Geothermics 55 (2015) 150–158

Contents lists available at ScienceDirect

Geothermics
journal homepage: www.elsevier.com/locate/geothermics

A reformulation of USGS volumetric “heat in place” resource

estimation method
Sabodh K. Garg a,∗ , Jim Combs b
a
Leidos Inc., 10260 Campus Point Drive, San Diego, CA 92121, USA
b
Geo Hills Associates LLC, 12945 Vanderbilt Dr. #207, Naples, FL 34110, USA

a r t i c l e i n f o a b s t r a c t

Article history: A reformulation of the USGS volumetric “heat in place” method is presented. More specifically, expres-
Received 4 December 2014 sions for “recoverable heat” are derived by considering specific power cycles, i.e., single-flash and binary.
Accepted 10 February 2015 The latter approach eliminates the ambiguities associated with specifications of a reference temperature
Available online 5 March 2015
and the utilization efficiency. Since the use of an arbitrarily low reference temperature such as the ambi-
ent temperature yields too optimistic an estimate for the recoverable heat, the abandonment temperature
Keywords:
instead of the ambient (or condenser) temperature should be assumed as the reference temperature in
Heat in place
order to obtain realistic estimates of recoverable heat. The standard USGS method will, however, yield a
Thermal recovery factor
Utilization efficiency
non-zero available work for temperatures between the ambient (or condenser) and abandonment tem-
Reference temperature peratures; in this case, the only way to conform the USGS method to reality is to require conversion
Monte Carlo (utilization) efficiency to also tend to zero. The probable range for the thermal recovery factor is con-
Single flash sidered, and it is argued that in the early exploration stage prior to deep drilling and well testing, the
Binary cycle proper lower limit for the recovery factor is zero. An illustrative example of the standard USGS method
Sumikawa vs. the present reformulation of the USGS method for resource estimation is presented using previously
published data from the Sumikawa geothermal field. A Monte Carlo procedure was employed to predict
the megawatt capacity at Sumikawa for three cases: (1) the reformulated USGS method, (2) the effect
of a zero minimum recovery factor, and (3) the standard USGS method. A total of 100,000 Monte Carlo
simulations were used in each case to compute the cumulative probability distribution. Comparing Cases
1 and 3, it is apparent that the original USGS method predicts a much larger capacity than the new for-
mulation. The main effect of a zero minimum recovery factor (Case 2) is to yield a considerably smaller
capacity at the 90% confidence level compared to Case 1. In a certain sense, a zero minimum recovery
factor encodes a lack of knowledge regarding the productivity of as yet undrilled wells.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction of combining estimates with uncertainties for the temperature,

During early stage exploration of geothermal resources associ-
ated with an identified hydrothermal convection system, it is often
desired to obtain an estimate of the potential electrical energy that
might be produced from the geothermal system in order to justify
the anticipated cost for development of the geothermal resource.
In the 1970s, researchers at the United States Geological Survey
(USGS) developed a methodology to quantify the uncertainty of
estimates of the geothermal resources associated with an identi-
fied hydrothermal convection system (e.g., Nathenson, 1975a,b;
Nathenson and Muffler, 1975; Muffler and Cataldi, 1977; Brook
et al., 1979). The USGS volumetric estimation methodology consists
∗ Corresponding author. Tel.: +1 858 826 1615; fax: +1 858 826 6044.
E-mail address: gargs@leidos.com (S.K. Garg).
area, and thickness of a geothermal reservoir into a prediction of the
stored energy (“heat in place”) with uncertainty. Probability density
functions for temperature, area, thickness, and thermal recovery
factor are assumed based on uncertain estimates in order to cal-
culate the probability distribution function for the “recoverable
heat”. The probability distribution function for the “recoverable
heat” can be obtained using a Monte Carlo approximation. Thus,
the USGS volumetric estimation method together with Monte Carlo
simulations is often used to provide estimates of the probable gen-
eration capacity of a geothermal resource. Taken at face value,
the method is deceptively simple. In two previous papers, the
authors focused on the difficulties of specifying a probability den-
sity function for thermal recovery factor (Garg and Combs, 2010)
and utilization efficiency (Garg and Combs, 2011) required to com-
pute the probable generation capacity. More recently, Williams
(2014) has re-examined the probable range for the thermal

