Predicting Soil Health from the Soil Microbiome. Cornell Soil Health Project

Agricultural soils are subjected to routine disturbances which affect microbial communities in a myriad of ways which we are only beginning to understand. We were interested in knowing whether a consistent set of microorganisms were favoured by agricultural management practices in a random sampling from across North America. We also tested whether physical, chemical and biological parameters of soil health correlated with changes in the microbiome. We used curated ecological databases to assess trends in populations of stress-tolerant or ruderal taxa and other functional guilds. We also used machine learning to test how well microbiome data, in this case bacterial 16S rRNA amplicon libraries, can predict dimensions of soil health.