Just share my suggestion according to my experience. First, you should measure emission from each main system which produce emission like CO2, CO, NxO, . Second, converting those compounds into CO2 (g) equivalent by using emission factor and formula which have been provided in TIER standards depend on kind of industry. Third, measure standard deviation of emission use CO2 equivalent data as variance.
Thank you, but the problem is that the emission from source A may be correlataded with that of source B and, thus, the variance of the total emission is not the sum of the individual variances. Therefore, the problem boils down to estimate the correlation of emissions. Does anybody know of references that propose methods for estimating these correlations?