~PUMENTS Of W.H.O.I. Sf:A·GRANT PRQ~ Limnol. Oceanogr., 25(6), 1980, 1044-1053 © 1980, by the American Society of Limnology and , Inc. Volatile organic compounds in seawater from the Peru regionL2

Philip Gschwend,3 Oliver C. Zafiriou, and Robert B. Gagosian Department of Chemistry, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543

Abstract The concentrations and distributions of individual volatile organic compounds in seawater were determined at six stations in the upwelling region off the of Peru by vapor phase stripping, glass capillary gas chromatography, and gas chromatography-mass spectrometry techniques. Saturated and unsaturated hydrocarbons, alkylated benzenes, and aldehydes were found. Individual compound concentrations varied from <2 ng· kg-' to about 60 ng· kg-'. All of the volatile compounds were present at higher concentrations in surface than in deeper , but each class of compounds had a unique horizontal distribution. Spatial distributions and correlations with hydrographic data suggest possible sources and transformations.

Several groups of specific volatile organ­ has provided information concerning ic compounds have been identified and some sources and transport processes of quantified in coastal seawater samples. dissolved organic matter in seawater Schwarzenbach et al. (1978) recovered (Sauer et al. 1978; Sauer 1978; Schwar­ alkanes, alkylbenzenes, naphthalene and zenbach et al. 1978; Gschwend 1979). alkylnaphthalenes, aldehydes, and di­ Further work on these volatile com­ methylpolysulfides from seawater col­ pounds may extend our insights and in­ lected from Vineyard Sound, Massachu­ dicate environmental transformations setts. Sauer et al. (1978) determined characteristic of their organic functional individual Cu-C14 alkanes, benzene, and groups. alkylbenzenes in coastal waters of the We report here a study of volatile or­ Gulf of Mexico. These workers used mild ganic compounds in seawater from the purging conditions to obtain their vola­ Peru upwelling region. The intense bio­ tile concentrates and focused on the oc­ logical productivity associated with up­ currence of individual specific organic welling provides an opportunity to focus compounds, the sum of which amounted on the distribution, sources, and sinks of to only 0.05-1.5 11-g C · kg-1 (Sauer 1978; biogenic volatile organic compounds. Gschwend 1979). Somewhat harsher Terrestrial and anthropogenic inputs of purging methods recovered about 50 11-g organic materials from the sparsely pop­ C · kg-1 from coastal seawater acquired off ulated, arid Peruvian coast should be Nova Scotia, but the composition of this minimal. We have analyzed a series of volatile extract is unknown (MacKinnon seawater samples from profiles along an . 1979). Although these volatile isolates offshore transect taken in March 1978 (a amount to a very small proportion of the period of active upwelling) using vapor total organic in seawater, study of phase stripping, glass capillary gas chro­ the spatial and temporal distributions of matography, and gas chromatography­ the specific volatile organic compounds mass spectrometry techniques for the "volatile organic compounds." These are operationally defined as those coin­ pounds which may be purged from ­ 1 This research was supported by NSF grants OCE 74-22781, OCE 77-26084 and by Department by bubbling with a gas and of Commerce, NOAA, Office of Sea Grant, 04-7-158- retained by a solid adsorbent (Schwar­ 44104. zenbach et al. 1978; Gschwend 1979). We 2 Woods Hole Oceanographic Institution Contri­ report the compounds identified, their bution 4482. distribution in the , and cor­ 3 Current address: Bldg. 48-420, Department of Civil Engineering, Massachusetts Institute of Tech~ relations with hydrographic, nutrient, nology, Cambridge 02139. and pigment data. 1044 Volatile organics in seawater 1045

4°C in the dark for 3-6 weeks before anal­ ysis at Woods Hole. Samples for , , nutrients, and chlorophyll a were drawn from the I If same Niskin bottle and analyzed on ­ board. Salinity was determined with a Guildline Autosal salinometer (model 8400), oxygen by a modified Winkler pro­ cedure (Carpenter 1965). Nutrients were measured with an AutoAnalyzer, and chlorophyll a was determined by the flu­ orometric technique (Yentsch and Men­ zel 1963; Holm-Hansen et al. 1965). The seawater samples were analyzed ~5 km by the "Grob" method (Grob 1973; Grob and Zurcher 1976; Schwarzenbach et al. 16°5 .______..______J 1978; Cschwend 1979). Just before anal­ 7~W 7~W ysis, 200 ml of water were poured off at Fig. l. Station locations, RV Knorr Cruise 73-2 a nearby isolated beach to provide clean­ (March 1978). air headspace; Schwarzenbach et al. (1978) showed that this procedure elim­ inated erratic airborne contamination of M. True and J. Alford collected the samples by lab air. The samples were samples. We are also indebted to Z. stripped for 2 h at 35°C by repeatedly cy­ Mlodzinska and T. Loder for nutrient cling the headspace air through the sam­ data, J. Kogelschatz for pigment analyses, ple and a charcoal microtrap to-collect the and G. Nigrelli for salinity measure­ volatile compounds. After the water was ments. D. Muller provided a sample bf stripped, a standard, 20 ng of 1-chloro-n­ authentic fucoserraten. N. M. Frew as­ octane in 2 ~1 of cs2' was added to the sisted in GC-MS analyses. We thank S. trap, which was then extracted with 15 ~1

