U.S. DEPARTMENT OF THE INTERIOR
U.S. GEOLOGICAL SURVEY
INVESTIGATIONS OF GAS SEEPS AND SPRINGS IN THE VICINITY OF THE GAS
ROCKS, SOUTH SHORE BECHAROF LAKE, ALASKA
by
Robert B. Symonds1, Beatrice E. Ritchie2, Robert G. McGimsey3, Michael H. Ort4 , Robert J. Poreda5, William C. Evans6, and Cathy J. Janik6
MJ.S. Geological Survey, Cascades Volcano Observatory, 5400 MacArthur Blvd., Vancouver, WA 98661 2Department of Geology, Portland State University, P.O. Box 751, Portland, OR 97207 3U.S. Geological Survey, Alaska Volcano Observatory, 4200 University Drive, Anchorage, AK 99508 4Department of Geology, P.O. Box 4099, Northern Arizona University, Flagstaff, AZ 86011 sDept. of Earth and Environmental Science, 227 Hutchison Hall, University of Rochester, Rochester, NY 14627 6U. S. Geological Survey, 345 Middlefield Road, Menlo Park, CA 94025
Open-File Report 97-127
This report is preliminary and has not been reviewed for conformity with U.S. Geological Survey editorial standards or with the North American Stratigraphic Code. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.
1997 CONTENTS
Page Abstract...... 1 Introduction...... 1 Field observations and sampling...... 2 Analytical methods...... 9 Gases...... 9 Waters...... 11 Results...... 11 Chemical andisotopic compositions of gases...... 11 Bulk composition...... 11 Carbon isotopes...... 15 Rare-gas isotopes...... 15 Chemical and isotopic compositions of waters...... 15 Bulk composition...... 15 Hydrogen and oxygen isotopes...... 20 Carbon isotopes...... 22 Gas hazards implications...... 23 Impact on local ecosystems...... 24 Conclusions...... 24 Acknowledgments...... 25 References cited...... 25
ILLUSTRATIONS
Page FIGURE 1. Location map showing sites investigated in the vicinity of The Gas Rocks and Ukinrek Maars...... 2 2. Photograph of the two largest gas vents off The Gas Rocks Peninsula...... 4 3. Photograph showing collection of bulk-gas samples...... 5 4. Photograph of the site UK-2 along the ridge that hosts the Ukinrek Maars...... 7 5. View of bubbling pool sampled at site UK-2 ...... 8 6. Photograph of an area of dead vegetation along the ridge that hosts the Ukinrek Maars...... 9 7. View of East Ukinrek Maar looking west...... 10 8. Plot of N2/Ar versus N2/(O2 + Ar) for the bulk gas samples from The Gas Rocks and Ukinrek Maars...... 13 9. Ternary diagram of Cl, SO*, and HCOa for selected waters from The Gas Rocks and Ukinrek Maars area...... 19 10. Plot showing selected elemental ratios for the 1977 Gas Rock spring waters...... 19 11. Modified Schoeller plot of major constituent concentrations for The Gas Rocks spring waters in 1977, 1981, and 1995...... 20 12. Plot of 8D versus 818O for waters collected in the vicinity of The Gas Rocks and Ukinrek Maars...... 21 TABLES
Page TABLE 1. Chemical compositions of gas samples collected in 1977 from The Gas Rocks and Ukinrek Maars areas...... 3 2. 1995 sites investigated in the vicinity of The Gas Rocks and Ukinrek Maars...... 6 3. Chemical compositions of gases in The Gas Rocks and Ukinrek Maars areas...... 12 4. Dry-gas and rare-gas compositions of rare-gas isotope samples from The Gas Rocks and Ukinrek Maars areas...... 14 5. Chemical and isotopic compositions of waters in The Gas Rocks and Ukinrek Maars areas...... 16 6. Trace element analyses of waters from The Gas Rocks and Ukinrek Maars areas...... 17 7. Chemical and isotopic compositions of waters collected in 1977, 1980, 1981, and 1994 from The Gas Rocks and Ukinrek Maars areas...... 18 8. Predicted 513C in DIG values for potential sources of DIG in water samples from The Gas Rocks and Ukinrek Maars area...... 23 ABSTRACT Rocks spring contain a NaCl-brine compo nent, which has decreased along with tem perature since the spring was first sampled In August 1995, we investigated gas in 1977. Additional saline springs may ex seeps and springs in the vicinity of The ist offshore from The Gas Rocks peninsula- Gas Rocks and Ukinrek Maars on the Similar to sedimentary brines, The Gas southern shore of Becharof Lake, Alaska. Rocks brine probably derives from seawa- At The Gas Rocks, a site of intense CO2- ter or meteoric waters that have reacted degassing since at least 1974, CO2-rich with the underlying marine sedimentary gases discharge into the lake from three rocks. All the saline springs discharge large vents and hundreds of small seeps. CO2-rich gas, which probably plays a role An onshore saline spring also discharges in transporting the brine to the surface. into the lake. No evidence exists for addi The main gas hazards to visitors in tional degassing saline springs that existed the vicinity of The Gas Rocks and the in 1981. Along the ridge that hosts the Ukinrek Maars are tent camping and dig Ukinrek Maars, weak or diffuse CO-2 de ging snow pits near the areas of gas dis gassing occurs from at least four sites, all charge, including the areas of dead and characterized by dead and dying vegeta dying vegetation along the Ukinrek Maars tion. Vegetation was apparently killed both ridge. Such enclosures may concentrate during the 1977 eruption and sometime CO2 to lethal levels. The potential for between 1977 and 1995, but the dying open-air CO2 asphyxiation at The Gas vegetation experienced trauma in 1995. Rocks and Ukinrek Maars has decreased Gas samples were collected at five since a visitor experienced asphyxia at The sites at The Gas Rocks and the Ukinrek Gas Rocks in 1977. Apparently the CO2 Maars ridge to determine bulk gas compo flux at The Gas Rocks has decreased since sition, 813C in CO-2, and rare-gas isotopes. 1977 as indicated by the reduction in the The samples contain 89.5 - 99.5 % CO2, number of onshore springs (or vents) and 0.5 - 10.5% air and air-saturated-water their visible degassing rate. This decrease gases (N2, O-2, Ar), and, for the Ukinrek may reflect progressive outgassing of the Maars ridge samples, 0.3 - 2.1% CH4 and CO2 source. However, the potential for N2 probably derived from heating of the open-air CO2 asphyxiation could poten sedimentary rocks beneath the maars. tially change at any time should mantle These results suggest that the sub-water- CO2 input increase. table gases are nearly 100% CO-2, except Areas of dead and dying vegetation for minor amounts of thermogenic CH« and on the Ukinrek Maars ridge are the main NQ in the Ukinrek Maars samples. Data for impact of the CO2 emissions on the local 813C in CO2 and 3He/ 4He suggest that CO2 onshore and aquatic ecosystems. In addi and He in these gases ultimately come tion, just offshore from The Gas Rocks, from a mantle source. subaqueous CO2 and saline-water dis Water samples were collected at charges produce elevated DIG, Cl, Na, SO^, most gas-sampling sites, The Gas Rocks SiO2, and H+ concentrations in lake waters. spring, the east Ukinrek Maar lake, and More work is needed to discern the soil-gas from a background site on Becharof Lake. CO2 concentrations in the plant-kill zones, Samples were analyzed to determine con the area! extent and flux of the diffuse CO2 centrations of major and selected trace emissions, and the spatial extent of the elements, 8D and 818O in H2O, and 813C in geochemical anomalies in Becharof Lake. dissolved inorganic carbon (DIG). The wa ter samples collected at sites of gas dis charge in Becharof Lake and along Ukinrek INTRODUCTION Maars ridge are dilute waters, except that they contain major amounts (300 - 1900 The Gas Rocks are located on the mg/L) of DIG (mostly as foCOa) from disso southern shore of Becharof Lake on the lution of the discharged CO2. But the sa Alaska Peninsula (Fig. 1). They are part of line waters discharging from The Gas a series of Quaternary volcanic plugs that 156°30'
Becharof Lake
The Gas Rocks
-57°50'
1 mile
