The Origin of Brines and Salts in Chilean Salars: A Hydrochemical Review

Earth-Science Reviews 63 (2003) 249 – 293
www.elsevier.com/locate/earscirev
The origin of brines and salts in Chilean salars:

a hydrochemical review
Francßois Risacher a,*, Hugo Alonso b, Carlos Salazar c
a
CNRS-IRD, Centre de Géochimie de la Surface, 1 rue Blessig, 67000 Strasbourg, France
b
Universidad Católica del Norte, Departamento de Quı́mica, Angamos 0610, Antofagasta, Chile
c
Dirección General de Aguas, Ministerio de Obras Públicas, Morandé 248, Santiago, Chile
Received 26 June 2002; accepted 20 February 2003
Abstract
Northern Chile is characterized by a succession of north – south-trending ranges and basins occupied by numerous saline
lakes and salt crusts, collectively called salars. Fossil salt crusts are found to the west in the extremely arid Central Valley, while
active salars receiving permanent inflows fill many intravolcanic basins to the east in the semiarid Cordillera. Sea salts and
desert dust are blown eastward over the Cordillera, where they constitute an appreciable fraction of the solute load of very dilute
waters (salt content < 0.1 g/l). The weathering of volcanic rocks contributes most components to inflow waters with salt content
ranging from 0.1 to 0.6 g/l. However, the average salt content of all inflows is much higher: about 3.2 g/l. Chemical
composition, Cl/Br ratio, and 18O – 2H isotope contents point to the mixing of very dilute meteoric waters with present lake
brines for the origin of saline inflows. Ancient gypsum in deep sedimentary formations seems to be the only evaporitic mineral
recycled in present salars. Saline lakes and subsurface brines are under steady-state regime. The average residence time of
conservative components ranges from a few years to some thousands years, which indicates a permanent leakage of the brines
through bottom sediments. The infiltrating brines are recycled in the hydrologic system where they mix with dilute meteoric
waters. High heat flow is the likely driving force that moves the deep waters in this magmatic arc region. Active Chilean salars
cannot be considered as terminal lakes nor, strictly speaking, as closed basin lakes. Almost all incoming salts leave the basin
and are transported elsewhere. Moreover, the dissolution of fossil salt crusts in some active salars also carries away important
fluxes of components in percolating brines.
Evaporative concentration of inflow waters leads to sulfate-rich or calcium-rich, near-neutral brines. Alkaline brines
are almost completely lacking. The alkalinity/calcium ratio of inflow waters is lowered by the oxidation of native sulfur
(reducing alkalinity) and the deposition of eolian gypsum (increasing Ca concentration). Theoretically, SO4-rich inflow
waters and their derived SO4-rich brines should be found in the intravolcanic basins of the Cordillera because of the
ubiquity of native sulfur, while Ca-rich brines should prevail in sedimentary basins where Ca-rich minerals are abundant.
This relation is perfectly observed in the salar de Atacama, the largest in Chile. However, several salars located within
the volcanic Cordillera belong to the Ca-rich group. Inflows and brines may have acquired their Ca-rich composition in
Pleistocene time when their drainage basins were mainly sedimentary. Later on, recent lava flows and ignimbrites
covered the sedimentary formations. Underground waters may have kept their early sedimentary signature by continuous
* Corresponding author. Tel.: +33-3-90240409; fax: +33-3-88367235.

E-mail addresses: [email protected] (F. Risacher), [email protected] (H. Alonso), [email protected] (C. Salazar).
0012-8252/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0012-8252(03)00037-0
250 F. Risacher et al. / Earth-Science Reviews 63 (2003) 249–293
recycling. However, the weathering of volcanic rocks tend to slowly shift the water compositions from the Ca-rich to the
SO4-rich type.
D 2003 Elsevier Science B.V. All rights reserved.
Keywords: Northern Chile; Hydrochemistry; Closed basin; Salar; Salt recycling; Brine evolution
1. Introduction hydrogeochemical trends, focusing on the origin of

salts and brine evolution. This approach is possible
The Central Andes in Chile, Bolivia and Argentina because all these salars belong to the same geograph-
contains a large number of closed basins whose ical, geological and climatological environment in
central depression is occupied by saline lakes and salt spite of their scattering over a 200,000 km2 area.
crusts (salars). Active research has mainly been The large number of basins and water samples
focused on the world’s largest salt crust: the salar de smoothes local irregularities and provides a global
Uyuni in the Bolivian Altiplano (Ericksen et al., 1978; view of their chemical trends. Fig. 1 presents the study
Rettig et al., 1980; Risacher and Fritz, 1991a, 2000; area and the location of all salars. Each basin or salar
Fornari et al., 2001; Baker et al., 2001). Information is identified in the text by a number (from Fig. 1)
on other small salars are not easily available. They can bracketed after its name or abbreviation. Table 1 gives
be found with difficulty in many technical reports, the detailed location and the physiographic and cli-
theses and communications at local symposia and bear matic features of the basins. All minerals cited in the
mostly on economic geology and water resources of text are listed with their formula in Table 2.
the basins. Nevertheless, Stoertz and Ericksen (1974)
and Chong (1984) described in detail the features of
the salar crusts of northern Chile, but gave little 2. Analytical data
hydrochemical information. Risacher and Fritz
(1991b) studied the geochemistry of most of the small This study is largely based on the analyses of 354
intravolcanic salars of the southern Bolivian Alti- inflow waters and 227 brackish lake waters and brines
plano. The origin of solutes in a few Chilean salars of Chilean salars (Risacher et al., 1999). However,
has been investigated by Alpers and Whittemore bromine is not included in this report. Inasmuch as the
(1990), Spiro and Chong (1996), Alonso and Risacher Cl/Br ratio is an important clue to the origin of
(1996), Garcés (2000a,b) and Carmona et al. (2000). solutes, we present in Appendix A the previously
Other studies of Chilean salars deal with paleohy- unreported Br analyses together with Cl concentra-
drography and paleoclimatology (Grosjean, 1994; tion. Both components were analysed by ion chroma-
Valero-Garcés et al., 1996; Bao et al., 1999; Grosjean tography (Dionex 4000i). A discussion of the
et al., 2001), sedimentology (Valero-Garcés et al., hydrogeochemical setting of each salar can be found
1999) and water quality (Caceres et al., 1992; Aguirre in Risacher et al. (1999). For the giant salar de
and Clavejo, 2000). Atacama, we rely on the comprehensive studies of
In this paper, we present a review of the hydro- Moraga et al. (1974, 206 analyses) and Ide (1978, 389
geochemistry of almost all active salars of northern analyses).
Chile: 52 saline lakes and salt crusts are distributed
along a 1000 km north – south transect in the Andes
Cordillera. The drainage basin of an important river 3. Regional geology
and its tributaries (Rio Lauca), feeding salar de
Coipasa in the Bolivian Altiplano, has been included. Northern Chile is characterized by contractional
We do not undertake a detailed geological, minera- structures and intensive block faulting due to the
logical, sedimentological and paleolimnological study subduction of the Nazca Plate below the South Amer-
of each salar, but to review as a whole the major ican Plate. A succession of north – south-trending
F. Risacher et al. / Earth-Science Reviews 63 (2003) 249–293 251
Fig. 1. Map of northern Chile and topographic sections showing the main morphostructural units and the location of salars. The elongated and
thin formation in the Pre-Andean Depression, between salar de Atacama and Domeyko Range, is the Cordillera de la Sal.
252
Table 1
Physiographic and climatic features of closed basins in northern Chile
Basin Code No. Latitude Longitude Altitude Alt Temperature Prec Evap TDS TDS Evol. Area Area Area
(Fig. 1) (South) (West) Salar max mean (jC) (mm/ (mm/ min max path basin Salar water
(m) (m) year) year) (mg/l) (mg/l) (km2) (km2) (km2)
LAGUNAS COT 1 18j11V08 69j13V11 4495 6342 1.9 379 1070 299 2002 SO4 119 6 6
COTACOTANI
LAGO CHR 2 18j14V48 69j09V13 4530 6342 1.9 338 1230 47 1633 SO4 273 22.5 22.5
CHUNGARA
RIO LAUCA LAA 3 18j27V00 69j14V54 3892 6063 4.2 370 1200 108 784 SO4 2374 – –
F. Risacher et al. / Earth-Science Reviews 63 (2003) 249–293

SALAR DE SUR 4 18j50V03 69j02V08 4260 5780 2.7 250 1280 108 285,000 SO4 574 144 9.5
SURIRE
SALAR PIN 5 20j31V38 69j40V17 980 5785 18.5 75 1630 199 365,621 SO4 17,150 377 0
DE PINTADOS
LAGUNA LGU 6 19j55V56 68j50V40 3900 5190 4.6 150 1490 267 1276.0 SO4 129 0.2 0.15
LAGUNILLA
SALAR DEL HCO 7 20j18V18 68j50V22 3778 5190 5 150 1260 108 113,093 SO4 1572 51 2.5
HUASCO
SALAR DE COP 8 20j38V12 68j39V33 3730 5220 5 150 1300 119 330,671 SO4 1116 85 5
COPOSA
SALAR DE MIC 9 20j58V48 68j33V09 4125 5407 3.5 200 1620 94 62,662 SO4 282 2.5 0.6
MICHINCHA
SALAR DE ALC 10 21j03V07 68j29V14 4250 5407 3.5 200 1620 116 99,361 SO4 128 3.8 0.75
ALCONCHA
SALAR DE CAR 11 21j22V36 68j20V55 3690 6176 5.8 125 1630 88 335,536 SO4 – Ca 561 108 3.5
CARCOTE
SALAR DE ASC 12 21j33V01 68j16V58 3716 6023 5.8 125 1630 89 119,853 SO4 – Ca 1757 243 18
ASCOTÁN
SALAR DE ATA 13 23j25V05 68j15V57 2300 6233 14 160 1800 243 339,719 SO4 – Ca 18,100 3000 12.6
ATACAMA
SALAR DE TAR 14 23j01V59 67j16V30 4400 5816 0 150 1500 287 176,486 SO4 2035 48 14
TARA
LAGUNA HEL 15 23j05V58 67j08V15 4300 5745 0 180 1500 118 308,546 SO4 221 5.8 5.8
HELADA
SALAR DE AC1 16 23j07V22 67j24V38 4280 5370 1 150 1500 1359 122,890 Ca 281 15 2.5
AGUAS
CALIENTES 1
SALAR DE PSA 17 23j12V26 67j30V35 4500 6046 1 150 1500 239 57,578 SO4 634 18 5
PUJSA
SALAR DE LOY 18 23j14V36 67j16V55 4150 5370 1 150 1500 163 242,456 Ca 676 80 5
LOYOQUES
LAGUNA TRI 19 23j24V01 67j23V51 4290 5060 0 200 1500 10,830 10,830 SO4 59 0.4 0.3
TRINCHERA
LAGUNA MUE 20 23j26V08 67j26V28 4295 5109 0 200 1500 30,916 97,153 SO4 41 0.03 0.03
CHIVATO
MUERTO
SALAR DE AC2 21 23j28V51 67j33V11 4200 6046 1 150 1500 2530 13,656 Ca 1168 134 9
AGUAS
CALIENTES 2
LAGUNA LEJIA LEJ 22 23j29V44 67j41V28 4325 5924 1 150 1500 1273 70,723 SO4 193 1.9 1.9
LAGUNA MIS 23 23j43V19 67j45V58 4120 5910 2 180 1500 217 5308 SO4 303 13.4 13.4
MISCANTI
LAGUNA MIN 24 23j45V46 67j47V30 4120 5910 2 180 1500 3702 10,912 SO4 27.5 1.6 1.6
MIÑIQUE
SALAR DE LACO LAC 25 23j51V15 67j24V24 4250 5852 1 200 1500 636 24,433 SO4 306 16.2 2.2
LAGUNA TUY 26 23j56V22 67j35V23 4010 5852 1 180 1500 6572 131,024 SO4 245 2.9 2.9

TUYAJTO
SALAR DE AC3 27 23j59V16 67j41V33 3950 5910 1 150 1500 2491 25,150 SO4 – Ca 476 46 2.5
AGUAS
CALIENTES 3
SALAR DE CPR 28 23j57V25 67j47V28 3950 5204 1 150 1500 6589 221,804 Ca 137 27 0.9
CAPUR
SALAR DE IMI 29 24j11V06 68j47V00 2949 3995 10 40 2000 565 90,008 Ca 189 9.8 0.16
IMILAC
SALAR DE PUN 30 24j35V38 69j00V00 2945 6739 10 150 2000 196 233,367 Ca 4263 250 0.1
PUNTA
NEGRA
SALAR DE AC4 31 25j00V16 68j37V15 3665 6739 2 180 1630 851 341,759 SO4 656 19.5 2
AGUAS
CALIENTES 4
LAGUNA DE LA LAZ 32 25j04V18 68j30V33 4250 5697 1 180 1630 3965 158,854 SO4 393 7.7 3.5
AZUFRERA
SALAR DE PAJ 33 25j08V33 68j48V50 3537 5488 5 115 1350 11,728 246,674 Ca 1984 104 1.4
PAJONALES
SALAR DE GOR 34 25j24V07 68j40V27 3950 5467 1 140 1000 2846 296,737 SO4 324 27 0.5
GORBEA
SALAR IGN 35 25j29V39 68j37V09 4250 5100 2 140 1000 4043 97,091 SO4 37.5 0.7 0.002
IGNORADO
SALAR DE LA AZU 36 25j28V36 68j48V30 3580 5488 3 120 1100 548 323,473 Ca 214 3.3 0.02
AZUFRERA
SALAR DE AGUA AMA 37 25j32V56 68j49V59 3558 5488 2 120 1100 7656 196,672 Ca 863 23 0.04
AMARGA
SALAR DE AGI 38 25j48V06 68j53V24 3320 5058 2 100 1100 177,044 334,882 Ca 589 71 0.0002
AGUILAR
SALAR DE LA ISL 39 25j44V50 68j38V00 3950 5761 0 130 1000 6229 329,693 SO4 858 152 2.4
ISLA
(continued on next page)
253
254
Table 1 (continued)
Basin Code No. Latitude Longitude Altitude Alt Temperature Prec Evap TDS TDS Evol. Area Area Area
(Fig. 1) (South) (West) Salar max mean (jC) (mm/ (mm/ min max path basin Salar water
(m) (m) year) year) (mg/l) (mg/l) (km2) (km2) (km2)
SALAR DE LAS PAR 40 25j49V33 68j30V17 3987 5761 0 140 1000 8907 333,942 SO4 676 40 1.2
PARINAS
SALAR GRANDE GRA 41 25j59V52 68j42V00 3950 5664 2 130 1000 8277 129,707 Ca 867 29 0.4

SALAR DE INF 42 25j58V32 69j03V26 3520 5012 2 100 1100 1677 318,744 SO4 – Ca 293 6.7 0
INFIELES
SALAR DE LA LGN 43 26j12V29 68j58V29 3494 5373 3 120 1100 3288 20,430 SO4 400 0.55 0.4
LAGUNA
SALAR DE PED 44 26j13V42 69j08V51 3370 6127 4 125 1200 85 326,745 SO4 – Ca 3620 335 0.6
PEDERNALES
SALAR DE PIE 45 26j20V12 68j44V49 4150 5857 2 140 1000 2908 82,327 SO4 388 28 0.2
PIEDRA
PARADA
LAGUNAS BRA 46 26j18V09 68j36V24 4250 6080 2 140 1000 2134 121,494 SO4 545 10 10
BRAVAS
LAGUNAS DEL JIL 47 26j23V10 68j39V41 4150 6127 2 140 1000 2925 19,592 SO4 119 3.4 1.5
JILGUERO
LAGUNA BAY 48 26j23V09 68j35V02 4250 6127 2 140 1000 2432 4785 SO4 221 1.2 1.2
DEL BAYO
SALAR DE WHE 49 26j41V35 68j37V32 4220 6146 1 140 1000 1466 158,739 Ca 466 6.3 4
WHEELWRIGHT
LAGUNA ESC 50 26j38V51 68j31V21 4353 5918 1 140 1000 2318 32,315 SO4 194 3.8 3.5
ESCONDIDA
LAGUNA VERDE LAV 51 26j52V39 68j27V49 4350 6893 1 170 1000 935 174,581 SO4 1075 15 15
SALAR DE MAR 52 26j55V25 69j04V31 3760 6893 2 120 1100 144 331,453 SO4 – Ca 3045 145 6
MARICUNGA
LAGUNA FRA 53 27j27V47 69j13V03 4110 6052 1 200 1000 130 327,885 SO4 933 24.8 24.8
DEL NEGRO
FRANCISCO
Climatic data are taken from the Balance Hidrico de Chile (DGA, 1987). The potential evaporation (Evap) measured in class A evaporimeters has been corrected by a factor 0.65. Alt
max is the highest elevation of the drainage basin. TDS min and max are the lowest and the highest salt contents of the waters and brines, respectively, found in each basin. Basin area
includes drainage area and salar area. Salar area includes lakes, ponds and dry surfaces (crusts, playas).
Table 2 tary, metamorphic and igneous rocks. The large cop-

