PROCEEDINGS, Thirty-Sixth Workshop on Geothermal Reservoir Engineering

Stanford University, Stanford, California, January 30 - February 1, 2012

SOIL GAS FLUX EXPLORATION AT THE ROTOKAWA GEOTHERMAL FIELD AND

WHITE ISLAND, NEW ZEALAND

Simon BLOOMBERG1, Clinton RISSMANN2, Agnes MAZOT3, Christopher OZE1, Travis HORTON1, Darren
GRAVLEY1, Ben KENNEDY1, Cynthia WERNER4, Bruce CHRISTENSON3 and Joanna PAWSON1
1
University of Canterbury, Pvt Bag 4800, NZ
2
Environment Southland, Pvt Bag 90116, NZ
3
GNS Science, Pvt Bag 2000, NZ
4
Cascades Volcano Observatory, Vancouver, Wa
E-mail: [email protected]

advanced considerably over the last 50 years since

ABSTRACT some of the first measurements of heat and mass flow
from volcanic hydrothermal systems. In particular the
A significant challenge to geothermal exploration is
measurement of CO2 flux and its use as a proxy for
accurate quantification of the heat and mass flow heat and mass transfer has greatly improved our
between deep reservoir(s) and the surface. Here, we understanding of the heat and mass release from high
use high resolution measurement of carbon dioxide
temperature reservoirs as well as the nature of
(CO2) flux and heat flow at the land surface to
structural controls over fluid flow (Chiodini et al.
characterise the mass (CO2 and steam) and heat
2005; Fridriksson et al. 2006; Viveiros et al. 2010;
released from the geothermal reservoir. Statistical
Werner & Cardellini, 2006).
characterisation of background soil respired CO2 flux
reduces the level of uncertainty when deriving mass Numerous soil CO2 flux surveys of volcanic and
(emissions) and heat flow estimates from high
magma-hydrothermal systems have been undertaken
temperature reservoirs.
over the last 20 years. Initially, flux surveys sought to
better quantify the contribution of diffuse emissions
We report the preliminary results of two soil gas and
from volcanic systems to the global carbon cycle.
heat flow surveys for an active andesitic From this initial work the value of flux surveys as a
stratovolcano (White Island) and a high temperature method for assessing heat and mass flow and for
geothermal field (Rotokawa), both of which occur
investigating structural controls to fluid flow was
within the Taupo Volcanic Zone, New Zealand. The
recognised. Increasingly CO2 flux surveys are being
flux and heat flow surveys include over 3000 direct
used by researchers of volcanic and volcanic-
measurements of CO2 flux, soil temperature and hydrothermal systems as well as by geothermal
current and historic fumarolic discharge samples. exploration geologists (Brombach et al., 2001;
Initial results include a total CO2 emission rate of 633
Cardellini et al., 2003; Chiodini et al., 1998;
16td-1 (2.5km2) for Rotokawa while at White Island
Dereindra & Armannsson, 2010; Evans et al. 2001;
we report a total CO2 emission rate of 116 2td-1 for
Mrner & Etiope, 2002; Lewicki et al., 2005;
the crater floor (.3km2). Using CO2:H2O molar ratios
Rissmann et al., 2012).
the thermal energy release associated with diffuse
degassing is 317MW and 54MW respectively. Here we present the preliminary results of detailed
CO2 flux and heat flow surveys for White Island
These preliminary results shed light on the heat and
(WI), an active andesitic stratovolcano, and
mass flow dynamics of an active andesite volcano
Rotokawa (RK) a high temperature magma-
(White Island) and a high temperature magma-
hydrothermal field, both within the TVZ of New
hydrothermal field (Rotokawa). Zealand. Total heat flow (MW) and emission rates
(td-1) are quantified for each survey along with
INTRODUCTION detailed maps of the spatial extent and magnitude of
Heat and mass energy transfer through volcanic and soil gas flux and heat flow. Normalised total emission
hydrothermal systems in the Taupo Volcanic Zone rates are contrasted with values reported for fields
(TVZ) of New Zealand has historically been within the TVZ and internationally (Mrner and
measured using surface geophysical, geochemical Etiope, 2002; Rissmann, 2010; Werner and
and geological techniques. Modern techniques for Cardellini, 2006) and spatial maps of surface flux and
surface measurement of heat and mass flow have heat flow are used to infer structural controls to fluid
flow. A comparison is also made between the by surface sulphur mining activity in the past 40
observed and CO2-derived heat flow estimates for years.
both fields. Finally, the source of CO2 at each field is
assessed on the basis of cumulative probability plots
and by comparison with CO2 flux and soil
temperature.

