The Devil is in the Details: Odor Modeling and AERMOD

Air dispersion modeling has historically played an important role in the design and evaluation of wastewater treatment plants (WWTPs) and their processes. It is a highly cost-effective design tool when used to determine the need for odor or air toxic control, and enabling the comparisons between potential control alternatives. Modeling is also frequently used by local state, and federal regulatory agencies to determine if a WWTP is a significant source of odors and air toxics as part of the environmental review and air permit processes. For over two decades, the Industrial Source Complex-Short Term (ISCST3) Model was the model approved by most regulatory agencies and routinely used in assessing the air quality impacts of WWTPs for permitting, risk assessment, environmental review, and odor. For better or worse, as the refined model of choice, ISCST3 provided a somewhat level playing field when comparing the impacts from different wastewater treatment plants and processes. However, many of the advances in the fields of atmospheric diffusion and modeling over that period were not incorporated into ISCST3. In order to include these scientific advances into regulatory modeling, the United States Environmental Protection Agency (USEPA) collaborated with the American Meteorological Society (AMS) in order to develop a model that could better characterize plume dispersion. The result of this collaboration was the AMS/EPA Regulatory Model (AERMOD). AERMOD incorporates state-of -the-practice planetary boundary layer (PBL) principles into a plume dispersion model. After years of development and enhancement, the AERMOD modeling system officially replaced ISCST3 in December of 2006 as the preferred and approved regulatory model for simulating the impacts of emissions from a variety of sources, including power plants, industrial facilities, landfills, hazardous waste facilities, etc. Because of its wide regulatory acceptance, AERMOD is also replacing ISCST3 in odor and air toxic modeling of WWTPs.Unlike ISCST3, the AERMOD modeling system is made up of three programs. The AERMOD meteorological preprocessor (AERMET) develops the meteorological input file from surface and upper air data, while the AERMOD terrain pre-processor (AERMAP) generates terrain and receptor grid input files from digitized terrain data. The AERMOD dispersion model then uses source information combined with the AERMET and AERMAP output to determine the concentrations at the specified receptor locations. Because AERMOD has the ability to characterize the profile of the PBL and the dispersion of pollutants within it more realistically, the model requires a significantly greater amount of site-specific input for the user to evaluate, and more options to select.Such major changes in model code are often associated with significant differences in the results and conclusions drawn from the earlier model. The change from ISCST3 to AERMOD has been especially challenging. It has been found that the results of modeling analyses with AERMOD can be highly dependent upon the selection of site-specific input. Although USEPA has provided some guidance on the appropriate application of AERMOD, this regulatory guidance on the use of AERMOD has not kept pace with its promulgation and continues to evolve; nor has the guidance expanded to cover the more non-regulatory applications of modeling, such as odor modeling. In addition, the refinements in the algorithms used in the AERMOD system came with a number of unintended consequences, such as inordinate model run times, often orders of magnitude greater than found using ISCST3.This paper discusses problems and issues that have been encountered when using AERMOD to model odors from WWTP processes. The paper focuses on the challenges encountered when performing an odor assessment for a major municipal wastewater treatment plant first modeled with ISCST3 and then with AERMOD, and presents the results obtained from different interpretations of the input selection. The wastewater treatment plant itself combined a variety of process source types onsite, including tanks (“area sources”), odor control stacks and building vents (“point sources”), andopen doors and weirs (“volume sources”). With impending plant upgrades, determining which sources may require control was a high priority. Unique issues (and some unintended consequences) emerged when modeling the process emissions from the plant using AERMOD. These included:selection of representative meteorological data,characterization of surface characteristics,selection of urban/rural options,effect of urban population on model output, andeffect of source characterization on model runs times.Differences in the model results depending upon the selection of surface characteristics (e.g., at the meteorological station versus the application site, and seasonal effects), selection of urban versus rural characteristics, determination of the appropriate urban population, especially in a large metropolitan area, problems when modeling large open area sources and comparisons of modeling results using alternative approaches are presented in the paper. When used for odor control evaluation and design, these differences in the model output can potentially lead to different conclusions regarding odor control and different designs.