Geochimica et Cosmochimica Acta, Vol. 66, No. 21, pp.

3827–3838, 2002
Copyright © 2002 Elsevier Science Ltd
Pergamon Printed in the USA. All rights reserved
0016-7037/02 $22.00 ⫹ .00
PII S0016-7037(02)00825-6
129
Origin of iodine in volcanic fluids: I results from the Central American Volcanic Arc
GLEN T. SNYDER* and UDO FEHN
Department of Earth and Environmental Sciences, University of Rochester, Rochester, NY 14627, USA

(Received July 11, 2001; accepted in revised form December 20, 2001)

Abstract—The largest reservoir of crustal iodine is found in marine sediments, where it is closely associated
with organic material. This presence, together with the existence of a long-lived, cosmogenic radioisotope 129I
(t1/2 ⫽ 15.7 Ma), make this isotopic system well suited for the study of sediment recycling in subduction
zones. Reported here are the results of 129I/I ratios in volcanic fluids, collected during a comprehensive study
of fluids and gases in the Central American Volcanic Arc. 129I/I ratios, together with I, Br, and Cl
concentrations, were determined in 79 samples from four geothermal centers and a number of crater lakes,
fumaroles, hot springs, and surface waters in Costa Rica, Nicaragua, and El Salvador. Geothermal and
volcanic fluids were found to have iodine concentrations substantially higher than values in seawater or
meteoric waters. 129I/I ratios in most of the geothermal fluids are below the preanthropogenic input ratio of
1500 ⫻ 10⫺15, demonstrating that recent anthropogenic additions are largely absent from the volcanic
systems. The majority of the 129I/I ratios are between 500 and 800 ⫻ 10⫺15. These ratios indicate minimum
iodine ages between 25 and 15 Ma, in good agreement with the age of subducted sediments in this region. In
all four geothermal systems, however, a few samples were found with iodine ages older than 40 Ma—that is,
considerably below the expected age range for subducted sediments from the Cocos Plate. These samples
probably reflect the presence of iodine derived from sediments in older accreted oceanic terraines. The iodine
ages indicate that the magmatic end member for the volcanic fluids originates in the deeper parts of the
subducted sediment column, with small additions from older iodine mobilized from the overlying crust. The
high concentrations of iodine in geothermal fluids, combined with the observed iodine ages, demonstrate that
remobilization in the main volcanic zone (and probably also in the forearc area) is an important part in the
overall marine cycle of iodine and similar elements. Copyright © 2002 Elsevier Science Ltd

1. INTRODUCTION rapid formation of nitrogenous organic iodine compounds

(Luther et al., 1995; Carpenter et al., 1999). This organic iodine
Iodine is an element strongly associated with organic mate-
eventually accumulates on the ocean floor through the settling
rial and is heavily concentrated in surficial reservoirs of the
of detrital material. The concentration of iodine in marine
Earth’s crust. A recent compilation of the Earth’s iodine dis-
tribution (Muramatsu and Wedepohl, 1998) showed that the particulates is at least three orders of magnitude greater than
largest reservoir of crustal iodine is found in marine sediments, that of seawater and ranges from 200 to 1300 ppm (Wong,
but that known iodine fluxes in and out of the marine cycle are 1991). The continual scavenging of iodine from the oceans has
not well balanced. In particular, subduction of marine sedi- given rise to the largest inventory of iodine on the Earth’s crust,
ments could carry a large fraction of iodine back into the that of marine sediments (Fig. 1). Roughly 70% of the total
mantle, resulting in a depletion of crustal and surface invento- crustal iodine inventory presently resides in marine sediments
ries. Iodine has a cosmogenic isotope, 129I, with a half-life (Muramatsu and Wedepohl, 1998). The breakdown of organic
(15.7 Ma) appropriate for investigations of sediment recycling sediments during diagenesis generally leads to sedimentary
in subduction zones. The purpose of this investigation was to pore-fluid concentrations between 10 and 144 ppm (Martin et
study the isotopic signature of iodine in volcanic fluids to al., 1993; Moran et al., 1995).
determine the origin of iodine in arc volcanoes and relate it to The flux of iodine from seawater to the ocean floor is roughly
subduction processes. The investigation was carried out as part equal to the flux of iodine into the atmosphere through biologic
of a comprehensive study of gases and fluids in geothermal and production of methyl halides, principally CH4I, in the photic
volcanic systems of the Central American Volcanic Arc in zone (Carpenter et al., 1999). A major difference, however, is
Costa Rica, Nicaragua, and El Salvador (Snyder, 2001). that the flux of iodine into the atmosphere is balanced by the
return of iodine through precipitation (Fig. 2), whereas the flux
1.1. Iodine in Natural Systems of iodine to the ocean floor represents a net loss, at least on the
short term. The residence time of iodine in the atmosphere is
Less than 1% of the Earth’s total crustal inventory of iodine short (14d; Moran et al., 1999) compared with the residence
presently resides in the oceans (Muramatsu and Wedepohl, time of iodine in the ocean of 340,000 yr (Broecker and Peng,
1998), where concentrations are fairly uniform at 56 ppb (Bro- 1982). Assuming that the average age of the ocean floor is 60
ecker and Peng, 1982). The production of H2O2 by phytoplank-
Ma, then the residence time of iodine in marine sediments must
ton in the presence of unsaturated carbon networks favors the
be much greater than in seawater. The residence time based on
the estimated size of the marine sediment reservoir and the
*Author to whom correspondence should be addressed deposition rate (Muramatsu and Wedepohl, 1998) is 30 Ma.
([email protected]). Box models of the global flux of iodine in marine environ-
3827
3828 G. T. Snyder and U. Fehn

Fig. 1. Principal iodine reservoirs (109 ton). Sources: (1) Muramatsu

and Wedepohl (1998); (2) NCRP (1983); (3) 10.5 ppb bulk silicate
earth concentration (Déruelle et al., 1992).

