Investigating Detection Limits of Antarctic Microbial Communities Using Remote Sensing Data

Autotrophic microbial communities drive carbon cycling in the McMurdo Dry Valleys (MDV), Antarctica. Dense microbial mats, consisting mainly of cyanobacteria and moss, occupy aquatic areas. Microbial communities also occur at much lower densities as surface soil biocrusts throughout the terrestrial landscape. Given the broad spatial extent of this terrestrial landscape, we anticipate a significant amount of carbon on the soils as a result of biocrust presence. Our previous research has shown that multispectral satellite data can be used to detect and estimate microbial mat biomass in high-density areas like streams. However, while remote sensing efforts originally derived for high-productivity areas were broadly applied to the entirety of the terrestrial landscape, the accuracy of this application was not fully explored. We have found that spectral parameters like the Normalized Difference Vegetation Index (NDVI) applied to WorldView-2 (WV-2) satellite imagery can be challenged by false detections from abiotic sources. While we have had success using NDVI in the past with dense microbial mats, now at such low abundances it is unclear how accurate NDVI is at detecting low-density biocrusts. Therefore, additional spectral techniques using both multispectral and hyperspectral data are required to quantify the detectability of biocrust in this region. Using the eight WV-2 multispectral reflectance bands and some traditional vegetation indices in a principal components analysis, there is separability between biotic and abiotic surfaces with a detection threshold between 5 and 28 mg/cm2 ash-free dry mass. By applying spectral linear unmixing models to downsampled hyperspectral measurements of biocrusts in the lab, we can detect biocrust abundances down to 1% (g/g). These methods allow detection of low-density biocrusts that also support diverse soil food webs (i.e., nematodes, tardigrades, rotifers, and ciliates). If biocrusts do supply a significant amount of carbon on MDV soils cumulatively, our work will be important for monitoring terrestrial carbon stocks and predicting future ecosystem dynamics over space and time.