Content uploaded by John Houston
Author content
All content in this area was uploaded by John Houston on Feb 15, 2016
Content may be subject to copyright.
0361-0128/11/3988/1225-15 1225
Introduction
R
ECENT DEVELOPMENTS IN the car industry have led to a
boom in lithium exploration and development for the new
generation of batteries that will power the cars of tomorrow.
In particular, one of the cheapest sources of lithium is the
brines that contain the metal in solution. Such brines also
contain other elements of commercial interest, most notably
potassium and boron (Ericksen and Salas, 1989). As a result,
there has been a plethora of new exploration projects focused
on the brines hosted in the aquifers of the intermontane
basins of the Central Andes.
Most of the exploration is being carried out by junior min-
ing companies with experience in metalliferous deposits that
are attempting to comply with existing European, Canadian,
or Australian stock market commission standards for report-
ing mineral resources and reserves. However, there is a fun-
damental problem for the companies attempting to comply
with the existing standards: these standards apply to solid
phase mineral resources, not to fluids. Because the resource
is in a fluid state, it has the propensity to move, mix, and re-
arrange itself relatively rapidly during the course of a project
lifetime. This is unlike any other type of mineral resource,
and hence a different approach to its investigation and evalu-
ation is required. There are no standard methodologies for
evaluating water resources, inasmuch as such resources are
rarely considered a commercial commodity. Hence, there is a
requirement for new or modified filing standards.
Here, we review the requirements for brine resource and
reserve evaluation, drawing on several examples from our ex-
perience in the Central Andes.
Continental Brine Occurrence
Worldwide
Brines have been exploited on a relatively small scale at Sil-
ver Peak, Nevada, and Bristol and Searles Lakes, California,
for many years for a variety of aqueous species. However, the
largest tonnages of Li and K are located in the Andes of Chile,
Argentina, and Bolivia, and in western China and Tibet. The
latter group, however, has complex chemistry, and the eco-
nomic recovery of the contained Li and associated elements
is proving difficult. This is not the case in South America,
where three operations (Soquimich and Chemetall in the
Salar de Atacama, FMC in Hombre Muerto) are responsible
for a very large percentage of the current world lithium sup-
ply. These sources, together with more recent discoveries, are
expected to meet much of the massive demand for the elec-
trification of motor vehicles (Evans, 2008, 2010).
Basin and host aquifer formation in the
Altiplano-Puna of the Central Andes
The Altiplano-Puna is the second largest high-altitude
plateau in the world and is the location of numerous brine
bodies containing elevated concentrations of Li among sev-
eral other species of economic interest (Table 1). The Andes
of western South America are the result of subduction
processes as the Nazca plate descends beneath the South
American plate (Oncken et al., 2006) (Fig. 1). The central vol-
canic zone, between 14° and 28°S, is the location of an ex-
tensive Neogene ignimbrite province (de Silva, 1989) that is
underlain by one of the largest magma bodies in existence,
known as the Altiplano-Puna magma body (Zandt et al., 2003;
de Silva et al., 2006). In the Central Andes, salt pans, known
locally as salares, form in closed topographic depressions (en-
The Evaluation of Brine Prospects and the Requirement for
Modifications to Filing Standards
JOHN HOUSTON,
1,†
ANDREW BUTCHER,
2
PETER EHREN,
3
KEITH EVANS,
4
AND LINDA GODFREY
5
1
Stuart Lodge, 273 Wells Road, Malvern, WR14 4HH, United Kingdom
2
British Geological Survey, Maclean Building, Crowmarsh Gifford, Wallingford, OX10 8BB, United Kingdom
3
Ehren-González Ltda., Pedro Pablo Rubens 2969, La Serena, Chile
4
11940 Adorno Place, San Diego, California 92128
5
Institute of Marine and Coastal Sciences, Rutgers University, 71 Dudley Road, New Brunswick, New Jersey 08901
Abstract
The recent increase in demand for lithium has led to the development of new brine prospects, particularly
in the Central Andes. The brines are hosted in closed basin aquifers of two types: mature, halite dominant, and
immature, clastic dominant. The estimate of elemental resources in these salars depends on a detailed knowl-
edge of aquifer geometry, porosity, and brine grade. The geometry of the aquifers can be evaluated by classi-
cal geophysical and drilling techniques, but because the resource is a fluid, with the attendant problems of in-
aquifer mixing and reorganization, existing codes for filing resource and reserve estimates need modification.
Total porosity is relatively straightforward to measure, but effective porosity and specific yield, which are re-
quired to estimate the resource, are more difficult. Recovery factors are low compared with most metallifer-
ous and industrial mineral deposits due to reliance on pumping of the brine from wells for extraction. These
and related issues lead us to believe that modifications to the existing standards for reporting mineral resources
and reserves are required for these prospects.
†
Corresponding author: e-mail, houston.jft@gmail.com
©2011 Society of Economic Geologists, Inc.
Economic Geology, v. 106, pp. 1225–1239
Submitted: January 11, 2011
Accepted: May 30, 2011
dorheic basins). Salars occur at all elevations from 1,000 to
more than 4,000 m above sea level (Fig. 2). They generally
represent the end product of a basin infill process that starts
with the erosion of the surrounding relief, initially depositing
colluvial talus and fan gravels, grading upward into sheet
sands, and playa silts and clays as the basin fills. There are
many variants to this model, and the tectonic and sedimentary
processes that led to the formation of such basins have been
widely addressed in the literature both generally (Hardie et
al., 1978; Reading, 1996; Warren, 1999; Einsele, 2000) and
specifically with regard to the Altiplano-Puna (Ericksen and
Salas, 1989; Alonso et al., 1991; Chong et al., 1999; Bobst et
al., 2001; Lowenstein et al., 2003; Risacher et al., 2003;
Vinante and Alonso, 2006).
Structure controls the compartmentalization of many An-
dean basins. North-south–aligned thrust faults, grabens, and
half grabens frequently create accommodation space, while
transverse strike-slip faulting may assist with basin closure,
offsetting basins against impermeable bedrock (Salfity, 1985;
Marrett et al., 1994; Reijs and McClay, 2003). Volcanism also
plays a significant role, both in basin infill (e.g., tuffs and ign-
imbrites) and in basin closure (e.g., volcanic necks and lava
flows).
The latitude of the Central Andes and its position under the
subtropical high-pressure belt for at least the last 55 m.y.
(Hartley et al., 2005) has influenced both the type of sedi-
mentary infill and its architecture within the basins. Basin clo-
sure is thought to have occurred frequently at about 14 Ma
(Vandevoort et al., 1995), although the majority of evaporitic
deposits appear to be less than 8 m.y. old (Alonso et al., 1991).
Recent evidence suggests that the Andes reached their cur-
rent elevation about 6 m.y. ago (Ghosh et al., 2006), and since
that time the climate has been dominated by hyperarid con-
ditions (Hartley and Chong, 2001), allowing ample opportu-
nity for evaporation of the influent water. However, during
the same period there have also been excursions into wetter
conditions (Fritz et al., 2004; Placzek et al., 2006; Rech et al.,
2010), potentially allowing salt recycling. During the course
of the aquifer formation, influent groundwater and surface
water have not always had the opportunity to escape from the
basin, leading to the formation of temporary lakes or wet-
lands. Because the influent waters contained dissolved solutes
as well as transported sediment load, evaporation resulted in
the precipitation of salts, leading to the deposition of a wide
range of evaporite deposits. Depending on the paleohydro-
logical history of the basin, the deposition of evaporites may
have taken place on more than one occasion, generating re-
peat sequences. There is a typical precipitation sequence
starting with carbonate (typically calcite) as the first mineral
precipitated, through sulfate (typically gypsum), to chloride
(halite). Of course, natural salars rarely conform to this ideal.
Asymmetry, gradational and changing boundary positions due
to climate change, tectonism, and sediment supply are normal.
Salar types
We recognize two types of host aquifers in the Altiplano-
Puna: mature halite salars and immature clastic salars (Fig. 3).
A classification of salar types in the Altiplano-Puna is pro-
vided in Table 1.
Immature salars may be characterized by their greater
moisture regimes (higher precipitation, lower evaporation),
1226 HOUSTON ET AL.
0361-0128/98/000/000-00 $6.00 1226
TABLE 1. Selected Salar Types and Brine Chemistry in the Altiplano-Puna Region
Salar Area Elevation MAP Salar type Brine type Cl Li K B
(km
2
) (masl) (mm) (typical values in g/l)
Uyuni 10,000 3,653 150 Immature Na-Cl-SO
4
190 0.42 8.7 0.24
Atacama 2,900 2,300 25 Mature Na-Cl-Ca/SO
4
210 2.55 27.4 0.82
Olaroz-Cauchari 550 3,900 130 Immature Na-Cl-SO
4
180 0.71 5.9 1.09
Huayatayoc-
Salinas Grande 2,500 3,400 180 Immature Na-Cl-Ca/SO
4
190 0.78 9.8 0.23
Rincon 280 3,740 63 Largely mature Na-Cl-SO
4
195 0.40 7.5 0.33
Arizaro 1,600 3,500 50 Immature Na-Cl-SO
4
190 0.08 4.0 0.12
Pocitos 435 3,660 60 Immature Na-Cl-SO
4
170 0.09 4.8 1.32
Antofalla 540 3,580 - ?Immature Na-Cl-SO
4
166 0.32 4.7 10.80
Hombre Muerto W 350 3,750 77 Mature Na-Cl-SO
4
195 0.68 6.3 2.06
Hombre Muerto E 280 3,750 77 Immature Na-Cl-SO
4
140 0.78 8.9 0.62
Maricunga 90 3,760 35 Mixed Na-Cl-Ca/SO
4
204 1.05 8.9 0.79
Abbreviations: MAP = mean annual precipitation, masl = meters above sea level
70 W
o
68 W
o
66 W
o
72 W
o
Altiplano
W
Co
rdillera
E Cordillera
Precordillera
Longitudinal
V
alley
Coastal Cordillera
Peru-Chile
Trench
64 W
o
20 S
o
18 S
o
22 S
o
24 S
o
26 S
o
28 S
o
Puna
Subandean
Range
s
S
ierras
P
a
m
peanas
S
an
t
a
Barba
ra
R
a
n
g
es
APMB
Fig. 2
100 km
FIG. 1. Location map of the Altiplano-Puna and the principal physio-
graphic features, including the Altiplano-Puna magma body of de Silva et al.,
2006 (dashed white line). Volcanoes and calderas of the western cordillera
are shown as solid black triangles and open circles, respectively.
and hence tend to be more frequent at higher elevations and
toward the wetter northern and eastern parts of the region.
