GEOPHYSICAL RESEARCH LETTERS, VOL. 25, NO. 15, PAGES 2853-2856, AUGUST 1, 1998

Carbon isotopes in aquatic plants, , as records of past hydrothermal and magmatic activity JohnB. Reid,Jr. 1 JesseL. Reynolds2 Nathan T. Connolly,Shari L. Getz,Pratigya J. Polissar3,and LawrenceJ. Winship

Schoolof Natural Science,Hampshire College, Amherst, Massachusetts

Laura J. Hainsworth 4

Lawrence Livermore National Laboratories, Livermore, California

Abstract. Hot andcold springs contribute "dead" (14C free) streambedplants as indicators of present and past magmatic dissolvedinorganic carbon (DIC) to the and Hot CO2emissions into the caldera's surfacewaters. Creek. Headwatersaquatic plants have modem •4C, but live The use of aquatic plants for this purpose came about plants downstreamof the intracaldera springs are depleted in accidentally.Where the Owens River skirts the northeast flank •4C,(as low as 19% modem,with apparentages up to 13.3 of the caldera's resurgent dome, its course has alternated kyrs). In an abandoned meander of the upper Owens River, betweentwo parallel meanderbelts, apparently in responseto preserved streambed plants are buffed by 600 year old Inyo tilting of its floodplain. One possible causeof the tilting is Craterspumice. Apparent •4C ages of theseplants exceed true the alternating inflation and deflation of the dome (Reid, ages by- 1100 years indicating that they also incorporated 1992). We hoped to date the river's occupation of these dead DIC as they grew. The preservedplants are downstreamof meanderbelts by radiocarbon-dating the remains of ancient Big Springs,whose elevated dead DIC may representmagmatic plants buried in abandonedchannels of both meander belts. For 'CO2. The buried plants incorporated -10% dead carbon, control, we also dated modem plants in the Owens River and although modem plants here have -50% dead carbon, Hot Creek. Modem plants throughout the caldera- and for at suggestingthat more magmaticCO 2 is now entering the upper least65 km furtherdownstream - have anomalous •4C ages (up Owens River than at the time of the Inyo Craters eruptions 600 to 13,000 years), particularly near the hot and cold springs. years ago. Whilethis findingmeant that buried plants cotild not be used for dating the hypothetical past behavior of the dome, it Introduction suggestedthat they might be useful in assessing modern and past emissions of dead CO2 into the caldera's river system. Long Valley caldera(Fig. 1), a 17 x 32 km depression on Other investigatorshave seen •4C deficienciesin modern the eastern boundary of the , California, formed plants due to magmatic activity (e.g.: Sveinbj6rnsd6ttir et al., about 760 kyr ago (Bailey, 1989). Magmatic activity since 1992; Srdocet al., 1986); As discussedbelow, Long Valley then has produceda 500 m high resurgent dome in the west- provides the additional opportunity of comparing past and central caldera floor. The most recent caldera-related volcanism present releases of magmatic CO2 to surface waters in a was the Inyo Crater eruptions, -600 years ago, and phreatic volcanic setting. explosions on ,-700 years ago (M.L. "Dead"(•4C free) dissolvedinorganic carbon (DIC) in Sorey, pers. comm, 1997). Since 1980, there have b6en hydrothermal and volcanic areas has been attributed to several indications of renewed magmatic unrest. Seismic magmatic inputs (e.g., Shevenell and Goff, 1993; Giggenbach, activity and changes in fumarole composition may indicate 1995; Rose and Davisson, 1996). While dissolution of shallow magmatic intrusion (Hill et al., 1990; Sorey et al., carbonate minerals at depth is a possibility, and the deep 1993). Leveling surveys along U.S. Highway 395 have shown reservoir of the Long Valley hydrothermal system may reside -60 cm of inflation of the dome between 1979 and 1984 in metamorphicmarine sediments(Sorey et al., 1991), CO2- (Castle et al., 1984). Diffuse emissions of magmatic CO2 near rich gas with magmatic He and C isotopic character has been Mammoth Mountain followed a swarm of seismic activity in enteringthe calderaas diffuse emissions (-1200 T CO2/day) at 1989 (Farrar et al., 1995). Magmatic CO2 may also be entering Mammoth Mountain since 1990 (Farrar et al., 1995; Sorey et the surfacewaters of Long Valley and taken up by aquatic al., 1996). The isotopic characterof DIC in wells and springs plants.In this report, we usethe anomalous•4C contentof near Mammoth Mountain shows that the groundwater is dissolving magmatic CO2 (Sorey et al., 1996). Hilton (1996) examined He and C isotopes of hydrothermal gases and • to whomcorrespondence should be sent 2 Now at Departmentof EnvironmentalScience, Policy and concludedthat the primary source of CO2 is magmatic gas, Management,University of California,Berkeley, CA 94701 although a small input from thermal decarbonizationof marine 3 Now at Departmentof Geosciences,University of Massachusetts, Amherst, MA 01003 carbonates could not be excluded. Taylor and Gerlach (1984) 4 Nowat ChemistryDepartment, Emory and Henry College, Emory, VA found that volcanicrocks are not a sourceof hydrothermalCO2. 24327 In addition,we canrule out decayof atmospheric•4C in the Copyright1998 by the AmericanGeophysical Union. groundwatersystem as the main reason for diminished•4C activity since residencetime is estimated at -1200 years (small Papernumber 98GL01854. comparedwith the half life of •4C) in the hydrothermal 0094-8534/98/98GL-01854505.00 reservoir (White et al., 1990).