http://dx.doi.org/10.1016/j.geothermics.2015.02.004
0375-6505/© 2015 Elsevier Ltd. All rights reserved.
 S.K. Garg, J. Combs / Geothermics 55 (2015) 150–158
recovery factor, and opined that the thermal recovery factor must
be greater than zero. In conformity with Garg and Combs (2010),
the authors maintain that a zero value for the thermal recovery
factor cannot be a priori excluded.
In this paper, a reformulation of the USGS volumetric “heat
in place” method is presented. More specifically, expressions for
“recoverable heat” are derived by considering specific power cycles,
e.g., single-flash and binary. The latter approach eliminates the
ambiguities associated with specifications of a reference temper-
ature and the utilization efficiency. The probable range for the
thermal recovery factor is considered, and it is argued that in the
early exploration stage prior to deep drilling and well testing, the
proper lower limit for the recovery factor is zero. Finally, the new
formulation along with Monte Carlo method is applied to assess
the power potential of an actual geothermal field, and results are
compared to those derived using the conventional USGS method.
2. Recoverable heat
Consider a geothermal reservoir with volume V. Heat stored in
the geothermal reservoir, qR , is given by:
qR = V c (TR − Tr )
where cl , cg , cr = heat capacity of liquid water, steam, and rock
grains; A = reservoir area; H = reservoir thickness; Sl , Sg = liquid
water and steam saturation in pores (Sl + Sg = 1); Tr = reference tem-
perature; TR = average reservoir temperature; V = reservoir volume
(=AH);  = porosity; c = volumetric heatcapacity of fluid-saturated
rock = Sl l cl + Sg g cg + (1 − ) r cr ; l , g , r = density of liq-
uid water, steam, and rock grains.
The choice of the reference temperature is very important since
it has a large effect on the computed value for the “recoverable
heat”. In previous applications of the USGS method, two choices
for Tr , i.e., the ambient (say ∼15 ◦ C) or the condenser (∼40 ◦ C) tem-
peratures, have been used (see e.g., Brook et al., 1979; Williams,
2014). While any value for reference temperature, including abso-
lute zero, may be employed to compute the “stored heat” in the
reservoir, the choice of a reference temperature is restricted for
calculating the “recoverable heat”. For the latter purpose, it is use-
ful to introduce the concept of an “abandonment temperature”,
which is defined as the temperature below which a geothermal
reservoir will not be produced. Values for “abandonment temper-
ature” will depend upon the power cycle. Thus, for a flash-type
power plant, the lower limit for the “abandonment temperature”
is given by the saturation temperature Tsep corresponding to the
separator pressure, and for a binary power plant it equals the so-
called pinch point temperature Tp (see Fig. 1). The pinch point is
the location in the heat exchanger with the least temperature dif-
ference T between the primary (water) and secondary fluids, and
corresponds to the bubble point Tb for the secondary fluid. Thus,
the pinch point temperature Tp is given by:
Tp = Tb + T
A geothermal thermal recovery factor Rg is defined as the ratio
of the heat recovered at the wellhead, qw , to the heat stored in the
reservoir, qR .
Rg = qw /qR
Combining Eqs. (1) and (3), the heat recovered at the wellhead
qw can be expressed by the following expression:
qw = ˛(TR − Tr )
where
˛ = Rg V c
151
Brine
Temperature
Pinch Point
Secondary Fluid
Pre-heater Evaporator
0% 100%
Heat Transfer
Fig. 1. Heat transfer in a simple binary cycle.
(1) A simple example will suffice to illustrate the importance of
specifying a reasonable value for the reference temperature Tr .
Consider a typical high temperature geothermal resource with an
average temperature of 250 ◦ C. It is assumed that a single flash
cycle with a separator pressure of 5 bars (saturation tempera-
ture = 151.831 ◦ C) is used for generating power. In this case, the
“abandonment temperature” equals 151.831 ◦ C. With a reference
temperature of 15 ◦ C, a value commonly used in previous applica-
tions of the USGS method, the heat recovered in Joules (J) at the
wellhead is:
qwUSGS = 235˛ J
Since the wells will not be discharged below the “abandonment
temperature”, a more reasonable value for the reference temper-
ature is 151.831 ◦ C, and the heat recovered at the wellhead in the
actual case is:
qwActual = 98.169˛ J
The two values for the heat recovered at the wellhead differ
by a factor of more than two. Use of an arbitrarily low reference
temperature such as the ambient temperature yields too opti-
mistic an estimate for the recovered heat. It is suggested that the
“abandonment temperature” instead of the ambient (or condenser)
temperature should be assumed as the reference temperature in
order to obtain realistic estimates of recovered heat.
3. Available work and conversion efficiency
(2) In a geothermal reservoir most of the heat is contained in the
reservoir rocks and not in the reservoir fluid. The fluid is needed
to mine the heat from the rocks. Injection of cooled brine after its
use in the power plant and/or influx of cold water from reservoir
boundaries is required to mine heat from the reservoir. The tem-
perature of the discharged fluid gradually declines from its initial
(3)
value TR to the “abandonment temperature”. Average temperature
of the produced fluid T̄ is in general different from TR .
Assuming isenthalpic flow in the wellbore and neglecting the
work required to raise water to the wellhead, the enthalpy of pro-
(4) duced fluid at the wellhead, hw , is equal to that of liquid water at T̄ .
Thus
 
(5) hw = hw T̄ (6)
152 S.K. Garg, J. Combs / Geothermics 55 (2015) 150–158

The amount of fluid produced at the wellhead, mw , is given by: The separated brine is injected into the reservoir, and the steam
is used to generate power. The mass of the fluid produced at the
mw = qw / (hw − hr ) (7)
wellhead is given by Eq. (8) with Tr = Tsep . The steam fraction of the
where hr is the enthalpy of liquid water at the reference tempera- produced fluid is:
ture, Tr . Substituting from Eq. (4) into Eq. (7), there follows:       
mstm = mw hw T − hw Tsep /hgl Tsep (12)
mw = ˛ (TR − Tr ) / (hw − hr ) (8)
where hgl (Tsep ) denotes the heat of vaporization. Combining Eqs.
The concept of availability plays an important role in the (8) and (12), there follows:
USGS volumetric “heat in place” method. Availability is defined    
mstm = ˛ TR − Tsep /hgl Tsep (13)
as the maximum work (or power) output that can theoretically
be obtained from a substance (water) at specified thermodynamic Substituting mstm for m in Eq. (10), the available work for the
conditions (wellhead) relative to its surroundings. DiPippo (2008) single-flash case is given by:
observes:  
“To achieve this ideal outcome, there are two thermodynamic ˛ TR − Tsep   
WAflash =   hstm Tsep − hw (Tc )
conditions that must be met: hgl Tsep
   
(1) All processes taking place within the system must be perfectly −TcK sstm Tsep − sw (Tc ) (14)
reversible.
in Eq. (14), Tc denotes the condenser temperature (◦ C) and TcK is
(2) The state of all fluids being discharged from the system must
the absolute condenser temperature.
be in thermodynamic equilibrium with the surroundings.”

