Possible sources of P and nitrogen (N) in ranchland runoff include fertilizer, manure, and local organic soil. The mobility of P and N from these sources is enhanced by drainage improvements, soil disturbance, or fertilizer-enhanced bacterial activity. Determining contributions of nutrients from multiple sources is complicated and in most cases involves controlled field studies, extensive multi-scale sampling, and mass balance calculations tied to the type and intensity of land use (Fluck et al., 1992; Halberg and Keeney, 1993; Spalding and Exner, 1993; Boggess et al., 1995). In some cases, nitrogen isotope ratios can be used to distinguish between isotopically distinct nitrogen sources from fertilizer or animal waste, but interpretations are generally complicated by bacterially-mediated processes of nitrification and denitrification that can modify nitrogen isotopic compositions (Heaton, 1986; Seiler, 2005). Similar isotopic identification of P sources is not possible because natural P is monoisotopic.

Figure 1. Map of the MacArthur Agro-Ecology Research Center (MAERC) showing locations of three soil cores and locations of two pastures (S5, W4) where runoff water was collected. [larger image]

Indirect evidence of nutrient sources in soil or runoff can be based on the presence (or absence) of other fertilizer-derived constituents such as uranium (U) from phosphate fertilizer or sulfur (S) from ammonium sulfate fertilizer. The sensitivity and reliability of U and S as indirect tracers are improved in areas with low background concentrations of these elements. In addition, U and S may serve as analogs for P and N to the extent that they have similar mobility in the environment under study. Although elevated concentrations of U or S in historically fertilized soils are suggestive of chemical fertilizer contributions (Rothbaum et al., 1979), stronger indications are based on the distinctive isotopic composition of fertilizer-derived U or S compared to their natural counterparts. For example, the combination of elevated U concentrations in peat and surface water and distinctive fertilizer-like ²³⁴U/²³⁸U activity ratios (AR) of 1.0 ± 0.05 have been used to identify fertilizer impacts in agricultural areas south of Lake Okeechobee (Zielinski et al., 2000).

Soil and water samples for this study were collected within the south half of the 4,171-ha Buck Island Ranch, Highlands County, central Florida (Figure 1). This fully operational cattle ranch is the main component of the MacArthur Agro-Ecology Research Center (MAERC), a division of Archbold Biological Station. The MAERC was established in 1988 as a center for research on the relation between cattle ranching and the native ecosystems of central Florida. In particular, MAERC provides opportunities for controlled, field-scale experiments that monitor the effects of various ranchland management practices on soils, runoff, and the local wetland and grassland ecosystems (Capece et al., 2006; Swain et al., 2006).

In this study, U concentrations and U isotopic compositions in collected samples are used to identify, and in some cases quantify, fertilizer-derived U in soil and surface runoff. Data from a soil profile in improved pasture, with a history of P fertilizer application are compared and contrasted with similar data from nearby semi-native pasture and undeveloped native grassland, with no history of P fertilizer application. Surface water runoff samples from the improved pasture are also compared to contemporaneous runoff samples from unimproved pasture. Finally, the mode of occurrence and mobility of U and S in soil and water are evaluated with respect to possible analogous behavior of P and N.