Henrichs, C. Lee, S. Wakeham, and J. of CS2 • Aliquots of this extract were ana­ Sharp for comments on the manuscript. lyzed by glass capillary gas chromatog­ raphy on a model 2101 Carlo Erba gas Sampling and analysis chromatograph equipped with a flame Seawater samples were collected with ionization detector, or by combined gas a single 30-liter Niskin bottle with Tef­ chromatography-mass spectrometry (GC­ lon-coated spring at six stations in March MS) with a Finnigan 3200 quadrupole 1978 on Cruise 73, Leg 2 of RV Knorr mass spectrometer. A single 20-m x 0.32- (Fig. 1). Immediately after retrieval, sam­ mm-i.d. SE-54 column (Jaeggi, Switzer­ ples were drawn from the bottle at the land) was used throughout. Electron im­ bow, heading into the wind, to min­ pact (EI) spectra were acquired at 70 eV imize contamination from the ship. Water ionization potential, chemical

flowed by gravity from the bottle through ionization (CH4-CI) spectra at 130 eV and a polypropylene tube to an inline 142- a source pressure of 950 ~· mm glass-fiber filter (previously cleaned As a check for possible contamination by baking at 450°C overnight) in a stain­ effects of filtering and poisoning, some less steel holder; these components were samples were not treated; others were rinsed with a portion of the the sample. poisoned only. Untreated samples were Glass reagent bottles (2 liters) with drawn first, followed by unfiltered sam­ ground-glass stoppers were then filled ples, and finally by filtered ones. Com­ completely, poisoned with 0.5 ml of sat­ parison of treated with untreated samples

urated HgCl2 solution (67 ~M Hg in the showed that filtering and poisoning pro­ samples), stoppered tightly, and stored at cedures did not contaminate the samples.

------~ ·, 1046 Gschwend et al.

0 IOkm '------' absolute quantities recoverable may STATION NUMBER change during sample storage, the rela­ 6 5 4 3 2 tive distributions within each compound should not be influenced. The analyses provided a detection lim­ 1 it of about 2 ng· kg- ; Schwarzenbach et al. (1978) found reproducibility to be about ±10%; blanks were <2 ng·kg-1 • Nonpolar and slightly polar compounds were recovered with >80% efficiency;

