Processes controlling the composition of precipitation at a remote

Document Sample
Processes controlling the composition of precipitation at a remote Powered By Docstoc
					                  JOURNAL OF GEOPHYSICAL            RESEARCH,       VOL. 101, NO. D3, PAGES 6883-6897, MARCH 20, 1996

Processescontrolling the compositionof precipitation at a
remote southern hemispheric location: Torres del Paine
National Park, Chile
JamesN. Galloway and William C. Keene
         of            Sciences,
Department Environmental       Universityof Virginia

Gene E. Likens
Instituteof EcosystemStudies,Millbrook, New York

                                 measuredat Torres del Paine National Park, Chile
Abstract. Precipitationcomposition
(TdP) (51ø10'S, 71ø58'W), between1984 and 1993 was acidic (volume-weighted-
average of 4.96) witha dilute      component. + wasthedominant
                             seawater      H                non-sea-
        cation; decreasing
salt(nss)     in                         nss
                         orderof abundance, anionswereHCOO-, CI-, SO4 =,
CH3COO-,andNO3-. Relative lowerlatitude,remote         concentrations
                                              locations,           and
        depositionsnss =, NO3-,andNH4+ at TdPwerelower;
per-event        of   SO4                                 of
     t                           t         and
       andthose CH3SO werehigher. Concentrations deposition
similar;       of   3-                        and         fluxesof
             nss =,      and
HCOOt,CH3COOt, SO4 CH3SO3-, NH4+ varied       with
                                       seasonally summer
maximaand winter minima. Carboxylicacidsprobablyoriginatedfrom both direct
          emissions oxidationof hydrocarbons
terrestrial       and                       emittedby marine and terrestrial
             =          3-
biota. NssSO4 andCH3SO originated   primarily            of       (CH3)2S
                                             fromoxidation biogenic
emitted               PacificOcean. Directemissions NH3 fromupwind
       fromthe southern                           of
         and     ecosystems
terrestrial marine              accounted most
                         probably      for          NH4+. NO3-
concentrations depositions    were highestduringthe latter part of australwinter and
springsuggesting                            of
                  abioticcontrols. Transport precursors   from lightning,biomass
burning,and fossil-fuelcombustion lower latitudes  andpossibly            of
                                                                 transport reactive
N from the stratosphere apparently           oxidizedN to the southern
                                   contributed                           Patagonian
                                         of            at
troposphere.Althoughthe ionic strength precipitation TdP is currentlyamongthe
                              are              of            in
world's lowest, future changes likely because increases local and regional
populationand energyand food production.

Introduction                                                  This paper reports the annual and monthly volume-
   Many importanttropospheric  compounds produced
                                           are          by weighted-average (VWA) concentrations and associated
terrestrialand marine organisms. Several of thesechemical wet-depositionfluxes of major chemical constituentsin
species their reaction   products)                                      at
                                   (e.g., SO4 CH3SO3-, precipitation TdP. On the basisof thesedata and related
NO3-, NH4+, HCOOH, CH3COOH haverathershort                                             the
                                                           information,we investigated major processes   controlling
atmosphericlifetimes (e.g., less than 1 week) and are lost tropospheric chemistry in the region with a focus on
from the atmospherevia deposition. For such species, biosphere-atmosphere        interactions. We also assessed the
measured concentrationsand deposition rates provide potential       impacts thatmay resultfrom projectedincreases in
important            on
           constraints sourcestrengths  associatedwith the anthropogenic  emissionsover the next few decades.
corresponding  biologicalprocesses. However, the perva-       These data were generatedunder the auspicesof the
siveness humaninfluences     over mostof the globecompli- GlobalPrecipitation  ChemistryProject (GPCP). From 1979
catesthe assessment natural linkages between the atmo- to 1994 the GPCP operateda network of stations deter-
sphereand the biosphere. In extremely remote locations,    mine the processes controllingprecipitationcomposition in
such as Torres del Paine National Park (TdP) in southern   regions of the world remote from the influencesof human
                                                                                   a         for       with
Chile (51ø10'S, 71ø58'W) (Figure 1), these relationships activitytherebyestablishing baseline comparison
can be investigatedwith reasonableconfidence. TdP is more populatedregions.
downwind          of the   southern   Pacific   Ocean   and   far   from
human-population industrialinfluences.                                     Methods

        1996by theAmerican
Copyright                         Union.
                         Geophysical                                          Site Description
Papernumber95JD03229.                                                        TdP is 400 km north of PuntaArenas (Figure 1). The
0148-0227/96/95JD-03229505.00                                              park covers approximately240,000 ha and, in 1978, was
6884                 GALLOWAY        ET AL.:     PRECIPITATION         AT A REMOTE            SITE IN CHILE

                                   80 ø               70 ø               60 ø                 50øW

                            30 ø                                                                  30 ø

                                                                                                  40 ø
                            40 ø                                                Atlantic

                           50øS                                                                   50øS
                                                                     Tierr del Fueg,

                                    90 ø         80 ø        70 ø        60 ø         50øW

                            1.     del   National Chile, surrounding
                       Figure Torres Paine     Park,   and       areas.
declared a member   of the International    Network     of Bio-     then       for     =                 t
                                                                        analyzed pH, SO4 , CI-, NO3-,HCOO (HCOO-+
sphere Reserves by the United Nations' Man and the                              t       4-      ),  +,
                                                                    HCOOH),CH3COO(CH3COO- CH3COOH NH4
Biosphereprogram. TdP and its surroundings formed                     +, ++, ++, K      3- measured
                                                                    Na Mg Ca and +. CH3SOwas     in
by a great diversity of landforms, glaciers, snowfields,            all samplescollectedafter December 15, 1989. Details of
valleys,lakes,rivers, moraines, plains,andmountains.The                          and quality-assurance
                                                                    the analytical                             are
                                                                                                     procedures reported
water bodies are of glacial origin and some lakes are still                                               3-
                                                                    by Likenset al. [1987], exceptfor CH3SO whichwas
connected glaciers. Mountainswith elevationsof 3,050                measuredusing proceduressimilar to thoseof Bardwell et
m dominatethe landscape. The impactof the Pacific Ocean                                     of         (nss) =,
                                                                    al. [1990]. Concentrationsnon-sea-salt SO4 CI-,
on precipitation TdP is diminished the AndesMoun-
                in                    by                                                            as
                                                                    Mg++, Ca++, andK+ werecalculated described by
tains, which are between the Pacific Ocean and TdP,                     et         using + asthereference
                                                                    Keene al. [1986]   Na                     and
creatinga continentalmicroclimate[Santanaet al., 1992].             Wilson's [1975] compositionof surface seawater. Nine
Solarradiationvariesseasonallydrivingstrongvariabilityin            samples were analyzedat the Institutefor EcosystemStudies
associatedmeteorological conditions. Average monthly                for orthophosphate   using the phosphomolybdenum    blue
temperatures between 1979 and 1984 rangedfrom a mini-               methodon an Auto Analyzer II [Murphy and Riley, 1962].
mum of 1.9 Cø in July to a maximumof 12.1 Cø in                     Thesesamples  were collectedin August, 1990 (1); October,
February (Meteorological Data Report, Torres del Paine              1990 (1); November, 1990 (2); January, 1991 (1); March,
NationalPark). Precipitation fairly uniformlydistributed
                           is                                       1991 (2); and April, 1991 (2).
over the year.
                                                                    Representativeness Database
Precipitation Collection and Analysis
                                                                       During the study period, 52% of all precipitationwas
  Precipitationwas sampledby storm event with an Aero-              sampledfor chemical analysis(Figure 2). The collection
chem Metrics collector at two sites in the park between                        for
                                                                    efficiencies individualyears typicallyvaried from about
March 23, 1984, and December 19, 1993. From March 23,               35% to 70%.            Becauseprecipitation          varied
1984, to January30, 1985, precipitation              2
                                        was sampled km              greatly over the year and, in some years, seasonalsub-
north of the Administration Building. After a nearby                sampling was not representative,it was impossibleto
wildfire, precipitationwas sampledat a similar elevation    reliablyanalyze interannual variability. We interpretedonly
approximately km away where it remained
               2                           from February    the average compositions    and depositionsfor the entire
1, 1985, to July 19, 1987. On July 20, 1987, the collection sampling  periodand for eachmonth(i.e., compositing     data
stationwas movedback to the original site.                  for all Januarys,etc.) assuming   that precipitationsampled
   After collection, mL CHC13was added 250-mL during each period was a representativesubset of all
                   0.5                          to
aliquots lessfor smallvolumeevents) prevent
         (or                            to         biologi- precipitationthat fell during that period. During January,
cal activity[Keene al., 1983;Keeneand Galloway,1984]; February, and Octoberthe efficiencieswere 27 %, 27 %, and
sampleswere subsequently    stored on site in a cool, dark 35%, respectively,           of
                                                                               because problems      with siteoperations
location and periodically (within one month of collection) usuallyin the earlier yearsof the record. Sampling  efficien-
shipped air to the Universityof Virginia. Samples    were cies in the other monthsrangedfrom 42 % to 68%.
                                   GALLOWAY               ET AL.'         PRECIPITATION        AT A REMOTE     SITE IN CHILE                         6885

        9O                                                                                                                                    120


        60                                                                                                                                    80 •

   =                                                                                                                                          60•
        4o                                                                                                                                      -g
  a• 20

                       19•         1985             19•            1987        1988       1989     1990       1991      1992         1993

             Figure2. Annual     precipitation          at                                          31,
                                               amounts Torresdel PainefromMarch23, 1984,to December 1993,
             (left axis) for total amounts  recorded(grey columns),amountssampled(opencolumns),and collection-
             efficiencypercentage   (solid line, right axis).

Quality Assurance                                                                                                          species cause
                                                                                          (e.g., HCOO- and SO4=)' unmeasured     can
   Of the 234 precipitation samples collected, 198 met                                              ionic imbalances.Thereforewe did not useionic
                                                                                          balances to assessthe data quality, only to investigate
quality-assurance criteria. We excluded datafor 36 samples
from the analyses: Precipitationgaugedatawere missing   for                               patternsin unmeasured ions. Specifically,a plot of the two
                                                                                          functions, Eanion-Ecation and Eanion/Ecation, showed that
4 samples;  supporting descriptivedata (e.g., date collector
                                                                                                                      anions(Figure 3). The
                                                                                          most samplescontainunmeasured
installed) were missing for 9; insufficient volume was                                    median value of Eanion/Ecation   was 0.80     and the median
sampledto allow completechemicalanalysesof 7; gross
contaminants  (e.g., by bird feces) were presentin 10; the
                                                                                              of           was                 the
                                                                                          value Eanion-Ecation-3.8 •eq L-1. Thus median
collector malfunctioned for 3; and 3 were apparently
                                                                                                     of          anions
                                                                                          concentration unmeasured       was 3.8 t•eq L-1,
contaminatedNa+ or Mg+ +.
         by                                                                  to
                                                              corresponding 20% of total anions. Low concentrations
   In pastanalyses GPCP data,we excluded         samples that of several speciesintermittently
                                                                                                   probably   contributed
failed to meet criteria involvingionic balances(e.g., samples to this pool of unmeasuredanions. Likely candidates
with ratios of the sum of anions to sum of cations outside a  included carboxylic  anions,suchaspropionate,  oxalate,and
definedlimit were excluded). This procedurewas appropri- pyruvate,       andinorganic  anions,suchas nitrite,bicarbonate,
ate for samplesfrom remote marine regions becausethe and bisulfite among others.
               of               is           by
ionic strength precipitation dominated seasalt; thus
an ionic imbalanceprovidesinsightsinto the quality of the
                                                              Results and Discussion
analysis of sea-salt constituents. However, in remote
regions immediately                to
                           adjacent oceans,  suchasTdP, the                               we
                                                                 In the following sections report variability in concen-
overall ionic compositionof precipitationis significantly trations                             of
                                                                      andper-eventdepositions major chemical     constit-
lower and controlled to a greater degree by nss species uents in precipitationat TdP and assessthe controlling
                                                              biogeochemical  processes. We also comparethe TdP data
                                                                                          with measured   and modeled   conditions    at other locations
                                                                                          within the region and around the globe. Caution is war-
   = 1.0                                                                                                            due
                                                                                          rantedin suchcomparisons to regionaldifferences the    in
   •-    0.8                                                                              frequency,size, and natureof precipitation    events;associ-
                                                                                          atedwater-accumulation rates (i.e., dilutioneffects);and, for
   •     0.6                                                                              data from snow and ice samples,post-depositional     changes
   ..• 0.4                                      ß
                                                                                          in composition.Despitethesepotentialcomplications,      these
                                                     !                                    comparisonsdo provide useful diagnostic information
   •4 0.2                                                                                 concerning major biogeochemical                in
                                                                                                                             interactions the remote
             0                I            I               I              I
                 -30         -20          -10             0               10
                                            [xeqL' 1
                                                                                            Precipitationat TdP was slightly acidic and contained
Figure 3. Pattern of 22anion/Ecation
                                   versusEanion-Ecation                                   both sea-salt (ss) and non-sea-salt components, which
in precipitationsampledat Torres del Paine.                                                     34.0/•eq -1and
                                                                                          averaged      L             L             The
                                                                                                              22.4/•eq -1,respectively.
6886                       GALLOWAY         ET AL.'        PRECIPITATION      AT A REMOTE         SITE IN CHILE

