Short-term direct and indirect effects of herbicide application on an agricultural soil microbiome

Chemical herbicides, including glyphosate, atrazine, and many others, are commonly used to control unwanted vegetation in agroecosystems and protect crop yield potential. While the effects of herbicides on aboveground weedy vegetation are well documented, by comparison, the influence of herbicides on belowground processes is poorly understood. Given the importance of the soil microbiome for providing ecosystem services, understanding microbial responses to herbicide application is of pressing interest. In this study, the short-term non-target effects of herbicide application were examined by comparing three commonly used chemical herbicides, glyphosate, dicamba, and a tank mixture of atrazine + mesotrione, to handweeded and non-treated controls. Changes in soil microbial function, nutrient pools, and microbiome composition in Zea mays were assayed over 20 days post-treatment. Direct effects of the examined herbicides on soil edaphic properties, microbial alpha-diversity, and enzyme activities were small and transient over this 20-day, short-term experiment. Changes in enzymatic activities, nutrient concentrations, and fungal community structure were predicted primarily by weedy vegetation cover, suggesting that the observed responses are indirectly mediated by herbicide application. Conversely, weed removal treatment was a significant predictor of bacterial community composition, and chemical herbicide application was associated with homogenization of the bacterial community over the short sampling period, a trend not observed in the non-treated and handweeded controls. Together, these results indicate that in heavily managed agroecosystems, the short-term direct effects of chemical herbicide application on microbial diversity and enzymatic function are modest, whereas the effects on bacterial community assembly, composition, and heterogeneity are more readily detectable. The observed trend of biotic homogenization following pesticide application, a pattern not observed in the handweeded and non-treated controls, warrants further examination over longer time scales and repeated herbicide applications.