High latitude oceans are vitally linked to ecosystems worldwide. Circumpolar deep water is a source of carbon for low-latitude oceans, and seasonal peaks in polar ocean primary productivity helps drive the ‘biological pump.’ Yet, the Amundsen Sea Polynya, which has some of the highest productivity in coastal Antarctica, and the surrounding sea ice, is one of the least studied regions in the world.
Here we ask, are Amundsen Sea bacterial communities homogeneous throughout the sea ice and the water column? We hypothesize there is stratification by depth, and dominant functional groups vary with nutrient composition, levels of primary production, and by association with particulate organic matter (particle association).
DNA was extracted from free-living (< 3 µm in size) and particle-associated (>3 µm) bacteria from four water-column stations and sixteen ice stations in and around the Amundsen and Ross Sea Polynyas. Hypervariable region (V6 and V3/V4) pyrosequencing was then used to get genetic profiles for each sample. Multidimensional scaling, analysis of molecular variance, cluster analysis, and similarity percentage analysis were used to test whether community composition varied with depth, location, type (ice, seawater, or brine), and particle association.
Results/Conclusions
There is distinct clustering of shallow water (≤100 m depth) bacteria by particle association. Interestingly, deep water (> 100 m) free-living bacteria cluster more closely to shallow types than to the deep water particle-associated group. Most (61-83%) of the variance in the water column can be explained by two variables, depth and particle association. However, 25% of the variation observed between communities is attributed to variation in the abundance of less than 2% of the operational taxonomic units (mostly gammaproteobacteria, alphaproteobacteria and flavobacteriales). Furthermore, we expect bacterial communities in the sea ice and brine will have different functional composition than those in the water column; but we anticipate these communities will also cluster by particle association and will vary across spatial scales.