Table 1. The Volume-Weighted-Average Concentrations                    Na +
of ChemicalSpecies Precipitationand Corresponding                                          of Na             L
                                                                         TheVWA concentration + was13.2/xeq -1,the
Annual Ratesof Wet Deposition(Basedon Annual                                deposition 99.0 eqha yr-1 (Table Na+
                                                                       annual       was         -1          1).
           of 0.75m yr-1)atTorres Paine
Precipitation                   del                                                   and
                                                                       concentrations per-eventdeposition      exhibiteda modest
                   Concentration,                     Deposition,      seasonalpattern(Figure 5a); concentrations  were somewhat
                         /•eq -1                         -1
                                                      eqha yr-1        lower duringaustralsummerand occasional     high concentra-
                Total           nss          Total             nss     tions and depositions  were measuredduring australspring
  H+             10.9           10.9             81.8         81.8                        are
                                                                       whenwind speeds highest         [Santana al., 1992].
  Na +           13.2             ---            99.0          ---        Sea salt was relatively lessimportantat TdP than at sites
  K+              0.4            0.2              3.0           1.4    closerto the oceanbut still more significantthan in regions
 Mg++             3.2              0.2           24.0           1.5          inland.Forexample, annual
                                                                       farther               the     VWA Na+ concen-
  Ca ++            1.1             0.5            8.3          3.8     tration Amsterdam
                                                                             at               Indian
                                                                                        Island,           is     L
                                                                                                    Ocean, 269/xeq -1
  NH4+            0.6              0.6            4.5          4.5     and      deposition
                                                                           annual                  -1        on
                                                                                        is 3000eqha yr-1(based annual
  NO3-            0.5              0.5            3.6          3.6
  C1-            17.0              1.6       128              12.0     VWA concentrations[Moody et al., 1991] and annual
   =              2.8              1.2           21.3          9.0                 amount[Miller et al., 1993]), aboutan orderof
     3-           0.19             0.19           1.5          1.5     magnitudegreater than at TdP (Table 1). In contrast,at
 HCOO-            4.9              4.9            ......               Katherine, Australia, a continental GPCP site about 200 km
 CH3COO-          0.5              0.5            ......                   from ocean, VWA Na+ concentration
                                                                               the   the                 is 3.3/xeq
                                                                           [Likens et al., 1987], a factor of 4 less than at TdP.
                   Concentration,                     Deposition,
                         t•mol -1                      ha -1
                                                     mol -1yr
                Total           nss          Total             nss     Cl'

 HCOO.            5.5              5.5           41.3         41.3                                         and
                                                                         AnnualandmonthlyVWA concentrations depositions
  CH3 t
    C(50           1.0             1.0            7.3          7.3                     the
                                                                       for C1- exceeded contributions
                                                                                                    expectedfrom seawater
 PO4              0.06           0.06             0.5          0.5              L              -1
                                                                       by 1.6/xeq -1 and12.0eqha yr-1 (Table andFigure
                                                                       5b),assuming Na+ wasanunbiased           species
   In table nss is non-sea-salt concentration.
                                                                       for sea salt [Keene et al., 1986]. Although typically
                                                                       referred to as nssCI-, seawaterwas probablythe sourceof
          was        by
nsscomponent dominated H + andHCOO, with                               most excessC1-. Poorly understoodchemicalprocesses
    contributionsSO4 CH3COO , and
smaller     by CI-, =,   t,NH44-                                       involving sea-salt aerosol in the marine boundarylayer
NO3-. Theannual                as
                 VWA H+ expressedpH was 4.96                                                     C1-
                                                                       (MBL) convertparticulate to one or more inorganic    C1
              L        1),   was
(H•: = 10.9/•eq -1,Table which within range gases
                                    the                 HC1,C12,
                                                including                 2
                                                               HOC1,andC1NO [e.g., Graedel
of valuesreportedfor precipitationin other remotemarine                and Keene, 1995]. The latter three gases are rapidly
(e.g., Amsterdam Island, 5.1; Moody et al. [1991]) and                 convertedto HC1 following photolysisand H-abstraction
continentalregions(e.g., Katherine, Australia, 4.7; Likens             reactionswith hydrocarbons. Assuming that the reacted
et al. [1987]). On the basis of ionic ratios, the maximum              aerosol                                      by
                                                                               and HC1 were efficientlyscavenged atmospheric
           of                     4,
contributions HCOOH, HC1, H2SO CH3COOH,and                             droplets,a shorteratmospheric                   dry deposi-
                                                                                                        lifetime against
HNO3 to precipitation
                    aciditywere45, 15, 11, 5, and5 %,                  tion for the aerosolrelative to HC1 would lead to an appar-
respectively. When we excludedan early period of record ent enrichment total (ss q- nss)C1-relativeto ss-Na in+
(March 23, 1984, to July 19, 1987) during which HCOOH     precipitation[e.g., Moody et al., 1991].
concentrations were significantly
                                lower (seecarboxylicacids                                and
                                                            The nss C1- concentrations depositions     were some-
sectionbelow), the maximumcontributions   changedslightly what more variable from month to month than were other
(51, 12, 8, 4, and 4%, respectively).                     species (Figures 5 and 6). This resultedin part from the
   Monthly                            ranged
            VWA H + concentrations from6.7/•eq relatively greater uncertaintiesin calculatednss C1- [e.g.,
             to          L
L-1 in June 19.7/•eq -1 in December              4).
                                          (Figure By Keene et al., 1986]. A generalpattern was evident, how-
subtracting concentrations HCOO- and CH3COO- ever, with higher concentrations
            the                 of                                                           and depositionstypical
from the concentrations measured H + we estimatethat      during the warmer months and lower values during the
         basis,        primarily stron•
ona monthly H+ contributed    by
mineral         from4.2 /•eq L-1 to 8.2/•eq L-I
(Figure 4). By comparison,the acidity from carbonicacid                         •'   20
            with      CO 2          L
in equilibrium atmospheric is2.5/•eq -1. The                                         16
strongseasonality both organicand mineral acidsat TdP
suggests  that seasonalvariability in ecosystem metabolism        * 12
and productivityled to distincttemporalpatternsin emis-
sionsof precursor  compounds   controllingacidityin precipi-                    • 8
tation. Seasonalvariability in oxidative capacityand in                                   I   i    t   I   t   I   I   I   I   I   I
other physical processes,such as wind velocity which                      J   F M A M J J A S O N D
partially controlssea-air exchange,may also have contrib-
uted to the observedpatterns. During the australwinter, Figure4. Monthly        volume-weighted-average H +(VWA)
concentrations of most anions from acids were lower but still
                                                                            for                               and
                                                              concentrations all acids(top line, opensquares) only
measurablesuggesting     nonbiogenicsources(local or re- strong    acids(middleline, closedsquares)  relativeto the pH
mote), residual local biogenic emissions,or long-distance of purewaterin equilibrium                      CO
                                                                                        with atmospheric 2 (bottom
transport  ofbiogenicmaterialfrom lower (warmer)latitudes. line, no squares).
                                               GALLOWAY                             ET AL.'                 PRECIPITATION                   AT A REMOTE                       SITE    IN CHILE                                              6887

                  a                                                                                                                           0.4                                                                                      16
       30                                                                                                                   1200
                                                                                                                                              0.3                                                                                      12 l:l
       20                                                                                                                   800
                                                                                                                                         • _•0.2                                                                                   8•
      lO [•
      0 ['•-] •
            ß                                                                                                       _
                                                                                                                                             o.o : ß ß ß .•.•.                •Z              BB
                                                                                                                                                                                               , ,                                 4'
                                                                                                                                                                                                                                  .o •
        4                                                                                                                    160

 • •, 3                                                                                                                      120•

  ••2   0                  ,       :       ß       :.       :       ß       ß       '       :-'.I 80        :               .0

                   d                                                                                                                                     d
                                                                                                                                               0.8                                                                                 35
                                                                                                                             80    •,
                                                                                                                             60    •           0.6
  •     2
                                                                                                                                               0.4                                                                                 20
                                                                                                                             4o '•
                                                                                                                                               0.2                                                                                 10

        o       ....

                       J       F       M       A








                                                                                                                                                                       M      A   M   J   J   A

Figure 5.                      Monthly volume-weighted-average
                                                             (VWA)                                                                       Figure 6.                                             (VWA)
                                                                                                                                                                 Monthly volume-weighted-average
concentrations     left  and                               (columns,left axes) and per-eventdeposition
                                   deposition concentrations
            (columns axes) per-event
       lines, a•es) (a)Na (b)nss (c) (vertical
(vertical right       of    •-,      CI-,                                            (b)       =
                                                     lines,rightaxes)of (a) CH3SO3-, nssSO4 , (c)
                                                  +                        del
HCOOt, (d)CH3COOatTorres Paine.Thevertical NH4 , and(d) NO3-at Torres Paine. vertical
      and           t     del                                                        The        lines
linesindicatepercentiles(10th at top, 50th at crossbar, 90th                                                                             indicatepercentiles(10th at top, 50th at crossbar, 90th at
at bottom)of per-eventdeposition.Nine samples    were taken                                                                              bottom) of per-event deposition. For all speciesexcept
in January,16 in February, 15 in March, 23 in April, 21 in                                                                                     the      of      are     as
                                                                                                                                         CH3SO3-, numbers samples thesame in Figure
May, 22 in June, 14 in July, 26 in August, 15 in September,                                                                              5:9 in January, 16 in February, 15 in March, 23 in April,
12 in October, 12 in November, and 13 in December.                                                                                       21 in May, 22 in June, 14 in July, 26 in August, 15 in
                                                                                                                                         September, 12 in October, 12 in November, and 13 in
                                                                                                                                                 For      6       weretaken January,
                                                                                                                                         December. CH3SO3-, samples        in
                                                                                                                                         15 in February, 12 in March, 16 in April, 9 in May, 7 in
winter. The seasonal
                   patternin nss-C1-
                                   was the opposite
                                                  of                                                                                     July, 19 in August, 7 in September,8 in October, 5 in
                  5a                   relationship November, and 10 in December.
thatof Na+ (Figures and5b) buttheinverse
may havebeenlargelyfortuitous.The VWA concentrations
               that                   to
for Na+ suggested seasaltwastransported thesite
more efficiently during the colder months (May through                                                                                     Like the situation on Amsterdam Island [Moody et al.,
October) than during the summer. In contrast, we expect                                                                                  1991] and at other marine locations [e.g., Graedel and
processes that dechlorinatesea-salt aerosol to be more                                                                                                                         of
                                                                                                                                         Keene, 1995], substantialconcentrations nss C1- suggest
important during the summer when concentrations     of                                                                                   thatHC1 may have significantly           to
                                                                                                                                                                       contributed the acidityof
reactant      such SO N gases, oxidants
       species,   as 2,      and      (e.g.,                             at
                                                            precipitation the site (Table 1). The ultimatesourceof
OH, peroxy radicals, etc.), were presumably higher. this acidity remainsan open question. Acid-displacement
Compared themoreremote
           to                  MBL, elevated concentrations reactionsinvolving      4
                                                                              H2SO andHNO3 [e.g., Brimblecombe
of such speciesin pollutedregionsare associated        with and Clegg, 1988] do not generateatmospheric      acidity.
greaterC1-deficitsrelativeto seasalt [e.g., Martenset al.,  Protonsare simply transferredfrom one classof mineral
1973; Keene et al., 1990]. Assumingthat dechlorination acids(H2SO and HNO3) to another           (HC1). If, however,
wasmoresignificant    duringwarmermonths,    higherconcen- photochemically  active C1 gases volatilize from sea-salt
trationsof volatile C1 and differentialatmospheric                 and
                                                  lifetimes aerosol produce                   in
                                                                              atomicchlorine significant   concentra-
againstdeposition   may have contributed the observed tions [e.g., Keene et al., 1990; Finlayson-Pitts,1993],
seasonal pattern in riss C1-.                               subsequent hydrogenabstraction         will
                                                                                          reactions generate   HC1.
6888                           GALLOWAY ET AL.'          PRECIPITATION AT A REMOTE SITE IN CHILE

