With the application of molecular biology techniques to investigate MDV biodiversity, a surprising level of bacterial (Cary et al. 2010, Takacs-Vesbach et al. 2010; Van Horn et al. 2013) and protist (Bielewicz et al. 2011; Xu et al. in review) richness has been revealed Although bacteria comprise a significant proportion of MDV biomass (Takacs and Priscu 1999; Foreman et al. 2007; Stanish et al. 2012), we know relatively little about their physiology and thus about their ecology and roles in biogeochemical cycling. To truly understand connectivity within the MDV (and hence, to address our hypotheses in MCM4) requires an understanding of the biodiversity, distribution, and functional roles of specific organisms within the environment and their responses to climate driven pulses and presses. Our approach is to determine spatial and temporal variations of microbial diversity, distribution, and function across all major habitats (cryoconites, streams, lakes, and soils) and determine changes in response to experimental manipulations (see sections on LakeICE and P3). Previous microbial 16S and 18S rRNA gene assays in the MDV have produced a large diversity of sequences from specific terrestrial and aquatic habitats (e.g., Gordon et al. 2000; Glatz et al. 2006; Barrett et al. 2006; Porazinska et al. 2009; Vick-Majors et al. 2013, Michaud et al. 2012). In MCM4, we are using our molecular data to investigate the role that microbial distribution and function play in ecosystem level process MDV-wide.

In Y1 of MCM4, we collected 425 samples from cryoconites, streams, lakes and soils throughout the Taylor, Wright, and Miers Valleys and analyzed 16S rRNA gene diversity to determine a baseline distribution of microbial communities throughout the McMurdo Dry Valleys. The data showed that bacterial alpha, beta, and gamma diversity is greatest in streams, followed by lakes and soils, and that phylogenetic diversity does not differ significantly between lakes and soils. Significant habitat filtering is observed among the samples (Fig. 1, Anosim significance of weighted and unweighted Unifrac clusters P