Minerals cited in the text and their formula per ore deposits of northern Chile are located in the
Amorpous silica SiO2 Precordillera. The Domeyko Range also contains
Antarcticite CaCl26H2O
continental evaporite formations on its eastern fringe.
Bischofite MgCl26H2O
Calcite CaCO3 The Pre-Andean Depression, also called the salar de
Gypsum CaSO42H2O Atacama Depression (not to be confused with the
Halite NaCl Atacama Desert), is a large intramontane basin, at
Mirabilite Na2SO410H2O 2500 m of elevation, filled with Tertiary to Holocene
Natron Na2CO310H2O
clastic and evaporitic sediments of continental origin.
Thenardite Na2SO4
Trona Na2CO3NaHCO32H2O It contains the largest evaporitic basin in Chile: the
Ulexite NaCaB5O98H2O salar de Atacama ([13], 3000 km2). Another important
salar, Punta Negra ([30], 250 km2), occupies the
southern edge of the basin. A prominent ridge of
ranges and basins alternate in the orogenic belt of the anticlinal folds and domes of gypsum and rock salt
Central Andes (Fig. 1). The main morphostructural borders the salar de Atacama on its western fringe: the
units are from West to East: (1) the Coast Range; (2) Cordillera de la Sal. It corresponds to an ancient salar
the Central Depression; (3) the Precordillera; (4) the of Tertiary age deformed as a consequence of tectonic
Pre-Andean Depression; (5) the Western Cordillera; events during the Cenozoic (Dingman, 1962, 1967).
(6) the Altiplano; (7) the Eastern Cordillera. The The Western Cordillera, of Miocene to Holocene age,
Coast Range is a moderately elevated mountain chain is an elevated plateau (4000 m) consisting of rhyolitic
(1500 m as an average) steeply dipping westward in ignimbrites and covered by numerous andesitic stra-
the Pacific Ocean. Marine sedimentary rocks and tovolcanoes reaching 6500 m in elevation. Many of
volcanic andesitic rocks of Mesozoic age make up these volcanoes contain native sulfur deposits, some
most of the range. The Central Depression (or Central of them of economic interest. The volcanoes often
Valley), at an altitude of 800 to 1400 m, is a thick fill delineate interior drainage basins occupied by saline
of detrital and lacustrine sediments (gravel, sand, silt, lakes and salt crusts. Most of the salars included in
clay) of middle Tertiary to Holocene age, often this study belong to this unit. The Bolivian Altiplano
covered with salt crusts remnants of ancient salars. is a major intramontane basin separating the Western
Roughly between 22j and 25jS latitude, the Central Cordillera from the Eastern Cordillera above 22jS
Depression is known as the Atacama Desert. The latitude. Its central trough is occupied by the giant
famous Chilean nitrate deposits are located in this salar de Uyuni (10,000 km2). The volcanic rocks of
depression where they impregnate many surficial the Western Cordillera are in contrast to the sedimen-
formations (Ericksen, 1981). On the east margin of tary rocks that dominate in all other units, from the
the valley, large alluvial fan systems make the tran- Coast Range to the Precordillera.
sition with the Precordillera Range. The Domeyko
Range, at an elevation of 3500– 4000 m, is the most
conspicuous part of the Precordillera because it is 4. Climate
clearly separated from the Eastern Cordillera by the
Pre-Andean Depression. The Precordillera Range is Climatic features of northern Chile are summarized
made up chiefly of Paleozoic and Mesozoic sedimen- in Table 3 (DGA, 1987). The potential evaporation
Table 3
Climatic features of northern Chile
Coast Range Central Valley Precordillera Pre-Andean Depression Western Cordillera
Precipitation < 10 < 10 10 – 25 25 – 50 50 – 300
Potential Evaporation 1000 1500 1500 1500 600 – 1200
Mean temperature 15 – 18 18 – 20 8 – 12 12 0
Precipitation and evaporation are in millimeters/year and temperature in jC.
given in Tables 1 and 3 is derived from measurements salars, from the Coast Range to the Highlands, is
in class A evaporimeters (DGA, 1987). Pan evapo- given by Stoertz and Ericksen (1974). Active Chilean
ration is usually greater than the actual evaporation salars of the Western Cordillera resemble those
from lake surface estimated by other methods (hydro- described in the Bolivian Altiplano by Risacher and
logic balance, energy budget). This is due to the Fritz (1991b). Three main types are found in northern
differences in size, heat distribution and water move- Chile as follows.
ment between large lakes and small and shallow (1) Saline lakes are permanent bodies of saline
evaporation pans. An empirical pan coefficient is waters. The water area occupies the whole salar area
generally applied to correct evaporimeter measure- (see Table 1). They range in depth from some deci-
ments. The many weaknesses of this approach has meters to some meters. Mirabilite often precipitates in
been discussed by Miller (1977, p. 266). Unfortu- winter because its solubility is markedly reduced by
nately, at most places, they are the only available data. low temperatures.
Reasonable values for pan coefficients may be (2) Playas are the most common type of salar in
extrapolated from studies by Jones (1965), Van Den- the Western Cordillera of northern Chile, as they are
burgh (1975), Yonts et al. (1973), Smith (1979) and in the Bolivian Altiplano. Shallow ( < 20 cm) and
Stanhill (2002). Most pan coefficients range from 0.6 ephemeral saline pools occupy variable areas of
to 0.8. In northern Chile, Grilli et al. (1986) selected a wide mudflats across the central depression of the
single 0.65 correction term, which will be used in this basin. Some meters below the playa surface, a
study. confined aquifer is filled with an interstitial brine,
A salient feature of the climate is the increase in whose capillary draw precipitates a variety of salts
precipitation from west to east across this region. The within the sediments (gypsum, mirabilite, halite,
coast range and the Central Depression are among the ulexite).
driest deserts of the world, in contrast with the (3) Salt crusts of two kinds are found in this region.
semiarid Western Cordillera where precipitation rea- Active salt crusts are presently forming from evapo-
ches 300 mm/year. In the whole area, however, the rating brines that have reached saturation with respect
potential evaporation greatly exceeds the precipita- to gypsum, mirabilite or halite. These crusts are of low
tion, which is a basic condition for the establishment thickness, rarely above 50 cm, and are found in the
of evaporitic environments (Hardie and Eugster, semiarid Western Cordillera. Fossil salt crusts are
1970; Eugster and Hardie, 1978). In the northern located in more arid areas, from the Coast Range to
area, 10% to 20% of the precipitation between 4000 the western fringe of the Western Cordillera. These
and 5000 m occur as snows. In contrast, in the crusts are of similar mineralogy to the active crusts,
southern area, snow constitutes as much as 50% to but are much thicker. The halite crust of salar de
80% of the precipitation at the same elevation (Vuille, Atacama is several hundred meters thick (Bevacqua,
1996). The mean annual temperature decreases from 1991; Bobst et al., 2001). However, fossil crusts are
north to south, which allows the snow blanket to more commonly of metric to decametric size. A 10-m-
persist for a longer time in the southern basins. The thick halite crust of Neogene age was described in the
dominant wind direction is from ocean to cordillera. Coast Range by Chong et al. (1999). Deep dissolution
Thus, sea salts and saline dust from the Central pits (halite, Pajonales [33]), outcrops above the salar
Valley are transported eastward over the Highlands level (thenardite, Gorbea [34]), tilted layers (gypsum,
by the wind. Pedernales [44]) are evidence of fossil, nonactive, salt
crusts.
Many salars are combinations of these types,
5. Main outlines of Chilean salars especially salt crusts and playas. Generally, salars in
the Coast Range, in the Central Valley and in the
5.1. Salar morphology Precordillera are fossil salt crusts, while those in the
Western Cordillera are predominantly playa lakes.
A comprehensive description and many photo- Mixed salars are found in the Central Depression
graphs of the surficial features of active and fossil and in the western fringe of the Western Cordillera,
close to the Precordillera. In this study, we have et al., 1984; Hastenrath and Kutzbach, 1985; Syl-
incorporated the largest salar of the Central Valley, vestre et al., 1999; Baker et al., 2001; Fornari et al.,
salar de Pintados, because its underground waters are 2001). It is likely that the neighbouring Chilean
actively fed by inflows originating in the Western basins were also filled with saline lakes during the
Cordillera. However, the salar itself is a large fossil same period. Fossil salt crusts stem from the desic-
salt crust. cation of such old saline lakes.
5.2. Inflows
6. Inputs
Four types of inflows feed Chilean salars are as
follows: 6.1. Atmospheric contribution
(1) Focused discharge through springs from the
drainage area with an observable flow rate generally Two fresh snows and seven late-lying snows were
between 0.1 and 100 l/s. Many springs emerge far sampled and analysed for major and minor compo-
above the salar surface. nents. The meltwater from one old snow has a salt
(2) Diffuse seeps percolate slowly to the salar content as high as 23 mg/l due essentially to Ca and
surface. Their flow rate is not noticeable, but their SO4, while the salt content of the six others ranges
lateral extension may reach tens or even hundreds of from 1.0 to 6.6 mg/l (mean: 3.6 mg/l). In compar-
meters along the salar shore. Total discharge through ison, meltwaters from fresh snows have salt contents
diffuse seeps may be significant but is impossible to of 3.1 and 3.5 mg/l (mean: 3.3 mg/l). Therefore, six
quantify. The slow flow rate may induce an incipient out of seven old snows did not receive significant
evaporation of the waters at the seepage face. dry fallout after their deposition, which suggests that
(3) Rivers are generally of short length and low these snows were probably sampled shortly after
discharge, generally below 500 l/s. They often in- their fall. Only one late-lying snow was markedly
filtrate in colluvium before reaching the salars, contaminated by eolian gypsum. It is therefore
where they eventually discharge through springs and possible to obtain a broad estimation of the flux of
seeps. each component brought by snowfalls in the basins
(4) Underground waters fill aquifers in the drain- by multiplying the volume of precipitation over each
age basins. They discharge in the central trough of the basin by the average concentration of the melt-
basins through springs and seeps. It was often possible waters.
to sample these waters in exploration and exploitation
wells drilled by mining companies. 6.2. Surface and groundwater inputs
5.3. Former lakes On the other hand, the flux of components entering
permanent lakes through springs, seeps, rivers and
Since Miocene time, the climate have remained underground discharges can also be estimated. The
arid or semiarid in northern Chile (Kött et al., 1995; hydrologic balance of a saline lake without surface or
Pueyo et al., 2001). Many basins have been occu- underground outflow is given by (Hutchinson, 1957;
pied by saline lakes (Stoertz and Ericksen, 1974; Langbein, 1961):
Geyh et al., 1999). Late Neogene lacustrine forma-
tions are known in the Central Valley (Sáez et al., Vinflows ¼ S ðE PÞ ð1Þ
1999; May et al., 1999) and in the Lauca Basin in
the Western Cordillera (Gaupp et al., 1999). Gros- where Vinflows = basin discharge to the lake (m3/year),
jean (1994) and Grosjean et al. (1995) described a S = surface of the lake (m2), E = actual evaporation (m/
late Pleistocene lacustrine phase in Laguna Lejia year) and P = precipitation (m/year).
[22]. Several large and deep saline lakes have Water loss from Chilean saline lakes through
repeatedly covered the Bolivian Altiplano during infiltration is generally small compared to basin dis-
Pleistocene time (Servant and Fontes, 1978; Lavenu charge (see discussion below on salt balance in
Section 8.2). The flux of components brought in the distinguished: lakes fed by dilute inflows whose salt
lake by all inflow waters is obtained by multiplying content is lower than 500 mg/l (CHR [2], HEL [15],
their average concentration by the volume of inflows. COT [1], LGU [6]) and lakes fed by inflows of higher
The uncertainty on the basin discharge depends on salt content: 500 –6500 mg/l (MIS [23], LEJ [22],
that of the climatic parameters E and P and on the FRA [53], BRA [46], ESC [50], LAV [51], MIN [24],
unweighted average concentration of inflow waters, TUY [26]). These two groups of waters differ also
except when the lake is obviously fed by one major significantly in relative ionic concentrations.
input.
Figs. 2 and 3 compare the fluxes of 11 components 6.3. Lakes fed by dilute inflows
brought by inflows to 12 permanent lakes to their
fluxes deposited by snows on their drainage basins. 6.3.1. Major components in dilute inflows
Lakes are ordered according to the increasing average The flux of major components brought by dilute
salt content of their inflows. Two groups can be inflows in lakes CHR [2], HEL [15], COT [1], LGU
Fig. 2. Fluxes of major components deposited by snows on the drainage basins (stars) of 12 permanent lakes compared to the fluxes brought into
the lakes by inflow waters (squares). Focusing our attention on the SO4 diagram, we can see that the flux of SO4 brought by dilute inflows is
only slightly more elevated than that brought by snows on the drainage basins (for LGU [6], the proportion is even reversed). In contrast,
concentrated inflows bring in the lakes significantly more SO4 than snows. The other major components show similar behavior although it is not
so easily noticeable because of the vertical logarithmic scale.
[6], are only 1 to 14 times higher than those deposited pheric inputs deposited on the land surface should
by the snows (Fig. 2). In LGU basin, all fluxes are then be lower than the flux brought from the basin by
even slightly higher in snows than in inflow waters. inflow waters. A lower flux from the basin could be
This means that for LGU, as well as for other lakes due to the removal of atmospheric components in
fed by dilute waters, atmospheric deposition could soils. We must also lay stress on the large uncertainty
represent a major, or the major, contribution to the on flux estimates.
lakes. This seems particularly clear for SO4 in very
dilute waters, which is probably derived from eolian 6.3.2. Minor components in dilute inflows
sulfates (gypsum and thenardite) and oxidized ele- With the only exception of As in Laguna Helada
mental sulfur eroded from the top of the volcanoes. (HEL [15]), the flux of minor components (Br, I,
Dissolved components in very dilute groundwaters NO3, As) deposited by the snows are 1 to 70 times
stem from rock alteration and atmospheric inputs that higher than those brought into the lakes by dilute
have been brought from the land surface to the inflows (Fig. 3: CHR [2], HEL [15], COT [1], LGU
groundwater by percolating waters. The flux of atmo- [6]), which suggests that most of the minor compo-
Fig. 3. Fluxes of minor components deposited by snows on the drainage basins (stars) of 12 permanent lakes compared to the fluxes brought into
the lakes by inflow waters (squares). Except for LGU, most of Si is brought in the lakes by inflow waters. In contrast, the fluxes of Br, I, NO3
and As deposited by the snows are higher than those brought in the lakes by dilute inflows. Not so for concentrated inflows, which bring in the
lakes more minor components than snows, with the exception of NO3.
nents in dilute waters are probably brought by 7. Origin of salts in inflow waters
precipitation. Br and I may be derived from sea
salts, and NO3 and I may be derived from surficial 7.1. Salt content of inflow waters
formations of the Central Valley. The removal of
atmospheric components in soils and plants could be The most conspicuous feature of inflow waters in
very efficient for minor components such as NO3 closed basins of northern Chile is their very broad
and I. Mineral deposits and geothermal areas are the range of concentration: from 47 to 79,740 mg/l with an
primary sources of arsenic in arid regions (Smedley average of 3230 mg/l. Northern inflows are globally
and Kinniburgh, 2002). Owing to the high aridity, As less saline (average = 2.4 g/l) than southern inflows
is not leached from the surficial formations and (average = 4.9 g/l). In comparison, White et al. (1963,
constitutes a noticeable component of desert dust, 1980) reported salt contents of underground waters in
which makes it a major health hazard in northern andesitic to rhyolitic aquifers of North America rang-
Chile (Smith et al., 1998; Queirolo et al., 2000a,b). ing from 70 to 473 mg/l. Alteration modeling of a
Its concentration in meltwaters ranges from 0.3 to 68 standard andesite (code Kindisp, Made et al., 1994)
Ag/l (mean: 18 Ag/l), compared to the world baseline leads to a maximum concentration of 340 mg/l of
concentration in rain and snow of 0.01 to 0.5 Ag/l dissolved solids. Thus, the alteration of volcanic rocks
(Smedley and Kinniburgh, 2002). In addition, copper does not seem to be the main controlling factor of the
smelting plants in northern Chile permanently release high concentration of many inflow waters.
large quantities of As in the atmosphere (Gidhagen The concentration of most components in inflow
et al., 2002). As for silica, in all basins but LGU [6], waters increases with the total salt content (Figs. 4 and
the flux brought by inflow waters is markedly higher 5). The correlation line between component concen-
(10 to 75 times) than the flux deposited by the tration and salt content shows similar slopes for
snows. Silica is likely to originate from the dissolu- inflows and lakes (Na, Cl, Li>1 and K, Mg, SO4,
tion of volcanic glass. Ca < 1). The concentration increase in lake waters is
easily explained by evaporative concentration, a
6.4. Lakes fed by concentrated inflows mechanism that cannot significantly account for most
inflow waters. The rate of evaporation of groundwater
Except for NO3, the fluxes of all components is much lower than in an open water and decreases
deposited by the snows are small in comparison with drastically with depth (Ripple et al., 1972; Hellwig,
those brought by concentrated inflows: approximately 1973; Miller, 1977, p. 304; Yechieli and Wood, 2002).
3 to 800 times lower. NO3 seems to be the only One may reasonably neglect evaporation of deep
component, among those studied, that originates in all groundwater within the unsaturated zone. However,
waters mostly from atmospheric deposition. Br, I, Si shallow groundwater ( < 3 to 5 m below surface) is
and As in these waters must come from some addi- subject to interstitial evaporation, especially when the
tional source. capillary fringe reaches the surface. A conspicuous
In the above discussion, we considered only the feature of this process is the deposition of efflorescent
flux of components brought in the basins by snowfall. salts on soils and plants around the salars. Therefore,
However, during the much longer dry season, dry evaporative concentration may increase the salt con-
fallout is also deposited on the drainage basins. tent of shallow groundwater at the margin of the
Therefore, the fluxes of all components brought by salars. However, many deep waters and springs with
atmospheric inputs are very likely to be higher than high discharges also have elevated salt content that
those estimated only from the snows. This strengthens cannot be related to direct evaporative concentration.
the conclusion that the lowest the concentration of an The simplest hypothesis is that dilute waters have
inflow water is, the higher is the contribution of mixed with underground brines of composition sim-
atmospheric salts to its salt load. Atmospheric salt is ilar to the present lake brines. The proportion of most
the main source of dissolved components in very components of the deep brines would be nearly
dilute waters. As concentration increases, other inputs conserved by dilution with freshwaters, which could
become progressively more significant. explain the similar slopes observed in Fig. 4.
Fig. 4. Concentration of major components in inflow waters and lake brines of Chilean salars as a function of total dissolved solids (TDS). Note
the overall correspondence between the slopes of the correlation lines of inflows and lakes.
Fig. 5. Concentration of minor components in inflow waters and lake brines of Chilean salars as a function of total dissolved solids (TDS).
7.2. Mixing with underground brines line could be due to the dissolution of a mixture of
sodium chloride with calcium and/or magnesium
Sodium + calcium ( 2) + magnesium ( 2) is chloride in dilute underground waters. Ancient evap-
plotted against chloride in Fig. 6. Dilute inflows orites containing halite and gypsum are present below
(Cl < 10 mmol/l) present a large scatter in values, the volcanic formations. However, calcium and mag-
while saline inflows (Cl>10 mmol/l) are gathered nesium chloride are very soluble salts (antarcticite:
along the isoconcentration line Cl = Na + 2Ca + 2Mg. 1650 g/kg [H2O] and bischofite: 1170 g/kg [H2O])
Rock alteration is likely to be the main control of the known to form either in extremely cold environments
composition of dilute inflows. North American waters or in very concentrated brines. They are unlikely to be
from andesitic rocks have compositions similar to the present in depth at relatively elevated temperature. A
dilute waters. This diagram is alike to those represent- more probable explanation is the mixing of dilute
ing any common evaporative system as illustrated in waters with Na –Ca – Mg/Cl underground brines sim-
the inset diagram showing a similar relation for ilar to those present in saline lakes as shown in the
evaporated lake waters. However, all waters repre- inset diagram in Fig. 6.
sented in the main diagram are inflow waters: ground-
waters, springs, seeps and rivers. Only seeps and 7.3. Mixing with subsurface brines
shallow groundwaters may have undergone some
evaporation just before they emerge. Nevertheless, Mixing of underground water with brines at the
seeps and springs present the same behavior, which margin of a salar follows the Ghyben – Herzberg
suggests that the evaporative trend shown in the main principle, which defines the geometry of a freshwater
diagram could have been inherited from ancient lens of low density lying on a large saltwater body
evaporated brines buried below the volcanic forma- (Heath, 1983). The typical case is a fresh groundwater
tions. Alternatively, the shift to the isoconcentration lens in an island. The Ghyben– Herzberg principle has
Fig. 6. Sodium + calcium ( 2) + magnesium ( 2) vs. chloride concentration for inflow waters (main diagram) and for lake brines (inset
diagram) of Chilean salars. North American waters from andesitic rocks are included for comparison. Both diagrams show the same overall
pattern, which suggests that inflow waters stem from the mixing of freshwaters with brines similar to those present in the salars.
received little attention in the case of playas and saline salar surface also have high salt content. Thus, mixing
lakes (Yechieli and Wood, 2002). It is difficult to at the margins is not likely to be the main process of
apply to salars in northern Chile. Firstly, the overall the high concentration of inflow waters.
pattern is inverted: brine lenses overlie freshwater
bodies. This is a basically unstable situation: the 7.4. Geochemistry of bromide
denser brine lens tends to move downward. The
Ghyben – Herzberg mixing zone slopes away from a Chloride-to-bromide ratios shown in Fig. 7 also
saline lake (see Yechieli and Wood, 2002), in contrast suggest that the composition of saline inflow waters
with a freshwater lens overlying a saline layer (see (Cl>10 mmol/l) results from the mixing of snow
Heath, 1983, p. 68). It is not easy to extrapolate very meltwaters with brines similar to those found in the
much in depth the interface between fresh and salt- present lakes. Two mixing curves between average
water below a saline lake. The geometry of the lens snow meltwater and lake brines are drawn: curve A
can be reasonably described only close to the surface, corresponds to the lake brine of lowest Cl/Br ratio and
near the zone of discharge. Secondly, the Ghyben – curve B to that of highest Cl/Br ratio. Any other
Herzberg principle applies only to an isotropic aquifer, mixing curve is constrained between these two end
which is not the case of volcanic formations with member curves. All saline inflows plot in this mixing
fracture porosity. Water and brine may be forced to area. By contrast, several dilute inflows (Cl < 10
move along unpredictable pathways. Thirdly, thermal mmol/l) plot outside the mixing area. Br content of
heating below the basin may considerably disturb the halite associated to Tertiary gypsum diapirs in the
interface between the water bodies. Anyway, brine neighbouring Bolivian Altiplano is around 2 ppm
mixing occurs at the margin of the salars and could be (Risacher and Fritz, 2000). Similar low values (1– 2
partially responsible of the elevated salt content of the ppm) have been found by Pueyo et al. (2001) in
seeps. However, many springs emerging far above the neogene salt rocks of northern Chile. The Cl/Br ratio
Fig. 7. Cl/Br ratio as a function of Cl for snows, inflow waters and lake brines. Two end member mixing curves between snow meltwater and
lake brines are drawn: Curve A corresponds to the lowest Cl/Br ratio (alc-9) and curve B corresponds to the highest ratio (isl-12). All other
possible mixing curves are located between these two curves. Curve C shows the Cl/Br ratio of the solution produced by dissolving halite in
snow meltwater. Curve C is well outside the field of saline inflows.
obtained by dissolving such low Br content halite in condensation does not counterbalance the evaporation
pure water is figured by an horizontal dashed line far in this arid region. Moser and Stichler (1980) carried
above all inflow points. Curve C represents the Cl/Br out experimental evaporation of snow samples which
ratio of the solution obtained by dissolving halite in led to a similar shift due to d2H and d18O increase.
average snow meltwater. The halite endpoint may be Inflow waters are clearly distributed along a line
put anywhere on the horizontal dashed line for Cl joining the lightest isotopic values to the heavily
values between 355 and 46,500 mg/l, which is the enriched ones observed in the lakes. The isotopic
highest Cl content measured in saline inflows. The values of inflow waters follow geographic and
dissolution curve C if well above all saline inflow hydrologic trends. (1) Northern inflows ( < 24j lat-
points. Thus, the high salt content of inflows cannot itude) have lower isotopic values than southern
stem from the dissolution of ancient salt rock over- inflows; only two northern inflows have high iso-
lapped by volcanic formations. topic values similar to those found in the south. (2)
Seeps are more enriched in d2H and d18O. The shift
7.5. Stable isotopes of inflow waters to the right of the meteoric line is
due to the conjunction of several processes reflected
The d2H and d18O values for 89 inflow waters, 6 by the two correlations. Snow prevails over rain and
lakes (Chaffaut, 1998) and 26 snows (Peña, 1989) are remains for a longer time in the south than in the
shown in Fig. 8. The meteoric water line for northern north of the study area (see Section 4). As a result,
Chile is taken from Aravena et al. (1999). The lakes the evaporation of the snows and their related iso-
were sampled in the central part of the studied area topic enrichment observed in the north by Peña
( f 23jS latitude) and the snows in the northernmost (1989) is likely to be significantly enhanced in the
sector ( f 18j). Most of the snows show a marked south. Seep waters are subjected to evaporation by
displacement to the right of the meteoric line. Accord- capillary draw before they emerge and by direct
ing to Peña (1989), the shift is due to sublimation and evaporation as they percolate very slowly to the land
evaporation of the snow blanket during the day. Night surface. There is no surprise if these waters have
Fig. 8. 18O and 2H contents in snows, inflow waters, and some lake brines of Chilean salars. The isotopic composition of the most concentrated
inflow of salar de Surire [4] that results from mixing of groundwater with brine is shown. The size of lake symbols (open circles) is not related to
their salt content.
Fig. 9. Ionic activity products of calcite, amorphous silica and gypsum as a function of total dissolved solids (TDS). Square brackets refer to
activities. Horizontal lines show the solubility product at 0 and 25 jC. A water is supersaturated with respect to a phase when its representative
point plots above the solubility product line. Below the line, the water is undersaturated with respect to the phase. Many inflow waters are
oversaturated, or close to saturation, with respect to calcite and amorphous silica. These two minerals seem to exert a main control on Si, Ca and
carbonates concentration of most inflow waters. As TDS increases, inflow waters tend to reach saturation with respect to gypsum. Eventually,
gypsum is saturated, or very close to saturation, in 3 springs and 10 seeps, which suggests gypsum dissolution, at least in spring waters.
high d2H and d18O values. In contrast, underground is obviously not due to evaporative concentration of
waters, springs and rivers (of short length in the the spring water, but to the mixing with a little volume
study area) do not undergo significant evaporation of hypersaline brine, which is confirmed by the d2H
before sampling, but also present a significative and d18O values plotting almost on the meteoric water
isotopic enrichment, particularly in the south. The line.
higher salt content of many of the southern waters
associated with heavier isotopic contents suggests 7.6. Mineral dissolution
rather a mixing process than an evaporative concen-
tration. Dilute and concentrated endpoints are snow Fig. 9 shows the ionic activity product of three
meltwaters and brines, respectively, similar to those important minerals involved in the salt load of inflow
found in present saline lakes. waters and in their evaporative evolution: calcite,
The relation between salt content and isotopic gypsum and amorphous silica. Solute and water
enrichment suggested in Fig. 8 is partially obscured activities are calculated with the ion interaction model
by the respective salt contents of the two endpoints. (Pitzer, 1979; Harvie et al., 1984). H4SiO4 activity and
For example, the water resulting from the mixing of a amorphous silica solubility are taken from Marshall
little volume of hypersaline brine with a large volume (1980), Marshall and Warakomski (1980) and Mar-
of dilute water has an isotopic composition close to shall and Chen (1982). The software used is EQL/
that of the dilute water. In contrast, the salt content of EVP (Risacher and Clement, 2001 and complete
the mixed water may be markedly higher than that of database references herein).
the dilute water. The volume of saline water is Most of the inflow waters are saturated, or close
negligible, but not the mass of dissolved salts. A to saturation, with respect to calcite and amorphous
typical example is that of the most saline inflow silica. Therefore, the solubility of amorphous silica
(17.7 g/l) to the salar de Surire [4]. The spring controls the concentration of silica in inflow waters.
emerges within the salar after passing through a thick Calcite is very likely an alteration mineral that
section of brine-filled sediments. The high salt content controls, at least partially, the concentration of Ca
Fig. 10. Temperature (jC) of springs and underground waters and mean annual air temperature as a function of elevation above sea level. Waters
that plot above the curve of mean annual air temperature + 5 jC present a thermal influence.
and HCO3 in inflow waters. Most inflow waters are could supply sufficient quantities of Ca and SO4 to
undersaturated with respect to gypsum. However, as reach saturation with gypsum.
concentration increases, gypsum becomes close to
saturation in 10 seeps and 3 springs. Concerning the 7.7. Thermal waters
seeps, it is possible that their Ca and SO4 concen-
tration has increased by evaporation. However, this Northern Chile is a region of high heat flow due to
cannot be the case for springs. For these waters, the recently active volcanism (Springer and Förster,
dissolution of gypsum is the simplest way to reach 1998). Geothermal systems, such as that of El Tatio
saturation. There are no other common minerals that (outside the study area), have been considered for
Fig. 11. Lithium and boron content of springs and underground waters [in wt.% of total dissolved solids (TDS)] as a function of temperature. No
correlation with temperature is observed.
energy production (Lahsen, 1988). Thermal influence has been found. Fig. 11 shows the concentration of Li
can only be estimated in springs and underground and B (in wt.% of the salt content) as a function of
waters. The temperature of seeps and rivers is con- water temperature. Li and B are generally enriched in
trolled by the atmosphere. The temperature of springs thermal waters (Berthold and Baker, 1976; White et
and underground waters is compared in Fig. 10 to the al., 1976). Nevertheless, in northern Chile, such an
mean annual air temperature (Grilli et al., 1986; DGA, enrichment is not observed. Therefore, hydrothermal
1987). If we reasonably assume that a water whose alteration does not seem to be a major process of
temperature is at least 5 jC higher than the mean solute acquisition.
annual air temperature presents a thermal influence, Fig. 12 shows the spatial distribution of inflows
then thermal heating has affected 78% of all sampled differentiated according to the difference between their
springs and underground waters. However, the ther- temperature and that of the atmosphere (Twater –Tatm;
mal influence is moderate: only 10 waters have seeps and rivers are not considered). In the northern
temperatures between 40 and 53 jC, which are the and central area (18 – 24jS latitude), both thermal and
highest values observed in the study area. No corre- nonthermal waters are intermingled. In the south (24 –
lation between temperature and solute concentration 27jS latitude), however, a clear separation can be
observed. Nonthermal springs are located westward
close to the boundary between the sedimentary Pre-
cordillera and the volcanic Western Cordillera. In
contrast, heated inflows are found preferentially east-
ward well within the volcanic Cordillera. This obser-
vation is in good agreement with the increase of heat
flow from the Coast Range to the active magmatic arc
of the Western Cordillera reported by Springer and
Förster (1998).
8. Origin of deep brines
8.1. Recycling of ancient salars
Volcanic formations in northern Chile range in age

