The hyperarid core of the Atacama Desert is considered the dry limit for life on Earth. Soils in this region have very low abundance of heterotrophic bacteria and are practically barren of photosynthetic microorganisms because of the extreme dry conditions (≤2 mm a−1 rainfall). However, relatively abundant endolithic communities of cyanobacteria (Chroococcidiopsis) occur within halite crusts in paleolake evaporitic deposits. By means of continuous monitoring of the microclimate conditions (temperature, relative humidity, water vapor density, wetness, and photosynthetically active radiation) inside and around the halite crusts, we demonstrate here that water vapor condenses within the pore space of the halite at relative humidity (RH) levels that otherwise hinder the occurrence of liquid water in the surrounding environment. Water condensation occurs at RH >75%, which corresponds to the deliquescence point of halite. We have estimated a total of 57 deliquescence events (i.e., water condensation) within the halite crusts, as opposed to only 1 liquid water event outside. These wet events resulted in a total of 213.8 h of potential photosynthetic activity for the endolithic microorganisms versus only 6 h for organisms outside the halite crusts. Halite crusts may therefore represent the last available niche for photosynthetic activity in extreme arid environments on Earth.
 The Atacama Desert (Chile) ranks as the driest desert on Earth, with long-term mean annual rainfall as low as a few millimeters in its driest zone [McKay et al., 2003]. Heterotrophic bacteria are virtually absent in soils from this hyperarid region [Navarro-González et al., 2003] and are also highly depleted in organic molecules partially because of nonbiological oxidation processes [Navarro-González et al., 2003; L. E. Fletcher et al., Variability of organic material in the hyperarid soils of the Atacama Desert, submitted to Journal of Geophysical Research, 2006]. Even hypolithic cyanobacteria, found in hyperarid stony deserts, are extremely rare in the hyperarid core of the Atacama Desert and exist in small spatially isolated islands amidst a microbially depleted soil [Warren-Rhodes et al., 2006]. The absence of soil and hypolithic bacteria is due to the extremely low availability of liquid water, which almost exclusively arrives in the form of fog and dew [Warren-Rhodes et al., 2006]. While the Antarctic Dry Valleys represent a cold extreme for microbial physiology and the Negev Desert represents a hot limit, the hyperarid core of the Atacama Desert represents the dry limit of photosynthetic activity and of primary production [Warren-Rhodes et al., 2006].
 The recent finding of diverse endolithic microorganisms inhabiting halite crusts in the driest part of the Atacama Desert [Wierzchos et al., 2006] shows, however, that even in such extreme arid conditions, ecological niches exist where life can grow in relative abundance and diversity. These halite crusts have a large spatial distribution and characteristic irregular shapes, which are the result of wind erosion and partial dissolution and reprecipitation of evaporitic deposits, during rare and transient wet events. The crusts are composed nearly exclusively of halite (96–99%) with minor amounts of gypsum (1–3%) and traces (∼1%) of sylvine and quartz [Wierzchos et al., 2006]. Colonies of endolithic photosynthetic cyanobacteria of the genus Chroococcidiopsis can be found 3–7 mm beneath the crust surface, distributed within pores and cracks [Wierzchos et al., 2006]. These colonies appear in association with rod-shaped heterotrophic bacteria, suggesting that the crusts can host relatively complex communities.
 The endolithic environment provides microorganisms with mineral nutrients and more favorable moisture regimes than if they were exposed directly to the atmosphere, as well as protection against harmful radiation [Golubic et al., 1981; Billi et al., 2000]. However, halite crusts are likely one of the most challenging environments for life on Earth. Any liquid water within the pore space of the halite will be saturated with NaCl (∼32% weight per volume), resulting in a water activity aw ≤ 0.75. The endolithic communities have to face large temperature oscillations, occasionally more than 50°C in a day cycle, and long periods (several years) without rain [McKay et al., 2003]. The abundance of microorganisms within the halite crusts and their paucity in the surrounding soil suggest that halite may possess particular physical and/or chemical properties that enable the occurrence and preservation of microbial ecosystems in such extreme conditions. Halite is a highly hygroscopic mineral and readily absorbs water vapor at the so-called deliquescence relative humidity (DRH). When DRH is reached, water vapor condenses into saturated aqueous solutions on the crystal surface and/or within the pore space between crystals. The relative humidity at which salts deliquesce is dependent on temperature and is characteristic to each salt mineral or assemblage of salt minerals. In the case of halite, DRH ∼ 75% at T = 25°C [Cohen et al., 1987; Ebert et al., 2002], equivalent to a water vapor density (Wd) of ∼20 g/m3. Rock deliquescence may play an important role in the occurrence of life in this hyperarid desert, where rain and fog are not significant sources of moisture. Halite crystals are also an effective scatterer of UV radiation when precipitated as a mass of small crystals [Cockell and Raven, 2004], and their translucent properties enable the penetration of photosynthetic light to depths of millimeters to centimeters.
