the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Feb 2025
Evaluation of Ozone and its Precursors using the Multi-Scale Infrastructure for Chemistry and Aerosols Version 0 (MUSICAv0) during the Michigan-Ontario Ozone Source Experiment (MOOSE)
Abstract. Surface ozone (O3) in Southeast Michigan (SEMI) often exceeds U.S. National Ambient Air Quality Standards, posing risks to human health and agroecosystems. SEMI, a relatively small region in the state of Michigan, contains most of the state’s anthropogenic emission sources and more than half of the state’s population, and is also prone to long-range and transboundary pollutant transport. Here, we explore the distribution of O3 and its precursors, such as nitrogen oxides (NOx) and volatile organic compounds, over SEMI during the summer of 2021 using the chemistry-climate model, MUSICAv0 (Multi-Scale Infrastructure for Chemistry and Aerosols, Version 0). Using the regional refinement capabilities of MUSICAv0, we created a custom grid over the state of Michigan of 1/16° (~7 km) to better understand the local-scale impacts of chemical and dynamic complexity in SEMI and compared it with a grid with 1/8° (~14 km) resolution over the contiguous United States. Model simulations are evaluated using a comprehensive suite of observations from Phase I of the Michigan-Ontario Ozone Source Experiment (MOOSE) field campaign. MUSICAv0 with higher horizontal grid resolution showed excellent skill in capturing peak O3 concentrations, but showed larger variation in the simulation of O3 precursors (e.g., NOx, HCHO, isoprene). In addition, we implemented a diurnal cycle for anthropogenic nitric oxide (NO) emissions, which is generally not included in global models. As a result, modeled nighttime O3 was improved because of lower NOx concentrations during the night. This work shows that when conceptualizing models in urban regions, it is important to consider a combination of high horizontal resolution and the diurnal cycle of emissions, as they can have important implications for the simulation of secondary air pollutants.
- Preprint
(6755 KB) - Metadata XML
-
Supplement
(1669 KB) - BibTeX
- EndNote
Status: open (until 11 Apr 2025)
-
RC1: 'Comment on egusphere-2025-228', Anonymous Referee #1, 26 Mar 2025
reply
The manuscript analyzes the distribution of ozone and some of its precursors over the Southeast Michigan (SEMI) region during summer 2021 based on model simulations with MUSICAv0 and observations from the MOOSE filed campaign. The authors discuss the impact of grid resolution and diurnal cycle of anthropogenic NO emissions and show that night-time ozone is mostly improved by applying diurnal cycles for NO emissions, while grid resolution is found to have more impact on ozone precursors. The study also shows that using a good conceptualization of grid resolution within MUISCAv0, with finer resolution could lead to more efficient computational costs, which could be beneficial for other local-scale studies including in other regions.
The paper shows the interesting potential of using global models with zooming capabilities like MUSCIAv0 to investigate air pollution characteristics even at specific small regions like SEMI. Overall, the paper is well structured and easy to read. However, the analysis and discussion sections are in some cases rather short and could be further improved in order to better identify the processes controlling summertime ozone in different parts of the SEMI region.
I recommend the manuscript to be accepted for publication after addressing the following comments and suggestions:
Section 2.1.1: Initial conditions are considered from a restart file based on MOZART-TS1. Which initial conditions are considered for the additional species in TS2 not included in TS1?
Section 2.1.3:
- Anthropogenic emissions are considered from CAMS_GLOB_ANTv5.1. A recent study from Soulie et al. (2024, ESSD) shows significant differences in the estimated emissions between CAMS_GLOB_ANTv5.1 and the EPA inventory in USA. In particular, EPA exhibits higher NMVOCs but lower NOx and SO2 emissions. Can the authors comment on the potential impact of such uncertainties in emissions on the model results?
- It is not clear how soil NOx emissions are considered for the simulations.
- The authors include calculated NO emissions from agriculture waste burning (AWB) in Table S1, but it is not clear if emissions from this sector are considered or not. This could lead to double counting of emissions with QFED, although the contribution of NO AWB emissions seems to be minor compared to other sectors.
Section 2.1.4:
- Can the authors comment why only NO diurnal distribution is considered, while diurnal distribution of other species like VOCs or SO2 could also impact the model results?
- Including a figure showing the diurnal distribution of NO emissions from different sectors, as used in the simulation is a useful information.
Section 3.1:
- This section is rather short and doesn’t fully cover the model’s ability to capture meteorological features in the considered region. In addition to the model evaluation, this section is also expected to contain a description of the meteorological situation that characterized the SEMI/MI region during the campaign period. This section can also be significantly improved by considering other meteorological variables (e.g. wind speed/direction), other networks or datasets (e.g. reanalysis).
Section 3.2:
- The authors could elaborate a bit more the discussion on the reasons behind the diurnal changes in the model bias and link with results in Sect. 4. For this, a map showing location of the stations could be very useful.
- The night-time NO2, in particular between 00 and 05 AM, although improved, remain high and the morning peak is less visible when NO diurnal cycle is applied. The authors should discuss the impact of potential uncertainties in the considered diurnal cycle, including the fact that this was applied only for NO.
Section 3.3:
- The authors relate the differences in simulated isoprene (and hence HCHO?) to potential changes in meteorological field leading to changes in calculated BVOCs. Although this could be true, no results (i.e. changes in metorology) are provided to assess this especially in the discussion in Sect. 3.1.
- Similarly, the discrepancies in other species (hydrocarbons and aromatics) is explained by misrepresentation of their anthropogenic sources in the CAMS inventory. The authors can assess such uncertainties in the considered emissions by comparisons with EPA emissions in SEMI.
