Gcwhimwu d (b,mrwhimrcu Aclu Vol. 52. pp. 2149-2765 00167037/88js3.

00 + a0
Copyright Q 1988 Pcrpnon Press pit. Printed in U.S.A.

Geothermal solute equilibria. Derivation of Na-K-Mg-Ca geoindicators

WERNER F. GIGGENBACH
Chemistry Division, DSIR, Petone, New Zealand

(Received December 14, 1987; accepted in revisedform September I, 1988)

Abstract-Relative Na, K, Mg, and Ca contents of thermal waters in full equilibrium with a thermodynamically stable
mineral system derived through &chemical ncrystallisation of an average crustal rock are, at a given temperature and
salinity, uniquely fixed. Together with the compositions of waters resulting from &chemical rock dissolution, they
provide valuable references for the assessment of the degree of attainment of fluid-rock equilibrium. In geothermal
systems, the pair K-Na reaches its equilibrium contents as governed by
Lt, = log (c&N,) = I .75 - (I 390/z-)
most slowly (c, in mg/kg, Tin K). The quotient c& adjusts much faster and to low temperatures (<lOOC) according
to
LLrn= log (&C& = 14.0 - (4410/T).
The system K-Ca is sensitive to variations in fa with
lt, = log (c&I?& = log /co, + 3.0.
These subsystems are combined to obtain graphical techniques for the evaluation of deep temperatures and COrpartial
ptessures by use of Na, K, Mg and Ca contents of geothermal water discharges.

INTRODUCTION fluids through intermediate to acid crustal rocks (GIGGEN-

BACH, 1984), four major alteration pmcesses and associated
IN DYNAMICHYDROTHERMAL alteration Systems Overall fluid-
environments were identified:
rock equilibrium is rarely attained and the fluids sampled
are likely to have reached some complex steady-state com- 1. Potassium metasomatism accompanied by silicification
position reflecting the combined effects of initial fluid com- in major upflow zones.
position, the kinetics of primary mineral dissolution and sec- 2. Hydrogen metasomatism caused by the attack of CO,
ondary mineral deposition at changing temperatures and on C-a-aluminium silicates leading to the formation of calcite
pressures, in addition to vapor loss, dilution and mixing with and Al-enriched, argillic to phyllic alteration assemblages
fluids of different origin. In the midst of all these complexities principally over parts of the hydrothermal system subject to
two limiting processes controlling the chemistry of thermal rapid conductive cooling or dilution adjacent to major upflow
fluids may be distinguished: The first of these more or less zones.
hypothetical pmcesses corresponds to &chemical dissolution 3. Sodium, magnesium, calcium metasomatism associated
of the rock material in contact with the rising fluid; the other with descending and heating solutions and, therefore, atliing
to eventual equilibration of the fluid with the thermcdynam- rock over the meteoric recharge zones at the periphery of a
ically stable alteration assemblage resulting from isochemical hydrothermal system.
recrystallisation of the primary rock at a given temperature 4. Isochemical tecrystallisation and hydration over close
and pressure, a process likely to come to completion only in to stagnant parts of a hydrothermal system.
stagnant systems of infinite age (GIGGENBACH, 1984). The
chemical composition of fluids and minerals resulting from In conjunction wlth findings on active volcanoes (GIGCEN-
both these reference pmcesses, however, are accessible to rig- BACH,1975a, 1987). a fifth alteration process, likely to prevail
orous evaluation. over the deep magmatic to hydrothermal transition zones of
By exploiting differences in the thermodynamic properties a hydrothermal system with close volcanic-magmatic asso-
and relative abundances of the four major cations encoun- ciations, may be defined. It consists of the attack of strong
tered in crustal rocks and geothermal waters, Na, K, Mg and acids (HCl, HzS04, HF) formed through absorption of mag-
Ca, an attempt is made to derive geoindicators allowing water- matic gases into deeply circulating groundwater and leads to
rock equilibration conditions and the position of a given water wholesale rock destruction, advanced argillic alteration and
sample with respect to the above two limiting states to be silicification.
evaluated. The alteration patterns thus delineated may be used to
construct a revised and more detailed version of an earlier
schematic diagram (GIGGENBACH,1981) depicting the thee
POTENTIAL PROCESSES AFFECITNC THE CHEMICAL
COMPO!SlTION OF THERMAL WATfSl?!S DURING retical distribution of alteration zones in a hydrothermal sys-
THEIR MOVEMENT THROUGH THE CRUST tern as shown in Fig. 1. Again the top of the degassing magma
is placed at an arbitrary but reasonable depth of 8 km, iso-
In an earlier investigation into mass transfer processes in therms were drawn by assuming uniform and isotropic per-
hydrothermal systems involving the rise of initially magmatic meability conditions and a geothermal gradient of 50/km
2149
2750 W. F. Giggenbach

HENLEY and MCNABB (1978), BRIMHALLand GHIORSO

(1983) and EUGSTER (1985). Fossil manifestations of such
highly acid alteration environments are characterised by si-
licification and advanced argillic alteration (SILLI IOE, 1973)
or even complete cation leaching as at the volcanic hosted
precious metal deposits at Summitville, Colorado (STOFFRE-
GEN, 1985) or Temora, New South Wales (THOMSON et al.,
1986). The physical processes likely to accompany gas release
from a solidifying magma body have been discussed by
CATHLES(1977) and BURNHAM(1979).
With regard to the chemistry of the rising fluids, the major
effects of their passing through the zone of primary neutral-
isation may be assumed to consist of the addition of Na, K,
Mg and Ca in proportions close to those of the original rock
and the conversion of much of the magmatic sulfur, initially
in the form of SO;! and HIS.. to pyrite at deep (GIGGENBACH,
1977; BRIMHALL,1983), or alunite and anhydrite at shallower
levels (GIGGENBACH,1975b; KRUPP and SEWARD, 1987).
The fluids emerging from the zone of primary neutralis-
ation are assumed to resemble closely those produced from
deep wells in active geothermal systems. These close to neutral
NaCl-KCI-brines still contain much of the originally mag-
FIG. 1. Schematic cross-section of a hydrothermal system with matic CO2 and some of the magmatic sulfur in the form of
close volcanic-magmatic associations. H& Their very low magnesium and calcium contents point
to early redeposition, possibly, still within the primary neu-
tralisation zones, in the form of Mg- and Ca-rich alteration
outside the thermal system. There, and within the uptlow products such as amphiboles, biotite, chlorite, anhydrite and
zones, fluid pressures were assumed to correspond to the hy- fluorite.
drostatic pressure of cold water, the temperature/depth dis- The major processes affecting the geothermal fluids proper
tribution is that of CO1-water mixtures in full equilibrium during their further rise to the surface are minor hydrogen
with K-mica, K-fekispar, chalcedony, calcite and the ther- metasomatism, due to secondary neutr~i~tion of COZ, and
rn~yna~~ly stable Ca-~uminium silicate (GIGGENBACH, potassium me~somatism. The very slow drop in temperature
1984). The model of a hydrothermal alteration system thus over the deeper central parts of the major fluid upflow zones
obtained corresponds closely to that derived by LOWELLand is also conducive to isochemical internal re-equilibration of
GUILBERT( 1970) on the basis of the lateral and vertical zo,ning the rock components and therefore to the attainment of full
of alteration in a wide range of prophyry ore deposits. Within fluid-rock equilibrium (GIGGENBACH,1984).
the present context, it is used to delineate the processes likely On approaching the surface or the margins of the major
to affect the chemical com~sition of hydrothe~~ fluids upilow zones, temperature gradients become much steeper
during their rise to the surface. leading to a rapid increase in the reactivity of CO* with respect
The composition of the fluids produced from the magma to hydrogen metasomatism. Separation of an increasingly
body at depth is not accessible to direct measurements. On COZ rich vapor phase and its absorption in shallow ground-
the basis of the compositions of high temperature gases re- water also generates an environment of increased acid alter-
leased from basaltic (GIGGENBACH and LE GUERN, 1976) ation and the production of waters high in bicarbonate; de-
and an&&tic magmas (GI~ENBACH, 1987), they are likely pending on the duration ofwater-rock intera~on, the relative
to consist largely of water vapor containing between 20-40% proportions of cations acquired by the waters in these mar-
b.w. C02, 5- 10%b.w. total sulfur, predominantly in the form ginal zones may again approach those of the unaltered rock.
of SO* at low pressures, but H$ at higher pressures, and 1 Superficial oxidation of H2S, especially in steam heated pools,
to 2% b.w, of HCl (ROSE et al., 1986). These fluids remain provides another source of acidity and an environment for
initially close to equilibrium with the rock. Because of the intense rock leaching (SCHOENet al., 1974). The waters thus
rapid increase in the dissociation constants of H$04 and produced are low in chloride, but high in sulfate (ELLlS and
HCI (RUAYA and SEWARD, 1987) with decreasing tempera- MAHON, 1377).
ture, absorption of these gases into deeply circulating Because of the dynamic nature of most hydrothermal al-
groundwater and cooling leads to the formation of acid, highly teration systems, the compositions of the waters present at
reactive solutions capable of virtually complete cation leach- any stage then are likely to reflect the transition between two
ing of the wall rock and the production of waters with chloride limiting end member processes: from close to isochemical
and sulfate as the major anions. The mineralogical conse- dissolution of crustal rock in highly aggressive, acid fluids,
quences of this primary neutralisation process were first to equilibration with the the~~ynami~ly stable mineral
evaluated by HEMLEYand JONES( 1964), the likely involve- assemblage resulting from isochemical recrystallisation of
ment of magmatic fluids containing free HCl in the transport such rock. In contrast to the intermediate waters, the com-
of ore metals was stressed among others by HOLLAND(I 972), positions of the waters produced by these endmember pro-
 Geothermal solute equilibria 2751