 Here we test the hypothesis that the hygroscopic properties of halite can potentially enable the survivability of endolithic microorganisms in extreme hyperarid conditions. We report a series of measurements of relative humidity (RH), Wd, T, wetness, and photosynthetically active radiation (PAR), obtained both inside and outside colonized halite crusts, for a period of 1 year in the hyperarid core of the Atacama Desert.
2. Site Description and Methods
 The study site is located in the Yungay area of the Atacama Desert (24°049′S, 69°591′W, Figure 1a). Yungay is located 60 km from the coast at an altitude of ∼1000 m, and it is flanked by two mountain chains, the coastal mountains to the west (∼1000–3000 m high) and the Domeyko Mountains to the east (∼4000 m high) (Figure 1b). Like the rest of the Atacama, Yungay receives negligible rain. In addition, the coastal mountains block most of the marine fog, the so-called camanchaca, from reaching Yungay except in very rare episodes; this is the only source of humidity.
 Geomorphologically, the study site is covered by 50–60-cm-high halite crusts with an irregular and rough surface and a brownish color that is the result of desert sand and dust trapped on the halite surface. These halite fields are common in the central Atacama and are the result of the evaporation of paleolakes and the subsequent precipitation of massive salt deposits [Pueyo et al., 2001]. Many of the halite crusts contain a 2–5-mm-thick layer of endolithic colonization, 3–7 mm below the crust surface [Wierzchos et al., 2006]. There is virtually no vegetation surrounding the study site, and the water table is typically at ≥25 m depth.
 To study the microweather conditions simultaneously inside and around the halite crusts, we used a HOBO® microweather station with a data logger equipped with two RH/T sensors (measurement range, 0–100%/−40°C–75°C and error, ±0.7°C/±0.7°C at 25°C), one PAR sensor for wavelengths of 400–700 nm (measurement range 0–2500 μmol m−2 s−1), and one leaf wetness sensor (measurement range 0 (totally dry) to 100% (totally wet)). The data logger is waterproof, has an operating range of −20°C–70°C, and has a typical use of 1 year. The station was emplaced on 16 June 2006. The data presented in this study correspond to 1 year of continuous measurements. We adapted the HOBO® Micro Station to measure the RH and T simultaneously outside and within a natural halite crust, colonized with endolithic microorganisms. To that end, a large (∼20 kg) halite crust was taken from the Yungay area and was transported to the facilities at the University of Antofagasta. Two holes were drilled on top of the crust with the same diameters as the PAR (3.2 cm) and the RH/T sensor (1.6 cm diameter). The PAR sensor was placed facing upward to obtain radiation readings at the level of the crust surface. The RH/T sensor was introduced ∼1 cm into the halite crust parallel to the rock surface to obtain readings close to the colonization zone. Once introduced, the sensor itself sealed the orifice. The crust was transported back to the halite field in the Yungay area and was placed in exactly the same spot where it was taken from. The leaf wetness sensor was initially placed within the rock, but after anomalous readings, probably due to salt, it was extracted and placed adjacent to the halite crusts. The external RH/T sensor was placed on the soil adjacent to the halite crusts containing the internal sensor and in the shadow. The temperature registered from this sensor was therefore a function of the air temperature and the radiation heat from the soil. The data logger, together with the batteries, was placed underneath the crust. All sensors were set to take a measurement every 10 min. Absolute humidity (AH) and Wd values inside and outside the rock were derived from the relative humidity and the temperature data. First, the saturation vapor pressure (Ps, in Pascals) for each data point was obtained following the Tetens formula [Buck, 1981]
where Tc is the temperature in degrees Celsius. Wd (in g m−3) is then obtained from
where AH = is the absolute humidity and Rw = 461.5 (J kg−1 K−1) is the gas constant for water vapor.
 Mean, maximum, and minimum values of the key microclimatic data outside and within the halite are shown in Table 1. Also listed in Table 1 are the mean seasonal data, the number of wet events, and the total hours of metabolic activity.
