Atmospherics

The atmospheric transport of biomass burning emissions is studied through a numerical simulation of the air mass motions using the CATT-BRAMS (Coupled Aerosol and Tracer Transport model to the Brazilian developments on the Regional Atmospheric Modeling System). CATT-BRAMS is an on-line transport model fully consistent with the simulated atmospheric dynamics. The sources emission from biomass burning and technological activities for several gases and aerosol may be defined from several published dataset and remote sensing. The mass concentration prognoses accounts also for convective transport by shallow and deep cumulus, wet and dry deposition and plume rise. Also, an additional radiation parameterization, which takes the interaction between aerosol particles and short and long wave radiation into account, was implemented. The model is applied for simulate carbon monoxide (CO) and particulate material PM2.5 transport during the SMOCC/RACCI campaign during the 2002 dry season. The model validation is presented with comparison of model results with remote sensing products and direct observations obtained during the SMOCC campaign.

We investigated the effects of fire-induced plume-rise on the predicted export of carbon monoxide (CO) over Africa during SAFARI 2000 using the NCAR Community Atmosphere Model (CAM) with a CO tracer and plume-rise parameterization scheme. The plume-rise parameterization scheme simulates the consequences of strong buoyancy of hot gases emitted from biomass burning, including both dry and cloud-associated (pyrocumulus) lofting. The scheme was first adapted from a regional model. The current implementation of the plume-rise parameterization scheme into the global model provides an opportunity to examine the effect of plume-rise on long-range transport. The CAM simulation with the plume-rise parameterization scheme shows a substantial improvement of the agreements between the modeled and aircraft-measured vertical distribution of CO over southern Africa biomass burning area. The plume-rise mechanism plays a crucial role in lofting biomass burning pollutants to the middle troposphere. In the presence of deep convection we found that the plume-rise mechanism results in a decrease of CO concentration in the upper troposphere. The plume rise depletes the boundary layer, and thus leaves lower concentrations of CO to be lofted by the deep convection process. The effect of the plume-rise on free troposphere CO concentration is more important for the source area (short-distance transport) than for remote areas (long-distance transport). The plume-rise scheme also increases the CO export fluxes from Africa to the Atlantic and Indian Oceans. These results further confirm and extend previous findings in a regional model study. Effective lofting of large concentration of CO by the plume-rise mechanism also has implication for local air quality forecast in areas affected by other fire-related pollutants. 1 Introduction Carbon Monoxide (CO) is one of the principal pollutants in the atmosphere and has an important impact on the chemical production of tropospheric ozone. CO has been used

Several strong statistical relationships quantifying local ozone generation are found which use only easily measured variables: nitrogen oxides (NO x), formaldehyde (HCHO), its photolysis (i.e., UV), and temperature (T). A parameterized regression developed for rural air was adapted to central Queens, New York City, i.e., considerable fresh emissions. Measurements of the radicals [HO 2 ] and [OH] were available. These provided explicit reference estimates of the predominant terms for chemical ozone production, P o (O 3) ¼ k [HO 2 ][NO], of the predominant chemical loss of nitrogen oxides, L(NO 2) ¼ k[OH][NO 2 ], and also their ratio. (This is termed a production efficiency for O 3 .) Chemical modeling supports a robust extension from P o (O 3) to total chemical production, P(O 3). The two regression variables, [NO] and j HCHO0rads  [HCHO], which best explain P o (O 3), have low correlation, R w 0.2 (variable, interacting urban plumelets?). In our analysis, R 2 for P o (O 3) (and an estimate for its rate-determining [HO 2 ]) was in the range 0.48e0.81. Signally, the method suggests a quantitative and very local application of descriptions of "VOC limitation" or "NO x limitation" to P(O 3) and L(NO 2), expressed as dimensionless sensitivity variables. Unexpected sources, transport, or chemistry may be highlighted using only HCHO, NO x , and UV radiation. More complex relationships are needed in a focused analysis of intermediate polluted situations, where timescales or individual sources may give trouble. Here, we find that T is informative, and cooperates with j  [HCHO] in defining [HO 2 ]. Sensitivities for radicals and NO for P o (O 3) are similar w0.4, but sensitivities for radicals and NO 2 for L(NO 2) emphasize NO 2. Remaining variability in the statistical estimates of P o (O 3) and L(NO 2) is modulated by incompletely understood, slowly varying gain factors. Understanding of these gain factors promises a better empirical indicator for P o (O 3)/L(NO 2). Complete 3-d simulations are not replaced, but this view helps separate sub-problems in the estimation of HO 2 and P(O 3).