cesses are accessible to rigorous evaluation. By comparing Temperatures for well discharges are those measured at depth, for
the chemical compositions of geothermal discharges with spring discharges actually measured surface temperatures are given.
Of the seven soda waters, only two, Radkersburg (RA) and Te
these theomtical compositions, the relative importance of rock Aroha (TE) are actively depositing calcite at the surface. This obser-
dissolution and equilibration in controlling the chemistry of vation is in agreement with values of +0.8 for the saturation index
thermal waters may be asses&. log (Q/K) based on the quotient (M+aHco;/aH+). The pHs reported
for all samples in Table I arethose measured in the laboratory and,
In earlier investigations into hydrothermal systems the redistri- therefore, may differ considerably from those at depth. The saturation
bution of ions between water and rock was frequently ascribed to an indices calculated by use of these pHs, therefore, are only valid at
ion exchangeprocess.Use of this term, however, should be reserved room temperature for the sample as analysed. &cause of the com-
to the reversible exchange of ions in solution without destruction or paratively shallowtemperatun dependence of the above equilibrium
change of the structure of the substrate. In hydrothermal systems, quotient, the values obtained may still reflect the original state of
this type of ion exchange is restricted to the interaction of waters calcite saturation of the waters. It is for all other scda and chloride
with a limited range of suitable clay minerals. The by far predominant spring samples with +0.4 close to equilibrium with calcite, except
processes facilitating the redistribution of cations among solutions for Lake Nyos (LN) which was found to be well undersaturated with
and minerals are rock dissolution (CHOUand WOLLAST, 1984;HOL- respect to this mineral (GIGGENBACH, 1988).
DREN and SPEYER, 1985) followed by solute transport and spatially In Fig. 2, the water are classified in terms of relative Cl-, SO;-
independent deposition of secondary phases (G~CUEL,1983;SAVAGE and HCO; contents. The set of samples covers the entire spectrum
et al., 1987; SCHRAMKE et al., 1987; PETITetal.. 1987). Also, the of naturally occurring waters from virtually pure chloride waters,
term isochemical is preferred for processes involving the entire rock over mixed chloride-sulfate waters to bicarbonate waters largely pro-
phase, the term congruent is reserved for the dissolution and recrys- duced from the soda springs.
tallisation of individual minerals.
A presentation in terms of the major cations is shown in
Attainment of water-rock equilibrium depends to some
Fig. 3. By comparison with the compositions of waters pro-
degree on the kinetics of individual mineral dissolution and
duced through &chemical dissolution of crustal rocks, such
deposition reactions, but may to a first approximation be
as basalt (BA), granite (GR) and the average crust (AC)
discussed in terms of the overall, steady-state rate of ad-
(TAYLOR, 1964) and the composition of waters expected for
vancement of an alteration front with the amounts of rock
equilibrium with an isochemically recrystallised, thermody-
components released during the dissolution stage being pro-
namically stable average crustal rock (GIGGENBACH,1984),
portional to those present in the original rock. The compo-
only the deep well discharges appear to achieve full water-
sition of full equilibrium waters is likely to vary also with the
rock equilibrium. The acid waters and the soda springs occupy
chemical and mineralogical composition of the original rock
positions suggesting varying degrees of approach to full equi-
(B~RISOV et al.,1985). For the present purpose, the com-
librium at lower temperatures. Most neutral chloride waters
position of waters in full equilibrium with intermediate to
appear to follow a common curved trend suggesting the tran-
acid rocks is assumed to be. determined only by temperature
sition from rock dissolution to rock equilibration to be a
and salinity (ANORSON efal.,1983a; GIGGENBACH,1984).
process more complex than simple end member mixing. A
more detailed evaluation of these rock-dissolution, rock-
CHEMICAL COMPOSITION OF THERMAL WATERS
equilibration processes requires the set of four cations to be
In order to simplify correlation of water compositions and subdivided into several subsets such as Na-K, K-Mg, K-Ca.
likely water-rock interaction environments, three well-char-
acter&d groups of waters were selected: COMPARISON OF ANALYTICAL AND THEORETICAL
ACIWITY QUOTIENTS INVOLVING Na-K-Mg
1. Highly acid chloride-sulfate waters likely to result from
the absorption of magmatic gases in groundwater followed The first ionic solute geothermometer is based on early
by close to &chemical dissolution of rock contacted. They observations of a general decrease in Na/K ratios of thermal
are assumed to reflect conditions within the deep, primary waters with increasing temperature (WHITE, 1957; ELLISand
neutralisation zone. WILSON, 1960; ELLISand MAHON, 1964). Initial calibration
2. Spring waters associated with copious discharges of CO2 of this Na-K geothermometer (WHITE, 1965; ELLISand MA-
(soda springs) and likely to be representative of the waters HON, 1967) led to relationships with quite shallow temper-
with high CO2 reactivities at the periphery of hydrothermal ature dependences (Fig. 4) largely due to the inclusion of
systems. incompletely equilibrated spring waters. By taking into con-
3. Discharges from deep geothermal wells and from as- sideration experimental data on cation exchange between
sociated neutral chloride springs. These waters are likely to coexisting alkali feldspars (ORVILLE, 1963; HEMLEY, 1967),
represent well-equilibrated fluids from the major upflow FOURNIERand TRUESDELL(1973) arrived at a revised cali-
zones. bration curve with a considerably steeper temperature de-
pendence. Two further curves proposed later on the basis of
The first three samples of the first group of Table I are obviously findings by White and Ellis (TRUESDELL,1975) and on geo-
volcanic waters, the White Island sample (WI) representing an emerald
green water discharged from springs on the crater floor (GIG(;ENBACH thermal well discharges in Iceland (ARNORSSONet al., 1983b)
and GLASBY,1977).that from Ruapehu Crater Lake (RU), a water led to only minor modifications (1983A, Fig. 4).
produced during an eruptive period (GIGGENBACH, 1975b).The next In 1979, Foumier provided a revised equation, now based
three samples are likely to be volcanic waters genetated at depth but on Na and K contents of waters with well established thermal
considerably diluted during their rise to the surface.
histories, such as deep geothermal well discharges and oil
The second group comprises comparatively low tempetature springs
discharging copious amounts of free CO2 (Soda Springs). The high field brines. The resulting correlation (1979) approaches very
chloride waters of the third group represent a wide range of deep closely that calculated by use of recent thermodynamic data
geothermal well discharges and associated neutral chloride springs. on the stability of Na- and K-feldspan (BOWERSet al.. 1984).
 Acid Springs

White Island. N7. WI 98O 0.6 5910 635 3800 3150 3870(! 4870 d
Kawa Idjen, Indon Ki 60" 0.6 1030 1020 680 170 21800 62400 d
Ml Ruapehu, NZ R" 38" 1 2 1120 170 1750 1380 13240 14'700 a
Tamagawa, Japan TA 98" 1 3 38 30 35 95 2970 2300 a
Guayabal, C. Rica Mv 64" 1.9 56 2 48 99 684 2830 a
B Georginas, G~tatern. ZU 74" 2 0 134 32 26 72 16 152" a

Soda Springs

Wnitangi, NZ WA 490 7.3 2R5 24 9 17 364 -I

9 202 a
0mapere. NZ NC 30" 6.9 62 9 10 35 25 Ii 29S il
Golden Spr NZ GS 43 7.0 224 20 7 11 51 R 670 a
Lake Nyos. Cameroon LN 23 5.7 19 7 58 45 1 <I 724 it
Acque Albulr, Italy AA 22 6 1 138 22 238 1042 163 1470 1403 d
Radkersburg, Austria RA 72 8.9 2115 182 47 3 264 398 4130 a
7~ Aroba, NZ TE 79 8.1 292 67 4 7 540 1260 6830 R

Geothermal Wells, Springs

Ngawha M-9, N7 NG 230 7.7 893 79 0.110 1260 18 185 a

Jubilee Pool. NZ NG 56 7.2 842 72 1.600 9 1180 174 393 a
Wairakei, WK-66. NZ Wx 240 8.5 995 142 0.040 17 1675 30 t5 a
Champagne Pool, NZ WK 99 8.0 1070 102 0.400 26 1770 26 76 b
Broadlands 88-11. NZ BR 260 7.4 875 130 0.011 1 964 41 376 a
Ohaki Pool. NZ BR 98 7.1 860 82 0.100 3 1060 100 680 b
Hiravalles 10.C. Rica Mv 250 7.8 ,750 216 0.110 59 2910 40 27 a
Sal Bagaces. C Rica M" 74" 8.6 2063 85 6.500 33 2700 102 739 a
Zunil ZQ-3, Guatem. 7, 295 8.1 933 231 0.012 1.5 1810 31 51 a
Zunil Spring 7.u 95 7.0 384 32 39 17 172 234 635 a
Cerro Prieto. Mexico CP 280" 7.3 5600 1260 0.270 333 10500 14 40 b
Spring NZ9. Mexico CP 80" 7.6 5120 664 4.600 357 8790 31 65 b
Tongonan. Philipp. TO 330" 7.0 3580 1090 0.200 128 6780 16 12 b
Banati Spring, Phil, TO 98" 8.3 1990 210 0.400 86 3400 74 lb
Morere Spring, NZ MO 47' 7.0 6690 84 79 2750 15670 <3 28 a
Salton Sea. Well, USA ss 330" 5.2 38400 L3400 10 22010 118400 4 140 b
Seawater SW 4" 7.8 10760 390 1290 410 19340 2710 140

800 4500 6700 28650 c

3300 200 1900 11240 c
2100 2300 4200 19740 c

refs a this work. b - Henley et al (1984). c Taylor (1964)

Another correlation (1983G), showing an even steeper tem- (1979) and FOURNIERand POTTER(1979). Assuming equi-
perature dependence was proposed by GIGGENBACH et al.. librium Na/K-values to be approached from below (Fig.
in 1983. It also is based on analytical data for a wide range 4), the latter line was obtained by connecting data points
of deep well dixhatges including those reported by FOURNIER representing maximum Na/K-ratios at a given temperature.
In addition to these empirical lines and the theoretical al-
bite-K-feldspar boundary, three horizontal lines correspond-
ing to the &chemical dissolution of an average basalt, an
average granite and an average crustal rock of intermediate
to acid composition (TAYLOR, 1964) are shown in Fig. 4.
The triangular space thus delineated is assumed to cover the
compositional range of waters affected to varying degrees by
isochemical rock dissolution and partial or complete equil-
ibration with a thermodynamically stable mineral assemblage
including albite and K-feldspar. Data points were plotted by
converting analytical concentrations q in mg/kg to activities
a, according to

(I, = ~,+,/I000 M, (1)

where +, is the total or stoichiometric ion activity coefficient

of the solute species i and M, its atomic weight. The e&lib
rium constant governing the reaction
FIG. 2. Relative Cl, SO, and HC03 contents (mg/kg) of waters
listed in Table I. albite + K = K-feldspar + Na (2)
 Geothermal solute equilibria 2153