       50                                                                                      of
                                                                                   composition incident precipitationduring the period of
 •     40                                      ß
                                                                                     Alternatively, changesin environmentalconditionsmay

                                           ß                                       have contributed to long-term temporal variability in
 E     30                                                ß

                                                                                       t.          above, precipitation
                                                                                   HCOO As mentioned    the          collector
                                                                                   was reinstailed at the original sampling site during July,
 •     20                                                        ß                 1987, after regrowthof nearbyvegetation thathad burnedin
                                    ß ß ß ß ß ß          e• mm                     a wildfire two years earlier. This move coincided with the
 L•    10                          ß    mm
                                       mm ß•
                                      .•. •.4:!. ;...! .                                   HCOO The fire mayhaveled to increases
                                                                                   increased   t.                              in
                   : ._
                •'__•_          •'•.•' ß '• • • ' • • •', •
                                                   local emissions of HCOOH or precursor hydrocarbons
            0             1•         2•0                           in
                                                   throughchanges flora (e.g., isoprene-emitting
                                                   3•0                       40•                 vegetation),
                 CumulativeJulian Day              fauna (e.g., formicine ants; although according to TdP
                                                   personnel, there were no large ant populations),or soil
             t             in individual
Figure7. HCOO concentrations          precipitationmicrobes. As discussed below, however, it is likely, given
eventsat Torres del Paine for cumulative                                           its close        to         thata portion HCOO
                                                                                            proximity the ocean,            of   t
January 1, 1984, to December 19, 1993.                                             deposited at the site originated from marine sources.
                                                                                   Therefore local influences such as fires probably did not
                                                                                   completelyaccountfor the observedtemporalpattern. We
Chemicalprocesses                 aerosolcouldthen be
                 involving sea-salt                                                also used standardlinear regressionanalysisto assess  the
an important source for atmosphericacidity in marine
                                                                                   relationship              in    t
                                                                                                   variability HCOO andtheintensity
regions.                                                                           of E1 Nifio periods(G. Bell, personalcommunication,1994)
                                                                                   but found no significantrelationship. On the basis of the
Carboxylic Acids                                                                   above analysis,we were not able to ascertainthe causeof
       t         an
  HCOO exhibited unusual  temporal
                                 pattern                    temporally
                                        overthe thelong-term,        variability HCOO
                                                                               in      t.
                                                  VWA concentrations annual
courseof the study: Significantlylower concentrations                                     of
                                                                               depositions HCOOt
        depositions HCOO weremeasured samples and
per-event        of     t             in                   t were
                                                    CH3COO 5.5 •molL-1 and1.0•molL-1 and
collected before July, 1987, relative to those collected                                 ha           7.3 ha                 (Table
                                                                                   41.3mol -• yr-• and mol -• yr-•, respectively
thereafter(Figure 7).                                                                    and                  of
                                This abrupt and puzzling change 1). The VWA concentrations per-eventdepositions
suggestedone of two possibilities: Either (1) the chemical                             t         t                highest
                                                                                   HCOO and CH3COO variedseasonally;    values
composition precipitation
             of             changed   significantly       were duringspring, summer,and autumnwhen temperature
studyperiod or (2) systematic  negative(beforeJuly, 1987) and solar insolationwere higher (Figures5c and 5d). This
or positive(afterJuly, 1987) artifactssignificantly       strongseasonal
                                                        the               cycle is typical of carboxylicacidsat many
data. We examined the available information for evidence  remote locations [e.g., Keene and Galloway, 1988] and
of artifacts. All plasticware (sample bucketsand bottles) suggeststhat biogenic emissions, including both direct
used in the GPCP were purchased   from the samevendors emissions[e.g., Graedel and Eisner, 1988; Talbot et al.,
and washed,qualityassured,   packaged, and shipped the 1990; Sanhueza and Andreae, 1991] and emissionsof
University of Virginia using standard,testedprocedures precursor hydrocarbonsfrom both marine and terrestrial
[Galloway et al., 1982, 1993]. After receipt, all GPCP sources[e.g., Jacob and Wofsy, 1988; Gallowayet al.,
sampleswere analyzed in the central laboratoryduring 1989; Madronichand Calvert, 1990] may control atmo-
mixed analyticalruns (i.e., samplesfrom numerous  GPCP spheric   concentrations.   Seasonal             in
                                                                                             variability the produc-
sites were included) incorporatinginternal and external tion of photochemical                                  to
                                                                                 oxidantsmay also contribute the
quality-assurance        Despiteoverlapping
                                               in                                             in
                                                                                   seasonality concentrations.        the
                                                                                                              Although aqueous-phase
concentrations, data for other GPCP sites did not exhibit                          oxidationof HCHO in cloudshad been thoughtto be a
similar                 in    t
            variabilities HCOO overtheperiod
                                           in                                                important
                                                                                   potentially             of
                                                                                                     source tropospheric HCOOH [e.g.,
question. This suggests
                      that, if artifactsdid occur, they                                        and
                                                                                   Chameides Davis, 1983; Jacob, 1986], recentmeasure-
were uniqueto the Chilean site and musthave resultedfrom                           mentssuggest    that this mechanism may be relativelyunim-
some on-site influence. However, available evidencebased                                                                         et
                                                                                   portant [e.g., Facchini et al., 1992; Sanhueza al., 1992;
on bothdetailed           of             and
                 queries siteoperators the operators'                              Keene et al., 1995a].
notes on forms returned with the samplesindicatedthat                                                 and                      at
                                                                                     VWA concentrations medianwet depositions TdP
standardprotocols were followed throughoutthe entire                               duringthe warm monthswere abouta factorof 2 greater
period operation.    Although                  and
                               cleanplasticware samples                            than those observed at Amsterdam Island in the southern
were storedat differentlocations TdP duringthe period,                             Indian Ocean [Moody et al., 1991] althoughwintertime
             in         did
the changes location not correlate      with the variability                       valuesweresimilar. Thusseasonal             in
                                                                                                                    excursions carboxylic
in the data. In addition,                exposed deionized                                at
                                                                                   species TdP were relativelymorepronounced.     Terrestrial
water blanksat storageand samplinglocations     and subse-                                                             higher concentrations
                                                                                   regionstypicallyexhibit substantially
quent analyses        no
               revealed significant
                                  HCOO contamina-of tropospheric
                                       t                      HCOOHandCH3COOH                to
                                                                                      relative marine
tion. Finally,               active
                   biologically   species + and regions
                                         (NH4                    greatersource
                                                        suggesting                      overcontinents
CH3COOt) notvarysignificantly         the
                                between earlyand [Keeneand Galloway, 1988]. DifferencesbetweenAmster-
           suggesting the samples
laterperiods        that        were adequately dam Islandand TdP may, therefore,indicatesignificant
preservedagainst microbial degradation. Therefore we                               contributions from terrestrial sources in southern Chile.
       that no obvioushandlingor analyticalartifacts Alternatively,
conclude                                                                 marinesources carboxylic
                                                                  stronger           for        acids
     the    change HCOOt; measured
caused abrupt    in     the     temporal upwind of TdP relative to thoseat AmsterdamIsland or less
          appeared reflectactual
variability      to                  in
                               changes thechemical efficientremovalof marine-derived         during
                         GALLOWAY       ET AL.'          PRECIPITATION        AT A REMOTE SITE IN CHILE                         6889

       5O                                                                                                       betweenHCOOH and
                                                                           factorsleadingto the high correlations
                                                                                           and         [e.g., Keeneand
                                                                           CH3COOHin aqueous vaporphases
   •] 40
                                                                           Galloway, 1988]. Since both acids appearto be controlled

   I• 30                                                                   by similarprocesses                     and
                                                                                                (biogenicemissions photochemical
                                                                           oxidationof precursors,direct biogenicemission,and short
   •   2o                                                                                               it
                                                                           lifetimesagainstdeposition), is not surprisingthat they are
                                                                           highly correlated. All these processes driven to greater
   • lO                                                                    or lesser extent by seasonal variability in atmospheric
                                                     !        I            physics (radiation, temperature, wind velocity, and
            0      1          2     3     4                                precipitation).
                     of     t     CH3COO in
Figure8. Concentrations HCOO versus    t
individual precipitationevents at Torres del Paine from                                      3-
                                                                              $04 = and CH3SO were the two major S species
March 23, 1984, to December 19, 1993, showing the                                  in                                   were
                                                                           measured TdP precipitation. VWA concentrations
reduced major axis regression(solid line).                                                                        and
                                                                           2.8 /•eqL-1 and0.19 /•eqL-1, respectively, annual
                                                                           deposition were            -1           -1
                                                                                             21.3eqha and1.5eqha respec-
                                                                                                    of              =
                                                                           tively. VWA concentrations ss and nss SO4 were
        over landmay alsohavecontributed the differ-
transport                              to                                                                       corresponding
                                                                           1.6/•eqL-1 and1.2 /•eqL-1, respectively'
ences between sites.                                                                rates              -1  9.0
                                                                           deposition were12.3eqha and eqha respec- -1,
  Concentrations carboxylic acids in ice from coastal tively (Table 1).
Antarctica[Legrandand Saigne, 1988] are 2 to 3 ordersof                   of      =
                                                            Concentrations nssSO4 in TdP precipitation  (Figure
magnitudelower than those at TdP. These results are 6b) were within the range of those at other high-latitude
consistentwith the lack of significantbiogenic sourceson areas in the southern hemisphere(Table 2, Figure 9).
the Antarctic continentand suggest                                         =              in
                                    that marine sourcesof However,nss SO4 concentrations six snowsamples
precursors upwindof eastern                                                  (60
                             Antarcticaare relativelylow from southPatagonia km north) (Table 2) are a factorof
compared other regions.
         to                                               2 lower than at TdP. These differences are probably
  Aqueous- vapor-phase
            and             concentrations HCOOH and associatedwith the closer proximity of TdP to marine
      are     highly
CH3COOH usually             suggesting both sources (CH3)2S below). In addition, Patagonian
                                                  of      (see                 the
                   correlated       that
compounds controlled similarprocesses
            are         by                [e.g., Keene                     samples         to
                                                                                  correspond winter snowfallduring periodsof
et al. , 1986; Keene and Galloway, 1988' Talbot et al. ,                   minimum                [Bates al., 1992]. Other
                                                                                   (CH3)2Semissions     et
1988, 1995]. A reduced major axis (RMA) regression                         factorspossiblycontributingto these regional differences
[Hirsch and Gilroy, 1984] for the TdP data yielded a slope                 include orographic influences on precipitation amount,
of 7.52 +_ 0.24, an intercept of-2.27 _+ 0.34, and a                                                                processes,
                                                                           variability in rain- versussnow-scavenging          and
correlation coefficient of 0.89 (Figure 8). A previously                   elevational gradients in precursor gas and aerosol
reportedslopeof 1.01 + 0.20 for the regression carbox-
                                                of                         concentrations.
            at                                             3-               has
                                                   CH3SO in precipitation been
ylic species TdP [Keeneand Galloway, 1986] corresponds                                     at
                                                                                    measured onlya few
                                            were high-latitude
to the early record when HCOOHt concentrations                sitesin the southernhemisphere. Concentra-
significantlylower (see above). Althoughabsolute concen- tions at TdP (Figure 6a; Table 1) were within the range of
trations HCOO and CH3COO in precipitation TdP measurementsfrom coastal Antarctica. Maupetit and
        of        t              t                in
were within the range of values at other remote and im- Delmas[1992] reportthat CH3SO concentrations 13      in
pacted sites, the slope for the TdP data was higher by snowsamples              in                  site
                                                                       collected 1984 at the coastal of Dumont
factorsof 2 to 5 [e.g., Keene and Galloway, 1986]. Since d'Urville range from below detectionto 44 ppb (0.46 /•eq
there would probablyhave been little spatialvariability in L-1)withan arithmetic       of
                                                                                 mean 6.5 ppb(0.068/•eqL-l).
the relative losses of the two acids through deposition, CH3SO in snowdeposited      duringJune-December,     1984,
sourcestrengths HCOOH upwindof TdP may have been on the coastat Mirny-Vostok ranged from about 3 ppb to
          higherthanthosefor CH3COOHrelativeto                                                                       et
                                                                           23ppb(0.03/•eqL-1 to 0.24 /•eqL-1) [Legrand al.,
                                                                            3-             in
many otherregionsof the Earth. There has beenrecurring 1992, Figure4]. CH3SO concentrations an ice core
          over the years concerning biogeochemical from
speculation                         the                    Dollerman      are                    et
                                                                     Island 1 to2/•eqL-1 [Mulvaneyal.,