They are characterized by an alternating sequence of fine-
grained sediments and evaporitic beds of halite and/or ulex-
ite, representing the waxing and waning of sediment supply
under a variable tectonic and climatic history. The contained
brines often barely reach halite saturation, suggesting that the
climate during their formation was not severely hyperarid.
The frequent occurrence of a thin surface halite crust in these
salars may not be a good climate indicator because they are
probably ephemeral and may get recycled during burial. The
alternation of drier and wetter climates may lead to inertial
disequilibrium between evaporation rates and the brine con-
centration, since an increase (decrease) in evaporation rate
will take a considerable time to cause the whole brine body to
increase (decrease) in concentration. The brines are normally
fully saturated with respect to gypsum, leading to the wide-
spread occurrence of gypsum (typically as selenite) through-
out the sequence. Past dry climate intervals are evidenced by
buried halite beds, suggesting that decreased precipitation in-
flow and/or increased evaporation may have lead to brines
saturated in halite. The presence of intercalated or underly-
ing beds of different permeability sometimes allows the trans-
mission of fresher waters from outside the salar margins
through to the center, where there is a tendency for the den-
sity differential with the nucleus brine to augment upward
flow of the brine, providing that the confining bed has suffi-
cient permeability to allow such leakage.
Mature salars have a lower moisture flux, and thus tend to
be more common in the lower and drier parts of the region.
They are characterized by a relatively uniform and thick se-
quence of halite deposited under varying subaqueous to sub-
aerial conditions (Bobst et al., 2001). Nevertheless, ancient
floods leading to widespread silty clay deposits and volcanic
fallout have led to thin intercalated beds that can be recog-
nized in cores and geophysical logs. Such layers of varying
permeability may lead to the formation of alternating aquifers
and aquicludes that pinch out around the margins of the nu-
cleus. Fresh groundwater in the higher-permeability layers
may be transmitted from outside the salar margins to the
edge of the nucleus where, once unconfined, it flows to the
surface as a result of the pressure differential with the nucleus
brine. The pressure differential is composed of two elements:
the imposed head and the density difference. Fresher waters
flowing to the surface dissolve halite in their ascent and lead
to the formation of pipes and salt dolines at the surface, es-
pecially in the marginal zones (Fig. 3). The contained brines
are invariably halite saturated throughout the brine body, al-
though the presence of multiple brine types, especially in the
larger salars, points to the hydrochemical variation of the con-
tributing source waters.
The distinction between salar types is maintained even
within the same basin, as at Hombre Muerto, where a mature
sub-basin exists to the west as a result of moderately evolved
brines decanting from the immature eastern sub-basin over a
subsurface bedrock barrier. Both types of salar may contain
THE EVALUATION OF BRINE PROSPECTS AND THE REQUIREMENT FOR MODIFICATIONS TO FILING STANDARDS 1227
0361-0128/98/000/000-00 $6.00 1227
-
Huayatayoc-
Salinas
Grande
-
Olaroz
Cauchari
Rincon
Pocitos
Arizaro
Antofalla
Hombre
Muerto
Maricunga
Atacama
Uyuni
FIG. 2. Digital elevation model (SRTM-90) of the Central Andes, showing the location of the Altiplano-Puna magma body
(dashed line) of de Silva et al., 2006, and the location of selected lithium-rich salars in closed-basin depressions.
commercially valuable brine resources, and while it might be
anticipated that mature salars contain more concentrated so-
lutions, this is not always the case. Elements such as Li, K,
and B may reach very high levels in immature salars (Table 1),
and, of course, clastic deposits possess considerably higher
porosities than halite.
The pattern and distribution of crustal types may allow the
identification of salar type in the field and on satellite im-
agery. Both types display the same range of features, from
high- reflectance re-solution crust, through salt polygons, to
low-reflectance pinnacle halite, representing a progression
from younger (<1 yr) to older (>10 yr) formation. However,
immature salars tend to have a much larger proportion of
their surface represented by re-solution crust and relatively
small areas of pinnacle halite.
Porosity and permeability
There is considerable misunderstanding of the terminology
related to porosity. Total porosity (P
t
) relates to the volume of
pores contained within a unit volume of aquifer material. Ex-
cept in well-sorted sands, some of the pores are isolated from
each other, and only the pores that are in mutual contact may
be drained. This interconnected porosity is known as the ef-
fective porosity (P
e
). Assuming that the P
e
is totally saturated,
only part may be drained under gravity during the pumping
process. This part of the porosity is known as the specific
yield, or sometimes the drainable porosity (S
y
). A portion of
the fluid in the pores is retained as a result of adsorption and
capillary forces and is known as specific retention (S
r
). These
parameters are related thus:
P
t
> P
e
P
e
= S
y
+ S
r
S
y
>
<
S
r
.
The relationship between S
y
and S
r
depends largely on
lithology (Fig. 4). In fine-grained sediments S
y
<< S
r
, whereas
in coarser-grained sediments S
y
>> S
r
. The determination of
these pore parameters is probably the most challenging as-
pect of brine resource estimation.
1228 HOUSTON ET AL.
0361-0128/98/000/000-00 $6.00 1228
SALAR SURF ACE
FEATURES
Re-solution crust
Polygons
Mature pinnacles
HYDROLOGY
Evaporation
Surface water
Groundwater
flow
flow
Transition from
freshwater
to brine
Halite
FACIES
Sands, silts, muds &
organic oozes
Palustrine clays with sulphates
Mudflat clays with carbonates
Sandflat sands
Braidplain sands &
fluvial channels
Talus
Pre-salar deposits
Bedrock or alluvial fan
high K
high K
low K
IMMATURE
CLASTIC SALAR
MATURE
HALITE SALAR
Salar
marginal
facies
Salar
nucleus
Salar
marginal
facies
Salar
nucleus
FIG. 3. Block models of mature and immature salars showing the distribution of facies and main hydrological components.
In the mature model, extension and recession of the marginal facies as a result of tectonism and climate variation lead to the
possibility of dilute waters being transferred into the nucleus. In the immature model, while the marginal conditions have
been simplified for clarity, the transmission of dilute waters into the nucleus is also possible. K refers to the hydraulic con-
ductivity of the different units.
The permeability of mature halite salars is very different
from that of immature clastic salars. Mature halite salars tend
to be rather more homogeneous and isotropic, and the per-
meability is controlled by matrix porosity and, in some zones,
by fissures. Unpublished data (Houston, 1987–1996) from the
Salar de Atacama and Hombre Muerto indicate that fissures
are most common at depths between 5 and 25 m in halite,
and, in these zones, permeability may reach extremely high
values (up to 10
5
m/d), whereas matrix values in the upper
30 m are normally in the 10 to 100 m/d range. Below 30 m,
the permeability of halite decreases significantly as a result of
cementation and crystal overgrowth (Casas and Lowenstsein,
1989). On the contrary, clastic sediments tend to be inhomo-
geneous and anisotropic, permeability being highly depen-
dent on lithology. Fissures are largely absent, so permeabil-
ity may range anywhere between 10
–2
and 10
2
m/d and
declines only slightly with depth (over the likely depth range
for extraction).
Brine chemistry
The chemistry of the brines is generally thought to result
from the weathering of surrounding rocks by the infiltration
and passage of precipitation through them, leading to the for-
mation of dilute inflow waters that subsequently concentrate
by evaporation. The brines hosted in salar aquifers show a
typical increase in concentration from their margins to the
nucleus. Highest concentration gradients occur where evapo-
ration is fiercest (Fig. 5), frequently reaching a plateau in the
nucleus, where evaporation rates are minimal as a result of
the reduced saturation vapor pressure gradient over brines,
coupled with the impermeable salt crust where present.
Concentration by evaporation leads to the development of
chemical divides (Hardie and Eugster, 1970), which create dif-
fering end points depending on the original ionic ratios, espe-
cially for the Ca-SO
4
system. The initial ion ratios in the influ-
ent waters depend on the passage of the water over and
through the basin rocks, and if several catchments drain to the
same salar, multiple brine types within the aquifer may result.