ments (NCRP, 1983; Muramatsu and Wedepohl, 1998) do not

adequately account for the source of iodine in marine sediments
or the release of organic iodine from sediments back into the
ocean. Because the loss of marine methyl iodide is roughly
balanced by the gain of iodine from precipitation, the flux of
iodine from the oceans into marine sediments must be balanced Fig. 3. Sampling locations in Central America. CR ⫽ Costa Rica;
NI ⫽ Nicaragua; SV ⫽ El Salvador. Circles ⫽ geothermal wells;
by the input from another source. The input from rivers is
stars ⫽ fumaroles and crater lakes; squares ⫽ rivers and lakes; dia-
insufficient to balance the net accumulation (Muramatsu and monds ⫽ hot springs and mud pots. 1 ⫽ Jaular Pond; Irazú; 2 ⫽ Poás
Wedepohl, 1998). Other sources may be the dissociation of Volcano; 3 ⫽ Cañas; 4 ⫽ Arenal Volcano; 5 ⫽ Miravalles Volcano;
methane hydrates (Fehn et al., 2000) or the dewatering of pore 6 ⫽ Rincón de la Vieja Volcano; 7 ⫽ Baños Calientes; Tipitapa; 8 ⫽
fluids in shelf deposits (Martin et al., 1993). Forearc dewatering Lake Managua; 9 ⫽ Momotombo Volcano; 10 ⫽ Hervidero; San
Jacinto; 11 ⫽ Chinameca; 12 ⫽ Berlı́n; 13 ⫽ Lempa River; 14 ⫽ San
of subducted sediments may also return a significant amount of Vicente Volcano; 15 ⫽ Ahuachapán geothermal field.
iodine to the oceans. Roughly 1.3 ⫻ 109 tons of sediment are
subducted annually along 2970 km of marine trenches (Plank
and Langmuir, 1998). At this point, all of the iodine must leave Central American Volcanic Arc, focused on the geothermal
the marine sediment reservoir either to return to the surface via systems of Miravalles, Costa Rica; Momotombo, Nicaragua;
remobilization in the forearc or main volcanic zone or be and Ahuachapán and Berlı́n, El Salvador.
subducted back into the mantle. Iodine remobilized from ma-
rine sediments should have a specific isotopic composition 1.2. Geologic Setting
representing the age of iodine in the subducted sediments.
Determination of the 129I/I ratio in volcanic fluids can thus be The Central American Volcanic Arc is a result of subduction
used to determine origin of iodine and relative contributions of the Cocos Plate under the Caribbean Plate (Fig. 3). Much of
from subducted sediments or other potential sources. We in- the subducted oceanic floor was initially produced at the East
vestigated the application of the 129I system for these questions Pacific Rise; however, from central Costa Rica southward, the
as part of a larger project on the origin of fluids and gases in the crust is younger and was produced at the Costa Rica and
Equador Rifts, which separate the Cocos and Nazca Plates. The
Cocos Ridge also trails along the southern margin of the
subducting plate. The combined effect of younger crust in the
south, plus warming of the plate as it passed over the Galápagos
hot spot, provides enough buoyancy to produce shallowing of
the Wadati-Benioff zone and changes in the overall stress
regime (Protti et al., 1995). Very low heat flow measurements
seaward of the trench in Costa Rica contradict this notion
somewhat (Langseth and Silver, 1996), although it is not
known to what degree seawater circulation in the oceanic crust
has reduced the overall thermal regime of the subducting plate.
The age of the oceanic crust at the Central American Trench
is approximately 25 Ma, and it is assumed that this is also the
age of the oldest sediments subducted at this location. The
Fig. 2. Iodine fluxes (103 tons/yr). Sources: (1) NCRP (1983); (2) sediment column undergoing subduction includes the entire
Muramatsu and Wedepohl (1998); (3) subducted sediment flux of range from these oldest sediments to sediments of Recent age
Plank and Langmuir (1998) and marine sediment concentration (15
ppm) of Muramatsu and Wedepohl (1998). (4) Calculated from MORB
at the top. By use of the subduction geometry and the conver-
concentrations of Muramatsu and Wedepohl (1998) (9.5 ppb) and gence rate of 4.8 cm/yr in this region, subducted sediment ages
corresponding subduction flux of Plank and Langmuir (1998). between 2 and 27 Ma are to be expected for this arc. This age
 Origin of iodine in volcanic fluids of Central America 3829

range should be reflected in the iodine found in volcanic fluids sediments just above the crust, given by the age of the con-
if they are derived from subducted marine sediments. The verging plate plus the time of subduction:
crustal material of the Central American Volcanic Arc is prin-
cipally composed of marine volcanics and sedimentary se- R sub ⫽ R ie ⫺␭共t2⫹ts兲. (2)
quences, which have been accreted against the pre-Mesozoic In the case of Central America, the total age is 27 Ma, which
Chortis Block (Ranero et al., 2000; Walther et al., 2000). These yields a ratio of 456 ⫻ 10⫺15 and represents a lower estimate
include Cretaceous organic-rich sediments (Astorga, 1997), of the subducted iodine isotopic signature. These two values
which have been overthrust on the margins of the Caribbean will be used to evaluate the meaning of the ratios observed in
Plate and in some cases consumed by subduction erosion the volcanic fluids.
(Meschede, 1999). Isotope ratios determined this way provide minimum values
because potential additions from other sources such as in situ
1.3. 129
I System production of fissiogenic 129I in the sediments and anthropo-
genic input have not been taken into account. The contribution
The cosmogenic radioisotope 129I (T1/2 ⫽ 15.7 Ma) is pro- of fissiogenic 129I is a function of the U concentration and the
duced by the spallation of atmospheric Xe and by spontaneous age of the sediments. Given typical U concentrations in marine
fission of 238U in crustal rocks (Fabryka-Martin et al., 1985). sediments (1 to 3 ppm), this addition is generally within the
Iodine has one stable isotope (127I), and its isotopic composi- uncertainties of the approach even in older sediments (Moran et
tion is reported as 129I/I. Accelerator mass spectrometry (AMS) al., 1995; Fehn et al., 2000). More important for the application
is the preferred method for the detection of cosmogenic radio- of the system in crater lakes and hot springs is the potential
isotopes (Sharma et al., 2000), and approximately 1 mg of presence of anthropogenic 129I. Although the concentrations of
iodine (precipitated as AgI) is needed to produce a reliable several other radioisotopes released during atmospheric weap-
sample for AMS determination. Recent marine sediments with- ons tests have returned to prebomb levels in surface reservoirs,
out anthropogenic 129I have a 129I/I ratio (Ri) of 1500 ⫻ 10⫺15 129
I has remained at elevated levels because of the continuing
(Fehn et al., 1986; Moran et al., 1998), which is more than two release from nuclear reprocessing plants (Schink et al., 1995;
orders of magnitude above the AMS detection limit. This ratio Santschi et al., 1996; Wagner et al., 1996). The major sources
is independent of latitudinal variations in the production of for anthropogenic 129I are in the Northern Hemisphere, but
cosmogenic 129I because of the long residence time of iodine in concentrations of 129I are elevated in all surface reservoirs on a
the oceans (340,000 yr; Broecker and Peng, 1982). global basis (Rao and Fehn, 1999; Fehn and Snyder, 2000;
The general 129I systematics in subduction zones are gov- Oktay et al., 2000). Depending on the location, the addition of
erned by the isotopic equilibrium between ocean water and the anthropogenic 129I has increased the isotopic ratios in surface
top layers of sediments, the subsequent decrease of the initial reservoirs by two to six orders of magnitude above the natural,
ratio as a function of sediment age, and the remobilization of preanthropogenic, ratios (Moran et al., 1999). This large
iodine from the subducted sediments under the main volcanic change in isotopic ratios is, however, counterbalanced by the
zone. Assuming that the magmatic end member is derived from large difference in iodine concentrations between rainwater and
a subducted source, the 129I/I ratio may be estimated by means magmatic fluids: although concentrations in rainwater typically
of two approaches. The first is to assume that sediment is are around 0.015 ␮M (2 ppb) (Takaku et al., 1995; Moran et al.,
deposited at a uniform rate during the lifetime of the oceanic 1999), concentrations in geothermal and volcanic fluids gener-
crust—in this case, from the time at which the Cocos Plate was ally are above 1 ␮M and can reach values in excess of 1000
formed at the East Pacific Rise to the time at which it was ␮M (Snyder et al., 2002).
subducted ⬃25 Ma later. It may also be assumed that the
subducted material travels a mean distance of 188 km at a rate 2. SAMPLING AND DETECTION METHODS
of 8.6 cm/yr and a subduction angle of 32° (Protti et al., 1995)
Reconnaissance sampling was initiated in Nicaragua and Costa Rica
to arrive at the volcanic front 2 million years later. The iodine during July and August 1997 to determine which fluids contained
ratio of the subducted component is given by (Fehn et al., 2002) enough iodine to carry out 129I isotopic studies. Potential sampling
localities included hot springs, mud pots, flank fumaroles, crater lakes,
R int ⫽ R ie ⫺␭共t1⫹ts兲关1 ⫺ e ⫺ ␭ 共t 2 ⫺ t 1兲 兲/共 ␭ 共t 2 ⫺ t 1 兲兴, (1) and geothermal production wells (Fig. 3). Subsequent sampling trips
focused on geothermal production wells in Costa Rica, Nicaragua, and
where Rint is the integrated ratio, t1 is the age of the youngest El Salvador because these generally have high iodine concentrations
(mostly ⬎1 ppm) and are less likely than surface samples to contain
subducted sediments (in this case, 0 Ma), t2 is the age of the anthropogenic 129I. Further sampling of fluids and gases was carried
oldest subducted sediments (in this case, 25 Ma), and ts is the out between 1998 and 1999. Samples from rivers, springs, lakes, and
calculated time between subduction and reemergence (in this mud pots were collected in 1-L HDPE bottles. Temperature measure-
case, 2 Ma). The resulting 129I/I ratio is 824 ⫻ 10⫺15 Ma. ments were also taken at the time of sampling. Two-liter samples were
collected from the Poás and Rincón de la Vieja crater lakes with bottles
Because this approach assumes that the bulk composition of the
tied to nylon rope. Condensates were sampled from low-temperature
subducted sediment is preserved, it is an upper limit. A more fumaroles (95°C ⬍ T ⬍ 200°C) with a hand pump, Teflon tubing, and
likely case is that initial compaction and dewatering in the a 1-L HDPE bottle cooled in a bucket of ice. Fluid samples were taken
forearc preferentially removes pore fluids in the shallow sedi- from geothermal production wells, either from the steam separators or
ments, which are then replaced by deeper fluids as compaction by attaching a Webber miniseparator to the steam lines between the
wellhead and the separators. One-liter fluid samples were condensed
proceeds. The bulk composition should therefore represent an from the miniseparator at atmospheric pressure by means of ice or
iodine isotopic signature of the deeper fluids. The second running water with a copper or stainless steel condenser.
approach simply reflects the decay of the input ratio Ri in the Chloride concentration was determined with a Dionex Bio-LC ion
3830 G. T. Snyder and U. Fehn

