Isotopic evidence for the source of Ca and S in soil gypsum, anhydrite and calcite in the Atacama Desert, Chile

Geochimica et Cosmochimica Acta, Vol. 67, No. 4, pp. 575–586, 2003 Copyright © 2003 Elsevier Science Ltd Printed in the USA. All rights reserved 0016-7037/03 $22.00 ⫹ .00

Pergamon

doi:10.1016/S0016-7037(02)01175-4

Isotopic evidence for the source of Ca and S in soil gypsum, anhydrite and calcite in the Atacama Desert, Chile JASON A. RECH,*,‡ JAY QUADE, and WILLIAM S. HART Department of Geosciences and Desert Laboratory, University of Arizona, Tucson, AZ 85721 USA (Received November 20, 2001; accepted in revised form August 26, 2002)

Abstract—The origin of pedogenic salts in the Atacama Desert has long been debated. Possible salt sources include in situ weathering at the soil site, local sources such as aerosols from the adjacent Pacific Ocean or salt-encrusted playas (salars), and extra-local atmospheric dust. To identify the origin of Ca and S in Atacama soil salts, we determined ␦34S and 87Sr/86Sr values of soil gypsum/anhydrite and 87Sr/86Sr values of soil calcite along three east-west trending transects. Our results demonstrate the strong influence of marine aerosols on soil gypsum/anhydrite development in areas where marine fog penetrates inland. Results from an east-west transect located along a breach in the Coastal Cordillera show that most soils within 90 km of the coast, and below 1300 m in elevation, are influenced by marine aerosols and that soils within 50 km, and below 800 m in elevation, receive ⬎50% of Ca and S from marine aerosols (␦34S values ⬎ 14‰ and 87Sr/86Sr values ⬎0.7083). In areas where the Coastal Cordillera is ⬎1200 m in elevation, however, coastal fog cannot penetrate inland and the contribution of marine aerosols to soils is greatly reduced. Most pedogenic salts from inland soils have ␦34S values between ⫹5.0 to ⫹8.0‰ and 87Sr/86Sr ratios between 0.7070 and 0.7076. These values are similar to average ␦ 34S and 87Sr/86Sr values of salts from local streams, lakes, and salars (⫹5.4 ⫾2‰ ␦34S and 0.70749 ⫾ 0.00045 87Sr/86Sr) in the Andes and Atacama, suggesting extensive eolian reworking of salar salts onto the surrounding landscape. Ultimately, salar salts are precipitated from evaporated ground water, which has acquired its dissolved solutes from water-rock interactions (both high and low-temperature) along flowpaths from recharge areas in the Andes. Therefore, the main source for Ca and S in gypsum/anhydrite in non-coastal soils is indirect and involves bedrock alteration, not surficially on the hyperarid landscape, but in the subsurface by ground water, followed by eolian redistribution of ground-water derived salar salts to soils. The spatial distribution of high-grade nitrate deposits appears to correspond with areas that receive the lowest fluxes of local marine and salar salt, supporting arguments for tropospheric nitrogen as the main source for soil nitrate. Copyright © 2003 Elsevier Science Ltd sen, 1981; Searl and Rankin, 1993, Berger and Cooke, 1997); 4) direct deposition of salts associated with volcanic emissions (sulfates) (Searl and Rankin, 1993; Berger and Cooke, 1997); 5) the decay of bird guano and remains of organic lake deposits (nitrate) (Ericksen, 1983). The importance of desert dust goes well beyond its influence on soil formation in the Atacama Desert. Atmospheric dust originates mainly in deserts and is transported to soils worldwide (Uematsu et al., 1983, 1985; Nakai et al., 1993; Rea, 1994, Tegen and Fung, 1994; Simonson, 1995; Kohfeld and Harrison, 2001). For example, dust from the deserts of Central Asia is transported to Hawaii (Dymond et al., 1974; Parrington et al., 1983) and Greenland (Biscaye et al., 1997), and dust from the Sahara Desert reaches Israel (Dan, 1991), South America (Prospero et al., 1981; Swap et al., 1992), and the Bahamas (Muhs et al., 1987). This fine-grained dust is composed of silicate minerals and soluble salts and is thought to play an essential role in sustaining many tropical ecosystems, where plant nutrients are rapidly weathered and removed from soil systems (Chadwick et al., 1999). There is also a component of atmospheric dust (e.g., nitrate), however, that forms directly in the atmosphere by various chemical reactions (Crutzen, 1974; Simonaitis and Heicklen, 1975; Noxon, 1976, 1978). Little is known about the deposition rates of these compounds that are commonly utilized by plants on the land surface or quickly removed by weathering processes. Local dust and aerosol sources, especially in arid regions, can also influence

1. INTRODUCTION

The Atacama Desert is likely the oldest and driest desert on earth, optimal for the accumulation of highly soluble soil salts. Moreover, landscape surfaces in the Atacama are exceptionally stable and thought to be largely mid-Miocene in age (Alpers and Brimhall, 1988), virtually unaltered by chemical weathering and erosion. Atacama soil salts, composed mainly of sulfate, nitrate, and chloride salts, are unique on earth (Ericksen, 1981), and their origin has been debated for over a century (see Ericksen, 1981, for citations between 1861 and 1980; van Moort, 1985; Chong, 1988; Alpers and Whittemore, 1990; Searl and Rankin, 1993; Ericksen, 1994; Berger and Cooke, 1997; Bo¨hlke et al., 1997). Some of the origins suggested by previous researchers include: 1) marine aerosols associated with upwelling in the Pacific Ocean immediately to the west of the Atacama (nitrates, sulfates, chlorides, and iodates) (Ericksen, 1981; van Moort, 1985; Chong, 1988); 2) dry and wet deposition of extra-local atmospheric aerosols (nitrate, iodate, perchlorate) (Claridge and Campbell, 1968; Ericksen, 1981; Bo¨hlke et al., 1997); 3) salts derived from the weathering of Andean volcanics (sulfates, chlorides) (Ericksen, 1961; Erick-

* Author to whom correspondence should be addressed ([email protected]). ‡ Present address: Department of Geology, Miami University, Oxford, OH 45056 USA 575

576

J. A. Rech, J. Quade and W. S. Hart

Fig. 1. Topographic map of the central Atacama Desert (22°–25°S), Chile, with sample locations and areas of high-grade nitrate deposits (as reported by Ericksen; 1981).

soils and the transport of nutrients across landscapes (Pe´ we´ et al., 1981; McFadden et al., 1987; Brimhall et al., 1988; Chadwick and Davis, 1990; Quade et al., 1995; Naiman et al., 2000). The Atacama Desert is an ideal location to study the nature and deposition of highly soluble and/or biologically important components of atmospheric dust, as well as sources and transport pathways for local dust. In this study we use stable isotopes of S and Sr to identify the origin of S and Ca in soil gypsum/anhydrite and Ca in soil carbonate along three eastwest trending transects in the Atacama. Our objectives are to determine the relative contribution of marine aerosols to soils; to identify the Andean signature for S and Sr and determine the importance of Andean weathering products to soils; to evaluate the importance of evaporite playas (salars) as source areas for soil salts; and finally to gain an understanding of the spatial distribution of the high-grade nitrate deposits.