0.80 8.28 0.95 420 6-v 1!.09 relatively polar compounds were re­ 10 10 11.34 16.44 13.12 13.47 10~~ covered at lower efficiencies, e.g. alde­ 5.52. J•.·< hydes =30%. Reported concentrations 100 14.53 19.23 ..-·:><:·.".'· ... 20 ...... are corrected for stripping efficiency and 40.51 storage effects by using results from mod­ ~ 1000 ·b. NITRATE i!: el compounds. Results reported here are ~ for filtered, poisoned samples. <::> 6.04 4.69 6.12 5.0:; 1.57 3/"' 10 4 3.80 3.02 3.28 2.22 _./1.99 1.37 Results 2 0. 100- 0.29 The hydrographic data showed active upwelling between stations 1 and 4 dur­ 1000 1.96 c. OXYGEN ing sampling (Fig. 2a). Nutrients were enriched at station 1 and dropped to rel­ ative minima in the surface seawater at 11.2 15.6 11.4 15~211.9 1.4 stations 4 and 6 (as shown for nitrate in 10 -----5----_ "' on. 2.1 ~2 10.4 1.9. Fig. 2b). An oxygen-deficient subsurface 0.3 0.1 0.2 100 0.08 o.s ...... layer was encountered ·at all statio~s at 0.07 about 100 m (Fig. 2c), reflecting the .... :.:.:':":"":. 1000 0.01 O.D2 d.: CHLOROPHYLL a strong remineralization activity which occurs at that depth. Chlorophyll a con­ Fig. 2. Sections showing sigma (J (%o), nitrate centrations were particularly high (> 10 1), 1), (MM · kg- oxygen (ml· kg- and chlorophyll a 1 1 ) (Mg· kg- ) in seawater from Peru upwelling region. J-tg· kg- in the surface waters offshore of Depth scale is logarithmic. station 2 (Fig. 2d), indicating intense bi­ ological productivity at these sites during sampling. A bloom was present in the region of stations 2 and 3 and a di­ Coastal seawater samples were fil­ noflagellate bloom at stations 4 and 5 (J. tered, poisoned, and spiked with model Hobbie pers. comm.). volatile ·compounds ( n-tridecane; 2,6-di­ The seawater samples contained rep­ methyl-3-heptene; 1-ethyl-2-methylben­ resentatives of several classes of volatile zene; 2-ethylhexanal) at 25 ng· kg-1 and organic compounds, including saturated stored for 4 weeks to evaluate sample and unsaturated hydrocarbons, alkylated storage effects. Stored samples showed benzenes, and aldehydes (Fig. 3_ shows reproducible (±8%) changes in volatile representative chromatograms). In gen­ concentrations from unstored replicates. eral, concentrations for all volatiles were All three hydrocarbons were recovered at higher in surface samples than in deepero 30-40% lower concentrations from the ones; however each class had a unique stored samples than from the unstored horizontal distribution. Normal penta­ ones. The aldehyde, however, was re­ decane was the only prominent n-alkane; covered at 50% greater efficiency after its C 13-C16 homologues were only found storage. The volatile compound varia­ at trace levels. Surface samples contained tions we discuss here are much greater the highest concentrations of pentadec­ ~han these storage effects. Although the ane, with a maximum at station 6 of more Volatile organics. in seawater. 1047 r-1 I

Q) L____JOkm Stn. 3, 5 meters ®® STATION NUMBER 6 5 4 3 ~ ~ j 57 18 25 1l 23 i@l 10 \_, 100

1000 <2 6

6 ~ 6. 9 6 10 10 · Stn. 3, 20 meters J----5 ~5' ' ' 4 100 4 I CD ~

(j4) --. 1000 3 ~ ® 12 ~ !!; <:::> 19 <2 ® 10 @ 3

100 <2

1000 2

Stn. 3. 100 meters 19 12 15 11 4 CD 10 . ~------10------100 5

® Qg @; 1000 3

Fig. 4. Sections showing pentadecane, 1,3- + 1,4-dimethylbenzenes, unknown (mol wt 108), and @: heptanal concentrations (all in ng · kg-1) in seawater from Peru upwelling region. Depth scale is loga­ rithmic.

20 20 50 80 110 1 0 than 50 ng· kg- (Fig. 4). CcC3 alkylated Temp ( () benzenes were found, but only the C2-al­ Fig. 3. Chromatograms showing volatile organic kylated benzenes were sufficiently abun­ compounds isolated from three depths at station 3. dant and free from blank problems to pro­ Compounds identified are: !-toluene; 2--hexanal; 3---ethyl benzene + unknown (mol wt 108); 4- vide interpretable data (deep-water 1,3- + 1,4-dimethylbenzene; 5---1,2-dimethylben­ toluene concentrations were similar to zene; 6---heptanal; 7-C3-benzenes; 8--octanal; 9--­ shallow sample levels). These C 2-aromat­ l-chloro-n-octane, internal standard; 10--nonanal; ic hydrocarbons were present at < 10 11-decanal; 12--l-chloro-n-decane, internal stan­ 1 ng· kg- • Surface seawater contained 2-3 dard; 13---l-chloro-n-dodecane, internal standard; 14-pentadecane. times more of these compounds than 1,000-m saJOples (as shown for 1,3-di­ methylbenzene + 1,4-dimethylbenzene in Fig. 4). 1048 Gschwend et al.

A relatively abundant unidentified centrations covaried; hexanal and hep­ compound, which cochromatographed tanal were the most abundant. As in the with ethylbenzene, was found in several case of the other volatiles, concentrations surface samples. A molecular weight of were highest in offshore surface samples 108 was assigned to this compound based (shown for heptanal in Fig. 4d). on the presence of m/e 109 (M + 1), 137 (M + 29), and 149 (M + 41) in the Discussion ~~' CH4-CI spectrum and m/e 108 in the EI The organic carbon recovered in the spectrum. Losses of 13 and 27 mass units "analytical window" defined by our (giving m/e 95 and 81) in the CH4-CI methods amounted to <0.01% ofthe total spectrum indicate a terminal vinyl group organic carbon in the seawater. Total

