Controls over the distribution of soil micro-invertebrates in the McMurdo Dry Valleys, Antarctica

Abstract

The McMurdo Dry Valleys of Antarctica are a cold desert ecosystem where soil micro-invertebrate communities are shaped by extreme abiotic conditions, particularly low soil moisture, organic carbon availability, and high salinity. In this thesis I combine long-term ecological data with spatial habitat modeling to investigate the environmental drivers of micro-invertebrate distributions and predict future distributions under variations in moisture availability associated with shifting climate regimes. Chapter 1 documents the creation and cleaning of long-term datasets compiled from the McMurdo Dry Valleys Long-Term Ecological Research (LTER) program, encompassing ~30 years of physical and chemical soil data linked to micro-invertebrate abundance. These datasets serve as the foundation for Chapter 2, which applies statistical and machine learning models, including logistic regression and Random Forest combined with remote sensing and GIS, to assess habitat suitability for key taxa. Results indicate that Plectus spp., Eudorylaimus spp., rotifers, and tardigrades are positively associated with higher soil moisture and organic carbon, while Scottnema lindsayae prefers dry, nutrient-poor soils consistent with previous literature. Projected increases in soil moisture favor habitat expansion for moisture-dependent taxa and reduce suitable habitat for taxa less tolerant of moisture such as Scottnema lindsayae. My findings suggest that climate-driven changes in soil water availability could shift species distributions and community composition in Antarctic soils. This work provides a framework for future distribution modeling in polar ecosystems and highlights the importance of long-term ecological data in predicting biodiversity responses to environmental change.