2. THE ATACAMA DESERT

2.1. Location and Environment The Atacama Desert is located between the central Andes and Pacific Ocean in southern Peru and northern Chile (Fig. 1). The core of the Atacama receives virtually no precipitation (⬍5 mm/yr), does not contain vascular plants, and supports few soil microbes (Cameron et al., 1966). The western margin of the Atacama receives consistent moisture in the form of coastal fog, locally known as camanchaca, between 300 and ⬃1000 m in elevation, which supports isolated coastal plant communities (lomas; Rundel et al., 1991). The eastern margin of the Atacama is at the base of the Andes (⬃2500 m). This margin grades into semiarid desert supporting desert scrub vegetation between 2700 to 3000 m in elevation and Andean Puna grasslands between ⬃3300 to 4000 m in elevation. Even the Andes

Isotopic evidence for the source of soil salts in the Atacama Desert, Chile

in this region, which include peaks ⬎6000 m in elevation, only receive ⬃200 mm/yr precipitation and lack perennial snow cover. The region of semiarid desert, between 2700 and ⬃6000 m in elevation and which receives between 10 to 200 mm/yr precipitation, is here referred to as the High Atacama. The eastern boundary between the Atacama and High Atacama has fluctuated in response to Quaternary climate change (Betancourt et al., 2000). Predominate wind direction in the Atacama is from the west, as indicated by eolian features associated with salars (Stoertz and Ericksen, 1974). 2.2. Geology The central Atacama (20° to 25°) is defined by three northsouth trending mountain ranges separated by alluvial basins: the Andes, the Cordillera Domeyko, and the Coastal Cordillera. The Andes range in elevation between 4500 to 6500 m in elevation and are composed mostly of Tertiary andesitic volcanics with some Jurassic and Cretaceous continental clastic sedimentary rocks. The Cordillera Domeyko is composed of Paleozoic and Mesozoic clastic and carbonate sedimentary rocks of continental and marine origin and middle Cretaceous to early Tertiary intrusive rocks. Large depositional basins, such as the Salar de Atacama and Salar de Punta Negra, separate the Andes from the Cordillera Domeyko and contain thick deposits of valley fill. The Coastal Cordillera is composed of Jurassic andesitic volcanics intercalated with Cretaceous marine limestone, shale, and conglomerates (Coira et al., 1982; Flint et al., 1993). 2.3. Surficial Salt Deposits Atacama soils with high concentrations of nitrate (7 to 15%), as well as iodate, perchlorate, chromate, and borate, have been the focus of much research. Ericksen (1981) provides a detailed description of these soils, their spatial distribution, and summarizes various theories regarding their genesis. Most highgrade nitrate soils are located between 19°S and 26°S on alluvial fans along the eastern margin of the Coastal Cordillera. Deposits range from a few km to 30 km in width (Ericksen, 1981). Our current understanding of the distribution of highgrade nitrate deposits is based upon the location of nitrate mines and mining claims from the 19th and early 20th century. It is unknown if high-grade nitrate deposits exist outside of these areas. Noncommercial nitrate deposits (1 to 7%) are more widespread, but their spatial distribution is not well known. Most commercial nitrate deposits are less than 2000 m in elevation, but deposits are known to exist up to 4000 m (Ericksen, 1981). These nitrate deposits are not restricted to alluvial fans, but occur on hilltops and in high valleys, as well as in crusts over some salars. Pedogenic nitrate deposits (termed caliche) occur in soils ranging between 3 and 13 m in thickness and consist of several horizons. Local soil horizon designations are chuca—a surface horizon (⬃30 cm) of powdery gypsum and anhydrite; costra—50 cm to 2 m thick horizon below the chuca composed of firmly-cemented gypsum and anhydrite; caliche—a 1 to 5 m firmly-cemented horizon below costra containing nitrate and other salts; conjelo—a horizon up to 2 m thick of saline-cemented regolith; and coba—a loose, unconsolidated regolith (Fig. 2) (Ericksen, 1981).

577

Little has been reported on the spatial distribution of soluble salts in Atacama soils outside of the high-grade nitrate deposits. Berger and Cooke (1997), studied the spatial distribution of soil salts on three alluvial fans in the Atacama and Andes, found higher concentrations of soil carbonate in the Andes, but recorded high levels of sulfates at all locations. The regional distribution of salts in soils is likely a combination of distance from local dust sources, deposition rates of local dust versus atmospheric dust, and precipitation gradients. The boundary between soluble soil salts in the Atacama and less-soluble soil carbonate, which is almost absent from salars and soils at low elevations but present at higher elevations in the Atacama, is also unknown. 3. METHODS 3.1. Transect Collection We collected ⬃60 soil gypsum/anhydrite and soil carbonate samples along three east/west-trending transects (Fig. 1). The first transect extends more than 200 km from the Pacific Ocean just south of Antofagasta, through Quebrada la Negra, to the Cordillera Domeyko. This area has a large breach in the Coastal Cordillera and therefore should be strongly influenced by marine aerosols. The second transect runs from Tocopilla, situated on the Pacific Coast ⬃180 km to the north of Antofagasta, into high-grade nitrate deposits in the Central Valley, a distance of ⬃65 km. In this area, the Coastal Cordillera abruptly rises to ⬃1400 m in elevation and blocks marine coastal fog from penetrating inland. The third transect runs from southwest to northeast for ⬃60 km from the Cordillera Domeyko, through the Llano de la Paciencia, and back into the Cordillera Domeyko and Andes. Most soil gypsum samples were collected from the top of the cemented costra horizon, from a depth of ⬃20 to 30 cm. In some cases, samples were collected from exposed costra horizons, or in soils containing less-developed (i.e., lacking costra) gypsic horizons. Soil carbonate analyzed in this study was collected from various depths and from soils displaying stage III carbonate development (sensu, Gile et al., 1966). We also collected water and lake/evaporite samples to identify the isotopic range of local Andean weathering. Water samples were collected from perennial streams in the Andes between 2500 to 3500 m in elevation. Evaporite salts were collected from lakes in the Andes (⬎4000 m), from salars at the base of the Andes (2300 to 3000 m), and from salars in the Central Valley (500 to 700 m). 3.2. Analytical Methods Approximately 200 mg of soil gypsum/anhydrite or evaporite salts was dissolved in warm 2 N HCl, decanted, and mixed with BaCl2 to precipitate all S as BaSO4. 2 N HCl was added to water samples to remove HCO3⫺, and mixed with BaCl2 to precipitate all BaSO4. Fifteen mg of BaSO4 was then combined with 60 mg Cu2O and 60 mg SiO2, combusted under vacuum at 1100° C, and cryogenically purified to SO2 (Coleman and Moore, 1978). SO2 gas was analyzed on a VG gassource mass spectrometer at the University of Arizona and is presented in ␦ 34S notation (34S/32Ssample/34S/32Sstd ⫺1) ⫻ 1000 as ‰, where 34 32 S/ Sstd is the Can˜ on Diablo Troilite (CDT) standard. For strontium analysis, ⬃50 to 100 mg of soil gypsum was dissolved in warm ultra-pure H2O in an ultrasonic bath for several hours. Soil carbonate was dissolved in doubly distilled 1 N acetic acid in an ultrasonic bath for half an hour. All solutions were passed through cation exchange columns to isolate strontium. Strontium isotopic ratios were measured on a VG Sector 54 mass spectrometer at the University of Arizona. An internal standard was used for 34S that was calibrated relative to NBS 127. 34S standards were within ⫾ 0.2 ‰ of the calibrated value. NBS-987 Sr standards run in association with samples (n⫽8) averaged 0.7102293 ⫾ 0.0000071. X-ray diffraction samples were mounted on a well slide with random orientation and analyzed on a Siemens D-500 X-ray diffractometer.

578

J. A. Rech, J. Quade and W. S. Hart

Fig. 2. Physical characteristics of Atacama soils. A) Typical soil surface devoid of plants and surficial drainage patterns, B) The top of a costra soil horizon that is actively being eroded by wind deflation. Polygonal costra blocks are between 25 to 50 cm across. C) Exposed upper 3 m of a well-developed Atacama soil (AT 623 to 626) downwind from Salar Mar Muerto. Soil is primarily gypsum with large fracture infillings of halite. D) Soil gypsum from the eastern Cordillera Domeyko.

