Air quality impacts from oil and natural gas development in Colorado

The development of hydraulic fracturing techniques has made it profitable to extract petroleum hydrocarbons from geologic shale formations. The application of this technology has caused a surge in new oil and natural gas (O&NG) drilling in shale basins across the United States, including in Colorado, where most of the activity has been in the Denver Julesburg Basin (DJB). In 2017, there were over 53,000 active O&NG wells in Colorado (Figure 1). The O&NG development is concentrated in a number of lower elevation basins, with ≈24,000 wells (January 2018) located in Weld County in the DJB, which in 2017 produced ≈90% of the oil in Colorado [Swain, 2018]. From 2010–2018, annual natural gas production in Weld County increased by a factor of 3.5, and annual oil production by a factor of ≈6.5 [Drilling-Edge, 2019].

Some of the growth of the O&NG industry has occurred within the periphery of urban and residential areas, raising concerns within communities. Proximity to O&NG operations has been associated with human health effects, with atmospheric emissions being a primary pathway of exposure. Types and causes of emissions are dependent on multiple variables and stages of the well development, and can arise, for instance, from heavy equipment use at the site and vehicle traffic, power generation, drilling operation, spillage and evaporation of fracking fluid, flowback of the extracted petroleum products, flaring, and fugitive or controlled hydrocarbon emissions during loading and transportation [Adgate et al., 2014]. Fracking fluid is a mixture of a multitude of synthetic chemicals, with the composition typically kept proprietary by operators. Silica, added to the fracking fluid, dust, and soot/particles from diesel engines contribute to particulate exposure. Gaseous emissions arise from fracking fluid additives, controlled venting, flaring, and leakage of equipment, storage tanks, and pipelines. Directly emitted gaseous pollutants of concern for human health are hydrogen sulfide and petroleum constituents, including aromatic and polycyclic aromatic hydrocarbons. Combustion processes cause emissions of carbon monoxide, nitrogen oxides, volatile organic compounds (VOCs), and soot/particulates [Adgate et al., 2014].

O&NG VOC emissions are a complex mixture of hydrocarbons. Some of the VOCs identified have the potential to affect the human endocrine system [Colborn et al., 2014]. Enhanced levels of VOCs have been observed in air near O&NG wells, including the known human carcinogen benzene [Verma et al., 2000; Macey et al., 2014; Sovacool, 2014; Halliday et al., 2016]. Human health risk assessments based on measured and modeled VOC concentrations near O&NG sites indicate increased risks for respiratory, neurological, and hematological health effects, as well as excess lifetime cancer risk above the U.S. Environmental Protection Agency (EPA) de minimis risk of one in a million for people living nearest to O&NG sites [McKenzie et al., 2012; McKenzie et al., 2018; McMullin et al., 2018; Holder et al., 2019]. Associations between proximity to O&NG sites and health effects, including congenital heart defects, childhood leukemia, asthma, low birth weight, and preterm births have been reported [HEI, 2019]. In Colorado, children with congenital heart defects and leukemia are more likely to be born or living in the areas with the densest O&NG activity [McKenzie et al., 2014; McKenzie et al., 2017; McKenzie et al., 2019]. While these studies indicate that VOC emissions from O&NG activities have the potential to and may be affecting the health of nearby residents, further study will be necessary to elucidate causality [HEI, 2019].

Furthermore, atmospheric oxidation of O&NG VOC emissions contribute to the photochemical formation of ozone and secondary aerosols, which pose additional health concerns. A recent EPA study estimated 1000 and 970 added premature deaths from particulates (PM2.5) and ozone exposure caused by O&NG pollutants in the U.S. by 2025, with 37 (from PM2.5) and 34 (from ozone) of those predicted to occur in Colorado [Fann et al., 2018]. Polluted air from the Northern Colorado Front Range (NCFR) can be lofted during upslope flow conditions on the eastern Rocky Mountain slopes to high elevation where it can impact alpine ecosystems, including Rocky Mountain National Park [Brodin et al., 2010; Thompson et al., 2015; Benedict et al., 2018]. Up to ≈20 ppb of additional ozone in air transport associated with O&NG emission has been measured at the Rocky Mountain National Park Longs Peak monitoring station [Benedict et al., 2019]. For southwestern Colorado, including Mesa Verde National Park, an maximum ozone enhancement of 9.6 ppb was estimated for the maximum daytime average 8-hour ozone (MDA8) from O&NG influences [Rodriguez et al., 2009].

