Geothermics Vol. 27, No. 2. pp.

183-196, 1998
C. 1998 CNR
Pergamon Elsevier Science Ltd
Printed in Great Britain. All rights reserved
0375-6505/98 $19.00+0.00

Pll:

CHANGES IN SILICA CHEMISTRY AND HYDROLOGY

ACROSS THE ROTORUA GEOTHERMAL FIELD, NEW
ZEALAND
RICHARD B. GLOVER I and EDWARD K. MROCZEK
Institute QfGeological and Nuclear Sciences, Wairakei Research Centre, P.B. Taupo, New
Zealand
(Received October 1996; accepted June 1997)

Abstract—A comprehensive review of field data from Rotorua geothermal

field, New Zealand, was undertaken to evaluate the mixing relationships of
cooler groundwaters with the primary fluid. Unlike previous studies the
reservoir well compositions were calculated using only the most reliable
measured downhole fluid temperatures. Inconsistencies in gas chemistry
showed that the hot fluid in the southeast could not be derived from a
postulated 2500 C parent by simple boiling, a mechanism proposed by
previous authors. Mixing relationships showed conclusively that across the
field from southeast to northwest, all the fluids were related along a mixing
line with low-chloride water at 1500C. Between 1 58 and 175 0 C the diluting
fluid changed to a low-chloride groundwater at 15 0 C. An SiOÜ temperature
plot of all the well data showed that silica values for the hot wells fall around
the quartz solubility line. The data for the majority of the cooler wells scatter
about the a-cristobalite solubility line. Plots of the Cl/Si02 ratio were used to
evaluate changes in silica chemistry across the reservoir by comparing
measured data with theoretical mixing lines, assuming partial equilibrium
with respect to the three silica polymorphs, quartz, a-cristobalite and
amorphous silica. The only assumptions made in the calculation were that the
150 and 15 0 C diluting fluids were in equilibrium with cristobalite and
amorphous silica, respectively. Surprisingly, the trends may be explained by
mixing with cooler silica-rich diluting fluids without requiring equilibration
with respect to any of the silica polymorphs. 0 1998 CNR. Published by
Elsevier Science Ltd.
Key words: geochemistry, silica, Rotorua, New Zealand.
 R. B. and E. K'. Mroczek
Present address: Glover Geothermal Geochemistry, 33c Brandon Rd., Glen Eden,
Auckland, New Zealand.
183
184 Glover

INTRODUCTION
Rotorua geothermal field, located in the Taupo Volcanic Zone, has been the focus of intense
geological, geophysical and geochemical investigations over the past few years because of the
conflict between increasing exploitation of the subsurface fluid and the effect on New Zealand's
last remaining geyser field (Allis and Lumb, 1992b). At the culmination of the monitoring
programme, the important conclusions of which were published in a special issue of
Geothermics (Allis and Lumb, 1992a), there was a comprehensive collection of interdisciplinary
results, on which the present study is based.

Geohydrolog.r
The Rotorua geothermal field is located near the southern margin of a caldera which collapsed
following the eruption of the Mamaku Ignimbrite, now dated at 220 ka (Wood. 1 992; Houghton
el al.. 1995). Following the emplacement of the ignimbrite, rhyolite eruptions built lava flow
and dome complexes ("Rotorua City domes" of Wood. 1992), mostly within the southern sector
of the caldera. The Rotorua basin is filled with mainly finegrained sediment and tephra deposits
which cover the Mamaku Ignimbrite surface and most of the Rotorua City domes. The main
aquifers exploited by the boreholes are the Rotorua City domes and the Mamaku Ignimbrite.
The sedimentary sequences generally have poor permeability and act as aquitards for the rhyolite
and ignimbrite aquifers (Wood, 1992). Thin beds of pumice gravels, sands and tephra cover the
siltstones, and comprise a shallow aquifer containing warm geothermal groundwater and
discharged bore emuent.
The Rotorua field (Fig. l) is the most extensively drilled geothermal system in New Zealand.
Since the 1920s over 950 wells have been drilled to supply domestic and commercial space-
heating and bathing water, and although there are only seven drillholes deeper than 300m (the
average well is about 120m deep), the hydrology of the system down to these levels is fairly
well known. In contrast to fields exploited for industrial and power developments, where deep
wells reach temperatures of 300 C but little is known about the shallow, cooler levels, Rotorua
h'as a large data base of fluids with temperatures less than 220 C. Chemical data (Glover, 1967;
Glover and Heinz, 1985; Stewart et al., 1992; Giggenbach and Glover. 1992) and physical data
(Grant et al., 1985: Allis and Lumb, 1992b) indicate upflow of 250 C water in the south and east
(Whakarewarewa and Ngapuna). There is also an upflow of > 200 C water in the northwest at
Kuirau Park (Wood, 1992). Some waters flow west and north towards Lake Rotorua.