                                     Composition Torresdel PaineCompared OtherHigh-Latitude
                Table 2. Precipitation         at                      to                 Sitesin the
                                         nss =                       NO3-               NH4+              Reference
                 Antarctica                                                                      Delmas [1996]
                  Median                      1.06                   0.66              0.06
                  Range                 0.39-3.94                 0.17-3.33         0.04-0.17
                  N                      34                         35                  4
                 Drake Passage                1.1                    0.6               2.6                et
                                                                                                 Pszenny al. [1989]
                 Amsterdam Island             3.4                    1.3               1.8       Moody et al. [ 1991]
                 Patagonia cap               0.6                     0.3               0.8       Aristarain and Delmas [1993]
                 Torres del Paine             1.2                    0.5               0.6       this paper

                      are/•eq -1.
                  Units     L
6890                                     GALLOWAY                            ET AL.'              PRECIPITATION      AT A REMOTE                   SITE IN CHILE

                                                                                                                  typically covary over marked seasonal cycles at
                                                                                                          lO      high-latitude, marine-influenced sites in the southern
   2.0                                                                                                            hemisphereincluding TdP (Figure 10a), Amsterdam Island
   1.6                                                                                                            (Figure 10b) [Nguyenet al., 1992], Cape Grim, Tasmania
   1.2                                                                                                            [Ayerset al., 1991], and Antarctica [Legrand et al., 1992;
                                                                                                                  Savoieet al., 1993]. Such relationships supportthe hypo-
                                                                                                                  thesis seasonallyvarying         of       (CH3)2S
                                                                                                                                           emissions biogenic
                                                                                                                                 oceans majorsources CH3SO3H
                                                                                                                  fromthe southern      are             for
   0.0           I       I       I       I       i       I       I       i        i       i       I
                                                                                                                          4              MBL. The equivalent
                                                                                                                  andH2SO in the overlying                  ratiofor
                                                                                                                                   of CH3SO tonss = atTdPduring
                                                                                                                  VWA concentrations        3-     SO4
                                                                                                                  the period when both specieswere measured was 0.17
                                                                                                                  (Figure 10c), within the range of 0.05-0.4 reported for
                                                                                                          8       coastalAntarctica [Berresheim,1987; Pszennyet al., 1989;
                                                                                                                  Legrandet al., 1992; Delmas, 1996] but substantiallyhigher
    1.2                                                                                                           than that for Amsterdam Island (0.04, calculatedfrom Table
    0.8                                                                                                           3). Monthly equivalentratios (based on VWA concentra-
                                                                                                          2 •     tions) at TdP were lower during the austral summer and
                                                                                                                  higher during the winter (Figure 10c); that is, relative
    0.0          I       I       I       i       I       I       I       I        I       I       I
                                                                                                                  seasonal          in        3-
                                                                                                                          variability CH3SO wasgreater     thanthatof nss
                                                                                                                      =        at
                                                                                                                  SO4 . Ratios Amsterdam     Islandexhibitedsomewhat less
                                                                                                                  seasonalvariability                          H
                                                                                                                                     (Figure10c). SinceCH3SO3 appears
                                                                                                                  to be moreefficientlyproduced  from(CH3)2S oxidationby
                                                                                                                  OH at colder temperatures  [e.g., Hynes et al., 1986], we
                                                                                                                  expected relatively more pronounced  seasonal excursion in
    0.6                                                                                                                   =
                                                                                                                  nssSO4 at TdP assuming                originated
                                                                                                                                             both species          exclu-
                                                                                                                  sivelyfrom (CH3)2S hadsimilaratmospheric     lifetimes.
    0.4                                                                                                           The greaterseasonal                     3-
                                                                                                                                       variabilityin CH3SO suggests   the
       0.2           _ •             :       :       =                                                            possibility a significant
                                                                                                                            of            anthropogenic            to
                                                                                                                                                       contribution nss
             J       F       M       A       M       J       J       A        S       O       N       D
Figure 9.                                             (VWA)
                         Monthly volume-weighted-average                                                                0.3
           of (a) SOn       +,   (c) 3-
concentrations nss = , (b)NH4 and NO at
Torres del Paine (open squares;left axes) and Amsterdam                                                               ;,0.2
                    right axes).
Island (solidsquares;                                                                                                •o.1
                                                                                                                        0.0       '       •    '     •       '       '       '       '       '       '       '

1992] (as reviewed by Delmas [1996]). For the Drake                                                                           b
Passage, Pszenny et al. [1989] report a range from 8
precipitation   of               L
           samples 0.02to 0.12txeq -1.                                                                               •.•
                                                                                                                                                                                                                         6 •
         sources precursors nssSO4 in precipita-
  Possible     of         for     =
tion at TdP include biogenic emissionsof reduced sulfur                                                              •0.2                                                                                                4,,=•
gases(primarily(CH3)2Sfromtheupwind      the
                                   ocean); long-
       transport combustion-derived2 fromremote,
distance        of                SO                                                                                 •0.0         ,       ,    ,     ,       ,       ,       ,       ,       ,       ,       ,           O•
                      sources'and volcanic emissions
primarily anthropogenic                            of
   2.      3-         to
SO CH3SO is thought originate   almostexclusively
fromtheoxidation biogenic
               of                   there
                        (CH3)2S.Since    wasno                                                                         0.4
                                                                                                                     '•" c
major volcanicactivity upwind of the site duringthe study                                                               0.3
period, volcanic emissionswere probably an insignificant
regional source of S (although we cannot discount the
possibilityof nonexplosiveventingof S). Thus the follow-
ing analysis will focus on assessingthe biogenic-versus-                                                                0.0                                                                                          ,
anthropogenic  sourcesof nss S in precipitationat TdP. At                                                                     J       F       M A        M       J       J       A       S       O       N       D
           of            is               reduced
the latitude TdP, (CH3)2S the mostimportant
S gas emittedto the atmosphereand marine ecosystems are Figure 10. Monthly volume-weighted-average     (VWA)
substantiallystronger sourcesthan terrestrial ecosystems concentrations (a) Torres del Paine and (b) Amsterdam
[Bateset al., 1992]. The oxidation (CH3)2Sby OH                                                                         of       3-                left
                                                                                                                  Island CH3SO (opensquares, axes)andnssSO4         =
                                                                                                                               rightaxes);(c) the CH3SO3-/nss = ratios
                                                                                                                  (solidsquares,                            SO4
produces           4
         both H2SO and CH3SO3H[Yin et al., 1990].      at Torres del Paine (open squares)and Amsterdamisland
Relative production appears to be strongly temperature (closedsquares). The AmsterdamIslanddataare for paired
dependent;           is          in
           CH3SO3H produced greater                at
                                         abundance measurements 198) from July 31, 1989, to March 28,
lower temperatures[Hyneset al., 1986].                 1994 (J. N. Galloway and W. C. Keene, University of
          3'            =
  CH3SO and nssSO4 relationships.Aqueous-        and Virginia, GlobalPrecipitation ChemistryProjectunpublished
              concentrationsCH3SO andnssSO4
                          of    3-        =                                                                       data, 1995).
                                 GALLOWAY                                            ET AL.'                      PRECIPITATION       AT A REMOTE              SITE IN CHILE                             6891

  =                at                     factor
SO4 in precipitation thesite. A complicating   may
involve differential seasonalvariabilities in atmospheric
lifetimes CH3SO andriss = . Relative riss = ,
CH3SO is associated coarseraerosol
                                   to     SO4
                                  (e.g., Pszenny
                                                                                                                                  .,5ø!• •o                                             ß
[1992] amongothers). Thus seasonal   variabilityin physical
processes,such as sea-salt aerosol production, associated
scavenging,and subsequent    dry deposition,could lead to
relative differencesin the seasonality removal rates for
                                                                                                                                      •               .      ' ... •.    ..         •       ß
the two species.                                                                                                                      • 10 ., :l
   C1 chemistry may also complicate interpretation of                                                                                              Ill ß
                                                                                                                                            0 '•[l•---II :: :l I I              :               '
                                                                                                                                                0.0          0.2              0.4           0.6
sourcesfor riss S in the MBL basedon presumedratios of
sulfonated                     of       (CH3)2S.
                fromtheoxidation biogenic                                                                                                                      CH3SO3', p•eqL'1
In contrastto OH, there is no evidencethat oxidationby
     C1     to            of
atomic leads the production CH3SO3H                              of HCOO versus
                                        et Figure12. Concentrations
                                   [Keene                              t      CH3SO in
al., 1996]. Thus latitudinalor seasonal
                                      variability in the                                                                     individual precipitationevents at Torres del Paine from
relativeimportance C1versus oxidation (CH3)2S December16, 1989, to December19, 1993, showing
                 of        OH          of                                                   the
maycontribute observed      in           CH3SO reduced
                     patterns tropospheric     3-     major axis regression(solid line).
andrissSO4 Sincethe spatial          distributions
of Cl-atom concentrations the MBL are very poorly
constrained [e.g., Graedel and Keene, 1995], it is impossi- at TdP using standard                 of
                                                                                 linear regression measured    concen-
ble, at present,to critically assess potentialimportance
                                   the                      trations versus an index of E1 Nifio intensity (G. Bell,
                                                            Climate Analysis Center, National Weather Service, NOAA,
                          in              S
of this chemicalprocess the tropospheric cycle.
                                                            unpublished                                        of
                                                                        data, 1994). Although concentrations both
       3-                             =
  CH3SO is enrichedrelative to nss SO4 in coastal
                                                               =         3-       to         duringE1
Antarctic ice during strong E1 Nifio Southern Oscillation nssSO4 and CH3SO appeared be greater
     suggesting increased
events        that      (CH3)2S                           (Figures11a and l lb), only nssSO4 was
                                       during Nifio periods
                               emissions                                                   =
                                                           statisticallysignificant. In addition, the ratios of the two
E1 Nifio periodsmay contributeto interannualvariability in
absoluteand relative concentrations sulfonated  products species    showedno apparenttrend (Figure 1l c). A similar
                                                           analysisof precipitationamountat TdP versusthe index of
[Legrand and Feniet-Saigne, 1991]. We investigatedthe
relationshipbetweenE1 Nifio and S species precipitation E1 Nifio intensity revealed no significant relationship.
                                                           Although our results are not completelyconsistent       with
                                                                                                                                        in     3-            =           in
                                                                                                                             variability CH3SO andnssSO4 measured Antarctic
                                                                                                                             ice [Legrandand Feniet-Saigne,1991], they do supportthe
                                                                                                                             hypothesis fluxesof (CH3)2Sfrom the southern     ocean
                                                                                                                             to the atmosphere increaseduring strongE1 Nifio periods.
                                                                                                                             Such variability may have important implications for the
                                                                                                                             direct       and     indirect   effects    of    S aerosol     on the   Earth's
                                                                                                                             radiative balance and climate [e.g., Charlsonet al., 1992].
                                                                                                                             Spatial heterogeneityin ecosystems,their responses E1 to
                                                                                                                             Nifio, atmospheric oxidation processes, or atmospheric
                                                                                                                             lifetimes may contribute to regional variability in tropo-
                                                                                                                                   CH3SO andnssSO4 relative E1Nifio events.
                                                                                                                             spheric    3-       =         to
    .,   0.5                                      ß                          ß
                                                                                                                             Apossibleexplanationfor the lack of consistency between
    • 0.4                                                               ß            ß                                       our data and those of Legrand and Feniet-Saigne[1991] is
                                                               ß        ß

    _-L 0.3
                                                       ß   ß                              ß                                  that our data spannedonly 10 years whereastheir analysis
                                                                                     ii       ß

                                                  i II                                    ß                                               to
                                                                                                                             corresponds a longer period (1922 to 1984).
                    ß                                                                                                                            3'
                                                                                                                               HCOOt and CH3SO relationships. HCOO and  t
    •    0ol
                             ß                                               ß
                                          ß       Ii                ß                     ß       ß

