Soil phosphorus variability: scale-dependence in an urbanizing agricultural landscape

We examine the hypothesis that human activity changes patterns of variance in soil P (Bray-1) concentrations across several spatial scales. We measured sod P concentrations and variability for each of four different land uses at three distinct levels of analysis. Land uses were remnant prairie, lawns, corn fields of cash grain farms, and corn fields of dairies in Dane County, Wisconsin (USA). For each land use type, levels of analysis were sites (an agricultural field, residential lawn or prairie, ranging in size from 100 m(2) to approximately 20 ha), 10-m plots within a site, and points within the 10-m diameter plot. The rank of mean soil P concentrations was cash grain > dairy > lawn > prairie. For all land use types, most of the variance was accounted for by site-to-site variation. Among-site variance was higher for human-dominated sites (0.55, 0.15, 0.14 [log (mg/kg)](2) for cash grain, dairy, and lawn sites, respectively) than it was for prairies (0.07 [log (mg/kg)](2)). However, prairies had the highest among-plot variation (0.04 [log (mg/kg)](2)) compared to other sites (0.01, 0.002, and 0.01 [log (mg/kg)](2) for cash grain, dairy, and lawn sites, respectively). The results indicate that in this watershed, human activity has increased the mean soil P and variance of soil P, and shifted the scale of variance to larger spatial extents. Human impacts on landscape pattern extend to soil properties that affect nutrient flow and eutrophication of surface waters. Because soil P turns over slowly, the legacy of altered soil P patterns may affect freshwaters for centuries.