Kenneth E Pickering

1 INTRODUCTION

In the early 1980s it was recognized that observed free tropospheric mixing ratios of some trace gases could not be explained simply by large-scale transport and eddy diffusion. Crutzen and Gidel (1983), Gidel (1983), and Chatfield and Crutzen (1984) hypothesized that convective clouds played an important role in rapid atmospheric vertical transport of trace species and tested parameterizations of convective transport in atmospheric chemical models. At nearly the same time evidence was shown of venting of the boundary layer by shallow fair weather cumulus clouds (e.g., Greenhut et al., 1984; Greenhut, 1986). Field experiments were conducted in 1985 that resulted in verification of the hypothesis that deep convective clouds are instrumental in atmospheric transport of trace constituents (Dickerson et al., 1987; Garstang et al., 1988). Once pollutants are lofted to the middle and upper troposphere, they typically have a much longer chemical lifetime and with the generally stronger winds at these altitudes they can be transported large distances from their source regions. Photochemical reactions occur during this long-range transport. Pickering et al. (1990) demonstrated that venting of boundary layer pollutants by convective clouds (both shallow and deep) causes enhanced ozone production in the free troposphere. Therefore, convection aids in the transformation of local pollution into a contribution to global atmospheric pollution.

Field studies have established that downward transport of larger O3 and NOx mixing ratios from the free troposphere to the boundary layer is an important process over the remote oceans (e.g., Piotrowicz et al., 1991), as well as the upward transport of very low 03 mixing ratios from the boundary layer to the upper troposphere (Kley et al., 1996). Global modeling by Lelieveld and Crutzen (1994) suggests that the downward mixing of 03 into the boundary layer is the dominant global effect of

Handbook of Weather, Climate, and Water: Atmospheric Chemistry, Hydrology, and Societal Impacts, Edited by Thomas D. Potter and Bradley R. Colman. ISBN 0-471-21489-2 © 2003 John Wiley & Sons, Inc.

deep convection. Some indications of downward transport of 03 from higher altitudes (possibly from the stratosphere) in the anvils of thunderstorms have been observed (Dickerson et al., 1987; Poulida et al., 1996; Suhre et al., 1997). Ozone is most effective as a greenhouse gas in the vicinity of the tropopause. Therefore, changes in the vertical profile of 03 in the upper troposphere caused by deep convection have important radiative forcing implications for climate.

More detailed discussion of observations of convective transport are presented in Section 2. Simulation of convective transport in cloud-resolving models and its parameterization in larger-scale models is discussed in Section 3, as well as implications for 03 production following convective redistribution.

2 OBSERVATIONS

Venting by Nonprecipitating Cumulus Clouds

Some fraction of shallow fair weather cumulus clouds actively vent boundary layer pollutants to the free troposphere (Stull, 1985). The first airborne observations of this phenomenon were conducted by Greenhut et al. (1984) over a heavily urbanized area, measuring the in-cloud flux of ozone in a relatively large cumulus cloud. An extension of this work was reported by Greenhut (l 986) in which data from over 100 aircraft penetrations of isolated nonprecipitating cumulus clouds over rural and suburban areas were obtained. Ching and Alkezweeny (1986) reported tracer (SF6) studies associated with nonprecipitating cumulus (fair weather cumulus and cumulus congestus). Their experiments showed that the active cumulus clouds transported mixed layer air upward into the overlying free troposphere and suggested that active cumuli can also induce rapid downward transport from the free troposphere into the mixed layer. A UV-DIAL (ultraviolet differential absorption lidar) provided space-height cross sections of aerosols and ozone over North Carolina in a study of cumulus venting reported by Ching et al. (1988). Data collected on evening flights showed regions of cloud debris containing aerosol and ozone in the lower free troposphere in excess of background, suggesting that significant vertical exchange had taken place during afternoon cumulus cloud activity. Efforts have also been made to estimate the vertical transport by ensembles of nonprecipitating cumuli in regional chemical transport models (e.g., Vukovich and Ching, 1990).