from Miocene to Pleistocene. During this entire time
span, the climate remained arid or semiarid (Kött et
al., 1995; Pueyo et al., 2001). Many salars must have
occupied ancient intravolcanic basins. Subsequently,
lava flows and ignimbrite sheets covered these saline
deposits. The buried brines may have mixed with
dilute meteoric inflows to produce brackish and saline
underground waters. However, this straightforward
explanation for the origin of deep brines comes up
against two difficulties. First, the brines must have
been preserved during the hot volcanic events. Ele-
vated temperatures may vaporize the lake waters.
Only underground brines protected by overlying sedi-
Fig. 12. Distribution of thermal springs and underground waters in ments could have been preserved. Secondly, these
salar basins of northern Chile. Note the exaggerated horizontal
scale. In the south, thermal inflows are more abundant eastward entrapped brines must have been rapidly exhausted as
within the Cordillera. A similar relation may also be suspected in they discharge with fresh inflows. The time of
the northwestern part of the study area. exhaustion depends of the size of the buried salars
and of their leaching rate by dilute waters. We may the lake brine. A steady state is attained at very high
intend a rough calculation of the exhaustion time concentration when the equilibrium water vapor pres-
based on present salars. Their wide range of morpho- sure exerted by the solution equals the mean relative
logical and chemical types make it difficult to define humidity of the atmosphere (Kinsman, 1976). In con-
an average salar. On the other hand, extreme values trast, if infiltration occurs through the lakebed, even at
of salar size and water composition may not be very a very slow rate, then a steady state is attained when
significative. Therefore, we have selected a broad the input flux balances the leakage flux of the con-
range of values for each parameter that covers a large servative component.
number of salars of different type: salar surface = 10 The difficulty is to determine if the steady state is
to 150 km2; lake surface = 1 to 15 km2; salt content of attained. A recent climatic, geologic or anthropic
inflows = 0.5 to 5 g/l; salt content of brines = 20 to event may have abruptly modified the lake composi-
200 g/l; evaporation = 1.33 m/year; precipitation, which could be undetectably reequilibrating
tion = 0.16 m/year. The only parameter that cannot through time with the new climatic or geologic con-
be easily estimated is the volume of underground ditions. One way is to compare the total mass of the
brines. We will assume a thickness of 100 m of brine- component in the lake to its annual input flux. The
filled sediments with a porosity of 40%. The mass of ratio between total mass and annual input has the
dissolved salts in deep brines of a buried salar can be dimension of a time. If this time is substantially lower
estimated by multiplying the salar surface by the than the age of the last disturbing event, then it is
sediment thickness, the porosity and the brine con- referred to as the residence time of the component in
centration. The flux of dissolved salts brought each the lake. A short residence time indicates a high
year in a present salar is obtained by multiplying the infiltration rate. The total mass of a component in
volume of inflows [see above, Eq. (1)] by their salt the lake is:
content. By dividing these two values, we obtain an
Mlake ¼ Slake Hlake Clake ð2Þ
exhaustion time ranging from 100 years (a small salar
rapidly leached, generating concentrated inflow where Slake is the surface of the lake, Hlake is its mean
waters) to 1.7 million years (a large salar slowly depth and Clake is the average concentration of the
leached, generating dilute waters). Small- and selected conservative component in the lake. The
medium-sized salars have been leached in a short annual input flux is [see Eq. (1)]:
time as compared with the age of volcanic formations
of Central Andes (Miocene to Holocene). Only large Minflows ¼ Slake ðE PÞ Cinflows ð3Þ
buried salars could still provide some salts in present where E is the actual evaporation, P is the precip-
inflow waters. itation and Cinflows is the average concentration of
the conservative component in inflows. The ratio
8.2. Brine leakage from present salars between the total mass in the lake and the annual
input flux is:
Brine leakage from present salars is another source
of salts in underground waters. The concentration of a Mlake =Minflows ¼ Hlake =ðE PÞ Clake =Cinflows ð4Þ
conservative component (Li, Br, Cl before halite
saturation) in a saline lake is controlled by the Table 4 presents this ratio for Cl in some character-
infiltration rate of the lake brine through bottom istic Chilean saline lakes. Most lakes belong to the
sediments. Water leakage as a controlling factor of Alconcha [10] – Lagunilla [6] types with ratios ranging
lake chemistry has already been stressed by Carmouze from a few years to some hundreds. Laguna Helada
and Pedro (1977), Alderman (1983), Sanford and [15] has the highest estimated ratio: 3370 years. The
Wood (1991) and Dutkiewicz et al. (2000). If a last drastic climatic change occurred 10,000 years ago
lakebed is completely impermeable, then the concen- when large lakes occupying the basins dried up (see
tration of a conservative component, and conse- Section 5.3). It seems therefore reasonable to assume
quently the salt content of the lake, will continuously that all these ratios represent the residence time of Cl
increase, which in turn reduces the evaporation rate of and that Andean saline lakes are under steady-state
Table 4
Rough estimates of residence times and infiltration rates of some typical north Chilean saline lakes
HEL [15] FRA [53] ALC [10] MIS [23] LGU [6]
Lake surface (km2) 5.8 14.3 0.75 13.4 0.15
Mean lake depth (m) 0.50 0.20 0.20 4.5 0.15
Mean inflow concentration (Cl: mmol/l) 0.64 2.0 0.80 1.8 21
Lake concentration (Cl: mmol/l) 5050 4810 1030 34 161
Residence time [years, Eq. (4)] 3370 411 220 73 1
Infiltration rate [mm/year, Eq. (5)] 0.15 0.5 0.90 62 150
Loss through infiltration (106 l/year) 0.86 7.0 0.68 830 23
Basin discharge [106 l/year, Eq. (1)] 6800 16,700 880 15,700 180
Percent infiltration relative to basin discharge 0.01 0.04 0.08 5.3 13
Cl is the conservative component used for the calculations. Lake levels may change during the year. Cl concentration corresponds to the lake
size given in this table. FRA surface is different from that given in Table 1. Actually, FRA is constituted of two separated lakes: one dilute lake
that acts as a preconcentration pool and a hypersaline lake which is considered here. MIS mean depth is taken from Valero-Garcés et al. (1996)
and FRA mean depth from Behnke (1987). In both cases, we have halved the maximum depth given in these studies to obtain an approximate
mean value inasmuch as the surface of both lakes reported by these authors was similar to that observed during our field trips. All other mean
depths were estimated by ourselves. The main uncertainty is the mean Cl concentration of inflows, difficult to estimate. Underground discharges
and ephemeral streams during the wet season are not taken into account.
regime as well as most subsurface brines. The salt brine (Imilac [29], Pajonales [33], Gorbea [34], Agua
balance for a conservative component (Cl, Li, Br) is Amarga [37], Aguilar [38], Pedernales [44]). These
then given by: thick crusts are quite different from those of centi-
metric to decimetric size forming by evaporation in
Vinflows Cinflows ¼ Vleakage Clake ð5Þ other salars (see Section 5.1). Numerous deep brine
ponds suggest that the thick crusts are undergoing
where C is the average concentration of the compo- active dissolution. They do not seem to be presently
nent and V is the annual flux of solutions. Eq. (5) forming. Under former wetter climate, deep salt
allows to estimate leakage rates from saline lakes. lakes occupied the basins (see Section 5.3). An
Table 4 gives infiltration rates of characteristic Chil- abrupt drying-up due to a climatic change induced
ean saline lakes. The volume of lake brine lost by the rapid evaporation of the lake and the precipita-
infiltration ranges from 0.01% to 12.8% of the basin tion of a thick salt crust. The residual lake is often
discharge. Therefore, it is generally justified to limited to a subsurface, intracrustal brine saturated
neglect the leakage in the water balance of a saline with respect to halite. The intracrustal brine infil-
lake although it is the main control of the salt trates through bottom sediments removing large
balance of the lake. Almost all dissolved salts enter- quantities of dissolved salts. In contrast, inputs to
ing the lake are lost by leakage. Only low soluble the salar are fresh or brackish waters largely under-
salts (calcite, gypsum, Mg silicates) are removed saturated with respect to halite, which tends to dilute
from the solution and stored in lake sediments. Their the intracrustal brine. In order to maintain saturation
mass is very small compared to the total input flux with respect to halite, the crust must dissolve. As
to the lake. long as the salt crust remains in the salar, the flux of
infiltrating salts is higher than the flux of incoming
8.3. Dissolution of present salt crusts salts, which induces a slow dissolution process
responsible for the numerous ponds pitting the crust.
Most of the incoming solutes are lost by infiltra- A new steady-state regime will be established only
tion, which precludes the formation of thick salt when the whole crust has vanished. In summary,
crusts. However, some salars, particularly those ancient salt crusts in present Andean salars are not
located close to the Precordillera, contain several likely to be in equilibrium with their climatic and
meter thick halite crusts filled with an interstitial hydrologic environment. They are an additional
source of salts in underground waters close to the mary, when a mineral precipitates, the concentration
Precordillera. of all its components cannot increase simultaneously
in the solution because the ionic activity product of
8.4. The salt cycle in Chilean salars the mineral must be kept constant and equal to its
solubility product. For example, in the case of a
Salts do not accumulate in most Andean salars. mineral containing one anion and one cation, if the
The salt cycle is only delayed. The delay is the anion activity increases in the solution, then the cation
residence time of the dissolved components. Most activity must decrease. The solution becomes enriched
incoming salts only pass through the saline forma- in some components and depleted in others, according
tions, where they undergo a concentration process and to the precipitated minerals and to the concentration
are further recycled in the hydrologic system. Infil- ratio of their components at the start of the precip-
trating brines are not necessarily recycled in the same itation.
basin. The intense and complex fracturing of volcanic Two approaches may be used to predict the fate
rocks could favor deep circulation from one basin to of an inflow water upon evaporation. The quantita-
the others. This is a continuous mechanism that tive approach calculates step by step the chemical
accounts well for the high salt content of inflow composition of the solution and the sequence and
waters over long time spans. Thermal convection is quantities of minerals produced during evaporation.
the likely driving force that mixes, moves and heats In this study, we use the EQL/EVP code (Risacher
the underground waters in this area of active volcan- and Clement, 2001) based on the ion interaction
ism. There is no need of a large salt reservoir below model (Pitzer, 1979; Harvie et al., 1984). The
the volcanic structures. Nevertheless, giant salt bodies qualitative approach only intends to predict which
are present in depth. Alonso et al. (1991) estimate at pathway a water will follow, without caring about
10,000 km3 the volume of deep halite in Central the precise chemical composition of the solution.
Andes. The Bolivian Altiplano is dotted with gypsum The main difficulty lies in that many analyses found
diapirs. The Cordillera de la Sal in northern Chile in the literature, and all of this study, report total
plunges to the east below ignimbrites and lava flows. alkalinity (Alk) as a routinely measured parameter.
Gypsum recycling through volcanic activity in the The first chemical divide produced by the early
Western Cordillera has been reported by Risacher and precipitation of calcite would apparently necessitate
Alonso (2001). Presently, only gypsum seems to to calculate the respective concentrations of CO32
supply part of Ca and SO4 to the salars. Owing to and HCO3. The first model established by Hardie
its high solubility, the top of halite bodies formerly in and Eugster (1970) was grounded on carbonate
contact with underground waters could have been speciation, neglecting ion association and using the
leached a long time ago and redeposited in the giant Debye – Hückel equation for ion activities. The
salars of Uyuni and Atacama. Only low soluble [Ca2 +] vs. [CO32 ] concentration relation is a rather
gypsum is still in contact with subsurface waters complex expression because [HCO3] and [OH]
and constitutes a present source of Ca and SO4. must also be taken into account. In addition, each
analysis must be recomputed so that the water is in
equilibrium with the atmospheric pCO2. It is possi-
9. Brine evolution ble to update these calculations by using the ion
interaction model and the EQL/EVP code. However,
9.1. The model this would be a useless repetition of the quantitative
approach. A much simpler and rigorous treatment
Waters undergoing evaporative concentration pre- directly based on the alkalinity concept has been
cipitate a sequence of minerals in order of increasing developed by Al-Droubi et al., (1980) for the first
solubility. Hardie and Eugster (1970) and Eugster and chemical divides in which carbonate and silicate
Hardie (1978) introduced the basic concept of chem- species are involved. We will summarize this
ical divide to account for the evolution of solute approach which does not require any computer code
concentration when a water evaporates. In brief sum- (see also Risacher and Fritz, 1991b).
The total alkalinity of a natural water can be If at the start of evaporation Alk0>2[Ca2 +]0, then the
expressed as (Dickson, 1981; Stumm and Morgan, right-hand side of Eq. (12) is positive and increases
1996): linearly with the concentration factor. Thus, as evap-
oration proceeds, the difference between alkalinity

