Revista internacional de contaminación ambiental
versão impressa ISSN 0188-4999
MENDOZA, Alberto e GARCIA, Marisa R.. Inverse modeling applied to the analysis of the Guadalajara metropolitan area (México) emissions inventory. Rev. Int. Contam. Ambient [online]. 2011, vol.27, n.3, pp.199-214. ISSN 0188-4999.
Emission inventories are one of the sources of highest uncertainty in the process of applying complex air quality models. If there are errors in such information, the conclusions that result from the model application can be clouded. This work presents the application of an inverse modeling technique to adjust the emissions inventory of the Guadalajara Metropolitan Area, Mexico, which is then compared against the original inventory. The modeling system estimates, by means of an iterative process, scaling factors to the emissions of a base case in such a way that after applying such corrections the error between observed and simulated concentrations is minimized. The obtained results indicate that, during the analyzed case study, the daily CO emissions needed to be globally varied from reductions in the order of ~16% to increments of ~60%. NOx and SO2 emissions are both apparently underestimated: NOx emissions were increased in the range of 100% to 150% with respect to the base case, while SO2 emissions increased in the range of 20% to 140% the nominal value. The changes suggested by the inverse model significantly increase the performance of the direct model, in the sense of its ability to replicate the observations, including O3, a secondary pollutant. However, not all discrepancies between observations and model-derived concentrations were solved due to the limitation of only scaling hourly domain-wide emissions but not by sub-domains since not enough monitoring stations were available and by the fact that the direct model is not perfect.
Palavras-chave : atmospheric modeling; air quality; ozone; emission sources.