Abstract. A database of 15,617 point measurements of dimethylsulfide (DMS) in surface waters along with lesser amounts of data for aqueous and particulate dirhethylsulfoniopropionate concentration, chlorophyll concentration, sea surface salinity and temperature, and wind speed has been assembled. The database was processed to create a series of climatological annual and monthly 1øxl ø latitude-longitude squares of data. The results were compared to published fields of geophysical and biological parameters. No significant correlation was found between DMS and these parameters, and no simple algorithm could be found to create monthly fields of sea surface DMS concentration based on these parameters. Instead, an annual map of sea surface DMS was produced using an algorithm similar to that employed by Conkright et al. [1994]. In this approach, a first-guess field of DMS sea surface concentration measurements is created and then a correction to this field is generated based on actual measurements. Monthly sea surface grids of DMS were obtained using a similar scheme, but the sparsity of DMS measurements made the method difficult to implement. A scheme was used which projected actual data into months of the year where no data were otherwise present.

We have determined the atmospheric concentrations of formic and acetic acid in the gas phase, in aerosols, and in rain during the dry season (Julyr-August 1985) in the Amazonia region of Brazil. At ground level the average concentrations of gas phase formic and acetic acid were 1.6 + 0.6 and 2.2 + 1.0 ppb, respectively. The diurnal behavior of both acids at ground level and their vertical distribution in the forest canopy point to the existence of vegetative sources as well as to production by chemical reactions in the atmosphere. Dry deposition of the gaseous acids appears to be a major sink. The concentrations of formic and acetic acid in the gas phase were about 2 orders of magnitude higher than concentrations of the corresponding species in the atmospheric aerosol. About 50-60% of the aerosol (total) formate and acetate were in the size fraction below 1.0 #m diameter. The highest levels of aerosol formate and acetate were found in haze layers derived from biomass burning. In precipitation, (total) formate and acetate represented about one half of the anion equivalents. This is in contrast to the atmospheric aerosol, where they contributed less than 10% of the soluble anionic equivalents. Furthermore, the precipitation contained considerable acidity (average 36 #eq L -• during the study period), again in contrast to the aerosol, which was acid-base neutral. The mean hydrogen ion concentration in rain was about 21-26 #eq L-• (pH 4.6-4.7). Most of the precipitation acidity can be attributed to the organic acids, with sulfuric and nitric acids contributing only about 10-20% of the hydrogen ion concentration. Aerosol scavenging can explain only a small fraction of the observed amounts of formate and acetate in rain. The observed levels of these ions in rain are most likely the result of a combination of chemical reactions in hydrometeors and scavenging of the gaseous acids by cloud droplets. Goldman et al. [1984] suggested preliminary values of about 0.4-0.6 ppb for gaseous formic acid at about 8-10 km altitude over the southwestern United States. At ground level in the southwestern United States, Dawson et al. [1980] and Farmer and Dawson [1982] measured gaseous formic and acetic acid concentrations that were typically in the range of 0.5-3 ppb. In the boundary layer over the tropical Atlantic and the tropical forest of Guyana, Talbot et al. [1986] and Gre•tory et al. [1986] measured particulate formate and acetate concentrations in the range of 0.5-40 ppt. (Unless otherwise specified, "formate" refers to "total formate," i.e., the sum of dissociated and undissociated forms in aerosols and hydrometeors.) The production of formic acid from the oxidation of formaldehyde in cloud water has been suggested as one of its major sources in the remote marine atmosphere [Chameides and Davis, 1983; Jacob, 1986]. However, the concentrations pre-Paper number 7D0289. 0148-0227/88/007D-0289 $05.00 dicted by these models are not high enough to explain the formate concentrations in rain over continental areas. Jacob and Wofsy [1...