Alk ¼ 2½CO2
3 þ ½HCO3 þ ½OH þ ½BðOHÞ4 and calcium concentration will continuously increase
½Hþ ð6Þ on behalf of alkalinity. Conversely, if Alk0 < 2[Ca2 +]0,
then calcium concentration will increasingly predom-
where square brackets refer to total concentrations in inate over alkalinity. Inasmuch as carbonate species
mol/l or mmol/l. This formulation corresponds to the are the main contributors to alkalinity in most natural
operational definition of alkalinity, but is difficult to waters, the behavior of alkalinity reflects that of
handle because all terms are interdependent. By carbonates. Therefore, the first chemical divide of a
combining Eq. (6) with the electro-neutrality equa- water composition upon evaporation can be very
tion, we obtain the alternative definition of alkalinity: easily and rigorously determined by examining the
simple ratio of the analytical concentrations Alk0/
Alk ¼ ½Naþ þ ½Kþ þ 2½Ca2þ þ 2½Mg2þ ½Cl 2[Ca2 +]0 of the initial inflow water.
2½SO2 Mg carbonate or Mg silicates also precipitate at an
4 ð7Þ
early stage of water evolution. The effect of copreci-
Alkalinity is the difference between the sum of the pitation of calcite and Mg salts on the evolution of an
conjugate cations of the strong bases and the sum of evaporating water may be determined by rewriting Eq.
the conjugate anions of the strong acids (Stumm and (7) as:
Morgan, 1996, p. 165). Calcite is the first mineral to
precipitate from most natural solutions undergoing Alk 2½Ca2þ 2½Mg2þ
evaporative concentration. We may rewrite Eq. (7) as: ¼ ½Naþ þ ½Kþ ½Cl 2½SO2
4 ð13Þ
Alk 2½Ca2þ ¼ ½Naþ þ ½Kþ þ 2½Mg2þ ½Cl which leads straightforwardly to:
2½SO2
4 ð8Þ
Alk 2½Ca2þ 2½Mg2þ
As long as calcite is the only precipitate, the concen-
tration of each component of the right-hand side of ¼ F ðAlk0 2½Ca2þ 0 2½Mg2þ 0 Þ ð14Þ
Eq. (8) increases linearly with the concentration factor
F of the evaporating solution: If Alk0>2[Ca2 +]0 + 2[Mg2 +]0, then the difference
between Alk and 2[Ca2 +] + 2[Mg2 +] is positive and
½Naþ ¼ F ½Naþ 0 ð9Þ continuously increasing. The evaporating water
evolves into an alkaline brine depleted in Ca and Mg
where and enriched in carbonates. Conversely, if Alk0 <
2[Ca2 +]0 + 2[Mg2 +]0, then the water follows a neutral
F ¼ ðH2 OÞ0 =ðH2 OÞ ð10Þ path depleted in carbonate. If Alk0>2[Ca2 +]0 and
Alk0 < 2[Ca2 +] 0 + 2[Mg2 +]0, then the evaporating
(H2O) is the mole number of water in the solution. water first follows the alkaline path as long as calcite
Subscript ‘‘0’’ refers to the initial solution at the start is the sole precipitate and reverts to the neutral path
of evaporation. Eq. (8) becomes: when Mg carbonate or Mg silicates reach saturation. It
must be emphasized that either Mg carbonate or Mg
Alk 2½Ca2þ ¼ F ð½Naþ 0 þ ½Kþ 0 þ 2½Mg2þ 0 silicates have the same effect upon the chemical divide.
½Cl 0 2½SO2
4 0 Þ ð11Þ Fig. 13 shows some initial evolutionary paths of north
Chilean inflow waters in a diagram Alk vs. 2[Ca].
which leads to: Waters following the neutral path become progres-
sively enriched in Ca and, in most cases, reach
Alk 2½Ca2þ ¼ F ðAlk0 2½Ca2þ 0 Þ ð12Þ saturation with gypsum (CaSO 42H2O). Gypsum
Fig. 13. Examples of evaporation paths of north Chilean inflow waters illustrating the qualitative approach (analyses in Risacher et al., 1999:
1 = asc-14; 2 = mic-1; 3 = ped-9; 4 = sur-26; 5 = sur-7). Solid circles are directly plotted from water analyses. Quantitative evaporation paths are
computed with the EQL/EVP evaporation program based on the ion interaction model and a rigorous treatment of the aqueous species. At the
beginning of evaporation, all waters remain undersaturated with respect to calcite. Ca and Alk behave conservatively and each water follows a
straight line path whose equation is 2[Ca]/Alk = 2[Ca]0/Alk0. 2[Ca]0 and Alk0 are the coordinates of the starting points (solid circles 1 to 5). The
precipitation of calcite causes an abrupt deviation of the linear evaporation path depending on which side of the iso-concentration line
Alk = 2[Ca] the starting point stands. The precipitation of magnesite in path 5 reverses the alkaline path to the neutral path because
Alk0 < [Ca]0 + 2[Mg]0 in inflow water 5. The precipitation of Mg salts also occurs in paths 3 and 4, but does not reverse the alkaline paths
inasmuch as Alk0>2[Ca]0 + 2[Mg]0 in these waters. The pH of the initial solution may be modified by CO2 exchange with the atmosphere or by
CO2 uptake by aquatic plants. The result is a displacement of the calcite appearance point along the initial conservative straight line, which does
not change the evolutionary trend.
induces a new divide of the neutral path producing Actually, Na and Cl are the main components of
either Ca-rich/SO4-poor waters or Ca-poor/SO4-rich most brines, but they do not induce a chemical divide
waters, according to their Ca/SO4 ratio at the begin- before halite saturation, which occurs at high concen-
ning of gypsum saturation (not at the beginning of tration once the major evolutionary trends of the brine
evaporation). have already been defined. Throughout the first cru-
The whole discussion is summarized in the flow cial steps of evaporative concentration, Na and Cl
sheet of Fig. 14 which accounts for most natural behave as conservative components. For the sake of
waters (modified from Hardie and Eugster, 1970). A brevity, waters, where SO4>Ca, will be referred to as
few minerals play a prominent part in the water SO4-rich, and those where Ca>SO4 will be referred to
evolution. The outcome of the processes is the gen- as Ca-rich.
eration of three main groups of brines: alkaline (Na/ The composition of a brine is strongly dependent
HCO3 –CO3 –Cl), sulfate-rich (Na/SO4 – Cl) and cal- on that of the initial dilute water. A small variation of
cium-rich (Na –Ca/Cl) brines. The flow sheet of Fig. the ratio of the critical components in the dilute water
14 defines four pathways: alkaline (I – IA); sulfate- {Alk0/2[Ca]0, Alk0/(2[Ca]0 + 2[Mg]0)} may cause
alkaline (I – IIA –III); sulfate-neutral (II – III) and cal- drastic changes in the final brine composition. In turn,
cic (II – IV). Pathways I– IIA –IV and II –(Na/CO3 – the composition of dilute inflow waters is mainly
Cl) have not been observed in nature. inherited from rock alteration. Therefore, there is an
Fig. 14. Flow diagram for evaporative concentration of dilute waters. Path numbers are those of Hardie and Eugster (1970): alkaline (I – IA);
sulfate-alkaline (I – IIA – III); sulfate-neutral (II – III) and calcic (II – IV). The dashed line shows a rarely observed direct path from the Mg salts
divide to the sulfate-rich brine group, bypassing the gypsum divide. It has no particular incidence for water evolution and is included in our
discussion in the sulfate-alkaline path I + IIA + III.
overall relation between lithology and brine compo- provide SO42 but also acidify the solution through
sition. Each broad type of lithology is reflected in the oxidation reactions such as:
composition of the dilute water which is further
amplified by evaporative concentration. In a pioneer- Sj þ 3=2O2 þ H2 O ! 2Hþ þ SO2
4 ð15Þ
ing study, Garrels and Mackenzie (1967) first threw
light on the (nonobvious) relation between igneous The alkalinity of the solution may be drastically
rocks and alkaline brines. High-purity igneous rocks reduced. Therefore, inflow waters draining moder-
are almost devoid of anionic components (Cl , ately mineralized igneous rocks tend to follow the
SO42 ). The only available anion that can balance sulfate alkaline evaporative pathway, while highly
the cations derived from the weathering of pure mineralized igneous rocks produce waters following
igneous rocks is HCO3, which stems from atmos- the sulfate-neutral pathway. Extreme cases of very
pheric CO2 and is brought by the rains as carbonic intense acidification leading to acid-sulfate brines
acid. However, igneous rocks often contain mineral- have been reported in Australia (McArthur et al.,
ization or Cl – SO4-rich fluid inclusions which lowers 1991; Long et al., 1992) and Chile (Risacher et al.,
the proportion of HCO3 in the weathering solution. 2002). Sedimentary rocks are often characterized by
Moreover, sulfide minerals and native sulfur not only high Ca (limestone, marl, dolomite, gypsum). Waters
in contact with such rocks generally tend to generate 9.3. Water evolution in Chilean salars
calcic brines.
9.3.1. The calcite divide
9.2. Discrepancies between model and observed brine As shown in Fig. 9, most inflow waters are close to
evolution saturation, or oversaturated, with respect to calcite.
The relations Alk vs. 2Ca and Alk vs. 2Ca + 2Mg for
The model relating lithology, inflow waters and inflow waters are shown in Fig. 15. At the start of
brine composition sets the geochemical and hydro- evaporation (upper diagram), 238 inflow waters
chemical baseline of evaporative concentration. It (67%) should follow the initial neutral pathway (II),
assumes direct evaporation of the inflow water as it while 117 (33%) should follow the opposite initial
would occur in an artificial pool protected from any alkaline pathway (I). The precipitation of Mg salts
contamination. Actually, several processes may disturb (lower diagram), probably Mg smectites as in Boli-
this scenario. The evaporating water often interacts vian salars (Badaut and Risacher, 1983), theoretically
with eolian dust, sediments or salts, which modifies its shifts 65 out of 117 waters from the initial alkaline
composition and may change its theoretical evolu- pathway to the sulfate-alkaline pathway (I – IIA),
tionary path. For example, the behavior of potassium remaining 52 waters (15%) that should end up as
is mostly controlled by exchange reactions on clay carbonate brines (I – IA). In contrast to the prediction,
minerals (Eugster and Jones, 1979). The content of in the whole study area, only one small brackish pond
SO4 may be lowered by bacterial reduction or freezing located in the drainage basin of laguna Helada [15] is
out of Na sulfate salts (Eugster and Hardie, 1978). Gac actually of alkaline composition (HEL [15] in the
(1980) carried out evaporation experiments of the lower diagram). All other 226 lakes and ponds contain
main tributary of Lake Chad with and without allow- low-alkaline or neutral waters of the Na/Cl– SO4 or
ing contact with atmospheric dust. The behavior of Mg Na – Ca/Cl groups including the main laguna Helada
was very different in each experiment. In Bolivian itself.
Andes, native sulfur is eroded from the top of the The discrepancy has already been observed and
volcanoes and deposited on the drainage basins where discussed in the neighbouring Bolivian salars and
its oxidation acidifies inflow waters enough to reverse explained by acidification due to the oxidation of
their initial alkaline evolutionary path to the sulfate- eolian sulfur in the drainage basins (Risacher and
alkaline or sulfate-neutral path (Risacher and Fritz, Fritz, 1991b). It is very likely that the same process
1991b). These interactions are most efficient at an occurs in northern Chile, where many volcanoes also
early stage of evaporative evolution, when inflow contain similar native sulfur deposits. However, the
waters are still dilute. Brackish and saline inflows proportion of inflow waters that should end up as
are much less sensitive to them. Another disturbing carbonate brines in Bolivian Andes is much higher
mechanism is the mixing of several inflows with than in northern Chile: 52% and 15%, respectively.
different evolutionary paths. Mixing in large basins Moreover, 6 Bolivian salars out of 30 are actually
of complex lithology is a more likely scenario than in alkaline. Therefore, an additional process must be
small monolithologic basins. Leakage through bottom involved to explain the lower proportion of alkaline
sediments has also been reported to significantly waters and brines in Chilean Andes. One of the most
modify the brine evolution (Sanford and Wood, likely is dry fallout of desert dust from the Central
1991). All these well-documented processes take place Valley on the Chilean Andes. The importance of
after discharge of inflow waters in the basin, while atmospheric inputs in the composition of very dilute
they undergo evaporative concentration. In this paper, waters has already been emphasized (see Section 6).
we describe a (apparently) new discrepancy between Gypsum is one of the main components of desert and
lithology and evolutionary paths of inflow waters. atmospheric dust. Its dissolution increases the con-
Several Chilean salars located in an exclusive volcanic centration of Ca of the inflow waters without modify-
environment are fed by inflow waters following the ing their alkalinity, which may shift the water
calcic path typical of sedimentary rocks. Their brines evolution from the alkaline path to the neutral or to
do not reflect the geological context. the sulfate-alkaline paths.
Fig. 15. 2Ca and 2Ca + 2Mg concentration vs. alkalinity for inflow waters in Chilean salars. Lake brines values are also plotted on the
2Ca + 2Mg diagram. To avoid redundancy, waters with Alk < 2Ca are not plotted in the lower diagram. Inset flow diagrams show the
evaporative concentration paths followed by the group of waters differentiated by the isoconcentration lines. There is only one lake brine (HEL)
in the alkaline-rich field although 52 inflow waters should end up after evaporation in the alkaline-rich field.
9.3.2. The gypsum divide precipitation of calcite reverses this ratio allowing
The upper diagram of Fig. 16 shows the Ca and the solution to become sulfate-rich. Two typical
SO4 contents of all inflows and lake waters. Lake evolutionary paths are drawn: one (line A) leading
waters are differentiated according to their Ca/SO4 to a Na/Cl– SO4 brine (but starting in the Ca>SO4
ratios: 143 are sulfate-rich and 82 Ca-rich waters. In area) and the other (line B) ending in a Na – Ca / Cl
contrast, inflow waters are differentiated according to brine. Owing to the large number of inflows (354), it
their evolutionary path calculated with the EQL/EVP is not possible to show the theoretical evolutionary
code (Risacher and Clement, 2001). Many inflow path of each of them and to compare it to the actual
waters that end up as sulfate brines have initial composition of the lake or pond it feeds. Actual
concentrations higher in Ca than in SO4. The early evolutionary paths after the gypsum divide are in
Fig. 16. SO4 and SO4 + Alk/2 concentration vs. Ca for inflow waters and lake brines in Chilean salars. Inset flow diagrams show the evaporative
concentration paths followed by each group of waters. The evolutionary path of each inflow water has been determined with the EQL/EVP
evaporation program. In contrast, lakes are simply differentiated according to their Ca/SO4 ratio. Many waters following the SO4-rich path are
actually in the Ca>SO4 field. The early precipitation of calcite lowers the Ca concentration and reverses the Ca/SO4 ratio. Line A shows such an
evolutionary path. Line B shows a typical Ca-rich path.
general agreement with the calculated paths. We occurs at rather high Ca and SO4 concentrations
could not detect reversion from one path to the once their ratio can no longer be easily modified.
other, as observed after the Mg salts divide when Nevertheless, the mixing of brackish waters of differ-
waters reverse from the alkaline path to the sulfate- ent Ca/SO4 ratio may produce unpredictable end
alkaline path due to eolian sulfur. The gypsum divide brines.
In total, 243 inflow waters follow either the Salar de Atacama, which is the largest in Chile,
sulfate-alkaline or the sulfate-neutral path which ends contains both Ca-rich and SO4-rich brines within the
up as SO4-rich solutions in 143 lakes and ponds, halite nucleus (Risacher and Alonso, 1996). The
while 60 inflows follow the calcic path leading to western part of the nucleus, close to the Cordillera
Ca-rich solutions feeding 82 lakes and ponds (Fig. de Domeyko where sedimentary rocks are predom-
16). In addition to the 243 inflows following the inant, is filled with Ca-rich brines, while the eastern
sulfate paths, 52 inflows that theoretically should part close to volcanic formations of the Cordillera is
follow the alkaline path up to carbonate brines also filled with SO4-rich brines. The Ca vs. SO4 diagram
end up as sulfate-rich solutions after reverting to the shows very clearly the gypsum divide and the relation
sulfate-alkaline path (Fig. 15). All the 225 lakes and between lithology and brine composition (Fig. 17;
ponds (143 + 82) are distributed in 52 salars. Table 1 data from Moraga et al., 1974; Ide, 1978).
shows to which chemical group each salar belongs. As can be seen in the upper diagram of Fig. 16, it is
Approximately two-thirds of the salars (34 out of 52, not possible to use the initial Ca/SO4 ratio of the
SO4 in Table 1) contain only brackish and saline inflow water to predict which evaporative path will be
waters high in SO4 and low in Ca. Twelve salars followed after the gypsum divide. In contrast, as
belong to the Ca-enriched group (Ca) and seven shown in the lower diagram, Ca vs. SO4 + Alk/2, the
contain both Ca-rich and SO4-rich solutions. Several isoconcentration line Ca = SO4 + Alk/2 separates per-
salars of the Ca-rich group have a few SO4-rich fectly the two groups of inflow waters. Those with
brackish ponds. However, their most concentrated initial Ca>SO4 + Alk/2 follow the Ca-rich path, and
brines are Ca-rich. The symmetric situation is not those with SO4 + Alk/2>Ca follow the SO4-rich path.
observed in SO4-rich salars, where Ca-rich brines are This is obvious if Ca>SO4 + Alk/2: whatever the mass
completely lacking. of CaCO3 and CaSO4 precipitated, Ca will remain in
Fig. 17. SO4 vs. Ca diagram illustrating the gypsum divide in salar de Atacama brines. (These data are not included in Fig. 16.) Points are
distributed along a tilted T. The ascending branch (+) corresponds to waters undersaturated with respect to gypsum. Waters saturated with
gypsum plot along the approximate gypsum saturation line and are divided into two groups: SO4-rich brines (circles) above the isoconcentration
line Ca = SO4, and Ca-rich brines (squares) below. The geograpical distribution of the three groups of solutions is shown in the Atacama outline
map. Analyses are from Moraga et al. (1974) for the SO4 – Ca diagram and from Ide (1978) for brine distribution in the salt nucleus. Note that
the approximate gypsum saturation line has no rigorous chemical meaning, inasmuch as the true saturation line can only be drawn in an activity
diagram.
Table 5 water. This happens in confined aquifers where the

