Sie sind auf Seite 1von 4

Hydrothermal element fluxes from Copahue, Argentina: A

beehive volcano in turmoil


Johan C. Varekamp
Andrew P. Ouimette*
Scott W. Herman
Department of Earth and Environmental Sciences, Wesleyan University, 265 Church Street, Middletown,
Connecticut 06459-0139, USA

Adriana Bermudez
Daniel Delpino
Consejo Nacional de Investigaciones Cientficas y Tecnicas, National University of Comahue, Neuquen, Argentina

ABSTRACT
Copahue volcano erupted altered rock debris, siliceous dust, pyroclastic sulfur, and rare
juvenile fragments between 1992 and 1995, and magmatic eruptions occurred in July
October 2000. Prior to 2000, the Copahue crater lake, acid hot springs, and rivers carried
acid brines with compositions that reflected close to congruent rock dissolution. The ratio
between rock-forming elements and chloride in the central zone of the volcano-hydrothermal system has diminished over the past few years, reflecting increased water/rock
ratios as a result of progressive rock dissolution. Magmatic activity in 2000 provided fresh
rocks for the acid fluids, resulting in higher ratios between rock-forming elements and
chloride in the fluids and enhanced Mg fluxes. The higher Mg fluxes started several weeks
prior to the eruption. Model data on the crater lake and river element flux determinations
indicate that Copahue volcano was hollowed out at a rate of about 20 00025 000 m3/yr,
but that void space was filled with about equal amounts of silica and liquid elemental
sulfur. The extensive rock dissolution has weakened the internal volcanic structure, making flank collapse a volcanic hazard at Copahue.
Keywords: hydrothermal fluids, volcanic processes, hydrochemistry, limnology, eruptions.

INTRODUCTION
Copahue is a 2997-m-high composite volcano (Province of Neuquen, lat 37.538S, long
71.108W), located at the east side of the South
Volcanic Zone of the Andes in western Argentina. The basaltic-andesitic products are of
Pleistocene to Holocene age (Delpino and
Bermudez, 1993). The volcano summit was
extensively glaciated in Pleistocene time, and
a modern small glacier provides meltwater to
the crater lake. The active cone consists of
surge and near-vent fallout deposits, whereas
the broad base consists of debris avalanches,
lahars, and lava flows (Goss, 2001).
Copahue activity has been reported since
the eighteenth century. A new eruptive cycle
started in July 1992; explosions continued in
1993, and major eruptions occurred in December 1994 and September 1995 (Bermudez and
Delpino, 1995; Delpino and Bermudez, 1993).
The crater lake explosions ejected hydrothermally altered rock fragments, siliceous white
dust, copious amounts of green and yellow
liquid sulfur (Delpino and Bermudez, 1995),
and some basaltic-andesitic juvenile frag*Present address: U.S. Geological Survey, Volcano Hazards Division, Menlo Park, California
94025, USA.

ments. Magmatic eruptions (VEI 12), starting with phreatomagmatic events, began in
July 2000; continuous degassing occurred between eruptive phases. Incandescent bombs
were ejected, and dark ash and chilled sulfur
fragments covered an area up to 50 km from
the source (Global Volcano Network, 2000a,

2000b); the larger ash and gas plumes were


detectable up to 250 km from the vent.
An acid crater lake, acid hot springs, and a
geothermal field (Martini et al., 1997) are the
surface expressions of an extensive volcanicmagmatic hydrothermal system (Fig. 1),
which has been the focus of our investigations

Figure 1. Map of Copahue region. Light gray area at left is steep volcanic cone. CPL
crater lake, CPacid hot springs. Gently sloping foothills are between dashed line and
Lake Caviahue, which is situated on flat caldera floor. Arrow shows location for river flux
measurements.

q 2001 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or editing@geosociety.org.
Geology; November 2001; v. 29; no. 11; p. 10591062; 3 figures; 2 tables.

