Preliminary evidence of a biomass-burning event from a 60

Transcription

1 Annals ofglaiology nternational Glaiologial Soiety Preliminary evidene of a biomass-burning event from a 6 year-old firn ore from Antartia by ion hromatographi determination of arboxyli aids R. UDST, S. BECAGL, R. TRAVERS, S. VERMGL, G. PCCARD Department of Publi Health and Environmental Analytial Chemist 1]" University oj Florene, Via G. Capponi 9, Florene, taly ABSTRACT An ion hromatographi method for the determination of fluoride and some organi anions (latate, aetate, glyolate, propionate, formate, methane sulphonate and pyruvate) at sub-lg 1-1 onentration levels is applied to a 22 m (63 years) firn ore drilled at Herules Neve, northern Vitoria Land, Antarti The first effetive distribution data of aetate, propionate and formate for this region are reported here. Mean values of9.9, 7.7 and 2.41lg 1 were alulated for aetate, formate and propionate, respetively, in the absene of partiular events (bakground values). The temporal ( ) onentration/depth profiles of these omponents showed a probable long-term biomass-burning event in the 193s that an be related bak to proesses in the Southern Hemisphere. Suh burning events may be demonstrated from the simultaneous onentration inrease of arboxyli aid, potassium, ammonium and fluoride and from a very high dust ontent. Similar inreasing trends in the same time period are not shown by other snow omponents usualli onsidered as soure indiators (methane sulphoni aid: biogeni marker; nsss4 : biogeni and volani indiator; Na + and Cl : sea-spray ontribution; Ca 2 +: rustal input), so it was possible to exlude suh alternative explanations. NTRODUCTON n areas of undisturbed snow aumulation it is possible to determine, by using the orret hemial and isotopi traers, individual snow layers aumulated layer by layer over time. By assuming a lose orrelation between the atmospheri aerosol omposition and the snow omposition (Lambert and others, 1983; Pourhet and others, 1983) and onsidering, where possible, the effets of eventual post-depositional proesses, a detailed history of the annual and seasonal variation of the atmospheri omposition from the snow hemial analysis an be reonstruted. n northern Vitoria Land, East Antartia, the relative loseness of plateau-like or high-altitude neve areas to the oastline, and to the talian Terra Nova Bay base, makes aess to undisturbed snow-aumulation sites easier. Chemial and physial analyses of firn and ie ores sampled in suh areas permit the observation of environmental variations over several deades with detailed analyses of the seasonal behaviour of the main and seondary soures of atmospheri aerosol. The study of arboxyli aid presene in Antarti snow an give useful information about soures and transport proesses of the oxidation produts of substanes related to biologial yles (Berresheim, 1987; Keene and Galloway, 1988; Legrand and Saigne, 1988; Andreae and others, 199; Watts and others, 199). Although many soures were postulated for arboxyli aids in polar atmosphere (suh as diret input from biologi ativity or seondary emission by oxidation of marine and ontinental biogeni hydroarbons), further studies are indispensable for understanding their atmospheri yles and their role in the weak aidity of snow preipitation (Legrand and de Angelis, 1995, 1996). Formate and aetate are diretly emitted by vegetation and biomass-burning or from anthropogeni ativity (Talbot and others, 1988, 199). The biomass-burning, usually underestimated, an represent a signifiant input of hydroarbons in remote areas. n partiular, the methane produed yearly by biomass-burning is about 7% of global methane emissions (Dignon, 1995). Formi aid prodution is related to the oxidation proesses of hydroarbons emitted by forest areas (mainly isoprene) and to the natural yle of formaldehyde Uaob, 1986; Jaob and Wofsy, 1988, 199). Measurements of aeti and formi aid in polar snow are sporadi. The first detailed investigation of arboxyli aid presene in Greenland an be found in two reent papers of Legrand and de Angelis (1995, 1996) onerning GRP projet samples. The presene of propionate in atmospheri preipitation was shown by measurements in rainwater (Tsidouridou and Puxbaum, 1987; Murray, 1989), but propionate measurements on snow samples from Antartia and Greenland are missing. n superfiial snow and ie samples olleted at Terra Nova Bay, East antartia, the onentration of propionate was always below the detetion limit (Udisti and others, 1991). n this paper an ion hromatographi (C) method for the determination of some organi anions and fluoride at sub-fig 1-1 onentration levels is applied to a 22 m firn ore from Herules Neve, Antarti The first data on distribution and temporal profiles (aboul 6 years) of some shorthain arboxyli aids (formi, aeti and propioni aid) 391

