MODELING ODORS AND VOCS: A COMPARISON OF MODELS AND MODELING TECHNIQUES

Wastewater treatment plants, landfills, feedlots, and sources of odors, VOCs, and air toxics have come under increasing pressure from surrounding communities as buffer zones around these sources shrink, and health and quality of life issues take on increasing importance. An important tool in understanding the transport and fate of these emissions is the dispersion model. Used for many years in the regulatory arena for the permitting of traditional sources, dispersion models basically use mathematical equations to convert the mass emission rate from a source to a mass ambient air concentration some distance downwind. When used for odors; however, a dispersion model may have to account for the difference in impact criteria and time frame between compound-specific emissions (a given mass concentration over a specified averaging time, generally on the order of 1 hour) and odors (short-term intensity or odor strength and frequency of odors, or a combination of all three). In addition, most regulatory modeling is predicated on the need to determine a maximum offsite impact, regardless of location. Impacts at sensitive receptor locations are generally the focus of odor evaluations.Various models, steady-state and puff, Gaussian and Lagrangian, have been used for odor assessment. As regulatory modeling continues to move towards in the direction of a single powerful and multipurpose model, it makes sense to model odors, if possible, with the same model. This common approach to overall air quality assessment reduces the burden on a facility to defend a methodology outside standard modeling practice, expedites the review required by the regulatory agencies, and produces more consistent results, especially when a given source emits both odors and VOCs/air toxics. Steadystate Gaussian models, like the Industrial Source Complex-Short Term (ISCST3) model that assume direct transport from a source to a receptor for every hour of meteorological data, generally yield the conservatism required of a regulatory application. For odors, however, a more realistic model that tracks a puff as it travels with the wind, bending, stretching, and turning as it travels downwind, should result in more accurate prediction of both magnitude and location of downwind odor impacts.The CALPUFF model, currently approved for long range transport, has these capabilities. CALPUFF is a Lagrangian puff model that accounts not only for the change in wind direction, but also includes the effect of terrain on wind flow. In addition, CALPUFF can consider nearfield effects, such as building wake and cavity flow. It can use onsite or representative meteorological data from one or more stations, as well as terrain and geographical data in the nearfield and farfield. CALPUFF also includes an odor algorithm based upon an exponential increase in impacts with a decrease in averaging time.This paper describes a preliminary analysis of the emissions from a hypothetical wastewater treatment facility located in a suburban/rural area using the two different dispersion models, ISCST3 and CALPUFF. The facility emits odors and VOCs/air toxics from the wastewater processes and onsite engines burning digester gas, as well as criteria pollutants (NO2, SO2, and CO) from the engines and boilers. Sources at the facility include aeration basins (area sources), odor control stacks (point sources), and engine generator stacks (point sources). Criteria for the study are the ambient levels of hydrogen sulfide (a state-sponsored odor standard), short-term and annual ambient air levels of air toxics/VOCs, and odor strength within the community.Since the facility had to meet local, state, and federal ambient air quality standards, as well as state air toxic requirements, the use of regulatory-approved models was required for these contaminants. However, odor strength within the community was also an important issue. The usual approach to this type of analysis, with its complex source contributions and regulatory component, is the use of the ISCST3 model; USEPA's preferred model for multisource facilities. An alternative, though more data-intensive approach is the use of the CALPUFF model. In this preliminary analysis, the emissions from the facility were modeled with ISCST3. In addition, the same facility set-up and emissions were also modeled using CALPUFF in the fully refined form. Predicted impacts for the two alternative models were compared.