New Mexico Environmental Department Ozone Attainment Initiative (NMED-OAI) Modeling Results

This Webpage contains access to the New Mexico Environmental Department (NMED) Ozone Attainment Initiative (OAI) Modeling Study modeling results. A separate NMED OAI Modeling Study Webpage contains reports and results of the Webinars that occurred during the NMED OAI Modeling Study. More details on the NMED OAI Modeling Study modeling results are contained in the links below. The modeling result displays are made available through an Image Browser with the data behind some of the displays and Excel Dashboard visualization tools also available through a Download Browser.

Introduction

The NMED Air Quality Bureau has authority over air quality management activities throughout the state of New Mexico, with the exception Bernalillo County and Tribal Lands. The New Mexico Air Quality Control Act (NMAQCA) requires the NMED to develop a plan to address elevated ozone levels when they exceed 95% of the ozone NAAQS level (74-3-5.3, NMSA 1978). To address the high observed ozone concentrations in New Mexico, the NMED has embarked on an Ozone Attainment Initiative (OAI) to protect the ozone attainment status of the state and ensure health and welfare of the residents of the state for future generations. The OAI was initiated in Spring 2018. As part of the OAI, NMED released a Request for Proposal (RFP#20 667 4040 0001) for the NM OAI Modeling Study and the NM OAI Modeling Study photochemical grid model (PGM) modeling was awarded to a contracting team of WESTAR and Ramboll.

New Mexico Ozone Air Quality

The 2008 ozone National Ambient Air Quality Standard (NAAQS) has a threshold of 0.075 ppm (75 ppb) with the relevant metric being the ozone Design Value (DV) that is expressed as the three-year average of the fourth highest Maximum Daily Average 8-hour (MDA8) ozone concentrations. The ozone NAAQS was revised in 2015 with a threshold of 0.070 ppm (70 ppb). Table 1 lists the observed ozone DVs at New Mexico monitoring sites from 2010 to 2019. The DVs are color-coded when they exceed the 75 ppb 2008 ozone NAAQS (red), 70 ppb 2015 ozone NAAQS (yellow) and within 95% of the 70 ppb ozone NAAQS (green, i.e., ≥ 67 ppb). In more recent years, the ozone DVs have exceeded the 70 ppb 2015 ozone NAAQS at several sites in southern New Mexico (e.g., in Doña Ana and Eddy Counties). In particular, the 2017-2019 ozone DVs are particularly high and even exceeded the 75 ppb 2008 ozone NAAQS at three sites in southern New Mexico (colored red in Table 1).

Overview of NMED OAI Modeling Study

The NMED OAI Modeling Study leverages the 2014 PGM modeling platform developed by the Western Regional Air Partnership (WRAP) and the Western Air Quality Study (WAQS) and enhanced it by adding a 4-km grid resolution modeling domain covering New Mexico and adjacent regions. The NMED OAI Modeling Study Air Quality Technical Support Document (AQTSD) documents the 2014 base case, 2028 future year base case and 2028 NM O&G control strategy modeling and includes a summary of the development of the 2014 PGM modeling platform with more details found on the NMED OAI Study Website that includes Webinars presented to the NMED during the course of the study and the following reports:

• NMED-OAI Photochemical Modeling Study – Draft Modeling Protocol (Ramboll and WESTAR, 2020a).
• NMED-OAI Photochemical Modeling Study – Draft Work Plan (Ramboll and WESTAR, 2020b).
• NMED-OAI Photochemical Modeling Study – 2014 Modeling Platform Development and Model Evaluation (Ramboll and WESTAR, 2020c).
• NMED-OAI Photochemical Modeling Study – Revised 2014v2 Base Case and Model Performance Evaluation (Ramboll and WESTAR, 2021a).
• NMED-OAI Photochemical Modeling Study – Air Quality Technical Support Document (Ramboll and WESTAR, 2021b).

Although the WESTAR/Ramboll team conducted the 2014 base year and 2028 base and O&G Control Strategy future year ozone modeling, the implementation of the proposed NMED O&G ozone precursor control strategy in the 2028 New Mexico oil and gas (O&G) emissions was conducted by Eastern Research Group (ERG) under separate contract with NMED. ERG documented the controls implemented on the 2028 NM O&G emissions in a Memorandum and Excel Spreadsheet that are available on the NMED OAI Webpage.

