Pore size distribution and water retention in colonized Antarctic Beacon sandstone

Abstract

We report on the pore size distribution and water retention of Beacon sandstone from Antarctica that harbours a cryptoendolithic community, predominantly lichens, just below the surface. We measured the pore size distribution, drying time and equilibrium relative humidity of sandstone samples that were colonized by lichens. The incremental pore volume distribution peaks at ~10 μm radius, but ~20% of the pore volume occurs for a radius < 0.02 μm. Water from snowmelt fills ~20% of the total pore volume. It takes ~4–5 days of evaporation to lose 90% of the water. As the rock loses water, the equilibrium relative humidity remains at 99% until an appreciable amount (80–90%) of the pore water is lost, after which the equilibrium relative humidity starts to decrease. The relative humidity remains at > 80% (68 h) long after the water content falls to < 98% (19 h) - the point at which liquid water is estimated no longer to be present. Lichens can remain active in air with high relative humidity (> 80%). Thus, the pore size distribution of the sandstone may play a role in explaining why lichens dominate these sandstones.

Introduction

The McMurdo Dry Valleys (MDVs) are the largest ice-free region in Antarctica. At high elevation (typically > 1000–1500 m) air temperatures are always below freezing and there is very little precipitation (Friedmann et al. 1987, McKay 2015, Marinova et al. 2022). Yet remarkably, just below the surface (~1 cm) of exposed Beacon sandstone, microbial communities of photosynthetic microorganisms thrive (Friedmann & Ocampo 1976, Friedmann 1982). In the summer, owing to solar insolation, the rocks are significantly warmer than the ambient air by up to 15°C (Kappen et al. 1981, McKay & Friedmann 1985, Friedmann et al. 1987), warm enough to melt snow on the surface of the rocks (Friedmann 1978, 1982, Kappen et al. 1981, Sun 2013). This liquid water is absorbed by the porous stone. The organisms grow no deeper than 1 cm, the depth of light penetration (Nienow et al. 1988, McKay 2012).

The effects of temperature and light on shaping the cryptoendolithic habitat have been extensively studied. The effect of water, however, has been much less studied. Whereas it is clear that snowmelt serves as the source of water for these lithobionts (Friedmann 1978, Sun 2013), the mechanisms that control the retention of water within the rock are poorly understood. In particular, the pore size distribution of the rock and its role in water retention is not yet quantified. Work by Wierzchos et al. (2012) suggests that this could be important. They showed that endoliths within halite rocks in the Atacama Desert obtain liquid water through spontaneous capillary condensation at relative humidity (RH) much lower than the deliquescence RH of NaCl due to small pores in the halite. They also showed that as the halite warmed and dried these small pores retained brine for longer than larger pores.

The relevant field studies in Antarctica have focused on the duration of moist conditions after snowmelt has wetted the rock. Kappen et al. (1981) placed the active element of RH sensors into deep holes (15 mm) in the rock and then sealed the holes with silicone. The holes had to be large to accommodate the sensor. RH inside the rock was high and remained at 80% or more for 5 days following snow (Kappen et al. 1981). Friedmann et al. (1987) and McKay et al. (1993) monitored moisture using the conductivity between two wires spaced ~1 cm apart in much smaller, shallower holes drilled into the rock. The change in conductivity of the rock showed a clear distinction between times when water channels provide or do not provide a continuous path of conductivity. High-conductivity periods were assumed to indicate a water activity suitable for growth (Friedmann et al. 1993, McKay et al. 1993). According to Friedmann et al. (1987), rock conductivity continues at high levels for more than 1 week. This was verified by direct field measurements during which rock samples still contained 0.45–1.15% water by weight ~5 days after a snow event (Kappen et al. 1981).

An interesting but not fully explained aspect of the cryptoendolithic colonization in the Antarctic Dry Valleys is that the rocks are dominated by lichens. It is known that lichens can utilize water at lower water activity. According to Lange (1969), chlorolichens begin respiration at a RH of 70% and require a RH of 80% for net photosynthesis. The values reported by Palmer & Friedmann (1990) for the cryptoendolithic lichens in the Dry Valleys were similar, pointing to the possibility that the growth of lichens is also associated with a water activity of < 0.98, as has been substantiated in the Negev Desert (Kidron et al. 2014, Kidron & Kronenfeld 2022) and for alpine endolithic lichens (Weber et al. 2007). In this context, and in line with numerous publications regarding a possible link between substrate texture and water retention (Su et al. 2015, Song et al. 2021) and the directly relevant previous work of Wierzchos et al. (2012), our goal is to investigate the pore size distribution and water retention of a lichen-dominated sandstone and to study the possible implications of water availability for lichen colonization.