Figure 1. Location map showing sites investigated in the vicinity of The Gas Rocks and Ukin- rek Maars. Base is from U.S. Geological Survey, Ugashik D2 quadrangle. Contour interval is 100 feet; dashed lines represent interpolated 50 foot contours. Approximate locations of the East and West Ukinrek Maars are shown for reference. Inset shows location of study area. lie along the Bruin Bay fault. Mesozoic prior to Miller's visit because vegetation marine sedimentary rocks underlie the defoliated by hot erupted mud was just area (Burk, 1965; Dettennan et al., 1987). leafing out during Miller's visit. During his The Gas Rocks area has been investigations near one of the springs, he known as a site of vigorous CO2 degassing was asphyxiated temporarily by a CO2-rich since at least 1974, when it was first inves gas pocket while kneeling over the vent. In tigated by US Bureau of Mines scientists late August, Barnes and McCoy (1979) (Bureau of Mines, 1976). About two miles sampled offshore gases and the hottest on south of The Gas Rocks are the Ukinrek shore spring (52.8°C) in the vicinity of The Maars, two monogenetic craters that Gas Rocks; they also sampled gases and formed during their March-April 1977 waters from a bubbling pool (80.9°C) in the eruption (Kienle et al., 1980; Self et al., West Ukinrek Maar. Their work shows that 1980). In July 1977, R. L. Smith visited the 1977 gases emanating from The Gas The Gas Rocks area and reported a new Rocks area contained up to 98% CO2 and onshore spring and much higher gas dis that the CO2 apparently derived from the charges than before Ukinrek Maars mantle as indicated by 813C data (Table 1). erupted (Barnes and McCoy, 1979). In Obviously, the primitive olivine basalt early August 1977, Tom Miller (personal erupted from the Ukinrek Maars also came communication, 1995) visited The Gas from the mantle. Rocks area and noted three vigorous off The pool in the West Ukinrek Maar shore gas vents, several degassing onshore and the hottest Gas Rocks spring were re- springs (saline and up to 55°C), and a fum sampled in 1980 and 1981, respectively ing onshore mud volcano. The mud vol (Motyka et al., 1993). At the time of these cano apparently erupted about one month visits, there were still several onshore
Table 1. Chemical compositions of gas samples collected in 1977 from The Gas Rocks and Ukinrek Maars areas. Samples collected in evaculated bottles and analyzed by gas chromatography. Data from Barnes and McCoy (1979).______Field numbers: CQ135GM77 CQ134GM77 CQ133GM77 Map site: GR-1A GR-1C Location: onshore Becharof Lake Pool, West spring at The near largest Ukinrek Maar Gas Rocks gas vent Date: August 25, August 25, August 24, 1977 1977 1977 Site temperature T°C 52.8 9.3 80.9 Chemical Species of Dry Gas (mol%)t C02 98.40 97.08 77.48 N2 0.54 2.52 15.14 02 0.02 0.59 3.19 Ar <0.02 0.03 0.34 0.04 0.08 1.88 C2H6 <0.05 <0.05 0.23 Selected ratios N2/Ar >27 84 45 N2/(02+Ar) >13.5 4.1 4.3 Carbon Isotope Composition -6.36 tSpecies below detection in all samples include He (<0.02), H2 (<0.01), and CO (<0.02). Figure 2. Photograph of the two largest gas vents off The Gas Rocks Peninsula. *The largest vent, about 1 meter in diameter, is site GR-1C. springs at The Gas Rocks, but the mud In August 1995, with partial fund volcano had apparently washed away. ing from the U.S. Fish and Wildlife Service Also, the pool in the West Ukinrek Maar (USFWS), Becharof National Wildlife Ref and the hottest Gas Rocks spring had uge, we investigated gas seeps and springs cooled to 16°C and 39°C, respectively. in the vicinity of The Gas Rocks and Ukin However, The Gas Rocks gases remained rek Maars. This is the first comprehensive CO2-rich (99.99% CO2) and 8«C and investigation of degassing at The Gas 3He/4He data indicated that both CO2 and Rocks since 1981. The purposes of our He derived from the mantle. work at The Gas Rocks and Ukinrek Maars Degassing on the ridge that hosts are to (1) investigate the gas seeps and the Ukinrek Maars was investigated par springs, (2) sample the gases and spring tially in 1994 by Tina Neal and Michael Ort waters to determine their present composi (personal communication, 1995). They tions and origins, (3) make a preliminary discovered a bubbling spring (site A-l) at evaluation of the potential hazards of de the base of the ridge. They also collected gassing to visitors, and (4) initiate study of water samples from the lake (15-16°C) in the impact of degassing on the local the East Ukinrek Maar. aquatic and onshore environments. Figure 3. Photograph showing collection of bulk-gas samples. Site is an offshore seep (site GR-1B) just off The Gas Rocks Peninsula. The pool to the left of the samplers is the onshore spring (site GR-1A). FIELD OBSERVATIONS AND circular upwellings up to one meter in di ameter (Fig. 2) and from hundreds of SAMPLING smaller vents that discharge single-bubble- width streams of gas. The onshore spring We investigated gas seeps and (16°C) fills a small (0.5 meter wide) pool springs in the vicinity of The Gas Rocks and contains reddish brown deposits of and along the ridge that hosts the Ukinrek iron oxyhydroxide (Fig. 3). Since the visits Maars (Fig. 1; Table 2). At The Gas Rocks, by Motyka et al. (1993) in 1981, the num visible degassing occurs offshore of The ber of onshore springs has decreased from Gas Rocks peninsula; minute amounts of several to one and the hottest spring has gas also vent through the onshore spring. cooled by 23°C. We lack evidence (e.g., dead vegetation, At The Gas Rocks, we collected gas bubbling in pools) for additional onshore samples to determine bulk-gas composi degassing in the vicinity of The Gas Rocks, tion from one of the smaller vents (site GR- but diffuse onshore degassing would be 1B; Fig. 3) and from the largest vent (site detected best by soil-gas studies (e.g., Far- GR-1C; Fig 2). Samples were collected fol rar et al., 1995; Gerlach et al., 1996). Off lowing the procedures of Fahlquist and shore, gases discharge from three large Janik (1992). A 15-cm funnel was held vents that disrupt lake-surface areas in beneath the water surface over the bub Table 2. 1995 sites investigated in the vicinity of The Gas Rocks and Ukinrek Maars. Figure 1 shows site locations.
Site Site description latitude T(°C) Sampling and measurements longitude accomplished GR-1A onshore spring at The N57° 52.00' 16.3 5-9 Gas Rocks W156° 30.18'
GR-1B small offshore gas seep N57° 52.00' 10.1 1-3 at The Gas Rocks; near W156° 30.18' onshore spring
GR-1C Largest gas vent at The N57° 52.00' 10.1 1,5-9 Gas Rocks W156° 30.18'
A-1 spring on same ridge as N57° 50.53' 10.2 1,4-9 the Ukinrek Maars; gas W156° 29.82' bubbling into pool; dead ______vegetation observed___ UK-1 spring-fed steam on the N57° 50.39' 7.5 1-9 same ridge as the W156° 29.87' Ukinrek Maars; gas bubbling into stream; dead vegetation observed UK-2 8-inch diameter pool (no N57° 50.50' 12.3 1,4-9 outlet) on same ridge as W156° 29.45' the Ukinrek Maars; gas bubbling into pool; dead ______vegetation observed___ UK-3 site of dead vegetation N57° 50.36' none along the same ridge as W156° 28.72' the Ukinrek Maars; no spring observed EM-1 Lake in East Ukinrek N57°50.10' 12.0 4-9 Maar; no degassing W156° 30.84' observed in lake B-1 Becharof Lake near N57° 51.10' 9.5 4-9 camp W156° 32.07'
1. gas sample for chemical analysis, 2. gas sample for 813C in CO2 analysis, 3. gas sample for rare-gas isotope analysis, 4. field conductivity, 5. laboratory conductivity and pH, 6. water sample for anion analysis, 7. water sample for cation analysis, 8. water sample for 8D and 818O analysis, 9. water sample for 813C in DIG analysis.
bling vent. This funnel was attached to a and opened the bottle's inlet valve. Regu short-length of Tygon tubing that led to an lation of the valve prevented water from evacuated double-port sampling bottle entering the sampling line. Occasional partly filled with a known volume of 4N shutting of the valve and shaking the bot NaOH solution. After purging air from the tle facilitated CCh gas dissolution. Gas was sampling line with a hand pump attached collected until the flow rate ceased or pre to the outlet port, we clamped the outlet cipitates (probably NaaCOs) formed, usually Figure 4. Photograph of the site UK-2 along the ridge that hosts the Ukinrek Maars. Note dead vegetation around the pools of water, some of which are bubbling (see Fig. 5). The Gas Rocks are in the background. within 3-8 minutes. At site GR-1B, we also ene bottle for analysis of anions, alkalinity filled two evacuated flasks (without NaOH excluded. Another sample was filtered solution) for analysis of 513C in CO2 and (with 0.45 nm filters), acidified (to a pH of rare-gas isotopes. The rare-gas isotope 2), and stored in a 250-ml polyethylene flasks were made of Corning-1720 glass bottle for cation analysis. A third sample and fitted with high-vacuum stopcocks to was collected into a 60-ml glass bottle with prevent leakage of He. a Polyseal closure for analysis of 8D and Oxygen concentration was below 618O in water. A fourth sample was col the analytical detection of our OQ meter (< lected in a 250-ml glass bottle with a 3 volume percent) in gases discharged from Polyseal closure for analysis of alkalinity the offshore seeps (Fig. 3). These mea and 613C in DIG. Equipment failure pre surements were verified by the total gas vented field measurements of conductivity analyses (below). at The Gas Rocks sites and pH at all sites We also collected water samples investigated so these parameters were de from the onshore spring (site GR-1 A; Fig. 3) termined in the laboratory (conductivity, and from the lake at site GR-1C. At each pH of near-neutral waters) or by thermo- site, we collected four samples. One sam chemical techniques (pH of acidic samples; ple was collected into a 250-ml polyethyl see below). Figure 5. View of bubbling pool sampled at site UK-2.