    L)     0        i             i

                                                                                                              ß                   3-
                                                                                                                             CH3SO in precipitation TDP were also significantly
                                                                                                                             correlated(Figure 12). A RMA regression    yielded a slope
                                                                                                                             of 53.1 _+ 3.4, an interceptof-2.35 _+0.79, and a correla-
                                                                                                                                             of                a
                                                                                                                             tion coefficient 0.70 suggesting possible  direct chemical
         1.0                                                                                                                 link betweenthe cycling of S and HCOOH. The oxidation
         0.8                                                                                                                 of (CH3)2S       bothHCHO andCH3SO [Yinet al.,
                                                                                                                                       produces                3-
                                                                                                                                                 vapor- or aqueous-phase
                                                                                                                             1990]' the subsequent                      oxidationof
                                                                                                                             HCHO generatesHCOOH [e.g., Jacob, 1986]. However,
         0.4                          ß                                          ß        ß
                                                                                                                             much of the HCHO in the rural and remote troposphereis
         0.2                                  .:•.:..:.                                   ß                                        to        fromthe oxidation CH4 [e.g., Logan
                                                                                                                             thought originate               of
           o        I             I ß                  ' lal                              I           '   I                  et al., 1981' Duce et al., 1983] and isoprene[e.g., Shepson
               -4   -3           -2                             -1                        0               1
                                                                                                                             et al., 1991; Munger et al., 1995]; thus, the relatively minor
                        EI Nino (increasingstrengthto left)
                                                       contributionfrom (CH3)2Soxidation    wouldnotbe expected
                          betweenthe E1 Nifio indexand to result in strong correlations
Figure 11. The relationship                                                              betweenCH3SO and   3-
themonthly  volume-weighted-average
                                 (VWA) concentrations  HCOOH via this pathwayalone. The major precursorsfor
of (a) riss =, (b) CH3SO3-, (c) CH3SO3-/nss = HCOOH in marine air are probably other hydrocarbons
           SOn               and                 SOn
ratio.                                                 emittedfrom the oceansurface. It is likely that the observed
6892                    GALLOWAY    ET AL.'   PRECIPITATION      AT A REMOTE SITE IN CHILE

          betweenCH3SO and HCOO reflectsthe Nitrogen
correlation          3-        t
common influence of physical processes on essentially
independent biogenicsources,the subsequent           +                       of NH4
                                                NH4 . TheVWA concentration + in TdPwas
                                                      L       1).
                                              0.6 tteq -1(Table This
oxidation of precursor compounds,and deposition. Like                     is       than
                                                                      value greater therange
           3-                                          in         (0.04          L
                       fromthephoto-oxidation observedAntarctica to0.17tteq -1)butless
HCOOH, CH3SO is produced                   of                                           than
a biogenic precursor and is lost from the atmospherevia that for other high-latitude,non-Antarcticsites(Table 2).
deposition the surface.
          to                                               A pronounced           cycle       +
                                                                         seasonal of NH4 concentrations        and
  Sulfur budget. Estimating the annual budget for nss per-eventdepositions maximaduringspringand fall and
sulfurrequires information emissions (CH3)2S,wet- a minimumduringwinter (Figure 6c). The seasonal
                           on           of                                                                 pattern
deposition rates of CH3SO and nss SO4 and dry-
                            3-              =,          at AmsterdamIsland [Moody et al., 1991] also peaksduring
deposition                       2,
           ratesof CH3SO3-,SO and nss SO4      =. The springbut absolutevalues are substantially  higher, suggest-
annualwet-deposition  ratesof CH•SO¾ and nss SO4 at ing greater sources in that region (Figure 9b). Strong
                                        seasonal   in                are        with
TdP 1.5 -• ha yr and 1 e•ha yr respectively patterns NH4+ wetdeposition consistent
  were1 ec[ -1 -1 9.0 1 TM -1,
(1.5 mol ha- yr- and 4.5 mol ha- yr- , respectively) (Table   biologicalsources. Both terrestrialand marine ecosystems
1).                                                                           and      1994]. Given similar-
                                                              emitNH3 [Denlener Crulzen,           the
   Estimates of wet-depositionrates for nss sulfur in polar   ity of natural emissionsfrom the marine and terrestrial
               rangefromabout molha yr-1 to systems mol -1yr-1and mol -1yr-1,respectively)
                                                     (15 ha             36     ha
                             1      -1
       -1             1985];
30molha yr-1 [Galloway,               rates
                           wet-deposition at [after Dentener and Crutzen, 1994], we expect that both
                                                                        to        +
TdP fell within this range. On the basisof eight precipita- contributed NH4 in TdP precipitation.          Agricultural
tion samplescollectedin the austral fall on a cruise in the materials (fertilizer, animal waste) can also be important
                                                                   of                   el
Drake Passage, Pszenny el al. [1989] estimate a wet- sources NH3 [Galloway al., 1995]. Theprimary                 ter-
          rate 2 1 3-
               for          and SO4              ttmol   -2
deposition CH3SO nss = of0.26 rn restrial ecosystemat TdP is grassland,some of which is
d- and 1.7 ttmol m- d- , respectively. These scaleup to used raise       sheep  whose  waste      NH
                                                                                            emits 3 [Schlesinger and
          -1           6.3      -1
0.9 molha yr-1and molha yr-1,respectively,          which Hartley, 1992].
are similar to the TdP data. Pszenny el al. [1989] and        The                      of
                                                                  wet-deposition NH4+ TdP 4.5eqhaat     was         -1
Galloway [1985] estimatethat dry depositionof nss sulfur yr (Table                     and
                                                                        1). Dentener Crutzen               use
                                                                                                    [1994] a three-
(SO +nss SO4 is aboutone quarterand lessthanone dimensional, chemical-transportmodel to calculate global
     2           =)
                                                                  of                and
half of the wet-deposition of nss SO4 respectively. distributions NHx emissions depositions.For the
                          flux            =,
Assuming dry-deposition                             =
                             ratesfor SO2 andnssSO4 at southwestcoast of Chile, they estimate emissionsto be
                                                                   20 to             -2 -1
TdP were onethird of the nssSO4 wet-deposition we approximately 50 mgN m yr (14 to 36 eqha
                                 =              rate,                                                      -1
       a               -1
calculatefluxof 1.5molha yr-1. Total                      estimated
                                          dry) yr-1). Their
                                   (wetplus                                 rates
                                                                 wet-deposition areof thesame
deposition nssSO4 was thusestimated be 6.0 mol magnitude,or about 3 to 8 times higher than our measure-
         of     =                 to
 -1            el                 that
ha yr-1. Pszenny al. [1989]estimate the dry- ments. There are several possibleexplanationsfor these
deposition of CH3SO is about thirdof wetdepo- differences' Dentenet and Crutzen's model may overesti-
           rate          3-         one
sition, which, at TdP, would have yieldeda dry deposition mateNH3 emissions' x might
                                                                          NH                      and
           -1             the    depositionnss
of 0.5 molha yr-1. Therefore total       of                           upwind of TdP' or there might have been substan-
     to          8     -1 -1,       3-
sulfur TdPwasabout molha yr withCH3SO tial dry deposition NH3. Langford al. [1992]notea
                                                        of             el
        about25% of the total with SO and nssSO4
providing                            2         =              rapiduptakeof emittedNH3 within the grasslandcanopy
providingthe balance.                                         andthus                    to
                                                                     littlenetNH3 emission theatmosphere abovethe
            [1987]estimates (CH3)9S 1
                          that  2
                                           from canopy. Such rapid recycling could possiblyaccountfor
the SouthernOcean average 4.9 ttmol m- d- over a year.        some apparent differences between the measured and
Bales al. [1992]estimate
     el                 (CH3)2S emissions marine modeled fluxes. In another study, Quinn el al. [1988]
ecosystems the same latitude range as Torres del Paine estimate NH 3 flux from the northeastern
                                                              an                              PacificOcean
       2.21 ttmolS m d-1 in winter 4.96ttmol during
areabout            -2           and              spring 7/•molm d-1 or about molha yr-1.
                                                       of       -2          25    -1
S m-2 d-1 in summer. The annualmarine fluxesof Berre- This flux is also substantially
                                                                                    higherthan our measurements.
      and       et           18        -1
sheim Bates al. areabout molha yr-1and13mol However, spatialand temporalvariabilitiesin emissions           and
  -1                  or      a        of
ha yr-i, respectively, about factor 2 greater     than associateduncertaintiesin extrapolationscompromisethe
ourestimated           flux           -1
              deposition of 8 molha yr-1. However, reliability of direct comparisons   between Quinn's estimates
TdP is not surroundedby ocean; it is 50 km inland and and our measurements.
separated from the oceanby a significantmountainchain.                                     of
                                                         NO3- The VWA concentration NO3- at TdP was
                                                              'L -1       1),
Thus we expectS depositionat TdP was influenced both 0.5/mq (Tableslightly thanthe
                                               by                                     higher at Patagonian
marine and terrestrial emissionsand by S depositionin the ice cap(0.3 tmqL-1, Table2) [Aristarain Delmas,
overland fetch. Bales el al. [1992] estimate S emissions 1993]. NO3- concentrationsothersouthern
                                                                                      at              high-latitude
from terrestrial ecosystems the same latitudeas TdP are sites fall within the range of Delmas' [1996] estimatesof
0.02ttmol rn d-1and
        S -2       0.19ttmol rn d in winter
                           S -2 -1                    •eq       a     of                to
                                          and 0.17-3.33 L-1,with median 0.66•eqL-1similar
summer, respectively. The annual rates would be about                                                  at
                                                              thatat TdP (Table2). NO3- in precipitation TdP (Figure
           1    1
0.4 mol ha- yr-, significantlyless than our estimated  S      9c) and at AmsterdamIsland [Moody el al., 1991] exhibit
depositionrates at TdP. Our estimatedtotal deposition of                    patterns. However, long-distance
                                                              similarseasonal                              transport
nss S at TdP falls between the estimatedbiogenic S emis-                 NO              Africa to Amsterdam
                                                              of pollutant 3- fromsouthern                 Island
sions in upwind marine and terrestrial ecosystems;   the                to
                                                              contributes the springtimepeak and larger annualVWA
consistency theseestimates
           of               providesindependentcorrobo-         3-        of 1.3   L         and
                                                              NO concentration /•eq -1[Galloway Gaudry,
ration of their reliability.                                  1984; Moody et al., 1991].
                     GALLOWAY ET AL.' PRECIPITATION AT A REMOTE SITE IN CHILE                                     6893