Waters from soda springs occupy positions close to those

of the acid waters. Their generally much lower acidity should
ahow ~uminium silicates to form, their increased Na/K-ratios
then may be ascribed to the removal of K in the form of K-
rich clays. Of the CO2 waters, that of Lake Nyos, Cameroon
(LN), has a composition compatible with dissolution of local
basalt (GIGGENBACH,1988). Only the Te Aroha (TE) water,
New Zealand, shows close approach to full equilibrium with
the rock.
Of the waters considered here, the deep well discharges
have Na and K contents compatible with attainment of water/
feldspar equilibrium. The acid and COrrich waters occupy
positions intermediate between rock dissolution and equili-
bration. The apparent partial re-equilibration thus indicated
may be due to preferential leaching of sodium or the removal
of K-rich alteration products such as alunite at high, clays at
lower acidities, but is unlikely to reflect attainment of equi-
librium with feldspar at temperatures much higher than dis-
charge temperatures. The acidity and high sulfate contents
of the first group of waters provide early warning signs that
Na/K-temperatures are likely to be unrealistic; in the case of
Fro. 3. Plot of lO~$(lO~ + cc.) versus lOc~(lOcx + c& of
the close to neutral, low sulfate, high CO2 waters no such
waters listed in Table 1 (cr in mg/kg). The commotions of waters
expected for attainment of full equilibrium with an average crustal ready warning signals are apparent.
rock as a function of temperature and of solutions of basalt (BA), The strong dependence of Mg contents on the temperature
granite (GR) and an average crustal rock (AC) are shown. For symbols of thermal discharges has long been recognised and has gen-
see Table I and Fig. 2. erally been ascribed to equilibration with chlorite (ELLIS,
1971) or Mg containing clays for which chlorite may serve
then becomes, with &+ = -&+, as a the~~ynamic proxy. By use of theoretical consider-
ations, two potential mineral systems likely to control mag-
nesium contents of geothermal discharges were identified,
5 1.7 C~++/CK+. (3) chlorite over parts of the hydrothermal system still affected
The activity ratio ax_r~r&pr/&rbirefor the secondary feldspars by hydrogen metasomatism, biotite as a component of a fully
may safely be assumed to be close to unity. equilibrated, isochemically recrystallised alteration assem-
Data points for deep gathers well discharges fall close blage (GI~ENBACH, 1984). Uncertainties in the activities of
to the theoretical albite/K-feldspar coexistence line suggesting the clinochlore and phlogopite endmembers, arising from
close attainment of fluid-rock equilibrium. Na/K ratios of the large compositional variations of naturally occurring
associated spring discharges are generally shifted to higher chlorites and biotites, the close similarity in the thermody-
values, pointing to some re-equilibration of the spring waters namic behaviour of the two mineral systems, and the ubiq-
during their rise to the surface. Corresponding Na/K equil- uitous presence of chlorite in hydrothermal alteration assem-
ibration temperatures, however, are only some 50 below blages suggest that the behaviour of magnesium in waters
deep well temperatures suggesting Na/K ratios to be quenched
at considerable depth. The sample MO represents a vigorous
discharge of a 47 high chloride water, its inferred equilibra-
tion temperature is with 90C also very low. Seawater (SW)
plots well within the stability field of K-feldspar and is ob-
viously far from ~u~ib~urn with albite.
Of the highly acid waters, two show NafK-ratios corre-
sponding to isochemical leaching of crustal rocks (KI, TA),
the remainder have increased Na/K-ratios suggesting pref-
erential leaching of sodium or removal of some of the po-
tassium in the form of secondary alteration products. Under
the acid conditions, K-containing aluminium silicates are
unstable, the most likely K-mineral to form is alunite (HEM-
LEY and JONES, 1964). Actual deposition of this mineral has
6 fiiil A.yERew CRUST
been reported for both the White Island (WI) and Ruapehu GRANITE
-0

(RU) waters (GIGGENBACH, 1975b). No surface deposition 0.

of alunite has been reported for the two remaining acid waters t(c) 100 XXI m
as discharged at Guayabal, Costa Rica (MV) and Bafios FIG. 4. Empirical, theoretical and analytical NafK-activity ratios
Georginas, Guatemala (ZU); their high Na/K-ratios may be of thermal waters as a function of temperature. For symboh see Table
ascribed to the deposition of K-Al-sulfates at depth. 1 and Fig. 2.
2754 W. F. Giggenbach

approaching attainment of full equilibrium may adequately The effect of the incorporation of muscovite and clinochlore into
be described, in terms of endmember components, by the naturally occurring illites and chlorites may be assessed by use of
their activities reported by CAPUANO and COLE(1982) for drill cut-
reactions
tings from the Roosevelt Hot Springs and by WALSHE(1986) for a
0.8 muscovite + 0.2 clinochlore + 5.4 silica + 2 Nat P wide range of hydrothermal layer silicates. For these samples, values
of LogGnu%foite vary from -0.73 to -0.15 with a mean of -0.33, log
2 albite + 0.8 K-feldspar + 1.6 water + Mg*+ (4) acrinachlore
from -4.7 to - 1.3 with a mean of -2.2. The activities of
the feldspar endmembers may safely be assumed to be close to unity,
and the logarithm of the activity product AP = log (~0,8,,,,,,a~i:.,~,,,)
then may vary from - 1.5to -0.4 with a mean of -0.7. The effect
0.8 muscovite + 0.2 clinochlore + 5.4 silica + 2K+ P on log (nMs2+/&+)corresponds to

2.8 K-feldspar + 1.6 water + Mg2+. (5)

The generally observed supersaturation of geothermal solu- as shown in Fig. 5 for the above limits. The rock dissolution process
tions with respect to silica is taken into account by assuming is indicated by two lines representing the dissolution of IO g and 100
its activities to be controlled by equilibrium with its micro- g of an averagecrustal rock in 1 kg of water.Comparison of analytical
crystalline form, chalcedony; water fugacities are controlled and theoretical information again is based on Eqn (1). For the iso-
coulombic reactions considered here, the activity coefficients for
by the presence of liquid water. Corresponding equilibrium mono- and di-valent ions can be assumed to be closely related through
constants then become +: = +2 (LINDSAY,1980).
0.2
K4 = (aMg2+/c&+)(u2
albtteaof
K feldsp./ amusc.~clinochlore
. ) (6)
In the case of uMgl+/&+ (Fig. 5a), deep well discharges
K5 = (uMs2+/u~+)(uh.8f,,dJp.lu~8Usc.a~i~och,ore). plot close to the theoretical lines albeit with a tendency to
(7)
deviate to higher relative Mg contents. Associated spring dis-
Values for K4 and K5 were obtained by use of the hydrolysis charges are shifted parallel to the theoretical trend apparently
constants provided by BOWERS et al. (1984), their resulting suggesting maintenance of water-rock equilibrium during
temperature dependences are shown in Figs. 5a and 5b for their rise to the surface. An exemption are the samples from
activities of unity of the endmember phases involved. the Zunil area in Guatemala (ZU) where the neutral spring
discharge appears to be strongly affected by admixture of the
nearby high sulfate water (B. Georginas).
All acid waters occupy positions far from the equilibrium
2, -nm

7 lines and over a range expected for rock dissolution. Of the

,z 3
91
soda springs, Te Aroha (TE) and Radkersburg (RA) appear
+
to approach equilibrium most closely. The compositions of
2 4
SO
Lake Nyos (LN), Acque Albule (AA) and Omapere Soda
B Spring (NG) are likely to be controlled by rock dissolution
5
-1 rather than equilibration. At a temperature of 4C, the com-
position of seawater would be compatible with full equilib-
6
-2 rium with respect to reaction (4).
The distribution of data points in Fig. 5a clearly shows
7
-3 that relative Na/Mg-contents of acid waters and of most of
the soda springs are close to those expected for dissolution
8
of crustal rock suggesting that preferential leaching of Na, to
-km give rise to the high Na/K-values of Fig. 4, is not important.
4 The increased Na/K-ratios, therefore, are more likely to be
1
due to the deposition of K-rich alteration phases.
;3 2
The K, Mg-contents of deep well discharges again point
NX
, to close attainment of equilibrium with respect to Eqn. (5)
cs
m.2 (Fig. 5b). Associated springs are shifted to higher u&u&al-
3
5 ues, again quench temperatures some 50 below their deep
Sl temperatures are indicated. Instead of being significantly re-
4
moved from the rock equilibration lines as in Fig. 4, the
C compositional ranges of dissolved rock now coincide with
5
equilibrium compositions over low to intermediate temper-
-1 atures. The position of data points for the acid waters close
6
to the rock equilibrium line, therefore, is not too surprising;
-2 the position of some well above the rock dissolution range
7
(MV, TA) again points to the deposition of K-rich minerals.
In contrast to Fig. 5a, the compositions of soda sprbs
FIG. 5. Empirical, theoretical and analytical activity quotients in- agree with attainment of equilibrium with respect to reaction
volving Na (A), K (B) and Mg. Chlorite-illite-feldspar coexistence
(5), suggesting that the system K-Mg approaches mineral-
lines am shown as a function of variations in mineral activities ac-
cording to Eqn. (8). For symbols see Table 1 and Fig. 2, AP = log fluid equilibrium at temperatures down to 25 even in a
(U%0,,PXc&Xe). comparatively aggressive environment. The thermodynamic
 Geothermal solute equilibria 2755