1059

TABLE 1. CRATER LAKE, ACID HOT SPRING, WHITE RIVER TRIBUTARY, AND LAKE CAVIAHUE
COMPOSITIONS IN 1999 AND 2000
Sample
CPL1
CPL2
CP1
CP2
CP3
CP3/CP1
WR1
CVL-N1

Date
(mo/yr)

pH*

Cl

SO4

Al

Fe

Mg

Ca

1/99
11/99
1/99
11/99
7/00

0.30
0.18
0.32
0.46
0.30
1
5.4
2.2

7978
9775
10,883
7018
9781
0.9
115
96

55,416
65,983
65,473
45,438
59,530
0.9
115
473

1375
1421
3236
1910
5478
1.7
0.5
32.4

748
794
2090
1067
3907
1.9
BD
20.6

279
274
539
310
2199
4
22.7
14.7

890
1038
871
968
1161
1.3
25.1
21.8

11/99
11/99

Si
61
98
102
98
NA
18.9
13.7

Na

361
342
819
413
1484
1.8
10.1
14.1

319
334
791
423
723
0.9
4.3
7.6

Note: Element values for crater lake (CPL), main acid hot spring (CP), White River tributary (WR), and Lake
Caviahue (CVL) samples are in parts per million. BDbelow detection, NAnot analyzed.
*If pH , 0.5, calculated by charge balance.

since 1997 (Herman, 1998; Ouimette, 2000).


The crater lake water varied in color from
dark-green to gray; clouds of HCl vapors
exsolved when the water was hot. The main
acid hot spring emerges about 100 m below
the lake level and feeds an acid river, the Upper Rio Agrio, which discharges ;12 km
downslope into Lake Caviahue, a large, acidified glacial meltwater lake.
About 300 water samples, collected between March 1997 and August 2000, were analyzed at Wesleyan University by inductively
coupled plasmaatomic-emission spectroscopy and ion chromatography. The water flux of
the Rio Agrio was measured repeatedly with
a flow meter, and element fluxes were determined from water flux and local river water
compositions.

COPAHUE CRATER LAKE AND ACID


HOT SPRINGS
Copahue crater lake contains an acid brine
(Table 1) with temperatures of 21 (January
1999) to 54 8C (March 1997). The hot springs
had higher temperatures and higher concentrations of rock-forming elements (RFEsCa,
Na, Mg, K, Al, Fe) than the lake waters (Table
1). Chemical and isotopic data show that the
crater lake and hot springs are both fed by
deep hydrothermal fluids, which contain up to
70% magmatic brine (Ouimette, 2000). A
shallow reservoir with a temperature of about
175 8C probably feeds the crater lake directly,
whereas the hot-spring fluids have undergone
additional water-rock interaction and conductive cooling prior to reaching the spring areas
(Ouimette, 2000). Both the lake and the hot
spring had high concentrations of toxic ele-

TABLE 2. FLUXES FOR THE CRATER LAKE, ACID HOT SPRINGS, UPPER RIO AGRIO,
AND COPAHUE VOLCANO
Location

Copahue crater lake


T (8C)
Cl (ppm)
Brine influx* (m3/s)
Energy input* (MW)
Net rock removal rate (m3/yr)
Copahue springs
T (8C)
Cl (ppm)
RFE/Cl (molar)
Water flux (m3/s)
Heat flux (MW)
Upper Rio Agrio watershed
Cl (ppm)
Water flux (m3/s)
Mg flux (t/yr)
Sulfur flux (kt S/yr)
Net rock removal rate (m3/yr)
So accumulation rate# (m3/yr)
Eq. volcanic SO2 input (t/d)
Whole Copahue system
Eq. volcanic SO2 input (t/d)
So accumulation rate (m3/yr)
Net rock removal rate (m3/yr)

Date (month/year)
3/97

3/98

1/99

54
10 157
0.087
45
7000

;33
8893

21
7978
0.01
7
480

33
9775
0.024
15
1150

No lake

68
10 617
1.0
0.031
9

63
10 166
1.07

83
10 883
0.84
0.035
12

72
7018
0.79
0.052
16

75
9781
1.61
0.063
20

1609
0.26
1310
21
7200
5870
174
656
22 200
14 300

950
0.48
1510
24
8300
6650
197

178
2.25
1990
25
10 900
6960
206

245
8300
8800

344
11 600
12 000

Note: N.A.not applicable because rock dissolution is no longer congruent.


*Energy steady-state model (Pasternack and Varekamp, 1997; Ouimette, 2000).

Taking silica precipitation into account (about 50% of rock mass is SiO2).

Assuming that all SO4 is derived from the hot springs.