Theoretical evaporative paths and final brine composition predicted water is not in contact with the atmosphere and
from the initial composition of inflow waters
therefore does not evaporate. In evaporitic environ-
Concentrations Evaporative path Brine group
ments, the CO2 dissolved in waters is close to equi-
in inflow water
librium with that of the atmosphere. Thus, the ratio
Alk>2Ca and alkaline (I – IA) Na/CO3 – Cl
(SO4 + Alk/2)/Ca of the initial water may be used with
Alk>2Ca + 2Mg
Alk>2Ca and sulfate-alkaline Na/SO4 – Cl good confidence to predict which of the Ca-rich or
Alk < 2Ca + 2Mg (I – IIA – III) SO4-rich path will be followed by an evaporating
Alk < 2Ca and sulfate-neutral Na/SO4 – Cl water after the gypsum divide. Table 5 summarizes
SO4 + Alk/2>Ca (II – III) the relations between the initial composition of inflow
Alk < 2Ca and calcic (II – IV) Na – Ca/Cl
waters and their evaporative path.
SO4 + Alk/2 < Ca
Path numbers as in Fig. 14. Concentrations are in mol/l or mmol/l.
9.3.3. Relation between evolutionary path and
temperature
excess. However, this rule cannot be rigorously dem- We have plotted in Fig. 18 the ratio (SO4 + Alk/2)/
onstrated in the case SO4 + Alk/2>Ca. If the precip- Ca of all springs and underground waters against the
itation of calcite is not occurring and if Ca>SO4, then difference between water temperature and mean
the Ca-rich path could be followed. The most com- annual air temperature (seeps and rivers are not
mon way to prevent, or to delay, the precipitation of considered). Positive Y-axis values (log scale) corre-
calcite is to increase the content of CO2 gas in the spond to the SO4-rich path and negative values
Fig. 18. Diagram showing the relation between temperature and the two main evolutionary paths SO4-rich and Ca-rich for springs and
underground waters. Waters are differentiated according to their salt content. Most hot inflow waters follow the SO4-rich path. Note that the
region labeled the ‘‘SO4-rich path’’ include the waters that have reversed their theoretical evolutionary path from alkaline-rich to sulfate-rich
after probable acidification by sulfur oxidation.
correspond to the Ca-rich path. Almost all high

thermal inflow waters (Twater Tatm>20 jC) belong
to the SO4-rich group. Only 3 out of 22 hot inflow
waters follow the Ca-rich path. It must be emphasized
that the evolutionary path is the only parameter for
which some relation with temperature has been
detected so far. None of the dissolved components
presents any relation with temperature (see Section
7.7).
9.3.4. Discrepancy between lithology and brine