behaviour of the K-Mg-clays, likely to form at these low tem- until kaolinite precipitates. From then onwards the system moves
peratures, appears to be still described adequately by their along the kaolinite boundary until the solution becomes saturated
with respect to solid silica (chalcedony, quartz). At fixed silica activ-
thermodynamic proxies muscovite and &m&lore. ities, ongoing rock dissolution will shift Aln-values to lower values
The obviously fast response to variations in temperature (SAVAGE, 1986) until a state corresponding to equilibrium with a
renders K, Mg contents highly suitable as the basis for a geo- thermodynamically stable assemblage formed from the original rock
thermometer. The temperatures obtained, however, will al- is reached. For an average crustal rock the minemlogicaJ composition
of this assemblage contains Na in the form of albite,Mg either as
ways be close to those measured anyway. On the other hand,
chlorite or biotite, Ca as one of the aluminium silicates, represented
the fast response of this system may provide the means to by the their thermodynamic proxies laumontite at low temperatures
differentiate among the various processes interfering with the and wairakite at high temperatures; K-feldspar and mica provide the
evaluation of deep equilibration temperatures by use of the buffer to accommodate any aluminium not incorporated into the
system Na-K. The findings provided by Figs. 4 and 5 also above mineral phases.The final stage of the alteration process, there-
fore, is delineated by the intersection ofthe quartz solubility line with
show that both rock dissolution and equilibration play an the K-mica-feldspar coexistence boundary (full equilibrium).
important part in determining the chemical compositions of According to this reaction sequence, secondary albite may be ex-
thermal waters. In order to assess the relative importance of pected to form well Wore K-feldspar. A look at Fm. 4 and 5, however,
these processes on the individual constituents Na, K, Mg, indicates that equilibrium with K-feldspar is obviously much more
readily attained than with albite, most of the lower temperature waters
Ca, a more detailed discussion of the evolution of thermal
olottinn well outside the stabilitv fields of Na-feldscar. The major
waters during the transition from rock dissolution to equil- reason for this discrepancy is the f&t that twodimensi&al project&s
ibration is required. of multicomponent systems such as Fig 6 provide only an indication
of the potential stability not necessarily the actual formation of a
VARIATIONS IN SOLUTE CW4TEN-I-S AS A given mineral. For a mineral to form, the solution has to be saturated
FUNCTION OF ROCK DlSSOLUlION with respect to all the components involved in its deposition. The
AND EQUILIBRATION Aln-silica-diagram of Fig. 6 takes into account only Al-, H and
SiQ-activities; albite, however, is only able to form if the activities
In an earlier investigation into mass transfer pmcesses in hydro- of the additional component Na+ are also high enough.
thermal alteration systems, the reaction path of solutions resulting During the initial interaction of an acid water with rock the amount
from the dissolution of and equilibration with an average crustal of sodium added to the solution is assumed to be proportional to the
rock was described in terms of AI = log (a&&) versus silica amount of rock dissolved. The amount of sodium required for the
diagrams (GIGGENBACH, 1984). A simplified version for 200C is stable formation of albite within the solute system Na-K-MgCa-Cl
given in Fig. 6. It was constructed by assuming dioctahedral alkali may be calculated by use of the simple charge balance
clay minerals to belong to a continuous solid solution series involving
the endmembers muscovite and pyrophyllite as described earlier mN,++ rnx. + 2mM$- + 2mc.l+ = m- (9)
(GIGGENBACH, 1985). In case recent suggestions are valid that illite
and a procedure based on thermodynamic equilibrium constants for
and smectite represent distinct phases (SASSer al., l987), the smooth
ion exchange reactions among the four cations in a fully equilibrated
feldspar-clay boundary curves would be replaced by a series of chords
system as described earlier (GEGENBACH, 1984). The distribution
marking feldspar-illite and fekkspar-smectite coexistence, without
of solutes as shown in Fig. 7 was obtained accordingly by assuming
changing the overall correlations in any significant way.
attainment of full equilibrium between the solution and an assemblage
During the initial dissolution stages, both Al and silica in solution
containingalbite, K-feldspar, muscovite, clinochlore and Iaumontite
are assumed to increase in proportion to the amount of rock dissolved
below, wairakite above 260C. Relative concentrations of monovalent
and divalent cations of Fig 7 differ considerably from those shown
earlier (GIGGENBACH, 1984) due to an arithmetical error in the cal-
culation of the latter. Figure 7 here., therefore, replaces that of the
earlier study. As already stated much earlier (HOLUND, 1972; SHI-
KAZONO,1978). relative contents of divalent ions (Iv@, Ca+) in-
crease at low salinities with the square of the chloride content. At all
salinities, however, full equilibrium Mg and Ca contents decrease
with temperature, potassium contents increase, with sodium at low
salinities remaining independent of temperature but increasing with
temperature considerably at high salinities. Cooling of fully equili-
brated, highly saline solutions, therefore, may lead to both K- and
Na-metasomatism; of dilute solutions (c 1000 mg/kg Cl) only to K-
metasomatism.

The full equilibrium cation contents thus derived may be

compared to those released from the rock as shown in Fig
8 for the reaction of a strong acid (HCl) with an average
crustal rock. Initial rock dissolution is evaluated in terms of
a titration according to
(,V&O) + 2H - 2/zM+ + Hz0
where IV+ is one of the metallic solution species Na+, K+.
M$+, Ca*+ present in the rock as the oxide component (M2,
,O). Release of silica does not consume protons, the rock
component Al203 is also assumed to be quantitatively con-
FIG. 6. The stat&y of Na- and K-ahrminium silicates as a function
of the activity of dissolved silica and Aln = log (u~~+/u$) at 200C. verted to hydroxy-minerals (AI(OH)r, AlOOH) and un-
Arrows depict likely path of solutions during the transition From rock charged aluminium silicates (kaolinite, pyrophyhite) without
dissolution to rockcquilibration. consuming protons. Initial rates of rock dissolution am likely
 FIG. 7, Distribution ofsolute species (in mg/kg) in full equilibrium hydrotbcrmal solutions as a function of temperature
and chloride contents of 10, 1000 and 100,000 mg/kg (0.00028,0.0283,2.83 m).

to be fast, sicwing down considerably as the equivatence

7 point is reached. The drastic drop in proton activities on
reaching this point will also lead to the immediate deposition
of charged aluminium silicate such as smectites and chlo-
rites. For the systems shown in Fig. 8, this point is reached
after the dissolution of 5.1 g of an average crust& rock in 1
kg of an i&i&fly 0.028 m solution of HCI (1000 m&kg of
chloride).
The amounts of I@ and Ca released by then are well above
those required for full equilibrium with chlorite and either
laumontite (100) or wairakite (3009, respectively. In ac-
cordance with Fig. 6, the minerals forming in this not-yet
full equiiibrium system, however, are likety to be various
layer silicates typical of argilfic to phylIic alteration assem-
blages. At both 1NY and 300, the amounts of sodium re-
leased at this stage are still well below those required to form
a full equilibrium assemblage involving albite. Because af the
essentially complete consumption of free protons further
release of Na to the solution relies now on protons generated
by the conversion of acid clays to mare *neutraE atumin-
ium silicates according lo

z(H-aluminium silicate) -t M --c

(N-aluminium silicate) + zH+. (I 1)
The ~~rn~~~ely low steady state proton ;tctivities main-
tained by this process can be expected to lead to very slow
rates of rock dissolution and, therefore, to very slow rates of
equilibration with respect to the system involving sodium.
The amount of average crustal rock required to produce a
1000 mgfkg chloride solution in equilibrium with an assem-
blage ~on~n~n~~bite is f 5.4gat IWC and 14.5 gat 3OOC
(Fig* 7). three times that consumed in the fast primary neu-
tralisation step. On attainment of full fluid-rock equilibrium,
further rock alteration consists of very slow isochemical re-
amount d rock dissolved (g/kg I 10 XXI crystallisation (metamorphism) of the remaining rock
(Fig. 8).
RG. 8. Amounts of Na, K, Big, Ca and SO, r&eased during the
diss&ution of increasing amounts of an average crustal mck in a During the secondary neutralisation process as symbol&d by
0.0283 m (IO@0 m&kg CQ so2ution of HCI at 100Dand 300C. On reaction (1 I], the rebxive concentrations of the three cations K,
reaching saturation with respect to duminium silicates, amounts of Me. and Cal+. can be expected to be initially controlled by the
Na, K, Mg and Ca in solution dewease due to incorporation into formation ofacid clays in a thermodynamic environment still dic-
respective alteration products. tated by the composition of the fluid phase (fluid dominated) but
 Geothermal solute equilibria 2757

rapidly approaching the rock dominated environment of the full too low. The system K-Mg appears to adjust most rapidly
equilibrium assemblage (Fig. 8). In spite of non-attainment of full and to attain equilibrium values even for COrrich waters.
equilibrium, the solution still being undcnaturated with respect to
The overlap of rock dissolution values with equilibrium values
albite. the chemistry of K and Mg may quite early on be discwed
in terms of paAal equilibration within the system K-mica/K-feldspar again introduces uncertainties. Each of the above three sub-
and chlorite. The chemistry of Ca is complicated by the likely early systems then has its peculiar advantages and disadvantages.
formation of the non-aluminium silicate calcite in CO1 containing By developing a technique allowing the simultaneous eval-
systems, it, therefore, will be discussed separately. The presence of uation of the effects of rock dissolution and equilibration on
sulfate was also not considered in the construction of Fig. 8. It is
likely to reduce both K and Ca contents through formation ofalunite
the combined system Na-K-M& it may be possible to elim-
and anhydrite and, therefore, would affect the arrangement of alter- inate some of the drawbacks associated with separate eval-
ation zones as given in Fig. 8 for simple chloride waters.. uation.
At temperatures above 200, well discharges were found
to be close to equilibrium with respect to both Na- and K- SIMULTANEOUS EVALUATION OF WATER/ROCK
feldspar (Fig. 4). The preservation of these high temperature EQUILIBRATION CONDITIONS FOR THE
SYSTEM Na-K-Mg
Na/K-ratios in associated surface spring discharges provides
the basis for the most commonly used geothermometer. The Initial evaluation of correlations among three independent
coincidence of high temperature equilibrium Na/K-values variables is most conveniently carried out by use of tri-linear
with those for obviously unequilibrated acid waters, however, or triangular diagrams. Their construction requires the con-
introduces large uncertainties in its application. In order to version of absolute to relative values through application of
be able to judge the reliability of Na/K-temperatures an in- a linear scaling or normalising factor. Because of the non-
dicator is required allowing a distinction to be made among linear thermodynamic correlations among Na, K and Mg,
waters likely to have equilibrated with the rock or not. comparison of analytical and theoretical information by use
The distribution of data points in Fig. 5a provides such a of standard triangular plots would only be valid for given
clear distinction between acid waters resulting from rock dis- values of mNn++ mK+ + 2mMp. In order to render the plot
solution and waters approaching full rock equilibrium. The applicable at all salinities, relative Na, K, Mgdata have to
underlying large difference in the theoretical position of data be plotted in terms of either mk.+, m&, muZ+ or rnNa+,
points for rock dissolution and equilibration suggests that the mK+,G, the latter conducive to a less spread-out pre-
system Na-Mg may form the basis for a parameter allowing sentation. In gaining the ability to represent data for solutions
equilibrated and une-quilibrated waters to be separated. Use of any salinity in one diagram, one of the most valuable fea-
of this system as a geothermometer, by itself, is obviously tures of triangular plots, the presentation of mixing relation-
impeded by the likely lack of equilibrium with respect to ships by straight lines, had to be sacrificed. For processes
albite at low temperatures. consisting essentially of variations in relative Mg-contents,
A solute system involving alteration phases forming early mixing lines still remain straight (Fig. 9).
on in the rock equilibration process is that linking potassium Adoption of a triangular plot involving the variables
and magnesium as shown in Fig. 5b. High temperature data mN&, mK* and 6 dlOWS SOkiOn COmpSitiOnS IZpre-
points again fall close to the theoretical line representing co- senting attainment of full water-rock equilibrium for all sa-
existence of layer silicates with feldspar (Eqn. (7)) at end- linities to be presented in terms of a single curve. In order to
member activities close to unity. Data points for lower tem- facilitate practical application, the concentration scale c, in
perature waters occupy positions corresponding to the likely mg/kg (Eqn. (I)) is used, in order to accommodate most geo-
attainment of equilibrium with naturally occurring K-Mg- thermal waters, sodium contents are divided by 1000, po-
layer silicates. The major distinction in the behaviour of the tassium by 100 and the square root is taken of magnesium
system K-Mg is the apparent attainment of solution-mineral values as measured. Because of their potential use as geo-
equilibrium at very low temperatures and even for CO&h thermometers, the activity quotients used in Figs. 4 and 5
waters as represented by the soda springs. The coincidence are converted to concentration quotients with a positive tem-
of the composition of some of the highly acid waters with perature dependence. The temperature dependence of
that expected for equilibrium, however, is likely to be for- c~+/+~+ adopted here is that proposed earlier (GIGGENBACH
tuitous and to result from the overlap of the compositional ef al.. 1983) and shown in Fig. 4 (19836) according to
range-s for rock dissolution with those for rock equilibration
= I .75 - (1390/T)
I+. = log (CK*/CN~+) (12)
at around 100C and introduces considerable uncertainties
in the application of the quotient Mg/K2 as a geothermom- where T is the absolute temperature in K. It is assumed to
eter. describe attainment of full water/feldspar equilibrium.
Summarising the three sub-systems considered above, the For the quotient &+Jc~~I+ a temperature dependence par-
one involving Na-K is likely to respond most slowly and, allel but 0.3 log-units above the theoretical line for the co-
therefore, to preserve deep equilibration Na/K-ratios. The existence of K-feldspar with clinochlore and muscovite at
production of similar ratios in waters obviously not in equi- endmember activities of unity is adopted (Fig. 5b). It is that,
librium with the rock generates large uncertainties in its ap slightly rounded off, reported earlier (GIGGENBACH et al.,
plication as a geothermometer. Use of the system Na-Mg as 1983)
a geothermometer is limited by its shallow temperature de-
Lk,,, = log (c&/c,.,r+) = 14.0 - (4410/T). (13)
pendence and its excessive sensitivity to the addition of non-
equilibrated waters. Addition ofeven minor amounts ofsuch Use of this temperature correlation is justified by the position
Waters 1eadSto apparent temperatures of equilibration by far of high temperature data points in Fig. 5b and the assumption
 W. F. Ciiggenbach

