Comparison of measured and simulated phosphorus losses with indexed site vulnerability

Nonpoint-source losses of agricultural phosphorus (P) at field and watershed scales must be quantified to facilitate selection and placement of P control measures. Quantification of P loss has been pursued through field monitoring, simulation models, and risk assessment indices. However, the intended users of these methods differ, impacting each method's functional design and ease-of-use. For example, the Pennsylvania P Index, a risk assessment tool for planners, requires less discipline-specific knowledge and more readily available data than the Soil and Water Assessment Tool (SWAT), a complex, watershed-level, research-based simulation model. This study compared measured losses of P from the outlet of a 39.5 ha mixed land use watershed (FD-36) in south-central Pennsylvania with watershed-level losses predicted by SWAT. Measured watershed exports of dissolved P (0.06 kg ha^-1) and total P (0.24 kg ha^-1) during the 7-month sampling period were similar in magnitude to SWAT-predicted losses (0.05 and 0.73 kg ha^-1, respectively). Additionally, the study compared field-level P losses predicted by SWAT with field-level vulnerabilities to P loss derived by the P Index. The P Index and SWAT categorized 73% of the 22 fields similarly in terms of vulnerability to P loss, with Pearson correlation significant at p = 0.07; all except one of the remaining six fields were over- or underpredicted by a single risk category. Results indicate that while actual P loss from FD-36 was small, three fields contributed a major proportion of this loss. Additionally, this study suggests that the P Index can provide land managers with a reliable assessment of where P loss occurs within a watershed, thus allowing more effective placement and selection of conservation practices, which lead toward improved downstream water quality.