Satellite observations may improve the areal coverage of particulate matter (PM) air quality data that nowadays is based on surface measurements. Three statistical methods for retrieving daily PM 2.5 concentrations from satellite products (MODIS-AOD, OMI-AAI) over the San Joaquin Valley (CA) are compared e Linear Regression (LR), Generalized Additive Models (GAM), and Multivariate Adaptive Regression Splines (MARS). Simple LRs show poor correlations in the western USA (R 2 y 0.2). Both GAM and MARS were found to perform better than the simple LRs, with a slight advantage to the MARS over the GAM (R 2 ¼ 0.71 and R 2 ¼ 0.61, respectively). Since MARS is also characterized by a better computational efficiency than GAM, it can be used for improving PM 2.5 retrievals from satellite aerosol products. Reliable PM 2.5 retrievals can fill in missing surface measurements in areas with sparse ground monitoring coverage and be used for evaluating air quality models and as exposure metrics in epidemiological studies.

The abundance of gas phase nitric acid in the upper troposphere is overestimated by global chemistry-transport models, especially during the spring and summer seasons. Recent aircraft data obtained over the central US show that mineral aerosols were abundant in the upper troposphere during spring. Chemical reactions on mineral dust may provide an important sink for nitric acid. In regions where the mineral dust abundance is low in the upper troposphere similar HNO3 removal processes may occur on biomass burning aerosols. We propose that mineral and biomass burning aerosols may provide an important global sink for gas phase nitric acid, particularly during spring and summer when aerosol composition in the upper troposphere may be greatly affected by dust storms from east Asia or tropical biomass burning plumes.

Significant concentrations of formic and acetic acids are known to exist in the troposphere. These acids contribute between 16 and 35% of the free acidity in North American precipitation and between 25 and 98% of the free acidity of precipitation in remote areas. Their origin is still relatively uncertain. For formic acid, several gas phase and aqueous phase chemical production mechanisms have been postulated, and seem to explain at least roughly the observed concentrations. However, for acetic acid, the analogous production mechanisms appear to be far too slow, and it has been suggested that this acid may be emitted directly by vegetative matter. Based on recent laboratory and theoretical studies, we have considered the possible gas phase production of acetic acid via the reaction of peroxy acetyl radicals with HOg and organic peroxy radicals. A highly detailed gas phase chemical mechanism coupled with a simple 1-D model was used to predict organic acid concentrations for one set of conditions (Amazon boundary layer) where extensive field measurements are available for model initialization and comparison. The results indicate that the gas phase photooxidation of non-methane-hydrocarbons can produce amounts of acetic acid in the range of several pans per billion, comparable to the observed values.

The N oxidation ratio, [HNO3]/[NOx], in the free troposphere is measured at •-1 to 9 (averaging-5), consistently lower than the values of !5-100 which known theory suggests to be reached rapidly. This discrepancy casts doubt on current models' abilities to predict effects of NO sources on remote [NO], and so also [03] and [OH]. HNO3-to-NO x recycling processes involving HCHO in solution are appealing explanations for the ratio. Known chemical theory are illustrated using a Lagrangian box model of the mid-troposphere which simulates characteristic episodes from NOx input to HNO3 washout. Ratio and budget constraint equations for NOy can usefully restrict conceivable explanations of missing chemistry of fast nitrogen cycling; most explanations could be called "fast-cycles" or "re-NOx-ification." These equations also show how current global models may suggest spuriously good N oxidation ratios and errors in OH. Aerosol or cloud droplets reactions with HCHO may produce (a) formic acid (which needs an appropriate source) and NO x, or alternatively, (b) hydroxymethyl nitrate (or methyl dinitrate), which might supply a large, variable amount of "missing NOy" not currently measurable except by NOy instruments. Current Theory This note concerns the ratio [HNO3]/[NOx], as observed and as correctly and incorrectly simulated. (NOx = NO + NO2, compounds which interchange rapidly, given daylight.) Figure 1 (discussed below) depicts the activity of tropospheric nitrogen. NO is the N compound whose activity is most important, as it determines tropospheric oxidizing capacity, directly, in its effect on instantaneous local OH, and also cumulatively, in its control on ozone [Chameides and Walker, 1973, Liu et ai., 1987]. Hence it has a key role with respect to future greenhouse gas and reactive chlorocarbon concentrations.

