Application of probabilistic risk assessment to agricultural nonpoint source pollution

ABSTRACT:

Current and proposed regulations require farms to be managed for environmental benefit, but tools are needed to explain the necessity of improved management to farm operators. One tool for exploring the risk associated with management practices is probabilistic risk assessment (PRA). A PRA for the discharge of nitrogen (N) from agricultural nonpoint sources (NPSs) to a stream was performed for a small watershed in Virginia. Using fault tree analysis, a type of PRA, probability of occurrence for excessive N discharge to the stream during a runoff event was evaluated both qualitatively and quantitatively. Qualitative analysis showed that proper management of cover, nutrients, and livestock is key to controlling the probability of agricultural NPS pollution occurrence. While this conclusion is not new, presentation of pollution causing events in a graphical format is new. Land use, management practice, stream flow, and water quality data, along with expert opinion, were used to assign probabilities to individual events within the fault tree and to calculate overall probability of N discharge to the stream. Quantitative assessment showed the impact of best management practices (BMPs) on reducing the likelihood of N discharge. Probability of N discharge during a runoff event for the watershed decreased by 0.08 from 1987 to 1993 due to the implementation of BMPs during that time. For the same time period, a subwatershed isolating two dairies showed a 0.09 decrease in the probability of N discharge. The fault tree method was useful in representing changes within the watershed. Data requirements needed to complete the PRA were similar to those of other integrative approaches to assessing NPS pollution.