Table 1. Comparison of Key Environmental Parameters Measured Simultaneously Inside and Outside the Halite Crustsa
Mean, maximum, and minimum values of the key meteorological data and number of possible photosynthetic events outside and within the halite crusts.
RHh > 75% inside the crusts, and RHa > 90% outside the crusts.
 Daily temperature values outside and within the halite crust were similar during the study period. Air temperatures close to the soil surface (Ta) were relatively high during the day (maximum Ta = 51.79°C) and relatively low, often below the freezing point of water, during the night (minimum Ta = −3.37°C). Maximum Ta values were reached between 1200 and 1600 LT, and minimum values were reached shortly before sunrise, between 0600 and 0700 LT. The maximum Ta values were not extreme when compared to other desert environments such as Death Valley, where air temperature exceeds 50°C for many days in the summer. Halite temperatures (Th) were similar to the Ta values. The maximum registered temperature within the halite crust was 48.49°C, and the minimum was −3.85°C. Maximum and minimum Th values were reached approximately at the same time of the day as Ta values. The maximum difference recorded between simultaneous Ta and Th readings was 13.24°C (Ta = 18.28°C and Th = 31.52°C). Ta values showed larger daily oscillations than Th values, with higher maxima and lower minima. Mean daily temperatures were always between 0 and 10% higher inside the halite crust. Mean annual temperature values were also similar outside and within the halite crust (18.60°C and 19.5°C, respectively). Finally, the temperature data showed a clear seasonality, with spring and summer temperatures approximately 7°C higher than mean autumn and winter temperatures.
3.2. Relative Humidity
 The mean annual and seasonal air relative humidity (RHa) and halite relative humidity (RHh) showed comparable values, albeit slightly higher in the spring and summer months. On the other hand, daily RH values outside and inside the crust showed clear and occasionally significant differences. Figure 2 shows three episodes during which the RHh values remained relatively high and constant for 1–3 d, while the RHa values showed a normal daily variation, with maximum values at night and minimum values during the day. These events occurred when the RHh reached a value >75%, which corresponds to the DRH of halite [Cohen et al., 1987; Ebert et al., 2002], which we interpret as water condensation (wet halite events (WHE)) within the pore space of the halite crusts. Four recorded WHE had a relatively high intensity and lasted between 30–60 h, whereas most WHE had a lower intensity (between 1 and 6.5 h) and a higher frequency (53 a−1), for a total of 57 WHE throughout the year. The longest recorded period between two consecutive WHE was 50 d. On the other hand, RHa reached values above 85% on nine occasions (22 h) and above 90% on only two occasions (13.2 h). A relatively intense humidity event, likely due to heavy fog and/or dew, occurred on 20 August 2007 starting at 2300 LT and lasted 13 h.
 Taking the PAR data, we have estimated the number of hours throughout the year during which the minimum conditions for photosynthetic activity are met (presence of accessible liquid water in halite and PAR > 0). Assuming that water always condenses above the DRH of halite (>75%), this results in 213.8 h a−1 of potential photosynthetic activity for the halite endoliths (Table 1 and Figure 3).
3.3. Water Vapor Density
 Water vapor density is a measure of the amount of water vapor in the air and provides important clues regarding the origin of moisture and humidity in the area of study. The mean and minimum Wd inside and outside the rock were very similar, whereas the maximum Wd showed clear differences (Table 1). Figure 4 shows a typical daily variation of Wd both inside and outside the halite crusts. Two maxima occur, one after sunset and one after sunrise, and a daily minimum occurs soon after noon. Wd remains relatively constant during the night and early morning, when westerly winds dominate; a slight decrease in Wd is often observed before sunrise, a phenomenon likely due to dew formation in the soil and within the halite crusts. Right after sunrise, as the soil is heated, dew evaporates, which results in a transient increase of Wd. As the daily temperature continues to increase, the moisture in the soil and the halite crusts dissipates, dry easterly winds become dominant, and Wd reaches a minimum between 1600 and 1700 LT. The arrival of westerly, humid sea breeze in the late hours of the day increases Wd and closes the daily cycle. Our data show that halite efficiently retains water vapor. This is particularly important in the early hours after sunrise, when Wd inside the crusts is significantly higher than outside. This difference (ΔWd) between Wd inside and outside the rock occurs between 0700 and 1200 LT and can reach values as high as 35 g m−3 (Figure 4).