2 Udisti and others: le determination ifarboxyli aids in a 6 year-oldfirn ore in northern Vitori a Land are reported here. Preliminary onentration/depth resul ts show a probable biomass-burning event, ourring in the 193s, evidened by high levels of arboxyli aids, fluorides, potassium and, with less evidene, ammonium. METHODOLOGY Sampling site During the tali an Antarti Campaign, a 22 m firn ore was drilled at H erules Neve station, northern Vitoria Land. Figure shows the geographi position of the sampling site. orthern Vitoria Land is a region with a omplex orography influened by strong katabati winds (personal ommuniation from B. Stenni, 1997), so an aurate preliminary study was neessary to isolate relatively undisturbed sampling areas. Satellite photos showed that Herules Neve is not influened by snowdrifts and sastrugi, so that this area seems to be unaffeted by wind redistribution effets. A previous shallow firn ore (8 m; Udisti, 1996) onfirmed an undisturbed annual snow deposition. Herules Neve is a plateau of approximately 11 km 2, about 3 m s. l. and about 75 km from the sea (Lady Newnes Bay). t omprises the largest ie ap of northern Vitoria Land with the adj aent Evans Neve. The site is loated on the ie divide between the glaiers that fl ow into the Ross Sea a nd into the Paifi O ean. Mean annual temperature values of about - 33 C were estimated for the sampling station from the temperature data (-34.5 and C at 7.5 and 1 m depth, respetively) measured during the firn oring (personal ommuniation from B. Stenni, 1997). Sampling and analyt ial proedures During the oring, sampling and analysis proedures, all preautions were taken to minimise ontamination, and the personnel wore lean-room lothing. T he firn ore was drilled using a SPRE rotating ie-orer (eletrial 1 W engine) with a stainless-steel orer tube (1 m diameter, 12 m long) and fibreglass orer rods (12 m long). This orer was aurately leaned before use and produed ore segments 8-9 m long. T hese were leaned externally on site, labell ed a nd pl aed in doubl e polyethylene sealed bags. The firn-ore setions were kept frozen and brought to the lean-old room of the Laboratoire de Glaiologie in G renoble, Frane (Maggi and Petit, 1998). The firn ore was leaned by mehanially removing a thin ex ternal layer, and subsamples were obtained by utting the firn ores about every 5 m. T hese were stored at - 2 C in preleaned polye thylene ontainers plaed in double polyethylene bags. Due to eletrial failure of a old store, the first 4m were not analyzed. n the analytial laboratory of the University of Florene the samples, sti ll in their ontainers, were melted just before analysis under a lass-loo laminar-flow hood and filtered on a pre-leaned.45 flm Teflon membrane immediately after melting. Partiular are was taken to open the ontainers as little as possible, in order to minimise any uptake of ammoni a and arboxyli aid from the laboratory atmosphere (Udisti and others, 1991, 1994). 392 N Et) ROSS SEA -; ,W 5 km Fig. 1. Geographi position if the sampling station: station 36, Herules Nivi (73" (/ S, 165"28' E; 296 m s. l), about 9 kmjrom the oastline. Experimental Apparatus T he hemial analysis was performed using a Dionex 4i C equipped with a gradient pump and a ondutivity detetor (CDM-l). T he ondutivity suppression was performed by an eletrohemial anion membrane suppressor (Dionex ASRS-l). The anion separation was obtai ned by a Dionex AGll guard olumn a nd a Dionex ASll separator olumn. Tetraborate.9 and 3 mm was used as eluent aording to a step-eluent separation proedure. The eluent was degassed and pressurised with helium by an eluent degas module (Dionex EDM-2). The hromatographi data were olleted and elaborated