chromatograph with an anion micromembrane suppressor (Dionex Br/Cl ratios do not reflect the composition of the degassing
AMMSI), a guard column (Dionex AG4A), an anion column (Dionex magma.
AS4A), an eluant made up of 1.8 mM sodium carbonate and 1.7 mM
Ratios of 129I/I are with few exceptions below that of the
sodium bicarbonate, and a regenerant of 25 mN sulfuric acid (Weiss,
1986). Standards were run on every third injection, and multiple runs of preanthropogenic marine ratio of 1500 ⫻ 10⫺15, indicating that
individual samples indicated that instrumental precision errors were anthropogenic contributions are not appreciable in Central
generally below 5%. Total iodine and bromine concentrations were American geothermal systems. Fumaroles, hot springs, mud
determined with a VG Plasmaquad II at the Rochester inductively pots, and crater lakes generally present compositions between
coupled plasma mass spectrometry (ICP-MS) laboratory by methods
described in Schnetger and Muramatsu (1996), Félinas et al. (1998), those of the geothermal fields and the meteoric samples col-
and Takaku et al. (1995). A 5% tertiary amine solution (Spectrasol lected from rivers and lakes. The Lake of Managua has anom-
CFA-C) was added to the solution matrix to improve the washout time alously high iodine and bromine concentrations and the lowest
of iodine and enhance signal strength on the ICP-MS. A washout time 129
I/I ratio of the surface samples, which is either caused by
of 3 min was sufficient to return the signal strength at mass 127 to
subsurface hydrothermal activities (the Momotombo geother-
background levels. Because ICP-MS records total element concentra-
tions, we will report concentrations for iodine and bromine, in contrast mal field is located on the shoreline) or by evaporation. Iodine
to chloride, determined by ion chromatography. Samples were prepared ratios were converted into minimum iodine ages solving for
for AMS at the Rochester Cosmogenic Isotope Laboratory by using time in the decay equation
methods outlined in Fehn et al. (1992). The AMS analysis of 129I was
carried out at the Purdue Rare Isotope Measurement Laboratory, Pur- R obs ⫽ R ie ⫺␭t, (3)
due University (Sharma et al., 2000). AMS determination of iodine
ratios generally has a 1␴ error between 5 and 10%; the individual errors where the values for the variables are the same as in Eqns. 1
and the resulting age resolutions are indicated. Specifics of sample
and 2. These are minimum ages because contributions from
locations, detection methods, and major ion concentrations can be
found in Snyder (2001). fissiogenic production or contamination from meteoric water
will produce higher ratios and thus anomalously young ages.
3. RESULTS Figure 6 shows the compilation of the ratios, converted to ages
and compared with the ratio calculated for the integrated sed-
Concentrations of I, Br, and Cl and isotopic data for Central iment column (Eqn. 1) and for the oldest subducted sediments
America are presented in Table 1. Concentrations of iodine, (Eqn. 2). The majority of the results fall between those two
bromine, and chloride were corrected for steam loss at the point ratios, but there are also brines with iodine ages younger than
of sampling by using recorded sampling pressures and a reser- the subducted Cocos Plate, as well as several wells with iodine
voir enthalpy corresponding to saturated steam at the temper- ages that are significantly older. Although many of the wells
ature indicated through Na-K-Ca geothermometry (Snyder, have iodine ages clustered between 15 and 30 Ma, there are no
2001). Despite the corrections, significant variability (more wells with iodine ages between 30 and 40 Ma, but a group of
than one order of magnitude) exists between both the iodine wells has ages between 40 Ma and 50 Ma, and one well in the
and chloride concentrations in each geothermal field. Nonethe- Momotombo geothermal field has an age of 65 Ma.
less, the ratio of iodine to chloride remains fairly constant in If the results are plotted according to geographic location
each geothermal field (Fig. 4). With the exception of the Berlı́n (Fig. 7), no systematic regional variations are apparent in the
geothermal field, most of the geothermal fields have I/Cl ratios 129
I/I ratios throughout the sample region, indicating that sub-
between 15 ⫻ 10⫺5 and 20 ⫻ 10⫺5, which are substantially duction angle or other tectonic parameters apparently do not
higher than that of seawater (8 ⫻ 10⫺7). The Berlı́n field is substantially influence the iodine time signature. The
nearly twice as high as the majority of the geothermal fields, Miravalles geothermal field has more wells with 129I/I ratios
possibly reflecting the addition of crustal iodine. Samples from approaching those of meteoric water. In contrast to the 129I/I
Poás and Rincón de la Vieja crater lakes have much lower I/Cl ratios, iodine concentrations show a more systematic variation
ratios, which is not unusual because both are open systems. The along the Central American Arc, with concentrations being
drop in I/Cl ratios may be due to any of a number of factors, generally twice as high in El Salvador than toward the south
including leaching of chloride, preferential sorption of iodine (Fig. 8).
onto sulfur mineralizations, or the volatile release of iodine Only small changes were observed in 129I/I ratios for the
from the lake surface. Central American wells that were sampled more than once
Groundwater generally has Br/Cl ratios of 1.7 ⫻ 10⫺3 to 8 ⫻ during the 3 years of field sampling. Eight geothermal wells
10⫺3, which is only slightly higher than average seawater and the Lake of Managua were all sampled on two or three
values of 1.5 ⫻ 10⫺3 (Davis et al., 2000). The geothermal occasions. Only four wells showed changes in 129I/I that were
samples are all enriched in bromide relative to chloride, but by greater than the precision error of the samples. Although the
less than an order of magnitude. Because bromine is less concentration of total iodine changed in the Lake of Managua,
readily incorporated in organic sediments than iodine but more the 129I/I ratio remained essentially the same.
readily than chloride (Martin et al., 1993; Egeberg and Dickens,
1999), this is not inconsistent with the derivation of halides 4. DISCUSSION
from a subducted source. Ratios of Br/Cl are also less in crater
lakes (Fig. 5). Substantial drops in Br/Cl ratios have been To determine whether the iodine isotopic signature is truly
documented in groundwater systems where dissolution of ha- indicative of a subducted source, it is necessary to assess the
lides is apparent (Davis et al., 2000), and in arid regions where contributions of other sources and their effect on 129I/I ratios.
leaching of salt crusts occurs (Risacher and Fritz, 2000). In the This is not a simple task because geothermal systems undergo
case of crater lakes, it is likely that leaching from the crater a variety of processes, the majority of which can alter the
walls is a source of the additional chloride, and that the low magmatic signature in terms of halide concentrations, and
 Origin of iodine in volcanic fluids of Central America 3831

129
Table 1. Halide composition and I/I ratios.
129
TNa-K-Ca Icorr Clcorr Brcorr I/I
Samplea Location (°C)b X steam (mM) (mM) (mM) (10⫺15)