3.4. Photosynthetically Active Radiation
 The central Atacama, where Yungay is located, is free of clouds for most of the year. PAR profiles have a daily maximum between 1200 and 1300 LT, typically 2–3 h earlier than the maximum daily temperature. The maximum value recorded was 2.37 mmol m−2 s−1, and the yearly mean value was 1.10 mmol m−2 s−1. The daily onset of PAR oscillates seasonally between 0600 and 0730 LT, and the daily offset oscillates between 1800 and 2030 LT. The maximum values of PAR were recorded during the summer.
3.5. Leaf Wetness
 The wetness sensor successfully recorded data between 28 February 2007 and 16 June 2007. During the first month, dew occurred daily between 2000 and 0900 LT but dissipated immediately at sunrise. Dew events became sparse during May and June, with periods of several days (maximum 6 d) of null water condensation. Similar results have been reported by McKay et al.  in the same area and can be considered typical for this region. In all cases, dew did not result in measurable moisture, and the air RH remained well below 90% for practically all the measurement period (see section 4). However, the data obtained with the leaf wetness sensor should be analyzed with caution because of the physicochemical differences between the sensor and the halite crust surface.
 The results presented in this study show that while the occurrence of liquid water is extremely rare in the hyperarid core of the Atacama Desert, water is often absorbed and condenses within halite crusts, providing endolithic microenvironments potentially suitable for photosynthetic activity. As explained in section 2, some of the sensors were introduced in the crust to measure the humidity and temperature conditions within the colonized zone. We took care to ensure that the measurements within the crust were as close as possible to the conditions of unaltered halite. The mean temperature differences inside and outside the rock were approximately 1°C, with a maximum difference of 13.2°C. These values are consistent with previously reported studies comparing the habitability conditions inside and outside the endolithic niche in polar environments [Cockell et al., 2003], and we argue that our temperature sensors were indeed recording the actual temperatures within the endolithic zone. The wet events shown in Figure 2 indicate water condensation within the pore space of the halite at RHh levels close to 75%. These values are consistent with the DRH of halite and suggest that the recorded RHh values closely relate to the conditions in the pore space of the halite and are not an artifact of our instrumental setup.
 Cyanobacteria need liquid water or very high relative humidity levels to attain positive net photosynthesis [Lange et al., 1994, 2001]. Laboratory experiments of net carbon exchange with hypolithic cyanobacteria show a rapid decline when the relative humidity falls below 93% [Schlessinger et al., 2003]. Similar moisture sensitivity is reported for endolithic algae in the Negev Desert [Palmer and Friedmann, 1990]. The mean annual rainfall in Yungay is <2 mm a−1 [McKay et al., 2003], and the air relative humidity rarely reaches values above 90% [see also McKay et al., 2003; Warren-Rhodes et al., 2006]. Therefore, the survivability and abundance of photosynthetic microorganisms in that region are severely constrained by both the prolonged deprivation of liquid water [Warren-Rhodes et al., 2006] and the low levels of atmospheric RH.
 The main source of moisture and humidity in Yungay is sea breeze from the Pacific Ocean [McKay et al., 2003]. Mean wind directions in the area are predominantly from the west and carry humid sea breeze from the ocean, except in the hours around noon (between 1000 and 1300 LT), when dry easterly winds dominate [McKay et al., 2003]. The sea breeze reaching Yungay between sunset and sunrise frequently results in dew formation and occasionally fog, which are the only sources of liquid water for soil and hypolithic microorganisms. These transient wet events do not produce measurable moisture and quickly dissipate after sunrise when temperatures rise (Figure 4) [see also McKay et al., 2003], providing 395 ± 18 h a−1 of accessible liquid water or <75 ± 15 h a−1 of liquid water during conditions suitable for photosynthetic activity [Warren-Rhodes et al., 2006]. For that reason, soil and hypolithic niches are practically barren of photosynthetic microorganisms, which can only be found in small and scattered islands [Navarro-González et al., 2003; Warren-Rhodes et al., 2006].
