At a macro level on-road vehicular emissions, including greenhouse gases (GHG), are estimated, requirements compliance verified, and the impact of technological and policies alternatives evaluated via inventory models as, for instance, the one recommended by the Intergovernmental Panel on Climate Change (IPCC) [1], the USEPA’s MOVES [2], the European Environment Agency (EEA)’s COPERT [3], and many others developed for this purpose.
Normally they are based on the fleets’ characteristics, conditions of use, and activity, i.e., on data as fleet size divided into classes (ex. cars, buses, and trucks), vehicle characteristics (ex. age distribution and emission control technologies used), and vehicles in use average statistics (ex. average speed, fuel consumption, and emissions). However, at first, these models are not able to reproduce short-term fleet activity variations caused by changes in the economy as, for instance, fuel price fluctuations. This paper shows the development of an inventory model aiming to take these economic factors into account. Despite being developed specifically for the Brazilian context, the same technique can be used in other bottom-up inventory models.
For this study, first, a set of worksheets was created to integrate all data and calculations of a typical bottom-up model. The set was interconnected and configured to facilitate, through an iterative process, the fine-tuning of more uncertain parameters, in such a way that total fleet consumption estimated by this model as much as possible reproduces the fuel consumption observed in the country.
Thereafter, based on economic parameters that had been identified as statistically significant, an econometric model to replicate this country's total fuel consumption was developed. And by merging this econometric model into the previous bottom-up model, it was created a hybrid model that, based on the observed fuel consumption, obtained statistically more robust results than the bottom-up model itself. Due to the difficulty to link emission sources to air quality measurements, the same approach cannot be used for emissions. But the consumption can be easily converted into carbon dioxide (CO2) emission. And as the hybrid model improves the fleet activity estimate, it also enhances the estimates of other emissions as well.