(-CH=CH2) (Field 1968). The loss of 15 levels detected in surface samples, mass units (giving m/e 93) in both the EI where primary productivity was intense, 1 and CI spectra suggested a methyl sub­ were about 100 ng C·kg- • Deep water stituent. These data are consistent with samples contained 20 ng C · kg- 1 or less. an octatriene or possibly a methyl­ These total levels are approximately branched heptatriene structure. Authen­ equal to or less than the lowest volatile tic trans,cis-1,3,5-octatriene showed the fraction concentrations found in the Gulf same major fragment peaks as the un­ of Mexico by Sauer (1978) or in Vineyard known in the sample. However, trans,cis- Sound, Massachusetts, by Gschwend 1,3,5-octatriene did not cochromatograph (1979), both of whom used methodology with ethylbenzene; thus the unknown similar to that used here. The maximum was not this particular isomer. concentration observed for an individual The peak heights of this (mol wt 108) volatile organic compound was only 60 unknown were estimated by subtracting ng· kg-1• the ethylbenzene contribution from the One hypothesis to explain the gener­ mixed GC peak, assuming that ethylben­ ally low levels of volatile organic com­ zene peak heights were proportional to pounds in these waters is that there may the peak heights of other C 2-benzenes in be a lack of sources. The coastal region the same sample. This assumption is rea­ of Peru near the area where the samples sonable since ethylbenzene levels covary were collected is a sparsely populated with other C 2-benzene isomer concentra­ desert. Therefore, terrestrial biogenic tions in most fossil fuels and in many oth­ and anthropogenic sources are unlikely er seawater samples (Gschwend 1979). to be large, despite offshore winds. How­ These estimates by difference agreed ever, primary production in this upwell­ very well with peak heights inferred from ing region was very high at the time of GC-MS mass chromatography data. sampling, and total volatile concentra­ The derived spatial distribution of the tions were higher near blooms. Thus, unknown compound (mol wt 108) is biogenic production is implied, and shown in Fig. 4c. Maximum values were therefore sinks for these volatile organic always found near the surface. Station 3 constituents must act to limit their con­ showed the highest level, with concen­ centrations to parts per trillion (w/w) tration decreasing slowly offshore. levels. Mass chromatograms for m/e 109 in the Concentrations of volatile compounds GC-MS analyses of CI-CH4 showed only (especially biogenic compounds) may be one "octatriene" at station 3, but addi­ limited by rapid remineralization. Men­ tional isomers appeared in the surface zel (1975) suggested that dissolved or­ seawater at station 6. Mass spectra indi­ ganic carbon is rapidly decomposed after cated that the compounds are probably pulse inputs, since primary productivity cis-trans or positional olefinic isomers. correlated only poorly with dissolved or­ The homologous series of straight­ ganic carbon values. Also, the low con­ chain aldehydes from hexanal to decanal centrations of specific organic metabo­ was also found. Individual aldehyde con- lites imply that these compounds are Volatile organics in seawater 1049