Chemical changes bet)l'een drillhole discharges'

Calculated downhole chlorides are highest in the south and east (up to about 1400 mg/kg) and
decrease in waters to the west and north. Previous authors (e.g. Stewart et al.. 1992) have
suggested that this is due to dilution by a deeply circulating groundwater (about 150 C) and also
some near-surface dilution. However, it was noted (Glover, 1967) that changes in the silica
content across the field did not present a clear pattern. Glover (1967) observed that the silica did
not decrease as much as the chloride, so that the Cl/Si00 ratio decreased. He suggested that this
may have been due to admixture of cooler water such as the Utahina Stream, which contained
little chloride but 60 mg/kg of silica. A second possibility was that cold water inflowing from
the west was becoming heated and dissolving silica from the rock more readily than chloride: a
subsequent admixture with the thermal
185
 Changes in silica chemistry and hydrology across the Rotorua geothermal field

Fig. l. Location of geothermal and groundwater wells across the Rotorua geothermal field.
Wells, except G, are classified according to the proximity of the downhole temperature and
silica concentrations to the equilibrium solubility curves as shown in Fig. 3.

water resulted in a larger decrease in chloride than in silica. These observations were
the stimulus for the present study.

DATA USED
Chemical analyses reported by Glover and Heinz (1985) included all available data at
that time. Data were obtained from geothermal (R R) drillholes, mostly sampled on
discharge, and also from shallow groundwater (G) wells, mostly sampled at depth ( <
10 m). Since then more data have been obtained and the present paper includes some
from
 R. B. and E. K'. Mroczek
186 Glover

geothermal drillholes that were sampled at depth (up to 242 m), along with temperatures
measured at the same time.
In order to obtain the downhole chemistry' for a particular discharging drillhole, the chemical
concentrations were multiplied by the water fraction at the surface. The latter was obtained from
the discharge enthalpy, based on the "best" downhole temperature. The temperatures used were
provided by C.P. Wood (pers. comm.) after an extensive search of all temperature
measurements. The downhole chemistry calculation retains the same chemical ratios as are
present in the analysed sample but neglects any species that have been lost in the steam fraction,
such as C.O. and HS.
Tables list well number, downhole temperature, concentrations of silica, chloride, sulphate
and total carbonic species expressed as HCO*, together with the weight ratios of Cl/Si02 and Cl
/HCOÄ, for the discharging wells, "G" wells, and downhole samples. respectively. Locations of
wells used in this study are shown in Fig. l.
The temperatures (and enthalpies) used in this study differ from those used by Stewart
et al. (1992), which were derived from the silica concentrations in the surface
discharges. Stewart et al. (1992) assumed that the waters above 190 C were saturated
with quartz, and waters less than 1 75 -C were saturated with a-cristobalite, and also
used some measured downhole temperatures to derive well enthalpies.

