A procedure to estimate variances and covariances on GHG emissions and inventories

This study presents a method for estimating the mean and variance of total CO2 emission from multiple sources used by a company. The procedure is also readily applicable to estimate these parameters for other greenhouse gases (GHG) inventories and to determine a reliable confidence interval for the total emissions of GHG of a company. Our method represents an improvement over the existing methods that assume independence between emissions from different sources. The foundation of the proposed method is an iterative decomposition process applied to analyze the emissions correlations among activities, raw materials and other inputs used in a company's operations. From these correlations and the individual estimates of means and variances of emission factors, we show how to generate a confidence interval for the total GHG emission of a company. The application of the method is illustrated for a hypothetical manufacturing plant of bicycles and car toys, whose total CO2 emission is estimated within a precise confidence interval.