  The seasonalityin VWA concentrationand per-event                          of
                                                             equal. Because the diverseand poorly quantifiedsources
deposition NO3- at TdP are differentfrom thoseof
         of                                                  of oxidizedN species,it was difficult to compareour results
  +,   t,     t,   =
NH4 HCOO CH3COOmsSO4 andCH3SO3-. 3- to estimated
                              NO                     as              SO4       +
                                              emissions we didfor nss =, andNH4 .
peaks during australwinter throughlate spring, the others                          provideda usefulconstraint the
                                                             However, theseestimates                        for
peak during the austral summer (Figures 5 and 6). The        combinedsourcesof oxidized N upwind of TdP.
        pattern NO3- suggests
different     of                 controls.Several
processes (fossil-fuelcombustion, biomass
                                        burning,tramport     Phosphorus
from the stratosphere, lightning)couldhavebeenrespon-                              of
                                                               The VWA concentration the nine samplesanalyzedfor
sible. Given the long fetch over the Pacific and modest
upwindemissions    from fossil-fuelcombustion, is unlikely
                                                                       (PO4) 0.06/•mol-1(Table with
                                                             orthophosphatewas       L       1),   a
                                                                 of        L             L            data
                                                             range 0.02/•mol -1to0.10/•mol -1. Phosphorus
   the         transportanthropogenic was
that long-distance    of         NOy a                       for other regionsare limited. At Lago Yelcho near Puerto
               to            in            at
majorcontributor NO3- measured precipitation TdP.                                           of              and
                                                             Monte, Chile, the concentrations orthophosphate total
However, under conditionsof cold temperaturesand low                     in
                                                             phosphorus six precipitationsamples   rangedfrom below
                                              southern detection
[NO]/[NO2] ratios typical of the high-latitude,                    to        L          0.1
                                                              limits 0.08/•mol -1andfrom to0.2/•mol
Pacific during the austral winter and spring, peroxyacetyl L-1, respectively Soto,
                                                                          (D.              data,
                                                                                  unpublished Universidad
nitrate(PAN' a reservoir         for              in
                         species NOx produced more Austral de Chile, 1994). At Katherine, Australia, the VWA
pollutedregions)is sufficiently                      long
                               stableto be transported
                                                           concentration ten
                                                                      for with samples during
                                                                          P?4           collected 1990
distances[e.g., Crutzen, 1979; Singh and Hanst, 1981' was0.10 /•molL-, in a range 0.08/•molL-1 to
Singhet al., 1985]. PollutantN oxidestransported PAN                L              unpublished Institute
                                                           0.14/•mol -1 (K. Weathers,       data,          of
             havecontributed thehigherNO3- concen-
may,therefore,             to
                                                          EcosystemStudies, 1994).
trationsin precipitationand associated             fluxes
at TdP during this period. NO producedby regional
biomassburning was a possiblesource [Fishmanet al.,
1991] (H. Levy II, personalcommunication,1995). In the
winter when wind speedsdecrease,burning is prevalent        The atmosphereand biosphere are integrally linked.
acrosssouthernChile (D. Soto, personalcommunication, Ecological processesin terrestrial and aquatic systems
                the                of
1994)although concentrationsnssK+ (a tracerof control,to varyingdegrees,emissions nitrogenandsulfur
biomass                                                             to
         burning;Andreae[1983]) at TdP were low (annual compounds the atmosphere. For instance,seasonal
       K                                  no
VWA nss + was0.2 /•eqL-1, Table1) andshowed relationships our datasuggest the biosphere
                                                        in                that            exerted
      signal.Transport NOyfromthestratosphere
seasonal             of                      considerable                                     of
                                                         control over atmosphericconcentrations
may have been important. Mayewskiand Legrand [1990]                     =,       HCOO and CH3COO at
                                                             NH4+, nssSO4 CH3SO3-,   t,        t
     higher 3-concentrations
report    NO              at DomeC andVostok
                                           in             TdP. In addition,atmospheric          is
                                                                                       deposition a major source
Antarctica during the austral spring possibly becauseof of water, nutrients, and toxic substances for aquatic and
transportfrom the lower stratosphere     associated with terrestrialecosystems[e.g., Gorham, 1961' Likens et al.,
formation of polar stratosphericclouds and, therefore, 1977]. Atmosphericdepositionmay be an important N
connected with the Antarctic ozone hole. Long-range sourcefor freshwaterecosystems the TdP region. Preli-
transport NOx (and its reaction    products)produced   by minaryinformation          that
                                                                            suggests freshwater            in
                                                                                                   systems TdP
lightningin the low-latitudetroposphereand lower strato- are oligotrophicand perhapsN limited [Sotoet al., 1994].
sphereand NOx produced N20 oxidation
                            by                 within the For example, Lago Toro, the largest lake in TdP, has a
lowerstratosphere alsobe important      sources NO3- water volumeof 30.2 km3 with a theoretical
                                                for                                                 turnovertime
in precipitationin high-latituderegions of the southern ofapproximately   13years.Thedominant          are +
                                                                                                cations Ca +
hemisphere [Legrand and Kirchner, 1990]. However, and Na+' the dominant                     are     3-
                                                                                   anions HCO and SO4           =
Likens et al. [1987] report no correlation between the [Campos al., 1994a, b]. BetweenSeptember1988 and
frequency locallightning NO3- in rainfallat Kather- January1991, the nutrientconcentrations NO3- range
          of               and
                      variability NO3- overthe high- between tzmol-• and tzmol-• and
ine, Australia. Seasonal         in                       0.36   L      2.0  L          of 4
                                                                                    those PO
latitude southernoceans is caused by multiple factors of           0.06  L
                                                             between tzmol -1 and        L
                                                                                 0.66tzmol -1. TheN:Pratio
unknown magnitude (see also, Savoie et al. [1993]).          by weight for thesedissolved                  is
                                                                                         inorganicnutrients usually
Relative to other species(e.g., nss SO4 CH3SO3-,
                                       =,                    below 5, as are the TN:TP molar ratios. In contrast, the
HCOOt,CH3COOt),            processes
                  biological        appeared beless
                                           to                                         for          in
                                                             N:P ratiosin precipitation thosesamples which both
important sources NO3- in TdP precipitation.
                for                                          elementsare measuredis 12, with a range of 3-37. These
  Theannual  wet-deposition of NO3-at TdPwas3.6 eq
                         rate                                ratiosindicatea potentialnitrogenlimitationfor phytoplank-
 -1 -1         -1
ha yr (3.6 molha yr-1) (Table1). Thisflux was                ton growth in Lago Toro [Sotoet al., 1994].
somewhat  lower thanLevy et al.'s recentmodelestimate(H.                                                        in
                                                                Like many high-latitudeterrestrial ecosystems the
Levy II et al., NOAA's GeophysicalFluid Dynamics       southernhemisphere[McGuire et al., 1992], terrestrial
Laboratory, Princeton University, Global depositionand              in
                                                       ecosystems TdP are nutrientpoor and someare N limited
                                                       [Soto et al., 1994]. These systemsreceive nitrogen from
distributionof reactive nitrogen in the troposphere,manu-
     in            of         -1 -1.
script preparation)6 molha yr However,            when severalsources,includingatmospheric  deposition,nitrogen
we accountedfor the differencesin precipitationamounts fixation, mineralization, and migrating animals. Net
     et                    we     75           the
(Levy al.use100cmyr-1; used cmyr-1), Levy primary production(NPP) by terrestrialecosystems the            at
             dropped 4.5 molha yr similar our latitude TdPrequires
et al. estimate    to        -1 -1,     to           of   2  1
                                                                 360-2,600 ha yr-1ofnitrogen
                                                                         mol -1
measured                             a
         value. Levy et al. estimate total NO,• deposition (0.5-3.7 g N m- yr- [Melillo et al., 1993]). Since the last
                    1                         •
of 14.5 mol ha-1 yr-; wet and dry fluxesare approximately glaciationabout 10,000 years ago, these ecosystems    have

                       Species Precipitation Remote
       Table 3. Chemical     in            at                         Sites
                         Torres del       AmsterdamIsland,      Katherine,        Barbados,        Lijiang,
                        Paine, Chile        IndianOcean          Australia       WestIndies         China

                                              average concentration,-1
                                    Volume-weighted (VWA)         L
        H + , total        10.9                   8.0             17.1              62               9.6
        H+, mineral         5.5                   4.5              5.4              23               5.9
        pH, mineral         5.26                  5.35             5.27             5 64             5.23
        K+                  0.2a                  0.4a             0.8b             0 3a             0.3b
        Mg++                0.2
                              a                    1.9
                                                     a             1.0
                                                                     b              23 a             0.4 b
        Ca++                0.5a                  1.3a             1.5b             4 5a             2.0b
                                                                                    28               4.0
        NH4+                0.6                   2.4              2.8
                                                                                    29               1.5
        NO3-                0.5                   1.0              3.6
        C1-                 1.6a                  7.3a             6.1b             5.4 a            1.9b
         4-                 1.2
                              a                    3.2
                                                     a             3.0
                                                                     b              4.0 a            5.7b
        CH3SO               0.19                  0.13             ---

        HCOO-               4.9                    3.0             9.6              2.7              2.9
                                                                                    1.2              0.8
        CH3COO-             0.5                   0.5              2.1
                                    Volume-weighted-average       i•mol
        HCOO.               5.5                    3.2            10.5              2.8              3.1
        CH3C(50             1.0                    0.8             4.2              1.8               1.4
                    of K          =,       Ca          del   are   this  for
         Concentrations+, Mg++, $O4 CI-,and ++ forTorres Paine from study; Amsterdam
       Island,IndianOcean,from GPCP unpublished (J. N. Galloway                                   of
                                                                        andW. C. Keene,University Virginia, 1995);
                                                                      from J. N. Gallowayand W. C. Keene (University
       for Katherine,Australia, from Likenset al. [1987]; for Barbados,
       of Virginia,unpublished                                               et
                               data, 1995);andfor Lijiang,ChinafromKeene al. [1995b].
         aNon-sea-salt concentrations.
         bTotal concentrations.

receivedmost nitrogen from external sourceseither by (1)      to      because enhanced
                                                         begin increase         of        nitrateformation and
fixationof N2, (2) atmospheric         of
                              deposition nitrogen fixed discharge[Hedin et al., 1995]. Hence, as the terrestrial
                                                                  at                     N,
                                 (terrestrialand marine) ecosystems TdP receiveadditional mobilization N
by lightningor by other ecosystems                                                                        of
and then transportedto TdP via the atmosphere,or (3) to the downstream   aquaticecosystemsmay increase.
migratingorganisms. Over the centuries much external
N has accumulatedin the TdP ecosystems      that internal
cycling (N mineralization)has becomethe primary source.                                 at
                                                                PrecipitationComposition TdP VersusOther
                                          boreal forests
For exampleN mineralizationin high-latitude
and         supplies 640to 1900 ha yr-1 (9- Regions
    grasslands     from           mol -1
         -1            et           which in the
27 kg N ha yr-1 [McGuire al., 1992]),   is       Precipitationcompositionin remote regions provides
same range as required for NPP. Current rates of N                                                       of
                                                                informationaboutthe likely composition precipitation  in
deposition  (Table 1) are relativelyminor nutrientsources
                                                        for     the northern hemispherebefore the industrial revolution.
biota when comparedto the N suppliedby mineralization.                         it
                                                                Consequently, can be used as a baselineagainstwhich
   Galloway et al [1994] estimatethat in 2020 NO depo-          precipitation              in
                                                                             composition more populatedareas can be
                   ß                                y
sition in this region will increaseto about40 mol ha-lyr-1      assessed [Galloway et al., 1982; 1984; 1987]. The concen-
      to       15                      exists. trations nssSO4 , NO3-,andNH4+ in precipitation
compared the --- molha-•yr-1thatcurrently             of     =                              at
Assumingfertilizer use and other agriculturalactivitiesin  TdP were within the rangesof those at other high-latitude
upwind regionsresultsin an equivalentincreasein NH x       sitesin the southernhemisphere (Table 2) but were substan-
          total N deposition the next few decades
deposition,                 in                        will tially lower thanthoseat marine(Barbados, AtlanticOcean,
beontheorder 100Tg N yr-•. Given           theirN-limited and Amsterdam Island, Indian Ocean) and continental
status,future increasesin N depositionwill probably first (Lijiang, China, andKatherine,Australia)locations lower
result in additionalproductivity of freshwater ecosystems. latitudes(Table 3). Galloway et al. [1984, 1987] compare
For terrestrialecosystems, projectionof 2020 N deposi- precipitationcompositionin remote regions to that in
tion is well below the estimatedmineralizationsupply rate populatedregions of the developedand developingworld.
             mol -1
of 640-1900 ha yr-•. However,                 rates
                                      increased of N In all cases,                and             of
                                                                    concentrations depositions nssSO4 and  =
deposition may indirectly enhanceprojectedincreases  in NO3-aresignificantly         in
                                                                              greater thepopulated   regions.The
NPP from risingCO2 andperhaps     temperature.                                 with theserelationships.
                                              Melillo et TdP data are consistent
al. [1993] suggest the response NPP in high-latitude
                 that              of                      In contrast to the situation for strong mineral acids,
ecosystems suchincreases    will be dampened because                   and                     of
                                                     of concentrations per-eventdepositions carboxylicacids
N limitation. If this limitation is modified by increased fell within the rangesof valuesat otherremoteand impacted
atmospheric deposition, then NPP may increase,  which, in locations [e.g., Keeneand Galloway, 1988]. Suchrelation-
turn, could affect earth's albedo, water balance, etc. In  ships        the           that
                                                                support hypothesis carboxylic        species primarily
addition, as N depositionto old-growthforestsincreases originatefrom biogenicprocesses       that involve either direct
fromlessthan70 molha yr-• to levels          greater  than emissions emissions precursorhydrocarbons.Anthro-
                                                                     or           of
300molha yr-1, nitrogen       lossesfromwatersheds    may pogenic processesappear to be relatively unimportant
                             GALLOWAY       ET AL.:     PRECIPITATION      AT A REMOTE         SITE IN CHILE                      6895