2853 2854 REID ET AL.: CARBON ISOTOPES, LONG VALLEY CALDERA

results in stepwise increases in concentration, such as those shownfor arsenicin Fig. 2a. The highestinputs of most of the 6180 ' o oCrates 4,,•o4830 •'.','&,,,.t,,• / 7.,o7180 12 analyzedelements occur at the major hot springs (especially / "• • As, Na, Rb, Cs, W); cold springs contribute the bulk of a few / ':+:....-i•:• , / % 40201--"-[ elements(eg: Mg, Ca, P, Mo, Zn). Sr, K and Si are addedfrom both springtypes about equally. These data are available upon request. : q••.-:•':•:.'•..••:•:!\:'rt-r• gt / • •-•' •----' ] HotCreek Gore • C ,, I-%370 'i%w Carbon isotopic data for Long Valley plants and water • c• ' _._--/-132•0/13 samplesare shown in Figs. 2b, 2c and 3. On a plot showing •4Cand '3C variation together (Fig. 3), •4Cranges widely from 0% modernin magmaticCO 2 to 113% in atmosphericCO•, and thusmixing is the main causeof •4C variation. •5'3Cvaries 2470• •n.,...••-•_Crc 24•• •! ak¾.e ...... • Rte6bridge . much more narrowly, mainly due to fractionation 8950 • • • • • reshopm•m [ o• n:9o•"419011__2 o.•c,•, :•_ -.-•• do•5•y.• accompanying photosynthesis. CO• samples from waters appearto be mixturesof atmospheric,magmatic and soil gas on era (shadedtriangle in Fig. 3). Plants show typical depletion in I L gValley Cald •5'3Cby about20-25 per mil relativeto their host waters Figure 1. Map of Long Valley caldera(after Bailey, 1989) (Hoefs, 1997), with plants near Big Springs containing the showingapparent ages (numbers in years) and percent"dead" C mostnegative d'3C, and thus the highestproportions of soil (solid portionsof horizontalbars) for streambedplants, waters gas.Variation in •4Cand •3C with distancedown each drainage (W) and an aquaticinsect casing sample (B) in the Owens River are shown in Figs. 2b and 2c. An aquatic plant sample from and Hot Creek. Lake Mary (headwatersof Hot Creek) has a full complementof •4C(113 % modem•4C (pmC)), but all otherstreambed plants Here, we report the results of carbon isotopic studiesof the we havesampled are depleted in '4C.Where data for both exist, aquaticplants and springwaters of Long Valley as measuresof plants and their waters are similar in pmC, with mismatches present and past magmatic CO2 release in the caldera. due probably to incomplete mixing of water types, Hainsworth et al. (1995) used tree rings at a fumarole near photosynthesis-inducedfractionation and possibly the fact Mammoth Mountain to document local variations in the emissions of magmatic CO2 over the past decade. In this report, we suggestthat apparent radiocarbonages of preserved Owens River Mammoth / Hot Creek aquaticplants may provideclues to magmaticC0• duringthe 160 BS HCD FHSHS LH? 140 A :i :$":½•i :: . 140 Inyo Craterseruptions 600 years ago, provided the plants can 120 be dated by other means. 1oo 120100 80 6O Methods 40 40 20