The three-dimensional (3-D) distribution of carbon monoxide (CO) over the southern Pacific during the NASA Global Tropospheric Experiment Pacific Exploratory Mission-Tropics (PEM-T) (August-October 1996) has been analyzed in comparison to other CO measurements. The following data sets have been used in the study: National Oceanic and Atmospheric Administration Climate Monitoring and Diagnostic Laboratory surface level sampling; Commonwealth Scientific and Industrial Research Organization aircraft measurements over Cape Grim, Tasmania; solar spectroscopic measurements at Lauder, New Zealand; and data from two spaceborne Measurement of Air Pollution From Satellite experiments. For the PEM-T mission back trajectories analysis and 3-D modeling of the CO transport have been performed. It has been demonstrated that CO measurements obtained by different in situ and remote techniques can be used to build the picture of the CO climatology over the large geographical area. The structure of the CO distribution over the western part of the southern Pacific during the austral spring is mainly controlled by emission from biomass burning in Australia and Africa and subsequent long-range transport. The prevailing westerly transport occurs in the middle and upper troposphere, whereas the marine boundary layer remains relatively clean and uniform. Barriers in the form of the Intertropical Convergence Zone and South Pacific Convergence Zone protect the equatorial area (equator to 10øS) from direct impact of biomass burning plumes from north and southwest. Consistency between the measurements taken in different years and modeling results indicates that the observed feature is a stable phenomenon. Outside the equatorial area the CO vertical distribution has a broad distinctive maximum at the altitude range 5-8 km and latitudes between 20øS and 30øS. This maximum is a stable feature, and its location indicates the area where the most intensive westerly transport occurs. 1.