quickly removed after release into sea­ water. Such atmospheric levels are very water. Turnover times of only a few days unlikely; e.g. Cautreels and van have been calculated for amino acids Cauwenberghe (1978) found only 60 ng (Williams and Gray 1970; Lee and Bada pentadecane · m-3 in the vapor phase of 1975, 1977) and glucose and acetate urban air. (Wright and Hobbie 1966) dissolved in In situ biological production is a prob­ natural waters and have been attributed able source of pentadecane. Many phy­ .\ to heterotrophic utilization. Concentra­ toplankters contain pentadecane (Blumer tions of biogenic volatile organic com­ et al. 1971). Clark and Blumer (1967) re­ pounds may also be limited by microbial ported between 300 and 600 ng penta­ degradation. decane · g-1 dry algae for three species. A second potentially important sink for To estimate the magnitude of this source volatile organic compounds is gas ex­ for the Peru upwelling region, we as­ change into the atmosphere. The mixed sume that average have layer of the open exchanges gases 500 ng pentadecane · g-1 dry weight and of low water with a residence that twice the surface particulate organic time of about 1 month (Broecker and carbon (POC) values approximate the dry Peng 1974). The shallow mixed layer off phytoplankton . For the Peru up­ Peru ( 10-20 m: Fig. 2a) should degas sev­ welling region, where values between eral times faster. Therefore, volatile or­ 500 and 2,300 J.Lg POC · kg-1 were found, ganic compounds produced by the highly these could ·contain about 1 ng active planktonic community of the up­ pentadecane · kg- 1 seawater. If we as­ welling region off Peru may be removed sume a 2-4 week degassing equilibration from the seawater into the atmosphere on time as a minimum sink, these hypothet­ a time scale of days. As a result, we can­ ical phytoplankters would have to pro­ not determine the relative importance of duce and release all of this estimated heterotrophic utilization and evaporation pentadecane stock into solution in sea­ as sinks for the volatile organic com­ water at a rate of once per day. Although pounds. this rate corresponds to algal doubling Pentadecane-The concentration of rates observed in the Peru upwelling re­ pentadecane in these surface waters (Fig. gion (Strickland et al. 1969; Beers et al. 4) approached its calculated solubility 1971; Walsh 1975), it seems unlikely that (based on an extrapolation of data from all of the pentadecane should be released McAuliffe 1966; Button 1976). It has also as dissolved organic matter. Thus, while repeatedly been found in coastal seawa­ phytoplankton may contribute some pen­ ter (Schwarzenbach et al. 1978), in Gulf tadecane to seawater, other biogenic of Mexico surface seawater (Sauer 1978), sources may also be important. and in Sargasso Sea and equatorial Atlan­ Another reasonable source of penta­ tic surface water (Gschwend 1979). decane might be the biochemical de­ Known sources for this hydrocarbon in­ carboxylation by zooplankton of phyto­ clude fossil fuels and terrestrial and ma­ plankton-produced hexadecanoic acid, 1 rine plants. The absence ( <2 ng · kg- ) of analogous to the pathway proposed for homologues-tetradecane and hexadec­ the formation of pristane (2,6,10,14-tetra­ ane-rules out a fossil fuel source. Land­ methylpentadecane) (Blumer et al. 1964; derived dust sampled over the ocean Avigan and Blumer 1968). Since phyto­ does not contain detectable amounts of plankton contain about 10 mg hexadeca­ pentadecane (Simoneit et al. 1977). The noic acid· g-1 dry weight (Ackman et al. 4 air-water partition coefficient (ca. 10 ) 1968; Chuecas and Riley 1969; Fisher ..' for this saturated hydrocarbon favors and Schwarzenbach 1978), the standing the vapor phase so strongly that ca. 100 stock of this proposed starting material mg·m-3 in air would be required to main­ could be up to 20 J.Lg· kg-1 seawater. Con­ tain the tens of nanograms per kilo­ version of 0.1-1% of this hexadecanoic gram concentrations observed in sea- acid would yield tens of nanograms of 1050 Gschwend et al. pentadecane per kilogram of seawater. tamination or simply the result of inputs Tetradecane and hexadecane would not from the research vessel. be found in similar abundance because Unknown (mol wt 108)-The surface phytoplankton contain very little of the maximum in the unidentified olefin con­ C 15 and C 17 fatty acid precursors. Conse­ centration at station 3 suggests a phyto­ quently, in addition to direct phytoplank­ plankton source, possibly associatedwith ton inputs, a zooplankton-mediated trans­ the diatom bloom. The discovery of what formation of hexadecanoic acid to seem to be isomers of the unidentified I pentadecane may contribute an impor­ olefin in the surface sample at stat; 6 tant fraction of the dissolved pentadec­ implies that the decline in concentration ane in seawater. between stations 3 and 6 is not due solely Alkylated benzenes-The CcC3 alky­ to dilution but also to isomerization re­ lated benzenes-common constituents of actions. petroleum and combustion exhausts­ This same unidentified compound was were recovered from the coastal waters found at high concentrations in Vineyard off Peru. The spatial distribution shown Sound seawater during late winter 1978 by 1,3- + 1,4-dimethylbenzene suggests (Gschwend 1979), but its source is not a surface water or atmospheric source known. Fucus spp., benthic brown algae, (but not general shipboard or analytical release an octatriene (trans,cis-1,3,5-oc­ 1 contamination above about 2 ng· kg- , the tatriene) into seawater as a sex attractant deep water levels). Since oxygen concen­ (Muller and Jaenicke 1973), but the mol trations in recently upwelled waters at wt 108 compound is not this triene. station 1 were quite low, the transfer of Knowledge of the detailed structure and gases from the atmosphere was far from the biological source of this occasionally complete. If the atmosphere were the abundant and geographically widespread source, dimethylbenzene air concentra­ compound would obviously permit very tions in this region would have to be sev­ interesting studies of the transport, trans­ eral times >2 ng·litec1 air [K(wt·vol-1 formation, and biological function of air)/(wt·vol-1 water)= 0.34], the value what may be a chemical messenger (as is calculated for equilibrium with observed the Fucus compound). C 2-benzene levels; this is greater than Aldehydes-The C 6-C10 straight chain the 3-5 ng·liter-1 air measured in rural aldehydes were found primarily in sur­ areas of Florida (Lonneman et al. 1978) face waters offshore of station 3 (Fig. 3d), and much greater than the 0.02 ng·litec1 indicating a biogenic source. Aldehydes air detected along the west coast of Ire­ have been found in other seawater sam­ land (Eichmann et al. 1979). It seems un­ ples (Sauer 1978; Schwarzenbach et al. likely that air at this remote Peruvian site 1978), in freshwater (Kikuchi et would be much more contaminated. al. 1974), and in a freshwater chrysophyte Direct surface water pollution of alkyl­ (Collins and Kalnins 1965). The hypoth­ benzenes from marine traffic therefore esis that they are produced directly by seems most likely. Sauer et al. (1978) phytoplankton is also supported by the have noted that the research vessel is a covariation of cell counts and aldehyde significant source of alkylated benzenes concentration in a time series of seawater to surface waters; nonetheless, taking samples from Vineyard Sound (Gschwend precautions to avoid ship-derived sample 1979). The linear correlation coefficient contamination, they also found between for heptanal with chlorophyll a in the ( 2 and 10 ng 1,3- + 1,4-dimethylben­ Peru upwelling region samples was 0. 76 zene · kg-1 seawater in nonpolluted Gulf (P < 0.001). of Mexico surface seawater. We cannot at On the other hand, these aldehydes present determine whether the recovery may be produced by heterotrophic oxi­ of low levels of these aromatic hydrocar­ dation of algal organic matter. To test this bons in this relatively remote region is hypothesis, we plotted the aldehyde con­ due to ubiquitous fossil fuel-related con- centrations against dissolved oxygen (e.g. Volatile organics in seawater 1051