o ! •o Surface waters and aquatic and terrestrial plants were 1 oo collected in June of 1994 and 1996. We measured pH and alkalinity for every site in 1996; discharge was 60 40 measuredat selectedsites near streamand junctions. We 20 precipitated DIC samples as SrCO3 using strontium chloride and ammonium hydroxide. Liquid scintillation spectrometry C ' (BetaAnalytic, Miami, Florida)was used to measure'4C of all , , ,0 plant samplesand all DIC precipitates but one. It, the conifer ::::•:? -25 charcoals and the buried plant remains were analyzed by ø35 .•:• -35 accelerator mass spectrometry (AMS) (Lawrence Livermore • 45ø25t , : : : • : : , , : ; : I • I!' I i I NationalLaboratory). Fractionation in '4C wascorrected using 40 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 measured•5'3C. We performed•3C analyses on additionalplants distanceupstream from Lake Crowley (kin) by combustion and mass spectrometry at the University of --n- Mammoth / Hot Creek o Hot Creek Delta Vermont Stable Isotope Laboratory. Inductively coupled + Fish HatcherySpring / Creek + Hot Springs plasma-atomic emission spectroscopy (Thermal Jarrell Ash -•- Little Hot Creek + Owens River ICAP 61) at Middlebury College and inductively coupled -a- Big Springs + Mono Tunnel plasma-mass spectrometry (Perkin Elmer Elan 6000) at x streamsideterrestrial plants HampshireCollege were usedfor dissolved cation analyses of Figure 2. (A) Arsenic concentrationsin spring and surface the 1994 and 1996 water collections, respectively. waters; this stepped pattern is typical of most analysed elements. Little Hot Creek = 248 ppb and Hot Springs = 19 1 Results ppb(off scale).(B) 14Ccontent of plantsand DIC. (C) •3Cof plants and DIC. All plants are aquatic except (x) which are The results of the elemental analyses show that the caldera terrestrial grasses immediately adjacent to the stream. "W" surface water chemistry is controlled by the hot and cold next to a point signifies DIC. The grey vertical lines represent springs (Fig. 2a). Most conservativeelements enter the stream input from Big Springs (BS), the Hot Creek delta (HCD), Fish in pronounced,discrete inputs at the major springs and have Hatchery Springs(FHS), Hot Springsin Hot Creek gorge (HS), essentially constant concentrationsbetween these inputs. This and Little Hot Creek (LHC). RED ET AL.: CARBON ISOTOPES, LONG VALLEY CALDERA 2855