A full-chemistry simulation of the Great African Plume gives one example of a broad conceptual model of the intercontinental pollution of the tropical middle troposphere by lofted biomass burning plumes. This two-dimensional idealization "calibrated" by carbon monoxide distributions links conventional estimates of burning emissions to oceanic concentrations of pollutants. This paper makes use of GRACES, a modular photochemical simulation system, in two forms. The results of the chemically intensive two-dimensional form, using idealized winds, mixing, deposition, and rainout, match the general concentration patterns of a three-dimensional GRACES model study of CO during the TRACE A/SAFARI period of October 1992 (reported separately). The study highlights the importance of simulating the vertical and diurnal variation of the planetary boundary layer and cloud activity. These correlate temporally with the intensity of tropical agricultural burning. We emphasize one situation, the drift northward and eastward of pollution into the interocean convergence region, where it rises by small-scale motions and fides out westward in the lower midtroposphere (<5 km). These effects help set in place large strata of enhanced CO, ozone, and other pollution over the equatorial Atlantic Ocean. Overall, our comparisons of simulations with the TRACE A data on the cycling of CO, NOx, and 03 in the tropical atmosphere suggest substantial agreement of current emission estimates and atmospheric concentrations. In certain regions, ozone is simulated slightly below observed levels. The striking major disagreements are in NOy (total reactive nitrogen) and HNO3, which are intimately related to CO and 03; this suggests that current theory omits at least one fundamental process. 24,279 24,280 CHATFIELD ET AL.: AFRICAN EMISSIONS, OCEANIC PLUMES OF 03, CO, PAN, AND SMOKE the role of localized organized convergence/pattems in the lower troposphere, e.g., an interocean front, to the north and east of the emissions. These aspects of the convective process are repeatedly suggested by our simulations, but evaluation will require further studies. These important effects support the importance of the "mix and cook" situations described by Chatfield and Delany [1990], in which situations of biomass burning followed within hours by deep cumulus convection was found to be particularly effective in elevating the tropospheric ozone column. Two large field experiments, Tropical Atmospheric Chemistry Experiment Atlantic (TRACE A) and Southern African Fire Atmospheric Research Initiative (SAFARI), were mounted in September and October 1992, to sample the pollutant sources and mid-oceanic accumulation [Andreae et al., 1994; Bachmeier and Fuelberg, this issue; Fishman et al. [this issue], Thompson et al., this issue]. The work reported here is intended to summarize one important process at work during that sampling period and to make some general comparisons to the TRACE A data. Outline of the Paper Our transport-chemistry model, GRACES (Global-Regional Atmospheric Chemistry Event Simulator), is used in two ways. In one mode it uses highly realistic three-dimensional winds and cloud transports to describe the motion of an inert tracer, one that behaves like carbon monoxide. A striking and repeated feature of this simulation seems to deserve more detailed chemical simulation. After an introductory description of this simulation we describe some preliminary results which motivated the main chemistry simulation that constitute this paper. Further threedimensional analysis will come in a future paper, but this striking feature of the three-dimensional simulations causes us to seek to formulate a simple but general model of biomass plumes. Thus motivated, we return to model formulation, and propose a two-dimensional Great African Plume model that follows one main pathway of the plume that has very visible effects in Atlantic Ocean. This model also attempts to have just enough description of transport to rationalize the concentrations seen in the atmosphere with the emission fluxes of pollutants and other trace species. Yet, it retains the simplicity to provide a reasonably simple conceptual framework for comparing the roles of important processes. In some important ways, it overcomes some limitations of the trajectory models which have also been applied to explain the TRACE A observations. In a few days the atmosphere significantly acts both to accumulate pollution like CO and to disperse it horizontally and to vent it vertically. Since it is difficult to simulate these processes in trajectory models, their results are more limited in addressing emission/concentration budget questions and dispersion. The trajectory models, of course, being conceptually simpler, are even more rapidly and broadly deployable and give some three-dimensional information consistent with resolvable analyzed wind fields [Thomspon et al., this issue; Fuelberg et al., this issue; Pickering et al., this issue(b)]. We expect that our proposed model has complementary capabilities and limitations. Following the description of the idealization there is a section on methods for emission and chemistry. The two-dimensional results are analyzed, first to describe qualitatively the very different biomass burning plumes that different species trace out and then to make initial comparisons to a small set of the TRACE A data showing general agreement. However, nitric acid in the continental plume and ozone production in the far-downwind ocean plume show disagreements with observations; these require explanations that are beyond our paper. This connects our simulation to the local-photochemical model work also performed for TRACE A [Jacob et al., this issue; Thompson et al., this issue; Pickering eta/., this issue. Motivation Three-Dimensional Simulation of TRACE A Burning Emissions How do ozone and CO accumulate so in the central Atlantic? Weather patrems vary sufficiently in the tropics that we attempt to describe a typical main route, with typical transit times, rather than one constituted on mean winds. Our current purpose is to provide an appropriate setting for a specialized two-dimensional model; that is, a generalized picture of emission, vertical mixing, and transport in a general downwind direction, and one typical pattern will serve. We draw that pattern from a three-dimensional simulation of a carbon monoxide-like tracer using reconstructed meteorological winds and patterns of dry (boundary layer) and moist convection. Methods for three-dimensional simulations MM5 runs. Plates 1 a and lb give a multilevel view of an important October flow pattern centered on southern and central Africa, taken from a run of the MM5 mesoscale-synoptic model [Grell et al., 1994, and references]. This figure portrays the weather-simulation and forecasting model, MM5, run in an assimilation mode, for 1300 UT on October 2, 1992. The model nudges the solutions toward the observed meteorology as recorded in the analyses of the European Centre for Medium-Range Weather Forecasting (ECMWF) [Grell et al., 1994]. Special analyses for the TOGA (Tropical Ocean Global Atmosphere) period were available for 1992 with better-than-normal observational support. A standard Blackadar technique described in the MM5 documentation was used to simulate the planetary boundary layer. Plate l c shows the diagnosed boundary layer height, above the surface, for that time. The Grell scheme [Grell, 1993; Grell et al., 1994, and references] developed for MM5 (version 1) was used to simulate convection. Plates ld and le show the simulated cloud updraft top in kilometers and the cloud base up&aft mass flux for the same time. Clearly, the cloud top diagnostic should be interpreted judiciously outside the regions.of larger mass flux. No separate, special low-cloud parameterization was employed; the difficulty of detailed low-cloud simulation justifies more specialized work. The convective mass flux information as well as a boundary layer height and convective velocity-scale information were saved for off-line use. All the wind, density, and convective information were transferred into our GRACES model, run in a three-dimensional mode. GRACES is our Global-Regional Atmospheric Chemistry Event Simulator, to be more fully described by R. B. Chatfield et al. (manuscript in preparation, 1996) (hereinafter referred to as C). Briefly, it is modularly constructed and runs in a generalized mapprojection coordinate system deriving from the description of Toon et al. [1988]. It employs the exponential-fit system of that paper for vertical transport, but an S-tuned nonoscillatory version of the Smolarkiewicz [1984] horizontal transport scheme. The chemistry is described by a reaction table, and the photochemistry is driven by delta-Eddington integrated intensities that are integrated with the absorptions and quantum yields of Madronich and Calvert [1990]. The chemistry is integrated using the DASSL algorithm employing backward difference formulae [Brenan et CHATFIELD ET AL.