6 20

_, 'II),""" 4 ;:::-. ->c '~:>, ~.: ""' ~ '"0.88 1:: 10 .· r= 0.82 :) 2 ~ C).. .., /. ~ 0 0 0 5 10 15 20 0 5 10 15 20 1 HEPTANAL (ng.kq 1) HEPTANAL (ng. kq )

Fig. 5. Heptanal vs. dissolved oxygen and vs. nitrate. r is linear regression coefficient.

heptanal: Fig. 5). Rather than the inverse (photo?) oxidation reactions (Wheeler correlation expected for a heten;>trophic 1972; Zafiriou 1977) that produce these oxidation mechanism, we (ound strong aldehydes. direct correlations [e.g. linear correlation The occurrence of aldehydes in sea­ coefficient of 0.88 (P < 0.001) for hepta­ water is noteworthy since they may be nal vs. oxygen]. This aldehyde-oxygen involved in the formation of refractory correlation indicates a source related to polymeric substances. They are highly the atmosphere, to primary production, reactive and may combine with other or­ or to chemical oxidation. Since_ there is ganic compounds through condensation also a strong inverse correlation for hep­ reactions and thus contribute to the intro­ tanal vs. nitrate ( -0.82, P < 0.001) in the duction of the extensive aliphatic char­ 5-, 20-,;and 100-m samples (Fig.,5), the acter of marine humic materials (Stuer­ source appears to be related to primary mer and Payne 1976; Gagosian and production rather than to atmospheric gas Stuermer 1977; Stuermer and Harvey exchange. 1978). Chemical oxidation of algal metabo­ lites released during intense primary Conclusion production may be responsible for the , The detection of volatile organic· com­ aldehyde distribution observed. U nsatu­ pounds in seawater from the Peru up­ rated fatty acids are potential metabolite welling region suggests their use as reactants for chemical oxidation to alde­ tracers of some sources and sinks of or­ hydes. Two of the most common fatty ganic matter in the sea. In situ biological acids in phytoplankton are cis-9-hexade­ processes dominate as the source of dis­ cenoic acid and cis 9-octadecenoic (oleic) solved volatile organic constituents in acid (Ackman et al. 1968; Chuecas and seawater from this region, and other pro­ Riley 1969; Fisher and Schwarzenbach cesses, po-ssibly heterotrophic reminer­ 1978). Autoxidation of cis-9-octadecenoic alization or gas exchange with the atmo­

acid has been shown to yield C 7-C10 al­ sphere, are responsible for limiting the ) kanals (Badings cited in Schauenstein et concentrations of these substances. An­ al. 1977). Corresponding degradation of thropogenic contamination must also be cis-9-hexadecenoic acid would produce considered for some compounds ob­