Increased surface ozone, at times exceeding health safety standards, have been observed in the NCFR for some 20 years. After 10 years of repeated exceedances of the 75 ppb U.S. ozone National Ambient Air Quality Standard (NAAQS), the Denver Metro Area (DMA) and NCFR, including the seven counties of Adams, Arapahoe, Boulder, Broomfield, Douglas, and Denver, as well as portions of Weld and Larimer counties, were designated as a ‘Marginal’ ozone nonattainment area (NAA) for the 2008 NAAQS. Because of a lack of progress in lowering ambient ozone, the area was bumped up from a “Marginal” to a “Moderate” NAA for the 2008 ozone standard in early 2016 [CDPHE, 2019a], and in December 2019, the EPA reclassified the area to a “Serious” NAA for the 2008 standard [EPA, 2019]. In consideration of new health exposure findings, the NAAQS was lowered from 75 to 70 ppb in 2015 to provide a stronger protection to communities. This lower threshold will make it even more challenging for the NCFR to reach compliance with the standard. Note, that even the 70 ppb standard is higher, and less protective than the ozone standard in many other nations, including Canada and the European Union, as well as the recommended guideline value by the World Health Organization (WHO, [WHO, 2019a]).

In December 2018, the Colorado Department of Public Health and Environment (CDPHE) petitioned to the EPA for deferral of the re-designation (Supplemental Materials). In their letter, the agency stated that “Colorado has seen a dramatic decline in ambient levels of oil and gas related VOCs” and that the “majority of ozone concentrations in the DMNFR (Denver Metro Northern Front Range) are the result of emissions outside of the State’s control, including naturally occurring emissions and emissions transported from other states and countries”. The arguments presented in the letter did not consider most of the considerable body of peer-reviewed literature on the impacts from O&NG industry operations on Colorado’s air quality that has emerged during the past 10 years. Shortly after the new Colorado governor Jared Polis took office in 2019, he withdrew this petition, stating “There’s too much smog in our air, and instead of hiding behind bureaucracy and paperwork that delay action, we are moving forward to make our air cleaner now” [Nicholson, 2019]. For directing this policy there is urgency to consider the current understanding of ozone precursor sources, atmospheric transport, and chemistry, from published literature. This article examines predominantly this peer-reviewed literature with the goal to provide the State agencies, boards, legislators, and policymakers a summary document for better assessing the role of O&NG industry emissions on NCFR-DMA air quality, with a particular emphasis on surface ozone.

This review and policy bridge provides an overview of the evolution of the understanding of atmospheric impacts and the current state of knowledge of O&NG emissions in Colorado. A comprehensive review and evaluation of health effects of atmospheric O&NG emissions in Colorado is intentionally not included. This review primarily builds on peer-reviewed journal articles. Included were all articles that were identified using various search strategies, regardless of the reputation of the journal or recognition of the published work by the community. There has been remarkable growth of literature over the past ten years (Figure 2). A good fraction of publications have arisen from the FRAPPE (Front Range Air Pollution and Photochemistry Experiment [FRAPPE, 2013]) and DISCOVER-AQ (Deriving Information on Surface Conditions from COlumn and VERtically Resolved Observations Relevant to Air Quality [DISCOVER-AQ, 2013]) campaigns that, in the summer of 2014, brought researchers from a wide array of disciplines and institutes to Colorado to study air pollution sources and chemistry in the NCFR. Outcomes of these studies are evident in the increase in published papers in subsequent years (Figure 2). Most research on O&NG impacts on air quality has centered in the NCFR, except a few studies that investigated methane emissions in the Four Corners region (region of the U.S. consisting of the southwestern corner of Colorado, southeastern corner of Utah, northeastern corner of Arizona, and northwestern corner of New Mexico). This review will focus on the NCFR research. A summary of published work arranged by publication year is provided in Table 1.