sourcesfor HCOOH and CH3COOH in the global                                    and
                                                                        aerosol otheratmospheric              Nss  =
                                                                                                  constituents. SO4 and
troposphere.                                                                  3-
                                                                        CH3SO primarily          from           of
                                                                                        originated theoxidation (CH3)2S
   Any comparison of precipitation acidity in different                 from marinebiogenicsources.NO3- probablycamefrom
regions is complicatedbecause pH is not conservative;                   several sources,includinglocal biogenic emissions,light-
carboxylic acids and, to a minor extent, the bicarbonate                ning, transport of emissions from biomass or fossil-fuel
systemprovide finite buffering capacity. Thus the addition              combustion in lower latitudes, and transport from the
of pollutantmineral acids to precipitationdoes not corre-
                                                        The           of         =,
                                            stratosphere. concentrations nssSO4 NO3-, and
     to            change H + concentrations
spond a proportionate    in                 NH4 in precipitationTdPwerewithin range those
                                                +             at             the    of
[Keene et al., 1983]. In addition, many programsdo not                  at other high-latitudesites in the southernhemispherebut
adequatelypreserveprecipitation samples againstmicrobial                were substantiallylower than those at marine (Barbados,
degradation after collection. Thus biologically important               Atlantic Ocean, and Amsterdam Island, Indian Ocean) and
          (e.g.,      acids
constituents carboxylic           +)
                           and NH4 are often continental (Lijiang, China, and Katherine, Australia)
metabolized lost from untreated
            and                    samples  beforeanalysis,             locations at lower latitudes.
which causes  systematicchanges (typically increases) the
                                                    in                    In the next few decades,            composition TdP
                                                                                                  precipitation          at
original pH [e.g., Keene and Galloway, 1984]. Becauseof                 will most likely change becauseof internal and external
thesecomplications, caremustbe exercised  whencomparing                 factors, includingpopulationgrowth, and energy and food
pH measured differentregionsby differentprograms.                       production. IncreasedN depositionhas the potentialto
  Likenset al. [1987] infer a background from mineral
                                         pH                                    the           of
                                                                        increase productivity N-limited aquaticecosystems.
acids of approximately 5.08 based on the analysisof
unpreserved aliquots of precipitationfrom Katherine,
Australia. This approachmay slightly overestimate        the
                                                                          Acknowledgments. Research    performedin remote regions
actual contributionof mineral acids to free acidity for two                     the         of
                                                                        requires cooperation manypeople                    In
                                                                                                            andinstitutions. Chile,
         (1)                       of
reasons: residualconcentrations carboxylicacidsare                                  the       of
                                                                        we appreciate support CarlosWeber, Corporacion     Nacional
occasionally present suchsamples
                    in               [e.g., Gallowayet al.,             Forestal(CONAF), Santiago,and of the staff of CONAF in Punta
1982], (2) NH4 is consumed Keene al., 1983]
     and      +         [e.g., et                                       Arena. We thank Guillermo Santana, Director of TdP National
                                                                        Park, and Jovito Gonzales and Cristina Yanez who collected the
possibly           to                    untreated samples maintained equipment. special of apprecia-
        contributing free acidity. Because                 and           the          A       note
aliquots samplefrom TdP were not routinelyanalyzed tion to Doris Soto B. who providesa bridgebetweenscientists
and to eliminatethe potentialfor the above problems,we                         two                and              in
                                                                        speaking differentlanguages who participated helpful
       a                 for       the
adopted differentapproach estimating background discussions the influence atmospheric
                                                             on             of                    on
                                                                                         deposition aquatic
pH corresponding mineralacids.
                to                                            In
                                                 ecosystems. the UnitedStates,we gratefullyacknowledgeTom
  TheVWA concentrationH + contributed carboxylic Butler, Alex Pszenny,KathleenWeathers,andBruceWiersmawho
                       of            by
                                                                               in             and            the
                                                                        assisted establishing maintaining site. Connie       Knighting,
and     acids precipitation was10.9/xeq-1
   mineral  in           atTdP        L                                 John Maben, Judy Montag, and Cheryl Reinhartanalyzedthe
(pH = 4.96). To compare withunpreserved  samples of                              Ishi
                                                                        samples; Buffamaidedin datareduction. Financialsupport
            from pollutedregions(and ignoring minor
precipitation                                                           was providedby the U.S. National Oceanicand Atmospheric
influences of the carbonate system), we subtracted the                  Administration and the U.S. Departmentof State. We also
contributions dissociated
           of          HCOOH andCH3COOHto free                                    the           of
                                                                        appreciate assistance GerryBell andBobWoodin obtaining
                                                                        the Southern OscillationIndex data; the helpful discussionswith
       the      fraction H+ contributed mineral
acidity; residual      of            by                                 Rick Artz, Lars Hedin, Hiram Levy II, JohnMiller, Bill Schlesin-
    was       L
acids 5.5/xeq -1(pH= 5.26).Similar          for
                                  calculations                          ger, and Kathleen Weathers; and the constructivecommentsof two
data from other remote sites in both marine and continental             anonymousreviewers.
regionsyielded a tight pH range of 5.23 to 5.64 for the
VWA contributionsof mineral acids to free acidity (Table
3). This estimated range reflects upper limits for the
influence     of   mineral   acids   from   natural   sources   since   References
anthropogenic  acidsundoubtedly    contributeminor fractions Andreae, M. O., Soot carbon and excessfine potassium:Long-
of free acidity at all thesesites [e.g., Moody et al., 1991].  range transportof combustion-derivedaerosol, Science,220,
We infer from thesedata that, on a volume-weighted     basis,  1148-1151, 1983.
strong mineral acidity in precipitationof eastern North Aristarain, A. J., and R. J. Delmas, Firn core study from the
America [e.g., Galloway et al., 1984] has increasedby          southern         ice
                                                                        Patagonia cap, South America,J. GIacioI., 39, 249-
                                                               254, 1993.
factors of about 10 to 15 as a result of anthropogenic
                                                                        Ayers, G. P., J.P. Ivey, and R. W. Gillett, Coherencebetween
                                                                          seasonalcycles of dimethyl sulphide, methanesulphonate and
                                                                          sulphate marine air, Nature, 349, 404-406, 1991.
                                                                        Bardwell, C. A., J. R. Maben, J. A. Hurt, W. C. Keene, J. N.
                                                                           Galloway, J. F. Boatman, and D. Wellman, A techniqueusing
   Precipitation TdP was slighly acidic (pH 4.96) with a
                at                                                         high-flow, dichotomous filter packs for measuring major
                                                                           atmospheric chemical            GlobalBiogeochem.
                                                                                               constituents,                 Cycles,
dilute seawater component. The long fetch across the                       4, 151-163, 1990.
Pacific Ocean and the strong seasonality the concentra-                 Bates, T. S., B. K. Lamb, A. Guenther, J. Dignon, and R. E.
                 ratesof manyspecies
tionsanddeposition                 (i.e., HCOOt,                           Stoiber, Sulfur emissionsto the atmospherefrom natural
     t,  SO4       and   )        that
CH3COOnss =, CH3SO3-, NH4+ indicated                                       sources,J. Atmos. Chem., 14, 315-337, 1992.
               were the primarycontrols. HCOOt and
naturalprocesses                                                                  H., Biogenicsulfur emissions
                                                                        Berresheim,                          from the sub-Antarctic
                                                                          and Antarctic Oceans,J. Geophys.Res., 92, 13,245-13,262,
     t       originated a combinationdirect
CH3COOprobably        from         of                                      1987.
biogenic emissionsfrom terrestrial ecosystems and the Brimblecombe, and S. L. Clegg, The solubilityand behaviour
                       emittedby marine and terrestrial
oxidationof hydrocarbons                                 of acid gasesin the marine aerosol,J. Atmos. Chem., 7, 1-18,
biota. Nss C1-was produced reactions  involvingsea-salt  1988.
6896                    GALLOWAY       ET AL.:     PRECIPITATION       AT A REMOTE        SITE IN CHILE

Campos,H., D. Soto, W. Steffen, G. Aguero, O. Parra, and L.         Hedin, L. O., J. J. Armesto, and A. H. Johnson, Patterns of
  Zuniga, Limnological studies of Lake del Toro, (Chile):              nutrient loss from unpolluted,old-growth temperateforests:
  Morophometry,    physics,chemistry             Arch. Hydro-
                                     and plankton,                     Evaluationof biogeochemical  theory, Ecology, 76, 493-509,
  biol. Suppl., 99 (1/2), 199-215, 1994a.                              1995.
Campos,H. D. Soto, W. Steffen, O. Parra, G. Aguero, and L.                                                                      line
                                                                    Hirsch, R. M., and E. J. Gilroy, Methodsof fitting a straight
  Zuniga, Limnological studies of Lake Sarmiento (Chile): A            to data: Examplesin water resources,Water Resour.Bull., 20,
  subsaline from ChileanPatagonia,
            lake                       Arch. Hydrobiol.Suppl.,         705-711, 1984.
  99 (1/2), 217-234, 1994b.                                         Hynes, A. J., P. H. Wine, and D. H. Semmes, Kinetics and
Chameides,W. L., and D. D. Davis, Aqueousphasesourceof                           of
                                                                      mechanism OH reactionswith organic sulfides,J. Phys.
  formic acid in clouds, Nature, 304, 427-429, 1983.                   Chem., 90, 4148-4156, 1986.
Charlson, R. J., S. E. Schwartz, J. M. Hales, R. D. Cess, J. A.     Jacob,D. J., Chemistryof OH in remotecloudsand its role in the
  Coakley,Jr., J. E. Hansen,andD. J. Hoffman, Climateforcing                     of                                  J.
                                                                      production formic acid and peroxymonosulfate, Geophys.
  by anthropogenic aerosols,Science,255, 423-430, 1992.                Res., 91, 9807-9826, 1986.
Crutzen, J., The role of theNO andNO2 in thechemistry the
       P.                                               of          Jacob, D. J., and S.C. Wofsy, Photochemicalproductionof
  troposphere stratosphere,
             and              Annu. Rev. Earth Planet. Sci., 7,                                             atmosphere, Acid
                                                                      carboxylicacids in a remote continental         in
   443-472, 1979.                                                                at
                                                                      Deposition High ElevationSites,editedby M. H. Unsworth
Delmas,R., The acidityof polar precipitation,
                                            Chapter10 in Acid          and D. Fowler, Kluwer Acad., Norwell, Mass., 1988.
  DepositionAssessment  Report,editedby D. M. Welpdaleand          Keene,W. C., andJ. N. Galloway, Organicacidity in precipitation
  M. S. Kaiser, in press,WMO/UNEP, Geneva, 1996.                    of North America, Atmos. Environ., 18, 2491-2497, 1984.
                                                      model of
Dentener, F. J., and P. J. Crutzen, A three-dimensional           Keene, W. C., and J. N. Galloway, Considerations         regarding
   theglobalammonia    cycle,J. Atmos.Chem.,19, 331-369. 1994.      sourcesfor formic and acetic acids in the troposphere,        J.
Duce, R. A., et al., Organic material in the global troposphere,    Geophys.   Res., 91, 14,466-14,474, 1986.
   Rev. Geophys.,  21, 921-952, 1983.                             Keene, W. C., and J. N. Galloway, The biogeochemical     cyclingof
Facchini,M. C., et al., Phase-partitioning chemicalreactions                                         the              An
                                                                    formicandaceticacidsthrough troposphere: overviewof
   of low molecular weightorganic             in
                                    compounds fog, Tellus,44B,      currentunderstanding,    Tellus,40B, 322-334, 1988.
   533-544, 1992.                                                 Keene,W. C., J. N. Galloway, andJ. D. Holden Jr., Measurement
Finlayson-Pitts, J., Chlorineatomsas a potential
                B.                                 trophospheric    of weak organicacidityin precipitation   from remoteareasof the
   oxidant themarineboundary      layer,Res. Chem.Interreed.,19,    world, J. Geophys.   Res., 88, 5122-5130, 1983.
   235-249, 1993.                                                 Keene, W. C., A. A. P. Pszenny, J. N. Galloway, and M. E.
Fishman,J., K. Fakhruzzaman, Cros, and D. Nganga,Identifi-                                                             of
                                                                    Hawley, Sea-saltcorrectionsand interpretation constituent
   cation of widespreadpollution in the southernhemisphere                                         J.
                                                                    ratiosin marineprecipitation, Geophys.      Res., 91, 6647-6658,
   deduced from satelliteanalyses, Science,
                                          252, 1693-1696,1991.      1986.
Galloway, J. N., The deposition sulfur and nitrogenfrom the Keene, W. C., A. A. P. Pszenny,D. J. Jacob,R. A. Duce, J. N.
   remote atmosphere,background paper, in Biogeochemical                                            H.
                                                                    Galloway,J. J. Schultz-Tokos, Sievering,      andJ. F. Boatman,
   Cyclingof Sulfurand Nitrogenin theRemote   Atmosphere,  edited   The geochemical   cyclingof reactivechlorinethroughthe marine
   by J. N. Galloway, R. J. Charlson, M. O. Andreae, and H.         troposphere,  GlobalBiogeochem.     Cycles,4, 407-430, 1990.
   Rodhe,pp. 143-175, D. Reidel, Norwell, Mass., 1985.            Keene,W. C., B. W. Mosher, D. J. Jacob,J. W. Munger, R. W.
Galloway, J. N., and A. Gaudry, The composition precipitation
                                                 of                 Talbot, R. S. Artz, J. R. Maben, B. R. Daube, and J. N.
   on Amsterdam Island, Indian Ocean, Atmos. Environ., 18, 2649-                                          at
                                                                    Galloway,Carboxylicacidsin clouds a high elevation       forested
   2656, 1984.                                                      site in centralVirginia, USA, J. Geophys.      Res., 100, 9345-
Galloway, J. N., G. E. Likens, W. C. Keene, and J. M. Miller,       9357, 1995a.
   The composition precipitationin remoteareasof the world,
                    of                                            Keene, W. C., Liu Jiaqui, R. S. Artz, J. N. Galloway, J. M.
   J. Geophys. Res., 87, 8771-8786, 1982.                           Harris, J. L. Moody, andB. J. B. Stunder,Processes    controlling
Galloway,J. N., G. E. Likens, and M. E. Hawley, Acid precipita-     variability in the chemical compositionof precipitationat a
   tion: Natural versusanthropogenic   components, Science,226,     remote,temperate    site in south-central China, posterpresented
   829-830, 1984.                                                   at FourthScientific  AdvisoryCouncil, International   Geosphere-
Galloway, J. N., D. Zhao, J. Xiong, and G. E. Likens, Acid rain:    Biosphere  Programme,     Meeting, Beijing, October, 1995b.
  China, United States, and a remote area, Science, 236, 1559-     Keene, W. C., D. J. Jacob, and S.-M. Fan, Reactive chlorine: A
  1562, 1987.                                                   potential sink for dimenthylsulfideand hydrocarbonsin the
Galloway, J. N., W. C. Keene, R. S. Artz, J. M. Miller, T. M.   marineboundary   layer, Atmos.Environ., in press,1996.
  Church, and A. H. Knap, Processes  controllingthe concentra-Langford, A. O., F. C. Fehsenfeld,J. Zachariassen,  and A. S.
       of =,           NH4
  tions SO4 NO3-, +, H+, HCOO and         T, CH3COO      T in   Schimel, Gaseousammoniafluxes and background      concentra-
  precipitation Bermuda,Tellus, 4lB, 427-443, 1989.
               on                                                                              of
                                                                tions in terrestrial ecosystems the United States, Global
Galloway,J. N., D. L. Savoie,W. C. Keene,andJ. M. Prospero,     Biogeochem.  Cycles,6, 459-483, 1992.
                                                ratiosfor nss Legrand,M. R., and C. Feniet-Saigne,
  The temporaland spatialvariabilityof scavenging                                                   Methanesulfonic  acid in
                                  and        in
  sulfate,nitrate,methanesulfonate, sodium theatmosphere        South Polarsnowlayers:A recordof strong   E1Nino?, Geophys.
  over the North Atlantic Ocean, Atmos. Environ., 27A, 235-250,       Res. Lett., 18, 1991.
   1993.                                                                                                              of
                                                                   Legrand,M. R., and S. Kirchner,Originsandvariations nitrate
Galloway, J. N., H. Levy II, and P.S. Kasibhatla,Year 2020:                                    J.
                                                                     in SouthPole precipitation, Geophys. Res., 95, 3493-3507,
  Consequencesof population growth and development on            1990.
  deposition oxidizednitrogen,Ambio, 23, 120-123, 1994.
              of                                              Legrand, M., and C. Saigne, Formate, acetateand methanesul-
Galloway, J. N., W. H. Schlesinger, Levy II, A. Michaels,and
                                  H.                             fonate measurementsin Antarctic ice: Some geochemical
  J. L. Schnoor,Nitrogen fixation: Anthropogenic enhancement-    implications,Atmos.Environ., 22, 1011-1017, 1988.
  environmental   response,Global Biogeochem.Cycles, 9, 235-  Legrand, M. R., C. Feniet-Saigne, E. S. Saltzman, and C.
  252, 1995.                                                                                              of
                                                                 Germain,Spatialandtemporalvariations methanesulfonic      acid
Gorham, E., Factors influencingsupply of major ions to inland    and non-sea-sulfatein Antarctic ice. J. Atmos. Chem., 14, 245-
  waterswith specialreference the atmosphere,
                              to               Geol. Soc.Am.     260, 1992.
  Bull., 72, 795-840, 1961.                                   Likens, G. E., F. H. Borman, R. S. Pierce, J. S. Eaton, and
Graedel, T. E., and T. Eisner, Atmosphericformic acid from       N.M. Johnson,                     of
                                                                                 Biogeochemistry a ForestedEcosystem,       146
  formicine                              Tellus,40B, 335-339,
             ants:A preliminaryassessment,                       pp., Springer-Verlag, New York, 1977.
  1988.                                                       Likens,G. E., W. C. Keene,J. M. Miller, and J. N. Galloway,
Graedel, T. E., and W. C. Keene, The tropospheric   budgetof     Chemistry of precipitationfrom a remote, terrestrial site in
  reactivechlorine,GlobalBiogeochem.   Cycles,9, 47-77, 1995.    Australia,J. Geophys.  Res., 92, 13,299-13,314, 1987.
                       GALLOWAY        ET AL.:    PRECIPITATION       AT A REMOTE         SITE IN CHILE                        6897