In addition, we investigated three Ritchie experienced a sharp pepper-like gas seeps (sites A-l, UK-1,2; Fig. 4) on the burning when the sampling tube dis ridge that hosts the Ukinrek Maars. All charged a small amount of gas in front of three seeps vent through spring-fed water her nose. All three sampling sites (A-l; (Fig. 5). At these sites, we measured field UK-1,2) and at least one other area (site conductivity, collected bulk-gas-composi UK-3) along the ridge exhibit vegetation tion samples, and obtained water samples that has been killed recently (e.g., dying for analyses of anions, cations, 813C in DIG, alders and grass) and vegetation that has and 8D and 618O in water. At site UK-1, we been dead for some time (Figs. 4 and 6). also filled two evacuated flasks to analyze Also, some alders lost their leaves earlier in 613C in CO2 and rare-gas isotopes. Field the season, and were just budding again in evidence for high CO2 concentrations in mid-August. At some sites, there are ap these seep gases include below-detection parently two ages of dead vegetation in readings on our Oz meter at all three sites cluding large logs that probably represent and formation of precipitates in the gas trees killed during the 1977 eruption and collection bottles at two sites. Moreover, at several-cm-diameter twigs that appear too site UK-1, the gas extinguished a match fresh to have died in 1977. Examination of (not tested at other sites) and Beatrice 1983 infrared photos of the Ukinrek area Figure 6. Photograph of an area of dead vegetation along the ridge that hosts the Ukinrek Maars.
shows that many of the dead-vegetation were collected for analyses of anions, areas existed in 1983, which is consistent cations, 513C in DIG, and 5D and 5 18O in with the vegetation dying during the 1977 water. eruption. Thus, we conclude that one set On 4 March 1996 Bill Smoke, the of vegetation was killed in 1977, a second U.S. Fish and Wildlife Service pilot, flew set was killed some years after 1977, and a over the area to check for visible thermal third set suffered trauma in 1995. We anomalies. He noted that the lakes in both speculate that CO2 is killing the vegetation the East and West Ukinrek Maars were fro as has been observed recently at Mammoth zen and, although the area had little snow Mountain, CA (Farrar et al., 1995). These cover, he saw no sign of any thermal areas of dead vegetation suggest that there anomalies at any of the above sites. are significant emissions of diffuse CO2 in The Gas Rocks-Ukinrek Maars area. A soil ANALYTICAL METHODS gas survey is highly recommended as a fu ture investigation for this area (see below). Gases Finally, we measured conductivity and collected water samples from the lake The bulk-gas samples were ana in the East Maar (site EM-1; Fig. 7) and, for lyzed by Thermochem in Santa Rosa, CA, background information, a sample of for CO2, H2O, N2, O2, CH4, AT, total S (SO2 + Becharof Lake (site B-l). Again, samples Figure 7. View of East Ukinrek Maar looking west. Sampling site EM-1 is located on the lake shore below the low point in the crater rim.
H2S), H2, HC1, HF, and NH3. Before collec stream was passed through a conductivity tion, bottles were partly filled with a known detector. The amount of H2O in each volume of 4 N NaOH solution, evacuated to sample was determined by subtracting the <0.026 bar, and weighed. After collection, cumulative weights of non-H2O species the sample bottles were reweighed, and from the total weight gained by the sample their head space volumes and pressures bottles during collection. were measured. Gas chromatography was The helium isotope samples were used to determine the composition of the analyzed by Robert Poreda using methods nearly insoluble (in NaOH solution) head- described elsewhere (Poreda et al., 1992; space gases (N2, O2, CH4, AT, H2). The Poreda and Farley, 1992). The gases in the composition of the gases dissolved in the flasks were extracted on a high vacuum NaOH solution (CO2, total S, HC1, HF, NH3) line constructed of stainless steel and was determined by acidification followed by Coming- 1724 glass. After removal of gas chromatography (CO2), oxidation of all sulfur species to SO42' followed by ion vapor and CO^ at - 90c an^ -195C, re chromatography (SO42-), ion chromatogra spectively, the no n- condensable gases (He, phy (C1-, F-), and, for NHs, by a flow injec Ne, AT, N2» CH^.) were concentrated using a tion analysis where the NHs difiused across Toepler pump in series with a Hg-difiusion a teflon membrane and the separated pump. The amount of non-condensable
10 gas was measured using a calibrated gas each sample through a 0.45 )Am-pore filter splitter fitted with, a capacitance manome membrane. A precipitate consisting mostly ter. Gas ratios (N2, Ar, CH^ were analyzed of SrCOa and SrSO* was formed. The pu on a Dycor Quadrupole mass spectrometer rity of the precipitate was determined by fitted with a variable leak valve. The re reacting an aliquot of the precipitate with sults are combined with the capacitance phosphoric acid in an evacuated flask and manometer measurement to obtain gas measuring manometrically the volume of concentrations. Prior to isotope analyses, liberated CO2. The CO2 was then trans N2 and Q<2 were removed by reaction with ferred in a vessel to Carol KendalTs labora tory (U.S. Geological Survey, Menlo Park, Zr-Al alloy (SAES-ST707) and Ar and Ne CA) for mass spectrometric 513C analysis were adsorbed on activated charcoal at 77 K and at 40 K, respectively. Helium iso using a Finnigan MAT 251. By knowing tope ratios and concentrations were then the mass of precipitate and its purity for each sample, we calculated the alkalinity analyzed on a VG 5400 Rare Gas Mass prior to degassing. Using the methods of Spectrometer fitted with a Faraday cup and Reed and Spycher (1984), we were then a Johnston electron multiplier. The Mass Spectrometer was also used to measure able to calculate the pre-degassing pH of concentrations of argon, neon, and krypton the degassed samples. The water isotope samples were isotopes. The 2a errors in the ^He/^He submitted to Tyler Coplen's laboratory ratios are ±0.3% for the reported helium (U.S. Geological Survey, Reston, VA) to de isotope value. termine 5D and 518O in water. William Evans purified the carbon isotope samples using liquid nitrogen traps on a high-vacuum line. A Finnigan MAT RESULTS 251 was used to determine 813C in CO2. Chemical and Isotopic Compositions of Waters Gases
The anion samples were analyzed Bulk composition by Chemical and Mineralogical Services Table 3 lists the chemical analyses (CMS) in Salt Lake City, UT, for C1-, SO42-, of the bulk-gas samples collected in 1995 F-, and conductivity. Cl- and SO42' were from The Gas Rocks and Ukinrek Maars analyzed using colorimetric methods, and areas. For each sample, we report the F- was determined using an ion selective composition of the dry gas and Xg, which is electrode. The pH was also determined by the mole fraction of dry gas to dry gas + CMS in anion samples from, sites B-l rhO in the bulk sample. The Xg values (Becharof Lake) and EM-1 (east Ukinrek have large uncertainties because of the as Maar ), but degassing of CO2 from the other sociated errors of calculating the apparent anion samples prevented accurate deter FhO concentrations from small differences mination of the pre-degassing pH. The (< 0.9 g) between the weight gains of the cation samples were analyzed by CMS for samples and the weights of non-PhO spe major cations using atomic absorption cies. Reported dry-gas species include spectrophotometry; they also analyzed se C02, Na, 02, Ar, CH4, total S, HC1, HF, H2, lected samples for trace elements using and NHs. inductively coupled plasma mass spectros- The dry gas compositions contain copy. 89.5 - 99.5% C02, 1.0 - 7.9% Na, 0.2 - Cathy Janik analyzed the carbon 2.5% O2, 0.008 - 0.11% Ar, and, for the isotope samples for alkalinity and prepared Ukinrek-Maars-ridge samples, 0.3 - 2.1% these samples to determine 813C in DIG. CH4. Total S, HC1, HF, Ha, and NHs are Prior to opening the bottles, the samples below detection in all samples. The com were refrigerated to redissolve any de positions of the 1995 samples from The gassed CO2. Then, 10 m.L of saturated so Gas Rocks sites GR-1B and GR-1C (Table lution of NH4OH and SrCb was added to 3) are very similar to the 1977 samples
11 Table 3. Chemical compositions of gases in The Gas Rocks and Ukinrek Maars areas. Field numbers: 950803-1 950803-(3-4) 950803-9 950804-1 950805-1 950807-(6-7) Map site: GR-1B GR-1B GR-1C A-l UK-2 UK-1 Location: small offshore small offshore Becharof Lake spring on bubbling pool spring-fed gas seep at The gas seep at The near largest gas Ukinrek Maars on Ukinrek stream on Gas Rocks Gas Rocks vent Ridge Maars Ridge Ukinrek Maars ridge ~o" Date: August 3, 1995 August 3, 1995 August 3, 1995 August 4> 1995 August 5, 1995 August 7, 1995 Site temperature and Xg T°C 10.1 10.1 10.1 10.2 12.3 7.5 Xg 0.68 0.72 0.66 0.57 0.59 0.60 Chemical Species of Dry Gas (mol%) CO2 98.5 99.5 89.5 90.5 94.9 95.9 N2 0.992 0.356 7.86 6.58 3.67 2.91 02 0.534 0.171 2.50 0.719 0.785 0.901 AT 0.0235 0.00782 0.110 0.0492 0.042 1 0.044 CH4 < 1.1x10-3 <7.3xlO-4 <4.0xlO-3 2.12 0.598 0.265 Slbul <0.054 <0.028 <0.053 <0.074 <0.087 <0.086 HC1 <0.044 <0.023 <0.044 <0.064 O.069 <0.070 HF <0.021 <0.011 <0.021 O.030 <0.032 <0.033 H2 <2.7xlO-3 < 1.7x10-3 <9. 6x10-3 <3.7xlO-3 <4.4xlO-3 <4.6xO-3 NHa < 1.6x10-3 <8.5xKH < 1.6xlO-3 <2. 3x10-3 <2. SxlO-3 <2.4xlO-3 Selected ratios N2/Ar 42 46 71 134 87 66 N2/(02+Ar) 1.8 2.0 3.0 8.6 4.4 3.1 Carbon Isotope Composition 8«C -6.50 -6.38 Xg is mol fraction of total dry gas in total dry gas + H2O; Srotai = H2S + SO2 + any other sulfur species.