#
Using a density of sulfur of 1800 kg/m3.

1060

11/99

7/00

176
3.25
7800
39
N.A.
10 900
323

ments; e.g., the spring had 7.9 ppm As, 4.3


ppm B, 3.2 ppm Pb, 2.8 ppm Zn, and 1.8
ppm Cr in 1997, and the spring discharge
into the Upper Rio Agrio is environmentally
undesirable.
The Cl concentrations in the crater lake increased from 2500 ppm Cl in 1920 to
.10 000 ppm Cl in the late 1970s (Casertano,
1964; Martini et al., 1997; Herman, 1998;
Ouimette, 2000). From 1997 to 2000, the Cl
concentrations in the lake initially decreased,
but during 1999 the concentrations and lakewater temperatures increased considerably.
The compositional variation in the crater lake
samples relates to the magnitude of meteoric
and glacial meltwater inputs and to changes in
the rate of volcanic input. The compositional
variations in the hot-spring fluids result from
mixing between meteoric waters and the volcanic brine at depth.
Power inputs into the lake were calculated
from steady-state energy budgets (Pasternack
and Varekamp, 1997; Ouimette, 2000) and
varied between 7 and 45 MW (Table 2; input
variables: lake diameter 125 m, lake temperatures, and meteorological parameters). Sizable hydrothermal inputs are needed to keep
the crater lake warm (average ambient temperature at the summit is 22 8C) and lake
seepage is an important process in keeping hydrological balance (Rowe et al., 1995).
Speciation-saturation simulations with the
program SOLVEQ (Reed and Spycher, 1984)
indicate that the crater lake and hot-spring fluids are saturated in silica and close to saturation with gypsum and anhydrite. The ratios
between RFEs in crater lake and hot-spring
fluids did not vary much during the twentieth
century and are close to those in average rock
of Copahue (Goss, 2001; Fig. 2). The similarity of the RFE ratios in average rock and fluids and the general lack of secondary mineral
saturation strongly suggest that congruent
rock dissolution was the dominant mode of
water-rock interaction in the Copahue volcanohydrothermal system. Almost all RFEs are
carried off by the hydrothermal fluids, except
for silica, which is highly depleted in the fluids relative to all other elements. With the calculated water input fluxes from the energy
model, lake cation concentrations, and an average of Copahue rock compositions (Goss,
2001), we calculated net rock-removal rates
for the lake (Table 2), assuming congruent dissolution with approximately quantitative silica
retention in the hydrothermal system.
The RFE/Cl values and the degree of neutralization of the fluids (as used by Varekamp
et al., 2000) have declined in the crater lake
fluids and hot spring during the past few years
(Table 2). We hypothesize that this resulted
from increased water/rock ratios in the hydrothermal system as a result of rock dissolution
GEOLOGY, November 2001

Figure 2. Concentrations of K and Mg in twentieth century Copahue fluids (solid circles) are uniform, and K/Mg ratios are similar
to those in average Copahue rocks (small circles, parts per million
values divided by 20). July 2000 hot-spring fluids have exceptionally high Mg concentrations, with different K/Mg ratios than 2000
magma or older Copahue fluids, suggesting noncongruent dissolution of newly intruded magma.

as well as that the rock protolith became covered with liquid sulfur and/or cristobalite,
slowing water-rock reaction. The July 2000
fluids show a dramatic change from this trend
(Tables 1 and 2); the RFE/Cl ratios increased,
and RFE ratios differ from those in older Copahue rocks and in the new magma (Fig. 2).
The composition of these fluids resulted from
the noncongruent dissolution of the newly intruded magma. Calculations with SOLVEQ
indicate saturation of alunite, anhydrite, and
silica phases at temperatures .150 8C, which
may explain why the K and Ca concentrations
have increased less than those of Fe, Mg, and
Na (Table 1, CP3/CP1 ratio column).
UPPER RIO AGRIO
The Upper Rio Agrio starts as pure hotspring water, which is cooled and diluted with
glacial meltwater and tributary inflows farther
downstream. Many tributaries derive water
from thermal springs lower on the slopes of
Copahue, which tend to be much less acidic
and have much lower sulfate concentrations
(e.g., WR tributary, Table 1). Annual element
and energy fluxes (Table 2) were calculated
from water flux measurements taken where
the river enters Lake Caviahue (Fig. 1); calGEOLOGY, November 2001