It is noteworthy that the majority of this work has been published in well recognized/high impact factor peer-reviewed science journals. Lead and contributing authors are from federal laboratories (i.e. NCAR, NOAA) and Colorado and out-of-state universities. The journal count is led by the Journal of Geophysical Research (18), followed by seven publications in Atmospheric Chemistry and Physics, and five articles each in Environmental Science and Technology and Elementa. This publication record can be deemed as a testimony of the recognized importance of this research, having attracted leading scientists from top U.S. research institutions.

Air quality monitoring and air sampling is conducted by the CDPHE, NOAA, the National Park Service, Boulder County, and the City of Longmont. A map showing the distribution of monitoring sites and measured species is shown in Figure 1. Ozone is monitored at the highest number of sites, followed by PM2.5, and nitrogen oxides. There currently is only one location with continuous VOCs monitoring (Boulder Reservoir [Helmig et al., 2020], Boulder County site in Figure 1); however, two more sites are anticipated to begin VOCs monitoring within the next year (Longmont Union Reservoir (the eastern of the City of Longmont sites in Figure 1) and Rocky Flats North (CDPHE site south of Boulder, Figure 1)). Continuous methane monitoring is currently conducted at the Boulder Reservoir and Longmont airport (the western of the City of Longmont sites in Figure 1); continuous methane monitoring at the Longmont Union Reservoir is planned to start in early 2020. Besides this real-time monitoring, methane and VOCs are also quantified in flasks and canisters samples collected at Niwot Ridge (NOAA) and at Platteville and downtown Denver (CAMPS) by CDPHE. There is a relatively high density of sites in the DMA, but a relatively sparse network within and along the periphery of the DJB O&NG area.

Nitrogen oxide (NO and NO₂ = NO_x) emissions associated with O&NG development arise from a number of sources, including flaring, on-site electrical power generation, heavy equipment operation at fracking sites, and the heavy truck traffic for moving equipment and fluids in and out of O&NG well sites. Other sources are compressor stations, and heavy tanker truck traffic for transporting produced oil and gas products from the site. NO_x emissions arising from these sources have been studied in a number of O&NG basins [Bogacki and Macuda, 2014; Field et al., 2014; Majid et al., 2017; Archibald et al., 2018]. However, this literature review did not identify any studies from the DJB, leaving the contribution from the DJB O&NG sector to the regional NO_x emissions uncertain. The Regional Air Quality Council emissions inventory lists the year 2017 total O&NG NO_x sources at 65.8 tons per day, which is a 59% increase over the year 2011 emissions [Brimmer, 2019].

Methane is emitted by a variety of sources, with wetlands, landfills, feedlots, seepage from geological reservoirs, and O&NG extraction, distribution, and industries being the most significant ones on a global scale. Methane is a potent greenhouse gas. The methane background in the global atmosphere has more than doubled since preindustrial [Kirschke et al., 2013]. The increase of methane from anthropogenic sources is the second ranked contribution (after CO₂) to radiative forcing from anthropogenic greenhouse gases. The oxidation of methane in the atmosphere is also a pathway for ozone production, which constitutes another climate forcing pathway. Ozone exerts stress on the natural and agricultural ecosystems and human and animal life (see below). It has been estimated that a reduction of global methane emission by 20% would reduce the ozone MDA8 by ≈1 ppb in the background atmosphere [West and Fiore, 2005; West et al., 2006]. The benefits of methane reductions are shared internationally. The 20% emissions reduction was estimated to avoid ≈30,000 premature deaths by 2030 globally [West et al., 2006].