Development of the CAMx 2014 36/12/4-km Modeling Platform

The NMED OAI Modeling Study used the Comprehensive Air Quality Model with extensions (CAMx) photochemical grid model to conduct 2014 current and 2028 future year modeling. The NM OAI Modeling Study CAMx 2014 36/12/4-km modeling platform was developed by enhancing the WRAP-WAQS 2014v2 36/12-km PGM modeling platform with the addition of a new 4-km grid resolution domain covering New Mexico and surrounding areas, especially the O&G production regions in the Permian and San Juan Basins. The WRAP-WAQS conducted CAMx 2014v1 and 2014v2 36/12-km base case modeling and model performance evaluation. Figure 1 displays the 36/12/4-km resolution domain nested grid structure used in the NMED OAI Modeling Study CAMx ozone modeling. The 36/12-km domains were defined to be identical to WRAP-WAQS CAMx 2014 36/12-km platform so that the WRAP-WAQS 2014 and 2028 CAMx-ready emissions for the 36/12-km domains could be used as is in the NMED OAI Modeling Study. The NMED OAI Modeling Study conducted their own 2014 WRF meteorological modeling for the 36/12/4-kim domains (Figure 1) and 2014 SMOKE emissions modeling for the 4-km domain. The 2014 emissions were based on the 2014 National Emissions Inventory (2014NEI) with updates from western states. Boundary Condition (BC) inputs were based on a WRAP-WAQS 2014 GEOS-Chem model simulation.

The NMED OAI Modeling Study conducted a model performance evaluation (MPE) for an initial CAMx May-August 2014 base case simulation (Ramboll and WESTAR, 2020c) and a revised 2014v2 base case simulation (Ramboll and WESTAR, 2021a). Additional MPE displays are available for the 2014v2 base case in the links below in the Modeling Results section of this Webpage.

The NMED OAI Modeling Study 2028 emissions were based on the WRAP-WAQS 2028OTBa2 emissions scenario with updated New Mexico 2028 oil and gas (O&G) emissions and using 2014 actual fires instead of RepBase fires. CAMx modeling was conducted for a 2028 Base Case and a 2028 O&G Control Strategy. The WESTAR/Ramboll Team provided the New Mexico 2028 O&G base case emissions to the NMED. The Eastern Research Group, under separate contract with NMED, implemented the emission controls as the 2028 NM O&G Control Strategy emissions inventory (Excel workbook from the separate external NMED 20.2.50 NMAC Oil and Gas Sector-Ozone Precursor Pollutants Rulemaking (EIB No. 21-27 (R)) hearing web page) based on their interpretation of the proposed State of New Mexico “Ozone Precursor Rule for Oil and Natural Gas Sector.” Those reductions were evaluated in the 2028 O&G Control Strategy modeling scenario which reduced O&G emissions in 7 counties in New Mexico as described in the NMED OAI Modeling Study AQTSD (Ramboll and WESTAR, 2021b).

Two CAMx 2028 O&G Control Strategy ozone source apportionment simulations were conducted as described in Chapters 10 and 11 of the AQTSD (Ramboll and WESTAR, 2021b) with additional displays provided in the links below.

Summary of Modeling Results from the NMED-OAI Modeing Study

Below we present links for downloading more extensive modeling results from the NMED OAI Modeling Study then are presented in the Webinars, 2014 base case modeling report and addendum and the AQTSD for three components of the study:

1. CAMx revised 2014v2 base case modeling model performance evaluation (MPE).
2. CAMx 2028 Source Sector APCA ozone source apportionment modeling.
3. CAMx 2028 VOC and NOx Sensitivity OSAT ozone source apportionment modeling.

This Webpage provides the user to many more displays and analysis tools of the results from the NMED OAI Modeling Study than available in the reports, but is not intended to fully document the technical approach that has details in the NMED OAI Modeling Study reports cited above.

CAMx Revised 2014v2 Base Case Model Performance Evaluation

The CAMx revised 2014v2 base case MPE was presented in an Addendum MPE report (Ramboll and WESTAR, 2021a) and Chapter 6 of the AQTSD (Ramboll and WESTAR, 2021b) for the NMED OAI Modeling Study. EPA’s Atmospheric Model Evaluation Tool (AMET) was used to evaluate the CAMx 2014v2 base case for ozone performance in the 4-km New Mexico modeling domain. AMET generated the following graphical and statistical displays of ozone MPE that are contained in the “AMET_v14” folder of the Image Browser:

• Scatter Plots
• Soccer Plots
• Stats Plots
• Time Series

AMET generated over 300 displays of ozone MPE of which only a small fraction were presented in the MPE Addendum and AQTSD. The links below provides access to the full suite of AMET MPE graphical displays products in the Image Browser with the data behind the graphical displays in the Download Browser. The Plots link contains images of modeling data and results, and the Data Files link contains plots of modeling data and source apportionment results.