by Dionex PeakNet integration software. A mlloop was used for all sample analysis. Reagents Stok standard solutions ( g - ) were purhased from Merk when available, or obtained from "reagent grade" Merk or Fluka reagents dissolved in ultra-pure water (resistivity > 18 MQ). Sta nda rd solutions with lower onentration were prepared daily in pre-leaned polyethylene bottles. RESULTS AND DSCUSSON Analytial method The orgaru anions (latate, aetate, propi onate, glyolate, formate, methane sulphonate and pyruvate, plus flu oride) are present in snow from Anta rtia at very low onentrations, ranging from sub-pg to pg 1 levels. Moreover, these anions are weakly retained into the C separator olumns used for the inorgani anion determination (eluent:

3 Udisti and others: C determination qf arboxyli aids in a 6 year-old firn ore.5 s.4 a :. NQ. Q.Ll). ';! mu'). Lf)... :g o CV Q) as 1ii J!! "' "' l <.... (/) 1ii E > "' '' - :::;; -'<Cl e 2 LL..» b... "!.3 2: -8. fs:. Lt) N "' 2: o (").l!!.. " l" Oi a:i.2 :g! E « (; (Jl g e LL :::;; T: -' C.3 i...2 :g g u: Minutes Fig. 2. (a) Chromatographi separation qf fluoride and some O1gani anions (standard solution). The omponents determined and their onentrations are reported at the top qf eah peak. Operative onditions and step -eluent programme: Separation and guard olumns: Dionex ASll + AGll; loop: 1 ml; eluent flow: 2 ml min -. Eluents: El.1 = H2 Milli-Q; El.2 = 3mM Na2B47, El.3 = 3 mm Na2B47. Time %E.l % E1.2 % E1.3 Comment. 7 3 Start of analysis Start of l eaning Stabilisation (b, ) Chromatogram qf a real sample from Herules Nivi station (b), ompared with operational blank (); same onditions as above. The omponents determined and their onentrations are reported at the top qf eah peak ( ppb = Jig ). arbonate/hydrogenarbonate buffer) so that their separation is diffiult. An le method for a omplete separation of the onsidered omponents and able to determine sub-pg - onentration levels is required. Various methods for the simultaneous determination of organi and inorgani anions are reported in the literature (Roklin and others, 1987; Cheam, 1989, 1992; Legrand and others, 1993). They are usually based on a gradient elution with NaOH solutions. Under the reported onditions, however, the separation between some omponents with similar retention times (in partiular, aetate/glyolate/propionate and methane sulphoni aid (MSA) /pyruvate) is insuffiient for their simultaneous determination. Moreover, the baseline drift, due to the NaOH onentration gradient, prevents the determ i nation of low onentrations. The detetion limits reported for the baseline subtration tehnique (Cheam, 1989) (1-8 lg ; loop = 5 ll ) are too high to analyze Antarti samples with pg or SUb-Jig onentrations. After a omparison of three r methods (ion exlusion and anioni exhange in isorati and gradient mode), Cheam (1992) onluded that the isorati method produes more reproduible results and lower detetion limits. Udisti and others (1994) proposed two alternative methods based on isorati or gradient separation using a Dionex AS5A-5 Jlm separator olumn and tetraborate as eluent. The gradient methods gave better separation (in partiular between aetate and propionate) but lower reproduibility by baseline drift. The isorati method, on the other hand, did not fully separate some omponents. n this paper, the isorati method was improved using a Dionex ASll separator olumn and.9 mm tetraborate as eluent. n effet, after the separation of the organi aids, a more onentrated Na2B47 solution (3 mm) is pumped into the system to elute the inorgani anions trapped in the olumn. Therefore, the above "isorati" method should be alled a "step-eluent" method. During the leaning step, the inorgani anions are quikly eluted without effetive separation. Na2B47 was preferred to NaOH beause of its lower apaity to absorb atmospheri CO2. The arbonate so formed inreases the eluent strength to the detriment of the measurement reproduibility. The use of degassed (helium) Na2B47 solutions kept retention times stable for more than a week (Udisti and others, 1991). The hromatogram of a standard solution ontaining Jig - ofalllhe analyzed omponents is shown in Figure 2, whih also shows the experimental onditions. Table shows the analytial performane of the method (linearity range, sensitivity, reproduibility and detetion limits) for aetate, propionate and formate. The reprodui- Table 1. Linearity range, reproduibility and detetion limit for aetate, propionate and formate. Same onditions as Figure 2 Component Lill. range ppb Sellsl:tivity ns/ppb Reproduibility COliC l'viean Std. dev. ppb ms % Delfetion limit Con. Alean Std. dev. Del.lil11. ppb ms % ppt Aetate Propionate Formate Notes: The linearity range was obtained with at least 12 alibration points. Mean and standard dviation valus were alulated using ten onseutin measurements ofth same standard. The detetion limit was alulated as twie the standard deviation obtained rrom ten repliate masurments of standard solutions with onentration of.5 ppb (aetate and formate) and 1. ppb (propionate). ppb = y \ ppt = ng 1 393

4 Udisti and others: le determination qfarboxyli aids in a 6 year-oldjirn ore Table 2. Statistial parametersjor aetate, propionate andjormate in all thefirn ore, m setion (bakground level) and m setion ( biomass-burning events) Aetate Allflm-ore 4- /4.3 m J m Allfirn-ore sam/lfes samples samples samples Propionate Formate m m Allflm-ore m m sam/jfes samples samples samples samples N Min. fig Max. fig Mean pg l Std. dev. pg l Median pg l % pgl % pgl bility values (better than 2.% at 5 J1g - ) and the detetion limits (lower than.15 /-tg - ) permit the effetive determination of these omponents at very low onentration levels in Antarti snow. The low detetion limits are due to the very stable baseline value, to high peak reproduibility and to low blank values. Figure 2b shows the ion hromatogram obtained for a real sample ontaining Ouoride, aetate, propionate, formate and MSA at /-tg or sub-j1g - onentration levels, with respet to a blank (polyethylene ontainer) plotted on the same sale. Legrand and others (1993) advised against the use of plastis, to avoid possible ontamination. n our experiene, aurately leaned plasti ontainers (either polyethylene or polypropylene ontainers, disposable and heaper than glass bottles ) showed low blank values: below the detetion limits for propionate and in the range.1-1. J1g - for aetate and formate. Firn-ore data dis u ssion Among the omponents analyzed in the rirn ore, in this paper we pay attention to aetate, formate and propionate. A reliable dating is neessary to obtain the temporal setting of the sam pies. The urn-ore dating was performed by a ombination o[seasonal and absolute markers. For seasonal snow-layer identifiation (relative dating), we used a visual omparison between 1) 18 depth profile (1-22 m ) and a hemial profile (4-22 m) obtained by a linear ombination of the normalised onentration/depth profiles of three summer markers (MSA, nsss4 2 and H 2 2 ) (Udisti,1996; personal ommuniation from B. Stenni, 1997). Absolute dating levels were determined by the tritium onentration! depth profile. This dating has been disussed by Stenni (personal ommuniation, 1997). The dating results showed 64 summer peaks along the firn ore (1-22 m below the snow surfae) orresponding to 63 years from 199 to Chemial data have been measured on a 4-22 m setion ( ). Table 2 shows the basi statistial data onerning the distribution of the three arboxyli aids in the firn ore. The statistial values an be onsidered reliable beause the three omponents have been determined on all 351 subsamples (only four values for propionate and one value for aetate are missing). These ompounds are present in the Antarti snow at very low onentrations. The mean values of 16.9± 15.9, 17.4± and 3.8±2.5J1gl- l, found respetively [or formate, aetate and propionate, are very lose to those reported by Legrand and others (1993) for Greenland preipitations (only formate and aetate; mean values re- 394 spetively 12 and 13 J1g - ). Legrand and de AngeJis (1995 and 1996) reported mean bakground Holoene values of 9.3 ± 1.4 and 1.7 ± 1.7 /-tg 1-1 for aetate and formate at Summit, Greenland. Figure 3 shows aetate, formate and propionate distribution graphs (box plots). Eah box ontains 5% of the data, with the median value displayed as a line. T he sample dispersions are quite high, as shown by the relatively large interval between the 25th and 75th perentile values (box 1 8 :.. 6 o ::. 4 : a b Propionate Formate Aetate 8 6 Ol =l 4 2 b!lo Propionate Formate Aetate :.. o ::1.4 Pr opionate Formate Aetate 2 1 Fig. 3. Box plots if propionate, formate and aetate Jor all samples (a), m samples (b) and m sample ( ). Eah box ontains 5 % qf the data, with the median value displayed as a line. The top and bottom qf the box mark the limits qf ± 25% qfthe variable population (25th and 75th perentile). The lines extendingjrom the top and bottom qfeah box mark the minimum and maximum values that Jall within an aeptable range (1.5 times the box width). Any value outside qf this range ( outlier ) is displayed as an individual point.,

