William C. Evans

The 1986 Lake Nyos gas disaster in Cameroon, West Africa

The sudden, catastrophic release of gas from Lake Nyos on 21 August 1986 caused the deaths of at least 1700 people in the northwest area of Cameroon, West Africa. Chemical, isotopic, geologic, and medical evidence support the hypotheses that (i) the bulk of gas released was carbon dioxide that had been stored in the lakes hypolimnion, (ii) the victims exposed to the gas cloud died of carbon dioxide asphyxiation, (iii) the carbon dioxide was derived from magmatic sources, and (iv) there was no significant, direct volcanic activity involved. The limnological nature of the gas release suggests that hazardous lakes may be identified and monitored and that the danger of future incidents can be reduced.

Origin of the lethal gas burst from Lake Monoun, Cameroun

Abstract On 15 August, 1984, a lethal gas burst issued from a submerged 96-m-deep crater in Lake Monoun in Cameroun, western Africa, killing 37 people. The event was associated with a landslide from the eastern crater rim, which slumped into deep water. Waters below 50 m are anoxic, dominated by high Fe 2+ (∼600 mg/l) and HCO 3 − (≥ 1900 mg/l), anoxic and supersaturated with siderite, which is a major component of the crater floor sediments. The unusually high Fe 2+ levels are attributed to reduction of laterite-derived ferric iron gradually brought into the lake as loess and in river input. Sulfur compounds are below detection limits in both water and gas. Gases effervescing from depressurized deep waters are dominantly CO 2 with minor CH 4 , having δ 13 C of −7.18 and −54.8 per mil, respectively. Bacterial decomposition of organic matter may account for the methane, but 14 C of lake water indicates that only 10% of the carbon is modern, giving an apparent age of 18,000 years. The dominant source of carbon is therefore attributed to long-term emission of CO 2 as volcanic exhalation from vents within the crater, which led to gradual build-up of HCO 3 − in the lake. The density stratification of the lake may have been upset by an earthquake and underwater landslide on 15 August, which triggered overturn of the lake and caused nucleation of CO 2 in the deep water. The resultant ebullition of CO 2 from deep lake waters led to a gas burst at the surface and locally generated a water wave up to 5 m high. People travelling through the gas cloud were asphyxiated, presumably from CO 2 , and suffered skin discoloration from unidentified components.

Carbon dioxide and helium emissions from a reservoir of magmatic gas beneath Mammoth Mountain, California

Carbon dioxide and helium with isotopic compositions indicative of a magmatic source ( δ13C = −4.5 to −5‰, 3He/ 4He = 4.5 to 6.7 RA) are discharging at anomalous rates from Mammoth Mountain, on the southwestern rim of the Long Valley caldera in eastern California. The gas is released mainly as diffuse emissions from normal-temperature soils, but some gas issues from steam vents or leaves the mountain dissolved in cold groundwater. The rate of gas discharge increased significantly in 1989 following a 6-month period of persistent earthquake swarms and associated strain and ground deformation that has been attributed to dike emplacement beneath the mountain. An increase in the magmatic component of helium discharging in a steam vent on the north side of Mammoth Mountain, which also began in 1989, has persisted until the present time. Anomalous CO2 discharge from soils first occurred during the winter of 1990 and was followed by observations of several areas of tree kill and/or heavier than normal needlecast the following summer. Subsequent measurements have confirmed that the tree kills are associated with CO2 concentrations of 30–90% in soil gas and gas flow rates of up to 31,000 g m−2 d−1 at the soil surface. Each of the tree-kill areas and one area of CO2 discharge above tree line occurs in close proximity to one or more normal faults, which may provide conduits for gas flow from depth. We estimate that the total diffuse CO2 flux from the mountain is approximately 520 t/d, and that 30–50 t/d of CO2 are dissolved in cold groundwater flowing off the flanks of the mountain. Isotopic and chemical analyses of soil and fumarolic gas demonstrate a remarkable homogeneity in composition, suggesting that the CO2 and associated helium and excess nitrogen may be derived from a common gas reservoir whose source is associated with some combination of magmatic degassing and thermal metamorphism of metasedimentary rocks. Furthermore, N2/Ar ratios and nitrogen isotopic values indicate that the Mammoth Mountain gases are derived from sources separate from those that supply gas to the hydrothermal system within the Long Valley caldera. Various data suggest that the Mammoth Mountain gas reservoir is a large, low-temperature cap over an isolated hydrothermal system, that it predates the 1989 intrusion, and that it could remain a source of gas discharge for some time.