Scatter Plots

The “scatterplot” folder contains two subfolders that contain scatter plots of predicted and observed MDA8 ozone concentrations for sites within the entire 4-km New Mexico domain (“4km_domain”) and sites just within New Mexico (“NM”). Scatter plots include model performance statistical metrics as follows:

• Mean Observed (Mn_O)
• Mean Modeled (Mn_M)
• Normalized Mean Bias (NMB)
• Normalized Mean Error (NME)

Emery and co-workers (Emery et al., 2016) have developed model Performance Goals and Criteria for MDA8 ozone NMB and NME as follows:

MDA8 Ozone	NMB	NME
Performance Goal	≤±5%	≤±15%
Performance Criteria	≤±15%	≤25%td>

The MDA8 ozone scatter plots and statistical performance measures are available with and without using a 60 ppm observed ozone cutoff. When a 60 ppb observed ozone cutoff is used, the scatter plot and model performance evaluation metrics are only presented for predicted and observed MDA8 ozone pairs (matched by time and location) for which the observed MDA8 ozone is above 60 ppb. This allows the evaluation for high observed ozone concentrations that are close to the NAAQS.

Figures 2 and 3 display the CAMx 2014v2 base case scatter plots and performance metrics for MDA8 ozone at sites located in New Mexico and the May 15 through August 31, 2014 modeling period without and with using a 60 ppb observed ozone cutoff concentration. Without using an ozone cutoff, the MDA8 ozone NMB is 2.9% and the NME is 9.7% both of which achieve the ozone Performance Goals (Figure 2). With using a 60 ppb ozone cutoff, the NMB is -5.8% that barely fails to achieve the bias Performance Goal (≤±5%) but achieves the Performance Criteria (≤±15%). The NME using the cutoff (9.1%) achieves the error Performance Goal.

Soccer Plots

Soccer plots plot NMB (x-axis) versus NME (y-axis) along with boxes corresponding to the Performance Goals and Criteria so it is easy to interpret when model performance achieves the goals and criteria. AMET was run to generate a soccer plot of monthly NME/NMB for MDA8 ozone across all sites in New Mexico as shown in Figure 4.

Stats Plots

Spatial maps of site specific performance statistics are generated for the following performance statistics:

• Corr – Correlation Coefficient (r)
• FB – Fractional Bias (%)
• FE – Fraction Error (%)
• MB – Mean Bias (ppb)
• ME – Mean Error (ppb)
• NMB – Normalized Mean Bias (%)
• NME – Normalized Mean Error (%)
• RMSE – Root Mean Squared Error (ppb)

The “statsplots” folder includes three subfolders that contain spatial maps of site-specific performance statistics for different regions: “4km_domain”, “NM” and “NM_regions”. The “NM_regions” folder has separate spatial maps for New Mexico sites within “Bernalillo_County” and sites north (“North_NM_sites”) and south (”South_NM_sites”) of Bernalillo County.

Figure 5 displays an example Stats Plots for NMB using no cutoff across all sites in the 4-km New Mexico domain. Stats Plots are also available using a 60 ppb cutoff and for just sites in New Mexico. The shading cut-points use the grey color when the NMB falls within the ozone bias Performance Goal (≤±5%) with the next color cut-point (bright yellow or green) indicating the NMB falls between the Performance Goal and Criterion. All sites but one in El Paso achieve the bias Performance Criterion and most sites in New Mexico achieve the Performance Goal.

The data behind the Stats Plots are also available in .csv files from the Download Browser.

Time Series

Time series plot the daily predicted and observed MDA8 ozone concentrations from May 15 to August 31, 2014 at New Mexico monitoring sites and their differences (i.e., daily bias). AMET was run to display time series for the original CAMx 2014 base case (2014Base) and final CAMx 2014v2 base case simulation. Figure 6 displays an example MDA8 ozone time series plot for Desert View in Doña Ana County with the upper panel showing the observed (red) and CAMx predicted 2014Base (green) and 2014v2 (blue) MDA8 ozone concentrations and the bottom panel showing the daily bias. The monitoring sites are identified by their AQS identification number where one of the previous cited NMED OAI Modeling Study reports has cross-references with common names and county locations.