The Hardie-Eugster model interprets the chemical evolution
in terms of a series of divides that leads to at least six alterna-
tive evolutionary pathways, although only two are common in
the Altiplano-Puna. The exact concentration and density at
saturation is dependent on the dissolved species, although
empirical guidelines can be provided. The first mineral to
precipitate is calcite (Fig. 6A), typically at concentrations of
about 30 g/l total dissolved solids (approximate density equiv-
alent 1.03 g/cm
3
). This is usually followed by gypsum (Fig.
6B), typically at concentrations of about 200 g/l (approximate
density equivalent 1.15 g/cm
3
), or, in alkaline conditions, by
magnesium-rich smectite (Badaut and Risacher, 1983). Halite
saturation begins only at total dissolved solid concentrations
of about 280 g/l (approximate density equivalent 1.22 g/cm
3
),
and an Na-Cl divide is rarely seen at concentrations of less
than 320 g/l. Once halite saturation is reached and starts to
precipitate, other species may concentrate further in the
residual brine, leading, for example, to high concentrations of
Li (Fig. 6C).
The most common brine types in the Altiplano-Puna, based
on data in Risacher et al. (2003) for northern Chile, Rettig et al.
(1980) and Risacher and Fritz (1991) for Bolivia, and Igarzábal
(1984) and unpublished data (Houston et al., 2009–2011) for
northwestern Argentina, are Na-Cl-SO
4
and Na-Cl-Ca. The
economically important elements Li, K, and B are normally
found to covary with Na and Cl, but there are some notable
deviations. As noted above, Li concentrations are much higher
than expected in some salars. At Cl concentrations above
about 160 g/l, the Cl/Li ratio is typically >300. However, in the
southern part of the Salar de Atacama, the Cl/Li ratio is as low
as 30, and in the Salar de Uyuni, close to the mouth of the rel-
atively dilute influent Rio Grande, and again close to fresh-
water inflows at Hombre Muerto, it is ca. 140. The reasons
for these very high lithium concentrations, especially where
they are so close to dilute inflows, are currently unknown.
The source of the salts, and particularly the economic com-
ponents such as Li, K, and B, is the subject of debate. The dis-
tribution of Li-K-B–bearing brines in salar host aquifers
THE EVALUATION OF BRINE PROSPECTS AND THE REQUIREMENT FOR MODIFICATIONS TO FILING STANDARDS 1229
0361-0128/98/000/000-00 $6.00 1229
clay
siltsand
gravel
10100
1
0.1
0.01 0.001
Mean grain diameter (mm)
10
20
30
50
40
60
Porosity (%)
T
O
TA
L
P
OROS
I
T
Y
EFF
E
C
T
IV
EP
OROSIT
Y
SPECIFIC
YIEL
D
SPE
CIFIC RETENT
I
ON
0
50
100
150
200
02550
SAND
HALITE
Depth
(m)
Specific yield (%)
A
B
A
FIG. 4. Porosity relations for sedimentary basins and salars. A) shows how grain size affects the different components of
porosity. Values are typical for largely unconsolidated sediments under lithostatic (normal) pressure and within the upper 30
m of the crust. B) shows how porosity changes with depth for sands and halite. Although halite is generally considered to
have almost no porosity below 30 m as a result of compaction and crystal overgrowth, open, brine-filled fissures have been
observed at depths of 100 m or more in several Andean salars.
appears to be coincident with the Altiplano-Puna magma
body (Fig. 2), and it has been suggested (Ide and Kunasz,
1989) that this is their ultimate origin, but the precise mech-
anism remains speculative. Numerous hypotheses have been
advanced. Based on their bulk chemistry, it appears that re-
cycling of earlier deposits/salars seems to be a common factor
in many salars (Risacher and Fritz, 1991; Orti and Alonso,
2000; Risacher et al., 2003), and this is supported by mass bal-
ance and accumulation rates analyzed in the Salar de Atacama
(pers. commun., Godfrey and Houston, 2010). Mixing with
pre-existing subsurface brines has also been suggested
(Risacher et al., 2003), but neither of these hypotheses point
to an ultimate source for the anomalous values of Li and other
minor elements. Weathering of Neogene volcanic rocks has
been suggested by Rettig et al. (1980), Orti and Alonso
(2000), and Kay et al. (2010), but simple weathering of glass
from these volcanic rocks does not account for the very high
Li/Mg ratios (often >0.2) that characterize many of the Cen-
tral Andean brines. Alternatively, weathering of Palaeozoic
basement has been suggested by Ortis and Alonso (2000) and
Kasemann et al. (2004). Hydrothermal activity is another can-
didate mechanism (Orti and Alonso, 2000; Gibert et al., 2009;
Lowenstein and Risacher, 2009), especially because the sim-
plest way to produce brine with relatively low Mg concentra-
tions relative to Li is to extract both elements from rocks at
high temperatures. Because chlorite forms in these systems,
Mg is sequestered, and Li stays in solution (Badaut and
Risacher, 1983).
Brine hydrology and water balance
The brine solution in all South American Altiplano salars
has been formed over many years (10
4
–10
5
yr) in endorheic
basins. Precipitation within the catchments drains toward the
center of the basin either as surface water or groundwater
that contains dilute solutions of commercially interesting ele-
ments. Around the margins of the salar, the drainage water is
either close to the surface or in the form of open water la-
goons. Extremely high rates of evaporation prevail in this
marginal zone (Fig. 7). Evaporation of the relatively fresh in-
fluent water causes concentration of the entrained dilute dis-
solved salts. Very small volumes of concentrated fluid then
enter the salar nucleus and replace very small amounts of
brine evaporated from the surface of the salar. Halite crusts
are impermeable when they are more than a few months old
(Kampf et al., 2005), and, within the nucleus of the salar, the
depth to brine is controlled by limited evaporation, the ex-
tinction depth for which is universally between 0.5 and 1.0 m
(Houston, 2006).
Although the brine is generally considered to be static or
“fossil,” the system is actually in dynamic equilibrium, with a
1230 HOUSTON ET AL.
0361-0128/98/000/000-00 $6.00 1230
23 S
o
23.5 S
o
68.5 W
o
68 W
o
Cordillera de la Sal
Chloride
nucleus
Marginal
carbonate
-sulfate
zones
20 km
1.23
1.20
1.00
1.10
SO rich
brines
4
Li-K rich
brines
R San Pedro
delta
1.00
1.22
1.20
1.10
1.00
Archibarca
fan
Chloride
nucleus
Rosario
delta
Marginal zones
10 km
10 km
67 W
o
25.25 S
o
25.5 S
o
1
.00
1
.10
1.20
1.20
1.21
0.9
1.0
1.1
1.2
1.3
0
100
200
300 400
total dissolved solids (g/ l)
density (g/cm
)
3
A
B
C
1
.21
Los Patos
delta
Eastern
nucleus
Western
nucleus
FIG. 5. Satellite imagery of the Salars de Atacama (A), Olaroz (B), and Hombre Muerto (C) showing their main surface
zones and mean brine density contours for the upper 30 to 40 m of the host aquifer. Inset plot shows relationship between
total dissolved solids concentration and density for the Salar de Olaroz.
slow turnover controlled by evaporation. The dynamic aspect
of these salars is also attested by the presence of dissolution
pipes within their nuclei, combined with halite-saturated
brine bearing no indication of evaporation, based on the sta-
ble isotopes of O and H (Fritz et al., 1978). Furthermore,
within the brine body a density-driven convection cell some-
times develops. Evaporation at the phreatic surface increases
the brine density, causing it to sink through the aquifer. This
sinking manifests in reduced heads with depth in the center
of the salar. Correlative increasing heads with depth toward
the edge of the brine body indicate return flow of brine to the
surface.
The water balance of these salars is given by the following
equation:
P + I
SW
+ I
GW
– E
M
– E
N
= ∆S,
where P is precipitation over the salar, I
SW
are surface water
inputs, I
GW
are groundwater inputs, E
M
is the marginal evap-
oration, E
N
is the nucleus evaporation, and ∆S is the change
in storage within the nucleus (manifested by increasing or de-
creasing brine levels).
By ignoring evaporation, because it does not remove salts
from the system, the water balance can be converted to a
mass balance thus:
P ⭈ C
1
+ I
SW
⭈ C
2
+ I
GW
⭈ C
3
= ∆M ⭈ (1 – P
t
),
where C represents the concentration of the respective fluid,
∆M is the mass change in precipitated material, and P
t
is the
total porosity of the precipitated material.
The transition from fresher waters to brine around the salar
margin creates an interface in the subsurface that is governed
by both layer permeability and the Ghyben-Hertzberg law
(Badon-Ghyben, 1888; Herzberg, 1901). The latter states that
within a homogeneous and isotropic medium, the depth to
the interface is a function of the fresh-water head and the
ratio of the fresh water to brine fluid density. This inevitably
leads to the expansion of the brine body at depth.
Brine Body Response to Extraction
Immediately when pumping (extraction) starts, redistribu-
tion of the resource begins because the fluid is being dynam-
ically stressed. Pumping induces a cone of depression in the
phreatic (water table) surface around the well field. The size
and shape of the cone of depression depend on the physical
properties (permeability and storage) of the aquifer, as well as
the pumping rate and duration. Thus, the cone of depression
expands outward (and downward) over time until a new dy-
namic equilibrium becomes established, which may take
many years before stabilizing.