The O&NG industry is the single most significant source of methane in Colorado [Pétron et al., 2012]. Quantifying the methane flux from the O&NG industry has been challenging as the geographic area of the O&NG activities overlaps with agricultural, beef, and dairy production, which all constitute significant methane sources. Methane to non-methane hydrocarbon (NMHC) relationships, in particular those with ethane and propane, and the stable isotopic signature of methane, have been used to decipher the O&NG contribution to the total methane flux. Point source measurements near emission sources, mobile lab ground surveying, and aircraft observations, in combination with inventory information, have been used to derive basin-wide O&NG methane flux estimates. Three NOAA studies, covering observations during three short time windows within the 2008–2015 period, are summarized in Table 2. Data in this table are scaled to annual flux estimates. As all these experiments relied on relatively short observation periods, there is an inherent uncertainty from the lack of knowledge if, and how representative these shorter observations were for year-round conditions. Further, variability in parameterizations and uncertainties in assumptions that go into these flux estimates cause relatively large uncertainty ranges in the results (column 3). The best estimate values of the three studies are relatively consistent, nonetheless, spanning 130–169 Gg yr^–1. According to the U.S. Energy Information Administration, Colorado households consume an average of 103 million BTU of natural gas per year [EIA, 2009], which converts to ≈2 tons of household natural gas consumption per year. Therefore, the Peischl et al. [2018] O&NG methane emissions estimate corresponds to the natural gas consumed by ≈84,000 Colorado households.

These data points are too few, and uncertainties are too large, to make statements about potential trends in the methane flux over this time window with statistical certainty. Considering the increase in natural gas production during this time period, the relatively flat total emissions would indicate a reduction in the relative fugitive emissions rate. Peischl et al. [2018] present a statistical analysis that results in an 83% likelihood of a reduced methane leakage rate during 2008–2015.

Methane emissions result in atmospheric concentration increases in the source regions and downwind. Due to the relatively long atmospheric lifetime of methane (≈9 years) in comparison to NMHCs, at ≈1900 ppb the methane background is 3–4 orders of magnitudes higher, and methane enhancements are moderate (≈10%) on a relative scale. The slow atmospheric oxidation of methane causes relatively little of the regionally emitted methane to be oxidized locally. Modelling work [Lindaas et al., 2019] estimated a 2% contribution from DJB-wide emitted methane oxidation to the regional ozone production.

The Four Corners area is another Colorado region that has received attention because of its recognized methane emissions. Based on satellite data analyses, Kort et al. [2014] reported “… the largest anomalous CH₄ levels viewable from space over the conterminous U.S. are located at the Four Corners region in the Southwest U.S.” Their work primarily pointed out discrepancies between inventory and these satellite data derived methane flux estimates. Follow-up studies have confirmed an abundance of fossil methane sources in the Four Corners regions, with contribution from coal shaft venting, natural seepage, and O&NG well and distribution sites [Frankenberg et al., 2016]. The methane flux estimate for the Four Corners region of 540 ± 200 Gg yr^–1 (1σ) [Smith et al., 2017] exceeds the methane O&NG flux estimates for the DJB by a factor of ≈3.

Petroleum NMHC are the dominant constituents of VOC emissions from O&NG sources. Atmospheric VOCs in the DJB are highly elevated, largely due to O&NG emissions. NMHC and the combined atmospheric carbon from all species exceeds those in major urban areas [Swarthout et al., 2013]. Relative abundances of VOC species scale inversely with molecular size; ethane is typically the NMHC emitted at the highest flux, followed by propane, then the butanes, and so forth. There are dozens of individual VOCs that have been listed in O&NG emissions. However, the bulk of the mass is contributed by a narrower count. For instance, in O&NG plumes identified in continuous monitoring at the Boulder Reservoir, the 16 most abundant VOC species account for approximately 90% of the total O&NG emitted VOC mass [Helmig et al., 2020]. O&NG VOC emissions also contain aromatic constituents, such as the BTEX species (benzene, toluene, ethylbenzene, xylenes) [Pétron et al., 2012; Gilman et al., 2013; Swarthout et al., 2013; Koss et al., 2017]. While these are relatively low, sub-1% constituents, they have received notable attention because of their recognized health impacts on humans.