Na/ 1000

K/100 dKJ
FIG. 9. Relative Na, K, Mg contents of thermal waters. Isotherms calculated by use of Eqns. (I 2) and (I 3) are given
together with the full equilibrium curve. For symbols see Table 1 and Fig. 2.

that measured temperatures of the lower temperature waters The position of data points for two soda spring discharges
are likely to be somewhat below actual water-rock equilibra- (TE, RA) points to attainment of partial equilibrium, the
tion temperatures. The deviation with regard to theoretical remaining soda spring waters and all the acid waters plot
behaviour may also be ascribed to uncertainties in the ther- close to the Mg-comer. Data points in this area may then be
modynamic data used for K-feldspar, muscovite and cli- taken to correspond to those of immature waters generally,
nochlore. At present there appears to be no reason to favor to waters unsuitable for the evaluation of K/Na-equihbration
a theoretical over the empirical trend. temperatures. For not too acid waters, K/Mg-temperatures
The internally consistent temperature dependence of the may still be valid. Of these acid waters, the composition of
quotient &,+/cM~+ is simply obtained by use of Kawa ldjen (KI) corresponds to the dissolution of 90 g of an
average crustal rock, the others show reduced alkali contents
L Ill= Lkm - i?Lk, = log (Cf,;,/&@+)
indicating removal of some of the Na and K, possibly in the
= 10.5 - (1630/T). (14) form of alunite, with Mg remaining unaffected by the de-
position of secondary minerals. These conclusions are sup
In a fully equilibrated system each of the above three rela- ported by observations on waters from Ruapehu Crater Lake
tionships is equally valid. As pointed out earlier, only the during the 197 1 active period (GICGENBACH,1975b).
two sub-systems K-Na and K-Mg are likely to provide the Evaluation of analytical Na, K and Mg contents by use of
basis for suitable geothermometers. They are presented in Fig. 9 allows a clear distinction to be made between waters
Fig. 9 by two sets of isotherms, their intersections correspond suitable or unsuitable for the application of ionic solute geo-
to the composition of waters in equilibrium with both mineral thermometers. At the same time it allows deeper equilibration
systems, the resulting curve is marked full equilibrium. temperatures and the effects of a variety of processes such as
The compositions of waters formed through &chemical dis- re-equilibration and mixing of waters of different origins to
solution of an average crustal rock are shown in Fig. 9 for be assessed for a large number of samples. The simultaneous
the dissolution of up to 1000 g in I kg of water. The rock evaluation facilitates also the delineation of trends and
dissolution line thus defined is well separated from the full groupings among water discharges and, from these, of vari-
equilibrium curve. ations in the nature and intensity of processes affecting the
As expected, high temperature well discharges plot some- rising waters.
what above actually measured deep temperatures on the full
equilibrium line. Associated spring water compositions are
REACTIONS INVOLVING CALCIUM
shifted to lower temperatures and off the full equilibrium line
indicating acquisition of Mg by the waters in response to The most popular ionic solute geothermometcr presently
decreasing temperatures to be faster than that of Na. The in use is the Na-K-Ca-geothermometer as devised by FOUR-
low temperature formation water discharge from Morere NIER and TRUESDELL(1973). By assuming the three cations
(MO) also plots on the full equilibrium line suggesting at-
to be. linked through the general reaction
tainment of water-rock equilibrium, given enough time, even
at quite low temperatures. K + solid e b/2 Ca2 + (I - j3)Na + solid (15)
 Geothermal solute equilibria 2759

they arrived at a value of /3 = % for high temperature solutions The most important reaction leading to the formation of
corresponding to the reaction calcite in geothermal systems is the conversion of Ca-aht-
minium silicates to calcite by CO,of deep, probably magmatic
K+ + solid Ft /6Ca2+ + 2/3Na+ + solid (16) origin, according to
and of /3 = 4/sfor low temperature systems Ca-Al-silicate + CO2 + HrO P
K+ + I/, Na + solid P /, Ca2+ + solid. (17) CaCO, + 2(H-Al-silicate) ( 19)

FOURNIER and TRUESDELL(1973) do not report any tem- the acid clays thus formed promote further rock dissolution
perature dependences for these reactions, from their discus- according to Eqn. (11). The minimum COr-fugacity required
sions it has to be concluded that they are taken to be inde- to allow reaction (19) to proceed is that corresponding to
pendent of fl and, therefore, are given by only one equation equilibrium of calcite with the full equilibrium mineral as-
semblage involving K-feldspar/K-mica according to

Ca-A12-silicate + K-feldspar + CO2 #

= (1647/T) - 2.24, (18)
CaCO3 + K-mica. (20)
a condition only valid if /I is seen as a statistical weighing
factor and not a stoichiometric coefficient. Its temperature dependence is adequately represented by log
The apparent success of the technique gave rise to a flurry ho2 = 0.01681- 3.78 where t is the temperature in C (GIG-
of studies trying to provide its theoretical or experimental GENBACH,1984).
justification (SHIKAZONO, 1976; MICHARD and FOUILLAC, Initial C02contents of deep geothermal fluids are likely
1976; JANECKIet al., 1986). Quite early on, however, PA&S to be externally controlled and to reflect largely variations in
( 1975) observed a systematic deviation for the Na-K-Ca gee- the relative proportions of CO&h magmatic fluids and C%
thermometer below 75C and proposed a correction term poor meteoric water making up the rising fluid (Fig. 1). For
involving Pa. In 1977, FOUILLAC and MICHARD also any such rising fluid, there exists, a range ofconditions where
pointed out problems arising in the application of this geo- &.-values are likely to be below those of reaction (20) and
thermometer to high CO2 waters. In a valiant attempt to therefore too low to induce calcite formation. Full equilibrium
extend its validity range to lower temperatures, FOURNIER fcol-values, however, drop rapidly with decreasing tempem-
and Porrz~ (1979) proposed a Mg-correction. The general ture (Eqn. 20) and any rising geothermal fluid with a given
incompatibility of a common temperature dependence for CO2 content can be expected to become reactive (GIGGEN-
two quite different reactions, one essentially based on the BACH, 1984) with respect to COrinduced hydrogen meta-
exchange of K for Na (Eqn. 16), the other of K and Na for somatism and deposition of calcite at s0me stage.
Ca (Eqn. I7), was pointed out by ARNORSSONet al. (1983b) Rather than starting from the high Aln-end of the silica
and BENJAMINet al. in 1983. The latter derived two inde- saturation lines in Fig. 6, the reaction path describing the
pendent equations, valid below and above 100C respectively interaction of rising COr-containing geothermal fluids then
and suggested that a range of alteration phases thermody- is likely to hover in the vicinity of the full equilibrium point,
namically similar to feldspars, rather than feldspars them- deviating significantly to higher Al-values only when relative
selves control relative cation contents. In another recent ex- rates of cooling of the fluids become much faster than those
perimental study, POPE et al. (1987) found that the Na-K- of re-equilibration with the rock. The chemical environment
Ca-geothermometer worked well for dilute chloride solutions prevailing during the conversion of CO2 to calcite, over large
but not in the case of NaHCOl solutions. parts of a geothermal system, therefore, is likely to correspond
Based on these findings, it may be concluded that many closely to that governed by the full equilibrium buffer and
of the problems associated with the application of the Na-K- the reaction linking COZ-fugacity and water chemistry may
Ca geothermometer arise from its sensitivity to variations in be written in terms of endmembers as
the C02-content of geothermal fluids especially at lower tern-
3K-feldspar + CO2 + Ca F?
peratures. Rather than forming the basis for a geothermom-
eter, &contents of lower temperature geothermal solutions, muscovite + calcite + 6 silica + 2K+ (21)
therefore, may to good advantage be used in the formulation
of a valuable COr-geobarometer. its equilibrium constant corresponds to
Well before the arrival of the Na-K-C?+geothermometer, KZI = (~~l~a~*J~X~musc.~~,Ite(l~~lcJ~~icroc~im). (22)
ELLIS(1970) had already pointed out the sensitivity ofcalcium
contents to variations in the CO2 content of thermal waters Again the assumption +: = yr is made, silica activities are
and had proposed a technique for the evaluation of deep assumed to be those of chalcedony solubility and (o,,,~~,,~,~u-
COr-partial pressures by use of Na-Ca-contents. The proce- dci,JUbiemlim) is %t to unity. Under these circumstances the
dure was based on the assumption of equilibrium of the waters values of L2,= log K2, as calculated by data given by BOWERS
with an assemblage containing feldspar, illite and calcite. et al. (1984) are with - 1.66 + 0.15 from 50 to 300 almost
Calcite is an ubiquitous mineral in geothermal systems temperature independent. Together with the conversion fac-
(BROWNE,1978) and geothermal discharges have been shown tor from activities to mg/kg (Eqn. I), the relationship linking
t0 be Close to saturation with respect to calcite on many 00 analytical K- and Ca-contents tofcs is for practical purposes
casions (ARNORSSONef al., 1983b, WHITE, 1986). adequately given by
1760 W. F. Giggenbach