evolution
The geographical distribution of SO4-rich and Ca-
rich salars is shown in Fig. 19. According to the
previous discussion, salars located in volcanic basins
are expected to contain SO4-rich brines, while those
surrounded by sedimentary rocks should be filled with
Ca-rich brines, as perfectly illustrated in salar de
Atacama [13]. In addition to Atacama, two other
salars are located in the Pre-Andean Depression.
where sedimentary formations of the Precordillera
predominate: Imilac [29] and Punta Negra [30]. Both
belong to the Ca-rich type. All other salars, except
Pintados [5], are in volcanic basins of the Western
Cordillera and are expected to be of the SO4-rich type
as a result of the abundance of native sulfur in most of
the volcanoes. This is effectively observed in the
north of the study area (18 – 21jS latitude). Salar de
Pintados [5], in the Central Valley, is mainly fed by
inflows originating in the volcanic Western Cordillera.
In the central area (21 –24jS latitude), both types are
intermingled. No simple geological argument can
explain the difference from one basin to the other.
In the south of the study area (25 – 27jS latitude),
salars are predominently of the Ca-rich type close to
the Precordillera and of the SO4-rich type eastward in
the volcanic Cordillera.
The chemistry of Ca-rich salars located in the
volcanic Western Cordillera seems to be controlled
by sedimentary formations. If some of them are
adjacent to the Precordillera (Pedernales [44], Mar-
icunga [52]), others are well inside the Western
Fig. 19. Outline map of north Chilean Andes showing the chemistry
of each salar: SO4-rich, Ca-rich or both. In the south, Ca-rich salars
predominate close to the Precordillera, while SO4-rich salars are
found eastward within the Cordillera (compare with Fig. 12). Note
the slightly exaggerated horizontal scale.
Cordillera, up to 50 km from its western boundary observed that many salars in the Western Cordillera
(Aguas Calientes 2 [21]; Wheelwright [49]). It is have been tilted downward to the northwest during the
difficult to conceive a major contribution of sedimen- Quaternary. Salares de Pedernales [44] and Maricunga
tary rocks, either from the west or from depth, in [52] are precisely among the most asymmetric ones,
salars located in such massive volcanic formations. A which points to an older age than that of nontilted
much better explanation is the recycling of present salars within the Western Cordillera. The ancient Ca-
infiltrating brines from the salars themselves (see rich underground waters may have been continuously
Section 7). However, such an explanation only pushes recycled up to now.
the problem further back in time. When did the waters The alteration of volcanic rocks progressively
acquire their Ca-rich composition? The drainage shifts the chemical type of the solutions. Both Peder-
basins of the Ca-rich salars, now mostly or exclu- nales and Maricunga belong to the mixed Ca/SO4
sively volcanic, must have been sedimentary in the salar group. Inflow waters result from the mixing in
past, before the establishment of recent Pleistocene variable proportion of recycled brines with dilute
volcanoes and ignimbrite sheets. Consequently, these weathering solutions. The weathering of sulfur-rich
salars should be notably old, of Pleistocene age or volcanic rocks increases the SO4 content, and then the
even earlier. Stoertz and Ericksen (1974) have already SO4/Ca ratio, of inflow waters. Alternatively, if the
Fig. 20. Very schematic representation of the recycling and mixing processes responsible for the chemical shift from Ca-rich to SO4-rich brines
in Chilean salars. For the sake of simplicity, we have restricted the cycle to one basin. Actually, deep water circulation may transport infiltrating
salts in several distant basins. Size of symbols only reflects the proportion of Ca and SO4 in each solution. The salar brine is of the Ca-rich type,
which means that the concentration of Ca is markedly higher than that of sulfate (as suggested by the thick bold characters used for Ca). The
infiltrated brine has the same composition. Heat flow recycles this Ca-rich brine along the Ghyben – Herzberg mixing zone, in contact with fresh
waters that have leached volcanic rocks of the drainage basin. Sulfur-rich volcanoes produce SO4-rich waters which mix with Ca-rich recycled
brines. Thus, the Ca/SO4 ratio of the mixed solution is lower than that of the salar brine (spring 1). The gap between Ca and SO4 concentration is
reduced. The weathering of sulfur-poor volcanoes produces alkaline waters. Calcite is a common weathering product whose precipitation lowers
the Ca concentration of the water. Here again, the mixing of Ca-poor waters with Ca-rich infiltrated brines reduces the gap between Ca and SO4
concentration in the resulting inflow water (spring 2). Therefore, on the one hand, inflows tend to reduce the excess of Ca over SO4 in the salar
brine; on the other hand, the evaporative concentration amplifies the difference between Ca and SO4 in the brine. A long time is likely to be
necessary to shift the Ca-rich brine composition to a SO4-rich composition.
sulfur content of volcanic rocks is low, then the underground waters. Thermal heating may have dif-
weathering solutions are enriched in HCO3 (see Sec- ferent results below the salar and below the moun-
tion 9.1), which also reduces the Ca concentration of tains. Below the salar, heating could tend to oppose
the mixed inflow waters due to calcite precipitation. the downward brine flux and favor recycling in the
Here again, the SO4/Ca ratio of the inflow water same salar. In contrast, below the volcanic formations,
increases. However, the evaporative concentration in the heating could increase the loss of water from the
the basin amplifies the difference on behalf of the basin to other basins (see Fig. 20). Ancient evaporites,
most concentrated component (Ca). Thus, the SO4/Ca especially halite and gypsum, are present in depth
ratio of the infiltrating brine is much lower (more Ca) below the volcanic formations. However, the Cl/Br
than that of the input water. This amplification mech- ratio of halite is much higher than that of inflow
anism may delay for a long time the shift from the Ca- waters, which precludes the dissolution of deep salt
rich to the SO4-rich type. Eventually, SO4 overtakes rock in underground waters. On the contrary, the Cl/
Ca in spring waters and the amplification process then Br ratio confirms the mixing of present lake brines
favors the rapid increase of SO4 in lake brines. Fig. 20 with snow meltwaters as the origin of saline inflows.
summarizes the whole cycle. Only ancient gypsum seems to contribute some Ca
and SO4 to inflow waters. Stable isotopes also sup-
ports the mixing hypothesis: the d2H and d18O values
10. Concluding remarks of inflow waters plot along a straight line, probably a
mixing line, joining snow meltwaters to lake brines.
Three main sources of salts have been detected in Chilean salars in the Andean Cordillera cannot be
inflow waters of Chilean salars. Atmospheric inputs, strictly considered as closed basins or terminal lakes.
through precipitation and dry fallout, affect the whole Most of the salts entering the salars are lost by
area but are noticeable only in a few salars fed by very infiltration. Only poorly soluble salts, such as calcite,
dilute inflows [below 100 mg/l total dissolved solids Mg salts and gypsum, remove dissolved components
(TDS)]. Most of the minor components in very dilute which slowly accumulate in saline sediments. The
inflows, and a significant part of the major ones, stem high concentration of many inflows does not reflect an
from sea salts and desert dust, especially NO3, Br and intense material transfer by weathering, evaporite
As. Volcanic rock alteration is the main source of salts dissolution or hydrothermal alteration. Nor does it
in inflow waters with concentration ranging roughly mean a high accumulation rate of salts in the basins.
from 100 to 600 mg/l TDS. However, the average salt Actually, true material transfer is mostly restricted to
content of all inflows (3230 mg/l) is much higher than atmospheric inputs and volcanic rock weathering.
that derived from the mere alteration of volcanic This is intuitively understandable: if all inflowing
rocks. Brine recycling is the main source of salts in salts would accumulate in Andean salars, very thick
most salars. Saline lakes are under steady-state salt crusts should be observed. Thick modern salt
regime. The residence time of conservative compo- crusts are not found in this region.
nents ranges from a few years to some thousands According to the model of brine evolution ob-
years. Most of the dissolved salts brought in the lakes served in most saline lakes worldwide, three main
by inflow waters are lost by infiltration through groups of brines may be generated by evaporation and
bottom sediments and return to the hydrologic cycle. chemical divide (Hardie and Eugster, 1970): alkaline
The infiltrating brines provide the major part of brines (Na/HCO3 – CO3 – Cl, pH>9, traces of Ca and
dissolved salts to underground waters. The dissolution Mg), sulfate-rich brines (Na/SO4 –Cl, low Ca, pH < 9)
of surficial salt crusts also contributes to the salt load and Ca-rich brines (Na – Ca/Cl, low SO4, pH < 9). In
of deep waters. Owing to the complexity of water northern Chile, alkaline salars are almost completely
circulation in fractured volcanic formations, infiltrat- lacking. Only one brackish pond out of 227 was found
ing brines may be transported to adjacent basins, or to be alkaline. All others belong to the SO4-rich and
even further. Thermal convection due to the high heat Ca-rich groups. This paucity is due to the abundance
flow is the likely driving force that moves the waters. of native sulfur in the Western Cordillera and to
Thermal heating affects about 78% of springs and atmospheric deposition of gypsum-rich desert dust
from the Central Valley. The oxidation of sulfur A complex correlation between lithology, thermal-
produces SO4 and acidifies the inflow waters, reduc- ism and salar chemistry is hinted at in the south of the
ing drastically their carbonate content. Gypsum dust study area. Ca-rich salars are preferentially located
enriches the waters in Ca, which conversely lowers close to the sedimentary Precordillera and their inflow
their carbonate contents because of the very low waters undergo little thermal influence. Conversely,
solubility of calcite. Therefore, the (SO4 + Alk/2)/Ca SO4-rich salars tend to be located well inside the
ratio was found to be a valuable parameter to class and volcanic Western Cordillera and are largely fed by
investigate Chilean salars. thermal waters. This is consistent with the sedimen-
The abundance of native sulfur in volcanoes of the tary to volcanic transition from west to east associated
Western Cordillera leads to sulfate-rich inflow waters, to the increase of heat flow in the same direction.
which in turn produce sulfate-rich brines by evapo- Obviously, many factors obscure this overall correla-
rative concentration. Conversely, sedimentary rocks of tion: the sedimentary formations contain volcanic and
the Precordillera contain Ca minerals (calcite, dolo- volcano-sedimentary units. The boundary between
mite, gypsum) whose weathering leads preferentially Precordillera and Western Cordillera is not so pre-
to Ca-rich waters and Ca-rich brines. Small differ- cisely delimited as shown in our schematic maps.
ences of SO4 and Ca concentration in dilute inflows Thermal waters may be cooled by mixing with fresh
are amplified in the final brines by evaporative con- meteoric waters. Deep circulation in fractured eruptive
centration. This overall relation between lithology and rocks can be very complex. The northern and central
brine composition is remarkably illustrated in the areas seem to be particularly affected by these draw-
largest Chilean salar: the salar de Atacama. However, backs.
several salars in volcanic basins of the Western Dilute meteoric waters are more contaminated by
Cordillera belong to the Ca-rich group. Their chem- salt recycling in areas containing many salars. Salts
istry may have been inherited from the weathering of lost from one salar may be widely redistributed in
ancient sedimentary formations during the late Ter- several other basins. As a result, exploration for fresh
tiary or the Pleistocene. Later on, ignimbrite sheets water should be focused in areas devoid of salars. In
and lava flows may have covered the sedimentary salar basins, the water table is close to the topo-
formations. Such a process may well account for Ca- graphic surface, which reduces significantly the drill-
rich salars adjacent to the Precordillera, at the western ing costs. The drawback is that such waters are often
fringe of the volcanic cordillera. of poor quality. Outside salar areas, waters are much
Most of the salt load of inflow waters does not deeper, but probably less affected by salt contami-
stem directly from rock alteration in the drainage nation.
basins, but from lake brine recycling. The recycling
process is likely to be as old as the salars themselves.
Infiltration rates tend to slowly decrease with time Acknowledgements
because of the progressive choking of bottom sedi-
ments by clay and silt particles. Thus, brine leakage We want to express our deep gratitude to the
was probably more effective in past times. Non- institutions and companies that gave us invaluable
equilibrium conditions seem to control important and unselfish logistic assistance: the regional Bu-
features of Chilean salars. Old salt crusts deposited reaus of the Dirección General de Aguas (Regiones
under very arid climate are presently undergoing I, II, III); Ejército de Chile; Minera Escondida;
dissolution in the semiarid Cordillera. The Ca-rich Codelco Chile: Division Chuquicamata and Division
composition of many waters and brines does not Salvador; Minera Mantos Blancos; Corporacion
reflect the volcanic lithology of the Western Cordil- Nacional Forestal; Minera Quiborax; Minera Doña
lera. The global hydrogeochemical cycle is not yet Inés de Collahuasi; and Minera Zaldivar. Thanks are
equilibrated with the presently dominant volcanic due to Professor Abraham Lerman and to an
lithology. The quantity of weathering material enter- anonymous reviewer who made constructive reviews
ing the hydrological cycle is much lower than the of the manuscript. This is EOST contribution no.
quantity of recycled salts. 2003.403-UMR7517.
Appendix A Appendix A (continued)

Name Type CL/BR
Cl (mmol/l) and Br (Amol/l) of waters and brines of ALC-7 P 2.03/2.7
Chilean salars. Other components in Risacher et al. ALC-8 L 14.9/24
(1999). N = snow, S = spring, P = seep, R = river, ALC-9 L 58.6/110
L = lake, U = underground water. ALC-10 S 0.328/1.0
AMA-1 L 538/31
AMA-2 L 284/17
AMA-3 P 129/9.2
Name Type CL/BR AMA-4 L 3410/200
NEV-1 N 0.015/0.05 AMA-5 S 279/52
NEV-2 N 0.021/0.09 AMA-6 L 1420/68
NEV-3 N 0.018/0.13 ASC-1 U 32.7/18
NEV-4 N 0.029/0.10 ASC-2 U 8.88/1.5
NEV-5 N 0.025/0.08 ASC-3 U 25.3/15
NEV-6 N 0.003/0.02 ASC-4 U 30.5/9.4
NEV-8 N 0.003/0.04 ASC-7 U 109/8.6
NEV-9 N 0.005/0.02 ASC-9 U 35.5/9.0
AC1-1 S 408/56 ASC-10 U 60.1/10
AC1-4 S 102/18 ASC-13 S 0.087/0.03
AC1-6 L 1820/280 ASC-14 S 0.081/0.05
AC1-8 L 2020/210 ASC-15 U 25.0/14
AC1-11 S 17.3/7.3 ASC-16 U 3.01/1.9
AC2-1 S 19.4/2.6 ASC-17 U 47.2/13
AC2-4 P 35.5/7.7 ASC-18 L 211/43
AC2-7 S 40.0/12 ASC-19 S 122/7.9
AC2-10 P 217/28 ASC-20 L 124/9.0
AC2-13 P 35.2/13 ASC-21 L 287/21
AC2-15 L 165/45 ASC-22 U 31.6/7.7
AC3-1 P 75.0/14 ASC-23 L 91.3/19
AC3-4 L 361/46 ASC-24 U 31.4/8.7
AC3-7 S 28.0/11 ASC-25 L 67.5/21
AC3-10 P 126/13 ASC-26 S 30.9/6.0
AC3-12 P 22.3/4.2 ASC-27 L 64.7/15
AC3-13 P 116/40 ASC-28 S 56.3/8.7
AC3-14 S 34.1/14 ASC-29 L 151/23
AC3-15 L 48.0/18 ASC-30 S 54.4/10
AC4-1 P 9.39/5.0 ASC-31 L 136/21
AC4-2 L 380/190 ASC-32 L 657/150
AC4-3 L 88.3/43 ASC-33 S 55.8/9.7
AC4-4 P 27.1/12 ASC-34 L 199/41
AC4-5 L 271/130 ASC-36 S 49.1/18
AC4-6 L 892/460 ASC-37 S 33.9/17
AC4-7 L 4930/2400 ASC-38 S 27.8/13
AC4-8 P 25.2/16 ASC-39 L 266/120
AC4-9 S 5.22/2.8 ASC-40 L 107/24
AC4-10 S 16.9/10 ASC-41 L 1650/550
AGI-1 L 2960/230 ASC-42 P 28.4/1.7
AGI-1A L 3200/230 ASC-43 L 916/170
AGI-2 L 5880/150 ASC-44 S 3.69/0.75
ALC-1 U 2.52/2.7 ASC-45 L 583/120
ALC-2 S 0.169/0.49 ASC-46 L 245/68
ALC-3 S 0.168/0.57 ASC-47 U 30.3/10
ALC-4 S 0.197/0.30 ATA-1 R 11.8/0.48
ALC-5 S 0.184/0.40 ATA-2 S 2.32/0.28
ALC-6 L 1030/370 (continued on next page)
Appendix A (continued) Appendix A (continued)

Name Type CL/BR Name Type CL/BR
ATA-3 R 7.10/1.6 CAR-21 L 90.5/11
ATA-4 R 17.0/1.3 CAR-22 S 6.74/2.2
ATA-5 U 27.0/18 CAR-24 P 109/8.4
ATA-6 R 1.50/1.0 CAR-25 S 419/22
ATA-7 R 8.50/5.4 CAR-26 L 1250/69
ATA-8 R 2.03/5.0 CHR-1 L 2.71/0.57
ATA-9 R 1.25/1.0 CHR-2 R 0.091/0.09
ATA-10 R 6.50/3.8 CHR-3 S 0.007/0.05
ATA-11 R 2.50/1.6 CHR-4 S 0.045/0.06
ATA-12 R 25.0/11 CHR-5 L 0.087/0.08
ATA-13 R 43.0/11 CHR-6 S 0.046/0.05
ATA-14 R 15.4/10 CHR-7 L 1.74/0.23
ATA-15 R 15.0/10 CHR-8 S 0.016/0.02
ATA-16 S 15.0/10 COP-1 S 1.72/1.8
ATA-17 R 3.50/2.1 COP-2 S 43.9/25
ATA-18 R 0.480/0.42 COP-3 U 0.269/0.38
ATA-19 P 10.5/7.4 COP-4 P 5.30/2.0
ATA-20 P 304/40 COP-5 S 4.89/2.2
ATA-21 P 25.3/11 COP-6 L 461/180
ATA-22 U 52.9/14 COP-7 U 28.5/4.3
ATA-23 S 7.76/4.2 COP-8 U 0.777/0.78
ATA-24 S 0.041/0.08 COP-9 U 21.2/2.6
AZU-1 P 27.4/6.4 COP-10 L 461/160
AZU-2 L 893/78 COP-11 L 1430/29
AZU-3 L 2190/280 COP-13 L 32.3/12
AZU-4 L 120/23 COP-14 L 121/42
AZU-5 L 292/56 COP-15 L 5.60/2.4
AZU-6 L 5700/870 COP-16 U 20.9/2.2
AZU-7 U 4.41/1.1 COP-17 S 13.7/3.6
BAY-1 L 28.9/23 COP-18 L 807/190
BAY-2 P 15.9/7.1 COP-19 L 187/46
BRA-1 P 35.9/15 COP-20 L 43.6/10
BRA-2 L 1810/790 COP-21 L 4410/830
BRA-3 P 21.8/11 COP-22 U 3.01/1.6
BRA-4 L 47.7/24 COP-23 U 1.13/1.0
CAR-1 S 0.269/0.37 COP-24 U 0.739/0.89
CAR-2 S 0.272/0.32 COT-1 S 0.168/0.14
CAR-3 L 1030/59 COT-2 L 0.838/0.32
CAR-4 L 3020/130 COT-3 L 0.903/0.16
CAR-5 S 47.1/3.1 COT-4 L 2.55/1.3
CAR-6 L 2020/95 CPR-1 L 197/25
CAR-7 L 5700/530 CPR-2 S 106/17
CAR-8 L 2900/190 CPR-3 L 334/40
CAR-9 L 31.7/2.3 CPR-4 S 102/16
CAR-10 L 190/8.8 CPR-5 L 847/76
CAR-11 L 1250/70 CPR-6 L 1190/100
CAR-12 S 126/13 CPR-7 L 3770/280
CAR-13 L 1800/190 ESC-1 L 41.1/33
CAR-15 S 17.4/2.3 ESC-2 L 199/180
CAR-16 L 5490/130 ESC-3 P 18.7/14
CAR-17 L 463/87 ESC-4 P 30.5/30
CAR-18 S 11.4/2.6 ESC-5 L 347/330
CAR-19 L 273/48 ESC-6 L 71.6/59
CAR-20 L 36.8/7.6 FRA-1 R 0.483/0.17

FRA-2 P 2.73/1.3 HEL-7 S 0.537/0.35
FRA-3 R 0.023/0.03 HEL-8 L 23.7/5.6
FRA-4 L 4810/2700 HEL-9 L 14.9/22
FRA-5 P 0.608/0.41 IGN-1 P 4.00/1.1
FRA-6 S 0.405/0.07 IGN-3 L 17.0/1.8
FRA-7 L 429/170 IGN-4 L 12.5/1.9
FRA-8 S 0.630/0.10 IGN-5 L 39.6/5.1
FRA-9 L 97.0/37 IMI-1 U 2.70/2.0
FRA-10 S 3.41/2.0 IMI-2 L 3.24/2.1
FRA-12 P 11.4/7.5 IMI-3 U 55.5/7.8
FRA-13 S 1.02/0.49 IMI-5 L 1410/8.4
FRA-14 S 0.339/0.30 IMI-6 L 713/8.9
FRA-15 L 20.1/3.2 IMI-7 L 507/11
FRA-16 R 3.05/2.5 IMI-8 P 150/5.7
FRA-17 R 0.322/0.14 IMI-8D U 181/6.9
FRA-18 S 83.1/88 INF-1 P 15.3/6.2
GOR-1 R 3.36/0.72 INF-2 L 142/41
GOR-2 P 54.5/41 INF-3 L 5540/250
GOR-3 L 111/86 INF-4 L 1230/97
GOR-4 L 258/48 INF-5 L 280/28
GOR-5 L 1350/230 INF-6 P 119/19
GOR-6 P 5.63/2.0 ISL-1 S 276/2.2
GOR-7 P 112/8.6 ISL-2 L 5310/28
GOR-8 L 104/100 ISL-3 L 1410/8.0
GOR-9 L 506/270 ISL-4 P 102/1.1
GOR-10 P 46.4/52 ISL-5 L 1540/11
GOR-11 L 3450/110 ISL-6 L 425/2.2
GOR-12 L 3490/80 ISL-7 L 818/11
GRA-1 P 119/12 ISL-8 U 155/3.6
GRA-2 L 305/30 ISL-9 S 262/4.3
GRA-3 L 2120/170 ISL-10 L 658/11
GRA-4 L 440/47 ISL-11 L 2130/21
HCO-1 S 0.819/1.2 ISL-12 L 5070/22
HCO-2 R 0.145/0.25 ISL-13 S 494/2.5
HCO-3 R 0.269/0.64 ISL-14 L 2920/16
HCO-4 R 0.256/0.47 ISL-15 L 1220/8.0
HCO-5 R 0.748/0.72 ISL-16 S 226/2.5
HCO-6 U 2.30/2.6 ISL-17 L 5030/38
HCO-7 S 0.100/0.40 ISL-18 S 88.3/1.5
HCO-8 R 0.088/0.36 ISL-19 L 2470/52
HCO-9 U 0.576/0.73 JIL-1 L 177/100
HCO-10 R 0.390/0.64 JIL-2 P 9.93/7.3
HCO-11 S 1.43/0.98 JIL-3 L 115/67
HCO-12 S 0.916/1.3 JIL-4 L 123/67
HCO-13 L 140/78 JIL-5 L 215/110
HCO-14 L 222/140 JIL-6 P 32.3/13
HCO-15 L 80.2/47 LAA-1 U 0.101/0.29
HCO-16 L 12.5/8.0 LAA-2 R 1.14/1.1
HCO-17 L 246/150 LAA-3 R 0.964/0.85
HCO-18 L 909/470 LAA-4 R 1.52/1.4
HEL-1 S 1.66/0.24 LAA-5 S 0.682/0.45
HEL-3 L 5050/940 LAA-6 R 0.956/0.51
HEL-5 S 0.250/0.31 LAA-7 S 0.246/0.32
HEL-6 S 0.109/0.18 (continued on next page)