3.2. Simple binary cycle

The first of these conditions amounts to neglecting losses due
to friction, turbulence, and other sources of irreversibility. The sec-
ond condition requires that any fluids discharged from the system
are in temperature equilibrium with the surroundings. None of
the real power cycles can meet these conditions, and the “elec-
trical energy” that is generated is always less than the “available
work”. The conversion efficiency (utilization factor) is the ratio of
the “actual electrical energy” to the “available work”.
Neglecting kinetic or potential energy effects, the maximum
energy output per unit mass of the substance e is given by (DiPippo,
2008):
e = h − hex − TexK (s − sex )
where h and s denote the enthalpy and entropy of the substance
(e.g., steam or hydrocarbon vapor) at turbine inlet conditions with
temperature T, TexK is the absolute exit temperature (Kelvin), and
sex is the entropy of liquid phase (water or hydrocarbon liquid) at
the exit temperature. For mass m of the substance, the available
work is therefore given by:
WA = me = m [h − hex − TexK (s − sex )]
The available work, as used in the conventional USGS method
(Brook et al., 1979), is computed by replacing mass m in Eq. (10)
by mw from Eq. (8), h by hw (TR ), and putting Tex = To . Here To is the
ambient temperature and is often assumed to be 15 ◦ C. Thus
˛ (TR − To )
WAUSGS = [hR − ho − ToK (sR − so )]
(hR − ho )
In general, mass that enters the turbine is not the same as mw
computed from Eq. (8). In single (or dual) flash systems, liquid brine
(and the energy contained in it) is rejected at the separator tem-
perature; only the separated steam is used to generate power. The
binary technology involves the use of a secondary fluid, heated by
the brine, to generate power; the brine is in any event rejected at a
higher temperature than the ambient temperature. For a real power
cycle, m, h and s in Eq. (10) should be evaluated at the turbine inlet
conditions rather than the wellhead, and the most appropriate exit
temperature (Eq. (10)) is the condenser temperature. In the follow-
ing, expressions are developed for available work for single-flash
and binary power cycles.
3.1. Single-flash
It is assumed that the produced fluid with an average tempera-
ture T̄ at the wellhead is separated at a separator temperature Tsep .
The produced brine at an average wellhead temperature T̄ is
used to heat the working fluid (e.g., isobutane), and the spent brine
is injected back into the reservoir. Given the turbine inlet pressure
pin , the corresponding saturation temperature of the secondary
fluid Tb may be determined from thermodynamic data for the fluid
(see e.g., NIST, 2010). The secondary fluid is assumed to be satu-
rated at the turbine inlet. With a temperature differential T at
the pinch point, the temperature of the brine at the pinch point Tp
is given by Eq. (2). The heat provided by hot brine to vaporize the
secondary fluid (from incipient bubbling at Tb ) is given by:
(9)       
Q = mw hw T − hw Tp = ˛ TR − Tp (15)
The mass of secondary fluid msf vaporized by heat Q is:
 
msf = ˛ TR − Tp /hsfgl (Tb ) (16)
where hsfgl (Tb ) denotes the heat of vaporization for the secondary
fluid at temperature Tb .
Substituting from Eq. (16) into Eq. (10), the available work for
(10) the binary cycle becomes:
 
˛ TR − Tp   
WAbinary = hsfg (Tb ) − hsfl (Tc ) − TcK ssfg (Tb ) − ssfl (Tc )
hsfgl (Tb )

