The ozone climate penalty, NAAQS attainment, and health equity along the Colorado Front Range

Ground-level ozone, a respiratory irritant and key component of smog, is linked to acute and chronic adverse health effects even below current regulatory standards. Ozone is produced via photochemical reactions involving VOCs and NOx, which are enhanced by heat and specific meteorological conditions. Although EPA tightened the ozone NAAQS in 2008 (75 ppb) and 2015 (70 ppb), the DMNFR region remains in nonattainment, with complex timing of standards meaning older, looser standards can still drive policy. Colorado’s Front Range has experienced persistent summertime ozone exceedances, influenced by local precursor emissions (e.g., oil and gas, traffic), regional and intercontinental transport, and meteorology (high temperatures, stagnation, sunlight). Because weather affects ozone formation, climate change can exacerbate ozone pollution, termed the "climate penalty" (often estimated at 1–3 ppb per °C). This study asks: How large is the historical ozone climate penalty (1950s to 2010s) along the Colorado Front Range, how does it affect NAAQS attainment timelines, and which populations bear the greatest burden?

Prior work attributes elevated Front Range ozone to both local and nonlocal sources. Local precursor emissions, particularly from oil and gas activities, can contribute up to ~30 ppb (Cheadle et al.). Regional and long-range influences include wildfire smoke transporting ozone precursors (Lindaas et al.), interstate transport from southern California (Huang et al.), and evolving trans-Pacific influences (Parrish et al.). Stratospheric intrusions affect springtime ozone in the western US, especially after strong La Niña winters (Lin et al.), but are less significant in summer. Meteorologically, high temperatures, low winds, and high insolation favor ozone formation in the Denver region (Reddy; Helmig). The concept of a climate penalty—higher ozone with warmer temperatures—has been quantified with factors of roughly 1–3 ppb per °C (Bloomer; NCA4/Nolte et al.; Rasmussen et al.). However, most studies emphasize future projections rather than quantifying a historical, locally resolved climate penalty relevant to regulatory attainment and health equity. This study addresses that gap for Colorado’s Front Range.

Design: Quantify a historical ozone climate penalty by comparing observed 2010s ozone (May–October) to a counterfactual ozone field representing the 2010s with daily meteorology adjusted to reflect 1950s climate, holding other factors constant. Modeling approach: Spatio-temporal land-use regression (LUR) using generalized additive mixed models (GAMMs) in R (mgcv). Meteorological variables: daily maximum temperature, mean wind speed (log-transformed), relative humidity (probit-transformed), and sea level pressure. Data sources: Weather from NOAA NCEI monitors (1950s and 2010s; 21 sites with varying temporal coverage). Ozone: EPA AQS daily 8-h maximum ozone at 26 sites (2010s; May–October). Additional spatial covariates: elevation (elevatr), population density (SEDAC GPWv4, 2015, 30 arc-second), traffic density (USDOT HPMS 2011–2017). Prediction grid: 1/48° longitude by 1/36° latitude, plus ozone monitor locations and 1003 census tract centroids. Steps: 1) For each meteorological variable, fit a GAMM across 1950s and 2010s with fixed effects capturing spatially indexed variables and smooth seasonal trends (up to 6 df/year; possibly interacting with spatial terms) and random effects by date. Fixed effects characterize decade-specific climate; full model captures daily weather. Model selection via leave-one-monitor-out cross-validation. 2) Predict daily meteorology across the grid and at ozone monitors. 3) Construct counterfactual 1950s weather fields by adjusting observed 2010s daily weather with the decade differences in climate fields at each location/day. 4) Fit ozone LUR (2010s) to daily 8-h max ozone at monitors with fixed effects from weather fields (including nonlinearities), latitude/longitude, elevation, log10 population density, log10 inverse-distance-weighted traffic density, categorical year, smooth spatial trend (2 df), smooth date trend (6 df/year), and interaction between location and day-of-year; random effects include date-specific coefficients for intercept, latitude, longitude, and elevation to capture non-meteorological time-varying influences (e.g., NOx/VOC). Final ozone model selected via leave-one-monitor-out cross-validation included a five-way interaction among the four weather variables and year, plus additive population density and traffic terms. 5) Predict observed 2010s daily ozone fields at the grid, monitors, and census tract centroids. 6) Input counterfactual 1950s weather fields into the ozone LUR to generate counterfactual 1950s ozone fields. 7) Define ozone climate penalty as the difference: observed 2010s ozone minus counterfactual 1950s ozone, per day and location. NAAQS analysis: Mapped average annual exceedances of the 2015 standard (70 ppb) for June–August under observed vs counterfactual climates. Estimated counterfactual monitor observations for the 14 attainment monitors by adjusting measured 2010s values with local penalty estimates, then computed 3-year design values for both observed and counterfactual series. Extrapolated linear trends to estimate attainment years for the 2008 (75 ppb) and 2015 (70 ppb) standards, assuming continued historical emission reductions and no further warming. Health equity analysis: Estimated tract-level penalties (population-weighted across grid cells when available; otherwise at tract centroid). Regressed penalties against 2018 ACS sociodemographics and Colorado Department of Public Health and Environment health burden indicators: race/ethnicity, foreign birth, child poverty status, asthma, diabetes, heart disease, overweight/obesity, healthcare coverage, and self-reported health status. Included indicator for overlap with the DMNFR nonattainment area. Tested univariate models and models adjusting for urbanicity (log10 tract area or log10 population density). Wyoming tracts (n=7) included for demographics but excluded from health burden analyses due to data source limitations.