GEOLOGICAL SETTING

White Island

White Island (Whakaari) is an active andesitic

stratovolcano located 50 km to the north east of
Whakatane, on strike with the eastern edge of the
TVZ. The volcano is one of New Zealands most
active volcanoes with frequent eruptions over the last
40 years (Houghton and Nairn, 1989) and numerous
eruptions in historic times (Cole and Nairn, 1975;
Wardell et al., 2001), the latest eruptive sequence
being phreatomagmatic. The island, though being
surrounded by seawater on all sides, has a semi-
sealed acidic hydrothermal system (Houghton and
Nairn, 1989). The gas chemistry is typical of active
arc volcanism with a high gas input from subducting
marine sediments (Giggenbach, 1995).

The lithology of the crater floor is comprised of

volcanic sediments and ash from eruptions, landslides
and ancestral crater lakes. The island is devoid of
vascular plant species with the exception of the
seaward flanks of the volcano. Thermal activity
consists of steaming vents, active fumaroles
(>220C), acid streams and pools, steaming and
boiling mud pots and pools, inactive and fossil
features are numerous. The western crater is filled
with a large boiling acid lake (pH -0.2) which is the
centre of modern day phreatomagmatic activity.
Subsurface permeability at WI is controlled by
hydrothermally-altered clay cap layers within which
ascending steam condenses and sulphur minerals
precipitate. Areas of high permeability create
degassing structures and are often linked with
landslide hummocks that have infilled the crater floor
during historic eruptions. Fumarolic activity has
been monitored for the last 50 years (Giggenbach
1975a, 1983; Rose et al. 1996).

Rotokawa

The geothermal system at RK has been under

scientific inspection for the best part of 60 years
(Hedenquist et al., 1988; references therein). The
principal thermal feature is a warm (~24C) acid (pH
2.2) lake with an area of ~0.62 km2. Directly to the
north lies an area of steaming ground, hot pools, sink
holes, sulphur banks, explosion craters and
fumaroles. This area has been extensively modified
Figure 1: Mapped locations of soil gas measurements coupled with an adaptive sampling component
at RK (bottom) and WI (top) with respect to the (Boomer et al. 2000). Adaptive sampling was
Taupo Volcanic Zone, New Zealand. initiated when a measurement exceeded the
background flux level (>3ppm s-1, ambient TC) at
Two geothermal power stations generate from the which point the resolution increased to 5 m spacing.
geothermal reservoir, Rotokawa A (34MW) and Nga Due to time constraints at WI, measurements were
Awa Purua (140 MW). The maximum fluid based on the systematic sampling approach only
temperature is ~320C, recorded within the (Figure 1).
Rotokawa andesite deep reservoir rock. Two-phase
fluid flows from the deep reservoir through to the Data Analysis Methods
surface as the local confining cap layer (the Huka
Falls Formation, a volcaniclastic lake deposit rich in Data Analysis
clays) has been removed in this area by either
hydrothermal eruptions or dissolution. Upon entry into spreadsheet the raw (field) CO2 and
H2S data are converted from ppm s-1 to g m-2 d-1 using
METHODS & MATERIALS an equation that takes into account the ambient
temperature and atmospheric pressure. The raw soil
Field methods temp values were converted from c to W (watts) m-2
(using formulas below). As diffuse CO2 flux from
Two gas measurement methods are used; soil flux soil can come from multiple sources
surveys and subsoil gas sampling for compositional (organic/atmospheric/magmatic) we use Sinclairs
analysis of fumarolic gases (Rissmann et al., 2012). (1974) method of plotting geochemical data on a log-
normal probability plot with a Gaussian distribution
Soil gas flux (CO2 and H2S) and temperature (cumulative probability), the dataset was broken into
measurement: populations and the applied method described in
Chiodini (1998) was used to find the end member and
Soil gases flux (CO2 and H2S) and soil temperatures mixing populations within this studies datasets. From
were measured in the thermal areas at RK and at this graphical statistical analysis (GSA), gas-source
locations within the WI crater floor. The populations were calculated. 5 populations were
accumulation chamber method for measurement of found for each dataset (Appendix: 1). Interpolation of
soil CO2 flux (lewicki et al., 2005; Chiodini et al., un-sampled areas were modelled using the sequential
2008; Rissmann et al., 2012) was used. Due to the Gaussian simulation (sGs) algorithm within the
effects of rainfall on soil gas, data was collected up to WinGsLib software toolbox (Deustsch & Journel,
4 days after heavy rain and 2 days after light rain 1998), following the methods of Cardellini et al.
(Lewicki et al. 2005). Due to the poor response of the (2003). Each simulation ran for 500 realisation at 5m
chemical sensor very few H2S flux values were cell size using variogram models (Appendix: 2) based
measured (~ 100). When in the field, the minor purge on the direct measurements.
time between each consecutive measurement site (< 2
min) did not enable the H2S sensor to recover. Total emission estimates from GSA were compared
Therefore, in the following paper all references to to those from sGs and the results show that GSA
flux relate to CO2 unless otherwise stated. overestimates the contribution of hydrothermal CO2
to the total flux, as is consistent with other studies
Soil gas flux was measured using a West Systems (Cardellini et al., 2003; Rissmann et al., 2012) so
accumulation chamber and LICOR LI-820 infrared therefore only sGs estimates for CO2 emissions were
gas analyser for CO2 (after Welles et al., 2001). A used, however the cumulative probability plots were
mean measurement error of 2% was assumed for helpful in delineating the background populations in
CO2 Flux as is discussed by Giammanco et al. (2007). lieu of isotopic data.
Soil Temperatures were measured to 10cm depth
within ~0.1m of the accumulation chamber footprint
using a Yokogawa TX-10 digital thermometer and a Heat and mass flow through soil (from soil
K-type thermocouple (measurement accuracy temperature)
0.5c). Coordinates for each site were logged using
a Garmin 60x GPS. Over 3000 soil temperature measurements were taken
at RK and WI to a maximum depth of 15cm though
The flux survey design at RK used a systematic most were taken at 10cm, this means when using the
sampling approach (spacing 10 m - 20 m, depending following formulas all deliverables will be
on flux cues such as ground alteration and odour) underestimates. If the boiling point temperature (98.7
at RK and 100 at WI) was reached before 10cm this plume the contribution of soil gas to total emission
was noted. Eqs. 1 & 2. were then applied to the data rates previously were 1% (Wardell et al. 2001) and
based on Dawson (1964) and calibrated at the was recalculated in this study.
Wairakei geothermal field near Taupo. It was
expected that similar conditions exist at RK and WI To calculate equivalent steam flow the molar ratio of
and therefore they can be applied without change. the vapour phase is applied to the soil gas flux data
using:
Where t15 (soil temperature at 15cm depth) is
<98.7C the soil heat flux (qs in Wm-2) is estimated Eq. 4.
by:

Eq. 1. Where is the total flux of CO2 in g s-1, is

the molar ratio in g, and Fstm(CO2) is the steam mass
Where t15 is 98.7C Dawsons (1964) second flow in g s-1. In this way we can calculate the quantity
equation allows estimation using the depth to 98.7C of steam mass flow that is condensed within the
as d98.7 in: subsurface by comparing Fstm(CO2) with the value
from heat flow through soil (Fs (HO), Eq. 3).
Eq. 2
An equivalent heat flow can be calculated for the
Heat flow was then simulated using the sGs method steam mass flow (Fstm(CO2)) derived from Eq. 4.
and the observed and simulated values were using the following equation:
compared. The individual point values are summed to
determine the total heat flow through soil from the Eq. 5.
study area.
Where Fstm(CO2) is from Eq. 4., and is the steam
Mass flow of steam through soil (FS,HO in kg s-1) for flux in g s-1, and hs,100c is the enthalpy of steam at
the study area can also be calculated using the heat 100C (2676 kJ/kg) and Hs is the heat flow in Watts.
flow through soil (Hs) and the heat in steam mass
(hs,100C - hw,t ): RESULTS

Eq. 3. FS (HO) = Diffuse soil gas fluxes and soil temperatures

Rotokawa
Where Hs is the calculated heat flow in Watts from
eq.1-2 normalised for the thermal area, hs,100C is the
enthalpy of steam at 100C (2676 kJ/kg) and hw,t is
the enthalpy of water at the mean annual temperature
for the field areas
(RK=23C;96kJ/kg,WI=10C:41.9kJ/kg;
NIWA.co.nz 2011; Schmidt and Grigull 1979).

Heat, mass flow and emissions calculations from

CO2 flux

In order to compare the contribution of soil gas flux

to total CO2 emissions (in addition to boiling hot
pools and fumaroles) the total volume of diffuse soil
gas flux simulated measurements are summed and
normalised to the area of thermal activity. In absence
of a plume (e.g. Crater Lake at WI), the minimum
contribution to total CO2 emission from diffuse soil
gas flux was 85% and the maximum, 100%
(Dereinda & Armannsson, 2010; Fridriksson et al.
Figure 2: Bi-modal Distribution of CO2 flux and soil
2006). At RK there are steam heated pools, active
temperature measurements at RK.
steaming vents and sulphur banks so it appears likely
that the contribution of soil gas flux to the total
emission value is less than 100%. At WI, which has a
Diffuse soil CO2 fluxes range from <0 to 322,100g some decoupling of steam flow likely due to high
m-2 d-1 from 2,545 direct measurements. Soil water tables adjacent to the lake (Figure 3b). To the
temperatures range from 9 to 100 C. Figure 2 shows north and northwest greater coupling between heat
the correlation between CO2 and soil temperature to flow (high soil temperatures) and CO2 flux indicates
be a general one with two modes of high flux with less scrubbing of the steam phase. The active thermal
low temperatures and the other with high area has been mapped previously using resistivity
temperatures. boundaries and the flux map doesnt produce any
anomalous flux outside of this mapped boundary.

