Ronald L. Siefert

Global distribution of atmospheric phosphorus sources, concentrations and deposition rates, and anthropogenic impacts

A worldwide compilation of atmospheric total phosphorus (TP) and phosphate (PO4) concentration and deposition flux observations are combined with transport model simulations to derive the global distribution of concentrations and deposition fluxes of TP and PO4. Our results suggest that mineral aerosols are the dominant source of TP on a global scale (82%), with primary biogenic particles (12%) and combustion sources (5%) important in nondusty regions. Globally averaged anthropogenic inputs are estimated to be similar to 5 and 15% for TP and PO4, respectively, and may contribute as much as 50% to the deposition over the oligotrophic ocean where productivity may be phosphorus-limited. There is a net loss of TP from many (but not all) land ecosystems and a net gain of TP by the oceans (560 Gg P a(-1)). More measurements of atmospheric TP and PO4 will assist in reducing uncertainties in our understanding of the role that atmospheric phosphorus may play in global biogeochemistry.

Chemical composition of aerosols collected over the tropical North Atlantic Ocean

Ambient aerosol samples were collected over the tropical northern Atlantic Ocean during the month of April 1996 onboard the R/V Seward Johnson. Dichotomous high-volume collector samples were analyzed for ferrous iron immediately after collection, while trace metals, anions, and cations were determined upon return to the laboratory. Data are analyzed with the aid of enrichment factor, principal component, and weighted multiple linear regression analyses. Average mineral aerosol concentrations amounted to 19.3±16.4 μg m^(−3) whereby the chemical characteristics and air mass back trajectories indicated the dust to be of a typical shale composition and Saharan origin. Calcite accounted for 3.0 and 7.9% of the mineral aerosol during the first and second halves of the cruise, respectively. Total iron concentrations (averaging 0.84±0.61 μg m^(−3)) are crustally derived, of which 0.51±0.56% is readily released as Fe(II). Eighty-six percent of this Fe(II) is present in the fine (<3 μm diameter) aerosol fraction and correlates with NSS-SO_4^(2−) and oxalate. Approximately 23% of the measured NSS-SO_4^(2−) in both size fractions appears to be biogenically derived, and the rest is of anthropogenic nature. Biogenic SO_4^(2−) /methanesulfonic acid (MSA) ratios could not be easily extracted by employing a multiple linear regression analysis analogous to that of Johansen et al. [1999], possibly due to the varying characteristics of the aerosol chemistry and air temperature during the cruise. Because of the presence of anthropogenic SO_4^(2−), the non-sea-salt (NSS)- SO_4^(2−)/MSA ratio, 37.4±6.4, is elevated over what would be expected if the NSS - SO_4^(2−) were purely biogenic. Cl^− depletion is seen in all samples and averages 18.3±9.1%. The release of Cl from the aerosol phase appears to occur through acid displacement reactions with primarily HNO_3 in the coarse and H_2SO_4 in the fine fraction.

Chemical characterization of ambient aerosol collected during the southwest monsoon and intermonsoon seasons over the Arabian Sea: Labile-Fe(II) and other trace metals

Atmospheric deposition of iron (Fe) to certain regions of the oceans is an important nutrient source of Fe to the biota, and the ability of the biota to uptake Fe is dependent on the speciation of the Fe. Therefore understanding the speciation of Fe in the atmosphere is critical to understanding the role of Fe as a nutrient source in surface ocean waters. Labile ferrous iron (Fe(II)) concentrations as well as total concentrations for Fe and other important trace metals, cations, and anions were determined over the Arabian Sea for two nonconsecutive months during 1995. Ambient aerosol samples were collected during the Indian Ocean intermonsoon and southwest monsoon seasons over the Arabian Sea. Sampling took place aboard the German research vessel Meteor in the months of May (leg M32/3; intermonsoon) and July/August (leg M32/5; southwest monsoon). Both cruise tracks followed the 65th east meridian, traveling for 30 days each (from north to south during leg M32/3 and from south to north during leg M32/5). A high-volume dichotomous virtual impactor with an aerodynamic cutoff size of 3 μm was used to collect the fine and coarse aerosol fractions for metal analysis. A low volume collector was used to collect aerosol samples for anion and cation analysis. The analysis for labile-Fe(II) was done immediately after sample collection to minimize any possible Fe redox reactions which might occur during sample storage. The analytical procedure involved filter extraction in a formate/formic acid buffered solution at pH 4.2 followed by colorimetric quantification of soluble Fe(II). Metals, anions, and cations were analyzed after the cruise. Total atmospheric aqueous-labile-Fe(II) concentrations during the intermonsoon were between 4.75 and 80%) was present in the fine fraction (<3.0 μm). During the southwest monsoon, atmospheric aqueous-labile-Fe(II) concentrations were consistently below the detection limit (<0.34 to <0.089 ng m^(−3), depending on the volume of air sampled). Air mass back trajectories (5 day, three dimensional) showed that air masses sampled during the southwest monsoon had advected over the open Indian Ocean, while air masses sampled during the intermonsoon had advected over northeast Africa, the Saudi Arabian peninsula, and southern Asia. These calculations were consistent with the results of the statistical analysis performed on the data set which showed that the variance due to crustal species during the intermonsoon samples was greater than the variance due to crustal species during the southwest monsoon. The factor scores for the crustal components were also greater when the back trajectories had advected over the nearby continental masses. Principal component analysis was also performed with the intermonsoon samples where aqueous labile Fe(II) was above the detection limit. Aqueous labile Fe(II) did not correlate well with other species indicating possible atmospheric processing of the iron during advection.