4. RESULTS AND DISCUSSION

4.1. Physical Characteristics of Soils 4.1.1. Antofagasta transect Soil development in the western Coastal Cordillera and along the ridge top below ⬃1000 m in elevation is minimal and confined mostly to alluvial fans. Secondary gypsum is present in the form of isolated crystals between 2 and 5 mm or as cemented horizons ⬃5 cm thick (e.g., soils associated with samples AT 241, 243, 245). This gypsum is relatively translucent. Extremely well-developed soils are present along the western slope of the Sierra del Tigre, just inland of the Coastal Cordillera (e.g., AT 234, 620, 621, 622, 247, 249). Soils are white in color and typically are composed of ⬎90% gypsum. The thickness of these soils is unknown, but is likely several meters. In most locations, soils contain 10 to 20 cm of fine, white, gypsic dust with some anhydrite on the surface (chuca). Below the chuca, there is a soft, weakly cemented and extremely porous horizon of gypsum. This horizon overlies a hard, well-cemented gypsum horizon (costra), commonly composed of large (20 to 50 cm) blocks (Fig. 2b). These welldeveloped gypsic soils blanket the landscape up to the western slope of the Cordillera Domeyko, ⬃100 km from the Pacific Coast. In this area of the Cordillera Domeyko, gypsic soils are

less developed and not as extensive across the landscape as areas closer to the coast. Samples of soil gypsum were found in road cuts, in exposures of active washes, and as gypsic crusts eroding on the surface. These gypsic horizons are generally ⬍25 cm thick and are commonly light brown in color due to the inclusion of detrital particles. 4.1.2. Tocopilla transect Soil development is minimal along the western slope of the Coastal Cordillera below ⬃800 m in elevation. The soil associated with sample G1, at 450 m in elevation, was weakly cemented and contained ⬍10% gypsum. However, sample G3 at 880 m in elevation was from a soil with a white, gypsic horizon ⬃30 cm thick and containing ⬃50% gypsum. Soils along the remainder of the transect, beginning with sample G5 at 1020 m in elevation, are well-developed soils of unknown thickness. This transect ends on alluvial fans on the eastern side of the Coastal Cordillera, an area of commercial nitrate deposits and the location of Estancia Maria Elena, the only nitrate works still in operation today. 4.1.3. Llano de la paciencia transect Gypsic horizons between 20 to 50 cm thick occur throughout the Llano de la Paciencia and up to ⬃2900 m in elevation in the

Isotopic evidence for the source of soil salts in the Atacama Desert, Chile

eastern Cordillera Domeyko. Weakly developed Stage I (sensu Gile et al., 1966) soil carbonate, however, is also present on young surfaces at 2500 m in elevation in the northeastern portion of the Llano de la Paciencia. Soils containing soil carbonate and/or soil gypsum overlap between 2500 to 2900 m in elevation. Soil with Stage III carbonate development occurs mainly in a zone between ⬃3000 to 3300 m in the eastern Cordillera Domeyko and Andes. Carbonate coatings on clasts and dispersed in soil matrix (Stages I-II) are found up to ⬃4000 m in elevation in the Andes. 4.2. X-ray Diffraction Seven samples (AT 247, 252, 258, 264, 282, 292, 622) were selected for X-ray diffraction to identify the main salt minerals in soil cements. These samples were selected on the basis of clarity/color (e.g., clear/translucent, white, light brown), texture (porous and non-porous), and their distance from the coast. All samples are from well-cemented salt horizons from a depth of 20 to 30 cm. Sample AT 622 is from a vertical fracture infilling of salts between costra blocks (Fig. 2C). All samples are mostly gypsum, except AT 622, which is halite. Samples AT 247, 252, and 264 contain lesser amounts of anhydrite (⬃5 to 20%), whereas AT 258 contains greater amounts of anhydrite (⬃40%). Samples AT 282 and AT 292 from soil along the eastern portion of the transect lack anhydrite. 4.3. Isotopic Results and Discussion Isotopic values of modern marine sulfur in sulfate and strontium are well known, with a ␦ 34S value of 20.9‰ (Rees et al., 1978) and a 87Sr/86Sr value of 0.7091 (Burke et al., 1982). However, isotopic values of biogenic sulfur such as dimethylsulfide (DMS) released from the upwelling Humboldt Current are not well known. Bao et al. (2000) report a range for ␦34S values of oceanic biogenic sulfur between 13 to 22‰. Below we present ␦34S and 87Sr/86Sr values for salts in streams, lakes, and salars in the Atacama and adjacent Andes to identify the mean and range of regional Andean weathering products. Andean ␦34Ssulfate and 87Sr/86Sr isotopic values should be similar to mantle or juvenile igneous rocks, with ␦34Ssulfate values of approximately ⫺5 to ⫹5‰ and 87Sr/86Sr isotopic values of 0.7050 to 0.7075. Thus, marine and Andean sources of Ca and S should be readily distinguishable from one another. Ratios of Ca/Sr in marine coastal fog and Andean salar salts were found to be roughly equivalent (Marine fog ⫽ 480; Schemenauer and Cereceda, 1992; Andean weathering ⫽ 730; based on XRF analyses of 7 lake/salar salts), allowing Sr to be used as a proxy for Ca in simple mixing calculations. Distinguishing between atmospheric sulfate, which has a ␦34S value ⬃2.5‰ (Castleman et al., 1974), and Andean sulfur is probably not possible. Isotope results are presented in Table 1.

are 0.707575 and 0.707566, respectively. An aquatic mollusk from Rı´o Salado has a 87Sr/86Sr value of 0.707625. These and other streams in the High Atacama are strongly influenced by ground-water recharge that has undergone significant water-rock interaction. This is evidenced by high concentrations of dissolved solutes in lakes and streams, especially SO42⫺, SiO2, and Ca2⫹, and low percent modern carbon (pMC) values (Fritz et al., 1978; Magaritz et al., 1989; Aravena and Suzuki, 1990). Rı´o Salado and Rı´o Tula´ n in particular are both strongly influenced by local geothermal activity. Rı´o Tula´ n begins within 50 km of the active Volc␣n Lascar and Rı´o Salado commences at the El Tatio geothermal field. Therefore, the above range of values in Rı´os Salado, Tula´ n, and Puritama should provide an average value of local Andean rocks. Three samples were collected from lakes in the High Andes (Laguna Miscanti, Laguna Lejı´a, and Salar de Aguas Calientes), three from salars at the base of the Andes (Salar de Atacama, Salar de Punta Negra, and a small, unnamed salar in the Llano de la Paciencia), and two from small salars in the Central Valley (Salar del Carmen and Salar de Navidad). All salars are evaporite salars, with salt crusts produced from the capillary migration and evaporation of ground water. Salar del Carmen, in the Central Valley, is one of the few salars that contained significant quantities of nitrate for mining (Stoertz and Ericksen, 1974). The ␦34S values of these samples are ⫹5.3‰, ⫹7.6‰, and ⫹2.9‰ for High Andean lakes, ⫹2.8‰, ⫹6.5‰, and ⫹4.5‰ for salars at the base of the Andes, and ⫹5.4‰, and ⫹3.0 for salars in the Central Valley. A sample of secondary vein gypsum from the Cordillera de la Sal, which is composed of Neogene Salar de Atacama evaporites, returned a ␦34S value of ⫹3.7‰. Salts from the High Andean lakes of Laguna Miscanti and Laguna Lejı´a have 87Sr/86Sr values of 0.707341 and 0.707600, respectively. The 87Sr/86Sr values of salts from salars at the base of the Andes are 0.708210 (Salar de Atacama), 0.707961 (Salar de Punta Negra), and 0.706682 (unnamed salar in Llano de la Paciencia). Salar salts from the Central Valley have 87 Sr/86Sr values of 0.706844 for Salar del Carmen and 0.707469 for Salar de Navidad. Carmona et al. (2000) reported 87 Sr/86Sr and ␦34S values that ranged between 0.7076 and 0.7082 and between ⫹5 to ⫹9‰ for the Salar de Atacama. Combining all these results, stream, lake, and salar salts in this study area give an average ␦34S value of ⫹5.4 ⫾ 2‰, ranging from ⫹2.9 to 8.7‰, and an average 87Sr/86Sr value of 0.70749 ⫾ 0.00045, ranging from 0.706682 to 0.708210 (Fig. 3). 87Sr/86Sr values from Salar de Atacama and Punta Negra are slightly higher than Andean stream water, and may represent water-rock interactions with marine rocks in the Cordillera Domeyko. In the next section of the paper these average values are used in a simple mass-balance calculation to determine the relative contribution of marine aerosols and Andean dust in soil salts: x(87Sr/86Sr)marine ⫹ (1 ⫺ x)(87Sr/86Sr)Andes ⫽ (87Sr/86Sr)soil salt (1)