The arithmetic declustered mean of CO2 gas flux was

274 g m-2 d-1 while the sGs produced a mean of 246 g
m-2 d-1. The sGs total emission rate of 340 4 td-1
was simulated in the modelled area (Figure 3a, 1.4
km2). The thermal area at RK is actually larger than
this, so the full thermal field minimum emissions
estimate is provided by normalising the emission rate
by 1.4 km2 (i.e., 8.03x104 t yr-1 km2) and
extrapolating to 2.5 km2. Extrapolation provides an
emission rate of 633 16td-1 or ~50% larger. This
normalised emission rate is purely for soil zone CO2
and does not include measurement of gaseous efflux
from any other fluxing thermal features at RK
(fumaroles, hot pools, mud pools). Removing the
3a background contribution to the total CO2 emission
produces a value of 611td-1 for hydrothermal-CO2
only.

White Island

691 measurements of diffuse degassing were taken at

WI with a range of 0.1 29896 g m-2 d-1 CO2 and
with soil temperatures (10 cm) of 15 to 100C. Figure
4 shows the relationship between CO2 and soil
temperature at WI.

3b

Figure 3a: Soil CO2 flux mapped within the thermal

areas resistivity boundary (Hedenquist et al. 1988),
the Central Field Fault (CFF) is the orange dotted
line. Figure 3b: Soil Temperatures at 10cm (C) are
mapped within the thermal areas resistivity boundary
and CO2 anomalies with high flux but low
temperatures are outlined in white. Surface thermal
features are mapped with black dotted lines in each
image.

Figure 3a shows the high flux (>300 g m-2 d-1)

anomalies are quasi-circular in nature but become
more diffuse and broadly distributed at lower flux
rates. There is a concentration of areas close to the
lake shore as well as along an E-W trending sub-
linear feature north of the lake. The highest flux Figure 4: The distribution of soil CO2 flux and
values occur along the lake shore and coincide with temperature (~10cm, C) at WI. A broad positive
lower average soil temperatures (<50C) indicating trend is observed though there is some weak
separation of high CO2 flux into low and high There are obvious flank and crater wall emissions at
temperature modes. WI, so for the total island soil diffuse emission an
area of 1.5km2 is multiplied by the normalised rate
The sGs models for CO2 flux (Figure 5a) show values (1.34x105 tyr-1km-2) and gives 576 8td-1 CO2, this is
generally diminishing from the Crater Lake to the a maximum total emission. Removing the
south eastern edge of the crater floor. The most background contribution to the total emission gives a
significant fluxes are centrally located in the crater value of 115td-1for hydrothermal emissions.
near hot mud pools, sulphurous mounds and steaming
ground, with the flux anomalies being concentrated H2S Flux and relationships with CO2 at Rotokawa
and constrained to the areas surrounding these
features. H2S flux values from 121 sites range from 0.2 to 156
g m-2 d-1. There is a varied spread in the data with no
Soil temperatures (Figure 5b) are in good agreement correlation to soil temperature due to the limited data.
with the mapped surface features and CO2 flux No H2S flux was ever measured without being in
anomalies; there are only three areas in the crater close proximity to a surface thermal feature. Out of
floor which show both low temperatures and CO2 the 121 sites only 46 have significant H2S flux (>5 g
Flux (Figure 5b). The arithmetic mean for m-2d-1), while 75 of the same sites had significant
declustered CO2 Flux was 367 g m-2 d-1 which was CO2 flux (>100 g m-2d-1). In Figure 7 we plot the
the same as the sGs mean. From the sGs Model an fluxes comparatively and find a threshold
emission rate of 116 2td-1 is computed for the relationship. Here H2S flux is minimal until CO2 flux
survey area (0.31 km2) within the crater floor. values exceed a threshold of ~100 g m-2d-1 at which
point H2S is detected and strong correlation between
both gaseous species is evident.

Two fumarole samples were used to calculate the

CO2:H2S ratio of surface advective steam flow (pipe
flow). The range of their ratios is plotted in Figure 7.
This provides an estimate of the distribution of the
gas composition for advective degassing. While a
few of the soil gas samples fall within this range most
of the data show a higher CO2:H2S ratio.

