Soot and SO<sub>2</sub> contribution to the supersites in the MILAGRO campaign from elevated flares in the Tula Refinery

Abstract

Abstract. This work presents a simulation of the plume trajectory emitted by flaring activities of the Miguel Hidalgo Refinery in Mexico. The flame of a representative sour gas flare is modeled with a CFD combustion code in order to estimate emission rates of combustion by-products of interest for air quality: acetylene, ethylene, nitrogen oxides, carbon monoxide, soot and sulfur dioxide. The emission rates of NO 2 and SO 2 were compared with measurements obtained at Tula as part of MILAGRO field campaign. The rates of soot, VOCs and CO emissions were compared with estimates obtained by Instituto Mexicano del Petróleo (IMP). The emission rates of these species were further included in WRF-Chem model to simulate the chemical transport of the plume from 22 to 27 March of 2006. The model presents reliable performance of the resolved meteorology, with respect to the Mean Absolute Error (MAE), Root Mean Square Error (RMSE), mean bias (BIAS), vector RMSE and Index of Agreement (IOA).WRF-Chem outputs of SO 2 and soot were compared with surface measurements obtained at the three supersites of MI-LAGRO campaign. The results suggest a contribution of Tula flaring activities to the total SO 2 levels of 18 % to 27 % at the urban supersite (T0), and of 10 % to 18 % at the suburban supersite (T1). For soot, the model predicts low contribution at the three supersites, with less than 0.1 % at three supersites. According to the model, the greatest contribution of both pollutants to the three supersites occurred on 23 March, which coincides with the third cold surge event reported during the campaign.

Impact of external industrial sources on the regional and local SO<sub>2</sub> and O<sub>3</sub> levels of the Mexico megacity

“…2). However, the model does not reproduce the peak on 24 March neither when using the model configuration of this work nor with the configuration of our previous study (Almanza et al, 2012).…”

“…A first estimate of the contribution from flaring activities from Tula refinery is presented. The mass flow rates of acety-lene, ethylene and NO x from our previous work (Almanza et al, 2012) were used to represent the precursor emissions of the elevated flares in the Refinery. They were estimated with a CFD combustion code.…”

“…Previous studies have shown the benefit of applying data assimilation in the model performance of retrospective air quality simulations (Lo et al, 2008;Otte, 2008;Ngan et al, 2012), including the uncertainty reduction of the horizontal transport in the lower troposphere (Gilliam et al, 2012). Related work during MILAGRO campaign reported that the inclusion of full diffusion and of observation nudging improved the accuracy of meteorological fields (Doran et al, 2008;Fast et al, 2009;Bei et al, 2010;Song et al, 2010;Karyidis et al, 2011). In the present study, the multi-scale FDDA strategy is applied in order to reduce the model error in the meteorological fields, since it is aimed to explain an episodic SO 2 event as well as to estimate the Detail of the extent of the Mexico City Metropolitan Area (light brown) including the location of the monitoring stations (black stars) and the Sierra de Guadalupe (SG) in the inner domain d03.…”

“…This potential contribution from cement (IMP, 2006). IMPei was used in our previous study focusing on flaring emissions (Almanza et al, 2012). It is based on information provided by the Mexican Ministry of Environment (SEMAR-NAT).…”

“…The Tula refinery has three elevated flares and some ground flares, but in this work the focus is on elevated flares. In order to represent flaring activities, the emission rates of some combustion by-products are based on our previous study (Almanza et al, 2012). Briefly, the flame of an equivalent elevated flare representing sour gas flaring emissions from the refinery is modeled with a computational fluid dynamics (CFD) combustion code.…”