from The Gas Rocks (Table 1). This sug the water table in The Gas Rocks and gests that there has been little change in Ukinrek Maars area are virtually 100% discharged gas composition from The Gas CO2, except for the small amounts of non- Rocks area over this 18-year timespan. atmospheric N2 and CH4 in the Ukinrek However, the flux of CO2 from The Gas Maars samples. Rocks area has apparently decreased since To explain the non-atmospheric N2 1977 as indicated by post-1977 aerial ob and CH4 in the Ukinrek Maars ridge sam servations by Tom Miller (personal com ples we consider four possible sources: munication, 1996) and the decrease in the number of onshore springs and their de (1) Rotting of vegetation in the swampy gassing rate. ground around the vents. This seems an Ratios of N2/Ar versus N2/(O2+Ar) unlikely source for the excess N2 and CH4 show that most of the 1995 gas samples lie because there is only a thin layer (< 1 me on a mixing line between air-saturated wa ter) of organic debris in the swampy ter (ASW) and air (Fig. 8). This indicates ground around the Ukinrek Maars ridge that N2, O2, and Ar in most samples derive seeps. Also, swamp gases should contain from ASW and air, both of which are mete some H2S, but H2S is below detection oric in origin. But the trend towards Ng/Ar (<0.09%) in the Ukinrek Maars samples ratios greater than air in UK-2, UK-1 (in (Table 3). the He isotope sample; Table 4), and espe cially A-l indicates that the samples col (2) They are magmatic along with the CO2. lected on the Ukinrek Maars ridge contain As indicated below the CO2 probably comes some non-atmospheric N2. These same from a mantle-derived magma. However, samples also contain minor amounts (0.3 - the magma may also contain volatiles de 2%) of CH4. After accounting for the near- rived from the subducted slab. Thermal surface N2, O2, and Ar in the samples, we decomposition of subducted organic mate infer that the gases emanating from below rial is commonly invoked to explain the
12 140 A-1 +
100
+ UK-2 Air A N2/Ar / GR-1C-I/
60 -
GR-1B '
GR-1B+
ASW(11 C)
20 4 6 10 N2/(O2+Ar)
Figure 8. Plot of N2/Ar versus N2/(O2 + Ar) for the bulk gas samples from The Gas Rocks and Ukinrek Maars. All data from Table 3. elevated N2 (e.g., N2/Ar > 84) in gases from which is driven to the right at low tempera convergent plate volcanoes (Matsuo et al., ture (<200°C), moderate pressure (> 100 1978; Giggenbach and Gougel, 1989). bars), and under reducing conditions. But However, the subducted slab is an implau such CH4-producing reactions are gener sible source for the elevated CH* in the ally much slower than the timeframe of Ukinrek Maars gases because magmatic volcanic degassing (Giggenbach, 1987). gases from convergent-plate volcanoes Another pitfall of the idea that N2 and CH* generally do not contain much CH4 are magmatic is that, assuming that The (Symonds et al., 1994). It is possible that Gas Rocks and Ukinrek Maars gases come magmatic CO2 reacts with H2O to produce from the same magmatic source, this idea CH4 in the following reaction: fails to account for the much lower concen trations of N2 and CH4 in The Gas Rocks C02 + 2H20 = CH4 + 2O2, (1) gases.
13 Table 4. Dry-gas and rare-gas compositions of rare-gas isotope samples from The Gas Rocks and Ukinrek Maars areas. Field number: Poreda and Craig 950803-2 950807-8 (1989); Motykaetal. (1993) Map site: GR-1A GR-1B UK-1 Location: The Gas Rocks spring small offshore gas spring-fed steam on seep at The Gas Ukinrek Maar ridge Rocks Date: July 2 1,1981 Augusts, 1995 August 7, 1995 T'C 39 10. 1 7.5 Bulk composition (mol%) CO3 + H3St 99.99 99.96 99.78 N3 0.0114 0.034 0.187 CH4 0.00045 0.00053 0.0282 03 0.0026 0.00053 0.00055 Rare gases (ppbv) «°Ar 2600 11200 18400 1000 72 11990 36.9 51.9 Ne 8.1 6.7 2.48 1.02 Rare gas ratios 6.91 6.42 6.55 He 1 Ne 114 31 6224 (He 1 Ne) A
6.95 6.61 6.55 Ne/^Ar ... 0.22 0.13 N3/Ar 44 30 101 N3/(03+Ar) 4.0 21 78 He/Ne - 9 1792 «>Ar/36Ar 304 355 4He/«°Ar ... 0.21 3.84 ^Represents the portion of the anydrous gas (COa + HaS) that was condensed at - 195°C during the analysis. Mostly consists of CO3 + HaS.
(3) They come from a subsurface hydrocar tary rocks might also produce NHa, but this bon reservoir. We consider this possible is probably scrubbed by groundwater en because sedimentary gases in the vicinity route to the surface. Assuming that HaO is of Becharof Lake contain minor amounts also scrubbed by groundwater, mixing of a (5-20%) of N2 and CH4 along with 70-80% small fraction of the resulting COz-CHi-Na COa (Bureau of Mines, 1976). gases with magmatic-meteoric gases (e.g. similar to those discharged from The Gas (4) Thermal breakdown of organic residue Rocks) can explain the composition of the in the sedimentary sequence beneath the Ukinrek Maars gases. We prefer this ex Ukinrek Maars. Magma emplaced most planation because it links CO2, Na, and likely before or during the 1977 eruption CH« to the underlying magma and its could heat the surrounding sedimentary thermally heated country rock. One poten rocks. Thermal decomposition of these tial pitfall, however, is that mixing between sedimentary rocks would probably produce magmatic and sedimentary-derived CO? is a gas dominated by HaO, CCh, CHt, and NQ, not substantiated by the 513C in CO2 or the latter demonstrated experimentally by 813C in DIG data for the Ukinrek Maars Matsuo et al. (1978). Heating of sedimen sites with the exception of the 813C in DIG
14 for the UK-1 site (see below). However, (3He/4He, Ne/36Ar, N2/Ar, He/Ne, this is a potential problem with the other 40Ar/36Ar, 4He/40Ar). All 3He/4He ratios are proposed sedimentary sources of N2 and reported as R/R\ values where R is the ra CH4, all of which should produce isotopi- tio in the sample and RA is the atmospheric cally light CO2 (e.g., 5 13C < -20). ratio, using air helium as the absolute standard. We also report 3He/4He ratios Carbon isotopes corrected for atmospheric contamination Table 3 also reports the 8 13C values (RC/RA values), using the equation after for CO2 in the carbon isotope samples col Craig et al. (1978a,b): lected from sites GR-1B (-6.50) and UK-1 (-6.38). These values are virtually identical = RN-RANA with the 1977 results (-6.36; Table 1) indi (2) cating that the source of the CO2 emissions ° N-N A at The Gas Rocks and Ukinrek Maars has not changed over this 18-year period. Po where R is 3He/4He, N is He/Ne, and sub tential sources for this CO-2 include ther script "A* denotes atmospheric ratios. mal decomposition of limestone or organic The 3He/4He (RC/RA) values for the material from underlying sedimentary 1995 samples collected from The Gas rocks, the atmosphere, or the mantle (or Rocks (site GR-1B; 6.61) and the Ukinrek mantle-derived magma). Thermal break Maars ridge (site UK-1; 6.55) are virtually down of limestone or organic material identical, indicating that the He in gas dis seems implausible because 513C values for charges from, both sites comes from a nearby limestone (+3.860/00; Barnes and similar source. Assuming that the mantle McCoy, 1979) or for biogenic material (5 13C has a 3He/4He ratio of 8 and the crust, rich = -24 to -34°/oo for most terrestrial plants; in radiogenic 4He, has a ratio of 0.02 (e.g., Smith and Epstein, 1971) are much higher Poreda and Craig, 1989), the 1995 samples and lower, respectively, than the observed contain 82-83% mantle He. Table 4 shows 5 13C values. The observed 5 13C values are that the 1981 sample from The Gas Rocks identical to the -6 to -7 °/oo values for at (site GR-1A) has a slightly higher 3He/4He mospheric CO2 (Anderson and Arthur, of 6.95, implying an 87% mantle He com 1983), but the atmosphere seems an un ponent. The 4-5% decrease in mantle He likely source because it requires provoking at The Gas Rocks between 1981 and 1995 a mechanism to concentrate CO2 from. implies that the 3He flux from the source -0.0360% in the atmosphere to over 99% magma has decreased since 1981. in the discharged gases. Instead, we con cur with Bames and McCoy (1979) that Chemical and Isotopic Compositions of CO2 from The Gas Rocks and Ukinrek Waters Maars derives from the mantle because ob served 813C values are within the range (-5 Bulk composition to -8) for mantle-derived CO2 (Allard, 1983; Analyses of the water samples col Taylor, 1986). Furthermore, a mantle lected in 1995 from The Gas Rocks and source is consistent with the geochemical Ukinrek Maars area are reported in Table arguments for mantle-derived magmas 5. Data are given for the site temperature, erupted from the Ukinrek Maars (Kienle et calculated (or laboratory determined) pH al., 1980). (see above), conductivity, major and minor anions and cations, 8D and 818O in H2O, Rare-gas isotopes and 813C in DIG. Table 6 reports trace ele Analytical results for the rare-gas ment analyses of samples collected from isotope samples are reported in Table 4. The Gas Rocks spring (site GR-1 A) and the For each sample, we report the bulk com East Ukinrek Maar Lake (site EM-1). For position (CO2 + KhS, Na, CH4, 02), the con comparison, Table 7 reports previous centrations of rare-gas isotopes (40Ar, 4He, (1977, 1980, 1981, 1994) chemical and iso- 36Ar, Ne, 84Kr), and selected rare-gas ratios topic analyses of waters from several sites