El Salvador
AH06-9908SA Ahuachapán Geothermal 243 0.183 0.0443 219 0.478 231 ⫾ 27
AH16A-9908SA Ahuachapán Geothermal 228 0.137 0.0501 82 0.481 253 ⫾ 33
AH17-9908SA Ahuachapán Geothermal 188 0.002 0.0015 4 0.013 ND
AH20-9908SA Ahuachapán Geothermal 241 0.165 0.0406 189 0.465 842 ⫾ 45
AH21-9908SA Ahuachapán Geothermal 247 0.181 0.0517 257 0.569 257 ⫾ 24
AH22-9908SA Ahuachapán Geothermal 238 0.172 0.0352 176 0.356 764 ⫾ 35
AH23-9908SA Ahuachapán Geothermal 240 0.165 0.0359 178 0.407 571 ⫾ 35
AH4B-9908SA Ahuachapán Geothermal 228 0.145 0.0378 169 0.471 802 ⫾ 45
TR02-9908SA Berlin Geothermal 267 0.236 0.0446 111 0.289 1069 ⫾ 55
TR4B-9908SA Berlin Geothermal 260 0.188 0.0375 87 0.247 257 ⫾ 25
TR4C-9908SA Berlin Geothermal 282 0.163 0.0522 98 0.394 564 ⫾ 34
TR09-9908SA Berlin Geothermal 274 0.245 0.0592 155 0.337 721 ⫾ 29
LV01-9908SA La Viejona, Chipilapa (FC) 147 0.087 0.000003 0.003 0.000039 ND
IF01-9908SA Infiernillos, Chipilapa (FC) 153 0.100 0.000002 0.007 0.000016 ND
IF01B-9908SA Infiernillos, Chipilapa (MP) 229 0.251 0.000023 0.009 0.000221 ND

Nicaragua
MT02-9802NI Momotombo Geothermal 246 0.193 0.0091 68.0 0.071 1120 ⫾ 120
MT27-9802NI Momotombo Geothermal 246 0.192 0.0094 66.3 0.090 810 ⫾ 90
MT27-9903NI Momotombo Geothermal 219 0.219 0.0078 50.8 0.080 730 ⫾ 50
MT35-9802NI Momotombo Geothermal 281 0.275 ND 78.4 ND 590 ⫾ 90
MT35-9903NI Momotombo Geothermal 259 0.228 0.0118 63.9 0.109 567 ⫾ 34
MT36-9708NI Momotombo Geothermal 308 0.329 0.0211 96.6 0.171 85 ⫾ 25
MT3643-9903NI Momotombo Geothermal 313 0.433 0.0194 84.6 0.142 200 ⫾ 14
MT38-9802NI Momotombo Geothermal 263 0.224 0.0132 88.5 0.106 630 ⫾ 90
MT38B-9903NI Momotombo Geothermal 245 0.273 0.0105 59.6 0.100 730 ⫾ 32
MT40-9708NI Momotombo Geothermal 255 0.350 0.0135 82.3 0.109 532 ⫾ 46
MT40-9802NI Momotombo Geothermal 266 0.232 0.0132 65.9 0.114 540 ⫾ 100
MT40B-9903NI Momotombo Geothermal 246 0.273 0.0100 57.7 0.096 750 ⫾ 60
MT42-9802NI Momotombo Geothermal 231 0.155 0.0171 97.9 0.150 441 ⫾ 41
MT42-9903NI Momotombo Geothermal 260 0.226 0.0145 78.0 0.129 523 ⫾ 29
MT43-9903NI Momotombo Geothermal 194 0.166 0.000020 0.094 0.0014 ND
AZ01-9802NI Azulfral, Momotombo (FC) 281 0.352 0.000003 0.013 0.00001 ND
ROB01-9802NI R. Obraje, (HS) ND NA 0.001331 4.570 0.006 88 ⫾ 5610
HSJ01-9708NI Hervidero, San Jacinto (MP) 137 0.057 0.000071 0.027 0.004 18,593 ⫾ 5961
BCT01-9708NI Tipitapa (HS) 149 0.080 0.000935 6.644 0.013 700 ⫾ 700

Costa Rica
CAS01-9708CR Finca ANDE, Cañas (HS) 176 0.145 0.000215 19.515 0.022 1048 ⫾ 1594
CAS02-9802CR Finca ANDE, Cañas (HS) 177 0.147 0.000457 19.570 0.022 ND
PGM01-9802CR Miravalles Geothermal 227 0.137 0.0123 82.6 0.127 500 ⫾ 36
PGM02-9908CR Miravalles Geothermal 228 0.231 0.087 36.1 0.124 ND
PGM03-9803CR Miravalles Geothermal 241 0.174 ND 85.6 ND 1760 ⫾ 200
PGM03-9802CR Miravalles Geothermal 227 0.140 0.0125 75.0 0.168 ND
PGM03-9908CR Miravalles Geothermal 223 0.133 ND 77.5 ND 1100 ⫾ 100
PGM05-9802CR Miravalles Geothermal 228 0.099 0.0131 100.2 0.183 690 ⫾ 80
PGM10-9802CR Miravalles Geothermal 237 0.157 0.0099 99.1 0.156 580 ⫾ 46
PGM11-9802CR Miravalles Geothermal 243 0.148 0.0117 80.5 0.158 677 ⫾ 51
PGM12-9802CR Miravalles Geothermal 227 0.146 0.0138 102.7 0.183 1180 ⫾ 60
PGM19-9708CR Miravalles Geothermal 237 ND ND ND ND 1580 ⫾ 120
PGM20-9802CR Miravalles Geothermal 229 0.134 0.0153 113.7 0.168 965 ⫾ 49
PGM21-9802CR Miravalles Geothermal 222 0.114 0.0136 109.7 0.173 950 ⫾ 70
PGM29-9708CR Miravalles Geothermal 210 0.075 0.0082 23.5 0.101 200 ⫾ 70
PGM29-9802CR Miravalles Geothermal 258 0.187 0.0151 83.1 0.179 264 ⫾ 45
PGM31-9802CR Miravalles Geothermal 242 0.180 0.0120 102.2 0.178 580 ⫾ 70
PGM45-9908CR Miravalles Geothermal 228 0.145 0.0162 80.8 0.227 1050 ⫾ 80
PGM46-9802CR Miravalles Geothermal 228 0.071 0.0169 129.4 0.228 1560 ⫾ 120
PGM46-9908CR Miravalles Geothermal 221 0.127 0.0164 88.7 0.231 1030 ⫾ 130
POS01-9708CR Miravalles Geothermal 197 NA 0.000001 0.0685 0.002 771,262 ⫾ 154,870
VMV019708CR Hornillas, Miravalles (MP) 247 NA 0.000031 0.0474 0.0004 800 ⫾ 600
VMV039708CR Miravalles steam, SAT 1 227 ND ND ND ND 512 ⫾ 59
VAR01-9708CR Quebrada Lava, V. Arenal 168 0.129 ND 3.7 ND ND
VAR02-9708CR Tabacón, V. Arenal 182 0.156 0.00019 7.6 0.008 142,304 ⫾ 7425
VPOCL-9704CR V. Poas (CL) 230 0.253 0.00190 106.0 0.145 ND

(continued)
3832 G. T. Snyder and U. Fehn

Table 1. Continued
129
TNa-K-Ca Icorr Clcorr Brcorr I/I
Samplea Location (°C)b Xsteam (mM) (mM) (mM) (10⫺15)

VPOCL-9708CR V. Poas (CL) 190 0.172 0.00091 126.4 0.134 2600 ⫾ 2000
VPOCL-9903CR V. Poas (CL) 204 0.200 0.00130 292.6 0.282 2620 ⫾ 1500
VRVCL9302CR Rincón de la Vieja (CL) 261 0.319 0.01792 356.2 0.627 ND
VRVCL9903CR Rincón de la Vieja (CL) 227 0.246 0.01050 300.5 0.302 1980 ⫾ 180

