Long-term Effects of Shrub Encroachment and Grazing on Soil Microbial Composition and Function

Abstract:

Drylands account for ca. 35% of terrestrial net primary productivity and thus play a significant role in global water and biogeochemical cycles. Replacement of grasses by shrubs has been widespread in these systems and has altered rates of erosion and native plant biodiversity and productivity. The net effect of these changes on biogeochemical cycling is not well understood. Projected warmer and drier conditions may further alter the function and stability of these ecosystems and soil resources through direct effects on soil microbiota and plant-microbe interactions. We quantified microbial community responses to long-term livestock grazing and shrub encroachment in a Sonoran Desert grassland. We sought to characterize tipping points where biotic controls over ecosystem processes shift from being ‘grass-driven’ to ‘shrub-driven.’ We asked: How do livestock grazing (the predominant land use in dryland ecosystems) and shrub invasion (a predominant land cover change) interact to influence microbial biomass and the relative abundance of bacteria, archaea, and fungi and their extracellular enzyme activities? Surface soil from bare-ground patches, native and invasive grass rhizospheres, and bole and canopy dripline locations in patches of mature mesquite trees in long-term grazed and long-term (70+ y) protected pastures were collected and analyzed for microbial community composition, biomass, potential exoenzyme activities, and a suite of biogeochemical characteristics. We found no differences in microbial communities or the soils associated with native vs. exotic grasses. Overall, mesquite bole patches differed from other patches in all soil characteristics except potential enzyme activity: soil temperature was significantly lower, and total carbon (C) and soil moisture were significantly higher. Potential activities were lowest for bare ground and highest at shrub dripline patches for all seven exoenzymes tested. Mean potential activities for C and phosphorous (P) hydrolyzing enzymes in long-term protected pastures (C: 21.4 ug activity g^-1 h^-1 ± 2.3; P: 29.8 ug activity g^-1 h^-1 ± 3.5) were significantly higher than those in grazed pastures (C: 16.6 ug activity g^-1 h^-1 ± 2.1; P: 15.8 ug activity g^-1 h^-1 ± 2.5), suggesting long-term effects of past land use on current soil microbial populations.