Simultaneous measurements of the oxides of hydrogen and nitrogen made during the NASA Subsonic Assessment, Ozone and Nitrogen Oxide Experiment (SONEX) afforded an opportunity to study the coupling between these two important families throughout the free troposphere and lowermost stratosphere. Moreover, the suite of measurements made during the campaign was unprecedented in its completeness, thus providing a uniquely detailed picture of the radical photochemistry that drives oxidation and ozone production in this part of the atmosphere. On average, observed hydrogen oxides (HOx = OH + HO2) agree well with both instantaneous and diel steady-state models; however, there is a persistent deviation of the observations that correlates with the abundance of nitrogen oxides (NOx = NO + NO2) in the sampled air mass. Specifically, the observed HO• tends to exceed the model predictions in the presence of high NO• concentrations, by as much as a factor of 5 (>500 pptv NOx), and is sometimes as little as half that expected by steady state at lower NOj levels. While many possibilities for these discrepancies are discussed, it is argued that an instrumental artifact is not probable and that the discrepancy may bespeak a shortcoming of our understanding of HO• chemistry. The consistently elevated HO• in the presence of elevated NOx leads directly to greater ozone production than expected, thereby extending the NOx-limited regime of the upper troposphere. These results could thus have bearing on the predicted impacts of increasing NO• emissions into this region of the atmosphere from, for example, the growth of global air traffic.

This paper explores some of the mechanisms governing the accumulation of passive tracers over the tropical southern Atlantic Ocean during the northern hemisphere fall season. There has been a pioneering observation regarding ozone maxima over the South Atlantic during austral spring. The understanding of the formation of this maxima has been the prime motivation for this study. Using a global model as a frame of reference, we have carried out three kinds of experiments during the period of the Transport and Atmospheric Chemistry Near the Equator-Atlantic (TRACE A) project of 1992. The first of these is a simple advection of total ozone (a passive tracer) in time using the Florida State University global spectral model. Integration over the period of roughly 1 week showed that the model quite closely replicates the behavior of the observed total ozone from the total ozone mapping spectrometer (TOMS). This includes many of the changes in the features of total ozone over the tropical and subtropical region of the southern Atlantic Ocean. These studies suggest a correlation of 0.8 between the observed ozone over this region and ozone modeled from "dynamics alone," i.e., without recourse to any photochemistry. The second series of experiments invoke sustained sources of a tracer over the biomass burn region of Africa and Brazil. Furthermore, sustained sources were also introduced in the active frontal "descending air" region of the southern hemisphere and over the Asian monsoon's east-west circulation. These experiments strongly suggest that air motions help to accumulate tracer elements over the z-•uctHtl•, Uk, CctH. Z-• tHltkl 31711173 Ut C)llOClJlllCllt3 •klkllCbb Wllttt fllay bo •HLII5111 required to improve the deficiencies of the vertical stratification of ozone predicted by the model over the flight region of the tropical southern Atlantic during TRACE A. Here we use the global model to optimally derive plausible accumulation of burn elements over the fire count regions of Brazil and Africa to provide passive tracer advections to closely match what was observed from reconnaissance aircraft-based measurements of ozone over the tropical southern Atlantic Ocean.