Lkc = log (ctc+/cc,z+)= log fro, + 3.0. (23) low to lead to COz attack on Ca-aluminium silicates and the
formation of calcite. The values of_/& obtained then provide
Application of Eqn. (23) again is, of course, limited to systems only a qualitative indication. The relative positions of most
for which close approach of equilibrium with all the mineral of the well and associated spring discharges, straddling the
phases involved in reaction (21) can be assumed. In the at+ full equilibrium line, however, clearly indicate attainment of
sence of calcite the correlation between Lkr and CO2 does COz-fugacities conducive to the formation of calcite at some
not apply. stage during the rise of these waters to the surface.
Notwithstanding the temperature independence of Lt,, The C02-fugacities derived by use of the K- and Ca-con-
C02-fugacities in geothermal systems are strongly dependent tents of some of the well discharges (Eqn. 23) may be com-
on temperature. In order to be able to correlate the C02- pared to those obtained on the basis of measured total dis-
fugacities obtained by use of Eqn. (23) to their likely equil- charge gas contents. Assuming the gases to have been dis-
ibration reference temperature, a geothermometer equili- solved in a single liquid phase, the compositions of steam
brating with similar speed and under similar conditions to discharged from Wairakei and Broadlands wells (GIGCEN-
that of the K-Ca-geobarometer is required. The nearest such BACH, 1980) correspond to COZ-fugacities of between 0.2 to
system is likely to be that involving K and Mg as described 1.I b and 7.0 to I7 b, respectively. The upper limits, likely
above. The overall geobarometer then again is based on three to be most representative of the undisturbed state of these
solution components and, therefore, would be suitable for systems, agree exactly with those obtained on the basis of K-
evaluation by use of a triangular diagram. Because of the and Ca-contents (Fig. IO).
intricate correlations among/co, and the CO1-contents of The data points for the acid waters fall all well below the
coexisting liquid and vapor phases, a more open presentation full equilibrium line, a behaviour not too surprising taking
as shown in Fig. 10 is preferred. into account the certain absence of calcite and feldspars in
Analytical values of Lt, are plotted against 4,. Again a the mineral suites coexisting with these waters. Comparison
line representing full fluid-rock equilibrium (Eqn. 20) is with the lines representing the dissolution of varying amounts
shown together with stability boundaries for Ca-Al-silicates of average crustal rocks (Table I ) clearly shows again that
(GIGGENBACH, 1984). Data points below the full equilibrium their composition is controlled by rock dissolution rather
line correspond to the compositions of comparatively im- than equilibration. For such waters, already identified as im-
mature fluids with higher than full equilibrium CO1 fugacities mature by use of Fig. 9, evaluation of equilibration j& by
and, therefore, being reactive with respect to hydrogen meta- use of their K-Ca-contents is not possible.
somatism. Values above this line (ZU, CP, TO) are likely to Of the eight soda spring waters, almost all plot close to the
reflect comparatively low proportions of deep CO+zh vapors rock dissolution lines. Again taking into account the im-
contributing to the geothermal fluid at depth. In these systems mature nature of these waters as indicated by Fig. 9 this is
the CO*-contents of the fluid tapped by the drillhole are too quite expected. The two soda spring waters showing some

co2
0
(bar)
L kc

001

010

100

10.0

100
0 neutral chloride woters

FIG. 10. Evaluation of CO1-fugacitiesin geothermal systems by use of K, Mg and Ca contents of their discharge
waters. For symbols see Table 1 and Fig. 2. The stability fields of some CYa-Al-silicates
are shown as a function of
jcq and temperature together with lines representing solutions of typical erustal rocks.
 Geothermal solute equilibria 2761

apparent approach to fluid/rock-equilibrium, those from Te involving much fewer components, would have given the same an-
Aroha (TE) and Radkersburg (RA) (Fii. 9), give equilibration swer.
temperatures of around 130 at CO*-fugacities of 0.9 and 10 Most of the drawbacks of single species and global tech-
b, respectively. Roth are associated with active deposition of niques are overcome by the use of isomolar concentration
calcite at the surface thus confirming calcite supersaturation, ratios in the case of gases (GWENBACH, 1987) or isocou-
at least under surface discharge conditions. Deposition of lombic concentration quotients, as discussed above, in the
calcite at shallow levels or even during sampling may then case of ionic solutes. These approaches are in accordance
be responsible for the very low Ca-content of the Radkersburg with the general rule that comparison of analytical and theo-
sample and the correspondingly very high apparent CO*-fu- retical information in geochemical systems should be carried
gacity. The COZ pressure indicated for the Te Aroha water out as much as possible on the basis of actually measured
of around 1 bar is that expected for such lower temperature, variables and variable by variable (GIGGENBACH, 1987). The
calcite depositing springs. two diagrams represented by Figs. 9 and 10, based on readily
From the above findings it can be concluded that evalu- available analytical information, fulfil this condition; the
ation of fcol by use of K-Ca contents of discharge waters is overall treatment essentially preserves the identity of each
only reliable for data points close to the full equilibrium line. variable and allows secondary processes affecting each variable
For data points deviating significantly, the full equilibrium to be assessed individually.
line may be taken to separate waters from two distinct alter- The above findings also show that one of the most im-
ation environments: those plotting below in Fig. 10 are likely portant additional requirements in the formulation of a
to come from an alteration system dominated by acid fluids, geoindicator is the strict delineation of its limits of applica-
those plotting above fro.m a rock dominated, COzdeficient bility. In the graphical evaluation of water-rock equilibration
environment. by use of Fig. 9, two areas marked partially equilibrated waters
and immature waters were distinguished. They are defined
PIUCIICAL APPLICATIONS by the full equilibrium line and another apparently quite
arbitrary curve separating the area where the Na-K-Mg-geo-
The wide range of chemical geothermometers hitherto thermometer may be applied with confidence from one rep
proposed and to varying degrees presently in use may be resenting immature waters likely to reflect the effects of rock
subdivided into several categories, depending on the number dissolution rather than rock equilibration. In order to facilitate
of components involved in their formulation. The most sim- numerical evaluation, a maturity index measuring the de-
ple ones are those based on uni-variant reactions such as the gree of attainment of water-rock equilibrium would be useful.
silica geothermometer and a number of gas geothermometers Such an index may be obtained by first combining the Na-
as proposed by ARNORSSONet al. (1983a,b) and discussed K- and K-Mg-geothermometers as represented by the equa-
by ARNORSSONand SVAVARSSON(1985). Their major dis- tions
advantage is their high sensitivity to secondary processes such
as dilution, and in the case of gas geothermometers, to vapor tkn = 1390/(1.75 - I&) - 273.2 (24)
loss or condensation and errors in the estimates of vapor-
tk,, = 4410/( 14.0 - Lk,,,) - 273.2 (25)
liquid ratios.
At the other end of the spectrum are global techniques to form a temperature independent equation for the full
involving a large number of constituents and based on the equilibrium curve according to
inherent assumption that all these constituents have been in
equilibrium simultaneously at some stage and that their Lkn = 0.315&m + 2.66 = &,,, + MI (26)
compositions are still representative of the conditions in this
deeper equilibration zone. Examples are gas geothermometers or
proposed by GIGGENBACH(1980) on the basis of reactions cNa = 457~~~~~~~ = mi&+c&. (27)
involving the formation of CH., and NH3, or those based on
the reconstruction of the deep equilibrium fluid as proposed, The coefficient s = 0.315 represents the ratio of vant Hoff
e.g., by MICHARD and ROEKENS (1983) and REED and slopes for the two geothermometers, the value of mi = 457
SPYCHER(1984). the concentration of sodium required for given values of
Application of these techniques is valid in the case of deep well c$~c~~~ to obtain full equilibrium. As discussed in detail
discharges likely to be representative of the deep equilibrium envi- above (Fig. 8), the major cause of non-attainment of water-
ronment, but questionable in the case of natural surface discharges. rock equilibrium in geothermal solutions approaching full
For these fluids, subject to secondary processes such as partial m-
equilibration to lower temperatures, dilution or steam loss, the results equilibrium is the slowness ofthe process supplying the com-
become inconclusive. Major disadvantages of these multicomponent paratively large equilibrium contents of Na especially in lower
techniques then are their dependence on virtually complete and rep temperature solutions. Lack of equilibrium, therefore, is likely
resentative analyses of thermal discharges and on the availability of to be reflected in Na-contents of the waters being too low.
reliable thermodynamic information for a large number of mineral
The apparently quite arbitrary line in Fig. 9 separating im-
and fluid phase species, their limited ability to allow any further
ready interpretation or intercomparison among a large number of mature from- partially equilibrated waters actually corre-
samples and their general unwieldiness. A useful aspect of these global sponds to a value of MI, the maturity index, of 2.0 ac-
techniques is their ability to reveal attainment of equilibrium among cording to
a large number of fluid and mineral components. Under these cir-
cumstances, however,any geoindicatorbased on a suitable subsystem, MI = 0.315&, - Lkn = 2.0 (28)
 W. F. Giggenbach

c,poo+ c,/mJ + 5 =
""~-Na"=~,,/lOS

K/100

0
L kc

- L 1 I 1 i.. t .._ !.