DISCUSSION
Changes in HCOR and C/
Glover and Heinz (1985) noted an increase in total HCO* as the waters moved from southeast
to northwest. As the waters also decreased in Cl there was a resultant decreasing trend in
CI/HCO*. They noted that dilution below approximately 500 mg/kg Cl did not change the
CI/HCO* significantly, deducing that the diluting water from the west is low in chloride and
bicarbonate.
Stewart el al. ( 1992) suggested that there were two diluting waters, a deep one causing
cooling and the formation of HC03, and a shallow one bringing in tritium and oxygen. The
shallow groundwater containing no chloride, with about 50 mg/kg HCOR and a 2 5 tritium ratio
(T R), must have a fairly direct route from the surface since the tritium mean residence time is
less than about 40 years (Stewart et al.. 1992). The shallow groundwater also contains air, which
accounts for the high sulphate concentrations in some of these dilute waters (see Fig. 2).
As waters move from southeast to northwest they cool from 250 to 150 I C by dilution, the
equilibrium dissolved C02 concentration with respect to the rock decreases greatly, while the
actual C02 in solution is only slightly decreased (Stewart et al., 1992). Hence the excess
dissolved CO. begins to react with the rock to form HC03 and re-establish the equilibrium (see
Fig. 8 in Giggenbach, 1984). The process requires a considerable residence time in the lower
temperature rock -water reaction zones as the carbonic acid is neutralised. On the other hand the
south and east waters cooled by boiling either in the ground or in the well during discharge are
greatly depleted in dissolved C02. which contributes to their high CI/HCOR ratios. Intermediate
Cl/HC03 ratios can be due to a small degree of boiling before dilution or to a smaller degree of
dilution and reaction of C02 with the rock.
Si02 vs temperature
Figure 3 shows a plot of downhole silica vs temperature using the data from the RR
wells. Equilibrium lines for quartz. chalcedony. a-cristobalite and amorphous silica are
also
 Changes in silica chemistry and hydrology across the Rotorua geothermal field
Table l . Chemical composition of samples from Rotorua wells calculated from surface
discharges (concentrations in mg/kg at downhole temperatures calculated from surface
discharge)

(Horer M
Table 2. Chemical composition of samples from the shallow "G" wells (mg/kg)
 R. B. and E. K'. Mroczek

Table 3. Chemical composition of dowr nhole samples (mg/kg)

shown. The wells are characterised by their proximity to the quartz, I-cristobalite, and
amorphous silica solubility curves. Data from the shallow groundwater (G) wells are also plotted
but are not characterised by their proximity to any of the solubility curves. All quartz category
wells, except one. lie in the south and east. The remaining well is in the north. close to the
inferred secondary upflow (Glover. 1967; Stewart et al., 1992; Giggenbach and Glover, 1992).
The trend show's a similar pattern to chloride, with lower concentrations at lower temperatures.
The intriguing question is whether the apparent change in control of the silica concentration
to a different silica polymorph is real or not. The apparent relationship may be due to the steep
shape of the equilibrium curves, i.e. most of the high-temperature wells lie near the quartz
equilibrium line and as the temperature is reduced the same silica con-
 Changes in silica chemistry and hydrology across the Rotorua geothermal field

10S04 5HC03
Fig. 2. Relative concentrations of Cl, S04, and HC03 (mg/kg) grouped according to
temperature. V: > 180 c; 150-180 0 c; o: 120-1500 0, A: < no o c.

250

Fig. 3. Si02 vs measured fluid temperatures and theoretical equilibrium solubility correlations
for the silica polymorphs. O: quartz; cristobalite; A: am. silica; V: G wells.

centration would appear to be in equilibrium with chalcedony and then with I-

cristobalite. This problem was resolved by compiling theoretical mixing lines that took
into account silica reactions and the available field data (see later).

