Beyond Deterministic Air Quality Modeling: A Probabilistic Screening Approach for Emission Inputs in AERMOD

https://doi.org/10.1093/inteam/vjaf098

Integrated Environmental Assessment and Management, Volume 21, Issue 6, November 2025

This article is part of the special series “Probabilistic Approaches for Environmental Risk Assessment, Decision-Making, and Regulatory Criteria Development.” The series presents a collection of articles that advance the understanding of probabilistic methodologies and their versatility for robust, transparent, data-based environmental risk assessment and standards derivation across a range of media that align with regulatory objectives to protect aquatic and terrestrial biota, human health, and vulnerable populations.

Authors
Tanvir Khan, PhD, NCASI
Zach Emerson, PhD, NCASI

Abstract
Traditional air dispersion modeling usually relies on deterministic frameworks that use multiple conservative assumptions as inputs. For example, atmospheric pollutant concentrations are often overestimated by using maximum emission rates for point sources, an approach recommended by regulatory agencies that may not reflect typical operating conditions, especially for sources with variable emissions. To better understand how emission variability affects modeled pollutant concentrations, this study presents a novel probabilistic modeling framework designed to estimate pollutant concentrations from industrial sources, with a focus on integrating variability in emission rates. The framework incorporates a Monte Carlo screening method combined with AERMOD to evaluate the atmospheric dispersion of emissions. This approach provides a more flexible and data-driven method for determining emission rates compared to traditional modeling methods. The utility of the method was demonstrated through an application to the pulp and paper industry that included modeling of nitrogen oxides (NO_x) emissions from a virtual kraft pulp mill. A base AERMOD simulation, using maximum emission rates, predicted the highest concentration of ambient nitrogen dioxide (NO₂), representing a worst-case scenario. In contrast, using emission rates derived from the Monte Carlo screening method, the estimated ambient NO₂ concentrations were substantially lower. The method can be further enhanced by incorporating additional sources of variability and expanding its application to other pollutants.

Keywords: AERMOD, Industrial Emissions, NAAQS, Probabilistic Modeling, Air Dispersion