−Vsf (Tc , psfb ) [pin − psfb (Tc )] (17)
(11)
The last term in Eq. (17) denotes the work required to raise
the pressure of the secondary fluid (liquid) from the condenser
pressure (=saturation pressure at condenser temperature) to the
turbine inlet pressure, Vsf is the specific volume, and subscripts g
and l refer to gas and liquid.
3.3. Illustrative examples
We consider two typical geothermal resources with temper-
atures of 150 ◦ C and 250 ◦ C. The lower temperature resource is
suitable for binary applications while the higher temperature
resource will most likely be exploited using flash technology. Before
embarking on an analysis of power cycles, it is useful to evaluate
available work using the USGS methodology (Eq. (11)). Available
work computed from Eq. (11) for the two resource temperatures
(150 ◦ C and 250 ◦ C) and a reference temperature of 15 ◦ C is pre-
sented in Table 1. All the water properties were evaluated along
the saturation line.
 S.K. Garg, J. Combs / Geothermics 55 (2015) 150–158
Table 1
Comparison of available work computed with USGS method (Eq. (11)) with that
obtained using power cycle Eqs. (14) and (17). For the sake of simplicity WA /˛ instead
of WA is given. See text for additional details.
Resource temperature TR (◦ C) Available work WA /˛ (◦ C)
USGS method
150 24.53
250 65.00
For the geothermal resource with an average temperature of
250 ◦ C, it is assumed that the produced fluid is separated at 5
bars (saturation temperature: 151.831 ◦ C). The separated brine is
injected into the reservoir, and the steam is used to generate power.
The turbine inlet pressure is set equal to the separator pressure, and
the condenser temperature is assumed to be 40 ◦ C. Evaluating the
water properties along the saturation line, Eq. (14) yields:
WAflash /˛ = 29.10 o C
The available work given by Eq. (14) is thus only about 45% of
that obtained using Eq. (11).
The produced brine from the geothermal reservoir with an aver-
age temperature of 150 ◦ C is used to heat a secondary working fluid
(isobutane), and the spent brine is injected back into the reservoir.
The turbine inlet pressure is assumed to be 20 bars; the satura-
tion temperature for isobutane at the latter pressure is 100.36 ◦ C
(NIST, 2010). Isobutane is assumed to be saturated at the turbine
inlet. With a temperature differential of 5 ◦ C at the pinch point,
the temperature of the brine at the pinch point is 105.36 ◦ C. Eval-
uating the isobutane properties using NIST tables (NIST, 2010) and
substituting in Eq. (17), there follows:
WAbinary /˛ = 10.47 o C
The available work obtained from Eq. (17) is about 43% of that
computed using Eq. (11).
3.4. Conversion (utilization) efficiency
Modern turbo-generators are quite efficient; the dry turbine
(turbines operating with dry steam or gas) efficiency is about 85%
(DiPippo, 2008; Zarrouk and Moon, 2014). Geothermal turbines
generally operate in the wet region. Presence of moisture lowers
the efficiency; thus a one percent (1%) moisture will lower the effi-
ciency by about 1% (DiPippo, 2008; Zarrouk and Moon, 2014). For
flash power plants, the efficiency is further reduced by the pres-
ence of non-condensable gases. Including losses in the generator,
it is suggested that a conservative value for conversion efficiency
of 70% to 80% be used together with Eqs. (14) and (17) to estimate
the electrical generating capacity of a geothermal reservoir.
The illustrative examples presented in Section 3.3 show that
the available work calculated using the standard USGS method
is much higher than that obtained considering the actual power
cycles. Thus, if Eq. (11) is used to compute the available work, the
conversion efficiency will need to be reduced from 70% to 80% to
no more than about 40%.
As an example, consider a geothermal turbo-generator supplied
with 5-bar steam and a condenser temperature of 40 ◦ C; in this
case, the available power per kg/s of steam from Eq. (14) is about
0.625 MW. With a 70% conversion efficiency, about 2.3 kg/s will be
required to generate 1 MWe; this value for steam consumption is
close to that used by modern geothermal power plants. Comparing
available work computed from Eqs. (11) and (14), the appropriate
conversion efficiency for use with the standard USGS method would
be about 31.3%.
DiPippo (2004) has analyzed the conversion efficiency for sev-
eral operating ORC power plants. Although second law efficiencies
153
of over 40% (referred to ambient temperature) have been reported
for a couple of power plants utilizing advanced power cycles, most
operating ORC power plants have relatively low (less than 25%) sec-
ond law efficiencies. GeothermEx (2004) used a utilization factor
of 0.45 (referred to 15 ◦ C) in a resource assessment of several low
Power cycle temperatures geothermal fields in Nevada and California. Although
10.47 the latter utilization factor appears to be theoretically possible,
29.10 resource characteristics (e.g., change in resource temperature over
time) and economic considerations will usually dictate a much
lower utilization factor.
It is apparent from Eqs. (14) and (17) that as the difference
between the average reservoir temperature and the separator (or
pinch point temperature) temperature tends to zero, available work
for conversion to electricity will also approach zero. The standard
USGS method (Eq. (11)) will, however, yield a non-zero available
work; in this case, the only way to conform USGS method to reality
is to require conversion (utilization) efficiency to also tend to zero.
Stated somewhat differently, the conversion (utilization) efficiency
used in concert with Eq. (11) must be assumed to be a function of
the reservoir temperature and the “abandonment temperature”. To
avoid ambiguities with a temperature-dependent conversion (uti-
lization) efficiency, it is recommended that the available work be
computed by considering the power cycle (e.g., Eqs. (14) and (17)).
4. Thermal recovery factor
The thermal recovery factor Rg was introduced in Section 2 as
the ratio of the heat recovered at the wellhead, qw , to the heat
stored in the reservoir, qR . The latter parameter depends on the per-
meability structure (fracture vs. matrix, permeability anisotropy,
faulting, etc.), production and injection well depths and patterns,
and the thermal (heat and fluid recharge from depth) and hydraulic
boundary conditions (recharge/discharge along the ground surface,
and assumed lateral boundaries for the geothermal reservoir). Since
many of the reservoir properties that affect the thermal recovery
factor are likely to be poorly known until the reservoir has been pro-
duced for several years, specification of the thermal recovery factor
for a specific reservoir is more often than not a matter of conjec-
ture. A major goal of reservoir engineering- including well drilling
and testing, data collection and synthesis, and detailed reservoir
modeling—is to obtain reliable estimates of the thermal recovery
factor for a particular reservoir.
Williams (2014) presents estimates of thermal recovery fac-
tor based on both theoretical grounds and data from operating
hydrothermal fields, and suggests that the appropriate range for
fracture-dominated geothermal reservoirs is from 0.08 to 0.20. For
sediment-hosted reservoirs, the suggested range is slightly higher
from 0.10 to 0.25. According to Williams (2014), if a permeable
reservoir exists, then Rg is non-zero; we agree.
During the exploration phase and prior to geothermal well
drilling and testing, it will not in general be possible to infer the
presence of a permeable reservoir. Indeed, many hydrothermal
exploration projects (e.g., Valles Caldera, Rye Patch, Lee Hot Springs,
Dixie Meadows, Newberry, Honey Lake, Mangakino) have been
abandoned due to lack of sufficient permeability (see also Grant,
2015). Ultra low formation permeability usually results in unac-
ceptably large pressure drop and uneconomic discharge rates. Thus,