15 Table 5 - Chemical and iso topic compositions of waters in The Gas Rocks and Ukinrek Maars areas. Field numbers*: 950803-(5-8) 950803-(10-13) 950804-(2-6) 950807-(l-5) 950805-(2-5) 950807-(9-13) 950808-(l-4) Map site: GR-1A GR-1C A-l UK-1 UK-2 EM-1 B-l Location: The Gas Rocks Becharof Lake spring on Ukinrek spring-fed stream bubbling pool on Lake, east Becharof Lake spring near largest gas Maars Ridge on Ukinrek Maars Ukinrek Maars Ukinrek Maar vent Ridge Ridge Date: August 3, 1995 August 3, 1995 August 4, 1995 August 7, 1995 August 5, 1995 August 7, 1995 Augusts, 1995 Conditions T('C) 16.3 10.1 10.2 7.5 12.3 12.5 9.5 pH* 5.22 5.58 5.29 5.72 5.72 7.45* 6.45* field cond. (fiS) ND ND 255 144 560 791 106 lab cond. (fiS) 7320 215 211 132 528 715 95 Chemical Species (mg/1) Na 650 40 3.2 5.7 12 39 3.5 K 49 1.7 2.3 1.7 3.8 3.2 0.9 Li 2.9 0.08 0.01 0.01 0.02 0.02 0.02 Ca 400 11 25 12 63 82 9.6 Mg 203 4.3 9.6 3.6 28 19 2.3 Fe 48 0.14 0.87 0.12 0.08 0.06 0.05 Mn 3.3 0.05 0.36 0.11 0.09 0.03 0.06 Al 0.3 <0.1 <0.1 <0. 1 <0.1 <0.1 <0.1 Cl 2300 55 <5 <5 <5 88 22 SO* 18 15 11 10 <10 75 <10 DIG* 1880 320 1420 295 1730 135 24 NH4 (as N) 2.6 ND 0.15 0.05 <0.05 <0.05 <0.05 SiO2 120 12 64 40 48 34 1.3 B 23 1 0.3 0.2 0.3 0.8 0.2 hydrogen and oxygen isotope composition (°/oo relative to SMOW) 8D -80.7 -81.1 -91.6 -91.8 -83.6 -82.4 ND 818O -10.79 -10.69 -12.19 -12.36 -11.14 -10.82 ND carbon isotope composition of DIG (°/oo relative to PDB) 813C -3.6 -6.9 -6.7 -8.0 -5.1 -9.2 -3.2 ^Several sub-samples were collected from each site to complete the requisite analyses. *pH determined numerically at the collection temperature using program CHILLER (Reed, 1982) *pH determined by electrode in the laboratory at 20°C "dissolved inorganic carbon (DIG) reported as HCOa ND =not determined Species below detection in all samples include HzS (< 0.01) and F (< 0.2). Table 6. Trace element analyses of waters ple) DIG, Ca, Mg, K, B, SO4, SiO2, Fe, Br, from The Gas Rocks and Ukinrek Maars Li, Sr, As, NH4, Ba, and I (Tables 5-7). areas. Although not all the gas seeps dis Field number: 950803-7 950807-10 charge saline waters, the saline springs al Map site: GR-1A EM-1 ways discharge gas. It seems likely that Location: The Gas Rocks East Ukinrek the acending CCb-rich gas plays a role in spring Maar lake transporting the saline waters to the sur Date: Augusts, 1995 August 7, 1995 Element (pg/1) face. Model equilibrium calculations show V 3 15 that the 1995 GR-1A sample is close to Cr 9 4 saturation with respect to CO2 gas. Al Co 35 <1 though some past samples (Table 7) of sa Ni 32 4 line waters from The Gas Rocks area are Cu 30 2 undersaturated with respect to CCb, we Zn 70 5 As 2876 62 think it likely that this probably reflects Se <5 13 improper determination of DIG rather than Br 19050 929 true undersaturation of CO2. Therefore, we Rb 215 5 suggest that gas rising through the Sr 9376 270 subsurface brine reservoir saturates the Zr 2 2 brine with gas, which decreases its density Mo 2 2 and enables the brine to ascend to the Cd 2 <1 Te <1 1 surface. I 293 27 The most saline water sample was Cs 42 <1 collected in 1977 from The Gas Rocks Ba 414 25 spring. This sample has similar salinities Ce 3 <1 and elemental ratios to sedimentary forma Sm 1 <1 tion waters (e.g., oil-field brines; Fig. 10; Hf 2 3 W 1 <1 White et al., 1963). But compared to Au 1 1 seawater, the 1977 sample has 1.5 times Tl 1 <1 the salinity, lower Mg/Ca and SO4/C1 ra Pb 1 <1 tios, and higher Li/Na, I/C1, and B/C1 ra Bi <1 1 tios (Fig. 10). However, simple heating of Elements below detection in all samples include Ti (<1), seawater to 300°C lowers Mg/Ca and Ag (<1), Sn (<1), Sb (<1), La (<1), Pr (<1), Nd (<1), Eu (<1), Gd (<1), Dy (<1), Er (<1), Yb (<1), Re (<1), and Hg SO4/C1 due to precipitation of anhydrite (CaSO4) and magnesium oxysulfate (Bischoff and Seyfried, 1978). So the brine in The Gas Rocks and Ukinrek Maars area. could represent concentrated seawater that The waters in the vicinity of The has reacted with underlying sedimentary Gas Rocks and Ukinrek Maars can be clas rocks at some elevated temperature, as sified as gas-fed waters, waters containing suming that this process could also ac a brine component, and dilute meteoric count for the elevated salinity, Li/Na, I/C1, waters (Fig. 9). The gas-fed waters are and B/C1 of the brine. The brine also dominated by DIG (mostly HbCOs but re could have formed by reaction of meteoric ported as HCO3 in Fig. 9 and Tables 5, 7) water with the underlying sedimentary obtained from the CCb-rich gases that dis rocks (see below). Regardless of the ulti charge through them. These include wa mate source of the water, the above obser ters collected in the proximity of the gas vations support the conclusions of Bames vents in Becharof Lake and waters from the and McCoy (1979) that the brine equili cold gas seeps on Ukinrek Maars ridge. In brated with the underlying Mesozoic ma contrast, waters from The Gas Rocks rine sedimentary rocks. spring and the West Ukinrek Maar pool Between 1977 and 1995, the tem contain a major NaCl brine component. perature and brine component in the hot Other constituents in this brine include (in test Gas Rocks spring have decreased dra decreasing abundance in the GR-1A sam matically (Tables 5,7) along with the num
17 Table 7. Chemical and isotopic compositions of waters collected in 1977, 1980, 1981, and 1994 from The Gas Rocks and Ukinrek Maars areas. Reference: Barnes and Motyka et al. Barnes and McCoy Barnes and Motyka et al. C. NealandM. Barnes and McCoy (1979) (1993) (1979) McCoy (1979) (1993) Thompson (pers. McCoy (1979) comm. 1996) Map site: GR-1A GR-1A GR-1C EM-1 Location: Gas Rocks spring Gas Rocks spring Becharof Lake near Pool, west Pool, west Lake, east Becharof Lake largest gas vent Ukinrek Maar Ukinrek Maar Ukinrek Maar Dale: August 25, 1977 August 17, 1981 August 25, 1977 August 24, 1977 August 24, 1980 August 6, 1994 April 20, 1977 Conditions T(°C) 52.8 39 9.3 80.9 16 14.7 ND field pH 5.9 5.7 4.4 6.3 9.0 6.9t 6.9t Chemical Species (mg/1) Na 17700 13900 7.1 3600 190 39 4.9 K 450 410 0.5 400 11 4.6 0.47 Li 44 42 <0.01 2.5 0.01 0.012 ND Ca 1500 1430 7.2 380 30 91 6.2 Mg 460 420 1.6 7.6 6.9 15 1.2 Fe ND 19 ND 0.13 0.07 0.13 ND Al ND ND ND ND ND ND ND Cl 32000 26100 10 6300 210 103 6 00 F 0.13 0.1 <0.1 0.72 0.6 0.54 <0.1 SO4 140 134 13 370 100 96 6 H2S 1.7 ND <0.1 1.2 ND ND ND field HCC-3 1240 72 19 28 199 124t 16t NH4 (as N) 13 ND <0.05 13 ND ND <0.1 SiOa 120 110 2 275 29 22 1 B 360 3.4 <0.1 33 1.8 0.13 0.1 Trace elements (ng/1) Br 92000 ND ND ND ND ND ND Rb 2100 ND <20 550 ND ND ND Sr ND 87000 ND ND 0.10 ND ND 1 5000 2600 ND 1000 ND ND ND Cs 3600 ND <100 100 ND ND ND hydrogen and oxygen isotope composition (°/
X X
gas-fed waters
SO4 HCO3
Figure 9. Ternary diagram of Cl, SO4, and HCOs for selected waters from The Gas Rocks and Ukinrek Maars area. Data from Tables 5 and 7.
