The use of bayesian networks and bootstrap to evaluate risks linked to the microbial contamination of leafy greens irrigated with reclaimed water in Southeast Spain

•Bayesian networks can build risk assessment models in cases with large information gaps.•Data augmentation through bootstrap can account for data uncertainty in the calculations.•The risk of irrigation by reclaimed water would be the same as with conventional methods.•The risk would be dominated by cross-contamination from the soil by splashing.