Using Bayesian Networks to Assess Effectiveness of Phosphorus Abatement Measures under the Water Framework Directive
Abstract
The EU Water Framework Directive requires all water bodies within the EU member states to achieve a "good status". Many economic assessments assume the "good status" is achieved using selected measures and evaluate only associated costs and benefits.
In this paper, Bayesian networks are used to test this assumption by evaluating whether the "good status" can be achieved with the selected abatement measures. Unlike in deterministic analysis, Bayesian networks allow effectiveness of measures of the same type to vary, which adds credibility to the analysis by increasing its robustness.
The approach was tested on Stanovice reservoir in Czechia using a set of 244 previously designed measures. The results show the target will be met with a probability of 72.4% using the most cost-efficient measures.
Based on the results, improvements to the measure selection process are suggested.