5 Udisti and others: lc determination ifarboxyli aids in a 6 year-oldfirn ore f! : JJ..::!!!:!:!.,L.w..:..:!L...::!!.J.b _li.1..gjjljjj. 1._::...: Depth (m) Fig. 4. Conentration/depth prifiles if formate (a), aetate ( b ), propionate ( ), potassium ( d), ammonium (e) and fluoride (j) at Herules NiDi station. height) and by the standard deviation value, of the same order of magnitude as the mean value. Moreover, these data dispersions are not symmetrially distributed, as shown by the median values (line in the box) shifted towards the bottom of the box (above all, for formate and aetate) and lower than the mean values. Formate and aetate median values are 58% and 7%, respetively, of the mean values. A few samples, in fat, have onentration values higher than the mean (up to 5.5 times for formate and aetate and 3.5 for propionate). Suh high values are not regularly distributed on all the firn ore. Figure 4a- show the onentration/depth profiles for the three omponents. n the first 15 m the onentrations for all three ompounds are lower than for the total firn-ore mean value ( m depth mean values of9.9 ± 4.8,7.7 ±5. and 2.4 ± 1. /lg 1- ' for aetate, formate and propionate, respetively). This part of the firn ore shows no partiular onentration/depth trend, and only a slight onentration inrease with depth an be observed for all the omponents. At 14.6 and 15.7 m depth (about 1947 and 1943, respetively; personal ommuniation from B. Stenni, 1997), two sharp spikes with relatively high onentrations (up to nearly 1 flg - ) are found for aetate and formate. Lower maximum values (about 1 fig ), but 3.5 times higher than the mean, are found at the same depths for propionate. For depths lower than 16. m, the onentration profiles show a ommon sharp inrease, with values onsistently higher than the bakground. The m depth (orresponding to the

6 Udisti and others: l C determination qfarboxyli aids in a 6 year-oldj irn ore 4 1 a d - - y = x R=.88 8 :;" 3 :- : a... 'Oi.. 4 < " " 11- hf : : <... 2.: _... o., Formate (g 1" 1) Formate g 1" 1) 6 2 b e --y = x R= ' : '.. : 4' Oi...,...- #t:' ;t... 1 e :> 's /. : 's.. :> :> g"..... '... g" 1 - t., J Formate (g r 1) Forma te (g r 1) C 1 2 :- 1 5 S S 1 S.. S e e f --y = x R=.388 S 5 S S S 5 < < Formate g r1) Formate ( g r1) Fig. 5. Conelations qf aetate, propionate and ammonium vs formate onent ration for m (a-) and m ( d-j) samples. 28 peri od ) mean values of 27.9 ± 13.6, 29.9 ± 16.9 and 5.8 ± 2.7 f1g, found for aetate, formate and propionate, respetively, a re 2.8,3.9 and 2.4 times higher than the orresponding m depth mean values. This behaviour is shown also by the box plots in Fi gure 3b-, where an inrease of median values and of data dispersion for the deepest firn-ore se tion is learly evident. Biomass-burning inputs a re a possible explanation. n partiula r, a long-term event seems to be indialed in lhe 193s by the steadil y high onentrations from 18 to 22 m ( ). Figure 5 shows the relati onships of aetale, propionate and NH4 + with formate (onsidered as the best biomassburning marker; Legrand and de Angeli s, 1995, 1996) in the m (Fig. 5a- ) and m (Fig. 5d- f) firn-ore setion. We an obse rve that a linear orrelation is miss ing in the upper firn-ore part, but it is learly evident for the organi aids in the bottom (orrelati on oeffiient R =.88,.89 and.39 for aetate/form ate, propionate/formate and ammonium/formate, respetively). The orrelation between ammonium and form ate is lower than that found by Legrand and de Angelis (1995, 1996) for biomass-burning events at Summit, Greenland. The linear regression slopes indiate a mean ratio of.64 and.14 for aetate/formate and propionate/form ate, respetively. An