that plants integrate over the growing season, while water of all analyzed terrestrial plants (--26%0) and aquatic plants samplesare instantaneous.Fractionation of '4C/'2C during- above the major springs (--18%o) are typical (Deines, 1980). photosynthesis,-45 %0 (Fritz et al., 1989), is small compared Two major springs, Big Spring and Fish Hatchery Springs to the rangeof values among plant samples (19 to 113 pmC). contributelighter carbon(lower •5•3C)to plants immediately The large amountsof deadDIC at Big Spring and Fish Hatchery downstream (Fig. 2c). The lighter carbon stems mainly from Spring, the two major cold springs, indicate that they receive biogenic soil CO•. As one progresseson downstream, the low dead carbon from magmatic gases,the hydrothermal system, or •5•3Cof the plantsi•eturns quickly to the "baseline"value of- contact with carbonate minerals. 18%o.•4C excursions are considerably larger, and persist for Total deadDIC leaving the calderacan be estimated from much greater distancesdownstream. simultaneousmeasurement of streamdischarge, alkalinity, and ]4Ccontent (estimated from plant isotopes).In June, 1996, -15 T/day total DIC and -6 T/day deadDIC were discharging Preserved plants below Big Spring as possible from the Owens River into Lake Crowley. These values are records of past C02 release probably slight underestimatessince by the time the water reachedthe lower Owens River some deadDIC may have been Dead DIC from the springs of Long Valley causesdramatic lost through photosynthesis and atmospheric mixing. The depletionsin '4C of live streamplants in the calderaand for maximum dead DIC fluxes of the Owens River and Hot Creek many kilometers downstream.If the deadDIC is magmatic in were 4.6 and 5.7 T/day, respectively, giving an estimate of origin, preservedstream plants could serve as recordsof past -10 T/day dead DIC enteringthe surfacewaters of the caldera. magmatic activity. Although the waters of Mammoth Creek (upper Hot Creek) MagmaticCO= is the likely sourceof deadDIC in the modern at US 395 chemically resemble those of the headwaterswith Long Valley hydrothermalsystem. As thermal waters saturated verylow cationconcentrations, the '4Ccontents of two aquatic with CaCO3rise, calcite precipitates and CO= evolves (White plantsthere are depleted (74 pmC).15'3C values and Peterson,1991). Hilton (1996) determinedthat this CO= is (-17.5 + 0.2%0) of these samplesare normal for aquatic plants primarily magmaticin origin. Big Spring is a cold spring with (Deines, 1980). There are no major springs upstream from this a high deadDIC concentrationbut probably does not receive point, and contributions from carbonates is excluded by the chemical inputs from the hydrothermal system (M. L. Sorey, low and nearly constant calcium and magnesium pets. comm., 1996). The possible sourcesof its deadcarbon concentrationsfrom Lake Mary to US 395. This '4Canomaly is are interactions with carbonates in marine metasediments, likely producedby fumarolic activity in a small valley -1.5 km magmatic emissions of CO2 to its rechargearea on the north northwestof the samplingsite, whereCO 2 from hydrothermal side of Mammoth Mountain (Heim, 1992), and decomposition fluids boiling at depth is entering the shallow groundwater. of ancient organic matter buried by the moat rhyolites (M. L. Unlike •4C,the •3Ccontents of aquaticplants are not useful Sorey, pers. comm., 1996). Since the molar Ca concentration as indicatorsof paststream carbon isotopic character. The •3C at Big Spring is about an order of magnitudelower than that of deadDIC, only a small portion of the C is from calcium carbonates. If Mg is included, at most 40% of C is from atmosphericCO2 carbonate minerals. Thus, there may be a large input of magmatic CO2 into the recharge waters of Big Spring. 120, ß . soil gas • ' Forthcoming helium isotope data from the U.S. Geological

100, 0 Survey shouldclarify this issue (M. L. Sorey, pers. comm., 1996). Heim (1992) estimatedthat Big Spring water averages -12 years old basedon tritium levels. Therefore,the '4C zE.0. o contentsof plants downstreamof Big Spring may be sensitive ,.• - 3 ,-•I indicatorsof magmaticCO2emissions. We have discoveredpreserved aquatic plant material that may have recordedthe isotopic characterof DIC in the upper OwensRiver at an important juncture in its recent past. We • 40, have found a pumice-filled abandonedchannel of the Owens