Deep Convection

Midlatitudes. The first unequivocal observations of deep convective transport of boundary layer pollutants to the upper troposphere were documented by Dickerson et al. (1987). Instrumentation aboard three research aircraft measured CO, 03, NO, NO„ NO,,, and hydrocarbons in the vicinity of an active mesoscale convective system near the Oklahoma/Arkansas border during the 1985 PRE-STORM experiment. Anvil penetrations about 2 h after maturity found greatly enhanced mixing ratios of all of the aforementioned species compared with outside of the cloud. Among the species measured, CO is the best tracer of upward convective transport because it is produced primarily in the boundary layer and has an atmospheric lifetime much longer than the time scale of a thunderstorm. In the observed storm CO measurements exceeded 160 ppbv as high as 11 km, compared with ~ 70 ppbv outside of the cloud (Fig. Ia). Nonmethane hydrocarbons (NMHC) with moderate lifetimes can also trace convective transport from the boundary layer. Ozone can also be an indicator of convective transport; in the polluted troposphere large ozone values will indicate upward transport from the boundary layer, but in the clean atmosphere such values are indicative of downward transport from the uppermost troposphere or lowermost stratosphere. In this case measured ozone in the upper rear portion of the anvil peaked at 98 ppbv, while boundary layer values were only ~ 65 ppbv (Fig. I b). It is likely that some higher ozone stratospheric air mixed into the anvil. Because lightning makes major contributions to reactive nitrogen in thunderstorms, NOx measurements are unsuitable as a convective tracer.

The large amount of vertical trace gas transport noted by Dickerson et al. (1987) cannot, however, be extrapolated to all convective cells. Pickering et al. (1988) reported airborne measurements of trace gases taken in the vicinity of a line of towering cumulus and cumulonimbus clouds that also occurred during PRE-STORM. In this case trace gas mixing ratios in the tops of these clouds were near ambient levels. Meteorological analyses showed that these clouds were located above a cold front that prevented entry of air from the boundary layer directly below or near the clouds. Instead, the air entering these clouds likely originated in the layer immediately above the boundary layer, which was quite clean. Enhanced values of ozone precursor gases were found in the upper troposphere during another PRE-STORM flight conducted in clear air (Pickering et al., 1989). These observations were identified through correlation analysis as indicative of air with a recent boundary layer source and were traced through back trajectory analysis to deep convection that occurred 600 km upstream. Luke et al. (1992) summarized the air chemistry data from all 18 flights during PRE-STORM by categorizing each case according to synoptic flow patterns. Storms in the maritime flow regime transported large amounts of CO, 03, and NOy into the upper troposphere, with the midtroposphere remaining relatively clean. During frontal passages a combination of stratiform and convective clouds mixed pollutants more uniformly into the middle and upper levels; high mixing ratios of CO were found at all altitudes.

Other flights in the vicinity of convective storms over the continental United States were reported by Kleinman and Daum (1991), showing a strong decrease of aerosol particles and water vapor with altitude. However, CO and NOy were more uniformly distributed in the vertical. Plumelike features, attributed to convective outflow, were noted at high altitude in which mixing ratios of boundary layer pollutants increased by 50% or more above background over a distance of several kilometers. Within these features aerosols and water vapor were enhanced over background values, but these soluble substances were always depleted relative to the insoluble species such as CO, suggesting in-cloud removal of the soluble material.

Surface

Troposphere

Surface

Figure I (a) Contour plot of CO mixing ratios (ppbv) observed in and near the June 15. 1985, mesoscale eonvective complex in eastern Oklahoma. Heavy line shows ihe outline of Ihe cumulonimbus cloud. Dark shading indicates high CO and light shading indicates low CO. Dashed contour lines are plotted according to climatology since no direct measurements were made in that area, (b) Same as (a) but tor ozone (ppbv). From Dickerson et al. [1987).

2 OBSERVATIONS 161

Poulida et al. (1996) reported observations taken prior to. in, and around a squall line over North Dakota that evolved into a mesoscale convective complex. In this case the anvil extended well into what used to be the stratosphere Air in the anvil was characterized by low concentrations of O3 (Fig. 2) and high CO, NO, and NO. relative to outside the cloud. This layer of tropospheric air lay above a tongue of stratospheric air. indicating that extensive stratosphere troposphere exchange had occurred. The flux of O3 into the troposphere and the fluxes of water vapor. CO. NO., and hydrocarbons into the stratosphere were estimated for the storm, ff only a small fraction of this material from such anvils remained in the stratosphere, it likely dominates the chemistry of the lower stratosphere in this midlatitudc region.

10.1 Jn 1 I---------nrnm---—1— —i--------—' 1 " '. » 1 ■ 1 1 ' ■ T ■ ■--n'? m p—1

46.50 a6 70 <16.90 47.10 47 30 47.50 47 70 47.90 48 10

LATITUDE CM)

10.1 Jn 1 I---------nrnm---—1— —i--------—' 1 " '. » 1 ■ 1 1 ' ■ T ■ ■--n'? m p—1

46.50 a6 70 <16.90 47.10 47 30 47.50 47 70 47.90 48 10

LATITUDE CM)

Figure 2 Ozone concentrations in tlic anvil of a mesoscale convective complex over North Dakota on June 28, 1989. Heavy line indicates flight track projected onto a vertical plane. Thin lilies are ozone isopleths every lO ppbv. Shading shows location of anvil based on aircraft icc particle measurements. Heavier shading indicates greater particle concentrations. From PouIida et al. (1996).