Mixing end members

In order to investigate how the different mixing processes and equilibria change with
temperature, the wells in Table I are grouped on the basis of temperature in Figs 4-6.
"G"
 R. B. Glover and E. K'. :Mroczek
250

400 800 1200 1600

Cl/(mg/kg)
Fig. 4. Cl vs measured fluid temperatures and mixing relationships. V: > 1 80 C; o: 1 5() 1 8() C; O • 1
20—1 50 C; V: G wells: A: downhole.

well and downhole samples (Tables 2 and 3) are also plotted and most fit in the following mixing
pattern.
Deep-water component. Some possible deep-water components are shown in Table 4. A case
could be made for a 25() C "parent" fluid. The composition of this fluid would depend on how
the analysed waters. particularly the hottest (213 C), high chloride fluid (found in well R R 889),
are assumed to be derived from it. Stewart el al. (1992) use adiabatic boiling and hence calculate
a parent water with 1250 mg/kg chloride [(e) in Table 4]. However, difficulties arise when the
gas analyses for the steam samples collected from RR889 are

250

Fig. 5. Cl/SiO. vs measured fluid temperatures and mixing relationships. V: > 180 C: o: 1 50—180 C: O:
1 20--1 50 C; A: < 120 c: V G wells: A: downhole•, A, B, C, D, E: see text.
 Changes in silica chemistry and hydrology across the Rotorua geothermal field 191
7

o 50 100 150 200 250

Fig. 6. Cl/Si02 vs measured fluid temperatures showing the effect of equilibration with respect to the
silica polymorphs. V: > 180 0 C; , O: 120-150 C; A: < 120 Q C; V: G wells; A: downhole.

used to derive the deep fluid chemistry. Combining the steam with the separated
water to form the single phase at 21 YC [(b) in Table 4] gives Cl = 1420 mg/kg, Si02 =
314 mg/kg and carbonic species equivalent to CO, = 1730 mg/kg. The minimum depth
to boiling (assuming 100 C at the surface) is 214m; the dissolved C02 adds 4.5 bar
pressure, giving a boiling depth of 267m (cf. well depth = 244m and casing depth = 235
m). This is reasonable, considering the uncertainties in the data. However, if it is
assumed that the 250 C water boiled to reach 21 YC and the steam formed was in
equilibrium with the 21 YC water, the composition of the 250 0 C fluid [(c) in Table 4]
would then be: Cl= 1289 mg/kg, Si02 = 487 mg/kg, and carbonic species equivalent to
C02 = 44 760 mg/kg. For this fluid the depth to boiling would be 1660 m. The high gas
concentration and depth to boiling suggest that the assumptions in the calculation are
flawed.
A second way of deriving the 21 YC fluid is by dilution from 250 D C [(d) in Table 4]. If the
diluting fluid was at 150 t 'C with zero chloride, then the chloride concentration in the 250 liquid
would have been 2304 mg/kg, C02 = 2810 mg/kg and depth to boiling 523 m.

Table 4. Calculated concentrations (mg/kg) in possible deep-water components

(b) 213 1420 314 267

(c) 250 1290 44 760 487 1660
(d) 250 2300 2810 487 523
250

(a) Total carbonic species expressed as C02.

(b) As calculated from analysed steam and water data.
 R. B. Glover and E.
(c) 213 'C fluid derived by adiabatic boiling from 250 C.
(d) 21 YC fluid derived by dilution with 150 0 C fluid with zero chloride and zero carbonic species. (e)
From Stewart et al. (1992).
1 92 K.

These characteristics are quite reasonable, but because of the uncertainties in calculating the
composition of a hypothetical 250"C parent fluid, in these calculations the 213 C fluid from
RR889 has been used as the deep-water component.
Diluting component. The plot of chloride vs temperature for the unboiled fluids discharged
by the drillholes is shown in Fig. 4. Regressions to zero chloride for drillholes with temperatures
above 1 50 C and for drillholes that are close to the quartz and chalcedony solubility lines in Fig.
3 intercept the temperature axis at 1 50.2 and 1 53.1' C, respectively. Thus, an initial diluting
fluid of 150 C and zero chloride concentration is assumed.
Regression to zero chloride in a plot of chloride vs silica for the same wells intercepted the
axis at 235 mg/'kg. This graph (not shown) had a large scatter in the data and since the possibility
arises that the silica is not conserved this would give an invalid correlation. Thus the I-
cristobalite solubility at 150 C (261 mg/kg) was used as the silica concentration of the 1 50 C.
zero chloride diluting component.
Cold diluting component. The plot of chloride vs temperature (Fig. 4) shows that most fluids
with temperatures below 1 50 C could be derived from dilution with a low-chloride. low-
temperature fluid. It is assumed here that the diluting fluid is at 15 C, has zero chloride
concentration and a silica concentration of 96 mg/kg, i.e. in equilibrium with amorphous