Spatial Maps of 2014 and 2028 Daily MDA8 Ozone and Ozone Design Values

The “2028_2014” folder in the Image Browser contains spatial maps of MDA8 ozone and ozone design values. Spatial maps of MDA8 ozone concentrations and their differences across the 4-km New Mexico domain for the 2014v2, 2028 Base Case and 2028 O&G Control Strategy scenarios were generated for May 15 to September 1, 2014. Spatial maps of daily MDA8 ozone concentrations and their differences for six of these days were presented in Chapters 7 and 8 of the AQTSD. They are available for all days in the “dailyMDA8” and “dailydiffMDA8” subfolders under “2028_2014” folder in the Image Browser. Figures 7 and 8 show examples daily MDA8 ozone spatial maps for June 5, 2014 and the, respectively, 2028 Base Case and 2028 O&G Control Strategy scenarios.

The CAMx 2014v2, 2028 Base Case and 2028 O&G Control Strategy modeling results were used with EPA’s Software for Modeled Attainment Test (SMAT) tool to generate spatial maps of current year (DVC) and future year (DVF) ozone Design Values and their differences. The SMAT Unmonitored Area Analysis (UAA) interpolates monitored ozone DVC across the 4-km New Mexico modeling domain and uses the CAMx results to generate spatial fields of projected 2028 ozone DVF under the 2028 Base and 2028 O&G Control Strategy. The folders “SMAT_UAA_2014Base_2028Base” and “SMAT_UAA_2028Base_2028Control” contain the spatial maps of 2028 ozone DVF and their differences following EPA guidance for the current year DVC that is based on observe date from 2012-2016 and are described in Chapters 7 and 8 of the AQTSD. Chapter 9 of the AQTSD describes results of the 2028 ozone DVF projections for the 2028 O&G Control Strategy that analyzes their sensitivity to the current year DVC by using DVCs based on observed ozone from 2015-2019 and 2017-2019 whose displays can be found in the “SMAT_DVC_2015_2019” folder.

CAMx 2028 Source Sector APCA Ozone Source Apportionment Modeling

The CAMx Anthropogenic Precursor Culpability Assessment (APCA) ozone source apportionment tool was used to obtain the contributions of Source Sectors from different regions to ozone concentrations under the 2028 Oil and Gas (O&G) Control Strategy emissions scenario. Details of the CAMx 2028 Source Sector APCA ozone source apportionment simulations are contained in Chapter 10 of the AQTSD.

The CAMx 2028 Source Sector APCA ozone source apportionment results were post-processed to generate spatial maps of ozone contributions from Source Sectors and Source Regions and the results were extracted at the monitoring sites to visualize ozone contributions. 

Design of the CAMx 2028 Source Sector APCA Ozone Source Apportionment Modeling

Separate ozone contributions were obtained from five source regions in the 2028 Source Sector APCA ozone source apportionment simulation as follows (Figure 9):

1. New Mexico;
2. Texas;
3. Colorado;
4. Remainder U.S. states plus coastal water out to 200 nautical miles (nmi) from the U.S. coast; and
5. International, which included Mexico, Canada and waters off their coasts plus waters more than 200 nmi off the U.S. coast.

1. The emissions inventory was divided up into 9 Source Categories in the 2028 Source Sector APCA ozone source apportionment modeling:
2. 1. Natural (biogenic, lightning NOx, windblown dust and oceanic [sea salt and dimethyl sulfide]);
3. 2. Fires (U.S. wildfires, wildland prescribed burns and agricultural burning and other [Mexico/Canada] fires);
4. 3. Oil and gas point sources (surrogate for midstream O&G);
5. 4. Oil and gas non-point sources (surrogate for upstream O&G);
6. 5. EGU point;
7. 6. Non-EGU point;
8. 7. On-road mobile;
9. 8. Non-road mobile; and
10. 9. Remainder anthropogenic.