Metoric
VIZ01-9708CR Laguna Jaular, Irazu NA NA 0.000003 0.033 0.000124 365,000 ⫾ 352,040
LMG01-9708NI Lake Managua NA NA 0.00027 5.898 0.016424 158,380 ⫾ 17,510
LMG02-9802NI Lake Managua NA NA 0.00017 6.478 0.008736 294,690 ⫾ 18,580
VPOLB-9708CR Laguna Botos, V. Poas NA NA 0.000012 0.016 0.000102 477,310 ⫾ 108,210
RCA01-9708CR Represa Cachı́, Cartago NA NA 0.000025 0.066 0.00159 117,250 ⫾ 50,080
VMV289802CR Ret. Tank. Miravalles NA NA 0.000004 0.146 0.000186 85,390 ⫾ 242,770
VMV269802CR Rio Cuipilapa, Miravalles NA NA 0.000015 0.173 0.000349 31,240 ⫾ 64,020
RL01-9908SA Rio Lempa NA NA 0.000006 0.080 0.000119 579,270 ⫾ 103,290
a
Sample types: FC ⫽ fumarolic condensate, MP ⫽ mud pot, HS ⫽ hot spring, CL ⫽ crater lake, ND ⫽ not determined. NA ⫽ not applicable.
Where steam fraction ⫽ NA, the concentrations are not corrected for steam loss. All others are corrected for the loss of the indicated steam fraction
at the time of sampling.
b
Temperatures derived from Na-K-Ca geothermometry (Fournier and Truesdell, 1973).

several of which can potentially change the iodine isotopic are 8.1 for Cl, 17.5 for Br, and 104 for I (Bureau et al., 2000).
signature. Partitioning during phase changes will affect the If magmatic degassing were the sole source of iodine and
relative abundance of the halides without changing the ratio of chloride in geothermal systems, then the I/Cl ratios of 15 ⫻
129
I/I beyond a small fraction of the AMS precision because the 10⫺5 to 20 ⫻ 10⫺5 observed in Central American geothermal
relative mass difference between 129I and 127I is small. The fields would require a magma body to have an I/Cl ratio of only
relative abundance of iodine, chlorine, and bromine in high- 1.2 ⫻ 10⫺5 to 1.6 ⫻ 10⫺5. Because this value is still much
temperature magmatic fluids is a function of the halide con- higher than that of seawater (I/Cl ⫽ 0.08 ⫻ 10⫺5; Broecker and
centration in the melt as well as the fluid/melt partition coef- Peng, 1982), direct derivation from seawater is not likely.
ficients for each of the halides. Experimental data in albitic The variability in concentrations and iodine isotopic signa-
melts suggests that the fluid/melt partition coefficients (Df/m) tures within each geothermal field suggests that partitioning of
the magma source is not as important as other shallow crustal
processes. As the magmatic fluids approach the surface, phase
separation may either increase or decrease concentrations, de-
pending on whether the fluid sampled is the residual brine or

Fig. 4. Chloride and total iodine concentrations. Concentrations have

been corrected for steam loss at the time of sampling. Open dia-
monds ⫽ Ahuachapán Geothermal; open circles ⫽ Berlı́n Geothermal;
open triangles ⫽ Momotombo Geothermal; open squares ⫽ Miravalles
Geothermal; solid triangles ⫽ hot springs, mud pots, and fumaroles;
solid squares ⫽ hyperacidic crater lakes. Rivers and lakes are plotted as Fig. 5. Chloride and bromide concentrations. Symbols as in Figure
asterisks near the origin. 4.
 Origin of iodine in volcanic fluids of Central America 3833

Fig. 6. Decay curve for Central American geothermal samples. Fig. 8. Iodine concentrations along the arc. High concentrations in El
Iodine ratios below preanthropogenic values (1500 ⫻ 10⫺15; Moran et Salvador are likely due to shallow crustal processes rather than changes
al., 1998) are plotted against the corresponding minimum iodine ages in the flux from subducted organic material. Symbols as in Figure 4.
based on Eqn. 3. Symbols are the same as Figure 4. Bars denote 1␴
errors in iodine ratios and corresponding age errors. Predicted values
for subduction of an integrated sediment column (Eqn. 1) and the oldest
subducted sediments (Eqn. 2) are indicated by dotted and dashed lines, neither of these processes will produce changes in 129I/I ratios.
respectively.
The release of iodine through water–rock interactions and
dissolution of recent igneous rocks by magmatic H2SO4 and
HCl is also unlikely to change the 129I/I ratios because the
the condensed vapor. As the fluids cool, the adsorption of
iodine signature of igneous rocks younger than 2 Ma is within
iodine on metallic sulfides may change the iodine concentration
the AMS precision error of the presently degassing magma
(Fuhrmann et al., 1998), as will mineralization involving iodine
bodies.
and chloride complexes (Gammons and Yu, 1997); however,
Two processes in the shallow crust can appreciably change
the 129I/I ratios, along with overall concentrations. Contamina-
tion of the magma body through thermogenic release of the
halogens in organic material will increase the concentration of
iodine in the volatiles and could potentially have a significant
effect on 129I/I ratios. The chloride concentration would for the
most part be unaffected, and bromide concentrations would
increase only moderately (You and Gieskes, 2001). Interaction
with groundwater may also decrease concentrations while in-
creasing 129I/I ratios. The relative effect that groundwater has
on the 129I/I ratios magmatic fluids depends on whether it
contains anthropogenic 129I. The addition of crustal and mete-
oric components to the magmatic source may be addressed by
looking at enthalpy and mixing relationships.

4.1. Enthalpy and Cl and I Concentrations

Both the flashing of steam in geothermal systems and dilu-
tion from groundwater can cause a decrease in measured con-
centrations of ions that readily partition into the fluid phase.
Enthalpy diagrams may be used to distinguish between the two
because steam flashing from magmatic systems will carry ex-
cess enthalpy with it, and groundwater will have low enthalpy.
Fig. 7. Iodine ratios along the Central American Volcanic Arc. The reservoir enthalpy in the Central American fields was
Distances are recorded as kilometers from the Guatemala-Mexico determined via Na-K-Ca geothermometry (Fournier and Trues-
border (Patino et al., 2000). Symbols as in Figure 4. Horizontal dotted
and dashed lines correspond to predicted ratios as in Figure 5. Contri-
dell, 1973) and assuming that the geothermal reservoir has
butions from an anthropogenic component are evident only in thermodynamic properties that are similar to steam-saturated
Miravalles, Costa Rica. water (Haar et al., 1984).
3834 G. T. Snyder and U. Fehn

Fig. 9. Iodine concentrations vs. enthalpy. Enthalpy values corre- Fig. 10. Chloride concentrations vs. enthalpy. Symbols as in Figure
spond to reservoir temperatures determined through Na-K-Ca geother- 4.
mometry (Fournier and Truesdell, 1973). Mixing lines are discussed in
the text. Symbols as in Figure 4.
the Lake of Managua, which shows evidence of high evapora-
tive loss, subsurface geothermal input, or both.
Diagrams were constructed for iodine (Fig. 9) and chloride Several trends can be distinguished in Figure 9. Fumaroles,
vs. enthalpy (Fig. 10). The position of the magmatic end hot springs, and mud pots all plot between rivers or lakes and
member was determined by first constructing a mixing line steam, indicating that the major source of heat in the system is
between pure steam and the geothermal well with the highest due to phase separation in the geothermal reservoir where the
chloride concentration (AH21). Although higher concentra- steam rises and condenses in shallow groundwater, contributing
tions of iodine are present in Berlı́n, they do not have corre- an excess of enthalpy to the shallow systems without adding
spondingly high chloride concentrations and may represent the iodine. Geothermal wells, on the other hand, show a bimodal
addition of iodine from a crustal source. A regression line was distribution, which is also reflected in the chloride concentra-
constructed for the Momotombo samples because they present tions (Fig. 10). The Ahuachapán and Berlı́n geothermal fields
wells with the highest enthalpy and the largest range of en- in El Salvador both show steam loss that has resulted in an
thalpy values within a given geothermal system. The enthalpy increase in iodine and chloride concentrations, as well as a
of the magmatic end member was averaged between the chlo- decrease in enthalpy. The most depleted end member (AH21)
ride diagram (H ⫽ 1642.5 kJ/kg) and the iodine diagram (H ⫽ has an enthalpy of 1071 kJ/kg and iodine and chloride concen-
1587.5 kJ/kg), and the average enthalpy (H ⫽ 1615 kJ/kg) was trations of 51 and 251 mM, respectively. The other wells in El
applied to the equation for the mixing line between pure steam Salvador show dilution of the steam-depleted end member by
and the steam-depleted well (AH21). The resulting magmatic groundwater. The Momotombo geothermal field in Nicaragua
concentrations of I and Cl⫺ are 34 and 170 mM, respectively, has apparently the least amount of steam loss, with direct
at a reservoir pressure of 152 bar and a temperature of 343°C. dilution of the magmatic source by groundwater. In the
Similarly, enthalpy of the groundwater end member was taken Miravalles field in Costa Rica, groundwater mixing occurs
as the average of the intercepts of the regression lines for the between both magmatic and steam-depleted wells. It is note-
chloride (H ⫽ 809.79 kJ/kg) and iodine (H ⫽ 797.72 kJ/kg) worthy that the mixing trends do not converge on groundwater
diagrams, and the average enthalpy (H ⫽ 804 kJ/kg) corre- derived from local rivers and lakes, but on groundwater that has
sponds to a pressure of 12 bar and a temperature of 185°C. attained thermal equilibrium at depth. The Poás and Rincón de
Concentrations of the groundwater end member (Table 2) were la Vieja crater lakes show extreme steam loss and high chloride
taken as the average of rivers and lakes, with the exception of concentrations. Iodine concentrations are not correspondingly

