Modelling the dispersion of SO2 emissions from the chevron (Cape Town) oil refinery using the US EPA dispersion models AERMOD and CALPUFF

Thesis submitted in partial fulfilment of the requirements for the degree Master of Technology: Chemical Engineering in the Faculty of Engineering at the Cape Peninsula University of Technology 2013In South Africa, the Department of Environmental Affairs (DEA) under the National Environmental Management Act, 1998 (Act 107 of 1998) (“NEMA”) sets out a series of environmental management principles that apply to the interpretation and application of all legislation that may affect the environment. Since 1998, various specific environmental statutes that fall under the NEMA framework have been promulgated, including the National Environmental Management: Air Quality Act, Act 39 of 2004 (NEM: AQA).NEM:AQA provides norms and standards for all technical aspects of air quality management. The National Framework (Sections 7 and 8 of NEM:AQA) must provide mechanisms, systems and procedures to promote holistic and integrated air quality management through pollution prevention and minimisation at source, and through impact management with respect to the receiving environment, from local scale to international issues. Among other measures, the NEM:AQA (Section 9) requires the establishment of Ambient Air Quality Standards and Emission Standards. These standards were promulgated in December 2009 and March 2010 respectively. Air quality monitoring stations, which sample and analyse pollutant concentrations continuously, are a common method of assessing air quality in a region. But a few continuous monitors located in source given region or airshed are inadequate for assessing compliance with ambient air quality standards – they are only able to monitor concentrations at a fixed site, not through the entire region of impact. In contrast, the ambient air quality standards are applicable everywhere. Air quality models estimate ground level ambient concentrations throughout the modelling domain, and in principle (subject to proper validation) provide better estimates of area-wide concentrations and hence the basis for assessing compliance with air quality standards. The United States Environmental Protection Agency (US EPA) approved atmospheric air dispersion models AERMOD and CALPUFF were used in this thesis to predict the ground level concentrations of SO2 emitted from Chevron Refinery (Cape Town), for the year 2010. The modelling is validated by comparing measured ambient concentrations with modelled concentrations. The results showed AERMOD-modelled annual average values for 2010, based on refinery emissions only, are in good agreement with monitored values at the Table View and Bothasig sites, predicting the monitored values by -11% and +17% respectively. The 24-hr average values similarly are in good agreement with monitored values, on average over-predicting by 9% at Table View, although the fit of the day-to-day modelled vs monitored values is comparatively poor (R2=0.32); at the Bothasig site the corresponding values are - 36% and R2= 0.089. The AERMOD-modelled isopleths imply that the 2010 annual average concentrations exceeded the South African Standard of 50 μg/m3 in a small area in the immediate vicinity of the refinery. The hourly and 24-hourly average standard concentrations of 350μg/m3 and 125μg/m3 respectively are exceeded in significantly larger areas. The allowable exceedences for hourly and 24-hourly averages are also exceeded, implying that the hourly and 24-hourly standards were exceeded. CALPUFF-modelled average values for 2010, based on refinery emissions only, are in comparatively poor agreement with monitored values at the Table View and Bothasig sites, under-predicting the monitored values by -20% and -61% respectively. Since the AERMOD-modelled concentrations are in far better agreement with monitored concentrations, only AERMOD was used for further analysis. The Emission Standards promulgated in March 2010 included emission limit values for sulphur dioxide emitted from oil refineries. If the actual 2010 emission rates were adjusted downwards to match the emission standards (to be complied with from 1 April 2015), AERMOD modelling indicates that the annual, 24-hourly and hourly Ambient Air Quality Standards would not be exceeded. Based on this case study, the current Emission Standard for SO2 emissions from existing crude oil refineries is therefore coherent with the Ambient Air Quality Standards. Regulatory air dispersion modelling practices in South Africa are being standardised for model applications regulatory purposes and to ensure that dispersion modelling practices are undertaken in a compatible form to ensure that results from one dispersion model study can be compared directly to those from another. In this study both AERMOD and CALPUFF modelling complied with the draft South African guidelines for Air Quality Modelling, yet the CALPUFF- modelled outputs differed significantly from the monitored values. This emphasizes the importance of the inclusion of modelling validation in guidelines for modelling for regulatory purposes. The 2012 draft regulation should be amended to make validation of regulatory dispersion modelling compulsory rather than optional as per the draft