Silica chemistry
Interpretation of the temporal and spatial changes in silica concentrations across the Rotorua
geothermal field requires an understanding of the hydrothermal chemistry of silica as well as the
hydrology. The deep source fluid is postulated to be at about 250 C, at which temperature the
quartz will control the silica solubility (Fournier, 1985). Rapid ascension of this fluid occurs
towards the southeast of the field, accompanied by decompressional boiling which would result
in an increase in solute concentrations. Although there is uncertainty as to the initial temperature
of the parent fluid and the temperature at which boiling may have commenced, at shallower
depths the cooler fluid is fully pressurised (no boiling or excess steam) and equilibrated at about
213 C with the reservoir rocks (viz. well RR889 discharges). If equilibrium with respect to quartz
were to be maintained during cooling of the parent fluid from 250 to 213 C by boiling,
conductive cooling or by dilution with a 150 fluid, then 40, 34 and 5 0 of the initial dissolved
silica, respectively, would need to precipitate as quartz.
Quartz controls silica solubility' at temperatures greater than about 180 C and most reservoir
fluids greater than 140 •C (Fournier, 1985). Although quartz is thermodynamically the most
stable phase, under certain circumstances the silica solubility may be controlled by the less stable
but more soluble phases, cristobalite or amorphous silica. If a fluid is saturated with respect to
quartz but encounters cristobalite or amorphous silica minerals, then the most soluble phase will
dissolve first and for a time control the solubility. Similarly, if the fluid becomes supersaturated
with respect to all three phases, the least stable phase, i.e amorphous silica, will deposit first
even though the fluid may be significantly more saturated with respect to cristobalite or quartz.
This consecutive sequence of precipitation reactions, from the phase with the highest solubility
to that with the least, is known as the Ostwald step rule and occurs because of the much faster
nucleation kinetics of the least stable phase (Stumm, 1992).
The rate at which the disequilibrium with respect to quartz (or one of the other polymorphs)
re-adjusts to equilibrium is dependent on the degree of saturation, the extent of
 Changes in silica chemistry and hydrology across the Rotorua geothermal field 193
the system (ratio of the rock surface area to water mass) and the rate constant (Rimstidt
and Barnes, 1980; Renders et al., 1995). The rate constant depends on temperature and
many other parameters (possibly unidentified), and is inversely and exponentially
related to the activation energy. The activation energy barrier for mineral reactions
occurring at the solid—liquid interface is large, and so the rates of precipitation (or
dissolution) of silica are expected to be extremely slow at low temperatures, even at high
supersaturation. Rimstidt and Barnes (1980) showed that because of these competing
effects the maximum deposition rate often occurs 25 to 500 C below the saturation
temperature. At temperatures below the temperature of maximum deposition, silica is
often quenched into solution. In an idealised example of fluid flow in a system
dominated by fractures they calculated that little silica precipitates as quartz at
temperatures below about 200 0 C. Similarly, slow deposition kinetics for cristobalite
and amorphous silica, at least where homogeneous nucleation is not a competing
process, means that silica may also be held in solution.
Their data also showed that the temperature dependence of the precipitation rate
constant was the same for quartz, cristobalite and amorphous silica, leading to the same
deposition rate at similar supersaturation. Renders et al. (1995) suggested that although
the temperature dependence of the precipitation rate constant of cristobalite is similar to
that of quartz (i.e same activation energies), the precipitation rate may be somewhat
faster due to a higher reaction entropy.
In terms of the developed hydrological model, fluid flows towards the northwest,
mixing with two cooler groundwaters and resulting in reservoir temperatures < 180 0 C.
The longer residence times may still be sumcient to ensure that fluid re-equilibrates with
respect to quartz. An SiO2 temperature plot of all the well data, using measured
temperatures and silica values calculated to reservoir conditions, is shown in Fig. 3. As
expected, the data for the hot wells fall around the quartz solubility line but the data for
the majority of the cooler wells scatter about the oc-cristobalite solubility line. This
suggests that either the fluid is passing through fractures containing crystalline
cristobalite, which is controlling the solubility, or metastable conditions with respect to
cristobalite are maintained by the fluid residence times being shorter than the time
required for nucleation/deposition of cristobalite.
Joints and fractures are presumed to be the dominant control of permeability of the
Mamaku Ignimbrite and Rotorua Rhyolite aquifers, though the latter also has a pervious
pumiceous bed at its top (Wood, 1992). Tectonic activity may open fresh fractures,
perhaps containing appreciable quantities of cristobalite, but this is considered an
unlikely scenario. In any event, the removal of cristobalite is very fast on a geologic
time scale. For example, at 1700 C the dissolution rate constant for cristobalite is 1.48x10
-9 mol/(m2 s) and the equilibrium solubility of quartz and cristobalite is 174 and 329
mg/kg, respectively. Assuming the fluid is initially in equilibrium with quartz, then the
rate of cristobalite dissolution is 1.3 g/(m2 yr) (Renders et al., 1995). At this high rate it
would be surprising if cristobalite (or glass and amorphous silica) dissolution was
controlling the silica solubility in the reservoirs.
Alternatively, the silica concentration may be controlled by the precipitation of
cristobalite, which will begin as its saturation solubility is exceeded. According to
Ostwald's step rule quartz will not precipitate, even though the solution is supersaturated
with respect to quartz, until the dissolved silica concentration reduces to near the
 R. B. Glover and E.
equilibrium solubility with respect to cristobalite. Depending on the flow rate, the
equilibrium solubility may never be reached because of the slow kinetics and
progressive dilution and cooling of the
194 K. Mroczek

