COS 235-6 - Modeling the distribution of microbial mats in the Fryxell basin of McMurdo Dry Valleys, Antarctica

Abstract: The McMurdo Dry Valleys are the largest ice-free area in Antarctica and represent one of the driest ecosystems in the world. In those ice-free areas, cyanobacteria-microbial mats are commonly found growing during the Austral summer when conditions are ideal. However, no studies to our knowledge have produced spatially explicit models linking environmental factors to the distribution of biologically active mats. Here, we conducted an ensemble species distribution modeling (SDM) analysis to understand where black microbial mats (subaerial microbial mat assemblages dominated by Nostoc) are distributed in the Fryxell basin region of the McMurdo Dry Valleys and what environmental factors drive their distribution in those areas. For the SDM, we modeled the distribution of black mats as a function of soil surface water, the presence of snow, elevation, slope, and aspect. The black mat presence data was derived from an abundance raster layer generated from a linear spectral unmixing model based on high-resolution multispectral satellite imagery. Soil surface water, presence of ice were also derived from the same satellite imagery. Elevation, slope and aspect were generated on those areas from data provided by Polar Geospatial Center (PGC). Our SDM models revealed that habitat suitable for black mats is scattered throughout the Fryxell basin region, identified mat hotspot areas along the southern and eastern edges of Canada Glacier, additional black mat habitat located in areas with local depressions where snow melt may provide water sufficient for mat growth north of lake Fryxell, and indicated the distribution of mat habitat in stream channels south of Lake Fryxell. The soil moisture variable was the most important predictor (importance score of ~60%) of the variation in the distribution of black mats followed by ice presence/absence (~20%) and aspect (~15%) variables. This is a first study attempting to understand microbial mats distribution in McMurdo dry valleys using the SDM method and is important for the understanding how the distribution of these sentinel taxa may respond in future climate scenarios.