the possibility of the thermal recovery factor being zero cannot
be discounted during the exploration phase. Therefore, we dis-
agree with Williams (2014) that a non-zero Rg is appropriate for
“resource” evaluation. In accordance with our earlier work (Garg
and Combs, 2010), it is suggested that the proper range for ther-
mal recovery factor is from 0 to 0.20 (the latter value is believed to
be the maximum credible value based on world-wide experience
with production from liquid-dominated reservoirs). Unfortunately,
154 S.K. Garg, J. Combs / Geothermics 55 (2015) 150–158
in most instances prior to geothermal well drilling and testing,
the minimum thermal recovery factor is assumed to be non-zero;
this more often than not leads to overly optimistic projections of
electrical generating capacity.
5. Electric generating capacity
The parameters required for the computation of electric capac-
ity with the “heat in place” method are provided in Table 2; these
parameters can be divided into two groups. The second of these
groups (Group 2 Parameters) contains parameters whose value does
not either vary substantially from case to case (volumetric heat
capacity, power plant or project life, power plant load factor) or
can be specified sufficiently accurately using available engineer-
ing data (reference temperature and conversion efficiency; see also
Sections 2 and 3). The situation is completely different as far as the
first group (Group 1 Parameters) of parameters is concerned, and
estimates for these parameters are highly dependent on the specific
phase of development of the geothermal reservoir.
Specification of statistical distributions for the parameters in
the first group (reservoir area, reservoir depth, reservoir thickness,
reservoir temperature, thermal recovery factor) is at best a difficult
task and as demonstrated below is highly dependent on the stage of
development of a geothermal system. Reservoir area, thickness, and
temperature are required to compute the “heat in place”. Reservoir
depth, usually ignored in Monte Carlo simulations, is important in
that it determines the depth to which the wells must be drilled in
order to access the geothermal resource. Thermal recovery factor
is needed to compute the fraction of “heat in place” that may be
recovered using a system of production and injection wells.
5.1. Exploration phase
Prior to deep well drilling and testing, data that may be used to
estimate reservoir parameters include shallow temperature gradi-
ent surveys (temperature data from heat flow holes usually less
than 100–200 m in depth), chemical geothermometer values for
fluid samples from any hot springs/fumaroles, and surface alter-
ation (e.g., sinter, carbonate) surveys. Surface alteration, if present,
can provide an indication of probable reservoir temperatures; as
an example, presence of sinter indicates fluid temperatures in
excess of about 175 ◦ C. Chemical geothermometer data, if available,
provide a reasonable first estimate of the reservoir temperature.
Extrapolation of the high shallow temperature gradient data to
great depth should be done with considerable care since the rela-
tively high permeability associated with a geothermal reservoir will
tend to produce a near isothermal (and low temperature gradient)
zone. Temperatures at depth, obtained by extrapolating shallow
gradient data and not supported by any other evidence, should be
regarded as an estimate of maximum reservoir temperature; a first
Table 2
Parameters required for the calculation of the electric generation capacity using the
USGS volumetric “heat in place” method.
Group 1 parameters
Reservoir area (km2 )
Reservoir depth (m)
Reservoir thickness (m)
Reservoir temperature (◦ C)
Thermal recovery Factor (%)
Group 2 parameters
Volumetric heat capacity (kJ/m3 -K)
Reference temperature (◦ C)
Conversion efficiency (%)
Plant or project life (years)
Plant load factor (%)
estimate of the minimum temperature may be computed using the
average regional temperature gradient.
During the exploration phase, chemical geothermometers
provide the best estimate of the possible range for reservoir tem-
peratures. Shallow temperature gradient data, in conjunction with
chemical geothermometers, may then be used to estimate the
depth range for the geothermal reservoir. A minimum estimate of
possible reservoir area may be obtained from the distribution of
hot springs/fumaroles and high temperature gradient areas. Geo-
physical data (e.g., resistivity surveys) or geological mapping (e.g.,
surface alteration, faulting, etc.) may be useful for estimating the
probable maximum geothermal reservoir area. As discussed in Sec-
tion 4, the appropriate range for the thermal recovery factor during
the exploration phase is from 0 to 0.20. The assumed value for the
minimum thermal recovery factor has a major influence on the pre-
dicted electrical power capacity at the 90% confidence level (Garg
and Combs, 2010). On the other hand, the power capacity at the 10%
confidence level remains more or less unaffected by the changes
in the minimum values for the recovery factor and the reservoir
thickness.
A geothermal well does not penetrate a region of uniform per-
meability. Very often a geothermal well is completed with a large
open-hole section (100 to 1000 or more meters); actual production
usually comes from one or more feed zones with a thickness of the
order of 1 to 100 m. In any event, during the exploration phase, the
reservoir thickness should be assumed to vary within a rather wide
range (say between 100 and 2000 m).
A triangular probability distribution is assumed for parameters
for which maximum, mode, and minimum values can be estimated.
If only minimum and maximum values are available, then a rect-
angular (i.e., uniform) probability distribution is used. It should
be pointed out that any particular choice or prescription of the
distribution of parameter probability over the postulated range
(minimum to maximum) will also impact the cumulative results,
and that this issue is not well-studied. Presumably, different dis-
tributions (uniform vs. triangular vs. Gaussian vs. log-normal, etc.)
will give different answers; thus, providing additional uncertainty
in the estimation process.
Adoption of the above procedure for specifying parameter
values during the exploration phase is liable to yield a wide dis-
tribution for estimated electrical capacity with a relatively large
ratio between the capacity at 10% confidence level and that at
90% confidence level. The estimated capacity at 10% level would
almost certainly be higher than what will ultimately prove to be
the case. Similarly, the estimated capacity at the 90% confidence
level will go up as we learn more about the geothermal reser-
voir. The major goal of the resource estimation at this stage should
be to determine whether an exploitable geothermal resource
exists and if it may be economically worthwhile to undertake a
program of geothermal well drilling and testing to reduce the
uncertainty in the geothermal reservoir parameters and thus
obtain a more reliable estimate of the probable resource megawatt
capacity.
5.2. Well drilling and testing phase
Drilling and testing of geothermal wells is essential for obtain-
ing reliable estimates of important reservoir parameters. Downhole
temperature and pressure surveys in deep wells can be used to
establish (1) depth to the top of the geothermal reservoir (defined
here as the transition from the linear conductive thermal gradi-
ent interval in the well to either a near isothermal or low thermal
gradient interval), (2) formation temperature, and (3) formation
pressure. Fluid samples from discharging geothermal wells may be
used to obtain chemical geothermometers temperatures; gener-
ally speaking, different chemical geothermometers yield a range
 S.K. Garg, J. Combs / Geothermics 55 (2015) 150–158 155