C 5-C8 alkanals. These two fatty 'acids are served at this remote site, but it is diffi­ found at J.tg ·'kg-1 concentrations in sea­ cult to be certain that this pollution did water (Jeffrey 1966) and are particularly not come from the research vessel. abundant in surface slicks (Kattner and Our study also suggests the occurrence Brockman 1978) which may be the site of . of three specific environmental transfor- 1052 Gschwend et al. mation processes of organic compounds netically transformed products in seawater. in seawater: carboxylic acids may be con­ Mar. Chern. 5: 60§-632. verted to hydrocarbons by zooplankton; GROB, K. 1973. Organic substances in potable water and its precursor. Part 1. Methods for polyolefinic moieties may interconvert to their determination by gas-liquid chromatog­ yield other geometric isomers; unsatu­ raphy. J. Chromatogr. 84: 255-273. rated lipids may be autoxidized to pro­ --, AND F. ZURCHER. 1976. Stripping of trace duce n-aldehydes. organic substances from water. Equipment and procedure' J. Chromatogr. 117: 285-294. References GscHWEND, P. M. 1979. Volatile organic com­ pounds in seawater. Ph.D. thesis, Woods Hole ACKMAN, R. G., C. S. TOCHER, AND J. MCLACHLAN. Oceanogr. lnst. 271 p. 1968. Marine phytoplankton fatty acids. J. HOLM-HANSEN, 0., C. J. LORENZEN, R. W. Fish. Res. Bd. Can. 25: 1603-1620. HOLMES, AND}. D. STRICKLAND. 1965. Fluo­ AVIGAN, }., AND M. BLUMER. 1968. On the origin rometric determination of chlorophyll. J. Cons. of pristane in marine organisms. J. Lipid Res. Cons. Int. Explor. Mer 30: 3-15. 9: 350-352. jEFFREY, L. M. 1966. Lipids in seawater. J. Am. BEERS, J. R., M. R. STEVENSON, R. W. EPPLEY, AND Oil Chern. Soc. 43: 211-214. E. R. BROOKS. 1971. Plankton populations and KATTNER, G. C., AND U. H. BROCKMAN. 1978. Fat­ upwelling off the coast of Peru, June 1969. ty acid composition of dissolved and particulate Fish. Bull. 69: 859-876. matter in surface films. Mar. Chern. 6: 233-241. BLUMER, M., R. R. GUILLARD, AND T. CHASE. 1971. KIKUCHI, T., T. MIMURA, Y. MORIWAKI, A. ANDO, Hydrocarbons of marine phytoplankton. Mar. AND K.-1. NEGORO. 1974. Metabolites of a dia­ Bioi. 8: 183-189. tom, Synedera rumpens Kutz. isolated from --, M. M. MULLIN, AND D. W. THOMAS. 1964. water in Biwa. Identification of odorous Pristane in the marine environment. Helgol. compounds, n-hexanal and n-heptanal, and Wiss. Meeresunters. 10: 187-201. analysis of fatty acids. Chern. Pharm. Bull. (To­ BROECKER, W. S., AND T.-H. PENG. 1974. Gas ex­ kyo) 22: 945-949. change rates between air and sea. Tellus 26: LEE, C., AND J. L. BADA. 1975. Amino acids in 21-35. equatorial Pacific Ocean water. Earth Planet. BUTTON, D. 1976. The influence of clay and bac­ Sci. Lett. 26: 61-68. teria on the concentration of dissolved hydro­ ---, AND ---. 1977. Dissolved amino acids carbon in saline solution. Geochim. Cosmo­ in the equatorial Pacific, the Sargasso Sea, and chim. Acta 40: 435-440. Biscayne Bay. Limnol. Oceanogr. 22: 502-510. CARPENTER, J. H. 1965. The Chesapeake Bay In­ LONNEMAN, W. A., R. L. SElLA, AND J. J. BUFALINI. stitute technique for the Winkler dissolved oxy­ 1978. Ambient air hydrocarbon concentration gen method. Limnol. Oceanogr. 10: 141-143. in Florida. Environ. Sci. Techno!. 12: 459-463. CAUTREELS, W., AND K. VAN CAUWENBERGHE. McAULIFFE, C. 1966. Solubility in water of par­ 1978. Experiments on the distribution of or­ affin, cycloparaffin, olefin, acetylene, cycloole­ ganic pollutants between airborne particulate fin, and aromatic hydrocarbons. J. Phys. Chern. matter and the corresponding gas phase. Atrnos. 70: 1267-1275. Environ. 12: 1133-1141. MACKINNON, M.D. 1979. The measurementofthe CHUECAS, L., AND J.P. RILEY. 1969. Component volatile organic fraction of the T()C in seawa- fatty acids of the total lipids of some marine ter. Mar. Chern. 8: 143-162. , phytoplankton. J. Mar. Bioi. Assoc. U.K. 49: MENZEL, D. W. 1975. Primary productivity, dis­ 97-116. solved and particulate organic matter, and the CLARK, R. C., }R., AND M. BLUMER. 1967. Distri­ sites of oxidation of organic matter, p. 659-678. bution of n-paraffins in marine organisms and' In E. D. Goldberg [ed.], The sea, v. 5. Inter­ sediment. Limnol. Oceanogr. 12: 79-87. science. COLLINS, R. P., AND K. KALNINS. 1965. Volatile MOLLER, D. G., AND L. }AENICKE. 1973. Fuco­ constituents of Synura peterssenii-1. The car­ serraten, the female sex attractant of Fucus ser­ bonyl fraction. Lloydia 28: 48-52. ratus L. (Phaeophyta). FEBS (Fed. Eur. Bio­ EICHMANN, R., AND OTHERS. 1979. n-Alkane stud­ chem. Soc.) Lett. 30: 137-139. ies in the troposphere. 1. Gas and particulate SAUER, T. C., }R. 1978. Volatile liquid hydrocar­ concentrations in North Atlantic air. Atmos. bons in the marine environment. Ph.D. thesis, Environ. 13: 587-599. Texas A&M Univ. 317 p. FIELD, F. H. 1968. Chemical ionization mass spec­ --, W. M. SACKETT, AND L. M. jEFFREY. 1978. trometry. 8. Alkenes and alkynes. J. Am. Chern. Volatile liquid hydrocarbons in the surface Soc. 90: 5649-5656. coastal waters of the ·Gulf of Mexico. Mar. FISHER, N. S., AND R. P. SCHWARZENBACH. 1978. Chern. 7: 1-7. Fatty acid dynamics in Thalassiosira pseudon­ SCHAUENSTEIN, E., H. ESTERBAUER, AND H. ZOLL­ ana (Bacillariophyceae): Implications for physi­ NER. 1977. Aldehydes in biological systems. ological ecology. J. Phycol. 14: 143-150. · Their natural occurrence and biological activi­ GAGOSIAN, R. B., AND D. H. STUERMER. 1977. The ties. Pion. cycling of biogenic compounds and the diage-- SCHWARZENBACH, R. P., R. H. BROMUND, P. M. Volatile organics in seawater 1053