Over the past five years, ethane has become an increasingly utilized tracer for natural gas VOC emissions. This increased attention to ethane has also been fostered by new instrumentation that has recently become available for sensitive and fast response ethane detection [Richter et al., 2015; Yacovitch et al., 2015; Yacovitch et al., 2017; Barkley et al., 2019; Kostinek et al., 2019]. Ethane has relatively weak non-O&NG source emissions, which makes it a sensitive tracer for identifying O&NG plumes and influences. The ethane to methane enhancement ratio has been used to characterize emissions from particular basins, and for scaling the VOCs flux to methane. In the DJB, natural gas on average has an ethane/methane molar ratio of 12–16% (ppb/ppb), equivalent to a 23–30% mass ratio [Peischl et al., 2018; Kille et al., 2019; Helmig et al., 2020], which is close to the estimated mean of all U.S. O&NG basin emissions [Helmig et al., 2016].

Atmospheric monitoring data show highly variable VOC concentrations. Large VOC enhancements can occur in air plumes originating from O&NG source regions [Swarthout et al., 2013; Rossabi et al., 2017; Helmig et al., 2018]. This effect is also evident in the large relative standard deviation of statistical analyses of the region’s VOCs data [Rossabi et al., 2000]. At the Boulder Atmospheric Observatory (BAO, Figure 3a), located in Erie at the transition between the more densely populated and industrialized DMA to the south, and the DJB O&NG and agricultural areas to the north, air composition was found to have variable urban and O&NG signatures from significant mixing and recirculation of air influenced by these different sources [McDuffie et al., 2016]. As with any surface-emitted source, VOC concentrations also vary significantly between day and night, with typically higher nighttime concentrations due to the nighttime absence of dilution of surface air from convective mixing [Swarthout et al., 2013; Halliday et al., 2016]. Concentrations drop with height [Swarthout et al., 2013], indicating that releases occur at the surface.

Several studies have demonstrated spatial gradients with increasing mean VOC concentrations and increasing variability with decreasing distance to O&NG wells and operations. Total alkanes concentrations increased by a factor of 10 from distances >1600 m to distances of <152 m [McKenzie et al., 2018]. Even higher gradients (up to a factor of ≈30) were found for BTEX species. The significance of O&NG sources to ambient BTEX was demonstrated in continuous surface measurements conducted at a site near Platteville (Figure 3a) during FRAPPE [Halliday et al., 2016]. Benzene values exceeding 10 ppb, with a maximum of 29.3 ppb, were observed on multiple occasions, particularly at night when emissions were trapped near the surface. The mean nighttime value (0.73 ppb) for August 2014 was above the 1:100,000 increased lifetime cancer risk threshold (0.5 ppb) listed by the World Health Organization [WHO, 2019b]. Benzene values above 10 ppb were also reported in nine canister samples collected from a mobile surface platform downwind of different O&NG facilities, including operations labelled as “waste water disposal well” and “oil waste dumping facility” [Pfister et al., 2017a]. Newer observations from mobile platforms presented by NOAA and University of Wyoming scientists point to even larger BTEX enhancements downwind of drilling operations and produced wastewater facilities: Madronich et al. [2019] found an abundance of benzene mixing ratios in the 10–50 ppb range near Greeley in the center of Weld County. Mielke-Maday et al. [2019], in their analysis of correlating wind with benzene data, found the highest benzene enhancements when winds originated from the direction of a nearby multi-well pad. Thus far, the highest concentrations were recorded in proximity of oil and gas wastewater disposal facilities in eastern Weld County, with maximum BTEX plume values approaching 500 ppb downwind of these facilities [Edie et al., 2019].