4.3.1. Streams and salars/lakes The ␦34S values of water samples from three streams along the western slope of the Andes are ⫹6.8‰ for Rı´o Salado, ⫹8.7‰ for Rı´o Puritama, and ⫹7.6‰ for Rı´o Tula´ n. The 87 Sr/86Sr values of an aquatic mollusk and tufa from Rı´o Tula´ n

579

and x(␦ 34S)marine ⫹ (1 ⫺ x)(␦ 34S)Andes ⫽ (␦ 34S)soil salt

(2)

where x is the proportion of sulfur or strontium contributed by

580

J. A. Rech, J. Quade and W. S. Hart

Fig. 3. ␦34S and 87Sr/86Sr values of stream water and lake/evaporite salts. Average ␦34S and 87Sr/86Sr values are represented as horizontal lines at 5.4‰ and 0.70749, respectively. Lower section of figure is a generalized topographic profile with distance inland.

marine aerosols, 87Sr/86Srmarine ⫽ 0.7091 (Burke et al., 1982), 87 Sr/86SrAndes ⫽ 0.70749, ␦ 34Smarine sulfate ⫽ ⫹20.9‰ (Rees et al., 1978), and ␦34SAndes ⫽ ⫹5.4‰. 4.3.2. Soils 4.3.2.1. Antofagasta transect The ␦34S and 87Sr/86Sr values of gypsum/anhydrite samples in soils along the Antofagasta transect systematically decrease with distance from the Pacific Ocean (Fig. 4). The ␦34S values of soil gypsum closest to the coast approach marine isotopic values, with ␦34S values as high as ⫹18.3‰ and 87Sr/86Sr ratios up to 0.708739. Soil gypsum within 50 km of the coast, and below 800 m in elevation,

Fig. 4. Antofagasta transect. ␦34S and 87Sr/86Sr values of soil gypsum/anhydrite with distance from coast. In Figures 4 – 6, horizontal lines represent the average ␦34Ssulfate and 87Sr/86Sr values of seawater, local Andean weathering, and a 50% mixing ratio between these two end members. Lower section of figure is the topographic profile along the transect.

generally has ␦34S values ⬎ ⫹14‰ and 87Sr/86Sr ratios ⬎ 0.70825. Mixing Eqn. (1) and (2) indicate that ⬎50% of gypsum/anhydrite S and Sr is of marine origin (Fig. 4). Gypsum in soils also in this coastal region but directly downwind (east) from salars (Fig. 1; AT 620, 623, 624, 625, 626) and soils north of the breach in the Coastal Cordillera (Fig. 1; AT 234, AT 621) have lower ␦ 34S values (Fig. 4). Sample AT 620, to the east of Salar del Carmen, has a ␦34S value of ⫹7.4‰, whereas two samples from soils just to the south (AT 619, AT 247) both have ␦34S values of ⫹18.1‰. Four gypsum samples (AT 623, ⫺250 cm; 624, ⫺150 cm; 625, ⫺60 cm; and 626, ⫺35 cm) from the exposed upper 3 m of a soil profile just east of Salar Mar Muerto (Fig. 2C) have values between ⫹7.5 and ⫹8.1‰, whereas two soil gypsum samples (AT 256, AT 258) from soils to the southwest and south have values of ⫹13.3‰ and ⫹10.9‰, respectively. Soil gypsum samples (AT 234, AT 621) to the north of the main breach in the Andes have ␦34S values of ⫹12.4‰ and 14.1‰. Soil gypsum/anhydrite beyond ⬃90 km from the Pacific Ocean, and above 1300 m in elevation, have ␦34S and 87Sr/86Sr values equivalent to Andean weathering (Fig. 4). The ␦34S values for soil gypsum spanning 90 to 220 km from the coast are between ⫹5.2‰ and ⫹7.5‰ and average ⫹6.5‰. A soil gypsum horizon that is 175 km from the coast has a 87Sr/86Sr value of 0.707210 (AT 282). These results indicate that marine aerosols are a major source of S and Sr in soils below 800 m in elevation and within 50 km of the coast and are a less important source of aerosols in soils located between 1300 m and 800 m in elevation and within 90 km of the coast. Moreover, S and Sr in gypsum/anhydrite in soils close to the coast, yet downwind from salars (east), are of mainly Andean origin. The relatively low ␦34S values of soils north of the main breach in the Coastal Cordillera indicate a lesser marine input than comparable soils along the main breach, indicating a topographic influence on the deposition of marine aerosols. The down-profile consistency of ␦34S values suggests that the processes regulating the deposition of salts to soils have not changed for an extended period of time.

4.3.2.2. Tocopilla transect The ␦34S and 87Sr/86Sr values of soil gypsum/anhydrite along the Tocopilla transect systematically decrease with distance from the coast, similar to the Antofagasta transect soils (Fig. 5). However, ␦34S and 87Sr/86Sr values are lower than values from soil gypsum/anhydrite along the Antofagasta transect that are from a comparable distance from the coast. For example, all samples within 50 km of the coast along the Antofagasta transect, except those downwind from salars or to the north of the main breach in the Coastal Cordillera, have ␦34S values ⬎14 ‰ and receive ⬎50% of their sulfur from the Pacific Ocean (Eqn. 2). In contrast, only one sample from the Tocopilla transect, sample G3 from an elevation of 880 m and 12 km from the coast, has a ␦34S value ⬎14 ‰. The 87 Sr/86Sr ratios of soil gypsum along the Tocopilla transect also indicate a relatively minor, and perhaps an even lower input of marine Sr than marine S to soils. Sample G 3 has a 87Sr/86Sr value of 0.707763, indicating a low input of marine Sr. Decreasing 87Sr/86Sr ratios with distance from coast, however, is suggestive of some input of marine Sr to soils (Fig. 5).

Isotopic evidence for the source of soil salts in the Atacama Desert, Chile

581

Fig. 7. ␦34S versus 87Sr/86Sr of lake/salar and soil gypsum samples. Shaded area represents the distribution of values reported by Carmona et al. (2000) for Salar de Atacama. Fig. 5. Tocopilla transect. ␦34S and 87Sr/86Sr values of soil gypsum/ anhydrite with distance from coast.