Spatial radon anomalies on active faults in California

Abstract Radon emanation has been observed to be anomalously high along active faults in many parts of the world. We tested this relationship by conducting and repeating soil-air radon surveys with a portable radon meter across several faults in California. The results confirm the existence of fault-associated radon anomalies, which show characteristic features that may be related to fault structures but vary in time due to other environmental changes, such as rainfall. Across two creeping faults in San Juan Bautista and Hollister, the radon anomalies showed prominent double peaks straddling the fault-gouge zone during dry summers, but the peak-to-background ratios diminished after significant rain fall during winter. Across a locked segment of the San Andreas fault near Olema, the anomaly has a single peak located several meters southwest of the slip zone associated with the 1906 San Francisco earthquake. Across two fault segments that ruptured during the magnitude 7.5 Landers earthquake in 1992, anomalously high radon concentration was found in the fractures three weeks after the earthquake. We attribute the fault-related anomalies to a slow vertical gas flow in or near the fault zones. Radon generated locally in subsurface soil has a concentration profile that increases three orders of magnitude from the surface to a depth of several meters; thus an upward flow that brings up deeper and radon-richer soil air to the detection level can cause a significantly higher concentration reading. This explanation is consistent with concentrations of carbon dioxide and oxygen, measured in soil-air samples collected during one of the surveys.

High CO2 emissions through porous media: transport mechanisms and implications for flux measurement and fractionation

Abstract Diffuse emissions of CO2 are known to be large around some volcanoes and hydrothermal areas. Accumulation-chamber measurements of CO2 flux are increasingly used to estimate the total magmatic or metamorphic CO2 released from such areas. To assess the performance of accumulation chamber systems at fluxes one to three orders of magnitude higher than normally encountered in soil respiration studies, a test system was constructed in the laboratory where known fluxes could be maintained through dry sand. Steady-state gas concentration profiles and fractionation effects observed in the 30-cm sand column nearly match those predicted by the Stefan-Maxwell equations, indicating that the test system was functioning successfully as a uniform porous medium. Eight groups of investigators tested their accumulation chamber equipment, all configured with continuous infrared gas analyzers (IRGA), in this system. Over a flux range of ∼200–12,000 g m−2 day−1, 90% of their 203 flux measurements were 0–25% lower than the imposed flux with a mean difference of −12.5%. Although this difference would seem to be within the range of acceptability for many geologic investigations, some potential sources for larger errors were discovered. A steady-state pressure gradient of −20 Pa/m was measured in the sand column at a flux of 11,200 g m−2 day−1. The derived permeability (50 darcies) was used in the dusty-gas model (DGM) of transport to quantify various diffusive and viscous flux components. These calculations were used to demonstrate that accumulation chambers, in addition to reducing the underlying diffusive gradient, severely disrupt the steady-state pressure gradient. The resultant diversion of the net gas flow is probably responsible for the systematically low flux measurements. It was also shown that the fractionating effects of a viscous CO2 efflux against a diffusive influx of air will have a major impact on some important geochemical indicators, such as N2/Ar, δ15N–N2, and 4He/22Ne.