of temperatures. Pressure transient tests (and in particular pres-

sure interference tests) and tracer tests can be used to establish
reservoir continuity in the region penetrated by the wells, and to
obtain the probable area of the permeable zone. Drilling records
(“mud losses”) are also helpful in determining depths of permeable
horizons. At the end of the initial well drilling and testing phase, it
should be possible to define the following parameters with a high
degree of confidence:

(1) Reservoir depth: Temperature profiles in wells (i.e., transition

from conductive to convective profiles) should be used to esti-
mate the minimum and maximum depths to the top of the
permeable zone.
(2) Reservoir temperature: Measured temperatures in wells
together with chemical geothermometry may be employed to
define the range of probable temperatures for the “hot region”
of interest in the geothermal system.
(3) Reservoir area: Since it is unlikely that the entire reservoir
area with elevated temperatures has been penetrated by the
geothermal wells, it is almost certain that the minimum reser-
voir area is greater than that investigated by drilling. Estimates
of maximum reservoir area may be obtained from pressure
transient data, geophysical surveys (e.g., electrical resistivity),
and geological mapping (e.g., alteration surveys).
(4) Reservoir thickness: Estimates of minimum and maximum
geothermal reservoir thickness may be obtained by examining
the convective zone from downhole temperature surveys.

Provided geothermal well drilling and testing has shown ade-

quate well productivity, it is justified at this stage to assume a Fig. 2. Sumikawa–Ohnuma area. Cross-sections AA and BB are shown in
non-zero minimum value (say 0.05) for the thermal recovery factor. Figs. 3 and 4, respectively.
At the end of geothermal well testing, it should thus be possible
to place much closer limits on reservoir parameters (i.e., Group 1
and Sarmiento (2005) postulated that while numerical simulation
Parameters) than those in the exploration phase. It need hardly be
is more sophisticated than the volumetric method, the latter can
stressed that even after initial geothermal well drilling and testing,
be readily conducted in a rigorously probabilistic way while the
it will only be possible to narrow (but not eliminate) the uncer-
former cannot; in principle, however, there is no reason why a
tainty in reservoir properties. Use of these “parameter ranges” will
Monte Carlo process cannot be used in conjunction with “detailed
result in a considerably narrower distribution of probable electric
numerical reservoir model” to assess the impact of uncertainty in
megawatt capacity than that obtained in the exploration phase. A
reservoir parameters on the probable electrical megawatt capacity.
measure of the success of the geothermal well drilling and testing
Availability of production and injection history as well as
would be a reduction in the ratio of the estimated electric megawatt
detailed reservoir modeling are essential for narrowing the possible
capacity at the 10% confidence level to that at the 90% confidence
ranges for important reservoir parameters (i.e., Group 1 Parameters),
level.
and in turn for obtaining a well-constrained estimate of the future
electrical megawatt capacity of the geothermal reservoir.
5.3. Production and injection phase

Following a positive outcome from the initial geothermal well
drilling and testing program, the production and injection well-
field must be developed. The knowledge of reservoir parameters
increases with each new geothermal well that is drilled and tested.
However, a production/injection history (and consequent changes
in reservoir temperature, pressure, and fluid state) is essential
for understanding important geothermal reservoir processes such
as boiling due to a reduction in pressure, recharge from the
boundaries, and possible short-circuiting between production and
injection wells.
At this stage, a detailed geothermal reservoir model may be
constructed by calibrating model predictions with available mea-
surements (e.g., temperature and pressure measurements, surface
heat and mass flows, production/injection induced changes in the
reservoir, etc.). The calibrated model may then be used to fore-
cast the electrical megawatt capacity of the reservoir. The current
practice in the geothermal industry is to develop “determinis-
tic” reservoir models. In large part, this practice is driven by the
“expert” time required to construct a geothermal reservoir model
that is in accord with most of the known facts. Although Sanyal
5.4. An illustrative example: Sumikawa geothermal field, Japan
The Sumikawa geothermal field is located in the Hachimantai
volcanic area, Honshu, Japan. Pritchett et al. (1989) described the
natural fluid circulation system at Sumikawa based on static pres-
sure and temperature logs in several wells drilled in the field, and
a limited amount of pressure transient testing. It is worth noting
that the latter study was published after several deep wells had
been drilled and tested but prior to the start of power production
in 1995.
The Sumikawa geothermal field is located in the western part of
the area shown in Fig. 2; to the east, the Ohnuma geothermal power
station has been generating about 10 MWe since 1974. The area lies
within a north–south graben; the Sumikawa field is located along
the western edge of the graben. Based on drilling logs, the geologic
sequence at Sumikawa consists of the following formations (see
Figs. 3 and 4):
(1) ST formation: Surficial andesitic tuffs, lavas, and pyroclastics of
recent origin.
156 S.K. Garg, J. Combs / Geothermics 55 (2015) 150–158

West East South North

1500 1500
B

A A’ B’
1000
1000 “ST”
“ST”
“LS”

500 “DA”
“DA” HM-3
500
“LS”
“MV” S-3
“MV”

z (m ASL)
“MV” 0
z (mASL)

S-3
“AA”
0 “MV” “AA”
S-2 SB-2
SM-2
–500 S-4

SB-3 KY-1 “AA”

–500 “GR” S-4
“AA”
SA-2
SN-5 SA-4 KY-2 –1000 “GR” SB-1

SA-1
“AA” “GR”
–1000 SB-1
“GR”
–1500

“GR”
Fault?