GSCHWEND, AND 0. C. ZAFIRIOU. 1978. Vol­ WHEELER, ]. 1972. Some effects of solar levels of atile organic compounds in coastal seawater. ultraviolet radiation on lipids in artificial sea­ Org. Geochem. l: 93-107. water. J. Geophys. Res. 77: 5302-5306. SIMONEIT, B. R., R. CHESTER, AND G. EGLINTON. WILLIAMS, P. }., AND R. W. GRAY. 1970. Hetero­ 1977. Biogenic lipids in from the trophic utilization of dissolved organic com­ lower atmosphere over the eastern Atlantic. pounds in the sea. 2. Observations on the re­ 267: 682-685. sponses of heterotrophic marine populations to STRICKLAND, J.D., R. W. EPPLEY, AND B. ROJAS DE abrupt increases in concentration. MENDIOLA. 1969. Phytoplankton populations, J. Mar. Bioi. Assoc. U.K. 50: 871-881. nutrients and photosynthesis in Peruvian coast­ WRIGHT, R. T., AND J. E. HOBBIE. 1966. The use al waters. Bioi. lnst. Mar. Peru (Callao) 2: 1- of glucose and acetate by bacteria and algae in 45. aquatic ecosystems. Ecology 47: 447-464. STUERMER, D. H., AND G. R. HARVEY. 1978. Struc­ YENTSCH, C. S., AND D. W. MENZEL. 1963. A tural studies on marine humus: A new reduc­ method for the detendnation of phytoplankton tion sequence for carbon skeleton determina­ chlorophyll and phaeophytin by fluorescence. tion. Mar. Chern. 6: 55-70. Deep-Sea Res. 10: 221-231. ---, AND J. R. PAYNE. 1976. Investigation of ZAFIRIOU, 0. C. 1977. Marine organic photochem­ seawater and terrestrial humic substances with istry previewed. Mar. Chern. 5: 497-522. carbon-13 and proton nuclear magnetic reso­ nance. Geochim. Cosmochim; Acta 40: 1109- 1114. WALSH, J. J. 1975. A spatial simulation model of the Peru upwelling ecosystem. Deep-Sea Res. Submitted: 26 December 1979 22: 201-236. Accepted: 12 june 1980

')