The controlling factor for extraction rates is the permeabil-
ity of the aquifer, with higher (lower) values being more (less)
conducive to higher (lower) yields. Thus, in mature halite
salars, the high permeability and relatively low S
y
result in a
flat cone that rapidly extends laterally (Fig. 8A). Immature
salars with relatively low permeability and higher S
y
result in
a steep cone that tends to deepen faster than it extends (Fig.
8B). Almost inevitably, the cone will extend to the boundaries
of the aquifer after a period of time, and the result depends
on the nature of the boundary. If it is permeable, off-salar flu-
ids will flow inward (Fig. 8A, solid line). On the other hand, if
the boundary is impermeable (Fig. 8A, B, dashed lines), the
cone will increase its downward rate of propagation.
In the event that dilute off-salar fluids are drawn into the
cone of depression, barren brine and/or “fresh” water may
eventually migrate to the wells. This is exacerbated if the in-
flow is of a lower density because the two fluids will tend to be
immiscible, with the inflow floating over the resource brine.
Ultimately, the inflow may pool in the cone of depression,
creating problems with the extraction of the reserve brine.
A related extraction problem that occurs in both mature
and immature salars is the ingress of fresh water at depth as a
THE EVALUATION OF BRINE PROSPECTS AND THE REQUIREMENT FOR MODIFICATIONS TO FILING STANDARDS 1231
0361-0128/98/000/000-00 $6.00 1231
0.1
1
10
100
1,000
0.1 1 10 100
1,000 10,000
Cl (mmol/ l)
Li (mmol/ l)
Halite saturation
Cl stabilizes
Li increases
1
10
100
1,000
10,000
100,000
0.1 1 10 100 1,000
Alkalinity (meq/ l)
2Ca + 2 Mg (mmol/
l)
Calcite precipitation
Ca increases in brine
carbonate decreases
Hardie-Eugster path II
A
0.1
1
10
100
1,000
0.1 1 10 100
1,000
SO (mmol/ l)
4
Gypsum precipitation
SO increases in brine
Ca decreases
Hardie-Eugster path VI
4
Ca (mmol/
l)
B
C
FIG. 6. Brine chemistry from the Salar de Olaroz showing its evolution
with increasing evaporation and concentration. Plots A, B, and C show the
chemical divides for calcite, gypsum, and halite, respectively.
result of higher-permeability layers or beds (for example, thin
sand and gravel beds) extending into the salar (Fig. 3). The
fresh water in these beds is usually sourced from recharge
outside the limits of the salar and transferred through the
higher-permeability bed as a result of an overlying confining
layer. Where the confining layer feathers out, the fresh water
is released and will rise as a result of its head and density dif-
ferential with the surrounding fluid. In halite, this upward
flow of fresh water may lead to the formation of solution pipes
that reach the surface. The importance of this mechanism can
be seen in Figure 9, where the concentrated in situ brine has
the potential to be somewhat diluted during pumping as a re-
sult of mixing with the fresh water that enters the well from
thin sand horizons.
For a given set of hydrogeological conditions and pump-
ing rates, a larger and more nearly circular salar will be bet-
ter able to support long-term extraction than a smaller or
elongate shape. In addition to aquifer size and shape, stor-
age exerts control over the scale of the recoverable reserve.
The ratio of extraction volume to storage volume will help
to determine the life of a project. This is particularly im-
portant since it is not widely appreciated that a rule of
thumb for the recovery factor of brine from an aquifer is
only about one-third.
Requirements for a Brine Evaluation
An in situ brine resource evaluation should consist of three
essential elements: the volume of the host aquifer, the S
y
of
the host aquifer, and the concentration (grade) of the ele-
ments of interest in the brine. The product of the geometry
and S
y
determines the volume of the brine resource. The el-
emental resource is determined as the product of the brine
volume and the grade of the element in the brine.
1232 HOUSTON ET AL.
0361-0128/98/000/000-00 $6.00 1232
San Pedro
de Atacama
Cordillera de la Sa
l
20 km
Chloride
nucleus
Clay
delta
M
a
rg
in
a
l
z
on
e
s
0
1
2
3
0
1
2
3
Depth to saturation (m)
Mean annual evaporation (m)
brine
(1.24 g/cm )
3
fresh/brackish
(1.0-1.1 g/cm )
3
0
1
2
1.0
1.1
1.2
1.3
Density (g/cm )
3
fresh water entering system
from aquifers comes close to surface
increasing evaporation
max evaporation
in marginal zones
and lagoons
further evaporation leads to brine
concentration and consequent
reduction in evaporation
Depth to saturation (m)
A
B
C
FIG. 7. Evaporation relations for the salars (modified from Houston, 2006). A) shows the typical increase in evaporation
as the water table comes close to the surface—note the difference in evaporation rates between dilute waters and brine. B)
shows the influence of water table depth and brine for the marginal zones of salars. C) shows a satellite image (bands 5, 4, 7
as RGB) of the Salar de Atacama, processed to reveal the main evaporating zone around the eastern margin.
0
1
2
1050510
0
10
20
30
40
50
MATURE AQUIFER
K=200 m/ d, Sy=0.05
IMMATURE AQUIFER
K=2 m/ d, Sy=0.1
Distance from wellfield centreto aquifer boundary (km)
depth (m)
(note scale dif
ference)
A
B
FIG. 8. Drawdowns around three wells, each pumping at 20 l/s for 25
years and spaced 1 km apart in a 100-m deep aquifer, showing the differences
that occur due to aquifer properties and extent. The solid lines represent the
cone of depression that occurs in a horizontally infinite aquifer (i.e., extends
beyond the bounds of the plots). The dashed line represents the cone of de-
pression that occurs in an aquifer with impermeable boundaries at 10 km
(i.e., the edges of the plots).
Host aquifer definition
Aquifer area and thickness, as well as boundary conditions
(e.g., faulted, gradational), are necessary to establish the lim-
its of the reservoir and the possible interactions between the
contained brine and surrounding groundwater. Within the
nucleus of the salar, the lithological variations of the aquifer
and its hydrostratigraphy will control both brine storage and
transmission. Initial studies often use geophysical techniques
such as gravity or seismic methods, but inevitably drilling will
be needed for a resource estimate. Drilling mature halite
salars is considerably easier than drilling immature clastic
salars, where underconsolidated formations are difficult to
drill and may require special equipment, such as sonic or
triple-tube diamond rigs.
Drill sites within claim areas should be on a grid spacing to
facilitate subsequent analysis. The size of the grid will depend
on the type of aquifer (salar) and on the level of confidence
required. Table 2 provides some guidelines that have been
used in several investigations by the authors. The spacing may
seem coarse when compared with that commonly used in in-
dustrial mineral exploration, but it is unlikely that horizontal
changes in brine chemistry will be more rapid. In addition to
defining the horizontal variations in the properties of the
claim areas, the vertical sampling interval for porosity and
brine chemistry is critical. Samples for both porosity and
brine chemistry should be taken at the same depths in order
to ensure compatibility during resource calculation. In some
cases, a vertical sample interval of 0.5 m has been chosen, but
with increased experience it would appear that 1.5 m is ac-
ceptable for a measured resource, expanding to 5 m for an in-
ferred resource, although this depends on hydrogeological
conditions and should be reviewed based on exploratory data.
High core recoveries are essential in environments where
rapid vertical changes in lithology may control both porosity
and brine variations. Geological logging of the cores needs
careful control to avoid subjective description bias in these
environments. Once the drill holes are at full depth, geophys-
ical logging will assist with hydrostratigraphic analysis. In this
regard, the most useful tool is usually natural gamma, since
the response of electric logs is generally masked by the ex-
tremely low resistivity of the brine.
Core sampling for the laboratory determination of porosity
must be undertaken on undisturbed samples. These may be
difficult to obtain in unconsolidated formations using conven-
tional methods; split spoons or Shelby tubes pushed ahead of
the drill face may be necessary. On-site laboratory determina-
tion of P
t
and P
e
is not difficult using gravimetric methods, pro-
vided initial moisture loss is prevented after core extraction
and subsequently controlled to a stable weight in the labora-
tory. It is essential not to dry the cores at too high a tempera-
ture, or gypsum will convert to anhydrite, and organic matter
may incinerate, creating erroneous measurements. Further-
more, due account must be taken of the salts precipitated
during the drying process as a result of brine evaporation.
In a recent study, samples from Lexan core barrels (9.4-cm
diam) were compared with samples from split spoon cores (3.4-
cm diam), both obtained by sonic drilling. The resulting P
t
de-
terminations (Fig. 10A) show wide scatter despite good corre-
lation (r = 0.65; p < 0.001). This scatter is partly a result of the
depth difference between the Lexan and split spoon cores (0.3
m) in a thin-bedded sequence with rapid alternations of lithol-
ogy. The veracity of these measurements can be seen when
compared with the neutron porosity log (N-Pt) in Figure 9.