A revised nitrogen budget for the Arabian Sea

Despite its importance for the global oceanic nitrogen (N) cycle, considerable uncertainties exist about the N fluxes of the Arabian Sea. On the basis of our recent measurements during the German Arabian Sea Process Study as part of the Joint Global Ocean Flux Study (JGOFS) in 1995 and 1997, we present estimates of various N sources and sinks such as atmospheric dry and wet depositions of N aerosols, pelagic denitrification, nitrous oxide (N2O) emissions, and advective N input from the south. Additionally, we estimated the N burial in the deep sea and the sedimentary shelf denitrification. On the basis of our measurements and literature data, the N budget for the Arabian Sea was reassessed. It is dominated by the N loss due to denitrification, which is balanced by the advective input of N from the south. The role of N fixation in the Arabian Sea is still difficult to assess owing to the small database available; however, there are hints that it might be more important than previously thought. Atmospheric N depositions are important on a regional scale during the intermonsoon in the central Arabian Sea; however, they play only a minor role for the overall N cycling. Emissions of N2O and ammonia, deep-sea N burial, and N inputs by rivers and marginal seas (i.e., Persian Gulf and Red Sea) are of minor importance. We found that the magnitude of the sedimentary denitrification at the shelf might be ∼17% of the total denitrification in the Arabian Sea, indicating that the shelf sediments might be of considerably greater importance for the N cycling in the Arabian Sea than previously thought. Sedimentary and pelagic denitrification together demand ∼6% of the estimated particulate organic nitrogen export flux from the photic zone. The main northward transport of N into the Arabian Sea occurs in the intermediate layers, indicating that the N cycle of the Arabian Sea might be sensitive to variations of the intermediate water circulation of the Indian Ocean.

Infrared spectroscopic study of the deliquescence and efflorescence of ammonium sulfate aerosol as a function of temperature

The deliquescence and efflorescence phase transitions of ammonium sulfate aerosols have been studied as a function of relative humidity (RH) over the temperature range from 234 K to 295 K. Polydisperse submicrometer ammonium sulfate particles produced by atomization were monitored in a temperature-controlled flow tube system using Fourier transform infrared spectroscopy. The relative humidity in the aerosol flow was controlled using a sulfuric acid bath conditioner and the addition of a known flow of dry nitrogen. The relative humidity was measured using a dew point hygrometer and infrared absorption features. The deliquescence transition was observed to be nearly independent of temperature, changing from 80% RH at 294.8 K to 82% RH at 258.0 K near the ice saturation line, in good agreement with previous results. The relative humidity at the efflorescence transition also increased slightly (32% to 39%) with decreasing temperature (294.8 K to 234.3 K). These results suggest that once a crystalline ammonium sulfate particle deliquesces, the droplet can exist as a metastable solution droplet over a broad region of temperature and water pressures under the conditions in the upper troposphere. The persistence of metastable ammonium sulfate solution droplets may have important implications for cirrus cloud formation and heterogeneous reaction rates in the upper troposphere.

Iron photochemistry of aqueous suspensions of ambient aerosol with added organic acids

Experiments to simulate cloudwater conditions were carried out by using ambient aerosol samples suspended in an aqueous solution. Electron donors known to exist in atmospheric cloudwater (oxalate, formate, or acetate) were then added to the simulated cloudwater, and the solution irradiated with ultraviolet light while important species were measured (i.e., H_2O_2, Fe_(total), Fe(II)_(aq), and pH). A total of four different ambient aerosol samples were used in the simulated cloudwater experiments; they were collected from (1) Whiteface Mountain, NY, (2) Pasadena, CA, and (1) Sequoia National Park, CA. Hydrogen peroxide (H_2O_2) photoproduction was observed in the simulated cloudwater experiments with added oxalate. Fe(II)_(aq) photoproduction was observed in the simulated cloudwater experiments with and without added acetate or added formate using ambient aerosol collected simultaneously with the ambient aerosol used in the added oxalate experiments. The production of Fe(II)_(aq) showed that Fe from the ambient aerosol was available for photochemical redox reactions. In all cases, the production rates for Fe(II)_(aq) and H_2O_2 in the light were greater than production rates in nonirradiated control experiments. The simulated cloudwater experiments (with four different aerosol samples) showed similar behavior to previous experiments carried out with synthetic Fe-oxyhydroxy polymorphs in the presence of oxalate, formate, or acetate. The Fe present in the ambient aerosol appears to be a critical component for the production of H_2O_2 in the simulated cloudwater experiments.