Tracing and quantifying magmatic carbon discharge in cold groundwaters: Lessons learned from Mammoth Mountain, USA

Abstract A major campaign to quantify the magmatic carbon discharge in cold groundwaters around Mammoth Mountain volcano in eastern California was carried out from 1996 to 1999. The total water flow from all sampled cold springs was ≥1.8×10 7 m 3 /yr draining an area that receives an estimated 2.5×10 7 m 3 /yr of recharge, suggesting that sample coverage of the groundwater system was essentially complete. Some of the waters contain magmatic helium with 3 He/ 4 He ratios as high as 4.5 times the atmospheric ratio, and a magmatic component in the dissolved inorganic carbon (DIC) can be identified in virtually every feature sampled. Many waters have a 14 C of 0–5 pmC, a δ 13 C near −5‰, and contain high concentrations (20–50 mmol/l) of CO 2(aq) ; but are otherwise dilute (specific conductance=100–300 μS/cm) with low pH values between 5 and 6. Such waters have previously escaped notice at Mammoth Mountain, and possibly at many other volcanoes, because CO 2 is rapidly lost to the air as the water flows away from the springs, leaving neutral pH waters containing only 1–3 mmol/l HCO 3 − . The total discharge of magmatic carbon in the cold groundwater system at Mammoth Mountain is ∼20 000 t/yr (as CO 2 ), ranging seasonally from about 30 to 90 t/day. Several types of evidence show that this high discharge of magmatic DIC arose in part because of shallow dike intrusion in 1989, but also demonstrate that a long-term discharge possibly half this magnitude (∼10 000 t/yr) predated that intrusion. To sustain a 10 000 t/yr DIC discharge would require a magma intrusion rate of 0.057 km 3 per century, assuming complete degassing of magma with 0.65 wt% CO 2 and a density of 2.7 t/m 3 . The geochemical data also identify a small (

Gas buildup in Lake Nyos, Cameroon: The recharge process and its consequences

The gases dissolved in Lake Nyos, Cameroon, were quantified recently (December 1989 and September 1990) by two independent techniques: in-situ measurements using a newly designed probe and laboratory analyses of samples collected in pre-evacuated stainless steel cylinders. The highest concentrations of CO2 and CH4 were 0.30 mol/kg and 1.7 mmol/kg, respectively, measured in cylinders collected 1 m above lake bottom. Probe measurements of in-situ gas pressure at three different stations showed that horizontal variations in total dissolved gas were negligible. Total dissolved-gas pressure near the lake bottom is 1.06 MPa (10.5 atm), 50% as high as the hydrostatic pressure of 2.1 MPa (21 atm). Comparing the CO2 profile constructed from the 1990 data to one obtained in May 1987 shows that CO2 concentrations have increased at depths to below 150 m. Based on these profiles, the average rate of CO2 input to bottom waters was 2.6 × 108 mol/a. Increased deep-water temperatures require an average heat flow of 0.32 MW into the hypolimnion over the same time period. The transport rates of CO2, heat, and major ions into the hypolimnion suggest that a low-temperature reservoir of free CO2 exists a short distance below lake bottom and that convective cycling of lake water through the sediments is involved in transporting the CO2 into the lake from the underlying diatreme. Increased CH4 concentrations at all depths below the oxycline and a high14C content (41% modern) in the CH4 4 m above lake bottom show that much of the CH4 is biologically produced within the lake. The CH4 production rate may vary with time, but if the CO2 recharge rate remains constant, CO2 saturation of the entire hypolimnion below 50 m depth would require ∼140a, given present-day concentrations.