Separate ozone contributions were also obtained for five Initial Concentrations (IC) and Boundary Condition (BC) Source Groups using results from the WRAP-WAQS 2014 GEOS-Chem global chemistry model simulations:

• BC_Intl -- contributions of international anthropogenic emissions
• BC_US -- contributions of U.S. anthropogenic emissions
• BC_Natural -- contributions of natural sources
• BC_Top -- contributions due to concentrations above the top of the modeling domain (~19-km or 50 mb).
• IC -- Initial Concentrations

With 5 Source Regions and 9 Source Categories plus 5 IC/BC Source Groups, separate ozone contributions were obtained from 50 different Source Groups (50 = 5 x 9 + 5).

CAMx 2028 Source Sector APCA Spatial Map Output

The CAMx 2028 Source Sector APCA source apportionment maps are contained in the “APCA_Source-Sector” folder in the Image Browser.

There are three subfolders in the “APCA_Source-Sector” folder:

• “Contribution_Plots”
• “Daily_plots”
• “SMAT_UAA_plots”

APCA Contribution Plots

The contribution plots display the highest (Max) contributions to MDA8 ozone for each of the 9 Source Categories from each of the 5 Source Regions, plus a combined “All_US” region. For the Mexico (Intl) Source Region, the Source Categories are not split out and all the anthropogenic emissions are in the “otherAnthro” Source Category. There are also highest contributions to MDA8 ozone from the IC and BC Source Groups in the “BCIC” folder.

Figure 10 shows an example Contribution Plot of highest MDA8 ozone contributions from non-point O&G sources from New Mexico (NM).

APCA Daily Plots

The daily plots display the ozone contributions of the 9 Source Categories from emissions in New Mexico (“NM”) and all of the U.S. (“All_US”) as well as IC and BC (“BCIC”) for six example days. Figure 11 shows an example of the New Mexico non-point O&G Source Category MDA8 ozone contribution for the 2028 O&G Control Strategy Source Sector APCA ozone source apportionment simulation on June 5, 2014.

APCA SMAT UAA Plots

The “SMAT_UAA_plots” folder has “NM_Sources” and No_International” folders that contains spatial maps of SMAT UAA 2028 ozone DVFs with contributions due to emissions from, respectively, New Mexico Source Categories and International sources removed. 2028 ozone DVF displays were generated using the 2012-2016 DVC as recommended in EPA’s modeling guidance, and for the 2028 DVF projection DVC sensitivity tests that used ozone DVCs based on 2015-2019 and 2017-2019 observed ozone data.

2028 Source Sector APCA Dashboard

The ozone source apportionment results for the CAMx 2028 O&G Control Strategy Source Sector APCA run were extracted at the New Mexico monitoring sites and loaded into an Excel Dashboard tool for visualization. The APCA Excel Dashboard visualization tool is available from the Download Browser. 

Figure 12 displays a screen shot of the APCA Excel Dashboard that has the following control features (also described in the README sheet in the Dashboard):

• The top left ‘Site_ID’ control panel is for selecting the monitoring site with a map provided just below.
• The 10 days selected for visualization is control by the middle ‘Top_ozone’ and ‘SMAT’ control panels. To displays ozone results for the 10 days used in SMAT to make ozone DVF projections, select 1 and deselect 0 in ‘SMAT” and 1-4 in ‘Top_ozone.’ To select top 10 modeled 2028 MDA* ozone days select 0 and 1 in ‘SMAT’ and 1 in ‘Top_ozone.’ Second top 10 (11-20) is done similarly only select 2 in ‘Top_ozone,’
• The pie chart and stacked bar chart on the right is are displayed for the region selected in the ‘control for Sources’ control pan in the top right of the Dashboard.

More details on the Source Sector APCA ozone source apportionment Dashboard are contained in Chapter 10 of the AQTSD.

CAMx 2028 VOC and NO_x Sensitivity OSAT Ozone Source Apportionment Modeling

The CAMx 2028 VOC and NOx Sensitivity OSAT ozone source apportionment simulations was performed to investigate the degree to which ozone formation in New Mexico is more sensitive to VOC or NOx emissions. When the OSAT version of the CAMx ozone source apportionment tool is used, the relative amounts of the OSAT O3V and O3N tracers provide an indication of whether ozone formation is more sensitive to VOC or NOx emissions, respectively.

Design of the 2028 VOC/NOx Sensitive OSAT Ozone Source Apportionment

The 2028 VOC/NOx Sensitivity OSAT simulations used the same 5 Source Regions as the 2028 Source Sector APCA ozone source apportionment simulations (i.e., NM, TX, CO, Remainder US and International, see Figure 9). Source apportionment was also obtained for the following Source Categories:

• U.S. Anthropogenic;
• International Anthropogenic; and
• Natural.

