Sensitivity analysis of the Pennsylvania phosphorus index for agricultural recycling of municipal biosolids

ABSTRACT:

Forty-nine states are developing phosphorus (P) site indices to assess vulnerability of agricultural fields to P loss. When a P index has multiple variables and mathematical operations, it is often difficult to predict the sensitivity of the final site rating to individual input variables. Sensitivity analysis is a powerful diagnostic technique for assessing the behavior of a P index, prioritizing data collection and research efforts to improve its predictive capabilities, and for focusing management efforts. In this study, nominal range sensitivity analysis was used to determine the effect of input factor perturbations on the final P index score for agricultural biosolids recycling. Baseline conditions (P index score = 116) for the sensitivity analysis were established from a detailed evaluation of representative Pennsylvania sites targeted for biosolids application in 2003. The base condition was evaluated for Sensitivity Coefficient and Swing to identify the importance of individual input factors on final P index rating. For biosolids sites, P loss vulnerability forecast by the Pennsylvania P index was found to be relatively insensitive to subsurface drainage and mineral fertilizer application rate and method (with Swing values of 23, 18, and 4 respectively). Accordingly, uncertainty in these Low Impact P index variables will have little adverse effect on the reliability of index scores for biosolids land application. For sites receiving biosolids, the Pennsylvania P index is most sensitive to the biosolids application rate and method, organic P source coefficient, and nature of the site buffer strip. Remarkably, varying the P source coefficient (alone) over its full range of plausible input values resulted in a P index rating Swing of 167 points. Thus, beyond changing biosolids application rates and methods, the greatest potential to reduce the P index score is in establishing riparian buffer zones and reducing biosolids P source coefficients. It is important to note that the differential sensitivities of fertilizer P source rate and organic P source rate in the Pennsylvania index do not imply that organic P sources are inherently more susceptible to offsite P export than chemical fertilizers. Research efforts should focus on High-Impact factors for improving the predictive capability of the P index.