reservoir fluids. The lower temperature fluids are not in equilibrium with quartz, which suggests
that the fluid residence times are much shorter than the quartz deposition kinetics, which is slow
at temperatures less than What appears surprising is that the fluid seems to be close to
equilibrium with respect to cristobalite to quite low temperatures. Assuming no dilution, a fluid
in equilibrium with quartz at 200 will only exceed the equilibrium solubility of cristobalite after
conductively cooling to below 148 L C. Further cooling will increase silica saturation and hence
precipitation; however, as discussed above, the precipitation rate constant is very small at these
low temperatures so that the silica will be effectively held in solution. In addition, precipitation
of cristobalite must imply widespread occurrence of this polymorph throughout the upper
reservoir. The hydrothermal precipitation of crystalline cristobalite can only occur by epitaxial
growth onto pre-existing cristobalite nuclei (Renders et al., 1995). Similarly, hydrothermal
quartz can only deposit on pre-existing quartz nuclei while amorphous silica can grow on almost
any substance (Rimstidt and Cole, 1983).

Changes in Cl/SiO,
The simple hydrological model derived above shows that the concentrations of the
conservative chloride may be adequately explained by progressive dilution of fluid flowing to
the northwest with two distinct low-chloride groundwaters. The Cl/Si02 temperature plot (Fig.
5) is used to evaluate changes in silica chemistry across the reservoir by comparing field data
with theoretical mixing lines assuming partial equilibration with respect to the three silica
polymorphs. The only assumption made in constructing the diagram was that the 150 and 15 -J
C diluting fluids were in equilibrium with cristobalite and amorphous silica, respectively. Fluid
interacting with pumiceous rocks can rapidly dissolve silica without a significant change in
chloride (Glover, 1967). Point A is the ratio of the typical fluid in equilibrium with respect to
quartz at 21 3 C. The line segments AB and AD represent mixing with the hot diluting fluid to
175 and 158 C, respectively, being the approximate bounds for the change from the hotter to the
cooler diluting fluid as deduced from the chloride temperature plots. Lines EB and ED are the
mixing lines with the cool diluting fluid while line BC represents an equal increment change
from 100% 150 C diluting fluid at 1 75 D C to 100% 15 0 C diluting fluid at 140 N C. Precipitation
of silica would shift the mixing lines up to higher Cl/Si02 ratios and conversely dissolution
would shift the mixing lines down to lower Cl/Si02 ratios. Figure 6 shows the curve labelled
ABCE in Fig. 5 and the shift from this line that would result if some equilibration occurred at
each mixing step for each of the three polymorphs (l and 100 0 0 equilibration shown).
To reach the amorphous silica equilibrium mixing line would require significant dissolution
of the solid amorphous silica across the whole temperature range. For similar reasons as
discussed above, amorphous silica is expected to persist for even less time than cristobalite in
the reservoir. The measured data give ratios well above the calculated amorphous silica
equilibrium mixing lines so that this phase is not controlling any of the silica reactions, except
possibly at low temperatures below 50 C. Comparison of the theoretical ratios with field data
for quartz and cristobalite shows that equilibrium with respect to these polymorphs is not likely
to have occurred at temperatures below 175 and about 12(Y'C. respectively.
Between 175 and about 120 C the cristobalite data are explained by mixing or by
