Abstract

"The Comprehensive Air Quality Model with Extensions (CAMx) has been applied to two summer photochemical episodes in the greater metro area of Santiago, Chile. It is shown that the CAMx model predicts reasonably well the spatial gradients of gaseous pollutants and the weekly cycle of variation for CO and NOx. The emission estimates for CO seem to be quite on agreement with actual emissions in 2000. NOx emissions are more appropriate for the 1998 scenario, whereas for 2000 they seem to overestimate actual emissions, so the mobile sources emission factors need to be verified. The SO2 simulations indicate that there relevant sources not accounted for in the inventory. There are problems with the VOC emissions for the area sources, because the model predictions for ozone are quite below what is being observed, for all network stations in 1998, a scenario with reliable NOx simulated levels. For the 2000 scenario, this problem is worsened because of excessive ozone scavenging. When we use the results of a comprehensive VOC campaign to produce a synthetic VOC profile of emissions for the whole city, simulated ozone gets closer to ambient measurements for the 1998 scenario. For the 2000 scenario, the overestimation in the simulated NOx levels is large enough to prevent any significant ozone sensitivity with respect to VOC increases. All things considered, CAMx has passed the ""proof of the concept"" for studying photochemical episodes at Santiago, Chile. The major sources of uncertainty are the emission inventory estimates for NOx and VOC. As a second source of uncertainty, meteorology was not properly represented on the second half of the January 1998 episode, and there are some problems with the stations closer to the Andes foothills overnight, particularly regarding ozone nocturnal profiles, that deserve further research."