We present the first directly retrieved global distribution of tropospheric column ozone from Global Ozone Monitoring Experiment (GOME) ultraviolet measurements during December 1996 to November 1997. The retrievals clearly show signals due to convection, biomass burning, stratospheric influence, pollution, and transport. They are capable of capturing the spatiotemporal evolution of tropospheric column ozone in response to regional or short timescale events such as the 1997-1998 El Niño event and a 10-20 DU change within a few days. The global distribution of tropospheric column ozone displays the well-known wave-1 pattern in the tropics, nearly zonal bands of enhanced tropospheric column ozone of 36-48 DU at 20°S-30°S during the austral spring and at 25°N-45°N during the boreal spring and summer, low tropospheric column ozone of <30 DU uniformly distributed south of 35°S during all seasons, and relatively high tropospheric column ozone of >33 DU at some northern high-latitudes during the spring. Simulation from a chemical transport model corroborates most of the above structures, with small biases of <±5 DU and consistent seasonal cycles in most regions, especially in the southern hemisphere. However, significant positive biases of 5-20 DU occur in some northern tropical and subtropical regions such as the Middle East during summer. Comparison of GOME with monthly-averaged Measurement of Ozone and Water Vapor by Airbus in-service Aircraft (MOZAIC) tropospheric column ozone for these regions usually shows good consistency within 1s standard deviations and retrieval uncertainties. Some biases can be accounted for by inadequate sensitivity to lower tropospheric ozone, the different spatiotemporal sampling and the spatiotemporal variations in tropospheric column ozone.

Biomass buming throughout the inhabited portions of the tropics generates precursors which lead to significant local atmospheric ozone pollution. Several simulations show how this smog could be only an easily observed, local manifestation of a much broader increase in tropospheric ozone. We illustrate basic processes with a one-dimensional time-dependent model that is closer to true meteorological motions than commonly used eddy diffusion models. Its application to a representative region of South America gives reasonable simulations of the local pollutants measured there. Three illustrative simulations indicate the importance of dilution, principally due to vertical transport, in increasing the efficiency of ozone production, possibly enough for high ozone to be apparent on a very large, intercontinental scale. In the first, cook-then-mix, simulation the nitrogen oxides and other burning-produced pollutants are confined to a persistently subsident fair weather boundary layer for several days, and the resultant ozone is found to have only a transient influence on the whole column of tropospheric ozone. In the second, mix-then-cook, simulation the effect of typical cumulonimbus convection, which vents an actively polluted boundary layer, is to make a persistent increase in the tropical ozone column. Such a broadly increased ozone column is observed over the the populated "continental" portion of the tropics. A third simulation averages all emission, transport, and deposition parameters, representing one column in a global tropospheric model that does not simulate individual weather events. This "oversmoothing" simulation produces 60% more ozone than observed or otherwise modeled. Qualitatively similar overprediction is suggested for all models which average significantly in time or space, as all need do. Clearly, simulating these 0 3 levels will depend sensitively on knowledge of the timing of emissions and transport. al., 1987; Crutzen, 1988]. Smog ozone in the tropics. Similar increases in ozone production in the tropics must now also be considered. Fishman and colleagues [Fishman et al., 1986; Fishman and Larsen, 1987] have described patterns in the total ozone column above the tropics which appear to reflect primarily patterns in tropospheric ozone, both on the mesoscale and the global scale. It is not yet clear that the largest scale patterns, covering the inhabited, more continental half of the global tropics, are due to pollution, although initial simulations suggest that meteorological and natural influences do not produce these patterns when and where tropospheric ozone is highest. In recent decades there has been an increase in industrialization and motorized vehicle traffic in many tropical developing countries [Oak, 1986]. There have been several