10-a
oil-field brines Gas Rocks spring ('77) seawater
0.000001 Ca/Na Mg/Ca K/Na Li/Na HCO3/CI SO4/CI F/CI Br/CI I/CI B/CI
Figure 1O. Plot showing selected elemental ratios for the 1977 Gas Rock spring waters (Table 7). For comparison, we plot seawater and the range of oil-field brines from White et al. (1963).
19 1000-13 1977
100-=
O" I o O 0.1 -= Gas Rocks spring Becharof Lake 0.01 \ I Si02 K Mg Ca Na Cl SO4 HCO3 Figure 11. Modified Schoeller plot of major constituent concentrations for The Gas Rocks spring waters in 1977, 1981, and 1995. For comparison, we plot the 1995 water sample from Becharof Lake (Table 5). All ionic species are in meq/L and SiC>2 is in mmol/L. her of onshore springs. Figure 11 shows Maar lake are dominated by local meteoric the decreasing concentrations of the major waters. In contrast, the 1977 and 1981 components (Cl, Na, Ca, Mg, K, SO4) in the samples from the hottest Gas Rocks spring brine over this timespan. These decreases are greatly enriched in deuterium and 18O reflect dilution of the brine by meteoric wa relative to local meteoric water, which re ter. We suggest that the declining gas flux flects the isotopic signature of the large at The Gas Rocks has caused a brine component in these samples. The corresponding decrease in the rate that the 1977 Gas Rocks sample displays the high brine ascends to the surface. est enrichments in deuterium and 18O, and the largest "chemical" brine component Hydrogen and oxygen isotopes (Table 7). The 1980 sample from the West Figure 12 shows the 5D and 818O Maar pool is also enriched in deuterium compositions of water samples from The and 18O, but this may reflect evaporative Gas Rocks and Ukinrek Maars area. The enrichment rather than a brine signature samples from the 1995 cold gas seeps, the because this dilute sample has a low Cl 1995 Gas Rocks spring, and the 1995 East concentration (210 ppm; Table 7). Finally, 20 10 'seawater brine -10- MW AW -30- G. R. spring '95 n -50- E. Maar lake '95 PMW 8D -] mixing -70- -90- + 77 hydrothermal water-rock Gas Rocks spring exchange n -110- x E. Maar lake other local waters W. Maar pool -130 1 I I -16 -8 818o 0 8 Figure 12. Plot of 5D versus 818O for waters collected in the vicinity of The Gas Rocks and Ukinrek Maars. All data are from Tables 5 and 7 in °/oo relative to Standard Mean Ocean Water (SMOW; Craig, 196 la). Shown for reference are SMOW, the meteoric water line (MW) of Craig (1961b), the field of primary magmatic water (PMW; Sheppard et al., 1969), and the field for andesitic water (AW; Giggenbach, 1992) that represents high-temperature volcanic gases from convergent-plate volcanic gases. The vectors indicate the 6D-518O trends expected for hydrothermal water-rock exchange and for mixing of the 1977 Gas Rocks spring waters with local meteoric waters. the 1977 sample from the West Maar pool the 1977 West Maar pool are also sup is enriched in 18O, probably due to hydro- ported by the pool's large amount of dis thermal water-rock exchange. Significant solved silica (275 ppm; Table 7). Clearly hydrothermal water-rock interactions in the 81°C temperature of the pool in 1977 21 enhanced water-rock interactions. charged gases (S13C = -6.4 to -6.5°/oo; Ta Relative to local meteoric waters, bles 1,3), (2) biogenic CO2 (S^C = -24 to - the 1977 waters from the hottest Gas 34°/oo for most terrestrial plants; Smith Rocks spring are enriched by 40-50°/oo in and Epstein, 1971), and (3) atmospheric deuterium and 10-12°/oo in 18O. Such en CO2 (813C = -6 to -7°/oo; Anderson and Ar richments are fairly typical for sedimentary thur, 1983). The nearly equal molar pro brines (e.g., oil-field brines from the Al portions of Ca plus Mg to DIC in waters berta, Gulf Coast, Illinois, and Michigan from sites GR-1A, EM-1, and B-l suggest basins and oil-field waters from California; that dissolution of carbonate mineral^ from Clayton et al., 1966; White et al., 1973). To underlying sedimentary rocks (813C for account for the 5D and 518O enrichment in nearby limestone is +3.86°/oo; Barnes and The Gas Rocks brine, we consider two ex McCoy, 1979) is another potential source planations: (1) The brine is a mixture of a of DIC in these samples. In addition to the more saline, 18O-enriched seawater brine source of carbon, the S13C in DIC values and local meteoric waters. Relative to the are affected by fractionation between the 1977 brine, standard seawater (Craig, source and the aqueous carbonate species. 196la) is less saline (see above), is en The main aqueous carbonate species in riched in deuterium by 38°/oo, and has The Gas Rocks and Ukinrek Maars waters similar 818O values (Fig. 12). However, re should be H2CO3 at pH < 6.4 and HCO3 at action of seawater with the underlying pH = 6.4 - 10.3 (Drever, 1981). sedimentary rocks could produce an 18O- To discern the origin of DIC in the enriched seawater brine (Fig. 12) that 1995 water samples, we predicted their could mix with local meteoric waters to hypothetical 813C in DIC values for the po produce the 1977 brine. Potential sources tential sources (above) using the estimated for the seawater include modern seawater 813C values for each source (above), the from the Pacific (25-30 miles away) or con calculated relative abundances of HiCOs nate seawater; the latter source suggested and HCOa, and the appropriate tempera by Hitchon and Friedman (1969) to explain ture-dependent fractionation factors for the origin of oil-field brines in western H2CO3(aq)-CO2(g), HCO3(aq)-CO2(aq), H2CO3(aqj- Canada. (2) The brine is meteoric water CaCO3, and HCO3(aq)-CaCO3 modified from that has reacted with clays and calcite in Deines et al. (1974). The equilibrium dis the underlying sedimentary rocks to pro tribution of foCOa and HCOa was deter duce the observed enrichments in 8D and mined using program CHILLER (Reed, 818O. Isotope exchange reactions with hy 1982) for the temperature, pH, and compo drous minerals (clays, gypsum) and calcite sition of the water samples (Table 5). We have been suggested as a mechanism to used the ambient collection temperature produce the observed 8D and S18O enrich for all calculations, except to model the ments in several oil-field brines (Clayton et dissolution of carbonates in the GR-1A al., 1966). sample, we used the Na-K-Ca equilibrium Unfortunately, we lack sufficient temperature (162°C; Fournier and Trues- data to distinguish between the above ex dell, 1973). planations for the origin of The Gas Rocks The predicted S13C in DIC values for brine. Nonetheless, the observed isotopic the potential sources of DIC in the 1995 (Fig. 12) and chemical trends in the 1977- water samples are shown in Table 8. For 1995 Gas Rocks spring waters are clearly sites GR-1C, A-l, and UK-2, there is close produced by mixing of local meteoric water agreement between the observed 813C in with a brine component. DIC values and those predicted for a dis charged-gas origin, suggesting that DIC in Carbon isotopes these waters comes mostly from mantle- The S 13C values for DIG in the 1995 derived CO2. The observed S13C in DIC for water samples range from -3.2 to -9.2°/oo these samples could also be explained by (Table 5). Potential sources for DIC in an atmospheric source, but this seems these waters include (1) CO2 from the dis very unlikely due to the large absolute 22 Table 8. Predicted 8 13C in DIG values for potential sources of DIG in water samples from The Gas Rocks and Ukinrek Maars area (see text). For comparison, we report the observed 5 13C in DIG values from Table 5. Site Observed 813C Predicted 813C in DIG (°/oo relative to PDB) for hypothetical sources in DIG (°/oo) discharged gases local carbonates biogenic COa atmospheric COa {8i3C--6.4--6.5%o) (813C - +3.86 °/oo) (513C--24- -34 % ) (8«C - -6 - -7°/oo) GR-1A -3.6 -6.4 - -6.5 +0.5t -24 - -36 -6.0 - -7.0 GR -6.9 -5.7 - -5.8 NA -23 - -35 -5.3 - -6.3 IG A-l -6.7 -6.4 - 6.5 NA -24 - -36 -6.0 - -7.0 UK-1 -8.0 -5.2 - -5.3 NA -23 - -35 -4.8 - -5.8 UK-2 -5.1 -5.2 - -5.3 NA -23 - -35 -4.8 - -5.8 EM-1 -9.2 + 1.8 - +1.7 + 1.0 -16 - -28 +2.2 - +1.2 B-l -3.2 -1.5- -1.6 -2.6 -19 - -31 -1.1 - -2.1 Estimated at the equilibrium temperature of 162°C (see text); NA = not applicable. concentrations of DIG in these waters CO2 at The Gas Rocks and the elevated (Table 5). The waters from, sites UK-1 and CO2 discharge along the Ukinrek Maars EM-1 have lower-than-mantle 813C in DIG ridge raise concerns about CO2 hazards to values, implying that DIG in these waters potential visitors. may be mixtures of mantle and biogenic CO2 concentrations