The significant differences between soils along the Antofagasta and Tocopilla transects demonstrate the strong blocking effect of the Coastal Cordillera, preventing the incursion of marine aerosols associated with the coastal fog. The relatively low ␦34S value of ⫹12.0‰ for sample G1, at 450 m in elevation and 8 km from the coast, is likely the result of local weathering of volcanic rocks in the Coastal Cordillera. This sample was from a thin (⬍5 cm) gypsic horizon that contained ⬍10% gypsum. 4.3.2.3. Llano de la Paciencia transect Soil gypsum/anhydrite samples along the Llano de la Paciencia transect have ␦34S and 87 Sr/86Sr values that are close to the averages of 5.4‰ and 0.70749 determined for Andean ground water and salars (Fig. 6). Two samples, AT 312 and AT 314, have ␦34S values of ⫹8.3 and ⫹11.0‰, which are above average for most Andean weathering. The 87Sr/86Sr values from soil gypsum (between 2500 to 2800 m) and soil carbonates (between 3000 to 3200 m)

in the Eastern Cordillera Domeyko and Andes are slightly lower than the average value determined for Andean weathering (Fig. 6). The ␦34S and 87Sr/86Sr values for soil gypsum/anhydrite and soil carbonate along this transect suggest that S and Sr are derived from Andean weathering, and possibly some in situ weathering of bedrock in this wetter region of the Atacama. The high ␦ 34S of samples AT 312 and AT 314 may reflect surface weathering of marine shales, or be the result of salts derived from the Salar de Atacama. Slightly lower 87Sr/86Sr values of soil gypsum/anhydrite and soil carbonate could also be the result of surficial bedrock weathering. However, it is not possible to distinguish between the relative contributions of local dust versus in situ bedrock weathering in most areas. 4.3.3. Summary Our 87Sr/86Sr and ␦34S results from Atacama soils, summarized in Figure 7, fall along a mixing line defined by seawater values (0.7091; 20.9‰) near the coast and average Andean weathering values (0.70749; 5.4‰), inland. In detail, however, local topography and transport pathways cause deviations from this general trend. The 87Sr/86Sr and ␦34S isotopic values of streams, lakes, and salars, which we used to identify the range of values for local Andean weathering, are generally between 0.7067 to 0.7076 and 3 to 7‰, respectively (Fig. 7). The two highest 87Sr/86Sr values for lakes and salars are from the Salar de Atacama and Salar de Punta Negra, with values of 0.708210 and 0.707961. These high values are likely the result of waterrock interactions within these deep aquifers with rocks that are not present along ground-water flow paths elsewhere. The intersection of the soil mixing lines with the lower 87Sr/86Sr values of stream, lake and salar salts suggests that most Andean weathering products in soil salts are from rocks with lower 87 Sr/86Sr values, such as Tertiary volcanics. 5. ORIGIN OF SOIL SALTS IN THE ATACAMA

Fig. 6. Llano de la Paciencia Transect. ␦34S and 87Sr/86Sr values of soil gypsum/anhydrite and soil carbonates with distance from coast.

Our results demonstrate that there are multiple source regions for pedogenic salts in the Atacama. The proximity of soils to local dust and aerosol sources, such as salars and marine coastal fog, significantly influences the source and deposition rates of salts in Atacama soils. In this section, we use our ␦34S and 87Sr/86Sr evidence from soil gypsum/anhy-

582

J. A. Rech, J. Quade and W. S. Hart

drite and soil carbonate in conjunction with results from previous studies to evaluate the contribution of the various source regions to Atacama soils across the landscape. Marine aerosols associated with coastal fog, located at an altitude of ⬃300 to 1000 m in this region of the Atacama (Schemenauer et al., 1987), have long been considered a source for sulfur, nitrogen, calcium, and other elements found in Atacama soil salts (Ericksen, 1981; Searl and Rankin, 1993; Berger and Cooke, 1997). However, there has been no strong evidence to support this argument. Our ␦34S and 87Sr/86Sr results from soils along the Antofagasta transect are the first conclusive evidence to demonstrate a significant marine influence to Atacama soils. This influence, however, is restricted to locations where the coastal fog can penetrate inland, i.e., where the Coastal Cordillera is ⬍1000 m. This excludes areas in the Atacama where soils contain high levels of nitrate, perchlorate, and iodate. The upwelling coastal waters of Chile and associated coastal fog contain high concentrations of nutrients and dissolved solutes. The average concentrations of dissolved solutes in coastal fog are 9.12 ppm SO4; 0.8 ppm NO3; 1.8 ppm Cl; 4.81 ppm Na; 0.37 ppm K; 0.96 ppm Ca; 0.62 ppm Mg; (Schemenauer and Cereceda, 1992). These solutes are incorporated into coastal fog by sea spray and the outgassing of biogenic sulfur, which forms on the sea floor by bacterial reaction (Watson, 1988). As fog moves inland over the Atacama, it evaporates and deposits all dissolved solutes. Annual average flux rates for coastal fog at 700 m in elevation are 3 L/vertical m2/d (Cereceda et al., 1997). Studies of the coastal fog have also identified a large, yet unquantified, component of marine aerosols that are deposited as dry deposition (Schemenauer and Cereceda, 1992). Assuming these values for concentrations of dissolved solutes and average flux rates for coastal fog, 1 m3 of soil gypsum at 700 m takes ⬃1.9 Ma to form from coastal fog (assuming Ca is the limiting ion, a topographic slope of 10% which would equate to a flux rate of 0.3 L/m2/d, 1.7 g/mL bulk density, and 50% pore space). This would be a maximum age because it does not account for dry deposition of marine aerosols, although it also assumes 100% salt retention, which is unlikely. In a study of soil sulfates in the dry valleys of Antarctica, Bao et al. (2000) found that soils from upland valleys and plateaus possessed lower ␦18O and higher ⌬17O values than soils from inland valley floors and sides. The authors attributed these differences to a greater contribution of sea salt sulfate, which is a coarse aerosol (micrometers), to areas adjacent to the coast, whereas biogenic (non-sea salt) sulfate, which is a fine aerosol (submicrometers), is transported further inland and to higher altitudes. We do not think that this explanation can account for the isotopic trends observed in our soil transects. Although biogenic sulfur compounds may influence soils further inland and above the elevation of the coastal fog zone, the close agreement between the ␦34S and 87Sr/86Sr values of inland soil gypsum and local Andean weathering products suggests that most inland soil salts are derived from salar salts that are redistributed by eolian processes. The bedrock of the Andes and Cordillera Domeyko contain significant sulfur, calcium, chlorine, and others elements found in Atacama soil salts. Surface weathering in the Atacama is limited by low temperatures at high elevations and by low

precipitation at low elevations. However, there can be significant chemical alteration in the subsurface from ground waterrock interactions. Because only one stream in the central Atacama flows into the Pacific Ocean, the Rı´o Loa with a discharge of ⬃1 m3/s, most ground water is discharged into salars where it eventually evaporates and deposits all dissolved solutes. Studies of a core from the Salar de Atacama (⬃3000 km2) show that evaporitic salts are deposited at rate of ⬃1 m/1000 yr (Bobst et al., 2001). Our results suggest that elements derived from the evaporation of ground water in locations where the water table is close to the surface, such as evaporite salars, are the main source of Ca, S, and likely many other salts found in Atacama soil salts. Two lines of evidence support this conclusion. Firstly, the average ␦ 34S and 87Sr/86Sr values that we found for Atacama and Andean ground water closely matches values found in non-coastal soil gypsum/anhydrite and carbonate. Secondly, soils downwind from salars receive large quantities of salarderived salt, as indicated by ␦ 34S and 87Sr/86Sr values. At present, we are not able to distinguish the relative role of highor low-temperature water-rock interactions in the Andes, the incorporation of volcanic gases in ground water, and waterrock interactions of ground water with local bedrock in the Atacama. All of these sources may provide significant levels of dissolved solutes to Atacama ground water. Two other sources for dissolved salts in Atacama soils that have been suggested are salts dissolved in precipitation (Berger and Cooke, 1997), and direct deposition of volcanic emissions associated with volcanic eruptions (Searl and Rankin, 1993; Berger and Cooke, 1997). We do not think these processes contribute significant amounts of salts to Atacama soils. Precipitation in the core of the Atacama is extremely low (⬍5mm/ yr), and therefore salts dissolved in precipitation (Risacher and Fritz, 1991) are negligible compared to those in ground water discharging into streams and salars. There have been numerous volcanic eruptions in the Andes since the mid-Miocene, such as the recent eruption of Volca´ n Lascar in 1993. However, most emissions associated with these eruptions are transported to the East by strong westerly winds. Therefore, we suggest that most sulfur and other elements associated with volcanic eruptions in the Andes are transported to the east, away from the Atacama. We favor eolian transport as a means of dispersing salts across the Atacama. Some early researchers and Searl and Rankin (1993) suggest that capillary action of ground water and/or Andean flooding can form Atacama salt deposits. Flooding or elevated ground water levels, however, could not have played a role in the formation of deposits located on hilltops or mesas. Also, salt deposits on the alluvial fans from the Coastal Cordillera are located well above any water table fluctuations that may have occurred during the Pleistocene. The dry deposition of extralocal atmospheric particulates on this extremely hyperarid, stable land surface has been suggested as an important source of nitrogen, and perhaps of perchlorate and iodate in the nitrate deposits (Claridge and Campbell, 1968; Ericksen, 1981; Bo¨ hlke et al., 1997). Ericksen (1981) ultimately rejected dry deposition of extra-local atmospheric particles because the explanation necessitated accumulation times of 1 to 2 Ma y, which he thought was too long. However, recent analyses of ␦18O in nitrate by Bo¨ hlke et al. (1997) identified values between ⫹31 and ⫹50‰ (V-SMOW),