The determination of P
e
, S
y
, and S
r
presents special prob-
lems, and there are no standards for this. Two methods are in
common use for the laboratory determination of P
e
: liquid
resaturation and helium injection. The former usually re-
quires the initial drying of the sample and subsequent resat-
uration using formation brine or an inert liquid such as iso-
propyl alcohol, although fully saturated samples can be
tested and then dried. The latter involves the injection of he-
lium into the sample under controlled conditions. P
e
is then
calculated using Archimedes’ principle and Boyle’s law, re-
spectively. The results (Fig. 10B) suggest that the liquid re-
saturation method may underestimate P
e
in fine-grained sed-
iments as a result of the extended drying times required. For
THE EVALUATION OF BRINE PROSPECTS AND THE REQUIREMENT FOR MODIFICATIONS TO FILING STANDARDS 1233
0361-0128/98/000/000-00 $6.00 1233
0
10
20
30
40
50
60
0 100 200
50 100
1.1 1.2 1.3
1000 100000
GR - API
N-Pt %
dens
g/cm
3
conc mg/ l
Ca SO
4
depth m
simplified lithology sand
silt & clay halite
FIG. 9. Data from a well in the Salar de Olaroz showing gamma log (GR),
neutron porosity log (N-Pt) overlain with on-site lab determinations of P
t
(gray dots), simplified lithological log, density, and Ca-SO
4
concentrations.
Two zones of relatively dilute water, associated with thin higher porosity beds
within a silty clay unit, occur at 36- and 48- to 51-m depth. Such thin dilute
zones have the potential to lower grade during recovery.
TABLE 2. Suggested Drill Spacing in Kilometers for Different Salar Types
and Levels of Resource Definition
Salar type Inferred Indicated Measured
Mature 10 7 3–4
Immature 7–10 5 2.5
Notes: An inferred resource is estimated on the basis of geological evi-
dence and limited sampling without being verified; an indicated resource is
based on sufficient quantity and grade data to allow the technical and eco-
nomic parameters to be estimated to support mine planning and evaluation
of the economic viability of the deposit—the estimate assumes sufficiently
detailed and reliable exploration and testing data so that geological and grade
continuity may be reasonably assumed; a measured resource is based on suf-
ficient data to confirm grade continuity with a high degree of confidence
S
y
, centrifuge techniques are used, but for reliable and re-
peatable results it is essential to control rotation speed to in-
duce a moisture tension of ca.
1
⁄3 atmospheric (the point at
which gravitational flow effectively ceases in medium-grained
sediments) and a temperature of 4°C (Johnson et al, 1963;
Lawrence, 1977). Independent checks on porosity may be
made using resin-impregnated cores, thin sectioned and
point counted under a microscope. Petrological inspection
also allows the extent of grain disturbance and salt overgrowth
to be estimated. Where such tests have been performed on
sonic cores that have been immediately sealed on arriving at
the surface, it can be shown that very little or no precipitation
occurs within the core, only on its outside surface during lab-
oratory testing, and this can be readily corrected.
In some cases, pumping tests have been used to determine
S
y
for the resource assessment. But it is essential to understand
that, while pumping tests on wells determine the S
y
of the
aquifer directly, the determination is made within the dewa-
tered cone of depression (Walton, 1970; Bear, 1972). This is
contrary to the source of the fluid being pumped at the same
moment, which originates from the saturated zone below the
cone of depression, hence precluding the use of pumping
tests as a means of establishing the in situ resource.
Neutron and density (gamma-gamma) logs provide a means
of converting the point measurements determined in the lab-
oratory to a continuous porosity profile, although it must be
remembered that such logs determine P
t
even after calibra-
tion for the varying rock types, so that algorithms for the con-
version of P
t
to P
e
and S
y
are required. Examples of such
algorithms are shown in Figure 10C and D for P
e
and S
y
, re-
spectively. Consequently, it is essential to provide a clear set
of protocols with appropriate QA/QC procedures when re-
porting porosity determinations for resource and reserve cal-
culations if their validity is to be assessed.
1234 HOUSTON ET AL.
0361-0128/98/000/000-00 $6.00 1234
-
0
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
for >0.18
= 0.014
P
SP
e
ye
-1.65
Effective porosity ( )P
e
Specific yield
(
)S
y
for <0.18
=
P
SP
e
ye
y = 0.8x
0
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
Total porosity ( )P
t
Effective porosity
(
)P
e
Split spoon total porosity
(
)P
t
0
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
Lexan total porosity ( )P
t
20%
20%
A
y = 0.9x - 0.05
0 0.2 0.4 0.6 0.8 1
Effective porosity by He injection ()
P
e
0
0.2
0.4
0.6
0.8
1
Effective porosity by liquid resaturation
(
)
P
e
B
C
D
FIG. 10. Porosity data from five wells drilled using sonic techniques in the Salar de Olaroz. A) Comparison of Lexan and
split spoon undisturbed samples separated by 0.3-m depth interval and analyzed by gravimetric methods. B) Comparison of
effective porosity determined by liquid resaturation and helium injection on same samples. C) Comparison of total (split
spoon) and effective (He injection) porosity on samples separated by 0.15-m depth interval. D) Effective porosity and spe-
cific yield on same samples. Gray lines are the 1:1 relationship, dashed line in A indicates percent variability, and black lines
in B-D are best fit regressions.
The permeability of the formations will be needed to assess
well yields and flow regimes under both natural and pumped
conditions in order to estimate a recoverable reserve. An im-
portant consideration is that data will be required from out-
side the claims blocks since fluid movement is no respecter of
claim boundaries. Permeability within the aquifer may be es-
tablished using pumping tests or laboratory determinations
on core, although the former is likely to be better able to pre-
dict flows at field scale. Strictly speaking, permeability (k) is a
function only of the matrix, with dimensions L
2
, whereas flow
through the aquifer is also controlled by the fluid density and
viscosity and should be properly termed hydraulic conductiv-
ity (K) with units of length per time. This becomes important
during simulation modeling of recoverable reserves, when
changes in density become a factor.
Brine body definition
The geometry of the brine body within the aquifer requires
definition. The boundaries will be controlled by the transition
zone to fresh water, which will conform to Ghyben-Herzberg
principles. Time-domain electromagnetic or audio-magneto-
telluric studies provide a means of establishing the approxi-
mate limits of the brine body, taking into account the uncer-
tainties introduced as a result of the lack of a unique solution
(Archie’s Law: Archie, 1950).
Brine sampling needs to be undertaken to conform to the
porosity sampling plan for the eventual resource estimate.
Samples may be taken as point samples or as zone samples
using packer arrangements to isolate a section of the well.
The latter becomes impossible in unstable formations, where
point sampling is the only option. Obtaining in situ point sam-
ples may prove as challenging as or even more so than the
porosity analysis. How can it be proved that the sample comes
from the formation at a specific depth and has not mixed with
drilling fluid or overlying brines leaking around the drill rods
or casing? Drilling dry using reverse circulation or sonic tech-
niques is an obvious prerequisite but may prove impractica-
ble at depths greater than ca. 100 m, because the rods may
get stuck. The use of a tracer in the drilling fluid and well
water prior to sampling may assist in determining whether the
sample is depth-specific or contaminated with over- (or
under-) lying brines. A wellpoint pushed ahead into the for-
mation will help, but during insertion, the wellpoint will most
likely fill with the mixed well fluid, requiring either evacua-
tion of the entire column above the wellpoint or low-flow
pumping from within the wellpoint. If adequate core samples
are available, the extraction of fluid from the matrix, after re-
moving the potentially contaminated skin of the core, and
centrifuging may provide sufficient uncontaminated in situ
sample for analysis (Fig. 11A, B). Experiments recently con-
ducted at the Salar de Olaroz in holes drilled without any
fluid additives in unconsolidated clastic sediments have
shown that, where the casing closely follows the drill bit, sam-
ples bailed from within the casing are representative of in situ
brines (Fig. 11B). Obtaining representative in situ brine sam-
ples may thus not be so much of a problem as might be an-
ticipated, but it is always necessary to check and to be able to
demonstrate that the samples are uncontaminated.
Certain parameters should be determined at the wellhead:
temperature, density, and pH as a minimum, since they may
change during transit to the analytical laboratory. Brine
analysis in the laboratory is not addressed here, but it is
noted that such analysis is not as straightforward as that of di-
lute waters, as several commercial laboratories have demon-
strated, and a set of documented QA/QC procedures is es-
sential to establish the validity of the results. In addition to
sampling within the claim boundaries, samples outside these
boundaries are required to investigate the grade and quality
of fluids that may potentially be drawn into the reserve dur-
ing extraction.
Putting it all together—in situ resources
For a resource estimate, the aquifer geometry will be de-
fined by the area of the claims or that of the brine body,
whichever is the smaller, coupled with the depth determined
from drilling results. Although this is required for valuation
purposes, it should be recognized that this is only a conve-
nience, since once extraction starts, resources may well be
pulled in from outside this area. Where two properties are ad-
jacent, this could lead to competitive extraction techniques,
requiring some form of mutually acceptable compromise or
arbitration.
It has been common practice in the past—for example, at
the mature salars of Atacama and Hombre Muerto—to use P
e
to define the resource and to base the reserve evaluation on
S
y
. In mature salars, P
e
is only slightly greater than S
y
, so that
the reserve base is only slightly less than the resource. The
term “reserve base,” when applied to a fluid, is used to dif-
ferentiate the potentially extractable volume from the actually
extractable reserve, which depends on the ability of the hy-
draulic parameters to allow extraction by pumping (see
below). Now that immature salars are being developed, how-
ever, the use of P
e
to determine the resource would lead to a
significant difference between the reserve base and the re-
source, potentially by as much as an order of magnitude dif-
ference. Thus, we suggest that S
y
be used to determine both
the resource and the reserve in all salar types. This dilemma
clearly compels the qualified person (QP) reporting the re-
source to ensure that sufficient detail and explanation be pro-
vided so that the result is not misleading.