Abstract. Aerosols in the size class <2.5 µm (6 daytime and 9 nighttime samples) were collected at a pasture site in Rondônia, Brazil, during the intensive biomass burning period of 16-26 September 2002 as part of the Large-Scale Biosphere-Atmosphere Experiment in Amazonia -Smoke, Aerosols, Clouds, Rainfall and Climate (LBA-SMOCC). Homologous series of dicarboxylic acids (C 2 -C 11 ) and related compounds (ketocarboxylic acids and α-dicarbonyls) were identified using gas chromatography (GC) and GC/mass spectrometry (GC/MS). Among the species detected, oxalic acid was found to be the most abundant, followed by succinic, malonic and glyoxylic acids. Average concentrations of total dicarboxylic acids, ketocarboxylic acids and α-dicarbonyls in the aerosol samples were 2180, 167 and 56 ng m −3 , respectively. These are 2-8, 3-11 and 2-16 times higher, respectively, than those reported in urban aerosols, such as in 14 Chinese megacities. Higher ratios of dicarboxylic acids and related compounds to biomass burning tracers (levoglucosan and K + ) were found in the daytime than in the nighttime, suggesting the importance of photochemical production. On the other hand, higher ratios of oxalic acid to other dicarboxylic acids and related compounds normalized to biomass burning tracers (levoglucosan and K + ) in the daytime provide evidence for the possible degradation of dicarboxylic acids (≥C 3 ) in this smoke-polluted environment. Assuming that these and related compounds are photo-chemically oxidized to oxalic acid in the daytime, and given their linear relationship, they could account for, on avCorrespondence to: K. Kawamura (kawamura@lowtem.hokudai.ac.jp) erage, 77% of the formation of oxalic acid. The remaining portion of oxalic acid may have been directly emitted from biomass burning as suggested by a good correlation with the biomass burning tracers (K + , CO and EC a ) and organic carbon (OC). However, photochemical production from other precursors could not be excluded.

Abstract. We investigated smoke emissions from fires in savanna, forest, and agricultural ecosystems by airborne sampling of plumes close to prescribed burns and incidental fires in southern Africa. Aerosol samples were collected on glass fiber filters and on stacked filter units, consisting of a Nuclepore prefilter for particles larger than -• 1-2 gm and a Teflon second filter stage for the submicron fraction. The samples were analyzed for soluble ionic components, organic carbon, and black carbon. Onboard the research aircraft, particle number and volume distributions as a function of size were determined with a laser-optical particle counter and the black carbon content of the aerosol with an aethalometer. We determined the emission ratios (relative to CO2 and CO) and emission factors (relative to the amount of biomass burnt) for the various aerosol constituents. The smoke aerosols were rich in organic and black carbon, the latter representing 10-30% of the aerosol mass. K + and NH• were the dominant cationic species in the smoke of most fires, while C1-and so•-were the most important anions. The aerosols were unusually rich in CI-, probably due to the high C1 content of the semiarid vegetation. Comparison of the element budget of the fuel before and after the fires shows that the fraction of the elements released during combustion is highly variable between elements. In the case of the halogen elements, almost the entire amount released during the fire is present in the aerosol phase, while in the case of C, N, and S, only a small proportion ends up as particulate matter. This suggests that the latter elements are present predominantly as gaseous species in the fresh fire plumes studied here.

Dimethylsulfide (DMS), sulfur dioxide (SO2), methanesulfonate (MSA), nonsea-salt sulfate (nss-SO~4-), sodium (Na+), ammonium (NH~), and nitrate (NO~) were determined in samples collected by aircraft over the open ocean in postfrontal maritime air masses off the northwest coast of the United States (3-12 May 1985). Measurements of radon daughter concentrations and isentropic trajectory calculations suggested that these air masses had been over the Pacific for 4-8 days since leaving the Asian continent. The DMS and MSA profiles showed very similar structures, with typical concentrations of 0.3-1.2 and 0.25-0.31 nmol m -3 (STP) respectively in the mixed layer, decreasing to 0.01-0.12 and 0.03-0.13 nmol m -3 (STP) at 3.6 km. These low atmospheric DMS concentrations are consistent with low levels of DMS measured in the surface waters of the northeastern Pacific during the study period.The atmospheric SO2 concentrations always increased with altitude from <0.16-0.25 to 0.44-1.31 nmol m -3 (STP). The nonsea-salt sulfate (nss-SO~4-) concentrations decreased with altitude in the boundary layer and increased again in the free troposphere. These data suggest that, at least under the conditions prevailing during our flights, the production of SO2 and nss-SO~4-from DMS oxidation was significant only within the boundary layer and that transport from Asia dominated the sulfur cycle in the free troposphere. The existence of a 'sea-salt inversion layer' was reflected in the profiles of those aerosol components, e.g., Na + and NO~, which were predominantly present as coarse particles. Our results show that long-range transport at mid-tropospheric levels plays an important role in determining the chemical composition of the atmosphere even in apparently 'remote' northern hemispheric regions.

