Land management practices such as regular burning of aboveground biomass as well as grazing by large ungulates impact ecosystem carbon (C) and nitrogen (N) cycles, which in turn affect soil microbial community structure and function. Understanding the relationship between historical land management and microbial diversity is central to predicting how management impacts microbially-mediated ecosystem services including decomposition and N mineralization. We specifically asked whether microbial distribution and heterogeneity differed among Konza Prairie surface soil samples (2.5 cm-wide cores, 0-20 cm depth) from watersheds that are burned either annually (1yB) or every 20 years (20yB), and are bison-grazed (G) or ungrazed (UG) (two-way factorial design); samples were collected along replicate transects in a log-distance design, allowing explicit evaluation of spatial heterogeneity from 10 cm to 1 km scales. Bacterial community composition (BCC) was measured using Illumina MiSeq sequencing of 16S rRNA genes; soil C, inorganic N, water content and pH was also measured. We predicted that soils from watersheds with historically higher N availability (G and 20yB) would have lower overall microbial diversity, due to greater competitive dominance of nitrophilic taxa, and that microbial heterogeneity in grazed soils would vary at different scales due to animal redistribution of nutrients and microorganisms.
Bacterial 16S rRNA gene sequencing of soils from differently managed watersheds revealed that both fire and grazing history had significant impacts on BCC within the watersheds (permutational ANOVA: Burn effect P=0.01, R2=0.048; Grazing effect P<0.0001, R2=0.108). Bacterial alpha richness in operational taxonomic units (OTUs) was highest in the annually burned, ungrazed watersheds (3741±61 > 3413±98, 3492±67, 3447±66 (1yB/UG > 20yB/UG, 1yB/G, 20yB/G, mean±SE); 2-way ANOVA: interactive effect P=0.056), but bacterial evenness was similar among all watersheds (2-way ANOVA: P>0.1). The UG watershed communities had more unique (unshared) OTUs (9.3%-10.0%) than the G watersheds (4.5%-5.3%). Thus, our predictions were supported in that the historical management treatment with lowest N availability had highest bacterial richness; however, there was no evidence for nitrophilic dominance. There was lower heterogeneity in grazed soils, in that there were fewer unique OTUs at the watershed scale. Further analysis will reveal at what geographic scale heterogeneity is highest in each treatment, and how diversity covaries with the soil environment. The clear differences in BCC between grazed and ungrazed soils may be related to soil texture and water content (10 cm scale variation), or plant community differences (meters scale variation) rather than generalized watershed-scale N availability.