Horizontal gradients in VOCs have also been demonstrated on a larger regional scale along transects from the periphery towards the center of the DJB. This behavior has been seen in surface [Thompson et al., 2014; Helmig et al., 2015; Rossabi et al., 2017; Rossabi et al., 2019] and aircraft data [Richter et al., 2015]. A representative example is illustrated in Figure 4, comparing data from a site in downtown Denver, data from Erie (southern border of the DJB), and from Platteville (Figure 3a). Mean mole fractions for the alkane NMHC are a factor of 10–50 higher at Platteville than in Denver, with Erie values falling in between. The Platteville ethane and propane values are among the highest ambient concentration values for these species ever published in the literature. For higher molecular weight NMHC and aromatic compounds, spatial gradients are less pronounced, indicating that emission sources for these compounds have a more even geographical distribution. Besides health concerns from direct exposure, these emissions are a major culprit for the regional photochemical ozone production that will be discussed in the next section.

Ozone has long been recognized as an important air pollutant. Breathing air with elevated ozone irritates the respiratory system and can cause acute and chronic respiratory cardiovascular health effects. People with asthma, children, and the elderly are particularly at increased risk. There is a rich literature on ozone health effects (i.e. [Fleming et al., 2018; Lefohn et al., 2018]). The risk increases with ozone concentration and length of exposure. Health effects have been noted in numerous studies below the current 70 ppb NAAQS [Fleming et al., 2018]. Exposure of communities to elevated ozone has been proven to increase mortality rates during and shortly after increased ozone events [Gryparis et al., 2004; Bell et al., 2005]. In an extensive study across 90 U.S. urban communities, a 0.5% increase in daily mortality was found for a 10 ppb increase in daily mean ozone [Bell et al., 2005].

Through plant respiration and surface uptake, ozone is also damaging to vegetation [Mills et al., 2018]. The stress on vegetation from ozone reduces plant growth and productivity, causing significant loss to U.S. and global farming industries and food supply [Van Dingenen et al., 2009; Lapina et al., 2016].

Ozone is not a directly emitted pollutant, but is formed in the atmosphere through a series of photochemical reactions that are fueled by emission of VOCs and NO_x in the presence of sunlight. The efficiency of this chemistry is rather complex, depending on other variables, including the ratio of VOC/NO_x, VOC speciation and reactivity, solar radiation, temperature, wind speed and dispersion conditions. This causes rates of ozone production to vary substantially, from single digits to tens of ppb h^–1 during mid-day hours.

Background ozone is generally higher in the western U.S. overall than in the eastern U.S. [Zhang et al., 2011; Cooper et al., 2012; Jaffe et al., 2018], causing air moving into the NCFR being on average higher in ozone than in many other parts of the country. An analysis by the U.S. EPA estimates the non-U.S. background contribution on days when ozone is relatively high (>60 ppb) at 38 ppb, which is the highest among 12 included comparison sites [EPA, 2008], and more than half of the current ozone NAAQS. The contribution of ozone from long-range transport is on average contributing more to the background in spring than during the primary summer ozone season [Cooper et al., 2012; Lin et al., 2017]. During 2000–2015, the resulting summer ozone background (range of median ozone during summer at Colorado rural, non-mountain monitoring stations) was 32–49 ppb, with mean and median values of 41 ppb ([Bien and Helmig, 2018]; Supplemental Materials). Downward folding of high troposphere/lower stratosphere air has been observed on a few occasions to bring elevated ozone to the surface. These conditions depend on the strength and location of the polar jet, are irregular, and have been reported exclusively for the spring [Langford et al., 2009; Lin et al., 2015]. Emissions from wildfires can contribute to ozone production, with the rate and total amount of ozone produced being sensitive to the fire and plume conditions [Jaffe and Wigder, 2012; Jaffe et al., 2013]. Overall, fire emissions are a minor contribution compared to the role of anthropogenic emissions to the larger geographic scale ozone buildup [Lin et al., 2017]. This influence is highly variable, and estimated to enhance the Intermountain West regional summer MDA8 by 0.3–1.5 ppb [Lu et al., 2016]. Regional ozone production is further promoted by the dry and sunny climate. Combined, these conditions make it more challenging for regions in the western U.S., including the NCFR, to control ozone, as it leaves less room than in other regions for local ozone production to exceed the standard [Cooper et al., 2015]. Notably, background ozone at remote high elevations sites across the western U.S. during summer has been declining during the most recent decade [Bien and Helmig, 2018; Jaffe et al., 2018], which should constitute favorable conditions for the NCFR on its path towards lowering surface ozone.