Measurements of Trace Metal (Fe, Cu, Mn, Cr) Oxidation States in Fog and Stratus Clouds

The oxidation state of four transition metals (Fe, Mn, Cu, and Cr) in cloudwater has been investigated during several cloud events at Whiteface Mountain (NY), one cloud event at San Pedro Hill (CA), and one fog event at Bakers-field (CA). Samples were collected and immediately analyzed for the oxidation states of four transition metals: Fe(II) [44 measurements], Cu(I) [30 measurements], Mn(IV) [27 measurements], and Cr(III) [3 measurements]. Extreme care was taken to minimize contamination and interferences when measuring these oxidation states. Par-ticulate and dissolved concentrations of these metals were also determined. Other measurements performed-relevant to the redox chemistry of these metals-included pH, total elemental concentrations (Fe, Cu, Mn, Cr, Al, K, Ca, Na, and Mg), organic anions (formate, acetate, glyco-late, oxalate), inorganic anions (chloride, sulfate, nitrate, sulfite), cations (sodium, calcium, magnesium, potassium), peroxides, and formaldehyde. These measurements were then used in thermodynamic-speciation models to understand the speciation of ambient fog and cloudwater. From this analysis, two different cases were found for Fe(III)soluble speciation. Fe(III) was found to exist either as Fe(OH)2+ or Fe(Oxalate)2. However, an unidentified strong chelating ligand with Fe(III) was also suggested by the data. Cu(I) and Cu(II) were calculated to be predominantly Cu+ and Cu2+ (with less than 10% as Cu(II)-oxalate complexes). A chemical kinetic model was also used to investigate the transition-metal chemistry. The model results indicate that Fe(II) should be the predominant chemical form of Fe during daylight conditions. This prediction is in agreement with the field measurements in which the highest ratios of Fe(II)/Fe total were found in samples collected during the day. The model results also indicated that Fe(III) should be the predominant form of Fe during nighttime conditions, which is also in agreement with the field measurements. In the model, Cu(II) and Mn(II) were the predominant oxidation states during daylight and nighttime conditions with Cu(I) and Mn(III) increasing during daylight conditions. Mn(III) concentrations were never high enough to influence the redox chemistry of Cr. Overall, Cr(VI) in cloudwater is predicted to be reduced to Cr(III) if free S(IV) is present.

Chemical characterization of ambient aerosol collected during the southwest monsoon and intermonsoon seasons over the Arabian Sea: Anions and cations

Ambient aerosol samples were collected over the northern Indian Ocean during two 1 month-long research cruises (German R/V Meteor) that took place during the intermonsoon (May) and SW monsoon (July/August) of 1995. A high volume and two small volume collectors were used to collect samples, which were subsequently analyzed for ferrous iron, 32 elements, and anions and cations. The present paper focuses on the bulk aerosol material, the ions, while utilizing some of the trace metal data that were presented in more detail in our previous paper [Siefert et al., 1999]. Data are analyzed and interpreted with the aid of principal component and multiple linear regression analyses. Intermonsoon samples were strongly influenced by continental material, both of crustal and anthropogenic origin. The crustal component (24.5±13% of the total suspended particulate mass (TSP), 6.0±4.4 μg m^(−3)) contained 3.2% gypsum (CaSO_4). While more than half of the TSP (21.2±9.6 μg m^(−3)) during the SW monsoon was sea-salt-derived due to the strong winds prevailing during this season, only 1.7±1.1% (0.7±0.4 μg m^(−3)) was found to be of crustal origin. Sulfate (SO_4^(2−)) sources were determined and quantified with linear regression analyses utilizing specific tracers for the independent variables. Lead (Pb) was found to be a more reliable surrogate for anthropogenic SO_4^(2−) compared to nitrate (NO_3^−) during the relatively polluted intermonsoon. Soluble calcium (Ca^(2+)) served as the tracer for gypsum, and methane sulfonate (MSA) served as the tracer for biogenically derived SO_4^(2−) during both seasons. On the basis of this analysis, 75% of the non-sea-salt sulfate (NSS-SO_4^(2−)) (0.8±0.2 μg m^(−3), representing ∼2.4% of TSP) was found to be of biogenic origin during the SW monsoon with the remaining 25% of anthropogenic origin. During the intermonsoon, NSS-SO_4^(2−) accounted for 2.1±1.2 μg m^(−3) (∼9.2% of TSP) and had a composition that was 65% anthropogenic, 21% biogenic, and 14% gypsum-derived. Linear regression analyses revealed that the bio-SO_4^(2−)/MSA weight ratios appear to be consistent with the temperature dependence proposed by Hynes et al. [1986]. In this case the yield of SO_4^(2−) increased relative to MSA with an increase in temperature. Three samples during the SW monsoon, near the coast of Oman, showed lower temperatures, due to coastal upwelling, than the rest of the samples; at 24°C the bio-SO_4^(2−)/MSA weight ratio was 6.8±0.5. The remainder of the SW monsoon samples were collected at an average temperature of 27.2°C, for which the bio-SO_4^(2−)/MSA weight ratio was 13.5±4.4. At an average temperature of 28.9°C during the intermonsoon, sampling gave a ratio of 17.7±4.8. These observations indicate a temperature dependence factor between 24° and 29°C of 2.2 (i.e., a 2.2 increase in the ratio of bio-SO_4^(2−)/MSA with every degree temperature increase). Cl− deficits determined during both seasons appear to indicate that different mechanisms may govern the observed depletion of Cl− in each season.