Table 2. End member compositions.

Source Symbol 129

I/I (10⫺15) 129
I (10⫺6 atom/L) Icorr (␮M) Clcorr (mM)

Preanthropogenic groundwater PRE 1500 0.0097 0.011 0.009

Anthropogenic groundwater ANT 260,000 1.69 0.011 0.009
Magmatic MAG 456 9.3 34 170
 Origin of iodine in volcanic fluids of Central America 3835

Fig. 11. Variations in reservoir enthalpy and iodine isotope ratios. Fig. 12. Mixing of groundwater with a magmatic iodine source.
Solid lines denote mixing trends with a magmatic end member con- Combining preanthropogenic (dashed line) and anthropogenic (solid
taining iodine from only the oldest subducted sediments. The ground- line) groundwater with the magmatic end member results in flat mixing
water end member consists of preanthropogenic meteoric water. Sym- curves due to the high concentration of iodine in the magmatic end
bols as in Figure 4. member. Steam loss affects only iodine concentrations; additions from
older crustal organics lower iodine ratios and raise iodine concentra-
tions. Asterisks indicate meteoric surface samples; other symbols as in
Figure 4.
high in crater lakes, perhaps indicating that iodine is not con-
servative in hyperacidic, sulfur-rich environments.
pogenic water, which in both cases would greatly lower the
4.2. Iodine Isotopic Composition and the Magmatic End concentration. Although the position of many of the wells
Member might suggest mixing between the magmatic end member and
anthropogenic water, it is more likely that steam loss has
129
I/I ratios were compared with the expected values for the affected iodine concentrations in the Miravalles, Berlı́n, and
deepest sediments and for the integrated sediment column (Fig. Ahuachapán fields and that the additional water source is
6). As mentioned earlier, the overall observations are that most preanthropogenic water in most cases. The Momotombo wells
of the ratios cluster in the area below those of the integrated show a decrease in iodine ratios with increasing concentration,
sediment column; that there are ratios with ages younger than indicating direct mixing between magmatic and meteoric wa-
the integrated values; and that a small but significant number of ters, although it is unclear from the ratios and concentration
samples have minimum ages considerably older than the oldest data whether the meteoric component is anthropogenic water or
sediments subducted in Central America. Because of the large whether the samples have mixed with preanthropogenic water
variation observed in the ratios, a closer look at mixing and after losing steam.
potential end members is required for the interpretation of the The 129I results indicate the presence of three dominant end
results. members: a magmatic end member with a ratio compatible with
Similar to the previous two diagrams, a diagram was con- derivation from the lower parts of the subducted sediment
structed for the relationship between reservoir enthalpy and column, a crustal contribution with an age older than 40 Ma and
iodine ratios (Fig. 11). Most of the wells indicate that the additions from meteoric sources either with preanthropogenic
magmatic fluids have undergone steam loss followed by dilu- or anthropogenic characteristics. The relationship between
tion by preanthropogenic groundwater. A number of samples these three end members is illustrated by plotting 129I vs. I,
show that the magmatic source has been contaminated by older both normalized to Cl in the wells, with slopes indicating the
crustal iodine as well, lowering both the iodine ratios and the potential end members (Fig. 13). The increase in 129I/Cl ratios
enthalpy. Figure 12 shows the effect of addition of preanthro- observed in each of the fields follows roughly the line indicat-
pogenic groundwater and anthropogenic groundwater on the ing mixing with preanthropogenic water. On the basis of these
magmatic end member. Because both anthropogenic and pre- results, we conclude that the iodine in the fluids, which is
anthropogenic groundwaters have very low iodine concentra- derived from both subducted and older crustal sources, has
tions, the magmatic brines can undergo a certain degree of been mixed with meteoric water, predominantly of preanthro-
dilution without producing a significant change in 129I/I ratios. pogenic age. The presence of the crustal end member may
Shifting the 129I/I ratios more than 10% (which in many cases possibly be related to sedimentation during the development of
is the precision error) would require admixture of at least 99% the Nicaraguan Depression on the margins of the Central
preanthropogenic water, or an addition of at least 35% anthro- American Chortis Block (Astorga, 1997; Ranero et al., 2000;
3836 G. T. Snyder and U. Fehn

of iodine is between 200 and 7700 tons/yr from arc volcanoes.

The large uncertainty reflects the variability of measured chlo-
ride fluxes in volcanoes around the world (Symonds et al.,
1988). Compared with the flux of subducted iodine (Fig. 2), the
release of iodine from arc volcanoes represents 1 to 38% of the
iodine that reaches the convergent margin. The presence of an
older age signature in Central America suggests that these
estimates are an upper limit for subduction recycling of iodine
(and possibly chloride) due to the incorporation of crustal fluids
as well. Barring incorporation of large amounts of iodine into
the mantle, the rest must be returned to the ocean in forearc
regions (Muramatsu et al., 2001).