aetate/formate ratio of around 1 was found by Talbot and others (1988) from atmospheri measurements during biomass-burning events. During the transport, formi aid prodution ours in the plume, so that the aetate/formate ratio reahes values lower than 1 (Legrand and de Angelis, 1995). The aetate/form ate ratio found in lhe firn ore (.64 w/w) agrees with this result. High onentration values for arboxyli aids, NH/, K + and F- are reported as indiati ng biomass-burning events (Cahier, 1995; Legrand a nd de Angelis 1995, 1996; Legrand and others, 1995). n order to onfirm analogous behavi our, lhe onentrations of K +, NH4 + and F- in the firn ore were plotted vs depth. Figure 4d- f show the relative onentration/depth profiles: a onentration inrease is evident starting from 14.3 m depth for these ompounds. This inrease is in-phase with arboxyli aid, although a orrelation with formate, aetate a nd propionate spikes at 14.6 and 15.7m depth is not visible for K +, NH4 + and F-. The mean onentrations in the bottom of the firn ore are 4.2 ± 6.4, 4.3 ± 2.9 and.67 ±.95 Jg 1-1 for K +, NH/ and F-, 1.3, 1.4 and 2.6 times higher, respetively, than the relative mean values in the m firn-ore part. For K +, the onentration inrease is more evident onsidering the median value (the K + bakground mean is affeted by some samples with high sea-spray ontent). An inrease of 2.5 times is measured in the firn-ore bottom setion with respet to the bakground. This means that a larger number of samples with hi gher onentrati on ours. Similarly, the 396

7 Udisti and others: le determination ofarboxyli aids in a 6 year-oldjirn ore onentration/depth profiles of Na + (sea-spray marker), nssso/ - (biogeni and volani indiator) and Ca 2 + (rustal input) were observed, exluding any alternative explanation of the experimental dat Mean and median values of these ompounds did not show any inrease trend in the m Urn-ore setion with respet to bakground. Therefore, a biomass-burning event seems to be the more probable explanation for arboxyli aid, K +, NH4 + and F ontemporaneous inreases. High dust levels found in the same firn ore at m depth (Maggi and Petit, 1998) ould reinfore the hypothesis of the biomass-burning effets. Preliminary data show that this partiulate has a predominant organi omposition (about 5%; personal ommuniation from V Maggi, 1997) and ould be onstituted by ash or other arboni material. A strutural analysis of the partiulate will be performed to onfirm the presene of ash. The presene of light partiles (suh as ash) ould explain the very long residene times of the above-mentioned hemial markers and dust in the atmosphere. CONCLUSONS The eluent-step le method used for the determination of aetate, formate and propionate in Antarti snow samples gives a sensitivity suffiient to obtain the first data onerning the distribution of arboxyli aid in northern Vitoria Land. The isorati separation step permits high peak resolution with better baseline stability and lower detetion limits with respet to gradient methods. The appliation of this method to a 22 m firn ore drill ed at H erules Neve gave bakground levels of about 1 and 811g - for aetate and formate (with a ratio lose to 1.) and of about 2{lg 1-1 for propionate in the first 14.3 m firn-ore depth. Preliminary data on relatively high inphase onentration values for arboxyli aids, fluoride, ammonium, potassium and dust evidened a long-term biomass-burning event. Further analyses on the pa rtiulate are needed to onfirm the latter so ure, beause of its persistene in atmospheri aerosol [or a relatively long period. On the other hand, biomass-burning events in Antartia an only ome from long-range transport. Therefore, their fingerprints should be different with respet to the observations of Leg rand and de Angelis (1995, 1996) and Legrand and others (1995) at Summit, Greenland, where high, sharp onentration spikes were found. The long-range transport smoothing effets ould give slight onentration inreases for long time p eriod s. ACKNOWLEDGEMENTS This researh was arried out within the framework of a proje t on glaiology and palaeolimatology of the Programma Nazionale di Rierhe in Antartide, and finania lly supported by Ente Nazionale Energia e Ambiente through an agreement with Universita' degli Studi