PS 20-11 - Stochastic assembly of soil bacterial communities moderates the effects of biologically-mediated plant disease control in agroecosystems

Abstract: Stochastic community assembly processes, such as dispersal and drift, are often invoked to explain spatial heterogeneity in the diversity of biological communities across unmanaged systems, but their relevance to applied agroecological contexts remains unclear. Soil microbiome composition can predetermine plant disease and underlie the efficacy of biological disease control methods. Thus, agricultural soils represent an underutilized, economically important system with which to evaluate the role of stochastic processes in microbial community assembly.



We examined whether anaerobic soil disinfestation (ASD), a biological crop disease control method used in specialty crop production, changes soil bacterial communities in ways that are contingent on preexisting, random spatial heterogeneity in initial communities. We hypothesize that (1) bacterial community composition displays a distance-decay pattern, (2) ASD increases dispersion of bacterial communities, and (3) the effect of ASD on bacterial alpha diversity is site-specific.



Soil samples (n = 45) were collected between 2011-2013 across 5 field trials in Florida spanning 220 km. These fields were managed with chemical fumigation prior to ASD trials. ASD was conducted by incorporating a mixture of molasses and chicken manure as a carbon source and covering the fields with plastic tarp to induce anaerobic conditions. Soil samples were collected before ASD and 18-21 days after the start of ASD. Bacterial communities were characterized by 16S rRNA sequencing and clustered into operational taxonomic units (OTUs) at 97% similarity, resulting in 2,312 OTUs.



We found a strong distance-decay relationship for bacterial community composition (operationalized as Bray-Curtis dissimilarity) in untreated soils (Mantel r = 0.71, P < 0.001). The dispersion of bacterial community composition increases under ASD treatment across sites (homogeneity of group dispersions test, P < 0.001) but does not consistently change within sites, suggesting that preexisting between-site compositional differences are amplified by ASD. Models with site-specific effects of ASD on bacterial alpha diversity outperform models assuming uniform effects of ASD across sites (likelihood ratio test, Prichness < 0.001; PShannon < 0.001). Our results suggest that stochastic processes are important in the assembly of bacterial communities in agricultural soils where the strong selective pressure of pesticide application has been replaced by the weaker selective pressures of biological disease control methods such as anaerobic soil disinfestation (ASD). Furthermore, we find evidence that preexisting variation in bacterial communities can lead to site-specific changes under ASD. Future research will investigate the relationship between these site-specific compositional shifts and the spatial heterogeneity of plant disease control outcomes.