Numerous modeling packages are available for resource
estimation, but it is always necessary to consider how the
data is prepared for these to work properly. Firstly, the
model cell density should approximate that of the drill holes;
increasing the cell density does not increase the precision of
the estimate, as some recently published work supposes.
Secondly, the ideal method for calculating resource ton-
nages will be based on the derivative at each data point, so
that,
G
z,x,y
= S
y
z,x,y
⭈ C
z,x,y
⭈ b
z,x,y
,
where G
z,x,y
is the unit volume tonnage, S
y
z,x,y
the specific
yield, C
z,x,y
the elemental concentration, and b
z,x,y
the unit
thickness (note that “unit” as used here is not a geological
unit—it is the cell unit). The superscript z refers to depth, and
x and y are horizontal coordinates. Because b and z values
should ideally be equivalent across the claim area, the re-
source can be considered as a 3-D matrix of j unit cells:
Resource= Σ
j
z,x,y = 1
G
z,x,y
.
THE EVALUATION OF BRINE PROSPECTS AND THE REQUIREMENT FOR MODIFICATIONS TO FILING STANDARDS 1235
0361-0128/98/000/000-00 $6.00 1235
It is tempting to average parameters for each well or litho-
logic unit and sum the n well results across the claim area, but
this results in the following:
Resource = Σ
n
well = 1
({[Σ
k
z,x,y = 1
S
y
z,x,y
]/k}
⭈
{[Σ
k
z,x,y = 1
C
z,x,y
]/k}
⭈ Σ
k
z,x,y = 1
b
z,x,y
)
,
where k is the number of depth samples. This will result in a
mathematically different estimate and is not the best use of
the data.
However, in immature salars, it is likely that samples for
different parameter evaluation (S
y
, permeability, brine chem-
istry) may be separated by vertical distance up to several
centimeters and, in cases where bed thicknesses are thin rela-
tive to the vertical sample interval, may result in parameters at
a specific depth interval being representative of different rock
types. Thus, it may be necessary to group the cells into a se-
ries of layers. Because in thin-bedded environments lithology
and flow tend to be near horizontal, such zoning should not
induce significant errors. Spatial analysis will then be carried
out layer by layer. Since the vertical sampling interval may be
on the order of 1 m, the maximum number of layers might
rapidly approach or exceed 100, equal to the number of sam-
ples in the vertical direction. However, this is probably un-
wieldy, and we suggest that 10-m layer thicknesses are a rea-
sonable compromise, although site-specific conditions must
always be taken into consideration. This layer thickness
proved to give useful results in the cases of the mature salars
of Atacama and Hombre Muerto, but the more variable im-
mature salars appear to require thinner layers to give the
most useful results. The use of layer resource estimates is also
beneficial when considering reserves, since over time there
will be a tendency to exploit increasingly deeper layers as the
cone of depression from the pumping well field expands.
Algorithms available for estimation of cell or layer tonnages
are numerous and well documented in the literature (e.g.,
1236 HOUSTON ET AL.
0361-0128/98/000/000-00 $6.00 1236
-
0
200
400
600
800
1000
1200
0 200 400 600 800 1000 1200
-
10%
+10%
Well fluid Li mg/ l
Pore fluid Li mg/ l
A
0
10
20
30
40
50
500 1000 1500
Li mg/ l
Depth
m
Well, casing and rods at 10-m depth
Well, casing and rods at 38-m depth
B
Pumped samples from push-ahead wellpoint
Depth samples from inside casing
Depth samples from inside rods
400
600
800
1000
1200
0 50 100 150 200
pumped vol L
Li m
g/
l
Well, casing and rods at 10m
Well, casing and rods at 38m
C
Initial pumped sample from push-ahead wellpoint
Pumped samples vs. volume pumped over time
from push-ahead wellpoint
FIG. 11. Validation of brine sample representativeness from wells in the Salar de Olaroz. A) Lithium concentrations for
pumped wellpoint samples versus pore fluid extraction from core sample at the same depth. B) Pumped samples were taken
every 1.5 m during drilling from a push-ahead wellpoint attached to the end of the drill rods and inserted inside the casing,
which follows the drill bit closely. Such samples were assumed to be representative of the in situ aquifer fluid. When the cas-
ing and rods were at a 10- and 38-m depth, a series of samples were taken inside the casing and inside the drill rods. These
samples confirm that they can only originate from in situ aquifer fluid at 10 and 38 m. C) While the casing and rods were at
10- and 38-m depth, low-flow pumping samples were collected over time from the wellpoint. The concentrations remain sta-
ble over time, providing further proof that fluid from the wellpoints in any specific depth sample can only be in situ fluid
from the same depth. These data confirm the precise origin of the sample and indicate no contamination from over- (or
under-) lying brines.
Cressie, 1993), so that no review is necessary here. However,
it is perhaps worth mentioning that Kriging has attracted
widespread use and approval for hydrogeological studies
(e.g., de Marsily, 1986; Kitanidis, 1997).
Putting it all together—extractable reserve
The potentially extractable volume of brine (reserve base)
is defined by the S
y
of the aquifer, or the proportion drainable
under gravity. As discussed above, coarser-grained sediments
and halite present S
y
only slightly less than P
e
due to minor
amounts of S
r
. However, for fine-grained sediments, P
e
di-
verges significantly from S
y
(Fig. 4) as a result of considerable
S
r
. Thus, for the purpose of comparing resources and reserve
base across different types of salar, it is recommended that
both the resource and the reserve base be calculated using S
y
.
Only a relatively small proportion of the reserve base is ex-
tractable by pumping. Under any circumstances, it is only
possible to pump out a fraction of the fluid from an aquifer
because the cone of depression around a well or well field can
never lower the water level throughout the aquifer to its base.
A rule of thumb for the recovery factor (the volume that can
be pumped out) of an infinite aquifer is one-third of the re-
serve base. In the mature salar example shown in Figure 8A,
assuming a circular aquifer of radius 10 km, with an imper-
meable boundary, it would require 60 wells, spaced at 1.3 km,
each pumping at 20 l/s for 25 years, to extract 33 percent of
the resource. For the immature salar in Figure 8B, the com-
parable number of wells would be almost 250, spaced at ca.
600 m. Once well inefficiency, drawdown limitations, possible
barren inflows, and economics are entered into the equation,
extraction of more than 33 percent is only possible under ex-
ceptional circumstances.
Many salar boundaries are permeable, and extraction causes
inflow to replace the volumes removed. Thus, the brine body
does not exist in isolation from its surroundings, so a broad
understanding of the catchment characteristics is required in
order to establish how the brine reservoir has become estab-
lished, how it is maintained, and how it will react to future
changes as a result of pumping. A monitoring program to
measure hydrometeorological parameters, surface water and
groundwater flows, levels, and fluid concentrations is re-
quired to establish baseline conditions against which future
changes can be compared. A conceptual model of the catch-
ment hydrology is the first step. Quantification of the model
parameters in space and time will allow a water balance to be
established.
For the reserve analysis, it will be essential to build a digi-
tal flow simulation model using one of several codes available
on the market. Modeling variable density flow is a highly spe-
cialized subject, at the limit of current knowledge, and re-
quires the services of experts in this field who are not neces-
sarily familiar with resource studies, so that the integration of
disciplines becomes paramount. Clearly, a well-tested and
documented code is to be preferred. The advantages and dis-
advantages of the various models are discussed, for example,
in Spitz and Moreno (1996) and Simmons et al. (2010), and
references therein, and are not considered in depth here. Two
important points do need to be considered, however: the re-
quirements to simulate both (reactive) solute transport and
changes in brine density. The former are necessary to establish
how grade may change over time and are in widespread use,
especially for contaminant studies, but as far as we are aware
no examples exist where solute movement has been linked to
aquifer matrix dissolution, as might well be the case where di-
lute waters enter halite aquifers. Reactive solute transport is
an area that needs further research. In addition to flow
caused by head differential, variations in brine density may
also cause flow. Such variations are also problematic inas-
much as for small differentials (<0.01 g/cm
3
), two fluids are
miscible, whereas at greater differential, they are not. Thus,
while hydraulic conductivity values determined from pump-
ing tests in the field and input into the model take care of
small variations, large variations in density require two-phase
flow to be simulated. A similar situation occurs in coastal
aquifers, where sea-water intrusion is explicitly modeled
based on Ghyben-Herzberg principles, but the geometry of
this, which often simulates a fresh-water lens sitting in an
oceanic island (Bear, 1972), is different from that of a brine
body in a halite aquifer surrounded by fresh water. The latter
is rarely modeled, to our knowledge only in the mature salars
of Atacama and Hombre Muerto, although even here no
comparison with a single-phase model was ever attempted.
Finally, it should be clear from the preceding discussions
that as a result of fluid reorganization, mixing, and inflows, a
reserve estimate is not a static figure, even assuming that it
could be calculated with precision. As extraction continues,
the reserve will change, over and above (or below) the
amount extracted. Thus, while it is obviously necessary to
have a prediction for the tonnage over time, we question
whether the use of a single reserve figure is the best way to
approach the issue. Water resource engineers are accustomed
to dealing with unstable dynamics, and we consider that re-
serve updates over time may also be a requirement for brine
development.