Methods

Two samples, shown in Figs 1 & 2, of colonized Beacon sandstone were collected from Battleship Promontory (76°55.3′ S, 161°05′ E, elevation 1294 m), within the Dry Valleys of Antarctica (Friedmann 1982). Due to a rich collection of endolithic microorganisms at this site, it has received considerable attention and has been the site of extended meteorological observations (McKay et al. 1993, Friedmann et al. 1994). Rock 1 (Fig. 1) has an air-dry mass of 586.1 g and is a loose rock collected from the ground. Rock 2 has an air-dry mass of 530.4 g and was chipped from a large sandstone outcrop. The colonized layer in Rock 2 is evident in a close-up of the sheared face, as can be seen in Fig. 2.

Fig. 1. Rock 1, a sample of Beacon sandstone from Battleship Promontory colonized by cryptoendolithic lichens. The background grid is 0.5 cm squares.

Fig. 2. a. Rock 2, a sample of Beacon sandstone from Battleship Promontory colonized by cryptoendolithic lichens. The background grid is 0.5 cm squares. b. A close-up image showing the black colonized zone.

The pore volume and area distribution of a small, representative piece of Beacon sandstone were measured by mercury intrusion porosimetry (Washburn 1921, Diamond 1970, ASTM 2018) performed by Micromeritics (Norcross, GA, USA). The surface tension of mercury is 15 times greater than that of water. The Micromeritics report (Supplemental Material) lists the parameters of the intrusion test. Briefly, mercury was forced into an 8.8 g sample over a pressure range that sampled pore size for diameters from 121 to 0.003 μm.

As a function of pore radius, the water activity, a_w (which equals the RH in the vapour phase) was determined with the Kelvin equation assuming a concave meniscus of water with a spherical shape and radius equal to the pore radius r (e.g. Camuffo 2014, ch. 5, p. 167):

(1)$${\rm ln}( {a_w} ) = 2\sigma Vm/( {rRT} ) $$

which can be conveniently expressed in terms of water potential per unit volume of water, Ψ  =  (RT/Vm)ln(a_w), as:

(2)$$r = 2\sigma /{\it \Psi} = 151 {\rm kPa}/{\it \Psi} $$

where σ is the surface tension of water (75.6 × 10^–3 N/m at 0°C), Vm is the molar volume of the water (i.e. Vm  =  18 cm³ =  18 × 10^–6 m³ for pure water), R is the gas constant 8.314 J mol^-1 K^-1, T is the absolute temperature and the water potential is in units of Pascals (Pa).

The infiltration pressure required to force a fluid into a pore of radius r is computed by the Young-Laplace equation from the theory of capillarity (Washburn 1921):

(3)$$P = 2\sigma cos\theta /r\;$$

where σ and θ are the surface tension and contact angle of the fluid, respectively, and r is the pore radius (Camuffo 2014, ch. 5, p. 172). Sandstone rocks are generally wettable with contact angles of ~20° at ambient pressure (Espinoza & Santamarina 2010, Kaveh et al. 2014, Zhang et al. 2020). Thus, we assume that the fraction of water that is held in the rock is held in the same pore size distribution as indicated by Hg intrusion. We assume this even though the pore volume filled with water is less than the pore volume indicated by Hg intrusion - presumably because Hg is forced into the pores at high pressure, which is not the case for the water. It is worth noting that Hg intrusion porosimetry assumes that all pores have a cylindrical shape. In addition, the Kelvin equation assumes a spherical pore. In the real sandstone, there are likely to be irregular shapes to the pores, and thus the pore size distribution represents a statistical averaging over these shape effects in terms of equivalent spheres. This is likely to be a reasonable approach if there is a large number of pores. There may also be pores that are closed off that contribute to the bulk porosity, but they are not of interest here because they do not affect the pore size distribution determined by Hg intrusion or the water content of the rock.

