N2 O5 on tropospheric aerosols: Impact on the global distribution of NOx , O3 , and OH, J. Geophys. Res. 98, 7149–7163. Emmons, L. K., Carroll, M. A., Hauglustaine, D. A., Brasseur, G. P., Atherton, C., Penner, J., Sillman, S., Levy, H., Rohrer, F., Wauben, W. M. F., VanVelthoven, P. F. J., Wang, Y., Jacob, D., Bakwin, P., Dickerson, R., Doddridge, B., Gerbig, C., Honrath, R., Hubler, G., Jaffe, D., Kondo, Y., Munger, J. W., Torres, A., and Volz-Thomas, A., 1997: Climatologies of NOx and NOy : A comparison of data and models, Atmos. Environ. 31, 1851–1903. Hack, J. J., 1994: Parameterization of moist convection in the NCAR community climate model (CCM2), J. Geophys. Res. 99, 5551–5568. Holtslag, A. and Boville, B. A., 1993: Local versus nonlocal boundary-layer diffusion in a global climate model, J. Clim. 6, 1825–1842. Huntrieser, H., Schlager, H., Feigl, C., and Holler, H., 1998: Transport and production of NOx in electrified thunderstorms: Survey of previous studies and new observations at midlatitudes, J. Geophys. Res. 103, 28,247–28,264. Lamarque, J. F., Brasseur, G. P., Hess, P. G., and Muller, J. F., 1996: Three-dimensional study of the reactive contributions of the different nitrogen sources in the troposphere, J. Geophys. Res. 101, 22,955–22,968. Lawrence, M. G., Chameides, W. L., Kasibhatla, P. S., Levy II, H., and Moxim, W., 1995: Lightning and atmospheric chemistry: The rate of atmospheric NO production, in H. Volland (ed.), Handbook of Atmospheric Electrodynamics, CRC Press, Boca Raton, FL, pp. 189–202. Martin, R. V., Jacob, D. J., Logan, J. A., Ziemke, J. M., and Washington, R., 2000: Detection of a lightning influence on tropical tropospheric ozone, Geophys. Res. Lett. 27, 1639–1642. Muller, J. F., 1992: Geographical distribution and seasonal-variation of surface emissions and deposition velocity of atmospheric trace gases, J. Geophys. Res. 94, 3787–3804. Nesbitt, S. W., Zhang, R., and Orville, R. E., 2000: Seasonal and global NOx production by lightning estimated from the optical transient detector (OTD), Tellus B 52, 1206–1215. Pickering, K. E., Wang, Y. S., Tao, W. K., Price, C., and Muller, J. F., 1998: Vertical distributions of lightning NOx for use in regional and global chemical transport model, J. Geophys. Res. 103, 31,203–31,216. Price, C., Penner, J., and Prather, M., 1997: NOx from lightning. 1. Global distribution based on lightning physics, J. Geophys. Res. 102, 5929–5941. Prinn, R., Cunnold, D., Rasmussen, R., Simmonds, P., Alyea, F., Crawford, A., Fraser P., and Rosen, R., 1987: Atmospheric trends in methyl chloroform and the global average for the hydroxyl radical, Science 238, 945–950. Rasch, P. J. and Williamson, D. L., 1990: Computational aspects of moisture transport in global models of the atmosphere, Quart. J. Roy. Meteorol. Soc. 116, 1071–1090. Ridley, B. A., Dye, J. E., Walega, J. G., Zheng, J., Grahek, F. E., and Rison, W., 1996: On the production of active nitrogen by thunderstorms over New Mexico, J. Geophys. Res. 101, 20,985– 21,005.

74

XUEXI TIE ET AL.

Seinfeld, J. H. and Pandis, S. N., 1998: Atmospheric Chemistry and Physics, John Wiley and Sons Inc., New York. Stith, J., Dye, J., Ridley, B., Laroche, P., Defer, E., Baumann, K., Hubler, G., and Venticinque, R., 1999: NO signature from lightning flashes, J. Geophys. Res. 104, 16,081–16,089. Tie, X., Kao, C. Y., Mroz, E. J., Cicerone, R. J., Alyea, F. N., and Cunnold, D. M., 1992: Threedimensional simulations of atmospheric methyl chloroform: Effect of an ocean sink, J. Geophys. Res. 97, 20751–20769. Tie, X., Zhang, R., Brasseur, G., Emmons, L., and Lei, W., 2001a: Effects of lightning on reactive nitrogen and nitrogen reservoir species in the troposphere, J. Geophys. Res. 106, 3167–3178. Tie, X., Brasseur, G., Emmons, L., Horowitz, L., and Kinnison, D., 2001b: Effects of aerosols on tropospheric oxidants: A global model study, J. Geophys. Res. 106, 22931–22964. Zhang, R., Sanger, N. T., Orville, R. E., Tie, X. X., Randel, W., and Williams, E. R., 2000: Enhanced NOx by lightning in the upper troposphere and lower stratosphere inferred from the UARS global NO2 measurements, Geophys. Res. Lett. 27, 685–688. Zhang, G. J. and McFarlance, N. A., 1995: Sensitivity of climate simulations to the parameterization of cumulus convection in the Canadian Climate Center general circulation model, Atmos. Ocean 33, 407–446. WMO, 1995: Scientific assessment of ozone depletion: 1994, Report No. 37.