Modifying Standards
The existing codes for mineral resource and reserve report-
ing (JORC, 2004; NI 43-101, 2005; PERC, 2008) have all
been prepared for solid phase minerals and, while they are
broadly applicable to brine resources and reserves, do not
deal with them specifically. It can be seen from the foregoing
review that there are several important differences that arise
when considering a fluid prospect.
The most important differences relate to the fluid nature of
the prospect:
1. Host aquifer property control—the resource is con-
trolled by aquifer porosity, and the reserve by permeability;
2. Brine body uniformity—because it is fluid, the brine
body is much more likely to have been homogenized during
formation, and at most a few brine types may occur relating
to different source (catchment) areas;
3. Fluid mixing—despite homogenization, dilute inflows
can penetrate deeply into the aquifer and mix with the in situ
brine to cause grade variation during production;
4. Influence from outside the claim area—during extrac-
tion, depending on the ratio of pumped to stored volume, ex-
traction may rearrange the brine body, necessitating investi-
gation beyond the claim boundaries to assess the water
balance and fluid properties throughout the catchment; and
THE EVALUATION OF BRINE PROSPECTS AND THE REQUIREMENT FOR MODIFICATIONS TO FILING STANDARDS 1237
0361-0128/98/000/000-00 $6.00 1237
5. Changes that occur over time with pumping—conse-
quent upon fluid mixing and reorganization, there is no single
resource or reserve value.
As a result, the guidelines and requirements for reporting
in the codes may in part be inappropriate. The “Contents of
the Technical Report” for NI 43-101, in particular, require
modification. For example, “Item 10, Deposit Types” does
not translate well to brine prospects; brine is not a “deposit,”
and while the host aquifer may be characterized as mature or
immature, within these broad categories aquifers form in an
almost infinite variety of volcanosedimentary deposits and
structural settings. Thus, we would prefer that this chapter be
headed “Host Aquifer” and contain all the investigations and
interpretations relating to host aquifer geology, structure, and
physical properties. “Item 11, Mineralization” also does not
translate well, implying as it does a fixed and time-invariant
orebody. We would prefer that this chapter be headed “Brine
Body” and contain information on grade, the existing geome-
try of the brine, and its relation to surrounding hydrogeology.
We also suggest that a new chapter is required in NI 43-101
to evaluate the water balance and to investigate (or predict)
potential changes to the resource during extraction. We sug-
gest that this chapter be called “Extraction Considerations.”
Some consideration of the water balance and potential grade
reorganization during extraction should be included in a re-
source statement. At the reserve level, this stage necessarily
requires the formulation, calibration, and running of a fluid
flow digital simulation model to predict grades during project
lifetimes. As previously discussed, this does not apply to a
mineral reserve as defined by the current codes, since a solid
phase mineral deposit will not change within project lifetimes.
Given the different requirements for brine prospect evalu-
ation, it becomes even more important to establish and docu-
ment protocols for data acquisition and analysis. Such proto-
cols should be included within the technical report to enable
third parties to understand what was done and to ensure that
any due diligence be properly informed.
Since the concept of the codes is based on transparency,
materiality, impartiality, and competence, we also recom-
mend that specialized hydrogeological knowledge be applied
to the investigation and reporting of these prospects.
Conclusion
Salars represent a relatively new and attractive source for
elements such as lithium, potassium, and boron. Existing op-
erations at Silver Peak, Searles Lake, Salar de Atacama, and
Hombre Muerto are among the few examples that may pro-
vide guidelines for exploration and development of brine re-
sources and reserves. Current reporting requirements are all
aimed at solid phase mineral deposits and are therefore not
ideally suited to fluid brine prospects. The resulting variation
in reporting standards leads to uncertainties and inconsisten-
cies in project evaluation, which can be confusing for analysts
and investors.
For any brine prospect, the fundamental issues for evalua-
tion are the host aquifer and the brine body. The resource is
estimated based on aquifer volume within the claim bound-
aries, its specific yield, and brine grade. A reserve, in addition
to economic and process aspects, requires estimation of
extractable grade variation as a result of in-aquifer mixing
with barren brine or fresh water.
Consequently, we believe that the current requirements for
disclosure and reporting of fluid mineral prospects need revi-
sion. It is essential, as we have shown, to be very clear which
porosity parameter has been measured, and we recommend
that the historical practice of using effective porosity to esti-
mate the resource and specific yield as the base for the re-
serve evaluation be discontinued, and instead specific yield
be used for both the resource and the reserve. We also be-
lieve that chapters headed “Host Aquifer” and “Brine Body”
should replace those entitled “Deposit Types” and “Mineral-
ization,” respectively, and that a chapter on extraction consid-
erations should be added to both resource and reserve filing
requirements for brines.
Acknowledgments
The concepts developed are largely based on the authors’
experience at most of the salars listed in Table 1, although ex-
amples in the text are restricted as a result of confidentiality
issues. The corresponding author is indebted to Richard
Seville of Orocobre Ltd for many thoughtful discussions that
helped to crystallize some of the ideas contained herein, as
well as for permission to publish data from the Salar de
Olaroz. In addition, the authors are grateful to Patrick High-
smith and Iain Scarr of Lithium One for permission to pub-
lish data relating to their properties at Hombre Muerto.
Addendum
While this paper was in press, the Ontario Securities Com-
mission issued Staff Notice 43-704, “Mineral Brine Projects
and National Instrument 43-101 Standards of Disclosure for
Mineral Projects,” dated July 27, 2011. This document pro-
vides guidance on reporting standards for brine prospects and
addresses many of the issues raised in this paper.
REFERENCES
Alonso, R.N., Jordan, T.E., Tebbutt, K.T., and Vandervoort, D.S., 1991, Giant
evaporite belts of the Neogene central Andes: Geology, v. 19, p. 401–404.
Archie, G.E., 1950, Introduction to petrophysics of reservoir rocks: American
Association of Petroleum Geologists Bulletin, v. 34, p. 943–961.
Badaut, D., and Risacher, F., 1983, Authigenic smectite on diatom frustules
in Bolivian saline lakes: Geochimica et Cosmochimica Acta, v. 47, p.
363–375.
Badon-Ghyben, W., 1888, Nota en verband met de voorgenomen putboring
nabij Amsterdam: Tijdschrift Koninklijke Instituut Ingenieurs, p. 8 –22.
Bear, J., 1972, Dynamics of fluids in porous media: Amsterdam, Elsevier, 784
p.
Bobst, A.L., Lowenstein, T.K., Jordan, T.E., Godfrey, L.V., Ku, T.L., and Luo,
S., 2001, A 106 ka paleoclimate record from drill core of the Salar de Ata-
cama, northern Chile: Palaeogeography, Palaeoclimatology, Palaeoecology,
v. 173, p. 21–42.
Casas, E., and Lowenstein, T.K., 1989, Diagenesis of saline pan halite: Com-
parisons of petrographic features of Modern, Quaternary and Permian
halites: Journal of Sedimentary Petrology, v. 59, p. 724–739.
Chong, G., Mendoza, M., García-Veigas, J., Pueyo, J.J., and Turner, P., 1999,
Evolution and geochemical signatures in a Neogene forearc evaporitic
basin: The Salar Grande (Central Andes of Chile): Palaeogeography,
Palaeoclimatology, Palaeoecology, v. 151, p. 39–54.
Cressie, N.A.C., 1993, Statistics for spatial data: New York, Wiley, 900 p.
de Marsily, G., 1986, Quantitative hydrogeology: Groundwater hydrology for
engineers: San Diego, Academic Press, 440 p.
de Silva, S.L., 1989, Geochronology and stratigraphy of the ignimbrites from
the 21
o
30’S to 23
o
30’S portion of the Central Andes of northern Chile:
Journal of Volcanology and Geothermal Research, v. 37, p. 93–131.
1238 HOUSTON ET AL.
0361-0128/98/000/000-00 $6.00 1238
de Silva, S.L., Zandt, G., Trumball, R., Viramonte, J.G., Salas, G., and
Jiménez, N., 2006, Large ignimbrite eruptions and volcano-tectonic de-
pressions in the Central Andes: A thermomechanical perspective: Geolog-
ical Society, London, Special Publication no. 269, p. 47–63.
Einsele, G., 2000, Sedimentary basins: Evolution, facies and sediment bud-
get: Berlin, Springer-Verlag, 792 p.
Ericksen, G.E., and Salas, R., 1989, Geology and resources of salars in the
Central Andes, in Ericken, G.E., Cañas Pinochet, M.T., and Reinemund,
J.A., eds., 1989, Geology of the Andes and its relation to hydrocarbon and
mineral resources: Circum-Pacific Council for Energy and Mineral Re-
sources Earth Science Series, v. 11, p. 151–172.
Evans, K., 2008, Know limits: An abundance of lithium: Industrial Minerals,
July v., p. 48–55.
——2010, Lithium’s future supply and demand: The Northern Miner, v. 96,
p. 11–12.
Fritz, C., Silva Henning, O., Suzuki, E., and Salati, E., 1978, Isotope hydrol-
ogy in Northern Chile, in Proceedings of the Symposium on Isotope Hy-
drology: Vienna, IAEA-SM-228/26, v. 2, p. 525–543.
Fritz, S.C., Baker, P.A., Lowenstein, T.K., Seltzer, G.O., Rigsby, C.A., Dwyer,
G.S., Tapia, P.M., Arnold, K.K., Ku, T.-L., and Luo, S., 2004, Hydrologic
variation during the last 170,000 years in the southern hemisphere tropics
of South America: Quaternary Research, v. 61, p. 95–104.