Max-Planck-lnstitut fiir Chemie Reactive surface emissions of reduced sulfur gases can produce SO2 in the middle and upper troposphere at the levels of 80 _+ 30 pptv measured high over the remote oceans. We present simulations with a two-dimensional "Staubsauger" or "vacuum cleaner" model that combines a photochemical model with a description of vertical transport of trace species by convective clouds within larger synoptic circulations. Emissions of 20-60 Tg (S)/yr of (CH3)2S, H2S, 0r CS2, may produce the observed SO 2. Roughly equal production rates of SO 2 and methane sulfonic acid may be expected. The amount and exact vertical distribution of the SO 2 produced remain uncertain: the greatest chemical uncertainties are the reaction yield of SO2 expectable under clean tropospheric conditions and also the liquid-phase removal of SO 2, and the oxidation rate. The amount of upper tropospheric SO 2 produced depends substantially on the proximity of strong reduced S sources to regions of active convection. However, the character of the solutions we present is invariably distinctly different from those obtained with one-or two-dimensional models employing the eddy-diffusion hypothesis. The results of the model point beyond its original conception, and stress the likely importance of the rainy tropical jungles and mid-latitude industrial regions, since both regions have large sulfur emissions and frequently active cumulonimbus convection. This process, however, should contribute mainly to upper-tropospheric SO 2. Other chemical implications are that tropospheric OH may depend critically on HOOH levels as well as the hydrocarbon and nitrogen oxide cycles. Cloud transport may play an important role in these cycles. The hydroxyl radical concentration depends as much on assumptions regarding HOOH reaction and transport as it does on NO levels. 1. INTRODUCTION Although SO2 and its acidic sulfate products are generally thought to arise from fossil fuel burning, recent measurements suggest that concentrations from 30 to 150 pptv are typical even in remote ocean regions. Measurements made on project GAMETAG [Maroulis et al., 1980] showed extremely variable SO: concentrations along oceanic flights between San Diego (30øN) and New Zealand (80øS), with midtropospheric concentrations perhaps 50% higher than boundary-layer ones. Some support for these observations comes from the measurements reported by Meixner [1981]; these were made over the eastern North Atlantic and suggested concent?ations of 35-50 pptv, again increasing somewhat with altitude. Measurements by Nguyen et al. [1974] suggest concentrations varying around 34 pptv at shipboard level. Sulfur dioxide is thought to be a major source of sulfate cloud condensation nuclei, which are also found to have a low background value distributed through the troposphere [Pruppacher and Klett, 1978]. Twomey [1980], among others, has suggested that these CCN constitute a significant control on climate via their strong influence on cloud droplet size spectra, precipitation efficiency of clouds, and the radiative properties of clouds. This is clearly a reason to establish the origin of the sulfate particles. A more fundamental reason to study SO2 is that it appears to be a relatively reactive tropospheric compound whose main natural sources appear to be even more reactive. How can it survive in measurable concentrations throughout the remote troposphere? Combustion sources (like industrial pollution) Copyright 1984 by the American Geophysical Union. Paper number 4D0378. 0148-0227/84/004D-0378505.00 cannot account for these SO 2 concentrations, for the SO 2 loss reaction SO2 + OH-• HOSO2'-• aerosol (1) should cause pollutant SO2 to decrease within 1 or 2 weeks travel time southward from the continents of the northern hemisphere. Hydroxyl-radical reaction and other processes (e.g., deposition onto the earth's surface and reaction and scavenging in clouds) act in the lower troposphere to give atmospheric lifetimes from one half to seven days [Nguyen et al., 1975; Prahm et al., 1976; Logan et al., 1979]. While the removal rate is not sufficiently well-known, it appears too fast to permit substantial transport to the southern tropics. Even though SO2 concentrations start at millions of parts per trillion, little should remain after months of interhemispheric travel and dilution [Rodhe and lsaksen, 1980]. 1.1. Are Unknown Sulfur Compounds Required? In recent years, there has been a search for a sulfur source compound which can release sulfur slowly into the troposphere, and in sufficient quantity to produce the observed SO2. It has been suggested that the source of this SO2 is reduced sulfur compounds emitted due to biochemical processes within bacteria living near the surface of the ocean [Granat et al., 1976; Lovelock et al., 1972; Lovelock, 1974]. Some of the candidate compounds and some approximate lifetimes are carbo•yl sulfide (OCS' months, years), carbon disulfide (CS 2' 2•-6 ::•ys), hydrogen sulfide (H2S' 2 days), dimethyl sulfide ((CF13)2S: 8 hours to 2 days), methyl mercaptan or methane thiol (CH3SH: 2 hours), and dimethyl disulfide (CH3SSCH3:15 min)(references in this work and Chatfield [1982]). Rodhe and lsaksen [1980] considered these reduced 7111