To determine the relation between RH and water contained in a rock, the rock sample and sensors were placed in an airtight 2.6 l chamber (Fig. 3) until equilibrium was reached, typically within a few days. The rock was then removed and its mass was determined.

Fig. 3. Photograph of the experimental setup to determine equilibrium relative humidity. The rock sample is in the chamber with two temperature/humidity sensors.

Rock mass was determined with a digital scale reading to 0.01 g. Tests with standard weights and the rocks indicated a relative precision of ±0.1 g and absolute accuracy of ±0.5 g between measurements taken over several days. For the rocks used here, with a mass of ~500 g and a water content of ~10 g, this implies a relative precision of 0.02% in rock mass and 1% in the mass of water retained during the drying experiments.

RH and temperature were measured with a self-contained ThermoPro TP49 Digital Hygrometer. Two or more units were inside the closed chamber during experiments (see Fig. 3). From this comparison we determined that the temperature accuracy was ±1°C (5–25°C) and humidity accuracy was ±3% over the range of primary interest (40–99%). The manufacturer's specifications are ±1°C and ±2–3% RH.

Oven dry mass was determined by heating the samples to 90°C for 24 h. Wetted water mass was determined after the sample was watered briefly and after it was submerged in water for 15 min and the container slightly vibrated by hand.

Results

As can be seen in Fig. 4, the incremental pore volume in the Beacon sandstone peaks at ~10 μm radius. However, as illustrated in the plot of normalized cumulative pore volume, pores with a radius of < 0.02 μm account for ~20% of the total pore volume. This is also clearly shown in the normalized cumulative area, which increases sharply for < 0.02 μm radius. The full analysis from Micromeritics is included as Supplemental Material, and the long tail of small pores is clearly shown in the figure on page 10 of this report on log differential intrusion vs diameter. A simple analogue for this size distribution is to compare the volume and area of one pore with a pore radius of 10 μm with 1 million pores of radius 0.1 μm. The two distributions have the same volume, but the small pores have 100 times more area. Also shown in Fig. 4 is the water activity, a_w (equivalent to RH), computed from the Kelvin equation (Equation 1). Over 83% of the pore volume corresponds to a_w > 0.98, and 10% of the pore volume corresponds to 0.80  <  a_w > 0.98. In other words, larger pores, representing 83% of total pore volume in the rocks, hold liquid water only at RH values > 98%, and they would quickly lose their water content to evaporation when RH falls below that level. In contrast, smaller pores, representing 10% of total pore volume in the rocks, still hold liquid water at RH values between 80% and 98%.

Fig. 4. The incremental pore volume as a function of pore radius (red line and red axis). The left vertical axis is for water activity (a_w) as a function of pore radius (dotted line). The normalized cumulative pore volume is the solid black line. The normalized cumulative pore area is the solid yellow line. Over 83% of the pore volume corresponds to a_w > 0.98.

Table I is a summary of the total porosity of the two sandstone rocks based on Hg intrusion and water uptake. The differences in pore volume per gram of rock for the two fluids can be understood from the Young-Laplace equation (Equation 3). The minimum pore size into which a fluid will flow will depend on the intrusion pressure and the surface tension. Although the surface tension of Hg is 15 times greater than that of water, the intrusion pressure is very high (Supplemental Material).

Table I. Summary of the properties of Beacon sandstone rocks.

In the wetted rocks, water only enters a fraction of the pore volume: 22% (0.0260/0.1162) for Rock 1 and 19% (0.0218/0.1162) for Rock 2. The results were the same for a brief wetting or for 15 min submergence.

Figure 5 shows the fraction of water retained in a rock as a function of the equilibrium RH as the wetted rock air-dries. As the rock loses water, the equilibrium RH remains at 99% until an appreciable amount of the pore water is lost (80–90%). The equilibrium RH then decreases gradually as the fraction of water retained becomes smaller. Thus, 80–90% of the pore water is contained in the large pores (where a_w > 0.98) and only a small fraction of the pore water is contained in the smaller pores (where 0.80  <  a_w > 0.98).

Fig. 5. The fraction of water retained in a rock as a function of the equilibrium relative humidity as the rock dries from wetted conditions to air dry. The square symbols and solid black line are for Rock 1; the circle symbols and dashed black line are for Rock 2. Red curve is calculated from the pore size distribution in Fig. 4.