LAA-8 R 0.268/0.38 MAR-15 P 13.8/2.3
LAC-1 P 5.43/1.8 MAR-16 U 62.1/25
LAC-4 U 5.22/1.5 MAR-17 U 26.1/15
LAC-5 L 325/45 MAR-18 U 0.838/0.54
LAC-7 P 24.2/5.9 MAR-20 L 130/13
LAV-1 L 2580/1300 MAR-21 P 38.5/6.3
LAV-2 R 7.60/3.8 MAR-22 L 66.8/12
LAV-3 P 78.7/72 MAR-23 L 159/38
LAV-4 L 56.4/58 MAR-24 L 1420/280
LAV-5 L 520/220 MAR-25 L 5750/1300
LAV-6 S 41.1/16 MAR-26 P 22.7/3.9
LAV-7 S 8.02/3.1 MAR-27 S 1.12/0.85
LAV-8 R 24.0/12 MAR-28 U 1.20/0.27
LAZ-1 S 37.5/28 MAR-29 P 0.185/0.05
LAZ-2 L 1580/990 MAR-30 R 10.6/4.0
LAZ-3 L 166/110 MIC-1 S 0.138/0.23
LAZ-4 P 28.0/15 MIC-2 R 0.594/0.48
LEJ-1 L 274/130 MIC-3 S 0.664/0.60
LEJ-3 P 3.14/1.0 MIC-4 L 3.95/2.4
LGN-1 P 22.3/9.3 MIC-5 L 181/98
LGN-2 L 73.9/41 MIC-6 L 20.9/8.8
LGN-3 P 100/54 MIC-7 L 50.0/22
LGN-4 L 232/160 MIC-8 U 0.511/0.57
LGN-5 S 28.6/30 MIC-10 U 0.661/0.75
LGU-1 S 0.906/1.2 MIC-13 U 0.397/0.65
LGU-2 L 4.53/3.7 MIC-14 S 0.157/0.44
LGU-3 U 0.268/0.88 MIC-15 L 5.21/5.5
LOY-1 R 38.4/3.5 MIN-1 L 74.6/30
LOY-5 P 230/19 MIN-3 L 67.1/24
LOY-9 P 1.50/0.53 MIN-6 P 23.4/9.0
LOY-13 R 0.515/0.33 MIN-8 L 62.6/24
LOY-16 P 3.38/0.68 MIS-1 P 1.57/0.90
LOY-20 L 64.3/6.2 MIS-3 L 32.4/13
LOY-24 P 108/20 MIS-5 P 3.86/1.4
LOY-28 P 59.0/4.8 MIS-8 L 34.0/12
LOY-32 U 57.2/10 MIS-9 U 0.089/0.37
LOY-35 P 62.0/5.5 MUE-1 P 364/38
LOY-39 L 4170/65 MUE-4 L 1090/130
LOY-40 S 1.11/0.51 PAJ-1 S 240/38
LOY-41 S 0.489/0.26 PAJ-2 L 907/140
MAR-1 S 0.501/0.22 PAJ-3 U 371/53
MAR-2 R 0.856/0.63 PAJ-4 L 723/120
MAR-3 R 5.29/3.1 PAJ-5 L 4250/620
MAR-4 R 29.5/17 PAJ-6 L 2820/270
MAR-5 R 1.00/0.07 PAJ-7 P 381/35
MAR-6 U 2.44/2.4 PAJ-8 S 193/59
MAR-7 R 12.2/7.3 PAJ-9 L 1890/670
MAR-8 R 0.096/0.04 PAJ-10 L 3720/820
MAR-9 R 2.20/1.2 PAJ-11 L 2390/690
MAR-10 L 61.0/14 PAR-1 P 139/3.6
MAR-11 L 3.48/2.0 PAR-2 L 568/15
MAR-12 P 35.3/19 PAR-3 L 3710/87
MAR-13 P 19.1/9.7 PAR-4 P 254/6.8
MAR-14 L 251/32 PAR-5 L 5290/180

PAR-6 S 133/4.3 PIN-12 S 0.740/1.4
PAR-7 L 708/23 PIN-13 U 0.931/1.6
PED-1 S 28.2/22 PIN-14 S 2.49/2.4
PED-2 R 29.9/13 PIN-15 U 3.12/3.6
PED-3 P 2.87/1.1 PIN-16 U 19.4/16
PED-4 R 8.62/2.5 PIN-17 U 4.09/4.5
PED-5 S 6.43/0.47 PIN-18 S 1.22/2.3
PED-6 S 61.0/6.1 PIN-19 S 1.47/2.7
PED-7 U 1.81/0.93 PIN-20 S 1.18/0.44
PED-9 U 0.070/0.05 PIN-21 S 1.15/2.2
PED-10 P 2.89/1.6 PIN-22 S 3.01/4.5
PED-11 L 3640/190 PIN-23 S 1.14/2.2
PED-12 P 1310/68 PIN-24 S 3.98/6.5
PED-13 S 40.3/2.4 PIN-25 U 6.82/6.7
PED-14 L 1790/50 PIN-26 U 19.1/17
PED-15 L 5450/97 PIN-27 U 21.9/19
PED-16 P 245/8.7 PIN-28 U 23.6/21
PED-17 S 4.87/5.6 PIN-29 U 6.97/6.5
PED-18 S 0.799/0.67 PIN-30 R 32.3/29
PED-19 L 5410/180 PIN-31 R 1.06/2.1
PED-20 U 0.157/0.16 PIN-32 S 1.88/2.9
PED-21 U 0.148/0.14 PIN-33 S 1.51/2.0
PED-22A L 175/35 PIN-34 R 7.69/5.8
PED-22 L 510/97 PIN-35 U 10.0/7.9
PED-23 L 1010/120 PIN-36 L 935/120
PED-25 U 0.252/0.27 PIN-37 R 3.65/3.4
PED-26 U 35.2/12 PIN-38 S 0.391/1.8
PED-27 U 33.3/12 PIN-39 S 0.406/1.7
PED-28 U 40.9/16 PIN-40 U 7.08/6.6
PED-29 U 35.5/13 PIN-41 L 236/39
PIE-1 P 41.6/8.2 PIN-42 U 12.3/9.3
PIE-2 L 353/75 PIN-43 L 4390/680
PIE-3 L 97.5/22 PIN-44 L 1440/190
PIE-4 P 84.4/29 PIN-45 L 115/24
PIE-5 L 304/110 PSA-1 R 0.983/0.61
PIE-6 L 492/130 PSA-5 L 343/230
PIE-7 L 269/49 PSA-8 P 6.36/3.5
PIE-8 P 38.8/6.9 PSA-11 P 0.559/0.20
PIE-9 L 1220/190 PUN-1 L 309/92
PIE-10 L 163/38 PUN-2 L 195/57
PIE-11 S 30.0/6.9 PUN-3 L 162/47
PIE-12 L 419/170 PUN-4 P 170/11
PIE-13 P 38.8/7.3 PUN-5 L 294/21
PIE-14 L 141/26 PUN-6 L 4010/120
PIN-1 U 2.10/2.3 PUN-7 L 155/3.8
PIN-2 U 3.70/3.1 PUN-8 U 112/2.2
PIN-3 U 2.60/2.7 PUN-9 L 706/17
PIN-4 S 6.00/5.0 PUN-10 P 238/9.8
PIN-5 U 5.96/5.3 PUN-11 U 33.6/11
PIN-6 U 17.4/17 PUN-12 S 2.12/2.4
PIN-8 S 5.26/4.4 PUN-13 S 5.17/2.8
PIN-9 R 74.5/45 PUN-14 S 35.6/31
PIN-10 S 1.80/2.1 PUN-15 U 37.9/32
PIN-11 U 23.4/18
(continued on next page)
Appendix A (continued) References