Logan,J. A., M. J. Prather,S.C. Wofsy, andM. B. McElroy,              phaseformic and acetic acids at a tropical cloud forest site,
  Atmospheric           A
              chemistry: global           J.       Res.,
                                perspective, Geophys.                Atmos. Environ., 26A, 1421-1426, 1992.
  81, 7210-7254, 1981.                                             Santana,A., H. Fuenzalida, and P. Aceituno, El inestableclima de
                                          reactions organic
Madronich,S., andJ. G. Calvert, Permutation       of                                        para el futuro, paperpresented A
                                                                      TdP hoy y un pronostico                             at
  peroxy       in             J.       Res.,95, 5697-
        radicals thetroposphere, Geophys.                             PatagoniaGem--TheEcologyand Natural History of a World
  5715, 1990.                                                         BiosphereReserve:                                 Chile,
                                                                                       TdP NationalPark, Chile, Santiago,
Martens, C. S., J. J. Wesolowski,R. C. Harriss, and R. Kaifer,        Jan. 10-12, 1992.
  Chlorine loss from Puerto Rican and San FranciscoBay area        Savoie,D. L., J. M. Prospero,R. J. Larsen, Huang F., M. A.
  marineaerosols, Geophys.    Res., 78, 8778-8792, 1973.             Izaguirre, Huang T., T. H. Snowdon,L. Custals,and C. G.
        F.,                      composition falling
Maupetit, andR. J. Delmas,Chemical         of                        Sanderson,                          in
                                                                                 Nitrogenand sulfurspecies Antarctica       at
  snowat Dumont D'Urville, Antactica, J. Atmos. Chem., 14, 31-       Mawson, Palmer Station,and Marsh (King GeorgeIsland), J.
  42, 1992.                                                           Atmos. Chem., 17, 95-122, 1993.
Mayewski, A., andM. R. Legrand,     Recent         in
                                           increase nitrate Schlesinger,
                                                                       W. H., and A. E. Hartley, A global budget for
  concentration Antarctic snow, Nature, 346, 258-260, 1990.              NH                 15,
                                                              atmospheric 3, Biogeochemistry, 191-211,1992.
McGuire, A.D., J. M. Melillo, L. A. Joyce,D. W. Kicklighter, Shepson,P. B., D. R. Hastie, H. I. Schiff, M. Polizzi, J. W.
   A. L. Grace, B. Moore III, and C. J. Vorosmarty,Interactions    Bottenheim,K. Anlauf, G. E. Mackey, and D. R. Karecki,
  between   carbon andnitrogendynamics estimating primary
                                         in          net           Atmospheric                 and
                                                                                concentrations temporal     variations C1-C3in
                for                    in
  productivity potentialvegetation North America, Global            carbonyl           at
                                                                            compounds two rural sitesin centralOntario,Atmos.
  Biogeochem.    Cycles,6, 101-124, 1992.                          Environ., 25, 2001-2015, 1991.
Melillo, J. M., A.D. McGuire, D. W. Kicklighter,B. Moore III,    Singh,H. B., and P. L. Hanst, Peroxyacetyl  nitrate(PAN) in the
   C. J. Vorosmarty,andA. L. Schloss,   Globalclimatechange and    unpollutedatmosphere:    An importantreservoir for nitrogen
   terrestrial primaryproduction,
             net                    Nature, 363, 234-240, 1993.    oxides, Geophys. Res. Lett., 8, 941-944, 1981.
Miller, J. M., J. L. Moody, J. M. Harris, and A. Gaudry, A 10- Singh,H. B., et al., Relationship  betweenperoxyacetly  nitrateand
   year trajectoryflow climatology Amsterdam       Island, 1980-   nitrogen oxidesin the cleantroposphere, Nature, 318, 347-349,
   1989, Atmos. Environ., 27A, 1909-1916, 1993.                       1985.

Moody, J. L., A. A. P. Pszenny,A. Gaudry, W. C. Keene, J. N.         Soto,D., H. Campos,W. Steffan,O. Parra, and L. Zuniga, The
  Galloway, and G. Polian, Precipitationcompositionand its             Torres del Paine lake district (Chilean Patagonia):A case of
  variabil)ty the southern
              in              IndianOcean:     Amsterdam     Island,   potentiallyN-limited lakesand ponds.Arch. Hydrobiol. Suppl.,
  1980-1987,J. Geophys.    Res., 96, 20, 769-20,786, 1991.             99, 181-197, 1994.
Mulvaney, R., E. C. Pasteur,D. A. Peel, E. S. Saltzman,and Talbot, R. W., K. M. Beecher, R. C. Harriss, and W. R. Cofer
  P. Y. Whung, The ratio of MSA to non-sea-saltsulphatein                                              of
                                                                       III, Atmosphericgeochemistry formic and acetic acids at a
  Antarctica Peninsula ice cores, Tellus, 44, 295-303, 1992.            midlatitudetemperatesite, J. Geophys.Res., 93, 1638-1652,
Munger, J. W., D. J. Jacob,B. C. Daube, L. W. Horowitz, W. C.           1988.
  Keene, and B. G. Heikes, Formaldehyde,glyoxal, and methyl- Talbot, R. W., M. O. Andreae, H. Berresheim, D. J. Jacob, and
  glyoxalin air and cloudwater a rural mountain      site in central    K. M. Beecher, Sources and sinks for formic, acetic, and
  Virginia, J. Geophys. Res., 100, 9325-9333, 1995.                     pyruvic acids over Central Amazonia, 2, Wet season, J.
Murphy, J., andJ.P. Riley, A modifiedsinglesolution     methodfor       Geophys.  Res., 95, 16,799-16,811, 1990.
                    of             in
  the determination phosphate natural waters, Anal. Chim.            Talbot, R. W., B. W. Mosher, B. G. Heikes, D. J. Jacob, J. W.
  Acta, 27, 30, 1962.                                                   Munger, B. C. Daube, W. C. Keene, J. R. Maben, and R. S.
Nguyen, B.C., N. Mihalopoulos,J.P. Putaud, A. Gaudry, L.                                                            atmosphere the
                                                                        Artz, Carboxylicacidsin the rural continental          of
  Gallet, W. C. Keene, and J. N. Galloway, Covariationsin               EasternUnited Statesduring SCAPE, J. Geophys.Res., 100,
  oceanicdimethyl sulfide, its oxidationproductsand rain acidity        9335-9343, 1995.
  at Amsterdam Island in the southern Indian Ocean, J. Atmos.                                                         of
                                                                     Wilson, T. R. S., Salinity and the major elements seawater, in
  Chem., 15, 39-53, 1992.                                               ChemicalOceanography,     editedby J.P. Riley and G. Skirrow,
Pszenny,A. A. P., Particle size distributions methanesulfonate          pp. 365-413, Academic, San Diego, Calif., 1975.
  in the tropicalPacificmarineboundary    layer, J. Atmos.Chem., Yin, F.-D., D. Grosjean, and J. H. Seinfeld, Photooxidationof
  14, 273-284, 1992.                                                    dimethylsulfideand dimethyldisulfide,I, Mechanism    develop-
Pszenny,A. A. P., A. J. Castelie, J. N. Galloway, and R. A.             ment, J. Atmos. Chem., 11, 309-364, 1990.
  Duce, A study of the sulfur cycle in the Antarctic marine
  boundary                  Res., 94, 9818-9830, 1989.
           layer, J. Geophys.                                                                                     of
                                                                      J. N. Gallowayand W. C. Keene, Department Environmental
Quinn, P. K., R. J. Charlson, and T. S. Bates, Simultaneous        Sciences,Clark Hall, University of Virginia, Charlottesville,VA
  observations ammoniain the atmosphere
              of                          and ocean,Nature,        22903. (email:;
  335, 336-338, 1988.                                                 G. E. Likens, Institute of EcosystemStudies, Box AB, Mill-
Sanhueza, E., and M. O. Andreae, Emission of formic and acetic     brook, NY 12545. (email:
   acidsfrom tropicalsavanna            Res.Lett., 18, 1707-
  1710, 1991.                                                                                      10, 1995;
                                                          (ReceivedJanuary31, 1995; revisedSeptember
Sanhueza,E., M. Santana,and M. Hermoso, Gas- and aqueous- acceptedSeptember22, 1995.)

Shared By:
xiaohuicaicai xiaohuicaicai