3D Photochemical Model Development

Project Updates

Updated March 2026

• Major Findings:
  
  • No findings to report yet.
    
    
• Current and Upcoming Work:
  • The SMOKE emissions model, which prepares emissions data for use in air quality modeling, has been successfully completed. We are currently using existing meteorological data—such as wind, temperature, and atmospheric conditions—as inputs to run the model and support ongoing air quality research and analysis. These efforts help improve our understanding of how emissions and weather together influence air pollution.
    
    
  • We are currently setting up the CMAQ air quality model to run detailed chemical simulations of the atmosphere. This model helps scientists understand how different pollutants form, interact, and move through the air. Running these simulations will support research aimed at improving our understanding of air quality and the processes that lead to pollution events.
    
    
  • The next step in the project is to run the CMAQ air quality model and compare its results with existing measurement data collected in the region. This evaluation will help determine how well the model represents real-world atmospheric conditions and will guide improvements to ensure it accurately simulates air quality and pollution processes.
    
    
• Problems:
  • This work is being conducted by graduate student Loknath Dhar. The project involves learning and implementing complex air quality modeling systems, which requires a significant technical learning process. While the work takes time due to this complexity, steady progress is being made.
    
    
  • We believe that one of the primary challenges in three-dimensional photochemical modeling in the Uinta Basin is accurately simulating meteorological conditions. At present, the most reliable meteorological dataset available for this region is from winter 2013. For this reason, our initial modeling efforts are focused on simulating conditions during the winter of 2013. While this approach has some limitations, it provides an important first step toward developing and improving modeling capabilities for the basin.

Ozone concentrations simulated using the WRF-CHEM 3D photochemical model at Ouray with the CB05 mechanism (WRFCHEM_RACM; red), and observed ozone concentrations (black). From Lyman et al. (2020a)

Aldehydes in ambient air (units of parts per billion of carbon, or ppbC) at the stations listed, simulated by the CAMx 3D photochemical model with the RACM2 versus the CB6r4 chemical mechanism.   From Tran et al. (2023)

 

Change in ozone production rate caused by increasing concentrations of each of the indicated carbonyls by 50%.  Results for four different chemical mechanisms are shown.  MCM is an explicit mechanism thousands of chemical reactions and can be considered the most reliable of the four.

 

Develop a Better 3D Photochemical Model of Winter Ozone

The computer model used to simulate meteorological conditions for most regulatorily-required photochemical modeling is the Weather Research and Forecasting (WRF) model. WRF simulates winter inversion events poorly (Neemann et al., 2015; Tran et al., 2018). Attempts to improve WRF performance have had some success but have not been able to simulate conditions that hold pollutants under the inversion layer as tightly as reality, and this results in an underprediction of winter ozone and other pollutants (Lyman et al., 2020a; Lyman et al., 2024b; Tran and Tran, 2021; Tran et al., 2018). WRF and similar models were not developed with complex mountainous terrain and winter inversion conditions in mind, and the extremely strong inversions experienced during some Uinta Basin winters (Mansfield and Hall, 2018) present a unique modeling challenge. We have shown that assimilating balloon-borne vertical measurements into WRF can improve its ability to simulate inversions (Error! Reference source not found.), but the improvements are inconsistent.

Simulated emissions and chemical reactions also contain large uncertainties and are areas our group has studied in past projects.  No model can ever perfectly represent the conditions that occur across the Uinta Basin during wintertime inversion episodes that lead to high ozone, but we will continue to study ways to make improvements, so the models used by industry, regulators, and others to understand the impacts of emissions-related actions can be more accurate and lead to better decision-making.  

We will use this new photochemical model to investigate how different chemical mechanisms perform in WRF-SMOKE-CMAQ.  A chemical mechanism is a list of all chemicals and reactions used in a chemical model. Thousands of chemical species and well over ten thousand chemical reactions are relevant to winter ozone photochemistry. Explicit simulation of all these reactions and species requires more computational resources than is possible for complex 3D photochemical models, which models are essential for several regulatory and permitting processes in the Clean Air Act, including some related to wintertime ozone.  3D models use simplified chemical mechanisms that group similar compounds and reactions together, to conserve computational resources. Several different simplified mechanisms exist, and all were developed for summer ozone, not winter ozone.

We have conducted several studies to investigate the impact of different chemical mechanisms on the performance of 3D photochemical models (Lyman et al., 2020a; Tran et al., 2015; Tran et al., 2023) and have shown that the choice of mechanism can have a large impact on the amount of simulated winter ozone (Error! Reference source not found.) and ozone precursors (Error! Reference source not found.). Unfortunately, the most comprehensive of these studies, Tran et al. (2023), used meteorological inputs to the model that allowed for unrealistically high mixing of pollution out of the inversion layer. This led to ozone production that was much lower in the model than in reality, which leaves lingering ambiguity as to the importance of the chemical mechanism in winter ozone simulations.

Recently, we have investigated winter ozone chemistry and chemical mechanisms with a box model.  3D photochemical models divide the atmosphere into thousands of cubes that cover the spatial extent of the modeled area. Differential equations for the chemical and physical processes of the atmosphere are solved for each cube, and transport is simulated across adjacent cubes. A box model is essentially the same, except that it contains only a single cube of indeterminate size. Because chemical reactions are computed for just one cube, box models can use explicit chemistry with thousands of chemical species and reactions. Transport and emissions processes are treated coarsely in box models, and meteorological conditions are usually forced to match measured values. 

We used the F0AM box model to simulate a 2019 winter ozone episode we have modeled previously. We compared the Master Chemical Mechanism (Saunders et al., 2003), a detailed mechanism with more than 1,000 species and 12,000 reactions, against simplified mechanisms used in 3D photochemical models, including CB6 (Yarwood et al., 2010), RACM2 (Tran et al., 2023), and SAPRC07 (Carter, 2007; Tran et al., 2015). We forced meteorological inputs to match measured values and kept those the same for all model runs. We followed standard species mapping guidance for converting measured chemical concentrations into model species for each mechanism and forced NO_X and organic compounds to match measured values (for the organic species we measured during the modeled episode). This exercise has shown that some simplified mechanisms are better able than others to replicate the chemical processes and outcomes of the Master Chemical Mechanism. The box model work is complete, and we are currently preparing a manuscript for publication.

Ultimately, reliable 3D photochemical models, not just box models, are needed by researchers, regulators, and industry. We will use results from the box model study to guide work with the WRF-SMOKE-CMAQ 3D photochemical model system to compare the SAPRC07, RACM2, and CB6 mechanisms.  We may also include a new mechanism, the acronym for which is CRACCM.  For the box model, SAPRC07 performed most similarly to the Master Chemical Mechanism.  Our 3D photochemical modeling work will investigate whether SAPRC07 is able to accurately simulate chemical conditions in a Basin-wide winter inversion simulation.