Isotopic evidence for the source of soil salts in the Atacama Desert, Chile Table 1. Location, ␦

Sample #

elevation (m)

latitude (S)

Streams AT640 2500 22° 19' AT106 2500 22° 19' AT641 3200 22° 45' AT642 2600 23° 48' AT41 2600 23° 48' AT63 2600 23° 48' Salar and lake salts AT386 4350 23° 31' AT390 4200 23° 31' AT398 4350 23° 45' AT25 2350 23° 38' AT643 2950 24° 35' AT310 2325 23° 11' AT239 500 23° 48' AT250 665 23° 39' AT644 2500 22° 55' Soil Gypsum (Antofagasta transect) AT245 425 23° 40' AT243 785 23° 42' AT241 710 23° 42' AT247 575 23° 45' AT619 550 23° 44' AT620 565 23° 43' AT234 665 23° 35' AT621 685 23° 30' AT249 710 23° 45' AT252 675 23° 46' AT254 710 23° 47' AT256 785 23° 50' AT623 715 23° 47' AT624 715 same AT625 715 same AT626 715 same AT258 845 23° 53' AT260 915 23° 55' AT262 1105 23° 59' AT264 1305 24° 02' AT267 1505 24° 04' AT274 2220 24° 11' AT278 3095 24° 14' AT280 3100 24° 14' AT282 2980 24° 14' AT283 3340 24° 13' AT288 3410 24° 09' AT292 3050 24° 06' Soil Gypsum (Tocopilla transect) G1 450 22° 05' G3 880 22° 06' G5 1020 22° 06' G7 1210 22° 06' G9 1380 22° 09' G11 1420 22° 13' G13 1450 22° 15' G15 1470 22° 16' G17 1330 22° 16' G19 1240 22° 16' G21 1210 22° 16' G25 1200 22° 11' G27B 1150 22° 06' G28 1130 22° 07' Soil Gypsum (Llano de la Pacienca transect) AT299 2850 23° 17' AT302 2900 23° 22' AT306 2985 23° 22' AT308 2555 23° 21'

34

S, and

longitude (W)

87

583

Sr/86Sr values of water, evaporites, and soil salts.

Distance from Coast (km)

␦34S (‰)

87

Sr/86Sr

⫹6.8

68° 68° 68° 68° 68° 68°

35' 35' 04' 07' 07' 07'

170 170 200 200 200 200

⫹8.7 ⫹7.6

67° 67° 67° 68° 68° 68° 70° 70° 68°

41' 36' 47' 21' 56' 31' 16' 04' 10'

280 300 280 200 160 200 20 40 190

⫹7.6 ⫹2.9 ⫹5.3 ⫹2.8 ⫹6.5 ⫹4.5 ⫹5.4 ⫹3.0 ⫹3.7

70° 22' 70° 21' 70° 18' 70° 17' 70° 16' 70° 15' 70° 13' 70° 11' 70° 10' 70° 04' 69° 59' 69° 54' 69° 46' same same same 69° 49' 69° 46' 69° 41' 69° 35' 69° 30' 69° 17' 68° 57' 68° 50' 68° 47' 68° 38' 68° 28' 68° 19'

4 6 8 17 15 15 17 22 25 35 45 55 60 60 60 60 65 75 80 90 100 120 155 165 175 185 205 220

⫹18.3 ⫹17.1 ⫹16.6 ⫹18.1 ⫹18.1 ⫹7.4 ⫹12.4 ⫹14.1 ⫹16.9 ⫹15.7 ⫹12.4 ⫹13.3 ⫹7.5 ⫹7.9 ⫹8.1 ⫹8.1 ⫹10.9 ⫹11.8 ⫹10.6 ⫹6.2 ⫹7.5 ⫹7.5 ⫹5.2 ⫹6.6 ⫹6.1 ⫹6.3 ⫹7.2 ⫹5.4

0.707625 0.707575 0.707566

70° 70° 70° 70° 69° 69° 69° 69° 69° 69° 69° 69° 69° 69°

09' 07' 05' 02' 59' 57' 53' 50' 45' 41' 38' 37' 38' 38'

8 12 15 20 23 25 32 43 50 55 65 65 65 65

⫹12.0 ⫹14.0 ⫹10.9 ⫹10.5 ⫹7.0 ⫹9.2 ⫹8.1 ⫹7.8 ⫹8.7 ⫹8.5 ⫹6.0 ⫹4.1 ⫹5.9 ⫹7.1

68° 68° 68° 68°

49' 48' 42' 38'

158 160 171 178

⫹5.6 ⫹4.3 ⫹5.5 ⫹6.6

0.707600 0.707341 0.708210 0.707961 0.706682 0.706844 0.707469

0.708739

0.708440 0.708267

0.707607 0.707363

0.707210

0.707763 0.707116 0.706989 0.707015

0.707071

*Description Rio Rio Rio Rio Rio Rio

Salado Salado (mollusk) Puritama Tula´ n Tula´ n (mollusk) Tula´ n (tufa)

Laguna Lejı´a Salar de Agua Caliente Laguna Miscante Salar de Atacama Salar de Punta Negra Salar (Llano de la Paciencia) Salar del Carmen Salar de Navidad Cord. de la Sal (vein) Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Vein Stage

1 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2 3 3 3 1 1 2 1 3 2

Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage Stage

1 3 3 3 3 3 3 3 3 3 3 3 3 3

Stage 1 Stage 3 Stage 3 (continued)

584

J. A. Rech, J. Quade and W. S. Hart Table 1. (Continued)

Sample # AT309 AT312 AT314 AT326 Soil Carbonate AT13 AT92 AT13b (250) AT329

elevation (m)

latitude (S)

longitude (W)

Distance from Coast (km)

␦34S (‰) ⫹5.4 ⫹8.3 ⫹11.0 ⫹5.6

2520 2400 2450 2850

23° 23° 23° 22°

17' 05' 03' 49'

68° 68° 68° 68°

36' 29' 27' 21'

180 193 196 207

3075 3200 3100 3180

22° 22° 22° 22°

41' 35' 21' 46'

68° 68° 68° 68°

29' 27' 21' 23'

193 196 190 203

87

Sr/86Sr

0.707072 0.706971 0.707161 0.707185 0.707034

*Description Surface Stage 3 Stage 1 Stage 3 Stage Stage Stage Stage

3 3 3 3

* Stage 1 through stage 3 development is modified from soil carbonate descriptions from the American Southwest (Gile et al., 1966); Stage 1 - soil gypsum coatings and partial filling of void space with gypsum, Stage 2 - soil gypsum completely filling void space and between 5 to 50 cm thick, Stage 3 - soil gypsum completely filling void space and ⬎ 50 cm thick.