1 2 3 4 7

I I. Blank diagrams for the evaluation of Na-K and K-Mg equilibration temperatures (A) and of COrfugacities
use of Na, K, Mg and Ca contents of thermal waters. The C&-contents of liquid and vapor phases as a function
and tempera& are only valid for low salinity systems. The rock dissolution line of Eig. 1 tb refers to solutions
100, 1000 g of an average crustal rock in 1 kg of water.
 Geothermal solute equilibria 2763

Of peratures. Simple dilution by non-mineral&d waters also

causes minor shifts in the position of data points towards the
Mg-comer. In the case of boiling chloride springs, even if
The curve thus de&& however, should only be taken as a unaffected by admixture of immature waters, K-Na-temper-
rou8h guideline as the reliability of the Na-K-Mg-geother- atures am likely to reflect minimum values, deep temperatures
mometer is likely to decrease gradually with MI, the reliabiity being often considerably higher as shown by the position of
of tkn decreasing more rapidly than that of tkm. The main data points for well and associated spring samples in Fig. 9.
value of the MI = 2.0 curve lies in its use of distinguishing Close attainment of water-rock equilibrium even at lower
waters suitable for the application of the K-Mg-Ca-geoba- temperatures, e.g. in essentially stagnant reservoirs of for-
rometer or not. mation waters, is indicated by the water discharged at Morere
Equations (24), (25) and (28) allow a purely numerical (MO). Seawater has a maturity index of 2.09 well below that
evaluation of the Na-K-Mg-geoindicator. The graphical expected for full equilibrium of 2.66. Both tkn of 162C and
technique of Fig. 9, however, has the advantage of allowing tkm of 97C possibly reflect the high teIqm?itUIeS of its in-
simultaneous evaluation of a large number of samples, a pre- teraction with basaltic material along mid-ocean ridges
liminary visual statistical assessment and the delineation of As pointed out above, before applying the K-Mg-Ca-geo-
trends and groupings. In order to facilitate application of the barometer the suavity of a water should be checked by use
graphical technique a blank version of Fig. 9 is given in Fig. of Fig. 1la or by determining its maturity index which should
1la. It may be photocopied or reproduced by use of the data be above 2.0. Again a blank diagram is presented in Fig. 1lb.
8iven in Table 2. Relative Na, K, Mg-contents required in The curves relating CO*-fugacities and the CO+ontents of
plotting the curve for MI = 2.0 by computer may be obtained coexisting vapor and liquid phases were drawn by use of data
from Eqn. (29) by setting cK to unity or by dividing the given in Table 2 assuming ideal gas behaviour and may,
amounts of Na given in Table 2 by 4.57. therefore, be applied only to low salinity systems (~0.1 m)
Figure 1 la also contains a curve marked weirbox, it de- at pressures < 100 bar. The relationships used are
scribes the position of waters subject to maximum steam loss
_k02 = k02.vfH20
through flashing from their full equilibrium temperature to
1OOC. Again lines for rock dissolution are given, they rep = fco2.1 BcozfHzo (30)
resent the composition of waters resulting from the dissolution
of varying amounts of crustal rocks (Table 1) in 1 kg of water. where r, are the xm&mttiOS &&nHfi in the vapor and
For waters plotting below the full equilibrium line, K-Mg- liquid phase respectively, fH20 the vapor pressure of water
temperatures are lower than K-Na-temperatures. Because of and BCO,the distribution coefficient rco,,V/rcoz,lof COz at a
the faster rate of equilibration in the system K-Mg, values of given temperature. The temperature dependences offHzo and
tkmam likely to reflect conditions at shallower levels, with tkn Bml (GIOGENBACH, 1980) are given, with adequate precision,
those at considerable depth. A number of additional processes by
such as admixture of immature waters with their generally
high Mg-contents will also lead to deviations from the full log fH&, = 5.5 1 - 2~8/(f + 273.2) (31)
equilibrium curve and to different K-Na- and K-Mg-tem-
log Bcol = 4.7593 - 0.01092t (32)
with tin C.
TABLE II
The CO+ontent of the fluids in wt% may be obtained
Data useful in the construction of Figs IlA and IlB, from the values of rcoz by use of
~1 in m&kg, fi in bar. Number of eqn in brackets.
%C02 b.w. = 244r~oJ(1 + 2.44rco1). (33)
t'c) CNa CK =MS fH20 Bco2 fco:!
(12) 113) (32) (201
The full equilibrium line is 8iven by Eqn. (20) and may be
expressed in terms of&, according to
20' 47 900 100 109880 0.02 34 750 .0004
Lkc = 1% hihb)
40' 48 75s 100 12035. 0.07 23 010 .0008
60' 26 404 100 1719.0 0.20 12 710 .0017 = log fCQ + 3.0
80" 15 327 100 306.10 0.47 7 690 .0037
100' 9 431 100 65.600 1.01 4 650 .0079 = 0.0168t - 0.78
120' 6 097 100 16.430 1.98 2 810 .0172
1400 4 111 100 4.7080 3.61 1 700 .0373 = 74.09/( 14.0 - 4,) - 5.37.
160' 2 875 100 1.5140 6.18 3 030 .0809 Again the ranges of water compositions resulting from the &so-
1600 2 075 100 0.5380 10.02 622 .I754 lution of 10, 100 and l#O g of an average crustal rock (Table I) in
2009 1 540 100 0.2090 IS.54 376 .3600 1 kg of water are shown. By applying Fig. 1lb only to waters with
220' 1 170 100 0.0874 23.16 227 .8240 MI > 2.0, those reflectingrock dissolution should have already heen
240' 909 100 0.0392 33.44 138 I.787 eliminated. In spite of its low maturity index of 2.06, just within the
2600 719 100 0.0187 46.86 83 3.670
applicability range of 2.0 to 2.66, the data point for seawater is also
260*
given in Fii. I I b. Its position again is far from the Ml equilibrium
579 100 0.0094 64.12 50 8.395
line but quite close to the basalt dissolution tine (Fii. 10) suggesting
300" 473 100 0.0049 85.81 30 18.20 that seawater represents a solute system far from any equilibrium
320-- 392 100 0.0027 112.70 18 39.45 involving minerals likely to control &id chemis&y in geothermal
340' 329 100 0.0016 145.90 11 85.50 systems.
2164 W. F. Giggenbach

Figure 10 also contains stability ranges of Ca-aluminium silicates BOWERST. S., JACKSONK. 1. and HELC~ESON H. C. ( 1984) Equilih-
as a function offco2 and temperature as derived earlier (GIGGEN- rinm Activify Diugrums. Springer-Veriag.
BACH, 1984). Minerals commonly observed in geothermal alteration BRIMHALLG. H. and GHIORSOM. S. (1983) Origin and ore-forming
assemblages (BROWNE, 1978). but with highly variable and frequently consequences of the advanced argillic alteration process in hypogene
unknown compositions, such as smectites and epidotes, have been environments by magmatic gas contamination of meteoric fluids,
omitted. Their potential stability ranges overlap with those of their l+on. Geol. 78, 73-90.
thermodynamic proxies laumontite and wairakite. With increasing BROWNEP. R. L. ( 1978) Hydrothermal alteration in active geothermal
ftol the latter are converted to kaolinite or pyrophyllite. The corre- Gelds. Ann. Rev. Earth Plant. Sci. 6, 229-250.
sponding boundaries describe maximum CO2 fugacities for the for- BURNHAMC. W. (1979) Magmas and hydrothermal fluids. In Geo-
mation of calcite from Ca-aluminium silicates according to reaction chemistry of Hydrothermal Ore Deposits (ed. H. L. BARNES),pp.
(19) and delineate, together with the full equilibrium curve (Eqn. 7 I- 136. J. Wiley & Sons.
20). the range of calcite formation in rising geothermal fluids and, CAPUANOR. M. and COLE D. R. (1982)Fluid-mineral equilibria in
therefore, the range of applicability of the K-Ca-geobarometer. A a hydrothermal system, Roosevelt Hot Springs, Utah. Geochim.
simplified eqn describing this upper fco, limit, valid from 80 to Cosmochim. Acta 46, 1353- 1364.
320C. is CATHLESL. M. (1977) An analysis of the cooling of intrusives by
logs,,, = 0.018% - 3.48. groundwater convection which includes boiling. Econ. Geol. 72,
(35)
804-826.
Data points outside this range are likely to represent waters from CHOU L. and WOLLASTR. ( 1984) Study of the weathering of albite
highly immature alteration systems dominated by acid fluids or from at room temperature and pressure with a fluidized bed reactor.
rock dominated alteration systems with low reactivities of CO?, as Geochim. Cosmochim. Acta 48,2205-22 Il.
indicated. The effects of a series of secondary processes such as con- ELLISA. J. (I 970) Quantitative interpretation of chemical charac-
ductive cooling, dilution and boiling onfco, have been discussed in teristics of hydrothermal systems. Geothermics 2, 5 16-528.
detail earlier (GIGGENBACH,1984). ELLISA. J. (1971) Magnesium ion concentrations in the presence of
magnesium chlorite, calcite, carbon dioxide, quartz. Amer. J. Sci.
The above techniques were derived for rising hydrothermal 271,481-489.
solutions where interaction with the rock can be assumed to ELLISA. J. and MAHONW. A. J. (1964) Natural hydrothermal systems
consist largely of initial rock dissolution in waters formed and experimental hot water/rock interactions. Geochim. Cosmo-
chim. Acta 28, 1323- 1357.
through absorption of magmatic vapors followed by more ELLISA. J. and MAHONW. A. J. (1967) Natural hydrothermal systems
gentle H-metasomatism associated with the conversion of and experimental hot water/rock interactions (Part II). Geochim.
CO2 to calcite and of acid clays to neutral aluminium Cosmochim. Acta 31, 5 19-538.
silicates. The other major alteration process is K-metaso- ELLISA. J. and MAHONW. A. J. (1977) Chemistrv and Geothermal
Systems. Academic Press, 392p. .
matism in major upflow zones in response to conductive or ELLIS A. J. and WILSON S. H. (1960) The geochemistry of alkali
isenthalpic (boiling) cooling (GIGGENBACH, 1984). For other metal ions in the Wairakei hydrothermal system. N. Z. J. Geol.
systems the above techniques may, in modified form, still be Geophys. 3, 593-6 17.
applicable. In the case of seawater driven alteration systems EUGSTERH. P. (1985) Granites and hydrothermal ore deposits: A
(SEYFRIED and BISCHOFF,198 1) the starting fluid is obviously
geochemical framework. Mineral Msg. 49, 7-23.
FOUILLACC. and MICHARDG. (1977) Sodium, potassium, calcium
seawater; in the case of interaction of fluids derived through relationships in hot springs of Massif Central. Proc. 2nd Intl. Symp.
leaching of evaporitic material, initial water compositions on Water-Rock Interaction, Strasbourg 3, 109-I 16.
may be characterised by the predominance of sodium with FOURNIERR. 0. (1979) A revised equation for the Na/K geother-
the full-equilibrium curve of Fig. 9 being approached from mometer. Geoth. Resources Council, Trans. 3, 22 l-224.
FOURNIERR. 0. and POTTERR. W. (1979) Magnesium correction
the Na-corner. to the Na-K-Ca chemical geothermometer. Geochim. Cosmochim.
Acta 43, 1543- 1550.
FOURNIERR. 0. and TRUESDELLA. H. (1973) An empirical Na-K-
Acknowledgements-The author wishes to thank R. L. Goguel, Ca geothermometer for natural waters. Geochim. Cosmochim. Acca
Chemistry Division, DSIR, for his expert help with atomic absorption 37, 1255-1275.
analyses and R. 0. Foumier and R. W. Henley who reviewed an GIGGENBACHW. F. (1975a) Variations in the carbon, sulfur and
earlier version for helpful comments. chlorine contents of volcanic gas discharges from White Island,
New Zealand. Bull. Volcanol. 39, 15-27.
Editorial handling: E. J. Reardon GIGGEI\;BACH W. F. (1975b) The chemistry ofcrater Lake, Mt. Rua-
pehu (New Zealand) during and after the I97 I active period. New
Zealand J. Sci. 17, 33-45.
REFERENCES
GIGGENBACHW. F. (1977) The isotopic composition of sulphur in
ARNORSSONS. and SVAVARSSON H. (I 985) Application of chemical sedimentary rocks bordering the Taupo Volcani Zone. In Geo-
geothermometry to geothermal exploration and development. chemistry 1977; NZ DSIR Bull. 218, pp. 57-64.
Geoth. Resources Council, Trans. 9,293-298. GIGGENBACHW. F. (1980) Geothermal gas equilibria. Geochim.
ARNORSSONS., GUNNLAUGSSONE. and SVAVARSSON H. (1983a) Cosmochim. Acta 44,202 l-2032.
The chemistry of geothermal waters in Iceland. II. Mineral equi- W. F. (198 1) Geothermal mineral equilibria. Geochim.
GIGGENBACH
libria and independent variables controlling water compositions. Cosmochim. Acta 45, 393-4 IO.
Geochim. Cosmochim. Acta 42, 523-536. GIGGENBACH W. F. (1984) Mass transfer in hydrothermal alteration
ARNORSSONS., GUNNLAUGSSON E. and SVAVARWDNH. (1983b) systems. Geochim. Cosmochim. Acta 48,2693-27 I I.
The chemistry of geothermal waters in Iceland. III. Chemical GIGGENBACH W. F. (1985) Construction ofthermodynamic stability
geothermometry in geothermal investigations. Geochim. Cosmo- diagrams involving dioctahedral potassium clay minerals. Chem.
chim. Acta 42,567-577. Geol. 49, 23 l-242.
BENJAMINT., CHARLESR. and VIDALE R. (1983) Thermodynamic GIGGENBACH W. F. (1987) Redox processes governing the chemistry
parameters and experimental data for the Na-K-Ca-geothermom- of fumarolic aas dischames from White Island, New Zealand. Appl.
eter. J. Volcanol. Geoth. Res. 15, 167-186. Geochem. 2,?43- 16 I. -
B~RISOV M. V., RYZHENKO B. N. and KRAYNOVS. R. (1985) In- GIGGENBACHW. F. (1988) Water and gas chemistry of Lake Nyos,
fluence of the acid-base properties of rocks on the composition of Cameroon, and its bearing on the eruptive process. J. Volcano/.
equilibrated aqueous solutions. Geochemistry Intl., pp. 87-94. Geoth. Res (in press).
 Geothermal solute equilibria 2765