5a

5b

Figure 5a: Soil CO2 Flux is mapped within the

Figure 7: Distribution of the acid soil gases at RK.
accessible crater floor; surface thermal features are
Data are grouped by CO2 GSA populations. The solid
mapped in both figures with black dotted lines.
grey bar is the estimated fluxes based on fumarolic
Figure 5b: Soil temperatures at 10cm (C) are
gas ratios.
mapped within the crater floor; areas of low
temperatures & CO2 flux are outlined in white.
The CO2:H2S ratio for the fumaroles has a mean of
7.9 while the mean ratio for significant soil gas flux
is 241, historic gas data from wells puts the ratio give an Fstm(CO2) of 10150 250td-1. From this
around 45 (Hedenquist et al. 1988). The trend figure a heat flow of 314 7MW is calculated. The
towards a high CO2:H2S ratio indicates a large 10150td-1 of estimated steam mass flow is equivalent
portion of H2S is either scrubbed by shallow to 117 3kgs-1 which is in good agreement with
groundwater and/or precipitates as elemental S before previous numerical model estimates for RK of
reaching the surface of the field. 105kgs-1 (Bowyer & Holt, 2010).

In order to produce a sulphur budget, the CO2 White Island

emissions associated with all fluxes in excess of 100
g m-2 d-1 are multiplied by the molar CO2:H2S ratio The total hydrothermal-CO2 emission of 115td-1 is
for the soil zone to produce a total H2S emission from multiplied by the molar ratio, 65gCO2/kgH2O or 15.3
diffuse soil degassing of 2.6 td-1. A previous study (Giggenbach & Matsuo 1988), to give a steam mass
from Lake Rotokawa found 10.38 td-1 , which flow (Fstm(CO2)) of 1760 25td-1, from which a heat
equates to 3.67 td-1 H2S (multiplying by the molar flow total of 54 1MW is calculated.
weight ratio, sulphur oxidation) (Hedenquist et al.,
1988; Werner et al., 2008). This gives a minimum DISCUSSION
total surface sulphur emission for RK of 6td-1 with a
minimum total emission of H2S from the reservoir at Diffuse degassing
80td-1 (multiplying the total CO2 by the CO2:H2S
fumarole ratio). The loss of ~74 t d-1 of H2S within Rotokawa
the subsurface is consistent with the extensive
elemental S deposits that characterise the RK field.
Soil gas flux measurements at RK geothermal field
Heat flow through soil cover a wide range of fluxes and indicate the
presence of a large gas anomaly within the thermal
ground area. The mean for the total CO2 population
Over 2,500 and 691 soil temperature measurements when compared with the background removed total
from RK and WI, respectively, were taken at depths (246 vs. 1082 g m-2 d-1) indicates that background
between 0 and 10cm. Using equations 1 and 2 these fluxes must be common spatially and when removed
were modelled with sGs then converted to heat flow the mean increases significantly(Figure 3a).
in Wm-2 and finally into a total megawatt (MW)
value. For RK heat flow measurements ranged The total CO2 emissions at RK calculated from the
between 0.03 16742 with a mean of 28 Wm-2. At sGs and GSA methods have produced the greatest
WI this ranged from 0.28 16742 with a mean of normalised (t yr-1 km-2) value in New Zealand for non
86.6 Wm-2. The total heat flow through soil is volcanic emissions, at 4 times Ohaaki (Rissmann et
estimated to be 73 2.5 and 27 1MW at RK and al., 2012) and 7 times Rotorua (Werner and
WI respectively. Cardellini, 2006). It does however compare to only
half the soil gas emissions of WI (Figure 8). This
Steam mass flow through soil indicates that though RK may not be a high gas
reservoir, all of the gas that is in the system is likely
The total heat flow through soil in watts can be used to end up degassing through the thermal area.
with Eq.3 to find the equivalent steam mass flow (Fs
(H2O)). At RK and WI the surface heat flow values
are equivalent to 2455 td-1 and 898 td-1 respectively.

Mass and heat flow from CO2 degassing

Using Eqs. 4 & 5, an equivalent steam mass flow

value is calculated from the CO2 flux and molar CO2:
H2O ratio for RK and WI fumaroles.

