Convection has two effects on trace gases. First, soluble species like nitric, sulfuric acid and hydrogen peroxide are rained out in precipitating clouds. (Droplets with high concentrations of nitric and sulfuric are the origins of acid rain.) Second, sparingly soluble gases like sulfur dioxide (SO2), carbon monoxide (CO), nitrogen oxides (NOx=NO + NO2) and hydrocarbons (methane, alkanes, alkenes, aromatics among others) are transported vertically with little loss.
Insoluble gases like methane and CO with relatively long lifetimes can travel far from the location of deep convection. The MAPS instrument on the Space Shuttle sees CO from savanna burning over the Atlantic far from burning regions of South America and Africa. The goal of our research effort is to use measurements and models to evaluate the impact of convection on ozone formation in the middle and upper troposphere (approx. 5-15 km above the surface). Although ozone formation near the surface (referred to as "smog" in the "boundary layer") is an environmental threat to agriculture and human health, ozone in the middle and upper troposphere acts as greenhouse gas:
[REF 3]
to follow the evolution of a particular cloud system that has been
observed in the field. Arrows show predominant flows from a cloud observed
in Amazonia, Brazil
in April 1987:
The effect on CO and/or NOx (shown below) is that a normal gradation (higher CO near surface and highest NOx near surface and in upper troposphere is altered as a cloud evolves. We refer to this as "cloud redistribution" of the trace gas:
Chemical models show that cloud redistribution is most effective in promoting ozone formation when distinct gradients form between mid-upper troposphere and the boundary layer and between the area of cloud and undisturbed air nearby [REF 4].
The trace gas with greatest impact on ozone formation is NO,
which forms in the boundary layer primarily from combustion (auto
exhaust, biomass or wood burning are primary causes
We use the term "convective enhancement factor" to quantify the effect of deep convection on the free troposphere (~5-15 km); it is the ratio of
Net Ozone Production (24-hr period post-convection)
CCF = ---------------------------------------------------
Net Ozone Production (24-hr period, no convection)
In other words, CEF >1 (<1) signifies that ozone formation is increased (decreased) by convection. From 1989 through 1993 we used the GCE and chemical models to analyze aircraft measurements in 3 episodes from the tropical ABLE 2 experiments, one from STEP/EMEX and one from the central US (PRESTORM-1985). A summary of the CEFs is as follows:
The highest values correspond to cases in which polluted air (NOx concentrations > 1 ppbv) was pumped from the boundary layer upward into a very clean troposphere (NOx< 0.2 ppbv). This occurred in some of the sampling of 26 April 1992 in Amazonia, when the NASA Electra encountered an urban plume from the city of Manaus, as well as during the 10-11 June 1985, PRESTORM case, which encountered midwest US pollution. The cases with CEF > 50 are hypothetical model sensitivity runs. The dynamical regime assumed in the calculations is based on 3 August 1985 ABLE 2A sampling but the chemical concentrations assumed corresponded to fresh biomass burning emissions. The latter would be normally be observed in late August or September, according to satellite statistics (from Setzer at INPE) that show peak burning usually later than the ABLE 2A campaign. Note two cases with CEF = unity. These occurred because overturning and vertical mixing were so effective that net redistribution after the event (6 hours) did not occur.
Major factors affecting the CEF are -
In addition to specific convective episodes, we have examined the impact of deep convection on regional carbon monoxide budgets. In one case we asked: "what fraction of carbon monoxide from biomass burning gets into the middle and upper troposphere by the process of deep convective pumping?"
To do this, we scaled up from very limited data to cover the biomass burning season in a specific region - the state of Rondonia, Brazil where aircraft sampling was conducted in ABLE 2.
For example, we used emissions factors (received from Andreae's group at Max-Planc) from selected fires studied in this area. We had calculated convective fluxes (rate of mass transfer from boundary layer to upper troposphere) for a prototype system (the 3 August 1985 event sampled on ABLE 2A; [REF 6]. This was scaled to a mass flux over area and time, using satellite classification of cloud type from ISCCP data. Likewise, fire counts from satellite (from Setzer at INPE) and field measurements of CO concentrations are used. The result is that 10-40% of CO released from one dry season's biomass burning in this part of Brazil would finds its way into the middle and upper troposphere as a result of deep convection.
We also have analyzed the convective role in a more urban- influenced CO budget [REF 7]. The point is to compare deep convective transport out of the boundary layer to other processes affecting CO - input from adjacent regions by advection (horizontal flow), surface sources including pollution combustion, plants or soils as appropriate, surface deposition (if applicable) and photochemical sources and sinks. Carbon monoxide, for example, is not only a pollutant but oxidation of nonmethane hydrocarbons, both anthropogenic and natural, adds to carbon monoxide. We used a climatology of midwest US aircraft CO measurements from the Univ. of Maryland Air Quality group.
[REF 8]. The benefit for the midwest is
that convection acts like a ventilator, letting the carbon monoxide
out of the boundary layer to travel onward. Thus, hazardous
concentrations do not build up as greatly in this populous area
as they would in the absence of convection.
Go to Part 2
Movie of Modeled Convective Activity on Air Chemistry
See Bibliography of Papers on Chemistry
and Convection from GSFC/U Md Collaboration (1990-present)
Chatfield, R. B. and A. C. Delany, Convection links biomass burning to increased tropical ozone: However, models will tend to overpredict O3, J. Geophys. Res., 95, 18,473-18,488, 1990.
Talbot, R., et al, Chemical characteristics of continental outflow over the tropical south Atlantic Ocean from Brazil and southern Africa, J. Geophys. Res., in press, 1996.