SC-1
–1500
–2000
Fault?

0 500 1000 1500 2000 2500 3000 3500

y (m North)

–2000
–1000 –500 0 500 1000 1500 2000 Fig. 4. A north–south cross-section across Sumikawa geothermal field.

x (m East)

Fig. 3. An east–west cross-section across Sumikawa geothermal field.

2000 m, and the average temperature is between 260 ◦ C and 320 ◦ C.
The Sumikawa power plant has a nominal capacity of 50 MW. The
initial design for the power plant called for the supply of 5-bar
(2) LS formation: Lake sediments; Pleistocene tuffs, sandstones, silt- steam. Ignoring any pressure drop between the separator and the
stones and mudstones. turbine inlet, the separator temperature (saturation temperature
(3) DA formation: Pliocene dacites, dacitic tuffs, and breccias. corresponding to 5 bar) is 151.831 ◦ C.
(4) MV formation: Interbedded Miocene dacites volcanic rocks and The parameters used to compute the megawatt capacity for the
marine shales and sediments. Sumikawa geothermal field are listed in Table 3. Triangular distri-
(5) AA formation: Altered andesitic rocks. butions are assumed for reservoir area and temperature. Reservoir
(6) BA formation: Crystalline intrusive rocks (mainly granodiorite thickness and thermal recovery factor are represented by rectangu-
and diorite). lar (uniform) distributions. All the other parameters have constant
values. Cases 1 and 2 were evaluated using the new formulation
The lake sediments have low permeability and separate the (Eq. (14)), and the original USGS formulation (Eq. (11)) is used
overlying cold groundwater from the hot geothermal reservoir for Case 3. Since the well productivity had been confirmed prior
below. Also, the MV formation is an aquitard. Geothermal wells to Pritchett et al.’s study, a non-zero minimum recovery factor is
mainly feed from DA, AA, and upper BA formations. justified; thus, the recovery factor is taken to range from 0.05 to
Using the information presented by Pritchett et al. (1989), it is 0.20 for Case 1. To illustrate the effect of a zero minimum recover
reasonable to assume that the Sumikawa geothermal field has a factor on the predicted capacity, Case 2 was also run; other than
minimum area of 3 square kilometers, and a maximum area of 5 the minimum value for the recovery factor, parameters for Case 2
square kilometers. The formation thickness ranges from 1000 m to remain unchanged from those for Case 1. The recovery factor and

Table 3
Parameters for Sumikawa geothermal field. See text for description of the three cases.

Parameter Case 1 Case 2 Case 3

Area (km2 ) Min: 3 Same as Case 1 Same as Case 1

Mode: 4
Max: 5
Thickness (m) Min: 1000 Same as Case 1 Same as Case 1
Max: 2000
Reservoir temperature (◦ C) Min: 260 Same as Case 1 Same as Case 1
Mode: 290
Max: 320
Volumetric heat capacity (kJ/m3 -◦ C) 2700 Same as Case 1 Same as Case 1
Thermal recovery factor Min: 0.05 Min: 0.00 Min: 0.08
Max: 0.20 Max: 0.20 Max: 0.20
Separator temperature (◦ C) 151.831 Same as Case 1 Not applicable
Condenser temperature (◦ C) 40 Same as Case 1 Not applicable
Ambient (rejection) temperature (◦ C) Not applicable Not applicable 15
Conversion efficiency 0.75 Same as Case 1 0.40
Plant life (years) 30 Same as Case 1 Same as Case 1
Plant capacity (Load) Factor 0.95 Same as Case 1 Same as Case 1