di Milano. REFERENCES Andreae, M.., R. W. Ta lbot, H. Berres heim and K. M. Beeher Preipitation hemistry in entral Amazoni J Geophys. Res.. 95 (D), 16,987-16,999. Berresheim, H Biogeni sulfur emi ssions from the subantarti and Antarti oeans. J Geop/!ys. Res., 92 (D ), 13,245-13,262. Cahier, H Combustion arbonaeous aerosols in the atmosphere: impliations for ie ore studi es. n Delm as, R.J., ed. e ore studies ofglo bal biogeohemial yles. Berlin, et., Springer-Vrlag, Cheam, V Determination of organi and inorgani aids in preipitation sa mpl es. J Chromatog., 482 (2), Cheam, V Comparison of ion hromatographi methods for the determination of organi and inorgani aids in preipitat ion sa n1ples. Analyst, 117, Dignon, J mpat of bi omass burning on the atmosphere. n Delm as, R.]., ed. e rore studies of global biogeollemiai yles. Berlin, et., Springer Verlag, (NATO AS] Series : Global Environmental Change 3.) J aeob, n] Chemistry of OH in remote louds and its role in the prodution of formi aid and peroxymonosulfate. J GeojJh)'S. Res., 91 (D9) J aob, n]. and S. C. Wor.y Photohemistry of biogeni emi ssion over the Am azon forest. J Geophys. Res., 93(D2), J aob, nj. and S. C. "Vofsy Budgets of reative nitrogen, hydroarbons and ozone over the Amazon forest during the wet season. J Geo ph)s. Res., 95 (D), 16,737-16,754. Keene, W. C. and]. N. Gall oway The biogeohemial yling of formi and aeti aids through the troposphere: an ovel'yiew of urrent understanding. TelLuJ, 4B (5), Lambert, G., B. Ardouin and A. l\1esbah-bendezu Atmosphere to snow transfer in Antarti in Pruppaher, H. R., R.G. Semonin and W. G. N. Slinn, eds. Preipitation savenging, dry deposition, and rfsllspension. New York, Elsevier Siene Publishers, Legrand, M. and M. de Angeli s Origins and variations of light arboxyli aids in polar preipitation. J Geo/lhys. Res., 1(D ), Legrand, M. and M. de Angelis Light arboxyli aids in Greenl and ie: a reord of past for es t fires and vegetation emi ss ions from the boreal zone. J Geojl/!ys. Res., 11 (D 2), Legrand, M. and C. Saigne Formate, aetate and methanesulfonate measurements in Antarti ie: some geohemial impliations. Atmos. Env;ron., 22 (5), Legrand, M., M. de Angelis a nd F. Maupetit Field investigation of ll,yor and minor ions along Summit (entral Greenl and) ie ores by ion hromatography. J Chromtog.. 64 (\-2), Legrand, M., M. de An ge lis, H. Cahi er and A. Gaudihel Boreal biomass burning over the last 8 yea rs reorded in a Summit-Greenland ie ore. n Delmas, R.]., ed. e ore studies ofglobal biogeohemial y les. Berlin, et., Springer-Verl ag, (NATO AS! Seri es!: Global Environmental Cha nge 3.) J\1aggi, V. and].-r. Petit Atmospheri dust onentration reord from the Herul es Neve fim ore (northern Vi toria Land, Antartia). Ann. Glaiol., 27 (see paper in th is volume). Murray, F Aid rain and aid gases in Australi Arh. Environ. Con tam.toxi., 18, Pourhet, M., ]. F. Pinglot and C. Lorius Some meteorologial appliations of radioative fa llout measurements in Antarti snows. J Geo '/!Ys. Res., 88(C), Roklin, R. n,. A. Pohl andj. A. Sh ibler G radient elution in ion hromatography. J Chromatog (1), Talbot, R. W., K. M. Beeher, R. C. Harriss and W. R. Cofer, ll Atmos pheri geohemistry of rormi and aeti aids at a mid-latitude temperate site. ] Geop")s. Res., 93 (D2), Talbot, R. W., M.. Andreae, H. Berresheim, n]. J aob and K..\1. Beeher Soures and sinks of formi, aeti and pyruvi aids over entral Amazoni 2. Wet season..j Geoph),s. Res., 95 (D), 16,799-16,811. Tsitouridou, R. a nd H. Puxbaum Appliation ofa portable ion hromatograph for field site measurements of the ioni omposition of fog water and atmospheri aerosols. lnt. J En viro n. Anat. Clwn., 31, Udisti, R Multiparamtri approah for hemial dating or snow layers from Antarti lnt. J En viron. Anal. Chem., 63, Udisti, R., E. Barbolani and G. Piardi Determination of some 1' ga rli and inorgani substanes present at ppb level in Antarti snow and ie by ion hromatography. Ann. Chi",., 81 (7-8), Udisti, R., S. Bell andi a nd G. Piardi Analysis of snow from Antartia: a ritial approah to ion-hro matographi methods. Fresenills' J Anal. Chem (4), Watts, S. F., P. Brimbleombe and A.]. Watso n Methanesulphoni aid, dimethyl sulphoxide and dimethyl sulphone in aerosols. Atmos. Environ., 24A (2),