1o River -3 km. upstreamof the Hot Creek confluenceapparently filled in a single event at the time of the Inyo Craters 20. 4 -• 15 eruptions,600 years ago. AMS •4C dates on 11 conifer magmaticCO2 2{) charcoalsin the pumice clusterbetween 600 and 700 ybp (Fig. 0 . , . • . • 40 4). Beneaththe pumice,we found the remains of aquaticplants -40 -30 -20 -10 0 with roots still attached to small river cobbles in anoxic •5•3C groundwaterin the thalweg stream gravel. AMS dates of these gr21waters plants plantremains range from 1100 to 2040ybp. Thus, when they 1 Big Springs ß Big Springs 2 Fish HatcherySpring ß OwensRiver aboveHC were buried by pumice, they contained --10% dead DIC, 3 Hot Creek@ Hot Springs O OwensRiver below HC compared with -50% in the plants of this section of the Owens 4 Hot Springs ß Hot Creek aboveFHS El Hot Creek below FHS River today. The only major inputsupstream of this site today are Big Spring and the Mono Craters Tunnel. Based on calcium Figure 3. '4Cand nC variationsin watersand plants of Long concentrations,the total upstream contribution from dissolved Valleycaldera. Wide variation in •4Cis dueto mixingof CO2 CaCO3 is no morethan one-third of the deadDIC. from three reservoirs(shaded triangle), while 1513Cvaries These resultsimply that (1) magmaticCO2 was entering the mainly due to fractionation during one or more stages of upper Owens River at the time of the 600 ybp Inyo Craters photosynthesis(see text). eruptions, and (2) that modern levels there (1994-1996) are 2856 REID ET AL.: CARBON ISOTOPES, LONG VALLEY CALDERA