Tropics. Several deep convection experiments with chemical measurements have been conducted in the tropics. Thompson et al. (1997) have summarized many of these results concerning convective transport of trace gases and their consequences for tropospheric ozone production. Garstang et al. (1988) reported measurements taken in front and behind a dry-season squall line over the Amazon rainforest during the NASA ABLE 2A (Amazon Boundary Layer Experiment) project in 1985. The importance of specific processes within the storm (updrafts and downdrafts) as well as the net result of convective transport (atmospheric overturning) were noted. Since the measurements were confined to the lowest 5 km, downward transport of chemical tracers (e.g., ozone) was the most evident feature (Fig. 3). A major emphasis was placed on sampling convective systems during the ABLE 2B wet-season experiment in the same region in 1987. Scala et al. (1990) reported on a locally occurring ABLE 2B convective system, showing that trace gases in the lower troposphere in the wake of the system were well mixed in the vertical. NO measurements behind the storm were greater than ahead of the system, indicating downward transport from above. However, the NO mixing ratios were low enough that ozone production/destruction rates were very small, allowing ozone to be considered a valid tracer of convective

10 20 30 40 50 60 70

O3 ppbv

Figure 3 Vertical profiles of ozone concentration along the east and west sides of the August 3, 1985, squall line observed in Brazil in ABLE 2 A. The mean profile of ozone from flights in undisturbed weather is shown with a dashed line. Means and standard deviations of ozone from UV-DIAL measurements are shown with symbols and horizontal lines. From Garstang et al. (1988).

10 20 30 40 50 60 70

O3 ppbv

Figure 3 Vertical profiles of ozone concentration along the east and west sides of the August 3, 1985, squall line observed in Brazil in ABLE 2 A. The mean profile of ozone from flights in undisturbed weather is shown with a dashed line. Means and standard deviations of ozone from UV-DIAL measurements are shown with symbols and horizontal lines. From Garstang et al. (1988).

Over remote marine areas the effects of deep convection on trace gas distributions differ from that over moderately polluted continental regions. Chemical measurements taken by the NASA ER-2 aircraft during the Stratosphere-Troposphere Exchange Project (STEP) off the northern coast of Australia show the influence of very deep convective events. Between 14.5 and 16.5 km on the February 2-3, 1987, flight, perturbations in the chemical profiles were noted that included pronounced maxima in CO, water vapor, CCN and minima of NOy and ozone (Pickering et al., 1993). Trajectory analysis showed that these air parcels likely were transported from convective cells 800 to 900 km upstream. Very low boundary layer mixing ratios of NOv and ozone in this remote region were apparently transported upward in the convection. A similar result was noted in CEPEX (Central Equatorial Pacific Experiment; Kley et al., 1996) where a series of ozonesonde ascents showed very low upper tropospheric ozone following deep convection.

Data from convective outflow in the NASA PEM-West A and B experiments (Pacific Exploratory Mission) have been reported by Newell et al. (1996) and by Kawakami et al. (1997). Newell et al. (1996) described sampling of a typhoon in the western Pacific. Boundary layer inflow contained low values of 03, CO, and hydrocarbons, but high values of dimethylsulflde (DMS). There was no evidence of downward entrainment of stratospheric air into the eye region based on ozone measurements. The DMS data suggested substantial entrainment of boundary layer air into the system, particularly in the eyewall region. Kawakami et al. (1997) reported very low NO, mixing ratios in the upper troposphere during the February PEM-West B flights between 1°N and 14°N. These measurements were accompanied by very low ozone and large mixing ratios of water vapor and CH3I, suggesting that the low NOF values were likely due to convetive transport of tropical marine boundary layer air. Other upper tropospheric measurements showed enhanced NO and high NOv/NO,. ratios accompanied by low CO, indicative of NO production by lightning.

Danielsen (1993) presented evidence from Darwin, Australia, ER-2 flights in STEP that rapid vertical irreversible transport of lower tropospheric air into the lower tropical statosphere occurs in convective cloud turrets and by large-scale upwelling in tropical cyclones. Suhre et al. (1997) reported 03 measurements from the tropical Atlantic upper troposphere (10-12 km) taken from commercial aircraft showing mixing ratios of 100 to 500ppbv at a horizontal scale of 5 to 80 km in the proximity of deep convection. It is hypothesized that there is either direct input of stratospheric 03 into the anvils of these systems or there is downward convective transport of 03-rich air that has been transported quasi-isentropically from the extratropical stratosphere.

0 0

Post a comment