The evolution of thermal structure and water chemistry in Lake Nyos

Abstract We collected a time series of physical and chemical data to gain a better understanding of the dynamics of Lake Nyos. Measurements of water and gas chemistry, and temperature made during January, March, and May 1987 are compared to data taken in September 1986 just after the initial CO 2 gas release. There is no pattern of change in overall heat content of the lake, although heat input to bottom waters (185–208 m) has occurred at a rate of 1600 mW m −2 . This increase in heat content translates to a change from 23.38 to 24.12°C at 200 m and can be explained by geothermal heat flow and addition of thermal spring water. Concentrations of Ca 2+ , Mg 2+ , Na + , K + , Fe 2+ and alkalinity have increased only in bottom waters. In situ lake processes such as sulfate and iron reduction are unable to account for the changes in alkalinity. Observed chemical changes are consistent with a scenario where slightly thermal soda water is being input to the bottom of the lake. Measurements of p CO 2 at depth ranged from 18 to 28% of saturation and exhibited horizontal variability. Overall recharge of CO 2 in bottom waters is negligible. Mainly because of increasing ion concentrations in bottom water, total stability of the water column increased 33% from 48,800 J m −2 in September 1986 to 64,700 J m −2 in May 1987. As long as CO 2 concentrations remain the same, this level of stability is higher than could be disrupted by common limnologic or meteorologic processes. There is thermal and chemical evidence that a buildup of dissolved iron and CO 2 in bottom waters must have preceded the August 1986 gas release. In addition, a survey of all crater lakes in Cameroon indicates that only Lakes Nyos and Monoun contain high concentrations of dissolved iron and CO 2 . Thus there is a low probability of any other Cameroonian lake releasing a substantial volume of CO 2 .

Excess nitrogen in selected thermal and mineral springs of the Cascade Range in northern California, Oregon, and Washington: sedimentary or volcanic in origin?

Abstract Anomalous N2/Ar values occur in many thermal springs and mineral springs, some volcanic fumaroles, and at least one acid-sulfate spring of the Cascade Range. Our data show that N2/Ar values are as high as 300 in gas from some of the hot springs, as high as 1650 in gas from some of the mineral springs, and as high as 2400 in gas from the acid-sulfate spring on Mt. Shasta. In contrast, gas discharging from hot springs that contain nitrogen and argon solely of atmospheric origin typically exhibits N2/Ar values of 40–80, depending on the spring temperature. If the excess nitrogen in the thermal and mineral springs is of sedimentary origin then the geothermal potential of the area must be small, but if the nitrogen is of volcanic origin then the geothermal potential must be very large. End-member excess nitrogen (δ15N) is +5.3‰ for the thermal waters of the Oregon Cascades but is only about +1‰ for fumaroles on Mt. Hood and the acid-sulfate spring on Mt. Shasta. Dissolved nitrogen concentrations are highest for thermal springs associated with aquifers between 120 and 140°C. Chloride is the major anion in most of the nitrogen-rich springs of the Cascade Range, and N2/Ar values generally increase as chloride concentrations increase. Chloride and excess nitrogen in the thermal waters of the Oregon Cascades probably originate in an early Tertiary marine formation that has been buried by the late Tertiary and Quaternary lava flows of the High Cascades. The widespread distribution of excess nitrogen that has been generated in low to moderate-temperature sedimentary environments is further proof of the restricted geothermal potential of the Cascade Range.

Radiocarbon studies of plant leaves and tree rings from Mammoth Mountain, CA: a long-term record of magmatic CO2 release

Evaluation of 14C in tree rings provides a measure of the flux of magmatic CO2 from Mammoth Mountain both before and after 1994 when copious diffuse emissions were first discovered and linked to tree kill. We analyzed the annual rings of trees with two main purposes: (1) to track changes in the magnitude of magmatic CO2 emission over time, and (2) to determine the onset of magmatic CO2 emission at numerous sites on Mammoth Mountain. The onset of CO2 emission at different areas of tree kill was determined to be in 1990, closely following the seismic events of 1989. At Horseshoe Lake (HSL), CO2 emission was found to have peaked in 1991 and to have subsequently declined by a factor of two through 1998. The tree-ring data also show that emissions of magmatic carbon from cold springs below the tree-kill areas occurred well before 1989. Trees located on the margins of the kill areas or otherwise away from zones of maximum gas discharge were found to be better integrators of magmatic CO2 emission than those located in the center of tree kills. Although quantitative extrapolations from our data to a flux history will require that a relationship be established between 14C depletion in tree rings and average annual magmatic CO2 flux, the pattern of 14C depletion in tree rings is likely to be the most reliable indicator of the long-term changes in the magnitude of CO2 release from Mammoth Mountain.