equilibration with respect to cristobalite. If the intermediate mixing line ABCE (no
equilibration) were to shift to the bottom of the mixing envelope (DE), deposition would
need
 Changes in silica chemistry hydrology across the Rotorua geothermal field 195
and
to occur to match the field data, and conversely dissolution of silica if the mixing line were to
move towards the top of the mixing envelope (BE).
Between 213 and 175 0 C the comparison between field and calculated data suggests
that mixing without precipitation of quartz best explains the data (since the starting point
"A" is not fixed).

Comparison with mineralogic data

The deductions from the field chemical data and mixing calculations may perhaps be
verified by comparison with known mineralogy and petrology of the Rotorua reservoir
rocks. Although a considerable number of wells have been drilled there are very few
cores, which severely limits the quality of the analysis (C.P. Wood, pers. comm.).
Cuttings have been retrieved from some wells, but there are many gaps in the data,
particularly for the high-temperature wells in the southeast. Many of the mineralogic
analyses are for bulk compositions only and do not contain sumcient detail for a valid
comparison with the present field data (Wood, 1970—1992).
Cristobalite is often, but not always, found at shallow depths below the capping rocks
at temperatures up to 1400C. At higher temperatures cristobalite is rare, and completely
absent above 180 0 C. Although hydrothermal precipitation of cristobalite cannot be
excluded, no textural evidence (epitaxial overgrowths) is presented in the mineralogic
records to support widespread precipitation of cristobalite from the mixing cooler fluids.
Drusy quartz crystals were found in cuttings from only one well, at a depth
corresponding to temperatures above 2000 C, and provide unambiguous evidence for the
hydrothermal deposition of quartz.

CONCLUSIONS
A comprehensive review of the field data in part supported previous deductions about
the hydrology. Unlike previous studies, we calculated the reservoir well compositions
using only the most reliable measured downhole fluid temperatures in the wells.
Examination of the C02 content of 213 0 C fluid in RR889 showed that this hot fluid in
the southeast could not be derived from a postulated 2500C parent by simple boiling, a
mechanism proposed by previous authors. Mixing relationships showed conclusively
that, across the field from southeast to northwest, all the fluids were related along a
mixing line with low-chloride water at 1500C. Between 158 and 175 0C the diluting fluid
changed to a low-chloride groundwater at 15 0 0.
The trends in the Cl/Si02 field data at temperatures less than 175 0 C, shown in Figs 5
and 6, can be explained by mixing with cooler silica-rich diluting fluids without
requiring equilibration of the mixed fluid with respect to any of the silica polymorphs.
Above this temperature limited equilibration with respect to quartz may have occurred
but the evidence from the mineralogic record for widespread deposition is not
conclusive. Therefore, the apparent correlation of the temperature and cristobalite
solubility shown in Fig. 3 may be coincidental. This conclusion was unexpected and is
 R. B. Glover E.
most likely the result of the unusual geohydrological features of the Rotorua geothermal
field: large fracture-controlled permeabilities with consequent lower surface-to-rock
ratios than might otherwise be expected, low reservoir temperatures and a relatively hot
(150 0 C) diluting low-chloride/high-silica
196 and K. Mroczek
Acknmvledgements The authors would like to thank M. K. Stewart, C.P. Wood, C. Panichi and
H. Kristmannsdottir for constructive comments that improved the manuscript.