Figure 3. Magnesium concentrations in Lake Caviahue as function of


time (month/year). Black symbols are measured Mg concentrations
at .20 m depth; large open circles are average deep-water values
for each year; small open circles are shallow-water (,20 m) data.
Evolution curves for Mg concentrations over time assume 5 yr residence time for water in lake and start in 1992 at 12 ppm Mg. Bottom
curve represents 1350 t Mg/yr input, whereas other curves show
compositional evolution with stepwise increase in Mg flux as measured in Upper Rio Agrio over time. July 2000 data plot above model
curve for 2000 t Mg/yr input and represent values that must be associated with much higher Mg input fluxes. Steep line is evolution
curve for flux of 7800 t Mg/yr as measured in July 2000 during eruptive period. That flux must have started at least six weeks prior to
eruption to create these observed bottom-water Mg concentrations.

culated net rock-removal rates range from


7000 to 11 000 m3/yr (Table 2).
The rate of elemental sulfur accumulation
in the volcano was calculated from the sulfate
flux and the stoichiometry of the SO2 disproportionation reaction (Kusakabe et al., 2000):
3SO2 1 2H2O 2HSO24 1 S8 1 2H1. We
assume that this is the main disproportionation
reaction in the system, given that H2S is virtually absent in the lake and hot-spring fluids.
This calculation provides for the Upper Rio
Agrio sulfate flux an approximate sulfur accumulation rate of ;6000 m3/yr (Table 2).
The amount of void space generated by the
rock dissolutionsilica precipitation mechanism is of the same magnitude as the sulfur
accumulation rate (compare Rowe et al.,
1992a), and rocks in the volcano interior are
replaced by roughly equal amounts of silica
and liquid native sulfur.
LAKE CAVIAHUE
Lake Caviahue is a two-finger glacial lake
with a volume of ;0.47 km3 (Rapacioli,
1985; Pedrozo et al., 2001). Water is supplied
largely by the mineralized Upper Rio Agrio
and the Aqua Dulce, a glacial meltwater
stream. The lake drains by an overflow in the

northern arm through the Lower Rio Agrio


(Fig. 1). The acidity and moderate concentrations of dissolved substances in Lake Caviahue (Table 1) are supplied by the Upper Rio
Agrio, and given the large volume of the lake,
the lake-bottom water compositions provide a
memory of the element fluxes of the Upper
Rio Agrio in the past. The average deep-water
(.20 m) compositions show a progression
from ;13.4 ppm Mg in 1997 to ;14.5 ppm
in 1999 and then they jumped to 15.524 ppm
Mg in July 2000 (Fig. 3).
We modeled the composition of a wellmixed lake as a function of time with the expression Ct 5 Css (1 2 e2t/R) 1 e2t/R Cinit,
where Ct is the concentration of a component
at time t, Css is the component concentration
at steady state, Cinit is the initial concentration,
t is the time elapsed, and R is the residence
time of the component in the lake at steady
state (Varekamp, 1988). The term Css can be
rephrased as RFin /V, where Fin is the input
flux of the component and V is the mass of
the lake. We used a bottom-water residence
time of 5 yr for Lake Caviahue, based on our
water flux measurements and models, taking
into account that the lake is thermally stratified in the Austral summer. The modeled time1061

