In-situ Environmental Sampling & Analyses

At 30 locations throughout the Fryxell basin, we documented surface type present (biocrust, soil, or oxidized granite) and collected a surface layer sample of soil or biocrust (if present) and rock from all sites for subsequent hyperspectral analysis in the laboratory. At 12 of these sites, we established 5 m x 5 m intensive sampling plots. We collected 5 surface layer soil or biocrust samples (128 cm²) from each corner and center of the plots for subsequent pigment analysis and organic matter content via ash-free dry mass (AFDM). We also collected underlying soil down to 10 cm below the surface for gravimetric water content (GWC), electrical conductivity (EC), pH, inorganic nitrogen (N) concentration in the form of ammonium (NH₄⁺) and nitrate (NO₃^-), inorganic phosphorus (P) concentration in the form of phosphate (PO₄^3-), total organic carbon (TOC), total nitrogen (TN), and invertebrate abundance (nematodes, tardigrades, and rotifers).

Ash-free dry mass (AFDM): The surface layer soil and biocrust samples were measured for AFDM by weighing a known area of sample, combusting at 550 °C for 24 hr using a muffle furnace, gently stirring samples halfway through combustion, and reweighing after cooling in a desiccator. The rehydration of clays was assumed negligible, so we did not rewet samples.

Pigments: We estimated pigment concentration on the surface ~1 cm layer soil and biocrust samples using a trichromatic spectrophotometric method for chlorophyll-a, carotenoids, and scytonemin at 663, 490, and 384 nm, respectively (Garcia‐Pichel and Castenholz, 1991). Throughout the process, care was taken to avoid exposing the samples to light. The samples were dried at 105 °C for 24 hr, sieved through a 4 mm sieve, and extracted for 24 hr at ambient temperature in 90% unbuffered acetone using a 3.75:10 soil to solvent ratio, based on protocols from Couradeau et al. (2016) and the McMurdo Dry Valleys Long Term Ecological Research Program (MCM LTER) standard methods. After centrifugation, the extracts were analyzed on a spectrophotometer using 10 mL cuvettes. The absorbances contributed by each pigment were calculated using the trichromatic equations outlined in Garcia‐Pichel and Castenholz (1991), and the pigment concentrations were calculated using the Beer-Lambert Law with the extinction coefficients of 89.7 L g^-1 cm^-1 for chlorophyll-a (Couradeau et al., 2016), 112.6 L g^-1 cm^-1 for scytonemin (Brenowitz and Castenholz, 1997), and 262 L g-1 cm-1 for carotenoids (Thrane et al., 2015).

Invertebrate counts: The number of soil organisms (nematodes, rotifers and tardigrades), divided by species, sex and maturity was determined at each of the 12 plots. Invertebrate abundance was enumerated using inverted light microscopy on soil solutions using a modified sugar-centrifugation extraction procedure described by Freckman and Virginia (1993).

Gravimetric water content: Water content was determined as the mass of water per mass of dry soil. This was conducted by weighing field moist soil, drying in an oven at 105 °C for 24 hr, and reweighing the soil after.

Electrical conductivity: Using the underlying 1-10 cm soil, we measured electrical conductivity using a 1:5 soil to DI H₂O slurry.

pH: Using the underlying 1-10 cm soil, we measured pH using a 1:2 soil to DI H2O slurry.

Inorganic N and P: We extracted inorganic N (NH₄⁺ and NO₃^-) in 2 M potassium chloride and inorganic P (PO₄^3-) in 0.5 M sodium bicarbonate; inorganic N and P were measured on extracts using a Lachat flow injection analyzer. 

Total N and organic C: We measured TOC and TN using an Elementar Vario Cube TOC/TN analyzer after fumigating samples with concentrated hydrochloric acid to remove the influence of carbonates on TOC values.
 

Remote Sensing Analyses

Hyperspectral reflectance of field samples: Spectra were acquired of samples from each of the 30 plots using an Analytical Spectral Devices (ASD) FieldSpec4 high-resolution hyperspectral reflectance spectrometer set up for use in a stable lab environment. Data were collected between 400 and 2500 nm at a 1 nm sampling interval. A halogen lamp was used to illuminate the samples at 30° off-nadir, while reflectance was measured at nadir using the ASD’s bare fiber optic cable roughly 3 cm above the sample surface. To minimize instrument noise, particularly at the longest and shortest wavelengths where the output of the halogen bulb is lowest, we averaged 50 individual spectra for each sample.

Hyperspectral reflectance of laboratory mixtures: Sample mixtures of varying percent biocrust and soil (not containing biocrust) were created by mixing a known amount of soil with a known amount of biocrust. The soil was collected from the Fryxell basin, Antarctica, and the biocrust was collected from Arizona, USA, and each were separately ground into a homogenous fine particle size. Hyperspectral reflectance data were gathered for dry samples with the same methods above. Samples were also moistened with DI water for damp readings as well. The targeted water content for the damp samples was approximately 0.25 GWC (g H₂O/g dry soil and biocrust material). These data were used to determine the nature of the observed spectral mixing relationship between soil and biocrust, validating our ability to confidently translate spectral unmixing to orbital data.