ODOR MODELING USING NON-STEADY-STATE METHODS

It is well known that odors are a major air quality issue of public concern which is highlighted by the large number of complaints received by government agencies. Dispersion modeling is being increasingly used to assess and quantify odor impacts. The focus of attention in the last decade has been in trying to establish odor guidelines in the hope of bringing a degree of consistency to the control and measurement of odors. Little effort has been spent assessing the suitability and applicability of dispersion models, such that the steady-state ISCST3 (ISC3) model has become the de-facto odor model despite its many shortcomings for this type of application. This paper aims to highlight the combined effect of a diagnostic meteorological model and its non-steady-state puff model to more realistically estimate the diffusion and dispersion of odor-causing pollutants compared to steady-state models such as ISC3.The limitations of the steady-state Gaussian plume assumptions to odor applications are discussed in terms of spatial and temporal variability, the importance of light wind dispersion, the treatment of calm wind conditions, causality effects and complex terrain effects. A case study of a complex terrain site and stable nighttime conditions is used to graphically represent some of the several important reasons why the non-steady-state modeling approach offers significant technical advantages over steady-state modeling. The non-steady-state CALPUFF model, recently adopted by the U.S. EPA as a Guideline Model is used in the example application as well as ISC3.Other technical considerations that make the CALPUFF modeling system appealing for use in odor applications include; the ease of applying short-term conversion factors directly to the concentration results, or, according to specific meteorological parameters directly within the input file. Secondly, an option to include state-of-the science turbulence-based dispersion parameterizations, the ability to allow variable emission factors and source parameters, and, the ability of the model to allow multiple effects within a single modeling framework.