Methods and Example for Development of a Probabilistic Per-Capita Emission Factor for VOC Emissions from Consumer/Commercial Product Use

Quantitative methods for characterizing variability and uncertainty are applied to a case study of developing a probabilistic national per-capita emission factor for volatile organic compound (VOC) emissions from consumer/commercial product use. VOC profiles of 20 product categories were investigated. For each category, measures of VOC weight percentage for different formulations in that category were analyzed. Beta distributions were fit to each data set to represent variability in the VOC content of different products. Bootstrap simulation methods were used to quantify uncertainty in the fitted distribution and uncertainty in the mean VOC content for each category. The uncertainty in the mean VOC content ranges from as much as minus 79 percent to plus 130 percent, which emphasizes the importance of recognizing and accounting for uncertainty in emissions estimates. The skewness in some uncertainty estimates illustrates the importance of using a numerical simulation approach that does not impose restrictive symmetry assumptions on the confidence interval for the mean. A challenge in developing per-capita emission factors is regarding how to handle the uncertainty in the product consumption and population data since only point estimations were available. The solution is based on consulting experts to elicit judgments for a probability distribution and an uncertainty confidence interval assigned to those inputs. The Monte Carlo technique then is used to simulate a group of samples that represent the uncertainty in the per-capita emission factor. One benefit of probabilistic analysis is that it enables the identification of key sources of uncertainty in the input assumptions, which in turn allows for prioritization of future data collection or other efforts to improve the emission factor. In this paper, we briefly present the probabilistic analysis method, the data sets, the results of the analyses, and key findings and recommendations.