Name Type CL/BR
PUN-16 U 43.9/36 Aguirre, I., Clavejo, J., 2000. Antecedentes fisicoquı́micos prelim-
PUN-17 R 43.4/32 inares de cuerpos de agua superficial del Altiplano de la III
PUN-19 U 4.84/0.77 Región de Atacama, Chile. Actas del IX Congreso Geológico
PUN-21 U 6.48/4.8 Chileno, (31 Julio – 4 Agosto 2000). Sociedad Geológica de
PUN-22 S 0.184/0.28 Chile, vol. 2, pp. 354 – 358.
PUN-23 R 0.491/0.35 Alderman, S.S., 1983. Geology of the Owens Lake evaporite de-
SUR-1 R 0.141/0.19 posit. In: Schreiber, B.C., Harner, H.L. (Eds.), Proceedings of the
SUR-2 S 47.9/36 Sixth International Symposium on Salt, the Salt Institute, vol. 1,
SUR-3 P 3.27/2.9 pp. 75 – 83.
SUR-4 S 1.49/1.2 Al-Droubi, A., Fritz, B., Gac, J.Y., Tardy, Y., 1980. Generalized
SUR-5 P 6.52/4.0 residual alkalinity concept; Application to prediction of the
SUR-7 S 0.653/0.97 chemical evolution of natural waters by evaporation. American
SUR-8 L 19.0/14 Journal of Science 280, 560 – 572.
SUR-9 S 39.3/32 Alonso, H., Risacher, F., 1996. Geoquı́mica del salar de Atacama,
SUR-10 P 23.2/19 parte 1: origen de los componentes y balance salino. Revista
SUR-11 P 0.793/0.79 Geológica de Chile 23 (2), 113 – 122.
SUR-12 S 1.20/1.0 Alonso, R.N., Jordan, T.E., Tabbutt, K.T., Vandervoort, D.S., 1991.
SUR-13 L 1040/680 Giant evaporite belts of the Neogene Central Andes. Geology
SUR-14 S 0.514/0.85 19, 401 – 404.
SUR-15 P 0.526/0.57 Alpers, C.N., Whittemore, D.O., 1990. Hydrogeochemistry and
SUR-16 S 6.12/4.6 stable isotopes of ground and surface waters from two adjacent
SUR-17 L 614/380 closed basins, Atacama Desert, northern Chile. Applied Geo-
SUR-18 S 0.375/0.53 chemistry 5, 719 – 734.
SUR-19 L 459/300 Aravena, R., Suzuki, O., Peña, H., Pollastri, A., Fuenzalida, H.,
SUR-20 S 226/170 Grilli, A., 1999. Isotopic composition and origin of the precip-
SUR-21 L 967/680 itation in northern Chile. Applied Geochemistry 14, 411 – 422.
SUR-22 L 50.6/33 Badaut, D., Risacher, F., 1983. Authigenic smectite on diatom frus-
SUR-23 L 4380/2700 tules in Bolivian saline lakes. Geochimica et Cosmochimica
SUR-24 L 1970/1300 Acta 47, 363 – 375.
SUR-25 L 89.3/66 Baker, P.A., Rigsby, C.A., Seltzer, G.O., Fritz, S.C., Lowenstein,
SUR-26 S 0.286/0.59 T.K., Bacher, N.P., Veliz, C., 2001. Tropical climate changes at
TAR-1 S 8.60/4.0 millennial and orbital timescales on the Bolivian Altiplano.
TAR-3 L 34.8/15 Nature 409, 698 – 701.
TAR-5 R 1.60/1.1 Bao, R., Sáez, A., Servant-Vildary, S., Cabrera, L., 1999. Lake-
TAR-6 P 7.25/3.5 level and salinity reconstruction from diatom analyses in Quil-
TAR-9 L 2760/1100 lagua formation (late Neogene, Central Andean forearc, north-
TAR-11 S 9.18/5.0 ern Chile). Palaeogeography, Palaeoclimatology, Palaeoecology
TAR-12 S 26.4/13 153, 309 – 335.
TAR-13 R 15.7/6.8 Behnke, R., 1987. Informe preliminar sobre la factibilidad técnica
TRI-1 L 131/14 ecológica de extraer aguas de la Laguna del Negro Francisco, III
TUY-1 R 88.9/9.7 Región. Informe técnico (Open-file report), Dirección General
TUY-4 L 2120/160 de Aguas, Departamento de Estudios, Santiago, Chile.
WHE-1 R 100/11 Berthold, C.E., Baker, D.H., 1976. Lithium recovery from geother-
WHE-2 L 2750/270 mal fluids. U.S. Geological Survey Professional Paper 1005,
WHE-3 S 48.7/5.2 61 – 66.
WHE-4 L 329/42 Bevacqua, P., 1991. Geomorfologı́a del Salar de Atacama y estrati-
WHE-5 L 2480/290 grafı́a de su núcleo y delta, Segunda Región de Antofagasta,
WHE-6 L 708/96 Chile. PhD Thesis, Universidad Católica del Norte, Antofagasta,
WHE-7 P 96.1/12 Chile.
WHE-8 L 175/25 Bobst, A.L., Lowenstein, T.K., Jordan, T.E., Godfrey, L.V., Ku, T.H.,
WHE-9 P 20.2/3.2 Luo, S., 2001. A 106 ka paleoclimate record from drill core of the
Salar de Atacama, northern Chile. Palaeogeography, Palaeocli-
matology, Palaeoecology 173 (1 – 2), 21 – 42.
Caceres, L., Gruttner, V.E., Contreras, R., 1992. Water recycling in
arid regions: the Chilean case. Ambio 21 (2), 138 – 144.
Carmona, V., Pueyo, J.J., Taberner, C., Chong, G., Thirlwall, M., Garrels, R.M., Mackenzie, F.T., 1967. Origin of the chemical com-
2000. Solute inputs in the Salar de Atacama (N. Chile). Journal position of some springs and lakes. Equilibrium Concepts in
of Geochemical Exploration 69 – 70, 449 – 452. Natural Water Systems. Advances in Chemistry, vol. 67. Amer-
Carmouze, J.P., Pedro, G., 1977. The contribution of the salt regu- ican Chemical Society, Washington, DC, pp. 222 – 242.
lation of a lacustrine system by geographical and sedimentolog- Gaupp, R., Kött, A., Wörner, G., 1999. Paleoclimatic implications
ical factors. In: Paquet, H., Tardy, Y. (Eds.), Proceedings of the of Mio-Pliocene sedimentation in the high-altitude intra-arc
Second International Sympsium on Water – Rock Interaction Lauca Basin of northern Chile. Palaeogeography, Palaeoclima-
(Strasbourg, France), vol. 2, pp. 73 – 84. tology, Palaeoecology 151, 79 – 100.
Chaffaut, I., 1998. Precipitations d’altitude, eaux souterraines et Geyh, M.A., Grosjean, M., Núñez, L., Schotterer, U., 1999. Radio-
changements climatiques de l’Altiplano nord-chilien. Thesis, carbon reservoir effect and the timing of the late-glacial/early
University Paris-Sud, Orsay. Holocene humid phase in the Atacama Desert (Northern Chile).
Chong, G., 1984. Die Salare in Nordchile, Geologie, Struktur und Quaternary Research 52, 143 – 153.
Geochemie. Geotektonische Forschungen 67, 146 pp. Gidhagen, L., Kahelin, H., Schmidt-Thomé, P., Johansson, C.,
Chong, G., Mendoza, M., Garcı́a-Veigas, J., Pueyo, J.J., Turner, 2002. Anthropogenic and natural levels of arsenic in PM10 in
P., 1999. Evolution and geochemical signature in a Neogene central and Northern Chile. Atmospheric Environment 36 (23),
forearc evaporitic basin: the Salar Grande (Central Andes of 3803 – 3817.
Chile). Palaeogeography, Palaeoclimatology, Palaeoecology Grilli, A., Vidal, J., Garin, C., 1986. Balance hidrologico nacional, II
151, 39 – 54. Region (Chile). Publicacion interna (Open-File report) S.D.E.H.
DGA, 1987. Balance hı́drico de Chile. Dirección General de Aguas, 86/1.
Ministerio de Obras Publicas, Santiago, Chile. Grosjean, M., 1994. Paleohydrology of the Laguna Lejia (north
Dickson, A.G., 1981. An exact definition of total alkalinity and a Chilean Altiplano) and climatic implications for late-glacial
procedure for the estimation of alkalinity and total inorganic times. Palaeogeography, Palaeoclimatology, Palaeoecology
carbon from titration data. Deep-Sea Research 28A, 609 – 623. 109, 89 – 100.
Dingman, R.J., 1962. Tertiary salt domes near San Pedro de Ata- Grosjean, M., Geyh, M.A., Messerli, B., Schotterer, U., 1995. Late-
cama, Chile. U.S. Geological Survey Professional Paper 450-D, glacial and early Holocene lake sediments, groundwater forma-
D92 – D94. tion and climate in the Atacama Altiplano 22 – 24jS. Journal of
Dingman, R.J., 1967. Geology and ground-water resources of the Paleolimnology 14, 241 – 252.
northern part of the salar de Atacama, Antofagasta Province, Grosjean, M., van Leeuwen, J.F.N., van der Knaap, W.O., Geyh,
Chile. U.S. Geological Survey Bulletin 1219. M.A., Ammann, B., Tanner, W., Messerli, B., Núñez, L.A., Va-
Dutkiewicz, A., Herczeg, A.L., Dighton, J.C., 2000. Past changes to lero-Garcés, B.L., Veit, H., 2001. A 22,000 14C year BP sediment
isotopic and solute balances in a continental playa: clues from and pollen record of climate change from Laguna Miscanti
stable isotopes of lacustrine carbonates. Chemical Geology 165, (23jS), northern Chile. Global and Planetary Change 28, 35 – 51.
309 – 329. Hardie, L.A., Eugster, H.P., 1970. The evolution of closed-basin
Ericksen, G.E., 1981. Geology and origin of the Chilean nitrate de- brines. Mineralogical Society of America Special Paper 3,
posits. U.S. Geological Survey Professional Paper 1188, 37 pp. 273 – 290.
Ericksen, G.E., Vine, J.D., Ballon, R., 1978. Chemical composition Harvie, C.E., Moller, N., Weare, J.H., 1984. The prediction of
and distribution of lithium-rich brines in salar de Uyuni and mineral solubilities in natural waters: the Na – K – Mg – Ca –
nearby salars in southwestern Bolivia. Energy 3, 355 – 363. H – Cl – SO4 – OH – HCO3 – CO3 – CO2 – H2O system to high
Eugster, H.P., Hardie, L.A., 1978. Saline lakes. In: Lerman, A. ionic strengths at 25 jC. Geochimica et Cosmochimica Acta
(Ed.), Lakes: Chemistry, Geology, Physics. Springer, New York. 48, 723 – 751.
Eugster, H.P., Jones, B.F., 1979. Behavior of major solutes during Hastenrath, S., Kutzbach, J., 1985. Late Pleistocene climate and
closed-basin brine evolution. American Journal of Science 279, water budget of the South American Altiplano. Quaternary Re-
609 – 631. search 24, 249 – 256.
Fornari, M., Risacher, F., Feraud, G., 2001. Dating of paleolakes in Heath, R.C., 1983. Basic ground-water hydrology. U.S. Geological
the central Altiplano of Bolivia. Palaeogeography, Palaeoclima- Survey Water-Supply Paper 2220, 84 pp.
tology, Palaeoecology 172 (3 – 4), 269 – 282. Hellwig, D.H.R., 1973. Evaporation of water from sand: 4. The
Gac, J.Y., 1980. Géochimie du bassin du lac Tchad. Bilan de l’alt- influence of the depth of the water-table and the particle size
ération, de l’érosion et de la sédimentation. Travaux et Docu- distribution of the sand. Journal of Hydrology 18, 317 – 327.
ments de l’ORSTOM, Paris 123, 251 pp. Hutchinson, G.E., 1957. A Treatise on Limnology, vol. 1, Geogra-
Garcés, I., 2000a. Geochemistry of Huasco salar, Chile. Origin of phy, Physics and Chemistry. Wiley, New York.
solutes and brine evolution. In: Geertman, R.M. (Ed.), Proceed- Ide, F., 1978. Cubicación del yacimiento salar de Atacama. Memo-
ings of the 8th World Salt Symposium, vol. 2. Elsevier, Am- ria para optar al tı́tulo de Ingeniero Civil de Minas. Universidad
sterdam, pp. 1159 – 1160. de Chile, Facultad de Ciencias Fı́sicas y Matemáticas, Departa-
Garcés, I., 2000b. The sodium sulfate and ulexite deposits: salar mento de Minas, Santiago. 144 pp.
of Surire, Chile. In: Geertman, R.M. (Ed.), Proceedings of Jones, B.F., 1965. The hydrology and mineralogy of Deep Spring
the 8th World Salt Symposium, vol. 2. Elsevier, Amsterdam, Lake, Inyo County, California. U.S. Geological Survey Profes-
pp. 1161 – 1162. sional Paper 502-A, 56 pp.
Kinsman, D.J.J., 1976. Evaporites: relative humidity control of Pueyo, J.J., Chong, G., Jensen, A., 2001. Neogene evaporites in
primary mineral facies. Journal of Sedimentary Petrology 46 desert volcanic environments: Atacama Desert, northern Chile.
(2), 273 – 279. Sedimentology 48, 1411 – 1431.
Kött, A., Gaupp, R., Wörner, G., 1995. Miocene to recent history Queirolo, F., Stegen, S., Restovic, M., Paz, M., Ostapczuk, P.,
of the western Altiplano in northern Chile revealed by lacus- Schwuger, M.J., Muñoz, L., 2000a. Total arsenic, lead, and
trine sediments of the Lauca basin. Geologische Rundschau cadmium levels in vegetables cultivated at the Andean vil-
84, 770 – 780. lages of northern Chile. Science of the Total Environment
Lahsen, A., 1988. Chilean geothermal resources and their possible 255, 75 – 84.
utilization. Geothermics 17 (213), 401 – 410. Queirolo, F., Stegen, S., Mondaca, J., Cortés, R., Rojas, R., Con-
Langbein, W.B., 1961. Salinity and hydrology of closed lakes. U.S. treras, C., Muñoz, L., Schwuger, M.J., Ostapczuk, P., 2000b.
Geological Survey Professional Paper 412, 20 pp. Total arsenic, lead, cadmium, copper, and zinc in some salt
Lavenu, A., Fornari, M., Sebrier, M., 1984. Existence de deux rivers in the northern Andes of Antofagasta, Chile. Science of
nouveaux épisodes lacustres quaternaires dans l’Altiplano Péru- the Total Environment 255, 85 – 95.
vo-Bolivien. Cahiers ORSTOM, Série Géologie 14, 103 – 114. Rettig, S.L., Jones, B.F., Risacher, F., 1980. Geochemical evolution
Long, D.T., Fegan, N.E., Lyons, W.B., Hines, M.E., Macumber, of brines in the salar of Uyuni, Bolivia. Chemical Geology 30,
P.G., Giblin, A.M., 1992. Geochemistry of acid brines: Lake 57 – 79.
Tyrell, Victoria, Australia. Chemical Geology 96, 33 – 52. Ripple, C.D., Rubin, J., Van Hylckama, T.E.A., 1972. Estimating
Made, B., Clement, A., Fritz, B., 1994. Modeling mineral/solution steady-state evaporation rates from bare soils under conditions
interactions: the thermodynamic and kinetic code KINDISP. of high water table. U.S. Geological Survey Water-Supply Paper
Computers and Geosciences 20 (9), 1347 – 1363. 2019-A.
Marshall, W.L., 1980. Amorphous silica solubilities: III. Activity Risacher, F., Alonso, H., 1996. Geoquı́mica del salar de Atacama,
coefficient relations and predictions of solubility behavior in salt parte 2: evolución de las aguas. Revista Geológica de Chile 23
solutions, 0 – 350 jC. Geochimica et Cosmochimica Acta 44, (2), 123 – 134.
925 – 931. Risacher, F., Alonso, H., 2001. Geochemistry of ash leachates from
Marshall, W.L., Chen, C.T.A., 1982. Amorphous silica solubilities: the 1993 Lascar eruption, northern Chile. Implication for recy-
V. Predictions of solubility behavior in aqueous mixed electro- cling of ancient evaporites. Journal of Volcanology and Geo-
lyte solutions to 300 jC. Geochimica et Cosmochimica Acta 46, thermal Research 109, 319 – 337.
289 – 291. Risacher, F., Clement, A., 2001. A computer program for the sim-
Marshall, W.L., Warakomski, J.M., 1980. Amorphous silica solu- ulation of evaporation of natural waters to high concentration.
bilities: II. Effect of aqueous salt solutions at 25 jC. Geochimica Computers and Geosciences 27, 191 – 201.
et Cosmochimica Acta 44, 915 – 924. Risacher, F., Fritz, B., 1991a. Quaternary geochemical evolution of
May, G., Hartley, A.J., Stuart, F.M., Chong, G., 1999. Tectonic the salars of Uyuni and Coipasa, central Altiplano, Bolivia.
signatures in arid continental basins: an example from the Upper Chemical Geology 90, 211 – 231.
Miocene – Pleistocene, Calama Basin, Andean forearc, northern Risacher, F., Fritz, B., 1991b. Geochemistry of Bolivian salars,
Chile. Palaeogeography, Palaeoclimatology, Palaeoecology 151, Lipez, southern Altiplano. Origin of solutes and brine evolution.
55 – 57. Geochimica et Cosmochima Acta 55, 687 – 705.
McArthur, J.M., Turner, J.V., Lyons, W.B., Osborn, A.O., Thirlwall, Risacher, F., Fritz, B., 2000. Bromine geochemistry of Salar de
M.F., 1991. Hydrochemistry on the Yilgarn Block, Western Uyuni and deeper salt crusts, Central Altiplano, Bolivia. Chem-
Australia: ferrolysis and mineralisation in acidic brines. Geo- ical Geology 167 (3 – 4), 373 – 392.
chimica et Cosmochimica Acta 55, 1273 – 1288. Risacher, F., Alonso, H., Salazar, C., 1999. Geoquı́mica de aguas en
Miller, D.H., 1977. Water at the surface of the Earth. An introduc- cuencas cerradas, I, II, III Regiones, Chile. Ministerio de Obras
tion to ecosystem hydrodynamics. International Geophysics Ser- Públicas, Dirección General de Aguas. Technical Report S.I.T.
ies, vol. 21. Academic Press, New York, 557 pp. No. 51, Santiago, Chile.
Moraga, A., Chong, G., Fortt, M.A., Henriquez, H., 1974. Estudio Risacher, F., Alonso, H., Salazar, C., 2002. Hydrochemistry of two
geológico del salar de Atacama, Provincia de Antofagasta. Bo- adjacent acid saline lakes in the Andes of northern Chile. Chem-
letı́n del Instituto de Investigaciones Geológicas, Santiago, Chile ical Geology 187 (1 – 2), 39 – 57.
29 (56 pp.). Sáez, A., Cabrera, L., Jensen, A., Chong, G., 1999. Late Neogene
Moser, H., Stichler, W., 1980. Environmental isotopes in ice and lacustrine record and palaeogeography in the Quillaga – Llamara
snow. In: Fritz, P., Fontes, J.Ch. (Eds.), Handbook of Environ- basin, Central Andean fore-arc (northern Chile). Palaeogeogra-
mental Isotope Geochemistry, vol. 1, The Terrestrial Environ- phy, Palaeoclimatology, Palaeoecology 151, 5 – 37.
ment A. Elsevier, Amsterdam, pp. 141 – 178. Sanford, W.E., Wood, W.W., 1991. Brine evolution and mineral
Peña, H., 1989. Mediciones de 18O y 2H en ‘‘penitentes’’ de nieve. deposition in hydrologically open evaporite basins. American
Estudios de hidrologia isotopica en America latina. IAEA-TEC- Journal of Science 291, 687 – 710.
DOC 502 (Vienna, Austria), 143 – 154. Servant, M., Fontes, J.C., 1978. Les lacs quaternaires des hauts
Pitzer, K.S., 1979. Theory: ion interaction approach. In: Pytkowicz, plateaux des Andes boliviennes. Premières interprétations palé-
R.M. (Ed.), Activity Coefficients in Electrolyte Solutions, vol. 1. oclimatiques. Cahiers ORSTOM, Série Géologie 10 (1), 9 – 23.
CRC Press, Boca Raton, FL, pp. 157 – 208. Smedley, P.L., Kinniburgh, D.G., 2002. A review of the source,
behaviour and distribution of arsenic in natural waters. Applied aquifer, Rainer Mesa, Nevada. U.S. Geological Survey Water-
Geochemistry 17 (5), 517 – 568. Supply Paper 1535-Q.
Smith, G.I., 1979. Subsurface stratigraphy and geochemistry of Yechieli, Y., Wood, W.W., 2002. Hydrogeologic processes in saline
Late Quaternary Evaporites, Searles Lake, California. U.S. Geo- systems: playas, sabkhas, and saline lakes. Earth-Science Re-
logical Survey Professional Paper 1043, 130 pp. views 58, 343 – 365.
Smith, A.H., Goycolea, M., Haque, R., Biggs, M.L., 1998. Marked Yonts, W.L., Giese, G.L., Hubbard, E.F., 1973. Evaporation from
increase in bladder and lung cancer mortality in a region of Lake Michie, North Carolina, 1961 – 71. U.S. Geological Survey
northern Chile due to arsenic in drinking water, American. Jour- Water-Resources Investigations 38-73, 27 pp.
nal of Epidemiology 47, 660 – 669.
Spiro, B., Chong, G., 1996. Origin of sulfate in the Salar de Ata- Francß ois Risacher is a graduate of the
cama and the Cordillera de la Sal, initial results of an isotopic Ecole Nationale Supérieure de Géologie
study. Proceedings of the Third International Symposium on de Nancy (1971). He holds a PhD in
Andean Geodynamics (ISAG), Saint Malo, France, Orstom Ed- Geochemistry from the University Louis
itions. Collection Colloques et Seminaires, Paris, pp. 703 – 705. Pasteur of Strasbourg (1992). He is a
Springer, M., Förster, A., 1998. Heat-flow density across the Cen- Senior Researcher at the Institut de
tral Andean subduction zone. Tectonophysics 291, 123 – 139. Recherche pour le Développement (a
Stanhill, G., 2002. Is the Class A evaporation pan still the most French cooperation agency formerly
practical and accurate meteorological method for determining known as Orstom). For more than 20
irrigation water requirements? Agricultural and Forest Manage- years, he carried out applied and funda-
ment 112, 233 – 236. mental research on evaporitic basins in
Stoertz, G.E., Ericksen, G.E., 1974. Geology of salars in northern Bolivia and Chile, focusing his interest on economical resources
Chile. U.S. Geological Survey Professional Paper 811, 65 pp. and geochemistry of saline lakes and salt crusts.
Stumm, W., Morgan, J.J., 1996. Aquatic Chemistry. Chemical
Equilibria and Rates in Natural Waters. Wiley Interscience, Hugo Alonso is a Chemical Engineer from
New York. the Universidad Católica del Norte (UCN),
Sylvestre, F., Servant, M., Servant-Vildary, S., Causse, C., Fournier, Chile (1965). Afterwards, he studied Soils
M., Ybert, J.P., 1999. Lake-level chronology on the southern and Hydrological Sciences in France. His
Bolivian Altiplano (18j – 23jS) during late-glacial time and research interests focus on soil reclamation
the early Holocene. Quaternary Research 51, 54 – 66. in the Atacama Desert, irrigation with
Valero-Garcés, B.L., Grosjean, M., Schwalb, A., Geyh, M., Mes- saline waters in native Andean commun-
serli, B., Kelts, K., 1996. Limnogeology of Laguna Miscanti: ities, arsenic and boron contamination in
evidence for mid to late Holocene moisture changes in the Ata- water and crops, preservation of water
cama Altiplano (Northern Chile). Journal of Paleolimnology 16, quality in mining areas. In 1995, he
1 – 21. received a National Government Award
Valero-Garcés, B.L., Grosjean, M., Kelts, K., Screier, H., Messerli, for his contributions to the use of waters in the arid zone of northern
B., 1999. Holocene lacustrine deposition in the Atacama Alti- Chile. Now, he is Professor of Environmental Geochemistry at UCN
plano: facies models, climate and tectonic forcing. Palaeogeog- and a consultant on water quality management for the National
raphy, Palaeoclimatology, Palaeoecology 151, 101 – 125. Environmental Commission, a Chilean government agency, and for
Van Denburgh, A.S., 1975. Solute balance at Abert and Summer several mining companies.
Lakes, South-Central Oregon. U.S. Geological Survey Profes-
sional Paper 502-C, 29 pp.
Carlos Salazar is a Civil Engineer from the
Vuille, M., 1996. Zur raumzeitlichen Dynamik von Schneefall und
Universidad de Chile (1981). He joined the
Ausaperung im Bereich des südlichen Altiplano, Südamerika.
Chilean National Water Resources Service
Geographica Bernensia (Univ. Bern, Switzerland) G45.
(Direccion General de Aguas), where he is
White, D.E., Hem, J.D., Waring, G.A., 1963. Data of geochemistry.
principally involved in water resources
Chemical composition of subsurface waters. U.S. Geological
development and global management as
Survey Professional Paper 440-F, 67 pp.
the Head of the Studies and Planning
White, D.E., Thompson, J.M., Fournier, R.O., 1976. Lithium con-
Department. He has special interests in
tents of thermal and mineral waters. U.S. Geological Survey
groundwaters and wetland systems in arid
Professional Paper 1005, 58 – 60.
areas.
White, A.F., Claasen, H.C., Benson, L.V., 1980. The effect of dis-
solution of volcanic glass on the water chemistry in a tuffaceous

The Origin of Brines and Salts in Chilean Salars: A Hydrochemical Review

Uploaded by

Copyright:

Available Formats

The Origin of Brines and Salts in Chilean Salars: A Hydrochemical Review

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Origin of Brines and Salts in Chilean Salars: A Hydrochemical Review

Uploaded by

Copyright:

Available Formats

Earth-Science Reviews 63 (2003) 249 – 293

The origin of brines and salts in Chilean salars:

* Corresponding author. Tel.: +33-3-90240409; fax: +33-3-88367235.

1. Introduction hydrogeochemical trends, focusing on the origin of

F. Risacher et al. / Earth-Science Reviews 63 (2003) 249–293

F. Risacher et al. / Earth-Science Reviews 63 (2003) 249–293

F. Risacher et al. / Earth-Science Reviews 63 (2003) 249–293

Table 2 tary, metamorphic and igneous rocks. The large cop-

8. Origin of deep brines

8.1. Recycling of ancient salars

Volcanic formations in northern Chile range in age

Table 5 water. This happens in confined aquifers where the

correspond to the Ca-rich path. Almost all high

9.3.4. Discrepancy between lithology and brine

Appendix A Appendix A (continued)

Appendix A (continued) Appendix A (continued)

Appendix A (continued) Appendix A (continued)

Appendix A (continued) Appendix A (continued)

Appendix A (continued) Appendix A (continued)

Appendix A (continued) References

You might also like