Heterogeneity of soil nutrients and subsurface biota in a dryland ecosystem

Dryland ecosystems have long been considered to have a highly heterogeneous distribution of nutrients and soil biota, with greater concentrations of both in soils under plants relative to interspace soils. We examined the distribution of soil resources in two plant communities (dominated by either the shrub Coleogyne ramosissima or the grass Stipa hymenoides) at two locations. Interspace soils were covered either by early successional biological soil crusts (BSCs) or by later successional BSCs (dominated by nitrogen (N)-fixing cyanobacteria and lichens). For each of the 8 plant type×crust type×locations, we sampled the stem, dripline, and 3 interspace distances around each of 3 plants. Soil analyses revealed that only available potassium (K_av) and ammonium concentrations were consistently greater under plants (7 of 8 sites and 6 of 8 sites, respectively). Nitrate and iron (Fe) were greater under plants at 4 sites, while all other nutrients were greater under plants at less than 50% of the sites. In contrast, calcium, copper, clay, phosphorus (P), and zinc were often greater in the interspace than under the plants. Soil microbial biomass was always greater under the plant compared to the interspace. The community composition of N-fixing bacteria was highly variable, with no distinguishable patterns among microsites. Bacterivorous nematodes and rotifers were consistently more abundant under plants (8 and 7 sites, respectively), and fungivorous and omnivorous nematodes were greater under plants at 5 of the 8 sites. Abundance of other soil biota was greater under plants at less than 50% of the sites, but highly correlated with the availability of N, P, K_av, and Fe. Unlike other ecosystems, the soil biota was only infrequently correlated with organic matter. Lack of plant-driven heterogeneity in soils of this ecosystem is likely due to (1) interspace soils covered with BSCs, (2) little incorporation of above-ground plant litter into soils, and/or (3) root deployment patterns.