Portions of the global cycles of sulfur and iodine could be intertwined in reactions resulting from the emissions of dimethyl sulfide (DMS) and methyl iodide. A two-dimensional photochemical model of a marine tropical tropospheric synoptic system is used to analyze the consequences of these emissions through their atmospheric transformation cycles to aerosols. Following one line of analysis, the reaction of IO+DMS-• DMSO+I could play a significant role,-30-50%, in destroying DMS, if the reaction proceeds at the published fast rate. Under these assumptions, the concentrations of DMSO (dimethyl sulfoxide), dimethyl sulfone (DMSO2) , and sulfate appear to set rather strict upper limits on the extent of the reaction over the global oceans, and hence, on [IO], which is < 2x10 5 molecule cm-3. DMS plays an important role in the iodine cycle by converting IO into I. The formation of a reservoir species HOI, following H abstraction by IO, may help explain the high concentrations and apparent strong diurnal variation of filter-sampled inorganic iodine gases and help to explain low [IO] values. In an alternate. preferred, line of analysis, the rate of reaction of IO+DMS may be much slower, < 10-12 cm 3 molecule-Is-I and the I and S cycles remain decoupled: in this case, it appears that the OH+DMS addition reaction could proceed-30% to form DMSO. The mean concentration and variations of gaseous and particulate iodine species are also better simulated. A simple aerosol model describing fine and coarse particles is used to close the S budget, and, if only a minor role is played by DMSO, the budget compares well with data. Simulated SO 2 concentrations for the upper troposphere are highly dependent on upper tropospheric [OH]. The paper concentrates on remote tropical ocean regions and includes simulations of the spatial and diurnal variation of iodine species I, IO, HI, HOI, IONO2, I20 2, and Itot(particulate); the sulfur species DMS, DMSO, DMSO2, MSA (methane sulfonic acid), SO 2, and SO 4=(particulate), as well as the species OH, HO 2, and H 202 .

We synthesize daily sonde (vertical) information and daily satellite (horizontal) information to provide an empirical description of ozone origins over the northern Indian Ocean during the INDOEX field campaign (February-March 1999). This area is shown to be a significant portion of the "high-ozone tropics." East-west O 3 features and their flow are identified, and ozone origins are compared to other tropical regions, using water vapor as a second tracer. In the study period, multiple processes contribute to O 3 column enhancements, with importance varying strongly by latitude: (a) Low-altitude O 3 pollution over the northern Indian Ocean mainly originates from the Indian subcontinent, and is traceable to high emission areas; convective activity south of Sri Lanka helps direct ozone outflow from the Northern Indian Subcontinent; (b) Middle tropospheric O 3 maxima over the Northern Indian Ocean originate from various sources, often transitioning within a few hours. Convective venting of Asian pollutants can add 20-50 ppbv to the middle troposphere at 5-10 N, alternating with stratospheric influence. (c) A number of cases suggest that strong mixing-in of stratospheric air along the subtropical jet raised tropospheric O 3 in early March by~40-50 ppbv, especially poleward of~10°N. (d) Influences of lightning and large-scale biomass burning were not strong during this period, in contrast to the situation in Africa and the South Atlantic or locally in Southeast Asia. This work illustrates successes and limitations in approaches to synthesizing disparate information on trace-gas distributions taken from satellite retrieval products and ozonesondes.

This work studies the atmospheric CO2 budget in the Amazon basin, focusing on the role of shallow and deep convective systems. The vertical redistribution of CO2 is numerically simulated using an Eulerian transport model coupled to the Brazilian developments on the Regional Atmospheric Modeling System (BRAMS). The transport model includes grid-scale advection, diffusion in the PBL (Planetary Boundary Layer) and convective transport by sub-grid shallow and deep moist convection. In the simulation, the mass conservation equation is solved for six tracers, including or not the shallow and deep moist convection terms. The rectifier effect is also showed through simulation of the transport to the free troposphere of PBL air masses with low CO2 concentrations due to assimilation by vegetation during the afternoon, when both CO2 fixation and convection are at their peak. The model is applied to simulate July 2001 with a 30 km grid resolution covering the northwest part of South America. We...