Rotokawa

In order to calculate the total heat input to RK, the

total hydrothermal (611td-1) CO2 flux is multiplied by Figure 8: Plot of thermal ground normalised CO2
the molar ratio (CO2:H2O) of deep fluid, emissions for this study in white and other examples
61gCO2/kgH2O or 16.6 (Hedenquist et al. 1988), to in grey. (Morner and Etiope, 2002; Rissmann 2011)
White Island ratio values produced estimates of 378 or 946MW
thermal.
The difference between the total and background-
removed (hydrothermal) total mean (368 vs. 723 g m-
2
d-1) indicates a significant area has background Spatial and structural relationships of diffuse
emission which is consistent with the flux map surface degassing and heat flow
(figures 5a). The range of flux values at WI, though
not as broad as at RK does have a higher mean value, Rotokawa
which is an indicator of the reduced contribution of
background sources to the total gas emissions In general the location of the thermal area at RK is
through the crater floor (4% background at RK vs. coupled to the location of small prehistoric eruption
1.3% background WI). craters and the major Lake Rotokawa eruption crater.
The location of the thermal area at RK coincides with
Heat flow variables a series of historic hydrothermal explosion craters.
The largest of which is the site of Lake Rotokawa.
For both RK and WI measured soil temperature data Vigorous surface thermal activity surrounds the
was converted to equivalent heat flow using Eqs. explosion craters and includes large outflows of
1&2, whereas heat flow based on CO2 emissions was acidic SO4-Cl boiling springs and steaming ground.
calculated using Eqs. 4&5. Both fields display lower Most diffuse degassing is coupled to these thermal
observed heat flow than that estimated from CO2 features and decreases rapidly in magnitude with
emissions. The discrepancies in heat flow values distance. There are some diffuse degassing areas
likely reflect the condensation of 77% (241td-1) and which show no relationship to Acid SO4-Cl thermal
50% (27td-1) of the steam phase within shallow features and even have low temperatures (<50C) that
meteoric groundwater that overlies the high occur along the lake margins (Figure 3a). Of note
temperature reservoirs at RK and WI, respectively. there is a major structure along a possible fault
escarpment which runs E-W for 300m across the
Fridriksson et al. (2006) report the condensation of thermal area.
87% of steam mass flow for the Reykjanes thermal
area as based on the discrepancy between observed Relationships between diffuse degassing and heat
heat flow and heat flow estimated from CO2 flux, flow at RK are neither purely spatial nor proportional
which matches closely with the discrepancy at RK. and there are some areas which though cold
White Islands hydrothermal source is shallower temperatures are present still produce high gas flux.
(~2km depth) than RK (~4km depth) and more active The locations of these zones near the lake shore
so there may be less steam condensate flowing out of indicate a cooling of the soil possibly due to the
the system. It is also notable that RK is an open water table sitting closer to the surface. The present
system at depth and cool inflows may condense hot spot activity at RK indicates multiple areas of
rising hydrothermal fluid. White Island has a sealed high permeability that channel fluid flow.
hydrothermal system (Houghton & Nairn 1989)
which might encourage more steam mass flow to A high number of these hot spots are located along
reach the surface in the crater area. the northern shore of Lake Rotokawa. Hot spots are
typical of permeability being controlled by alteration
Historic heat flow studies at RK presented in of clay and the movement of heat and gas through the
Hedenquist et al. (1988) put the thermal energy medium. As an area is flooded with hydrothermal
release between 218 and 610MW, with their study fluid and heat, the local strata will alter to clays over
finding 236MW. These studies are all based on time and at some point a feature which once
chloride measurements which can be less robust than exploited an area of high primary permeability will
CO2 and less mobile than the gas phase. For seal itself.
groundwater having reached saturation with respects
to CO2, the majority of the gas passes through to the The hydrostatic pressure of the lake might
surface. This relatively conservative behaviour of act as a barrier which forces the gas to migrate
CO2 makes it a better proxy of heat release from the laterally to the lower pressure shoreline and where
shallow reservoir than Cl or measurement of the gas can rise to the surface, this is evident by
observable heat flow (Chiodini et al., 2005). The bubbling zones within 5m of the shoreline where the
main error associated with any CO2 based estimate of water is shallow (<2m) continuing up on to the land.
reservoir heat release is the selection of a
representative molar H2O/CO2 ratio. The molar ratio Historic mining of sulphur has also affected the
of 16.6 is considered representative, though other topography of the thermal area by creating low-lying
surfaces which now intersect with the water table (2008) it is assumed that a similar process is
allowing hot springs and steaming ground to formed. occurring at RK, where during low CO2 fluxes,
In these areas hydrothermally altered clays from the diffusion is the main mechanism and H2S is removed,
shallow subsurface have been deposited on the but when there is high CO2 flux modest amounts of
surface which reduces permeability. H2S reaches the surface.