Daily Spatial Maps of 2028 VOC/NOx Sensitivity OSAT Ozone Source Apportionment

Information about ozone formation sensitivity to VOC and NOx is only available for the 15 emissions based Source Groups defined by the 5 Source Regions and 3 Source Categories. Ozone from the IC and BC Source Groups do not contain any information on ozone formation sensitivity since they were defined from output of the GEOS-Chem global chemistry model. A Percent NOx Sensitivity (%NOxSens) metric was developed and used to display spatial maps of the percent ozone formed under NOx sensitive conditions [i.e., %NOxSens = 100 x O3N /(O3V + O3N)]. The %NOxSens metric can be displayed for any combination of Source Region and Source Categories, although we focused on all emissions and anthropogenic emissions Source Categories. In order to interpret the %NOxSens metric, it is important to also know the amount of ozone associated with it to understand whether it is important or not.

The “OSAT_VOC-NOx-Sensitivity” folder in the Image Browser contains 7 subfolders with each folder containing 110 spatial maps of either daily MDA8 ozone or %NOxSens for the May 15 – September 1, 2014 period and the 2028 O&G Control Strategy emissions scenario as follows:

• “dailyMDA8.tot_ozone” folder contains spatial maps of the total MDA8 ozone concentrations for that day that includes contributions from all Source Groups including IC and BC. Figure 13 shows an example for June 5, 2014.
• “dailyMDA8.all_emis” contains maps of MDA8 ozone due to all the emission Source Groups (i.e., no IC and BC). Figure 14 shows an example for June 5^th.
• “dailyMDA8.all.emis.perc” contains maps of the %NOxSens metric from all emission based Source Groups showing the percent ozone formed under more NOx sensitive conditions with an example for June 5^th given in Figure 15.
• “dailyMDA8.all_anthro” includes maps of MDA8 ozone due to all anthropogenic emissions Source Groups (i.e., no IC and BC and fire and natural emission ozone contributions).
• “dailyMDA8.all_anthro_perc” are maps of %NOxSens metric for all anthropogenic emissions Source Groups.
• dailyMDA8.NM_anthro” contains maps of MDA8 ozone due to the New Mexico anthropogenic emissions Source Group with an example for June 5^th shown in Figure 16.
• “dailyMDA8.NM_anthro_perc” are maps of %NOxSens metric for New Mexico anthropogenic emissions Source Group with an example for June 5, 2014 shown in Figure 17.

Across most of New Mexico, the MDA8 ozone is more NOx sensitive than VOC sensitive as indicated by the yellow-orange-red colors in Figure 15 (i.e., %NOxSens > 85%). For anthropogenic emissions from New Mexico the percent NOx Sensitive Ozone is even higher over most of New Mexico on June 5^th (i.e., %NOxSens > 95%). However, there are four areas in New Mexico that have lower %NOxSens on June 5^th: San Juan Basin (San Juan County), Southern New Mexico neighboring El Paso, Albuquerque and Permian Basin (Figure 15). San Juan County and Southern New Mexico neighboring El Paso are also the areas in New Mexico with the highest MDA8 ozone concentrations on June 5^th that are approaching or even exceeding the 70 ppb ozone NAAQS (Figure 13). The %NOxSens metric in the high MDA8 ozone areas in San Juan County and El Paso are light blue indicating that VOC and NOx sensitive ozone are approximately equally important. This is also true for New Mexico anthropogenic emissions and ozone in San Juan County, but not for the ozone in El Paso due to New Mexico anthropogenic emissions (Figure 17).

2028 VOC/NOx Sensitivity OSAT Dashboard

The CAMx 2028 VOC/NOx Sensitivity OSAT ozone source apportionment results at New Mexico monitoring sites were loaded in an Excel Dashboard for visualization. Figure 18 displays a screenshot of the OSAT Dashboard. It displays stacked bar charts for 10 days of MDA8 ozone and pie charts of the average of the 10 days showing NOx sensitive ozone using warmer colors (e.g., red) and VOC sensitive ozone using cooler colors (e.g., blue). The user selects a site (‘Station’), the 10 days to visualize (‘SMAT’ and ‘Top_Grps’) and which Source Category (‘Source Groups’) and Source Region (‘Region’) to display the bar/pie charts for. Section 11.3 of the AQTSD discusses the OSAT Dashboard and provides several examples.