5. CONCLUSIONS

Iodine isotopic ratios and concentrations of iodine and chlo-

ride demonstrate that the magmatic end member is predomi-
nantly derived from the oldest subducted sediments. Although
the sediment column may be subducted nearly in its entirety,
dewatering and compaction have apparently removed the up-
permost portion of the fluid mobile elements. This observation
Fig. 13. 129I and I normalized to chloride concentrations. Slopes
represent iodine ratios for preanthropogenic meteoric (Tmin ⫽ 0 Ma), suggests on the one hand that iodine is still present in marine
subduction of iodine from the entire sediment column (Tmin ⫽ 13.5 sediments at the main volcanic zone, and on the other hand that
Ma), subduction of only the oldest iodine (Tmin ⫽ 27 Ma), and the a large fraction of it is already removed in the forearc region.
oldest recorded iodine age in Central America (Tmin ⫽ 65 Ma). Mixing Overall, the results indicate that remobilization of iodine from
trends are indicated between preanthropogenic (PRE) or anthropogenic
(ANT) water and magmatic fluids (MAG). marine sediments occurs during the subduction process and
prevents most, if not all, of the iodine from being introduced
into the mantle. Although these data alone do not resolve
Walther et al., 2000). In all the cases, the lowest ratios were inconsistencies with the overall global iodine budget, they do
found on the margins of the individual geothermal systems, suggest that the majority of the iodine that arrives at the ocean
whereas the portions of the fields that are most heavily ex- trenches is returned to the oceans in the forearc region, and a
ploited for geothermal energy had ratios closest to the mag- small portion is actually subducted and returned in geothermal
matic end member. This observation suggests that the bulk of fluids associated with island arc magmatism.
the geothermal fluids is derived from deeper magmatic sources, Steam loss has increased iodine concentrations significantly
reflecting the iodine characteristics of subducted sediments, in the Berlı́n and Ahuachapán geothermal fields, and to a lesser
whereas fluids at the margins derive part or all of their iodine degree in the Miravalles field, without changing the 129I/I
from older, crustal sources. ratios. The relationship between enthalpy and geochemistry
indicates that variations within each geothermal field are due to
shallow crustal processes, rather than heterogeneities in the
4.3. Iodine Flux from Volcanic Arcs
subducted source. After the loss of steam, all of the fluids have
The flux of iodine from volcanic centers may be calculated subsequently undergone varying degrees of dilution by ground-
from ultraviolet correlation spectroscopy (COSPEC), SO2 flux water. The present position of the wells in diagrams of 129I/I vs.
determinations, previously published ratios of HCl/SO2, and iodine and chloride concentrations cannot determine with cer-
the I/Cl readings from this study. Poas Volcano has an HCl flux tainty whether the slight elevation in 129I/I ratios is caused by
of 0.46 kg/s (Casadevall et al., 1984). If we assume that the I/Cl a very small addition of anthropogenic groundwater or a large
ratio for high temperature fumarolic systems is similar to that addition of preanthropogenic groundwater accompanied by
of Miravalles (15 ⫻ 10⫺5), then Poas has an iodine flux of 7.59 concentration through steam loss.
ton I per year, which is nearly identical to the flux of 7 tons/yr The recent influx of 129I from bomb testing and nuclear
reported at White Island, New Zealand (Tedesco and Toutain, reprocessing has not substantially changed the 129I signature of
1991), and only slightly lower than 12 tons/yr estimated for the geothermal fluids because of the elevated total iodine concen-
Satsuma-Iwojima Volcano in Japan (Snyder et al., 2002). Al- tration in these regions. All of the Central American geothermal
though COSPEC data are not available for Momotombo, the fields contain iodine that is derived from a crustal source with
HCl flux from Masaya Volcano, ⬃50 km to the south of ages between 10 and 35 million years older than the subducted
Momotombo, is 7.8 kg/s (Burton et al., 2000). Assuming the crust. In the case of Momotombo, this older iodine is probably
same I/Cl ratio as Momotombo (20 ⫻ 10⫺5), the resulting flux associated with sediments deposited within the Sandino Basin,
of iodine from Masaya is 170 tons/yr—that is, considerably which has been subsequently remobilized during a period of
higher than for the other cases. By extrapolating the Central westward migration of the present island arc system (Ranero et
American range of typical I/Cl ratios (15 ⫻ 10⫺5 to 20 ⫻ 10⫺5 al., 2000; Walther et al., 2000). Another possibility is that
mol/mol), and by using the global flux of chlorine from volca- portions of the Caribbean Plate have undergone subduction
noes (13 to 350 kg/s HCl; Symonds et al., 1988), the global flux erosion (Meschede et al., 1999) carrying down sediments with
 Origin of iodine in volcanic fluids of Central America 3837

high total organic carbon, such as shales of the Upper Creta- Fabryka-Martin J., Bentley H., Elmore D., and Airey P. L. (1985)
ceous Lomachumico Formation (Astorga, 1997). The source of Natural iodine-129 as an environmental tracer. Geochim. Cosmo-
chim. Acta 49, 337–347.
organic matter in these fluids is discussed in greater detail by
Fehn U., Holdren G. R., Elmore D., Brunelle T., Teng R., and Kubik
Snyder (2001), which deal with the nitrogen and methane P. W. (1986) Determination of natural and anthropogenic I-129 in
composition of geothermal fluids. marine sediments. Geophys. Res. Lett. 13, 137–139.
Fluids in forearc and island arc regions are a significant Fehn U., Peters E. K., Tullai-Fitzpatrick S., Kubik P. W., Sharma P.,
reservoir for both total iodine and 129I. The results of this Teng R. T. D., Gove H. E., and Elmore D. (1992) 129I and 36Cl
concentrations in waters of the eastern Clear Lake area, California:
investigation suggest that remobilization of iodine and of ele- Residence times and source ages of hydrothermal fluids. Geochim.
ments with similar geochemical behavior constitutes an impor- Cosmochim. Acta 56, 2069 –2079.
tant part in the marine cycle of volatile elements. Similar results Fehn U. and Snyder G. (2000) 129I in the Southern Hemisphere: Global
from a few volcanic systems in Japan (Snyder et al., 2002), redistribution of an anthropogenic isotope. Nucl. Inst. Met. Phys.
Res. B. 172, 366 –371.
Argentina, and New Zealand (Fehn et al., 2002) suggest that
Fehn U., Snyder G. T., and Egeberg P. K. (2000) Dating of pore fluids
iodine recycling from marine sediments is common and can be with 129I: Relevance for the origin of marine gas hydrates. Science
detected by using the 129I system. 289, 2332–2335.
Fehn U., Snyder G. T., and Varekamp J. (2002) Detection of recycled
Acknowledgments—We thank the Central American institutions in marine sediment components in crater lake fluids using 129I. J.
charge of overseeing and carrying out geothermal operations: Instituto Volcanol. Geotherm. Res., in press.
Costarricense de Electricidad (ICE), Empresa Nicaragüense de Elec- Félinas Y., Krushevska A., and Barnes R. M. (1998) Determination of
tricidad (ENEL), and Geotérmica Salvadoreña (GESAL). Particular total iodine in nutritional and biological samples by ICP-MS follow-
thanks go to Guillermo Alvarado, Eddy Sánchez, Karla Miranda, ing their combustion within an oxygen stream. Anal. Chem. 70,
Ernesto Martı́nez, José Tenorio Mejı́a, and Carlos Pullinger, without 1021–1025.
whose assistance we would not have been able to carry out our Fournier R. O. and Truesdell A. H. (1973) An empirical Na-K-Ca
research. We also thank D. Elmore and P. Sharma of the Purdue Rare geothermometer for natural waters. Geochim. Cosmochim. Acta 37,
Isotope Measurement Laboratory Lab for the 129I determinations. R. 1255–1275.
Poreda and A. Hunt of the University of Rochester Rare Gas Facility Fuhrmann M., Bajt S., and Schoonen M. A. A. (1998) Sorption of
also provided useful suggestions and helped with sampling. The com- iodine on minerals investigated by X-ray absorption near edge struc-
ments and insights of J. C. Varekamp and an anonymous reviewer were ture (XANES) and 125I tracer sorption experiments. Appl. Geochem.
of great assistance in improving this article in manuscript. Iodine and 13, 127–141.
Bromine determinations were carried out at the ICP-MS Laboratory of Gammons C. H. and Yu Y. (1997) The stability of aqueous silver
the University of Rochester. G.T.S. gratefully acknowledges support bromide and iodide complexes at 25–300°C; Experiments, theory,
from the Geological Society of America and from Sigma Xi. This and geologic applications. Chem. Geol. 137, 155–173.
research was supported by NSF grant EAR-9902919 to U.F. The article Haar L., Gallagher J. S., and Kell G. S. (1984) NBS/NRC Steam Tables.
is a contribution to the special issue honoring H. D. Holland. NSRDC. Hemisphere Publishing.
Langseth M. G. and Silver E. A. (1996) The Nicoya convergent
margin—A region of exceptionally low heat flow. Geophys. Res.
Associate editor: M. A. McKibben Lett. 23, 891– 894.
Luther G. W., Wu J., and Cullen J. B. (1995) Redox chemistry of iodine
REFERENCES in seawater: Frontier molecular orbital theory considerations. In
Aquatic Chemistry: Interfacial and Interspecies Processes (eds. C. P.
Astorga A. (1997) El Puente-istmo de América Central y la evolución Huang, C. R. O’Melia, and J. J. Morgan), pp. 135–155. Advances in
de la placa Caribe (con énfasis en el Mesozoico). Profil. Universität Chemistry Series 244. ACS.
Stuttgart, 12. Ph.D. thesis. Martin J. B., Gieskes J. M., Torres M., and Kastner M. (1993) Bromine
Broecker W. S. and Peng T. H. (1982) Tracers in the Sea. Eldigio and iodine in Peru margin sediments and pore fluids: Implications for
Press. fluids origins. Geochim. Cosmochim. Acta 57, 4377– 4389.
Bureau H., Keppler H., and Métrich N. (2000) Volcanic degassing of Meschede M., Zweigel P., and Kiefer E. (1999) Subsidence and ex-
bromine and iodine: Experimental fluid/melt partitioning data and tension at a convergent plate margin: Evidence for subduction ero-
applications to stratospheric chemistry. Earth Planet. Sci. Lett. 183, sion off Costa Rica. Terra Nova 11, 112–117.
51– 60. Moran J. E., Fehn U., and Hanor J. S. (1995) Determination of source
Burton M. R., Oppenheimer C., Horrocks L. A., and Francis P. W. ages and migration of brines from the U.S. Gulf Coast basin using
129
(2000) Remote sensing of CO2 and H2O emission rates from Masaya I. Geochim. Cosmochim. Acta 24, 5055–5069.
Volcano, Nicaragua. Geology 28, 915–918. Moran J. E., Fehn U., and Teng R. T. D. (1998) Variations in 129I/127I
Carpenter L. J., Sturgen W. T., Penkett S. A., and Liss P. S. (1999) ratios in recent marine sediments: Evidence for a fossil organic
Short-lived alkyl iodides and bromides at Mace Head, Ireland: Links component. Chem. Geol. 152, 193–203.
to biogenic sources and halogen oxide production. J. Geophys. Res. Moran J. E., Oktay S., Santschi P. H., and Schink D. R. (1999)
104, 1679 –1689. Atmospheric dispersal of 129Iodine from nuclear fuel reprocessing
Casadevall T. J., Rose W. J., Fuller W. H., Hunt W. H., Hart M. A., facilities. Environ. Sci. Technol. 33, 2536 –2542.
Moyers J. L., Woods D. C., Chuan R. K., and Friend J. P. (1984) Muramatsu Y. and Wedepohl K. H. (1998) The distribution of iodine
Sulfur dioxide and particles in quiescent volcanic plumes from poas, in the Earth’s crust. Chem. Geol. 147, 201–216.
arenal, and colima volcanoes, Costa Rica and Mexico. J. Geophys. Muramatsu Y., Fehn U., and Yoshida S. (2001) Recycling of iodine in
Res. 89, 9633–9641. fore-arc areas: Evidence from the iodine brines in Chiba, Japan.
Davis S. N., Fabryka-Martin J., Wolfsberg L., Moysey S., Shaver R., Earth Planet. Sci. Lett. 192, 583–593.
Alexander E. C. Jr., and Krothe N. (2000) Chlorine-36 in ground NCRP (National Council on Radiation Protection and Measurements).
water containing low chloride concentrations. Ground Water 38, (1983) Iodine-129: Evaluation of Releases from Nuclear Power
912–921. Generation. Report 75. NCRP.
Déruelle B., Dreibus G., and Jambon A. (1992) Iodine abundances in Oktay S. D., Santschi P. H., Moran J. E., Sharma P. (2000) The
oceanic basalts; implications for Earth dynamics. Earth Planet. Sci. 129-Iodine bomb pulse recorded in Mississippi River Delta sedi-
Lett. 108, 217–227. ments: Results from isotopes of I, Pu, Cs, Pb, and C. Geochim.
Egeberg P. K. and Dickens G. R. (1999) Thermodynamic and pore Cosmochim. Acta 64, 989 –996.
water halogen constraints on gas hydrate distribution at ODP site 997 Patino L. C., Carr M. J., and Feigenson M. D. (2000) Local and
(Blake Ridge). Chem. Geol. 153, 53–79. regional variations in Central American arc lavas controlled by
3838 G. T. Snyder and U. Fehn