[1] We investigated aerosol optical properties, mass concentration, and chemical composition over a 2 year period at a remote site in the Negev desert, Israel (Sde Boker, 30°51 0 N, 34°47 0 E, 470 m above sea level). Light-scattering measurements were made at three wavelengths (450, 550, and 700 nm), using an integrating nephelometer, and included the separate determination of the backscatter fraction. Aerosol coarse and fine fractions were collected with stacked filter units; mass concentrations were determined by weighing, and the chemical composition by proton-induced X-ray emission and instrumental neutron activation analysis. The total scattering coefficient at 550 nm showed a median of 66.7 Mm À1 (mean value 75.2 Mm À1 , standard deviation 41.7 Mm À1 ) typical of moderately polluted continental air masses. Values of 1000 Mm À1 and higher were encountered during severe dust storm events. During the study period, 31 such dust events were detected. In addition to high scattering levels, they were characterized by a sharp drop in the Å ngström coefficient (i.e., the spectral dispersion of the light scattering) to values near zero. Mass-scattering efficiencies were obtained by a multivariate regression of the scattering coefficients on dust, sulfate, and residual components. An analysis of the contributions of these components to the total scattering observed showed that anthropogenic aerosol accounted for about 70% of scattering. The rest was dominated by the effect of the large dust events mentioned above and of small dust episodes typically occurring during midafternoon.

27Aerosol particles (PM 2.5 ) were collected during the day (n=6) and nighttime (n=9) 28 from a tropical pasture site in Rondônia, Brazil during an intensive biomass burning 29 period (16-26 September, 2002 the aerosol phase at lower temperatures and higher humidities, as well as possibly 33 production through hydrolysis of N 2 O 5 on aqueous aerosol particles. About 4.2-7.5% of 34 OC (5-13% of water-soluble organic carbon (WSOC)) could be characterized at the 35 molecular level using GC-MS and GC-FID. Among the detected organic compound 36 classes, the relative abundances of anhydrosugars and aromatics were higher in night 37 samples, but sugars/sugar alcohols, diacids, oxoacids and α-dicarbonyls were more 38 abundant in day samples. Consecutive day and night samples showed that δ

Abstract.Measurements of aerosol optical properties (aerosol optical depth, scattering and backscattering coefficients) have been conducted at two groundbased sites in Northern Greece, Ouranoupolis (40 • 23 N, 23 • 57 E, 170 m a.s.l.) and Thessaloniki (40 • 38 N, 22 • 57 E, 80 m a.s.l.), between 1999 and 2002. The frequency distributions of the observed parameters have revealed the presence of individual modes of high and low values, indicating the influence from different sources. At both sites, the mean aerosol optical depth at 500 nm was 0.23. Values increase considerably during summer when they remain persistently between 0.3 and 0.5, going up to 0.7-0.8 during specific cases. The mean value of 65±40 Mm −1 of the particle scattering coefficient at 550 nm reflects the impact of continental pollution in the regional boundary layer. Trajectory analysis has shown that higher values of aerosol optical depth and the scattering coefficient are found in the east sector (former Soviet Union countries, eastern Balkan countries), whereas cleaner conditions are found for the NW direction. The influence of Sahara dust events is clearly reflected in the Angström exponents. About 45-60% of the observed diurnal variation of the optical properties was attributed to the growth of aerosols with humidity, while the rest of the variability is in phase with the evolution of the sea-breeze cell. The contribution of local pollution is estimated to contribute 35±10% to the average aerosol optical depth at the Thessaloniki site during summer. Finally, the aerosol scale height (aerosol optical depth divided by scattering coefficient) was found to be related to the height of the boundary layer with values between 0.5-1 km during winter and up to 2.5-3 km during summer.