Giggenbach,W. F., Variationsin the chemicaland isotopiccomposition 14C age anomaly dueto -50% magmatic Ctoday of fluids dischargedfrom the Taupo Volcanic Zone, New Zealand, •••::::••_:::•::-:.: ...... Journalof Volcanologyand GeothermalResearch, 68, 89-116, 1995...... -, .... _• ..... • ...... Hainsworth,L. J., C. D. Farrar,J. R. Southron,Magmatic CO 2 record in modernaquatic plants in anomalyfrom pineneeds and growth rings at MammothMountain, California, Eos, upperOwens River --10% magmatic C 76, 688, 1995. Heim, K., Hydrologicstudy of theBig Springs,Mono County,California, unpublishedbachelor's thesis, California State University, Fullerton, preservedaquatic plants buried by 1992. Hill, D. P., W. L. Ellsworth, M. J. S. Johnston,J. O. Langbein, D. H. Oppenheimer,A.M. Pitt, P. A. Reasenberg,M. L. Sorey, and S. R...... #: ...... -.**• ..... McNutt, The 1989 earthquakeswarm beneathMammoth Mountain, California:an initial look at the 4 May through30 Septemberactivity, Bulletin- SeismologicalSociety of America,80, 325-339, 1990. Hilton, D. R., The helium and carbonisotope systematics of a continental 7t }00 6000 5000 4000 3000 2000 1000 0 geothermalsystem: results from monitoringstudies at Long Valley Caldera, Chemical Geology, 127, 269-295, 1996. measured radiocarbon age (years bp) Hoefs,J. StableIsotope Geochemistry, Springer, Berlin, 1997. Reid, J. B. Jr., The Owens River as a Tiltmeter for Long Valley Caldera, Figure 4. Apparentages of modemand preserved California, The Journal of Geology,100, p. 353-363, 1992. Rose, T. P. and M. L. Davisson, Radiocarbon in hydrologic systems plants in the upper OwensRiver along with true ages for containingdissolved magmatic carbon dioxide,Science, 273, 1367- conifer charcoals collected in a pumice-filled abandoned 1370, 1996. channel of the fiver. Charcoalscluster about 600 years in age, Shevenell,L. and F. Goff, Addition of magmaticvolatiles into the hot springwaters of Loowit Canyon,Mount St. Helens, Washington, suggestingthe pumiceis Inyo Cratersejecta. Modemplants USA, Bulletinof Volcanology,55, 489-503, 1993. contain about 50% magmaticC, five times higher than the Sorey,M. L., B. M. Kennedy,W. C. Evans,C. D. Farrar, and G. A. magmaticC contentof plantsgrowing in the fiver at the time Suemnicht,Helium isotopeand gas dischargevariations associated with crustal unrestin Long Valley caldera, California, 1989-1992, of the 1400 AD eruptions. Journal of GeophysicalResearch, 98, 15871-15889, 1993. Sorey,M. L., B. M. Kennedy,and L. J. Hainsworth,Carbon dioxide and helium emissions from a reservoir of magmatic gas beneath Mammoth Mountain, California (abstract),Eos Trans. AGU, 77, F830, five times higher than those recordedby the plant remainsat 1996. the time of the last caldera-related eruptions. Sorey,M. L., G. A. Suemnicht,N. C. Sturchio,and G. A. Nordquist,New evidence on the hydrothermal system in Long Valley caldera, California, from wells, fluid sampling,electrical geophysics,and age Acknowledgments.We thank Mike Sorey and Roy Bailey for determinationsof hot-springdeposits, Journal of Volcanologyand stimulatingdiscussions; Sean Pack, Micah Jessup, and Carla Christensen Geothermal Research, 48, 229-264, 1991. for field assistance;and Paul Bierman,Ray Coish,Dan Dawson, Darla Srdoc,D., I. Krajcar-Bronic,N. Horvatincic,and B. Obelic, Increase of Heil, and AndreaLini for the use of analytical facilities. The work was •4C activityof dissolvedinorganic carbon along a river course, supportedby a NationalScience Foundation grant to JBR (EAR9628115) Radiocarbon, 28, 515-521, 1986. anda HowardHughes Medical Institute grant to HampshireCollege. Sveinbj6rnsd6ttir,A. E.;. E., J. Heinemeier,N. Rud, and S. J. Johnsen, Radiocarbon anomalies observed for plants growing in Icelandic geothermalwaters, Radiocarbon, 34, 696-703, 1992. References Taylor, B. E., andT. M. Gerlach,Mantle CO2 degassingat Long Valley, SteamboatSprings at the CosoRange (abstract),EOS TransAGU , 65, 1153, 1984. Bailey, R. A., Geologicmap of Long Valley caldera,Mono-Inyo Craters White,A. F, M. L. Peterson,H. A. Wollenberg,and S. Flexser,Sources volcanic chain, and vicinity, Eastern California, U.S. Geologic and fractionationprocesses influencing the isotopicdistribution of H, SurveyMap 1-1933, 1989. O, and C in the Long Valley hydrothermalsystem, California, U.S.A., Castle,R. O., J. E. Estrem,and J. C. Savage,Uplift acrossLong Valley Applied Geochemistry,5, 571-585, 1990. caldera, California, Journal of GeophysicalResearch, 89, 11507- White, A. F, M. L. Peterson,H. A. Wollenberg, and S. Flexser, Sources 11516, 1984. and fractionationprocesses influencing the isotopicdistribution of H, Deines, P., The isotopiccomposition of reduced organic carbon, in ,.O, and C in the Long Valley hydrothermalsystem, California, U.S.A., Handbook of EnvironmentalIsotope Geochemistry,Vol. 1, The Applied Geochemistry,5, 571-585, 1990. Terrestrial Environment,A, P. Fritz and J. Ch. Fontes (Eds.), pp. 329- 390, Elsevier: Amsterdam, 1980. J)B.Reid Jr. z, J.L. Reynolds 2,N.T. Connolly, S.L. Getz, P.J. Polissar 3, Farrar, C. D., M. L. Sorey,W. C. Evans,J. F. Howle, B. D. Kerr, B. M. and L.J. Winship, School of Natural Science, Hampshire College. Kennedy,C.-Y. King, andJ. R. Southon,Forest-killing diffuse CO2 at Amherst, MA, 01002. MammothMountain as a sign of magmaticunrest, Nature, 376, 675- L.J. Hainsworth4, Lawrence Livermore National Laboratories, 678, 1995. Livermore, CA 94551. email: [email protected] Fritz, P., J-Ch. Fontes,S.K. Frape, D. Louvat, J-L. Michelot, and W. Balderer, The isotopegeochemsitry of carbon in groundwater at (ReceivedNovember 3, 1997; revised April 7, 1998; accepted May 20, Stripa,Geochimica et CosmochimicaActa, 53, 1765-1775,1989. 1998)