REFERENCES
Allis. R. G. and Lumb, J. T. ( 1992a) Rotorua geothermal field, New Zealand. Geothermics

Allis, R. G. and Lumb. J. T. ( 1992b) The Rotorua geothermal field, New Zealand: its physical
setting, hydrology, and response to exploitation. Geothermics 21, 7—24.
Fournier, R. O. (1985) The behaviour of silica in hydrothermal solutions. In Geology)' and
Geochemistr.v o/Epithermal Svstems, Reviews in Economic Geology, eds B. R. Gerger and
P. M. Bethke, Vol. 2, pp. 45-72.
Giggenbach, W. F. (1984) Mass transfer in hydrothermal systems --a conceptual
approach. Geochimica et Cos•mochimica Acta 48, 2693 271 1 .
Giggenbach, W. F. and Glover, R. B. (1992) Tectonic regime and major process governing the
chemistry of water and gas discharges from the Rotorua geothermal field, New Zealand.
Geothermics 21, 12 1 140.
Glover, R. B. (1967) The chemistry of the thermal waters at Rotorua. Nevi' Zealand Journal of
Science 10, 70-96.
Glover, R. B. and Heinz, H. (1985) Chemistry of Rotorua waters. In The Rotorua Geothermal
Field. Technical Report Q/ the Geothermal Monitoring Programme 1982—1985, pp. 295—
354. Ministry of Energy, Wellington.
Grant, M. A.. McGuiness, M. J. , Dalziel, S. R. , Y. Razali and O'Sullivan, M. J. ( 1985) A model
of Rotorua geothermal field and springs. In The Rotorua Geothermal Field. Technical Report
of the Geothermal Monitoring Programme 1982—1985, pp. 471—493. Ministry of Energy,
Wellington.
Houghton, B. F. , Wilson, C. J. N.. McWilliams, M. O.. Lanphere, M. A. , Weaver, S. A. , Briggs,
R. M. and Pringle, M. S. ( 1995) Chronology and dynamics of a large silicic magmatic system:
Central Taupo Volcanic Zone, New Zealand. Geology 23, 1 3—1 6.
Renders, P. J. N. , Gammons. C. H. and Barnes. H. L. (1995) Precipitation and dissolution rate
constants for cristobalite from 1 50 to 300 -C. Geochimica et Cosmochimica Acta 59,

Rimstidt, J. D. and Barnes, H. L. (1980) The kinetics of silica water reactions. Geochimica et
C'osmochimica Acta 44, 1683—1699.
Rimstidt, J. D. and Cole, D. R. ( 1983) Geothermal mineralization l: The Mechanism of
Formation of the Beowawe, Nevada, Siliceous Sinter Deposit. American *Journal o/ Science
283, 861 875.
Stewart, M. K. , Lyon, G. L.. Robinson, B. W. and Glover, R. B. (1992) Fluid flow in the Rotorua
geothermal field derived from isotopic and chemical data. Geother;nics 21, 141 164.
Stumm, W. ( 1 992) Chemistry o/ the Solid-Water Interface. John Wiley, New York, 428 pp.
Wood, P. (1992) Geology of the Rotorua system. Geothermics 21, 25—4 1 .
Wood, P. (1970- 1992) Unpublished mineralogic and petrographic analyses of Rotorua wells.
New Zealand Geological Survey Files.