• Meteorological shifts (1950s→2010s): Maximum daily temperature increased by ~2.0 °C (3.6 °F) during warm months; mean wind speed decreased by ~0.5–1.5 knots (largest at higher elevations); relative humidity and sea level pressure showed minimal average changes. - Ozone climate penalty magnitude and timing: For daily max 8-h ozone, the penalty averaged ~0.5 ppb before mid-July, rising to ~1.0 ppb after day 200 (late July through August). Over June–August, low-elevation urban areas (Boulder, Fort Collins, Greeley, Denver metro within the DMNFR nonattainment area) experienced average penalties around 0.75 ppb per day. - Spatial patterns: Observed/predicted ozone was lower adjacent to roadways (titration by NO) and higher in suburban belts and the less-populated southern domain (not constrained by monitors). The penalty was highest in major urban centers at the foot of the Front Range. - NAAQS exceedances: Along the I-25 urban corridor (Colorado Springs to Fort Collins), the observed 2010s climate yielded roughly 15 exceedance days per year of the 2015 standard (70 ppb). Under the counterfactual 1950s climate, these urban centers would have experienced approximately 3 fewer exceedances per year (June–August). For the 2008 standard (75 ppb), roughly one additional exceedance per year was attributable to the climate penalty (June–August). - Design values and attainment: The climate penalty increased DMNFR design values by about 1 ppb relative to the counterfactual 1950s climate. Extrapolated attainment timing was delayed by ~2 years: 2008 standard (75 ppb) from 2023 (counterfactual) to 2025 (observed), and 2015 standard (70 ppb) from 2033 to 2035. - Health equity associations: Tract-level penalties were positively associated with % Hispanic/Latino, % American Indian/Native Alaskan, % children at 100–200% of the federal poverty level, % with asthma, % with diabetes, and % reporting fair or poor health. % with healthcare coverage was negatively associated (i.e., higher penalties where coverage is lower). Tracts overlapping the DMNFR nonattainment area had higher penalties. Associations persisted after adjusting for urbanicity (by tract area or population density).

This study demonstrates that historical climate change has already increased summertime ozone along Colorado’s Front Range, consistent with prior estimates of 1–3 ppb ozone increase per °C. The identified climate penalty—intensifying later in the summer and concentrated in urbanized low-elevation areas—elevates ozone design values and increases exceedances, thereby complicating and delaying attainment of both the 2008 and 2015 ozone NAAQS. The findings indicate that even modest increases in long-term ozone can have meaningful health impacts, particularly for respiratory outcomes in children, and that the burden is not evenly distributed: communities with higher proportions of Hispanic/Latino and American Indian/Native Alaskan residents, greater child poverty, higher asthma and diabetes prevalence, worse self-reported health, and lower insurance coverage experience larger climate-related ozone increases. These patterns align with existing environmental justice concerns, as many frontline communities are situated near emission sources of ozone precursors. The results highlight the dual benefit of reducing precursor emissions: directly lowering ozone concentrations and indirectly mitigating future warming that drives additional ozone formation. Without substantial intervention, continued warming is likely to exacerbate ozone control challenges and health inequities.

By constructing a counterfactual 1950s-climate scenario using spatio-temporal LURs, this study quantifies a historical ozone climate penalty of roughly 0.5–1.0 ppb in the Colorado Front Range, with the largest impacts in urban centers later in the summer. The penalty elevates design values by ~1 ppb and delays projected attainment of the 2008 and 2015 ozone standards by about two years, underscoring the regulatory implications of climate change for air quality management. The disproportionate associations with sociodemographic and health burden indicators show that climate-driven ozone increases exacerbate health inequities, implying that climate mitigation and targeted precursor emission reductions could yield substantial benefits for historically disenfranchised and frontline communities. Future research should quantify the full health impacts (morbidity and mortality) attributable to the ozone climate penalty, refine local-scale exposure estimates in areas lacking monitors, and evaluate policy scenarios that jointly target greenhouse gases and ozone precursors to accelerate attainment and reduce inequities.

• Spatial smoothness assumption in LURs may miss local-scale processes affecting ozone, potentially underestimating localized exceedances. - The approach is empirical and does not explicitly model underlying physical/chemical mechanisms; if key processes occur primarily in areas without monitors, predictions there may be less accurate. - Some regions of the domain lack ozone monitoring (e.g., southern and western extents), limiting model constraint; however, these areas are sparsely populated and do not affect attainment determinations. - Attainment-year extrapolations assume continued historical rates of precursor emission reductions and no further warming, which may not hold. - Health burden data were unavailable for Wyoming tracts, limiting those analyses to Colorado.