Section 3.4:
- The discussion on evaluation of modeled HCHO columns contradicts a bit the conclusion in Section 3.3: the authors say there is a combined effect of grid resolution and application NO diurnal cycle on HCHO (Line 429), whereas Sect. 3.3 states no obvious impact of NO on HCHO in the model (Line 383).
- The section could be improved by discussing the link between the location of the stations/sites and the changes in HCHO (e.g. induced impact from isoprene emissions under different Nox-regimes).
- Like for other Sect. 3 subsections, it would be useful to include the location of the monitoring sites and link the results with those discussed in Sect. 4.
Section 3.5:
- Surface maps for winds, temperature and other meteorological parameters could be added to the Supplement to better understand the conditions during the analyzed days and times.
Section 4:
- The significant changes in isoprene emissions from MEGANv2.1 depending on the grid resolution is linked to the induced changes in meterological parameters. This needs to be supported by maps of meterological fields showing these changes.
- The link between Sect 4. and Sect 3. should be strengthened in either or both sections to better understand what drives the changes in the different sites, locations, etc.
- The discussion section is rather short and could/should be improved by strong arguments on e.g. what controls O3 in different parts of the SEMI region and what mitigation strategies could be adopted to reduce the pollution.
Citation: https://doi.org/10.5194/egusphere-2025-228-RC1 -
RC2: 'Comment on egusphere-2025-228', Anonymous Referee #2, 28 Mar 2025
reply
This manuscript presents a good showcase of the use of the next-generation global model MUSICAv0 with regional refinements to study the summertime distribution of ozone and its precursors in the Southeast Michigan region (SEMI) evaluated against the observations from the MOOSE campaign in 2021 and ground-based measurements. As one of the first studies to evaluate MUSICAv0 simulations with an extended campaign, this manuscript would be a notable publication. The study discusses in particular the effect on model performance of using higher grid resolution and implementing a diurnal cycle of anthropogenic NO emissions. It is shown that higher grid resolution is more important for simulating the distribution of O3 precursors than O3 itself, while implementing a diurnal cycle of anthropogenic NO emissions can improve model performance for nighttime O3. This conclusion is in agreement with the other modelling studies.
While this study clearly shows the advantage of using a global model with regional refinements, such as MUSICA, over the conventional global model, the manuscript does not discuss how these new generation models might improve on regional models. Perhaps the authors can add a short discussion on this issue and how these new generation models can be applied to better study regional air quality problems.
I recommend that this manuscript be published with the following comments and suggestions:
Section 1
Line 56: A brief description of the instruments involved in the MOOSE campaign could be included here to provide a more comprehensive introduction to the campaign.
Line 100: Please explicitly mention CAMS-GLOB-ANTv5.1 here, as there are a number of CAMS emission datasets. The resolution of the emission data (0.1 degree, ~10 km) can also be mentioned here to illustrate that the emission resolution is comparable to the model grid resolution. Please add a reference to the CAMS emissions dataset used here.
Line 103: Please add "emissions" after the end of the sentenceSection 2.1.2
- Can the authors explain why the ne30x8 configuration covers the entire CONUS instead of just over Michigan?
Line 135: Can you include a reference to the Community Mesh Generation Toolkit?Section 2.1.3
- A more updated version of the CAMS-GLOB-ANT dataset should be considered in the future. For temporal profiles, the CAMS-GLOB-TEMPO datasets may be useful.Section 2.1.4
- Apart from the diurnal cycle of NOx emissions, the evolution of the PBL probably plays a role in the daytime and nighttime O3 and NOx concentrations. Can the authors comment briefly on this?Section 2.2.1
Table 2: First column, first row: "Selected" VOCs
Section 3.1
- The diurnal cycle of NO emissions probably plays little role in meteorology. Please consider focusing the discussion on the effect of model grid resolution and select specific time periods where the simulations of the two resolutions show significant discrepancy for discussion
Section 3.2
Line 326: Please state in the text that Fig. 4 is a time series of hourly averaged diurnal profiles of ozone concentrations over a specific time period.
Figures 4 and 5: Please try to show time series of O3 and NO2 concentrations from the same stations in the same order for better comparison
Line 358: Can the authors explain in more detail how O3 concentrations are affected by the aforementioned effect on NO2 concentrations?Section 3.3
- Can the authors also discuss whether there are significant differences between daytime and nighttime concentrations between the four simulations?Section 3.4
- The authors should better illustrate how the Pandora measurements can be related to the stations and AML measurements and how these comparisons can lead to the different performance of simulated O3 concentrations.Section 3.5
- Instead of narrative in the text, the authors could include wind vectors or maps of meteorological variables to illustrate how the different models capture the NO2 plumes at the different times shown, as the readers may not be familiar with the geographical locations shown.Section 4
Figures 13 and 14: The authors should explain why they show the conservatively regridded model outputs in panel (c).
Line 512: The authors should explain the consequences of the regridding method not being able to reproduce the higher resolution simulation results
Figure 15: Please consider also including the observed concentrations of O3, NO and NO2 in the time series to better illustrate which model configuration is closer to the observed values.Section 5
Line 618: In addition to the South Korean simulation, can the authors compare their work with other studies using MUSICA over other regions of the USA?
- The authors should consider briefly discussing how MUSICA or similar next-generation global models with regional refinement capability can be used to formulate regional/local air pollution monitoring and mitigation strategies.Citation: https://doi.org/10.5194/egusphere-2025-228-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 118 | 21 | 4 | 143 | 8 | 6 | 5 |
- HTML: 118
- PDF: 21
- XML: 4
- Total: 143
- Supplement: 8
- BibTeX: 6
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 82 | 55 |
| China | 2 | 13 | 8 |
| Germany | 3 | 11 | 7 |
| Canada | 4 | 6 | 4 |
| France | 5 | 6 | 4 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 82