MOLECULAR INVESTIGATION OF COMPOSITION AND FUNCTION OF THE SOIL MICROBIOME IN RESTORED FEDERAL WETLANDS FROM LANDSCAPE TO LAB

Abstract

Microorganisms have contributed directly to the development of contemporary conditions that sustain life on Planet Earth, through the modification of oceanic redox states, control of organic carbon flux in aquatic and terrestrial environments, and the cycling of key nutrients such as Nitrogen (N). Anthropogenic disturbances to terrestrial ecosystems have both direct and indirect impacts on the balance of biogeochemical cycling through changes in land-use, land management practices, and land degradation. Historic trends in global fertilizer use have indicated steady increases in both demand and application that has led to the pollution of waterways in a process known as eutrophication. Conservation research has shown that the restoration of global wetlands can serve to reduce the impacts of eutrophication by sequestering runoff nutrients such as N and providing unique conditions for biogeochemical cycling. Building predictive models of biogeochemical cycling in restored wetlands is a major focus for ecosystem ecologists and microbial ecologists alike and will serve to inform conservation land-use and best management practices. The microorganisms that perform biogeochemical cycling in soils offer direct and unique insight into nutrient process rates and the potential to mitigate eutrophication on a global-scale. The overall objective of my work is to understand how the soil microbiome influences N cycling across the landscape and in the lab. Previous work has suggested that microorganisms may be used as bioindicators of biogeochemical processes rates, therefore, I am interested in understanding how taxonomic composition and functional gene presence can be used to predict N cycling in restored wetlands. To address these objectives, I sampled soil bacterial assemblages in restored federal wetlands in Kentucky and Tennessee and utilized DNA sequencing to characterize the soil microbiome compositionally and functionally across the landscape, and the soil microbiome response to NH4-N in the lab. Results indicate that both functional gene copy numbers and microbiome composition can be used to predict N-flux rates on the landscape using machine learning and deep learning techniques. Additionally, I tested for an effect of NH4-N concentration on microbiome community assembly and found that carbon source, rather than N concentration, had the greatest impact on stabilized communities.