Quantifying Variability and Correlation in Biomarker and Mineralogical Measurements: Lessons from Five Astrobiological Mars Analogue Expeditions in Iceland

Studies in planetary analogue sites correlating remote imagery, mineralogy, and biomarker assay results help predict biomarker distribution and preservation. The FELDSPAR team has conducted five expeditions (2012-2017) to Icelandic Mars analogue sites with an increasingly refined battery of physicochemical measurements and biomarker assays. Two additional expeditions are planned; here we report intermediate results.The biomarker assays performed represent a diversity of potential biomarker types: ATP, cell counts, qPCR with domain-level primers, and DNA content. Mineralogical, chemical, and physical measurements and observations include temperature, pH, moisture content, and Raman, near-IR reflectance, and X-ray fluorescence spectra. Sites are geologically recent basaltic lava flows (Fimmvörðuháls, Eldfell, Holuhraun) and barren basaltic sand plains (Mælifellssandur, Dyngjusandur). All samples were 'homogeneous' at the 1 m to 1 km scale in apparent color, morphology, and grain size.[1]Sample locations were arranged in hierarchically nested grids at 10 cm, 1 m, 10 m, 100 m, and >1 km scales. Several measures of spatial distribution and variability were derived: unbiased sample variance, F- and pairwise t-tests with Bonferroni correction, and the non-parametric H- and u-tests. All assay results, including preliminary mineralogical information in the form of notable spectral bands, were then tested for correlation using the non-parametric Spearman's rank test.[2] For Fimmvörðuháls, four years of data were also examined for temporal trends.Biomarker quantification (other than cell count) was generally well correlated, although all assays showed notable variability even at the smallest examined spatial scale. Pairwise comparisons proved to be the most intuitive measure of variability; non-parametric characterization indicated trends at the >100 m scale, but required more replicates than were feasible at smaller scales. Future work will integrate additional mineralogical measurements and more specialized modeling. [1] Amador, E. S. et al. (2015) Planet. Space Sci., 106 1-10. [2] Gentry, D. M. et al. (2017) Astrobio., in press.