Abstract: The urbanization process substantially alters every aspect of the soil environment. In this study, we compared soil microclimate, chemistry, and physical characteristics of unmanaged natural soils with managed soils of three common urban land uses (stormwater natural treatment systems, ornamentally landscaped areas, and lawns) across three University of California campuses. Over the course of 1-year, average monthly soil temperatures among land uses showed fewer than expected differences. Average monthly soil moisture reflected wet and dry seasonal changes, but this pattern was muted in managed land uses compared to natural soils due to irrigation. From April through December, lawns and landscaped areas were significantly wetter than natural soils (e.g. 1.5–3 times higher in August and September). Soil organic matter, total carbon, and total nitrogen were significantly higher in lawns compared to other land uses, while their bulk density was significantly lower. Principle component analysis confirmed that natural and lawn soil properties were distinct from each other. These differences in the managed soils, particularly lawns, are attributable to typical urban land management practices such as fertilization, irrigation, and the installation of organic-rich sod. As urbanization continues to change the native landscape of semi-arid regions, these changes to soil microclimate, chemistry, and physical characteristics are important to consider for urban best practices and sustainable development.

The synergetic effects of metal(loid)s and soil characteristics on bacterial antibiotic resistance genes (ARGs) in green stormwater infrastructure (GSI) has been relatively understudied. Surface soil samples from six GSIs in Southern California over three time periods were assessed for selected ARGs, class 1 integron-integrase genes (intI1), 16S rRNA genes, and bioavailable and total concentrations of nine metal(loid)s, to investigate the relationships among ARGs, soil characteristics, and co-occurring metal(loid)s. Significant correlations existed among relative gene abundances (sul1, sul2, tetW, and intI1), total metal(loid)s (arsenic, copper, lead, vanadium, and zinc), and bioavailable metal(loid) (arsenic) (r = 0.29-0.61, p_adj < 0.05). Additionally, soil texture, organic matter, and nutrients within GSI appeared to be significantly correlated with relative gene abundances of sul1, sul2, and tetW (r = -0.57 to 0.59, p_adj < 0.05). Multiple regression models significantly improved the estimation of ARGs in GSI when considering multiple effects of soil characteristics and metal(loid)s (r = 0.74, p_adj < 0.001) compared to correlation results. Total arsenic was a significant (positive) correlate in all the regression models of relative gene abundances. This work provides new insights into co-dependencies between GSI ARGs and co-occurring metal(loid)s, indicating the need for risk assessment of metal(loid)-influenced ARG proliferation.