A compelling case demonstrating the influence of O&NG emissions on surface ozone in the NCFR was first published in the Denver Metropolitan Area and North Front Range 8-Hour Ozone State Implementation Plan in 2008 [CDPHE, 2008]. Combining summer ozone data from four sites along the NCFR with air transport back trajectory analyses showed that for elevated ozone events during mid-May to mid-August 2006, air transport from the center of the DJB was associated with the highest ozone values, whereas transport from surrounding areas, including the DMA, brought in air with lower ozone levels (Figure 3b). The geographical overlap of the source footprint with highest ozone with the area of highest O&NG well density (Figure 3a), provided credence to the argument that O&NG industry emissions played an important role in ozone production and high ozone occurrences. Daytime summer ozone production rates of 7–8 ppb hr^–1 have been seen in ambient diurnal ozone data [CDPHE, 2008; Cheadle et al., 2017]. Direct measurements of the ozone production capacity in Golden during FRAPPE found maximum late morning ozone production rates about two times that high [Baier et al., 2017]. Assessing the relative benefit of VOCs versus NO_x controls is extremely challenging in the NCFR. VOC/NO_x ratios vary widely across the region, with lower ratios present in the DMA, and higher ratios in the VOC-rich DJB. These different air masses can mix during transport and recirculation, causing a wide range of spatial and temporal differing conditions and ozone production regimes.

Two studies estimated the contribution of O&NG VOCs to the total reactivity with the OH radical using VOC speciation and atmospheric concentrations at the BAO. This variable can serve as a metric for the chemical reactivity of air and its potential for producing ozone. A NOAA study estimated that 55 +/– 18% of the reactivity was attributable to O&NG emissions [Gilman et al., 2013]. Swarthout et al. [2013], using a similar approach but with independently collected data, determined a value of 57%. While OH reactivity does not directly translate to ozone production, based on these results, both groups predicted that O&NG VOC emissions would enhance and play a significant role in the regional ozone budget. It should be noted that these measurements were conducted in the late winter, when ozone production is relatively moderate in the NCFR. Therefore, these findings represent, for example, lower influence from biogenic VOCs.

A NOAA study, i.e. McDuffie et al. [2016], went a step further and incorporated VOC speciation and VOCs reactivity in a photochemical model. Their findings showed “that O&NG alkanes contribute over 80% to the observed carbon mixing ratio, roughly 50% to the regional VOC OH reactivity, and approximately 20% to regional photochemical ozone production.” Using observations from BAO for correlation analyses and modeling of oxidation chemistry, Lindaas et al. [2019] stipulated that O&NG emissions contribute to ozone production on high ozone days; however, that study fell short of providing a quantitative estimate. Another modeling study by NCAR scientists [Pfister et al., 2017b], building on FRAPPE data for the wider NCFR area, concluded that on average, O&NG emissions contribute 30–40% to the local ozone production on high ozone days. It needs to be emphasized that all of these studies derived estimates for the ozone produced regionally, not the total ozone, which is also determined by the background that is transported into the region (see above).