GIGGENBACHW. F. and GLASBYG. P. (1977) The intluence of ther- momary and studies of boiling and dilution. Geochim. CosmcF
mal activity on the trace metal distribution in marine sediments chim. Acta 48, 1479-1492.
around White Island, New Zealand. NZ DSIR Bull. 218. 121- RUSE W. I., CHUAN R. L., GIGGENBACHW. F., KYLE P. R. and
126. SYMONDSR. B. (1986) Rates of sulfur dioxide emissions from
GIGGENBACHW. F. and LE GUERNF. (1976) The chemistry of mag- White Island volcano, New Zealand, and an estimate of the total
matic UBXSfrom E&Ale. Ethiooia. Geochim. Cosmochim. Acta flux of major gaseous species. Bull. Volcanol.48, I8 I- 188.
40,25:30. RUAYAJ. R. and SEWARDT. M. (1987)The ion-pair constant and
GIGGENBACHW. F., G~NFIANTINI R., JANGI B. L. and TRUESDELL other thermodynamic properties of HCl up lo 350C. Geochim.
A. H. (1983) Isotopic and chemical composition of Parbati Valley Cosmochim. Acta 51, 121-130.
geothermal discharges, NW-Himalaya, India. Geothermics12,199- SASSB. E., ROSENBERG P. E. and KH-IRICK J. A. (1987) The stability
222. of illite/smectite during diagenesis: An experimental study. Gee-
WUEL R. L. (1983) The rare alkalies in hydrothermal alteration chim. Cosmochim. A& 51,.2 103-2 115.
at Wairakei and Boradlands geothermal fields, NZ. Geochim. Cos- SAVAGED. (1986) Granite-water intemctions at 100. 50 MPa: An
mochim. Acta 47.429-437. experimental &dy. Chem. Geol. 54,8 I-95.
HEMLEY1. J. (1967) Aqueous Na/K ratios in the system K20-Na@- SAVAGED., CAVE M. R., MILDOWSKIA. E. and GEORGEI. (1987)
A&O,-Si02-H20. Program, 1967 Ann. Meeting, Geol. Sot. Amer. Hydrothermal alteration of granite by meteoric fluid: An example
New Orleans,pp. 94-95. from the Cammellis Gmnite, United Kingdom. Contrib.Mineral.
HEMLEY1. J. and JONES W. R. (1964) Chemical aspects of hydm Petrol.96, 39 I-405.
thermal altemtion with emphasis on hydrogen metasomatism. SCHRAMKEJ. A., KERRICKD. M. and LASAGAA. C. (1987) The
Econ. Geol. 59,538-369. reaction muscovite+quartz * andalusite+K-feldspar+water. Part
HENLEYR. W. and MCNABBA. (1978) Magmatic vapor plumes and I. Growth kinetics and mechanism. Amer. J. Sci. 287, 5 17-559.
groundwater interaction in porphm copper emplacement. Econ. SCHOEN R., WHITE D. E. and HEMLEYJ. J. (1974) Argillization by
Geol. 73, I-20. descending acid at Steamboat Springs, Nevada. ClaysClayMineral.
HENLEYR. W., TRUF~DELLA. H., BARTONP. B. and WHITNEY 22, l-22.
J. A. (I 984) Fluid-mineral equilibria in hydrothermal systems. Rev. SEYFRIEDW. E. and BISCHOFFJ. L. (1981) Experimental seawater-
in Econ. Geol. 1,267. basalt interaction at 3OOC,500 bars; chemical change, secondary
HOLDRENG. R. and SPEYERP. M. (1985) pH dependent changes mineral formation and implications for the lranspofl of heavy
in the rates and stoichiometry of dissolution of an alkali feldspar metis. Geochim. Cosmochim. Acta 45, 135-147.
at room temperature. Amer. J. Sci. 285, 994- 1026. SHIKAZONO N. (1976) Thermodynamic interpretation of Na-K-Ca
HOLLANDH. D. (1972) Granites, solutions and base metal deposits. geothermometer in the natural water system. Geochem. J. 10,47-
Econ. Geol. 67,28 l-301. 50.
JANECKID. R., CHARLE!SR. W., BAYHURSTG. K. and BENJAMIN SHIKAZONON. (1978) Possible cation buffering in chloride-rich geo-
T. M. (1986) The physicochemical basis of the Na-K-Ca geother- thermal waters. Chem. Geol. 23, 239-254.
mometer. Los Alamos Report LA-10806 MS. IOp. SILLITOER. H. (1973) The tops and bottoms of porphyry copper
KRUPP R. E. and SEWARDT. M. (1987) The Rotokawa geothermal deposits. Econ. Geol. 68, 799-8 15.
system, New Zealand: An active epithermal golddepositing en- STOFFREGEN R. E. (I 985) Genesis of acid-sulfate alteration and Au-
vironment. Econ. Geol. 82, 1109-I 129. Cu-Ag mineralisation at Summitville, Colorado. Ph.D. thesis. Univ.
LINLXAY T. W. (1980) Estimation of concentration quotients for of California, Berkeley, 205~.
ionic equilibria in high temperature water: The model substance TAYLOR S. R. (I 964) Abundance of chemical elements in the con-
approach. Proc. Intl. Water Cot& Pittsburgh,pp. 284-294. tinental crust. a new table. Geochim. Cosmochim. Acfa 28, 1273-
LOWELLJ. D. and GUILBERTJ. M. (1970) Lateral and vertical al- 1285.
teration-mineral&ion zoning in porphm ore deposits. Econ. Geol. THOMPSONJ. F. H., LESSMANJ. and THOMPSONA. J. B. (1986) The
65.373-408. Temora gold-silver deposit: A newly rrcognised style of high sulfur
MICHARDG. and FOUILLACC. (1976) Remarques sur le thermom&re mineral&ion in the Lower Paleozoic of Australia. Econ. Geol.
Na-K-Ca. J. Volcanol.Geoth. Res. 1.297-307. 81,732-738.
MICHARDG. and ROEKENSE. (1983) Modelling of the chemical TRUESDELLA. H. (1975) Geochemical techniques in exploration.
composition of alkaline hot waters. Geothermics12, 161-169. Proc. 2nd UN Symp. on the Developmentand Use of Geothermal
ORVILLEP. M. (1963) Alkali ion exchange between vapor and feldspar Resources I, 53-86.
phases. Amer. J. Sci. 261, 201-237.
WALSHE1. L. (1986) A six-component chlorite solid solution model
PACES T. (1975) A systematic deviation from Na-K-Ca geother-
and the conditions ofchlorite formation in hydrothermal and geo-
mometer below 75C and above lo- atm PC&. Geochim. Cos-
thermal systems. Econ. Geol. 81.681-703.
mochim. Acta 39.54 l-544.
PETITJ. C., GIANANTONIO D. M., DRANJ. C., SCHOTTJ. and BER- WHITE D. E. (1957) Magmatic, connate, and metamorphic waters.
NER R. A. ( 1987) Mechanism of diopside dissolution from hydrogen Geol. Sot. Amer. Bull. 69. 1659-1682.
depth profiling. Nature 325.705-707. WHITEA. (I 986) Chemical and Isotopic characteristics of fluids within
POPE L. A., HAJASHA. and POPP R. K. (1987) An experimental the Baca geothermal reservoir, Valles Caldera, New Mexico. J.
investigation of quartz, Na-K, Na-K-Ca geothermometers and the Geophys.Rex 91, 1855-1866.
effects of fluid composition. J. Volcanol.Geoth. Res. 31, 15 1- I6 I. WHITE D. E. (1965) Saline waters of sedimentary rocks. In Fluids
REED M. and SPYCHERN. ( 1984) Calculation of DH and mineral in Subsufice Environments. Symp. Amer. Assoc. of Petroleum
equilibria in hydrothermal waters with applicaiion lo geother- Geologists, pp. 342-366.