Deep field faulting can cause hot spots of high

permeability that allows boiling fluids to rise. There CONCLUSION
is a large field fault running through the middle of
RK, the Central Field Fault (CFF) of Winick et al.
(2011) indirectly influences the diffuse degassing The hydrothermal systems and their respective
surface thermal expressions at RK and WI emit an
structures with historic eruptions craters that appear
extraordinary quantity of CO2 and thermal energy per
to focus thermal activity occurring along its strike.
unit area. This study continues with the style and
The magnitude of the gas fluxes in the vicinity of
approach of recent soil gas surveys at New Zealand
these craters and along the strike of the CFF suggests
a deep-seated connection between surface eruption geothermal systems (Rotorua, Werner & Cardellini
craters and the CFF which may be channelling fluids 2006; Ohaaki, Rissmann et al. 2012) and has found
that RK emits the most CO2 per unit area of any
from depth.
measured geothermal field in New Zealand.
White Island
This study has produced a steam flux that matches
the upflow rate from numerical models based on well
The hydrothermal system at WI is actively degassing gas chemistry.
through permeable areas of the crater floor, crater
walls, crater rim, old crater rims and through mound Areas of high flux are constrained and confirm that
structures on the crater floor (Figure 5a). The mounds deep seated permeability as major normal faults
host numerous small fumarolic vents that are and/or crater rim faults and are paired with surface
encrusted with sulphur condensate, are of high thermal features. The magnitude of both heat flow
temperature and comprised of highly altered clays and CO2 flux decrease with distance from each
and are analogous to the diffuse degassing structures feature.
(DDS) characterised by Chiodini et al., (2005). The
vigour, size and number of DDS along with steaming At RK deep field faults appear to have control over
ground and acidic-Cl outflows increases with the location of thermal activity by controlling the
proximity to the Crater Lake and the modern day location of hydrothermal eruption craters and
centre of volcanic activity. subsequent gas release at the surface. At WI the
greatest concentration of thermal activity is
Most CO2 is exhausted through the Crater Lake and concentrated above the buried crater rim and along
around its rim there is a high flux zone. In the middle the margins of the modern day crater which suggests
of the crater floor there is an arcuate channel of high a deep-seated channel of fluids along crater rim
flux that aligns closely with an old crater rim. This faults-fracture zones.
area is one of the most permeable at WI, with boiling
mud pots and steaming ground common. Diffuse A mechanism for H2S fixing exists in areas of diffuse
degassing, fumaroles, and steaming ground are soil degassing as the concentration of H2S:CO2 in
present at breaks in slope (crater floor/wall, crater fumarole gas is more than is measured in soil gas.
wall/rim, crater floor/mound) which are generally
places of deposition or erosion, so strata would likely Acknowledgements
have high permeability.
Thanks to Mighty River Power Limited (Linda Price,
Elsewhere on the island there are areas of capping
Simon Addison, Jeremy OBrien, the staff at NAP),
clays at around 1m depth that limit degassing locally.
Tom Powell, GNS Science (Jeremy Cole-Baker,
These are hypothesised to be the remnants of the old
crater floor, consisting of impermeable and altered Karen Britten), University of Canterbury (Jim Cole,
volcaniclastic material. Anekant Wandres, Heather Bickerton, Jelte Keeman)
and Thrainn Fridriksson. This Study contributes and
is funded by the UC-MRP Source 2 Surface joint
Sulphur Budget (Rotokawa)
venture research programme with additional funding
H2S occurs in background concentrations until CO2 from the Foundation for Research, Science, &
flux becomes advective (Figure 7). Based on Technology through a TechNZ Scholarship.
interpretations of this phenomenon by Werner et al.
REFERENCES Geochem. 6, 125141.

Hedenquist, J.W., Mroczek, E.K., Giggenbach, W.F., 1988,

Bergfeld, D., Evans, W.C., Howle, J.F., Farrar, C.D., 2006. Carbon Geochemistry of the Rotokawa Geothermal System: Summary of
dioxide emissions from vegetation-kill zones around the resurgent Data, Interpretation and Appraisal for Energy Development: DSIR
dome of Long Valley Caldera, eastern California, USA. J. Chemistry Division Technical Note 88/6, 64
Volcanol. Geotherm. Res. 152, 140156.
Hochstein, M.P., Bromley, C.J., 2005. Measurement of heat ux
Boomer, K., Werner, C. and Brantley, S.L., (2000): CO2 emissions from steaming ground. Geothermics 34, 131158
related to the Yellowstone volcanic system; 1, Developing a
stratified adaptive cluster sampling plan. Journal of Geophysical Houghton, B.F., Nairn, I.A., 1989. A model for the 197682
Research, 105(B5): 10817-10830. phreatomagmatic and Strombolian eruption sequence at White
Island volcano, New Zealand. In: Houghton, B.F., Nairn, I.A.
Bowyer, D., and Holt, R., 2010, Case Study: Development of a .Eds. , The 197682 Eruption Sequence at White Island Vol-
Numerical Model by a Multi-Disciplinary Approach, Rotokawa .cano Whakaari , Bay of Plenty, New Zealand. N. Z. Geol.
Geothermal Field. Proc. World Geothermal Congress 2010 Bali, Surv. Bull. 103, pp. 127137, Rotorua
Indonesia, 25-29 April.
Mrner, N.-A., Etiope, G., 2002. Carbon degassing from the
Brombach, T., Hunziker, J.C., Chiodini, G., Cardellini, C., Marini, lithosphere. Global Planet. Change 33, 185203.
L., 2001. Soil diffuse degassing and thermal energy uxes from the
southern Lakki Plain, Nisyros (Greece). Geophys. Res. Lett. 28, Rissmann, C.F., 2010. Using Surface Methods to Understand the
6972. Ohaaki Field, Taupo Volcanic Zone, New Zealand. Unpubl.
Doctoral Dissertation. Univ. Canterbury, Christchurch, New
Cardellini, C., Chiodini, G., Frondini, F., 2003. Application of Zealand.
stochastic simulation to CO2 ux from soil; mapping and
quantication of gas release. J. Geophys. Res.108 (B9), 13 Rissmann, C., Nicol, A., Cole, J., Kennedy, B., Fairley, J.,
Christenson, B., Leybourne, M., Milicich, S., Ring, U., Gravley,
Chiodini, G., Cioni, R., Guidi, M., Raco, B., Marini, L., 1998. Soil D., 2011. Fluid ow associated with silicic lava domes and faults,
CO 2 ux measurements in volcanic and geothermal areas. Appl. Ohaaki hydrothermal eld, New Zealand. J. Volcanol. Geotherm.
Geochem. 13, 543552 Res. 204, 1226.
.
Chiodini, G., Avino, R., Brombach, T., Caliro, S., Cardellini, C., Rissman, C., Christenson, B., Werner, C., Leybourne, M., Cole, J.,
De Vita, S., Frondini, F.,Granirei, D., Marotta, E., Ventura, G., Gravley, D., 2012, Surface heat ow and CO 2 emissions within
2004. Fumarolic and diffuse soil degassing west of Mount the Ohaaki hydrothermal eld, Taupo Volcanic Zone, New
Epomeo, Ischia, Italy. J. Volcanol. Geotherm. Res. 133, 291309. Zealand, Appl. Geochemistry 27, 223-239.
Chiodini, G., Granieri, D., Avino, R., Caliro, S., Costa, A., Werner, Schmidt, E., Grigull, U., 1979. Properties of Water and Steam in
C., 2005. Carbon dioxide diffuse degassing and estimation of heat SI-units: 0800 C, 01000 bar. Springer-Verlag, Berlin Hei-
release from volcanic and hydrothermal systems. J. Geophys. Res. delberg, R. Oldebourg, Mnchen
110 (B8), 17.