of > 30% cause components. In contrast, the heavier- rapid unconsciousness and, if sustained than-mantle 813C value for the GR-1A for several minutes, death in humans sample suggests that DIG in this sample is (Stupfel and Le Guern, 1989). However, a mixture of DIG from the mantle and DIG when CO2 is discharged into the atmos from a brine endmember. The 813C in DIG phere, it normally dissipates rapidly. values for ambient Becharof Lake water Therefore, the main hazard to humans (site B-l) is consistent with admixtures of a generally occurs when CO2, which is heav heavier source (mantle, carbonates, at ier than air, concentrates in depressions, mosphere) with lighter biogenic CO-2. snow, or manmade structures. During Given the low concentration of DIG (24 field study, none of the sites at The Gas mg/L) in the B-l water, consistent with an Rocks or Ukinrek Maars presented direct atmospheric source, we favor the atmos asphyxiation hazards as tested by measur phere as the source of the heavier DIG. If ing the Os concentrations in ambient air air supplies the heavy DIG, then only a around the vents. We developed no symp small fraction (5-10%) of biogenic DIG is toms of asphyxia during field study, nor required to account for the observed DIG did we observe any dead animals, which for the B-l sample. are sometimes seen at sites of lethal CO2 discharge (e.g., "Valley of Death* in Indo GAS HAZARDS IMPLICA nesia, Loudon, 1832; Hekla volcano in Iceland, Thorarinson, 1950). This mostly TIONS reflects the open-air geometry of these sites, but the lower post-1977 CCh flux at This study shows that CO2-rich The Gas Rocks also helps miriimize as gases continue to discharge near The Gas phyxiation hazards as experienced by Tom Rocks, although the current CO2 flux ap Miller in 1977. Thus, our preliminary ob pears lower than just after the maars servations suggest that manmade enclo erupted in 1977. COs-rich gases also dis sures near the gas seeps might be the charge from at least four Ukin main gas hazard in the vicinity of The Gas rek-Maars-ridge sites (Table 2), all of which Rocks and Ukinrek Maars. In particular, display newly killed and dying vegetation we recommend against tent camping and indicating recent increases in soil-gas CCh digging snow pits near the areas of gas dis concentrations. The continued release of charge, including sites of dead vegetation 23 on the Ukinrek Maars ridge, because tents assess the overall impact of the CO? emis and snow pits can concentrate CO?. We sions on the aquatic ecosystem, we suggest also recommend that visitors exercise cau a water-composition survey to help define tion around all gas seeps because we do the spatial extent of the DIG anomaly. not know how static the degassing is, and The extent of the saline-water dis the flux could change at any time as indi charges and their impact on the Becharof cated by the evidence of recent vegetation Lake ecosystems needs further study. Sa damage and mortality in the areas of CO? line waters discharge into Becharof Lake discharge. from The Gas Rocks spring. As discussed above, the major to trace constituents in IMPACT ON LOCAL ECOSYS the saline Gas Rocks spring water pres ently include (in decreasing abundance) Cl, TEMS DIG, Na, Ca, Mg, SiO2, K, Fe, B, Br, SO4, Sr, Li, As, NH4, Ba, I, and Rb (Tables 5 and This study shows that CO?-nch 6). The impact of these saline-water dis gases and saline water discharges continue charges on the Becharof Lake ecosystem to be anomalous geochemical inputs into depends on their flux into the lake. During the aquatic and onshore ecosystems in the time of field study, The Gas Rocks spring vicinity of The Gas Rocks and Ukinrek (site GR-1A) discharged a seemingly insig Maars. The CCh-rich gases discharge from nificant amount of saline water (< 1 li offshore seeps in the vicinity of The Gas ter/minute). However, additional saline Rocks and from several seeps along the springs may exist on the lake floor along Ukinrek Maars ridge. There are also dif the Bruin Bay fault zone that extends fuse COs emissions along the Ukinrek across the lake (Detterman et al., 1987). Maars ridge as indicated by the areas of For instance, subsurface springs probably dead and dying vegetation. These vegeta exist near the largest gas vent (site GR-1C) tion-kill areas are the most obvious impact because the water sample collected from of the COs emissions on the local ecosys this site contains anomalous concentra tems. We expect that soil gases in these tions (>2 times those of the background B- vegetation-kill zones contain high (>20- 1 sample) of several brine components in 30%) CO? concentrations based on soil-gas cluding Cl, Na, SC»4, and SiO2 (Table 5). sampling by Farrar et al. (1995) in the tree- Again, a detailed water-composition survey kill areas at Mammoth Mountain, CA. We could help define the spatial extent of the encourage soil-gas studies at Ukinrek saline-water discharges. Maars ridge to help define the exact CCh concentrations in the kill zones, as well as the area! extent and flux of the diffuse CO? CONCLUSIONS emissions. A complete soil-gas survey us ing the LI-COR COs analyzer and the This 1995 study confirms that CO?- methods of Gerlach et al. (1996) would be rich gases continue to discharge from three an ideal way to conduct such a study. large vents and hundreds of smaller seeps A less obvious impact of the gas in the vicinity of The Gas Rocks, as they discharges is the elevated DIG content of have since at least 1974. CO? also dis lake water in the vicinity of The Gas Rocks. charges from three seeps along the Ukin Compared to the background Becharof rek Maars ridge. In addition, CC«2 degasses Lake water sample (site B-l), the 1995 wa diffusely along the ridge as indicated by ter sample collected near the largest gas areas of dead and dying vegetation at all vent (site GR-1C) contains 13-times more three seeps and at least one other ridge DIG and has a 0.9-unit-lower pH (Table 5). site. The discharged gases from The Gas Clearly some of the discharged CO? dis Rocks and Ukinrek Maars ridge contain solves in and acidifies lake water in and nearly 100% CC«2, excluding minor around The Gas Rocks. However, we do amounts of air and ASW gases (N?, O?, Ar) not know the extent of the elevated DIG in all samples and thermogenic CH4 and N? and H+ contents in lake waters. To help in the Ukinrek Maars ridge gases. The 513C 24 and 3He/4He isotopic data suggest that Na, SO4, SiO2, and H* concentrations. We CO2 and He in these gases derive from, a suggest a lake-water-composition survey to mantle source help define the spatial extent of these geo- The main gas hazards to visitors in chemical anomalies. the vicinity of The Gas Rocks and Ukinrek Maars are tent camping and digging snow ACKNOWLEDGMENTS pits near the areas of gas discharge, in cluding the areas of dead and dying vegeta This project grew in part out of tion along the Ukinrek Maars ridge. Such AVO's invited participation in a enclosures may concentrate CO2 to lethal USFWS-sponsored meeting of the Bristol levels. The potential for open-air CCb as Bay/Kodiak Ecosystem TEAM. This work phyxiation at The Gas Rocks and Ukinrek was viewed as part of a baseline inventory Maars has decreased since Tom Miller ex of the Becharof Lake Ecosystem under the perienced asphyxia at The Gas Rocks in leadership of USFWS in King Salmon. 