AERMOD and CALPUFF are well-known air dispersion models that are used worldwide for assessing the impacts of air emissions. There are important differences between the two that make them applicable for different types of scenarios.
AERMOD Model is EPA Preferred
The AERMOD model, developed and maintained by the U.S. EPA, has become one of the most important and widely used air dispersion models in the world since the release of its first full version in 2006. Since that time, stakeholders including government, industry, academia, and consultants have collaborated to improve the model on a regular basis. It is currently designated as the preferred model by U.S. EPA, meaning it can be used for most modeling applications without specific approval. Other models, including CALPUFF, typically require specific approval from the appropriate regulatory authority before they can be used for modeling in the U.S.
AERMOD Uses Plume Model - Predictions up to 50km
AERMOD is a steady-state Gaussian dispersion model that represents the current state-of-science, including advanced planetary boundary layer (PBL) parameterizations. Gaussian plume models assume that continuously released material is transported directly opposite to wind direction, and they assume this horizontal plume centerline travels downwind in a straight line indefinitely. They also assume that time-averaged spreading of the pollutants produces a normal (Gaussian) distribution of pollutant concentrations, both horizontally and vertically through the pollutant plume. Regardless of varying temporal or spatial conditions (e.g., meteorological or terrain variations), Gaussian models assume a constant, uniform set of meteorological conditions across the model domain for each period (e.g., hour of meteorological data) modeled, meaning the plume will never curve or change directions within a given hour. While a simplification of the real atmosphere, this approach yields a good match between modeled and observed results in the near-field because meteorological conditions are typically fairly uniform over short distance and time scales. However, because meteorological conditions can change quite drastically over longer distances or time scales (e.g. due to terrain effects or sunrise/sunset), the US EPA does not recommend using AERMOD for long-range transport modeling (>50 km) as the results will not be accurate.
CALPUFF is the Alternative Model Used for Predictions up to 200km
By comparison, CALPUFF is a Lagrangian puff model that accounts for the effects of varying temporal or spatial conditions on pollution transport, transformation, and removal. From 2003-2017 CALPUFF was the US EPA-promulgated model for long-range transport modeling, however, with the 2017 revisions to the Guideline on Air Quality Models (Appendix W to 40 CFR Part 51), it was reclassified as an alternative model that can be used in a screening approach for long-range transport assessments. As such, if a modeling assessment requires that impacts beyond 50 km be analyzed, then a modeler in the U.S. must contact their local regulatory authority to request the use of CALPUFF.
CALPUFF Utilizes Gridded WRF Meteorological Data Files
While AERMOD typically uses representative observational meteorological data (e.g., airport data) taken from a single location, CALPUFF utilizes meteorological data both from multiple observation sites and from meteorological grid models. This means that preparation of meteorological data for CALPUFF requires more expertise, more time, and greater computational resources than for models like AERMOD. There are several meteorological grid models, including the MM5 model, that can be used to develop inputs for air quality models, but the most commonly used meteorological grid model by the U.S. EPA and the modeling community is the Weather Research and Forecasting (WRF) model. The WRF model pulls in observations and archived meteorological model data from the region around the modeling site and uses the same physical equations of atmospheric motion and thermodynamics that are used in weather forecasting to model the historical weather conditions at the specific project location. WRF data sets can use high resolution to represent meteorological changes associated with variations in terrain and therefore provide more accurate characterization of atmospheric dispersion in complex environments, which often leads to more realistic modeling results. By incorporating gridded meteorological data and dispersing pollutants as discrete parcels of air, CALPUFF is able to account for the change in directions due to meteorological or terrain variations, making it able to predict concentrations up to 200-300 km whereas AERMOD is limited to no more than 50 km.
CALPUFF Model Run Times Are Much Longer Than AERMOD
This comes at a cost, however - processing meteorological data for CALPUFF modeling is very time- and computer-intensive due to the several processing steps and large data files involved. Similarly, because CALPUFF is a non-steady state puff model that simulates the effects of time- and space-varying conditions on pollution transport, transformation and removal, CALPUFF model runs can be very long and computer-intensive. On the other hand, AERMOD model run times are typically much shorter and the model is more computationally efficient. This, combined with good performance against field measurements and the incorporation of advanced algorithms, is why more and more regulatory agencies across the globe have started promulgating AERMOD for near-field ( <50 km) dispersion modeling assessments.
For most modeling applications, AERMOD is preferred over CALPUFF due to its relative ease of use, widespread regulatory acceptance, speed, and accuracy in the near-field. However, for some specific applications, especially those involving long distances (>50 km), complex chemistry, and certain complex terrain/coastal environment situations, the increased complexity of CALPUFF will yield more accurate results and is well worth the additional time and effort.