concentration curves for Mg fluxes of 1350 to


1500 to 2000 t Mg/yr (Table 2) cover the averages of the deep-water data for 19971999
quite well (Fig. 3). Surface-water data show
evidence for imperfect source mixing when
sampled close to incoming rivers. The much
higher Mg concentrations in deep lake waters
measured only 10 d after the onset of the July
2000 eruption indicate that the high Mg flux
of the Upper Rio Agrio must have preceded
the eruption by some time. Using the July
2000 Mg flux (7800 t Mg/yr) and a conservative average of 17 ppm Mg for the July
2000 deep-water Mg concentration, we derive
that the enhanced Mg flux of the Upper Rio
Agrio preceded the eruption by at least six
weeks (the lake is well mixed during the austral winter months). The Mg flux thus was an
excellent precursor to the eruption and a tracer
for the rise (and dissolution) of new magma
into the acid hydrothermal system.
CONCLUSIONS
Copahue volcano acts as a beehive volcano, accumulating liquid sulfur and silica
phases in void spaces generated by rock dissolution. Part of that accumulated sulfur and
silica was ejected during the 19922000 eruptions. The rock dissolution leads to higher water/rock ratios and therefore declining RFE/Cl
values in the fluids over time. The 2000 intrusion of fresh magma into the hydrothermal
system led to increased RFE/Cl values in the
fluids, and water compositions became especially enriched in Mg.
The Rio Agrio sulfate flux and associated
native sulfur storage rate can be expressed in
equivalent volcanic SO2 inputs into the hydrothermal system; these ranged from 174 to 323
t SO2/d (Table 2). These are typical SO2 emission rates for passively degassing volcanos
(Andres, 1998), and Copahue is an example
in which hydrothermal SO2 scrubbing is very
significant (Symonds et al., 2001).
Integration of the measured river flux data
and the modeled volcanic fluxes into the crater
lake provide the following parameters for the
whole Copahue system in November 1999: an
energy flux of ;32 MW, an equivalent sulfur
gas input of ;344 t SO2/d, a net rock removal
rate of about 12 000 m3/yr, and an elemental
sulfur accumulation rate of about 11 600 m3/
yr. Estimates of bulk hydrothermal rock alteration rates at Poas volcano are of a similar
magnitude (;25 000 m3/yr; Rowe et al.,
1995).
The evolution of the crater lake fluids during the twentieth century suggest a gradual
awakening of Copahue, and future activity can
be expected. Monitoring the RFE/Cl in the hot
springs and temperatures in the newly reestab-

1062

lished crater lake (January 2001) will be useful in assessing future volcanic activity (e.g.,
Rowe et al., 1992b). The water composition
data from Lake Caviahue serve as a memory
of the magnitude of element fluxes of the Upper Rio Agrio, and strongly suggest that enhanced Mg fluxes preceded the 2000 eruptions
by several weeks. We began a weekly monitoring of the composition of the Upper Rio
Agrio in February 2001.
The bulk rock dissolution process (20 000
25 000 m3/yr) is comparable to drilling a 1km-deep hole with a diameter of 2.5 m in the
volcano each year, a hole that is then partially
filled with equal volumes of silica and liquid
sulfur. The weakened structure creates risks of
flank collapse (Lopez and Williams, 1993; van
Wijk de Vries et al., 2000), especially taking
into account the fracture system that is present
in the east flank of the volcano. Flank collapse
with the snow and ice cap would cause severe
lahar hazards, and the rounded shape of the
volcano and abundance of lahar deposits on
its lower slopes indicate that such collapses
have occurred in the past.
ACKNOWLEDGMENTS
This research was supported by National Science
Foundation grants INT-9704200 and INT-9813912.
We appreciate the contributions of Caniche, Jane
Coffey, Jelle deBoer, Noah Garrison, Adam Goss,
Rob Kreulen, and Danielle Piraino.
REFERENCES CITED
Andres, R., 1998, A time-averaged inventory of subaerial volcanic sulfur emissions: Journal of Geophysical Research, v. 103, p. 25 25125 261.
Bermudez, A., and Delpino, D., 1995, Mapa de los
peligros potenciales en el area del Volcan CopahueSector Argentino: Neuquen, Argentina, Province of Neuquen Geological Survey,
scale 1:50 000.
Casertano, L., 1964, Some reflections on the fumarolic manifestations of the Los Copahues
crater: Bulletin of Volcanology, v. 27,
p. 197215.
Delpino, D., and Bermudez, A., 1993, La actividad
del Volcan Copahue durante 1992. Erupcion
con emisiones de azufre piroclastico, Provencia
del Neuquen, Argentina: Congreso Geologico
Argentino XII y Congreso de Exploracion de
Hidrocarburos II, Actas, IV, p. 292301.
Delpino, D., and Bermudez, A., 1995, Eruptions of
pyroclastic sulphur at crater lake of Copahue
volcano, Argentina: International Union of
Geodesy and Geophysics, General Assembly
XXI, Abstracts, p. B410.
Global Volcano Network, 2000a, Frequent ash explosions and acidic mudflows starting on July
1: Global Volcano Network Bulletin, v. 25,
p. 6.
Global Volcano Network, 2000b, Continued ash explosions and tremor during AugustOctober:
Global Volcano Network Bulletin, v. 25, p. 9.
Goss, A.R., 2001, Magmatic evolution of Volcan
Copahue, Neuquen, Argentina [undergrad.
thesis]: Middletown, Connecticut, Wesleyan
University, 209 p.
Herman, S.W., 1998, The hydrogeochemistry of Co-

pahue volcano, Argentina [undergrad. thesis]:


Middletown, Connecticut, Wesleyan University, 94 p.
Kusakabe, M., Komoda, Y., Takano, B., and Abiko,
T., 2000, Sulfur isotopic effects in the disproportionation reaction of sulfur dioxide in hydrothermal fluids: Implications for the d34S
variations of dissolved bisulfate and elemental
sulfur from active crater lakes: Journal of Volcanology and Geothermal Research, v. 97,
p. 309327.
Lopez, D.L., and Williams, S.N., 1993, Catastrophic
volcanic collapse: Relation to hydrothermal
processes: Science, v. 260, p. 17941796.
Martini, M., Bermudez, A., Delpino, D., and Giannini, L., 1997, The thermal manifestations
of Copahue volcano area, Neuquen, Argentina: Congreso Geologico Chileno VIII, Actas,
v. 1, p. 352356.
Ouimette, A.P., 2000, Hydrothermal processes at an
active volcano, Copahue, Argentina [M.A.
thesis]: Middletown, Connecticut, Wesleyan
University, 220 p.
Pasternack, G.B., and Varekamp, J.C., 1997, Volcanic lake systematics I. Physical constraints:
Bulletin of Volcanology, v. 58, p. 528538.
Pedrozo, F., Kelly, L., Diaz, M., Temporetti, P., Baffico, G., Kringel, R., Friese, K., Mages, M.,
Geller, W., and Woelfl, S., 2001, First results
on the water chemistry, algae and trophic status of an Andean acidic lake system of volcanic origin in Patagonia (Lake Caviahue):
Hydrobiologia (in press).
Rapacioli, R., 1985, Lake Caviahue and its basin
[Technical report, EPAS]: Argentina, Government Office, Province of Neuquen, 72 p.
Reed, M.H., and Spycher, N.F., 1984, Calculation of
pH and mineral equilibria in hydrothermal waters with application to geothermometry and
studies of boiling and dilution: Geochimica et
Cosmochimica Acta, v. 48, p. 14791492.
Rowe, G.L., Jr., Brantley, S.L., Fernandez, M., Fernandez, J.F., Borgia, A., and Barquero, J.,
1992a, Fluid-volcano interaction in an active
stratovolcano: The crater lake system of Poas
volcano, Costa Rica: Journal of Volcanology
and Geothermal Research, v. 49, p. 2351.
Rowe, G.L., Jr., Ohsawa, S., Takano, B., Brantley,
S.L., Fernandez, J.F., and Barquero, J., 1992b,
Using crater lake chemistry to predict volcanic
activity at Poas volcano, Costa Rica: Bulletin
of Volcanology, v. 54, p. 494503.
Rowe, G.L., Jr., Brantley, S.L., Fernandez, J.F., and
Borgia, A., 1995, The chemical and hydrologic structure of Poas volcano, Costa Rica: Journal of Volcanology and Geothermal Research,
v. 64, p. 233267.
Symonds, R.B., Gerlach, T.M., and Reed, M.H.,
2001, Magmatic gas scrubbing: Implications
for volcano monitoring: Journal of Volcanology and Geothermal Research (in press).
Van Wijk de Vries, B., Kerleg, N., and Petley, D.,
2000, Sector collapse forming at Casita volcano, Nicaragua: Geology, v. 28, p. 167170.
Varekamp, J.C., 1988, Lake-pollution modelling:
Journal of Geological Education, v. 36,
p. 49.
Varekamp, J.C., Pasternack, G.B., and Rowe, G.L.,
2000, Volcanic lake systematics. II. Chemical
constraints: Journal of Volcanology and Geothermal Research, v. 97, p. 161179.
Manuscript received March 12, 2001
Revised manuscript received July 9, 2001
Manuscript accepted July 27, 2001
Printed in USA

GEOLOGY, November 2001

Das könnte Ihnen auch gefallen