 On the other hand, the presence of abundant endolithic cyanobacteria in the halite crusts, a hygroscopic mineral with DRH ∼75% [Cohen et al., 1987; Ebert et al., 2002], suggests that mineral deliquescence may facilitate primary productivity within the crusts. Halite deliquescence reduces the saturation water pressure within the pore space of the crusts, thereby facilitating water condensation at relatively low humidity levels. Throughout 1 year of microweather monitoring, we have recorded up to four intense episodes of water condensation within the crusts, which provided relatively constant moist conditions for periods of several days (longest recorded episode of 58 h), while the air humidity followed a normal daily trend that inhibited the occurrence of liquid water outside the crusts. The moist conditions within the crusts and the resulting NaCl-saturated solutions could be used by the endolithic microorganisms for primary productivity and growth. These long moist episodes were rare and mainly occurred in the wintertime. We have also measured more frequent (∼60 a−1), but shorter (1–8 h) deliquescence episodes, which are likely more relevant to the survivability and growth of halite endoliths. These short-duration events typically took place at sunrise and resulted in the simultaneous occurrence of liquid water and sunlight, thereby enabling photosynthetic activity within the crusts (Figure 5). If deliquescence occurs every time that RHh > 75%, we have estimated a total of 290 h a−1 of water condensation within the halite crusts. Assuming that the minimum requirements for the onset of photosynthetic activity are the presence of liquid water and light, this resulted in longer and more stable conditions for potential photosynthetic activity inside the halite crust than in the soil or under stones (213.8 h a−1 versus <75 ± 15 h a−1, respectively). Of all the liquid water that reaches the soil in the Yungay area, only 20% is available for photosynthetic activity because it occurs simultaneously with light [Warren-Rhodes et al., 2006]. On the other hand, ∼70% of the water that condenses within the halite crusts occurs simultaneously with photosynthetically available radiation and can be used by the organisms for their metabolic activity. Halite deliquescence is therefore an efficient mechanism for trapping air moisture and enabling photosynthetic activity in extreme hyperarid environments. These repetitive, short-duration episodes of possible photosynthetic activity explain the presence of halite endoliths in an otherwise life-depleted environment and strongly support the idea that halite crusts may be the last available niche for primary productivity in extreme hyperarid conditions.
 The time window for photosynthetic activity and growth inside the halite crusts is, however, very limited. The growth rate of Chroococcidiopsis is relatively low, with a typical generation time of about 16 d [Billi and Grilli Caiola, 1996], although some fast-growing strains have mean generation times of 3–4 d. For endoliths inhabiting halite, the conditions for growth are only met for ∼10 d a−1, during the transient deliquescence episodes, which also result in NaCl-saturated pore water solutions with a very low water activity (≤0.75). This argues that the primary producers within the halite are in a survival state, in which metabolic activity is possible, but very limited growth can be expected within the crust.
 Deliquescence also occurs on the surface of the halite crusts, resulting in the partial dissolution of the salt and its reprecipitation after the condensed water evaporates. The long-term effect of these surface processes on the morphology of the halite crusts, and the endolithic communities of cyanobacteria, has not been studied in detail. Dissolution and reprecipitation of salt will arguably alter the shape of the crusts and, together with wind erosion and sand abrasion, could explain the rounded shapes that characterize the crusts. This may in turn alter the position of the optimal zone for photosynthesis within the crusts, thereby forcing the endolithic communities to reposition accordingly. Given the extreme aridity in the region, these processes are likely to operate in timescales of tens to hundreds of years.
 The results presented here may be relevant to the search for life on Mars. During its transition from a relatively humid to an extremely dry planet, the surface of Mars likely once resembled the hyperarid regions of some parts of the Atacama Desert. Data provided by the Mars Exploration Rover Opportunity indicate that large regions of Mars such as Meridiani Planum once hosted evaporitic environments that resulted in the precipitation of salt-rich deposits [Squyres et al., 2004]. Halite has been identified in mineral assemblages of SNC Martian meteorites [Bridges and Grady, 2000; Treiman et al., 2000]. Minor quantities (up to several percent) of halite have also been found in rinds and certain rock coatings at Meridiani [Yen et al., 2006]. Although extensive deposits of halite similar to those in the Atacama Desert have not yet been discovered, they may be buried beneath accessible depths. The results presented here show that if deposits of halite exist on Mars, they may represent one of the last environments where life could have withstood the increasingly arid conditions. Considering the Atacama as an analog, these microhabitats could be localized, easily identifiable, and relatively abundant. Halite crusts on Mars may still host traces of endolithic microorganisms and may therefore represent important targets for future life search missions on the planet.
 A.F.D. wishes to thank the NASA Post-Doctoral Program and the Oak Ridge Associated Universities for financial support. This work was also supported by grants CGL2007-62875/BOS and CGL2006-04658 from the Spanish Ministry of Education and Science and grant PROIM 1337-1 from the Instituto del Desierto, Universidad de Antofagasta. We would also like to thank the Spaceward Bound NASA program for logistic aid during the June 2006 expedition to the Atacama Desert and two anonymous referees for their valuable comments and suggestions.