variations in subducted sediment input. Contrib. Mineral. Petrol. Sharma P., Bourgeois M., Elmore D., Granger D., Lipschutz M. E., Ma
138, 265–283. X., Miller T., Mueller K., Rickey G., Simms P., and Vogt S. (2000)
Plank T. and Langmuir C. H. (1998) The chemical composition of PRIME Lab AMS performance, upgrades and research applications.
subducting sediment and its consequences for the crust and mantle. Nucl. Inst. Met. Phys. Res. B 172, 112–123.
Chem. Geol. 145, 325–394. Snyder G. T. (2001) Volatiles of the Central American Arc: Source
Protti M., Güendel G., and McNally K. (1995) Correlation between the determinations through iodine-129, noble gas and stable isotope
age of the subducting Cocos Plate and the geometry of the Wadati- analyses. Ph.D. thesis. University of Rochester.
Benioff zone under Nicaragua and Costa Rica. In Geologic and Snyder G. T., Fehn U., and Goff F. (2002) Iodine isotope ratios and
Tectonic Development of the Caribbean Plate Boundary in Southern halide concentrations in fluids of the Satusuma-Iwojima Volcano,
Central America (ed. P. Mann), pp. 309 –326. Special Paper 295. Japan. Earth Planets Space, in press.
Geological Society of America. Symonds R. B., Rose W. I., and Reed M. H. (1988) Contribution of Cl-
Ranero C. R., von Huene R. V., Flueh E., Duarte M., Baca D., and and F-bearing gases to the atmosphere by volcanoes. Nature 334,
McIntosh K. (2000) A cross section of the convergent Pacific margin 415– 418.
of Nicaragua. Tectonics 19, 335–357. Takaku Y., Shimamura T., Masuda K., and Igarashi Y. (1995) Iodine
Rao U. and Fehn U. (1999) Sources and reservoirs of anthropogenic determination in natural and tap water using inductively coupled
iodine-129 in western New York. Geochim. Cosmochim. Acta 63, plasma mass spectrometry. Anal. Sci. 11, 823– 827.
1927–1938. Tedesco D. and Toutain J.-P. (1991) Chemistry and emission rate of
Risacher F. and Fritz B. (2000) Bromine geochemistry of Salarde volatiles from White Island volcano (New Zealand), Geophys. Res.
Uyuni and deeper salt crusts, Central Altiplano, Bolivia. Chem. Geol. Lett. 18, 113–116.
167, 373–392. Wagner M. J. M., Dittrich-Hannen B., Synal H.-A., Suter M., and
Santschi P. H., Schink D. R., Corapcioglu O., Oktay-Marshall S., Schotterer U. (1996) The increase of 129I in the environment. Nucl.
Sharma P., and Fehn U. (1996) Evidence for elevated levels of Inst. Met. Phys. Res. B 113, 490 – 494.
iodine-129 in the deep Western Boundary Current in the Middle Walther C. H. E., Flueh E. R., Ranero C. R., von Huene R., and Strauch
Atlantic Bight. Deep-Sea Res. 43, 259 –265. W. (2000) Crustal structure across the Pacific margin of Nicaragua:
Schink D. R., Santschi P. H., Corapcioglu O., Sharma P., and Fehn U. Evidence for ophiolitic basement and a shallow mantle sliver. Geo-
(1995) 129I in Gulf of Mexico waters. Earth Planet. Sci. Lett. 135, phys. J. Int. 141, 759 –777.
131–138. Weiss J. (1986) Handbook of Ion Chromatography. Dionex.
Schnetger B. and Muramatsu Y. (1996) Determination of the halogens, Wong G. T. F. (1991) The marine geochemistry of iodine. Rev. Aquatic
with special reference to iodine, in geological and biological samples Sci. 4, 45–73.
using pyrohydrolysis for preparation and inductively coupled plasma You C-F. and Gieskes J. M. (2001) Hydrothermal alteration of hemi-
mass spectrometry and ion chromatography for measurement. Ana- pelagic sediments: Experimental evaluation of geochemical pro-
lyst 121, 1627–1631. cesses in shallow subduction zones. Appl. Geochem. 16, 1055–1066.