Table 4
Megawatt capacity vs. probability for Sumikawa geothermal field.
Minimum MW capacity with a probability > Case 1
90% 33
50% 65
10% 110
10% (MW)/90% (MW) 3.3
Cumulative Probability of Energy Reserves
1
0.9
0.8
0.7
Cumulative Probability
0.6
0.5
0.4
0.3
0.2
0.1
0
0 50 100
Power (MW)
Fig. 5. Cumulative probability of energy reserves for Cases 1–3. See text for descrip-
tion of different cases.
utilization efficiency for Case 3 are the ones recommended by
Williams (2014).
A Monte Carlo procedure was employed to predict the megawatt
capacity for the three cases (Table 4). A total of 100,000 simulations
were used in each case to compute the cumulative probability dis-
tribution; results are illustrated in Fig. 5 and Table 4. Comparing
Cases 1 and 3, it is apparent that the original USGS method predicts
a larger capacity at the 90% confidence level by about 55% than the
new formulation. The main effect of a zero minimum recovery fac-
tor (Case 2) is to yield a much smaller capacity at the 90% confidence
level compared to Case 1. In a certain sense, a zero minimum recov-
ery factor encodes a lack of knowledge regarding the productivity
of as yet undrilled wells.
Published data (Fuchino, 2000; Kawazoe and Shirakura, 2005;
Sugino and Akeno, 2010) indicate that the Sumikawa power plant
generated about 40 MWe between 1998 and 2007; this is about
20% less than the nominal plant capacity. The predicted capacity
at the 90% confidence level for Case 1 (Case 3) at 33 MW (51 MW)
is somewhat smaller (larger) than the actual power production. At
the 50% confidence level, predicted capacity for Case 1 (65 MW) is
much closer to the actual one than that obtained for Case 3 (82 MW).
6. Summary and conclusions
Geothermal reservoir assessment and the resulting prediction
of its electrical megawatt capacity should be regarded as a con-
tinuing process—from the early exploration phase to the time
when the reservoir becomes depleted. During the early exploration
phase, estimates of important reservoir parameters are poorly con-
strained; application of the USGS “heat in place” evaluation method
along with Monte Carlo simulations yields a rather wide distribu-
tion for the probable electrical megawatt capacity. As initial deep
geothermal wells are drilled and tested, it becomes possible to
refine estimates of reservoir parameters (i.e., narrow the range of
possible values), and in turn obtain a much narrower probability
S.K. Garg, J. Combs / Geothermics 55 (2015) 150–158 157
distribution for the electrical megawatt capacity. After the start of
large-scale production and injection operations and the availability
Case 2 Case 3 of data on production-induced changes in the geothermal reservoir,
10 51 it should in principle be possible to further constrain the range of
53 82 probable electrical megawatt capacity.
104 127 The USGS volumetric “heat in place” method together with
10 2.5 Monte Carlo simulations is an important tool for assessing the elec-
trical capacity of a geothermal reservoir. The method should be
modified in several respects to obtain valid results.
case #1 Firstly, available work and conversion efficiency should be eval-
case #2
case #3 uated using parameters for the power cycle (binary or flash) that
may be used for power conversion.
Secondly, specification of the probability distributions of the
reservoir parameters should take into account the stage of the
development of the project. Very often, these parameters are pre-
scribed based on data from other geothermal reservoirs.
Thirdly, at least during the early exploration phases and prior to
deep well drilling and testing, the minimum thermal recovery fac-
tor should be assumed to be zero.
At present, there exist insufficient data in the public domain
to specify probability distributions for most reservoir parameters.
Moreover, conditions vary widely between and within the vari-
ous geothermal provinces around the world. For this reason, it is
150 essential that as far as possible, actual field data should be used
when prescribing reservoir parameters. Without data-driven reser-
voir parameters, use of Monte Carlo simulations is liable to generate
only unreliable estimates of reservoir megawatt capacity.
Acknowledgments
We thank Malcolm Grant and Joel Renner for their careful read-
ing of the original manuscript and suggesting improvements.
References
Brook, C.A., Mariner, R.H., Mabey, D.R., Swanson, J.R., Guffanti, M., Muffler, L.J.P., 1979.
Hydrothermal convection systems with reservoir temperatures ≥90 ◦ C. In: Muf-
fler, L.J.P. (Ed.), Assessment of Geothermal Resources of the United States—1978,
790. U.S Geological Survey Circular, 170p.
DiPippo, R., 2004. Second Law assessment of binary plants generating power from
low-temperature geothermal fluids. Geothermics 33, 565–586.
DiPippo, R., 2008. Geothermal Power Plants: Principles, Applications, Case Studies
and Environmental Impact, second ed. Elsevier, Amsterdam, 493p.
Fuchino, H., 2000. Status of geothermal power generation in Japan. In: Proceedings
World Geothermal Congress 2000. Tohoku and Kyushu, Japan, 6p.
Garg, S.K., Combs, J., 2010. Appropriate use of USGS volumetric “Heat in Place”
method and Monte Carlo Calculations. In: Proceedings 34th Workshop on
Geothermal Reservoir Engineering. Stanford University, Stanford, CA, 7p.
Garg, S.K., Combs, J.,2011. A Reexamination of USGS volumetric “Heat in Place”
method. In: Proceedings 35th Workshop on Geothermal Reservoir Engineering.
Stanford University, Stanford, CA, 5p.
GeothermEx, Inc., 2004. New Geothermal Site Identification and Qualifica-
tion, Consultant Report for the California Energy Commission PIER Report
No. P500-04-051, April 2004, located at http://www.geothermex.com/CEC-
PIER Reports.htm.
Grant, M., 2015. Resource assessment, a review, with reference to the Australian
code. In: Proceedings World Geothermal Congress 2015, Melbourne, Australia,
6p.
Kawazoe, S., Shirakura, N., 2005. Geothermal power generation and direct use in
Japan. In: Proceedings World Geothermal Congress 2005, Antalya, Turkey, 7p.
Muffler, P., Cataldi, R., 1977. Methods for regional assessment of geothermal
resources. U.S. Geological Survey Open-File Report 77-870, 78p.
Nathenson, M., 1975a. Some reservoir engineering calculations for the vapor-
dominated System at Larderello, Italy. U.S. Geological Survey Open-File Report
75-142, 47p.
Nathenson, M., 1975b. Physical factors determining the fraction of stored energy
recoverable from hydrothermal convection systems and conduction-dominated
areas. U.S. Geological Survey Open-File Report 75-525, 38p.
Nathenson, M., Muffler, L.J.P., 1975. Geothermal resources in hydrothermal convec-
tion systems and conduction-dominated areas: In: White, D.E., Williams, D.L.
(Eds.), Assessment of Geothermal Resources of the United States—1975, 726.
U.S Geological Survey Circular, 160p.
158 S.K. Garg, J. Combs / Geothermics 55 (2015) 150–158

NIST, 2010. Thermophysical Properties of Fluid Systems, Available at
http://webbook.nist.gov/chemistry/fluid/.
Pritchett, J.W., Garg, S.K., Maki, H., Kubota, K., 1989. Hydrology of the Sumikawa
geothermal prospect, Japan. Energy Sources 11, 251–262.
Sanyal, S.K., Sarmiento, Z.F., 2005. Booking geothermal energy reserves. Geotherm.
Resour. Counc. Trans. 29, 467–474.
Sugino, H., Akeno, T., 2010. 2010 Country update for Japan. In: Proceedings World
Geothermal Congress 2010, Bali, Indonesia, 7p.
Williams, C., 2014. Evaluating the volume method in the assessment of identified
geothermal resources. Geotherm. Resour. Counc. Trans. 38, 967–974.
Zarrouk, S., Moon, H., 2014. Efficiency of geothermal power plants: a worldwide
review. Geothermics 51, 142–153.