Ghosh, P., Garzione, C.N., and Eiler, J.M., 2006, Rapid uplift of the Altiplano
revealed through
13
C-
18
O bonds in paleosol carbonates: Science, v. 311, p.
511–515.
Gibert, R.O., Taberner, C., Sáez, A., Giralt, S., Alonso, R.N., Edwards, R.L.,
and Pueyo, J.J., 2009, Igneous origin of CO
2
in ancient and recent hot
spring waters and travertines from the northern Argentinean Andes: Jour-
nal of Sedimentary Research, v. 79, p. 554–567.
Hardie, L.A., and Eugster, H.P., 1970, The evolution of closed basin brines:
Mineralogical Society of America Special Paper no. 3, p. 273–290.
Hardie, L.A., Smoot, J.P., and Eugster, H.P., 1978, Saline lakes and their de-
posits: A sedimentological approach: International Association of Sedimen-
tologists, Special Publication no. 2, p. 41–47.
Hartley, A.J., and Chong, G., 2001, Late Pliocene age for the Atacama
Desert: Implications for the desertification of western South America: Ge-
ology, v. 30, p. 43–46.
Hartley, A.J., Chong, G., Houston, J., and Mather, A., 2005, 150 million years
of climatic stability: Evidence from the Atacama Desert, northern Chile:
Journal of the Geological Society, London, v. 162, p. 421–424.
Herzeberg, A., 1901, Die Wasserversorgung einiger Nordseebäder: Dagboek
Gasbeleucht Wasserversorg, v. 44, p. 815–819.
Houston, J., 2006, Evaporation in the Atacama Desert: An empirical study of
spatio-temporal variations and their causes: Journal of Hydrology, v. 330, p.
402–412.
Ide, F., and Kunasz, I.A., 1989, Origin of lithium in Salar de Atacama, north-
ern Chile, in Ericken, G.E., Cañas Pinochet, M.T., and Reinemund, J.A.,
eds., Geology of the Andes and its relation to hydrocarbon and mineral re-
sources: Circum-Pacific Council for Energy and Mineral Resources Earth
Science Series, v. 11, p. 165–172.
Igarzábal, A.P., 1984, Estudio geológico de los recursos mineros en salares
del NOA (Puna Argentina): Proyecto de Investigación, Consejo de Investi-
gación, Universidad Nacional de Salta, 154 p.
Johnson, A.I., Prill, R.C., and Morris, D.A., 1963, Specific yield—column
drainage and centrifuge moisture content: U.S. Geological Survey, Water
Supply Paper no. 1662-A, 74 p.
JORC, 2004, Australian code for reporting of exploration results, mineral re-
sources and ore reserves: The Joint Ore Reserves Committee of the Aus-
tralian Institute of Mining and Metallurgy, Australian Institute of Geosci-
entists, and Minerals Council of Australia, 20 p.
Kampf, S.K., Tyler, S.W., Ortiz, C.A., Muñoz, J.F., and Adkins, P.L., 2005,
Evaporation and land surface energy budget at the Salar de Atacama,
northern Chile: Journal of Hydrology, v. 310, p. 236–252.
Kasemann, S.A., Meixner, A., Erzinger, J., Viramonte, J.G., Alonso, R.N., and
Franz, G., 2004, Boron isotope composition of geothermal fluids and bo-
rate minerals from salar deposits (Central Andes, NW Argentina): Journal
of South American Earth Sciences, v. 16, p. 685–697.
Kay, S.M., Coira, B., Wörner, G., Kay, R.W., and Singer, B.S., 2010, Geo-
chemical, isotopic and single crystal
40
Ar/
39
Ar age constraints on the
evolution of the Cerro Galán ignimbrites: Bulletin of Volcanology, doi:
10.1007/s00445-010-0410-7.
Kitanidis, P.K., 1997, Introduction to geostatistics: Applications in hydroge-
ology: Cambridge, Cambridge University Press, 249 p.
Lawrence, A.R., 1977, Determination of specific yield using the centrifuge
method: Institute of Geological Sciences, London, Technical Report
WD/ST/77/7, 13 p.
Lowenstein, T., and Risacher, F., 2009, Closed basin brine evolution and the
influence of Ca-Cl inflow waters: Death Valley and Bristol Dry Lake, Cali-
fornia, Qaidam Basin, China, and Salar de Atacama, Chile: Aquatic Geo-
chemistry, v. 15, p. 71–94.
Lowenstein, T., Hein, M.C., Bobst, A.L., Jordan, T.E., Ku, T.L., and Luo, S.,
2003, An assessment of stratigraphic completeness in climate-sensitive
closed-basin lake sediments: Salar de Atacama, Chile: Journal of Sedimen-
tary Research, v. 73, p. 91–104.
Marrett, R.A., Allmendinger, R.W., Alonso, R.N., and Drake, R.E., 1994,
Late Cenozoic tectonic evolution of the Puna Plateau and adjacent fore-
land, northwestern Argentine Andes: Journal of South American Earth Sci-
ences, v. 7, p. 179–207.
NI 43-101, 2005, National instrument 43-101—standards of disclosure for
mineral projects, form 43-101F1 and companion policy 43-101CP:
Toronto, Canadian Securities Authority, 13 p.
Oncken, O., Chong, G., Franz, G., Giese, P., Götze, H., Ramos, V., Strecker,
M., and Wigger, P., 2006, The Andes: Active subduction orogeny: Berlin,
Springer, 569 p.
Orti, F., and Alonso, R.N., 2000, Gypsum-hydroboracite association in the
Sijes Formation (Miocene, NW Argentina): Implications for the genesis of
Mg-bearing borates: Journal of Sedimentary Research, v. 70, p. 664–681.
PERC, 2008, Pan-European code for reporting of exploration results, min-
eral resources and reserves: The Pan-European Reserves and Resources
Reporting Committee, 53 p.
Placzek, C., Quade, J., and Patchett, P.J., 2006, Geochronology and stratigra-
phy of late Pleistocene lake cycles on the southern Bolivian Altiplano: Im-
plications for causes of tropical climate change: Geological Society of
America Bulletin, v. 118, p. 515–532.
Reading, H.G., ed., 1996, Sedimentary environments: Processes, facies and
stratigraphy: Oxford, Blackwell Science, 688 p.
Rech, J.A., Currie, B.S., Shullenberger, E.D., Dunagan, S.T., Jordan, T.E.,
Blanco, N., Tomlinson, A., Rowe, H., and Houston, J., 2010, Evidence for the
development of the Andean rain shadow from a Neogene isotopic record in
the Atacama Desert: Earth and Planetary Science Letters, v. 292, p. 371–382.
Reijes, J., and McClay, K., 2003, The Salinas de Fraile pull-apart basin, NW
Argentina: Geological Society, London, Special Paper no. 210, p. 197–209.
Rettig, S.L., Jones, B.F., and Risacher, F., 1980, Geochemical evolution of
brines in the Salar de Uyuni, Bolivia: Chemical Geology, v. 30, p. 57–79.
Risacher, F., and Fritz, B., 1991, Quaternary geochemical evolution of the
salars of Uyuni and Coipasa, Central Altiplano, Bolivia: Chemical Geology,
v. 90, p. 211–231.
Risacher, F., Alonso, H., and Salazar, C., 2003, The origin of brines and salts
in Chilean Salars: A hydrochemical review: Earth Science Reviews, v. 63, p.
249–293.
Salfity, J.A., 1985, Lineamientos transversales al rumbo Andino en el
noroeste de Argentino: Sernageomin, Congreso Geologico Chileno, IV,
Antofagasta, Chile, Transacciones, no. 2, p. 119–137.
Simmons, C.T., Bauer-Gottwein, P., Graf, T., Kinzelbach, W., Kooi, H., Li, L.,
Post, V., Prommer, H., Therrien, R., Voss, C.I., Ward, J., and Werner, A.,
2010, Variable density groundwater flow: From modeling to applications:
International Hydrology Series, Cambridge, Cambridge University Press,
p. 87-118.
Spitz, K., and Moreno, J., 1996, A practical guide to groundwater and solute
transport modeling: New York, Wiley, 461 p.
Vandervoort, D.S., Jordan, T.E., Zeitler, P.K., and Alonso, R.N., 1995,
Chronology of internal drainage development and uplift, southern Puna
plateau, Argentine Central Andes: Geology, v. 23, p. 145–148.
Vinante, D., and Alonso, R.N., 2006, Evapofacies del Salar de Hombre
Muerto, Puna Argentina: Distribucion y genesis: Revista de la Asociación
Geológica Argentina, v. 61, p. 286–297.
Walton, W.C., 1970, Groundwater resource evaluation: New York, McGraw-
Hill, 664 p.
Warren, J., 1999, Evaporites: Their evolution and economics: Oxford, Black-
well, 438 p.
Zandt, G., Leidig, M., Chmielowski, J., Baumont, D., and Yuan, X., 2003,
Seismic detection and characterisation of the Altiplano-Puna magma body,
Central Andes: Pure and Applied Geophysics, v. 160, p. 789–807.
THE EVALUATION OF BRINE PROSPECTS AND THE REQUIREMENT FOR MODIFICATIONS TO FILING STANDARDS 1239
0361-0128/98/000/000-00 $6.00 1239
