Introduction: The index of new cars’ climate impact (Trafikverket 2010) calculated a reduction factor to quantify the positive impact of renewable fuels on CO₂ emissions for new cars in Sweden 2009. Although uncertainty for many input parameters is acknowledged, all calculations were made with point estimates.

Aim: The aim of this study was to verify consumer recommendations based on carbon dioxide emissions of new cars by re-assessing the CO₂ emissions considering the uncertainty in input data. The data uncertainty was quantified by using probabilistic risk assessment methodology. The objectives were (1) to calculate the possible reduction of CO₂ emissions for new cars in Sweden accounting for uncertainties; (2) to identify the most influential factors by conducting a sensitivity analysis; and (3) to assess the implications of interval analysis and probabilistic modelling for epistemic uncertainty in this context and thereby to contribute to the discussion on which method is the most appropriate to support decision making.

Method: Calculations were performed with the deterministic model proposed by Trafikverket (2010) using probability distributions (Bayesian probability distributions assigned by expert judgements) and intervals as model input. Probabilistic modelling was implemented as Monte Carlo simulation with uniform distributions and triangular distributions in Matlab. Interval calculations were done by hand.

Results: The best estimate from probabilistic modelling of CO₂ reduction when using E85 as a fuel of 30% (95% confidence interval = 10% to 52%) is in a similar range as the 20% given by Trafikverket (2010). The best estimate of 28% decrease for gas cars (95% confidence interval = 3% to 44%) is lower than the originally proposed 42%, but still in a similar range. The difference is due to the large extent of optimistic values taken by Trafikverket (2010). The interval analysis produced a wider range. For ethanol driven cars, a CO₂ reduction of 68%, an increase of 14% or anything in between is calculated, for cars operated on gas the result is a CO₂ reduction of 59%, an increase of 29% or anything in between.

Conclusions: The use of biofuels most likely reduces CO₂ emissions. The CO₂ emissions from the production of the biofuel are the most influential factors. The recommendations of Trafikverket to the consumers are still valid after probabilistic recalculation. Probabilistic modelling is more appropriate than interval analysis to guide authority and consumer decisions as it allows conclusions to be drawn and therefore can support decision making.