significantly higher than that of atmospheric O2 (⫹23.5‰). These high values are unlikely to result from nitrification of reduced nitrogen, but instead resemble values of atmospheric nitrogen (30 to 70‰, see Bo¨ hlke et al., 1997), thus providing strong evidence for an atmospheric source for nitrates in the Atacama Desert. It is very likely that the iodate and perchlorate associated with the nitrate in these deposits are also atmospheric in origin. We suggest that nitrate mined from salars, such as Salar del Carmen (Ericksen, 1981), is nitrate from soils that has been mobilized during infrequent precipitation events and transported into local ground-water systems. Salar del Carmen is situated at an elevation of 500 m, below the temperature inversion, and therefore receives greater amounts of precipitation than other areas in the Atacama. The Baqaden˜ o Nitrate District represents a large portion of the catchment area for Salar del Carmen, thus providing a significant source of soil nitrate. 6. CONCLUSIONS

The presence of thick, well-developed soils in the Atacama Desert that contain a variety of soluble salts is the result of prolonged hyperaridity in the Atacama and extremely stable land surfaces. The type and origin of soil salts in the Atacama varies according to distance from local dust sources, rates of local versus extra-local dust deposition, and precipitation gradients. We suggest that most salts in Atacama soils result from ground water-bedrock interactions both along ground-water flowpaths from the Andes and in regional aquifers within the Atacama. Salts are deposited in salars when ground water discharges onto the surface or moves to the surface by capillary action and evaporates, leaving behind all dissolved solutes. These salts are then transported to soils by eolian processes. Most salts are transported to soils downwind (east) of salars by westerly winds in the Atacama, but lesser amounts of salar salt are also transported to soils in all directions by local wind gusts. The similar values of both ␦34S and 87Sr/86Sr values of ground water and soil gypsum/anhydrite and calcium carbonate found in this study, and evidence for high deposition rates of salar salts to soils east of salars (downwind), support these conclusions. Although volcanic emissions and salts dissolved in precipitation would have similar ␦34S and 87Sr/86Sr values to those

found among inland Atacama soils, we do not think that these processes contribute significant amounts of salts to soils in the Atacama. Another significant source of local salts to Atacama soils is marine aerosols associated with coastal fog that is present between ⬃300 to 1000 m in altitude along the Chilean coast. The input of marine aerosols to Atacama soils, however, is restricted to locations where the coastal fog can penetrate inland, i.e., locations where there is a breach in the Coastal Cordillera. This input of marine aerosols to soils is visible in the decreasing ␦34S and 87Sr/86Sr values of soil gypsum/anhydrite with distance from coast along our Antofagasta transect, where there is a breach in the Coastal Cordillera, and a minor trend of decreasing isotopic values in our Tocopilla transect, where the Coastal fog is mostly blocked by the Coastal Cordillera. Soils that are both directly downwind from salars and within the influence of the coastal fog possess ␦34S values similar to salars, indicating a greater contribution of salar salts to these soils. Soils with the lowest local-dust deposition rates in the hyperarid core of the Atacama contain the greatest proportions of extra-local dust. We suggest that this region corresponds to the areas containing high-grade nitrate soils, which have been identified as being mostly of atmospheric origin (Bo¨ hlke et al., 1997). These areas are generally located on the leeward side of the Coastal Cordillera and upwind of salars and receive very little rainfall. Outside these settings, high input of marine or local salar salts will tend to dilute atmospheric nitrate fall-out, producing low nitrate concentrations in soils. Perchlorate and iodate, which are found in association with nitrate in soils, are also probably of atmospheric origin. The relative proportion of local versus atmospheric Ca and S in these soils is undetermined. We suggest that most Ca and S come from Andean weathering, although significant amounts of atmospheric Ca and S may be present. Acknowledgments—NSF Grant EAR-9904838 and a grant from the Arizona Geological Society supported this research. A NASA spacegrant administered through the University of Arizona also provided funding for Jason A. Rech. We would like to thank Chris Eastoe for his help with the sulfur isotopic analyses, Wes Bilodeau for X-ray diffraction analyses, Mary Kay O’Rourke and Mari Bartlett for XRF analyses,

Isotopic evidence for the source of soil salts in the Atacama Desert, Chile Jeff Pigati and Christa Placzek for field assistance, and Robert Schemenauer for helpful discussions on the Coastal Fog of Chile. This paper greatly benefited from the comments of two anonymous reviewers and from Allan Chivas. Associate editor: T. E. Cerling REFERENCES Alpers C. N. and Brimhall G. H. (1988) Middle Miocene climatic change in the Atacama Desert, northern Chile: Evidence from supergene mineralization at La Escondita. Geol. Soc. Amer. Bull. 100, 1640 –1656. Alpers C. N. and Whittemore D. O. (1990) Hydrogeochemistry and stable isotopes of ground and surface waters from two adjacent closed basins, Atacama Desert, northern Chile. Applied Geochem. 5, 719 –734. Aravena R. and Suzuki O. (1990) Isotopic evolution of river water in the northern Chile region. Water Res. Res. 26(12), 2887–2895. Bao H., Campbell D. A., Bockheim J. G., Thiemens M. H. (2000) Origins of sulphate in Antarctic dry-valley soils as deduced from anomalous 17O compositions. Nature 407, 499 –502. Berger I. A. and Cooke R. U. (1997) The origin and distribution of salts on alluvial fans in the Atacama Desert, northern Chile. Earth Surf. Proc. Land. 22, 581– 600. Betancourt J. L., Latorre C., Rech J. A., Quade J., and Rylander K. A. (2000) A 22,000-year record of monsoonal precipitation from northern Chile’s Atacama Desert. Science 289, 1542–1546. Biscaye P. E., Grousset F. E., Revel M., Van der Gaast S., Zielinski G. A., Vaars A., and Kukla G. (1997) Asian provenance of glacial dust (stage 2) in the Greenland Ice Sheet Project 2 Ice Core, Summit, Greenland. J. Geophys. Res. 102, 26,765–26,781.. Bobst A. L., Lowenstein T. K., Jordan T. E., Godfrey L. V., Ku T. L., and Luo S. (2001) A 106 ka paleoclimate record from drill core of the Salar de Atacama, northern Chile. Palaeogeol., Palaeoclimatol., Palaeoecol. 173, 21– 42. Bo¨ hlke J. K., Ericksen G. E., and Revesz K. (1997) Stable isotope evidence for an atmospheric origin of desert nitrate deposits in northern Chile and southern California, USA. Chem. Geol. 136, 135–152. Brimhall G. H., Lewis C. J., Ague J. J., Dietrich W. E., Hampel J., Teague T., and Rix P. (1988) Metal enrichment in bauxites by deposition of chemically mature aeolian dust. Nature 333, 819 – 824. Burke W. H., Denison R. E., Heatherington E. A., Koepnick R. B., Nelson N. F., and Otto J. B. (1982) Variation of seawater 87Sr/86Sr throughout Phanerozoic time. Geol. 10, 516 –519. Cameron R. E., Gensel D. R., Blank G. B. (1966) A. Soil Studies— Desert Microflora. XII. Abundance of microflora in soil samples from the Chile Atacama Desert. JPL Space Programs Summary No., 37–38:IV 140 –147. Carmona V., Pueyo J. J., Taberner C., Chong G., and Thirlwall M. (2000) Solute inputs in the Salar de Atacama. J. Geochem. Explor. 69 –70, 449 – 452. Castleman A. W. Jr., Munkelwitz H. R., and Manowitz B. (1974) Isotopic studies of the sulfur component of the stratospheric aerosol layer. Tellus. 26, 222–234. Cereceda P., Schemenauer R. S., and Velasquez F. (1997) Variacio´ n temporal de la niebla en El Tofo-Chungungo, Regio´ n de Coquimbo, Chile (1987–1985). Rev. Geograf. N. Grande 24, 103–111. Chadwick O. A. and Davis J. O. (1990) Soil-forming intervals caused by eolian sediment pulses in the Lahontan basin, northwestern Nevada. Geol. 18, 243–246. Chadwick O. A., Derry L. A., Vitousek P. M., Huebert B. J., and Hedin L. O. (1999) Changing sources of nutrients during four million years of ecosystem development. Nature 397, 491– 497. Chong G. (1988) The Cenozoic saline deposits of the Chilean Andes between 18° and 27° South. In The southern Central Andes (eds. H. Bahlburg, C. Breitkreuz, and P. Giese), pp. 135–151. SpringerVerlag. Claridge G. G. C. and Campbell I. B. (1968) Origin of nitrate deposits. Nature 217, 428 – 430. Coira B., Davidson J., Mpodozis C., and Ramos V. (1982) Tectonic and magmatic evolution of the Andes of northern Argentina and Chile. Earth Sci. Rev. 8, 769 –790.