1977. Apparently, the CO2 flux at The Gas Funding for this project was provided by Rocks has declined since 1977 probably the USGS Volcano Hazards Program, the due to outgassing of the CO2 source. How USFWS Becharof Lake Ecosystem Project, ever, the potential for open-air CCh as the USFWS King Salmon Fishery Re phyxiation could potentially change at any sources Office, and Northern Arizona Uni time should mantle CO2 input increase. versity. Ron Hood, Bill Smoke, and others Besides the CO2 seeps, saline wa at the USFWS helped make our field work ters discharge from an onshore spring on successful and safe. Tina Neal played an The Gas Rocks peninsula. This spring also instrumental role in soliciting funds from discharges CCh-rich gas, which helps USFWS, coordinating logistics, and review transport the brine to the surface. The wa ing an earlier version of this work. An ad ters derive their salinity from a NaCl-brine ditional review by Jake Lowenstern helped component. Additional saline waters may clarify discussions on the potential origins discharge offshore from The Gas Rocks of the gases and brines. peninsula, perhaps along the Bruin Bay fault zone. The temperature and brine component of the spring have decreased REFERENCES CITED significantly since first measured in 1977, probably due to the corresponding de Allard, P., 1983, The origin of hydrogen, crease in the gas flux in The Gas Rocks carbon, sulfur, nitrogen, and rare area. The brine is similar to sedimentary gases in volcanic exhalations: evi brines and apparently comes from the un dence from isotope geochemistry, in derlying marine Mesozoic sedimentary Tazieff, H., and Sabroux, J. C., eds., rocks. The brine probably originates from Forecasting Volcanic Events: Am seawater or meteoric waters that reacted sterdam, Elsevier, p. 337-386. with the underlying marine sedimentary Anderson, T. F., and Arthur, M. A., 1983, rocks. Stable isotopes of oxygen and car The main effect of the CO2 gas bon and their application to sedi- emissions on the onshore and aquatic eco mentologic and paleoenvironmental systems are the vegetation-kill zones on problems, in Arthur, M. A., Ander the Ukinrek Maars ridge. These zones in son, T. F., Kaplan, I. R., Veizer, J., dicate areas of diffuse CO2 emissions with and Land, L.. S., Stable isotopes in soil-gas CO2 concentrations probably ex sedimentary geology: Society of ceeding 20-30%. We propose a soil-gas Economic Paleontologists and Min survey to help define (1) the exact CO2 eralogists Short Course No. 10, p. concentrations in the kill zones and (2) the 1-151. area! extent and flux of the diffuse CO2 Barnes, I., and McCoy, G. A., 1979, Possi emissions. In the vicinity of The Gas ble role of mantle-derived CO2 in Rocks, subaqueous CCb-gas and saline- causing two "phreatic" explosions in water discharges produce elevated DIG, Cl, Alaska: Geology, v. 7, p. 434-435. 25 Bischoff, J. L., and Seyfried, W. E., 1978, J., Prentice-Hall, Inc., 388 p. Hydrothermal chemistry of seawater Fahlquist, L., and Janik, C. J., 1992, Pro from 25° to 350° C. American Jour cedures for collecting and analyzing nal of Science, v. 278, p. 838-860. gas samples from geothermal sys Bureau of Mines, 1976, Oil and gas seeps tems: U.S. Geological Survey Open in Alaska: U.S. Bureau of Mines File Report 92-211, 19 p. Report of Investigations 8122. Farrar, C. D., Sorey, M. L., Evans, W. C., Burk, C. A., 1965, Geology of the Alaska Howle, J. F., Kerr, B. D., Kennedy, Peninsula-island arc and continen B. M., King, C.-Y., and Southon, J. tal margin, pts. 1,2, and 3: Geologi R., 1995, Forest killing difluse CO2 cal Society of America Memoir 99, emission at Mammoth Mountain as 250 p. a sign of magmatic unrest: Nature, Clayton, R. N., Friedman, I., Graf, D. L., v. 376, p. 675-677. Mayeda, T. K., Mcents, W. F., and Foumier, R. O., and TruesdeU, A. H., 1973, Schimp, N. F., 1966, The origin of An empiracal Na-K-Ca geother- saline formation waters: Journal of mometer for natural waters: Geo Geophysical Research, v. 71, p. chimica et Cosmochimica Acta, v. 3869-3882. 37, p. 1255-1275. Craig, H., 196la, Standard for reporting Gerlach, T. M., Doukas, M. P., McGee, K. concentrations of deuterium and A., Litasi-Gerlach, A., Sutton, J. A., oxygen-18 in natural waters. Sci and Elias, T., 1996, Ground Efflux ence, v. 133, p. 1833-1834. of 'Cold' CO2 in the Long Valley Cal- Craig, H., 196Ib, Isotopic variation in me dera and Mono-Inyo crates volcanic teoric waters: Science, v. 133, p. chain vicinity of eastern California: 1702-1703. EOS, Transactions of the American Craig, H., Lupton, J. E., and Horibe, Y., Geophysical Union, v. 77, p. F830. 978a, A mantle helium component Giggenbach, W. F., 1987, Redox processes in Circum-Pacific volcanic gases: governing the chemistry of fuma- Hakone, the Marianas, and Mt. Las- rolic gas discharges from White sen, in Alexander, E. C., and Ozima, Island, New Zealand. Applied Geo M., eds., Terrestrial Rare Gases: chemistry, v. 2, p. 143-161. Toyko, Japan Scientific Society Giggenbach, W. F., 1992, Isotopic shifts in Press, p. 3-16. waters from geothermal and vol Craig, H., Lupton, J. E., Welhan, J. A., and canic systems along convergent Poreda, R., 1978b, Helium isotope plate boundaries and their origin. ratios in Yellowstone and Lassen Earth and Planetary Science Let Park volcanic gases: Geophysical ters^. 113, p. 495-510. Research Letters, v. 5, p. 897-900. Giggenbach, W. F., and Gougel, R. L., Deines, P., Langmuir, D., and Harmon, R. 1989, Collection and analysis of S., 1974, Stable carbon isotope ra geothermal and volcanic water and tios and the existence of a gas gas discharges. Department of Sci phase in the evolution of carbonate entific and Industrial Research Re groundwaters: Geochimica et Cos- port No. CD2401, Petone, NZ, p. 69- mochimica Acta, v. 38, p. 1147- 81. 1154. Hitchon, B. and Friedman, L, 1969, Geo Detterman, R. L., Case, J. E., Wilson, F. H., chemistry and origin of formation Yount, M. E., 1987, Geologic map of waters in the western Canada the Ugashik, Bristol Bay, and west sedimentary basin I. Stable iso ern part of Karluk quadrangles, topes of hydrogen and oxygen: Geo Alaska: U.S. Geological Survey Map chimica et Cosmochimica Acta, v. 1-1685, scale 1:250,000. 33, 1321-1350. Drever, J. I., 1982, The geochemistry of Kienle, J., Kyle, P. R., Self, S., Motyka, R. natural waters: Englewood Cliffs, N. J., and Lorenz, V., 1980, Ukinrek 26 Maars, Alaska, I. April 1977 erup Journal of Volcanology and Geo tion sequence, petrology, and tec thermal Research, v. 7, p. 39-65. tonic setting: Journal of Volcanol- Sheppard, S. M. P., Nielson, R. L., and ogy and Geothennal Research, v. 7, Taylor, H. P., 1969, Oxygen and hy p. 11-37. drogen isotope ratios of clay miner Loudon, A., 1832, Visit to the Valley of als from porphyry copper deposits: Death, in the Island of Java, Letter Economic Geology, v. 64, p. 755- to Professor Jameson: The Edin 777. burgh New Philosophical Journal, Smith, B. N., and Epstein, S., 1971, Two October 1931-April 1932, p. 102- categories of *3C/ 12C ratios for 105. higher plants: Plant Physiology, v. Matsuo, S., Suzuki, M., and Mizutani, Y., 47, p. 380-384. 1978, Nitrogen to argon ratio in vol Stupfel, M., and Le Guern, P., 1989, Are canic gases, in Alexander, E. C. Jr., there biomedical criteria to assess and Ozima, M., eds., Terrestrial an acute carbon dioxide intoxica Rare Gases. Cent. Acad. Publ. Ja tion by a volcanic emission?: Jour pan, Tokyo, p. 17-25. nal of Volcanology and Geothermal Motyka, R. J., Liss, S. A., Nye, C. J., and Research, v. 39, p. 247-264. Moorman, M. A., 1993, Geothennal Symonds, R. B., Rose, W. I., Bluth, G. J. S., resources of the Aleutian Arc, and Gerlach, T. M., 1994, Volcanic- Alaska.: Alaska Division of gas studies: methods, results, and Geological and Geophysical Surveys applications, in Carroll, M. R., and Professional Report 114, 17 p. Holloway, J. R., eds., Volatiles in Poreda, R., and Craig, H., 1989, Helium Magmas. Reviews in Mineralogy, v. isotope ratios in circum-Pacific vol 30., p. 1-66. canic arcs: Nature, v. 338, p. 473- Taylor, B. E., 1986, Magmatic volatiles: iso- 478. topic variation of C, H, and S, in Poreda, R. J., and Parley, K. A., 1992, Rare Valley, J. W., Taylor, H. P., and gases in Samoan xenoliths: Earth O'Neil, J. R., eds., Stable isotopes in and Planetary Science Letters, v. high-temperature geological pro 113, p. 129-144. cesses: Reviews of Mineralogy, v. Poreda, R. J., Craig, H., Arnorsson, S., and 16, p. 185-225. Welhan, J. A., 1992, Helium iso Thorarinson, S., 1950, The eruption of Mt. topes in Icelandic geothermal sys Hekla 1947-1948: Bulletin Vol- tems: I. 3He, gas chemistry, and 13C canologique, Ser. II, v. 10, 157-168. relations: Geochimica et Cosmo- White, D. E., Bames, I., and O'Neil, J. R., chimica Acta, v. 56, p. 4221-4228. 1973, Thermal and mineral waters Reed, M., 1982, Calculation of multicom- of nonmeteoric origin, California ponent chemical equilibria and re Coast Ranges. Geological Society of action processes in systems involv America Bulletin, v. 84, p. 547-560. ing minerals, gases and an aqueous White, D. E., Hem, J. D., and Waring, G. A., phase: Geochimica et Cosmo- 1963, Chemical compositions of chimica Acta, v. 46, p. 513-528. subsurface waters, in Fleischer, M., Reed, M. and Spycher, N., 1984, Calcula ed., Data of geochemistry: U.S. tion of pH and mineral equilibria in Geological Survey Professional Pa hydrothennal waters with applica per 440-F, 67 p. tion to geothermometry and studies of boiling and dilution: Geochimica et Cosmochimica Acta, v. 48, p. 1479-1492. Self, S., Kienle, J., and Huot J.-P., 1980, Ukinrek Maars, Alaska, II. Deposits and formation of the 1977 craters: 27