Determination of photochemically available iron in ambient aerosols

Experiments to determine the concentration of photochemically available Fe in ambient aerosol samples were carried out using a novel photochemical extraction procedure. Ambient aerosol samples, which were collected on Teflon filters, were suspended in an aqueous solution within a photochemical reactor and irradiated. Under these conditions, which were favorable to the photochemical weathering of aerosol particles, the relative amount of Fe(II)_(aq) to Fe_(total) was shown to increase. The extent and rate of Fe(II)_(aq) photoproduction was used to characterize the Fe in aerosol samples collected from Whiteface Mountain, New York, Pasadena, California, San Nicholas Island, California, and Yosemite National Park, California. Photochemically available Fe concentrations found ranged from <4 ng m^(−3) (0.07 nmole m^(−3)) to 308 ng m^(−3) (5.52 nmole m^(−3)), Fe_(total) concentrations ranged from 10 ng m^(−3) (0.18 nmole m^(−3)) to 3400 ng m^(−3) (61 nmole m^(−3)), and the percentage of photochemically available Fe to Fe_(total) ranged from 2.8% to 100%. Aerosol samples were also collected during biomass burning events in southern California; these samples showed insignificant changes in the photochemically available Fe (compared to nonbiomass burning samples) in conjunction with large increases of Fe_(total). Calculations based on these experiments also provide further evidence that redox reactions of Fe in cloudwater could be an important in situ source of oxidants (OH, HO_2/O_2^−). The estimated oxidant production rate in cloudwater based on these experiments is between 0 and 60 nM s^(−1), with an average value of 16 nM s^(−1). This estimated in situ oxidant production rate due to Fe chemistry is approximately equal to previous estimates of the oxidant flux to cloudwater from the gas phase.

Photoreduction of iron oxyhydroxides and the photooxidation of halogenated acetic acids.

The photolytic reduction of ferrihydrite (am-Fe_2O_3*3H_2O), lepidocrocite (γ-FeOOH), goethite (a-FeOOH), hematite (α-Fe_2O_3), maghemite (γ-Fe_2O_3) and iron-containing aerosol particles (Fe_(aerosol)) in the presence of a series of halogenated acetic acids has been investigated. The fastest rates of photoreduction of Fe(lll) to Fe(ll) were achieved with ferrihydrite as an electron acceptor and fluoroacetic acid as an electron donor. The relative rates of photooxidation of the monohalogenated acetic acids with ferrihydrite in order of decreasing reactivity were as follows: FCH_2CO_2H > CICH_2CO_2H > BrCH_2CO_2H > ICH_2CO_2H; for multiple substituents the relative order of reactivity was as follows: FCH_2CO_2H > F_2CHCO_2H > F_3CCO_2H. With respect to the iron oxide electron acceptors, the relative order of reactivity toward monohaloacetate oxidation was am-Fe_2O_3-3H_2O > γ-Fe_2O_3 > γ-FeOOH ≥ α-Fe_2O_3 ≥ Fe_(aerosol) > α-FeOOH. Strong kinetic isotope effects observed for the photooxidation of CICD_2CO_2H suggest that the oxidation of the mono- and disubstituted haloacetic acids proceeds via hydrogen-atom abstraction by surface-bound hydroxyl radicals to produce haloacetate radicals, which in turn yield the corresponding halide and glycolic acid. Fully halogenated haloacetic acids appear to be oxidized via a photo-Kolbe mechanism to yield the corresponding halo acids and CO_2.