585

Coleman M. and Moore M. P. (1978) Direct reduction of sulfates to sulfur dioxide for isotopic analysis. Anal. Chem. 50, 1594 –1598. Crutzen P. J. (1974) A review of the upper atmosphere photochemistry. Can. J. Chem. 52, 1569 –1581. Dan J. (1991) The effect of dust deposition on the soils of the land of Israel. Quat. Intl. 3, 107–113. Dymond J., Biscaye P. E., and Rex R. W. (1974) Eolian origin of mica in Hawaiian soils. Geol. Soc. Amer. Bull. 85, 37– 40. Ericksen G. E. (1961) Rhyolite tuff, a source of the salts of northern Chile. USGS Professional Paper. 424-C: C224 –225. Ericksen G. E. (1981) Geology and the origin of the Chilean nitrate deposits. Geol. Soc. Amer. Spec. Paper 1188, 37. Ericksen G. E. (1983) The Chilean nitrate deposits. Am. Sci. 71, 366 –374. Ericksen G. E. (1994) Discussion of a petrographic study of the Chilean nitrates. Geol. Mag. 131, 849 – 852. Flint S., Turner P., Jolley E. J., and Hartley A. J. (1993) Extensional tectonics in convergent margins basins: An example from the Salar de Atacama, Chilean Andes. Geol. Soc. Amer. Bull. 105, 603– 617. Fritz P., Silva Hennings C., Suzuki O., Salati E. (1978) Isotope hydrology in northern Chile. Proceedings of the symposium Isotope Hydrology, 525–543. Gile L. H., Peterson F. F., and Grossman R. B. (1966) Morphological and genetic sequences of carbonate accumulation in desert soils. Soil Sci. 101, 347–360. Kohfeld K. E. and Harrison S. P. (2001) DIRTMAP: the geologic record of dust. Earth-Sci. Rev. 54, 81–114. Magaritz M., Aravena R., Pen˜ a H., Suzuki O., and Grilli A. (1989) Water chemistry and isotope study of streams and springs in northern Chile. J. Hydrol. 108, 323–341. McFadden L. D., Wells S. G., and Jercinovich M. J. (1987) Influences of eolian and pedogenic processes on the origin and evolution of desert pavements. Geol. 15, 504 –508. Muhs D. R., Crittenden R. C., Rosholt J. N., Bush C. A., and Stewart K. C. (1987) Genesis of marine terrace soils, Barbados, West Indies: Evidence from mineralogy and Geochemistry. Earth Surf. Proc. Landforms 12, 605– 618. Naiman Z., Quade J., and Patchett P. J. (2000) Isotopic evidence for eolian recycling of pedogenic carbonate and variations in carbonate dust sources throughout the southwest United States. Geochim. Cosmochim. Acta 64, 3099 –3109. Nakai S., Halliday A. N., and Rea D. K. (1993) Provenance of dust in the Pacific Ocean. Earth Planet. Sci. Lett. 119, 143–157. Noxon J. E. (1976) Atmospheric nitrogen fixation by lightning. Geophys. Res. Lett. 3, 463– 465. Noxon J. E. (1978) Tropospheric NO2. J. Geophys. Res. 83, 3051– 3057. Parrington J. R., Zoller J. R., and Aras N. K. (1983) Asian dust: seasonal transport to Hawaiian Islands. Science 220, 195–197. Pe´ we´ T. L., Pe´ we´ E. A., and Pe´ we´ R. H. (1981) Desert Dust: Characteristics and rates of deposition in central Arizona. Geol. Soc. Am. Spec. Paper. 186, 169 –190. Prospero J. M., Glaccum R. A., and Nees R. T. (1981) Atmospheric transport of soil dust from Africa to South America. Nature 289, 570 –572. Quade J., Chivas A. R., and McCulloch M. T. (1995) Strontium and carbon isotope tracers and the origins of soil carbonate in South Australia and Victoria. Palaeogeog., Palaeoclimatol., Palaeoecol. 113, 103–117. Rea D. K. (1994) The paleoclimatic record provided by eolian deposition in the deep sea: the geological history of wind. Rev. Geophys. 32, 159 –195. Rees C. E., Jenkins W. J., and Monster J. (1978) The sulphur isotopic composition of ocean water sulphate. Geochim. Cosmochim. Acta 42, 377–381. Risacher F. and Fritz B. (1991) Geochemistry of Bolivian salars, Lipez, southern Altiplano: origin of solutes and brine evolution. Geochim. Cosmochim. Acta 55, 687–705. Rundel P. W., Dillon M. O., Palma B., Mooney H. A., Gulmon S. L., and Ehleringer J. R. (1991) The phytogeography and ecology of the coastal Atacama and Peruvian Deserts. ALISO. 13, 1– 49.

586

J. A. Rech, J. Quade and W. S. Hart

Schemenauer R. S., Cereceda P., and Carvajal N. (1987) Measurments of fog water deposition and their relationships to terrain features. J. Clim. Appl. Meteorol. 26, 1285–1291. Schemenauer R. S. and Cereceda P. (1992) The quality of fog water collected for domestic and agricultural use in Chile. J. Appl. Meteorol. 31, 275–290. Searl A. and Rankin S. (1993) A preliminary petrographic study of the Chilean nitrates. Geol. Mag. 130(3), 319 –333. Simonaitis R. and Heicklen J. (1975) Perchloric acid- A possible sink for stratospheric chlorine. Planet. Space Sci. 23, 1567–1569. Simonson R. W. (1995) Airborne dust and its significance to soils. Geoderma 65, 1– 43. Stoertz G. E. and Ericksen G. E. (1974) Geology of salars in northern Chile. Geol. Surv. Prof. Paper 811, 1– 67. Swap R., Grastang M., Greco S., Talbot R., and Kallbert P. (1992) Saharan dust in the Amazon Basin. Tellus 44(B), 133–149.

Tegen I. and Fung I. (1994) Modeling of mineral dust in the atmosphere: Sources, transport, and optical thickness. J. Geophys. Res. 99, 22,897–22,914. Uematsu M., Duce R. A., Prospero J. M., Chen L., Merrill J. T., and McDonald R. L. (1983) Transport of mineral aerosol from Asia over the north Pacific Ocean. J. Geophys. Res. 88, 5343–5352. Uematsu M., Duce R. A., and Prospero J. M. (1985) Deposition of atmospheric mineral particles to the north Pacific Ocean. J. Atmos. Chem. 3, 123–138. van Moort J. C. (1985) Natural enrichment processes of nitrate, sulphate, chloride, iodate, borate, perchlorate, and chromate in the caliches of northern Chile. Congreso Chileno. 4(3), 674 –702. Watson A. (1988) Desert gypsum crusts as paleoenvironmental indicators: A micropetrographic study of crusts from southern Tunisia and the central Namib Desert. J. Arid Env. 15, 19 – 42.