Figure 6 shows the results of the drying experiments for both rocks under warm (20°C) and cold (5°C) conditions, the latter of which, for obvious reasons, being more relevant to Antarctic conditions. It seems that under this range of temperatures there is little temperature dependence. The rocks dry rapidly in ~24 h and then more slowly for the next 48 h and even slower thereafter. The blue line in Fig. 6 is the equilibrium RH of the rock as it dries. This is computed using the results given in Fig. 5. The equilibrium RH is > 98% for the first 19 h and > 80% for the first 68 h.

Fig. 6. The fraction of water retained as the rock dries, shown as black lines. The square symbols and solid black lines are for Rock 1; the circle symbols and dashed black lines are for Rock 2. Unfilled symbols are for drying at 20°C; filled symbol are for drying at 5°C. The blue solid line is the equilibrium relative humidity of Rock 1 during the drying process computed from the results in Fig. 5. The arrows indicate the times at which the relative humidity falls below 98 (19 h) and below 80% (68 h).

Discussion

The experimental results obtained in this study on drying times and equilibrium RH are generally consistent with prior works based on field measurements of RH and conductivity in lichen-colonized Beacon sandstone (Kappen et al. 1981, Friedmann et al. 1987, 1993, McKay et al. 1993). A precise comparison is not possible between the drying time of our small rocks (in a windless indoor environment) and the drying time of the bedrock surfaces in the outdoors. Nonetheless, our results indicate a several-day-long drying time after a rock is wetted by snowmelt, consistent with the measurements of Kappen et al. (1981) and Friedmann et al. (1987). Kappen et al. (1981), for instance, found a range of water contents between 0.45% and 1.15% under field conditions. We found a water content of ~2% under wetted conditions, with no difference between a brief watering and submersion in water for 15 min. Rock 2 has a slightly lower water content than Rock 1 (Table I), probably because it has a higher proportion of cemented crust. A surface crust may inhibit water entry into the rock, but for the rocks studied here the crust covers only a fraction of the surface. Rocks completely covered with silicified crusts may not take up snowmelt at all (Friedmann & Weed 1987).

The typical values for the fraction of water retained over the drying time range from 0.1 to 0.6, giving water contents of 0.2–1.2%. The measurements of pore size distribution by Hg intrusion provide a microphysical basis for understanding the results primarily through the control of RH by the size of the water-holding pores.

The results of the pore analysis indicate that 83% of the water contained in the pore volume corresponds to a_w > 0.98, implying significant liquid water retention by the rocks. The results also indicate that ~10% of the water contained in the pore volume corresponds to 0.80 <  a_w < 0.98, which is within the range usable by lichens. Using these values, we find that the amount of time that the Beacon sandstone rocks would be suitable for lichen growth after snowmelt is three times greater than the amount of time that bulk water may be present within the rock pores. Whilst under the conditions used in the current experiment growth conditions may last for 3–4 days, under the low temperatures that characterize the sandstones during most of the day in Antarctica these conditions may last substantially longer, as reported by Kappen et al. (1981). The efficient water retention of the rocks can be attributed to the population of the small pores within the Beacon sandstone that hold ~10% of the pore water at RH values > 80% but < 98%. We assume that the small pores are uniformly distributed in the rock. The lichens may not necessarily live inside these small pores, but the water vapour from these pores would be available. This is entirely consistent with the work of Wierzchos et al. (2012), which reported a similar effect in the small pores of halite rocks in the Atacama Desert.

Conclusion

We have measured the pore size distribution, drying time and equilibrium RH as a function of the water content of samples from the Antarctic Beacon sandstone colonized with lichens. Based on these results, we draw the following conclusions. Water from snowmelt infiltrates but does not completely fill the pore space of the Beacon sandstone; meltwater fills ~20% of the total pore volume. As the rock dries, the equilibrium RH decreases sharply once the water content falls below 83% of its initial content; however, it remains in the range that lichens can utilize for a longer period - effectively tripling the time during which lichens can grow. We find that the pore size distribution in the rock is consistent with and can explain these results. Thus, the pore size distribution of the sandstone may play a key role in explaining why the cryptoendolithic community in the Antarctic Beacon sandstone is dominated by lichens.