Collar, R.J. and Browne, P.R.L., 1985, Hydrothermal eruptions at Sinclair, A.J., (1974): Selection of thresholds in geochemical data
the Rotokawa Geothermal Field, Taupo Volcanic Zone, New using probability graphs. J. Geochem. Explor. 3, 129149.
Zealand. Proc. 7th NZ Geothermal Workshop. . 1-5.
Wardell, L.J., Kyle, P.R., Dunbar, N., Christenson, B., 2001. White
Dawson, G.B., 1964. The nature and assessment of heat ow from Island volcano, New Zealand: carbon dioxide and sulfur dioxide
hydrothermal areas. N. Z. J. Geol. Geophys. 7, 155171 emission rates and melt inclusion studies. Chem. Geol. 177, 187
200.
Deutsch, C.V., Journel, A.G., 1998. GSLIB: Geostatistical
Software Library and Users Guide. Applied Geostatistics Series. Welles, J.M., Demetriades-Shah, T.H. and McDermitt, D.K., 2001.
Oxford University Press, New York, Oxford. Considerations for measuring
ground CO2 effluxes with chambers. Chemical Geology, 177(1-2):
Fridriksson, T., Kristjansson, B.R., Armannsson, H., 3-13.
Margretardottir, E., Olafsdottir, S., Chiodini, G., 2006. CO 2
emissions and heat ow through soil, fumaroles, and steam-heated Werner, C., Cardellini, C., 2006. Comparison of carbon dioxide
mud pools at the Reykjanes geothermal area, SW Iceland. Appl. emissions with uid upow, chemistry, and geologic structures at
Geochem. 21, 15511569. the Rotorua geothermal system, New Zealand. Geothermics 35,
221238. Werner, C., Brantley, S.L., Boomer, K., 2000. CO 2
Giammanco, S., Parello, F., Gambardella, B., Schifano, R., emissions related to the Yellowstone volcanic system. 2. Statistical
Pizzullo, S., Galante, G., 2007. Focused and diffuse efuxes of CO sampling, total degassing, and transport mechanisms. J. Geophys.
2 from mud volcanoes and mofette south of Mt. Etna (Italy). J. Res. 105, 10,83110,846.
Volcanol. Geotherm. Res. 165, 4663.
Werner, C., Hochstein, M.P., Bromley, C.J., Manville, V.R.,
Giggenbach, W., Matsuo, S., 1991. Evaluation of results from Tilyard, D., 2004. CO 2-ux of steaming ground at Karapiti
Second and Third IAVCEI field workshops on volcanic gases, (Wairakei, NZ). Geol. Soc. N. Z. 117A, 115116
Mt. Usu, Japan and White Island, New Zealand. Appl.
APPENDIX

1: Cumulative Probability Plots

Individual populations are broken down in terms of mean (M i), proportion (fi) and standard deviation (i).

White Island

Rotokawa

2: Variogram Models
Red dotted line is the declustered, normal scores variogram for CO 2 Flux.
Rotokawa White Island