These predictions from reactivity consideration and modeling are backed by a series of observational studies: Evans and Helmig [2017], using a correlation analysis of ambient ozone and wind data from BAO and the South Boulder Creek regulatory monitor, found that during 2009–2012, 65% (average between both sites) of elevated ozone events were associated with transport from O&NG production regions. Cheadle et al. [2017], analyzing selected cases of observations near Greeley during FRAPPE, estimated that O&NG emissions contributed up to ≈20 ppb to ozone production on high ozone days. Oltmans et al. [2019] conducted an in depth analysis of the conditions on high ozone days at BAO. Their analysis showed an association of high ozone days with transport from sectors with intense O&NG production towards the northeast. The authors concluded that O&NG emissions were an important source of ozone precursors and are crucial in producing peak ozone events in the NCFR. The ozone production chemistry is primarily driven by VOCs of anthropogenic origin; biogenic emissions appear to have a minor contribution to the NCFR ozone production chemistry [Cheadle et al., 2017; Lindaas et al., 2019].

Ozone enhancements in air enriched with O&NG emissions have been measured all the way up the Rocky Mountain National Park in upslope flow along the eastern slopes of the Rocky Mountains. Benedict et al. [2019] estimated that high ozone events associated with O&NG emissions contributed ≈20 ppb of additional ozone at the Rocky Mountain National Park Longs Peak air monitoring station. This poses the question of ecosystem impacts of the elevated ozone on the natural environment, including Rocky Mountain National Park. Ozone has long been known to damage crops, and reduce yields in agriculture [Heck et al., 1982; Van Dingenen et al., 2009; Avnery et al., 2011]. For sensitive crops, those losses can be well more than 10%, accounting to a significant revenue loss to the farming industry [Morgan et al., 2003; Avnery et al., 2011]. The DJB O&NG production overlaps with a region that is also considered the agricultural heartland of Colorado. Colorado’s agricultural industry provides over $40 billion to the state economy [USDA, 2018]. There have been no assessments to date on the revenue loss to this industry from the elevated ozone caused by O&NG emissions within the State.

The DMA has a history of air quality problems from particulates pollution that goes back to the 1970s and 80s [Waggoner et al., 1983]. The episodic wintertime occurrences of reduced visibility from accumulation of gaseous and particulate matter near the surface have been named the ‘Denver Brown Cloud’ [Neff, 1997]. They are tied to the peculiar topographical and meteorological conditions in the NCFR, where during the winter, shallow (<500 m) boundary layer heights, low convective mixing, also promoted by snow cover and cold soils, can promote accumulation and buildup of particular emissions over several days. A series of studies revealed a mix of sources, with traffic, urban, and agricultural emissions [Wolff et al., 1981]. Most of the visibility reduction was found to be associated to particulates smaller than 2.5 μm [Groblicki et al., 1981]. Secondary aerosol production, particularly growth of organic aerosol, as air recirculates over areas with different source signatures, was identified as a major contributing mechanism for aerosol buildup [Sloane et al., 1991]. This secondary production of aerosol and ozone occurred on winter days, with temperatures as low as 6°C [Ferman et al., 1981].

Despite the population growth of the DMA/NCFR, occurrences of Denver Brown Cloud episodes and the aerosol loading (CDPHE, 2019, unpublished results) have seen a gradual decline during the past two decades, most likely thanks to stricter air pollution control measures.

O&NG operations are sources of atmospheric aerosol in several ways. Heavy equipment operation, soot emissions from diesel engines and power generation, unloading and handling of silica that can be added as a fracking fluid constituent, and soot from oil and gas flaring, are some of the important primary emissions sources. There is also potential of secondary aerosol formation from the atmospheric oxidation of gaseous emissions, such as H₂S, SO₂, NO_x, and VOCs. These gases are known to produce less volatile chemicals during their atmospheric oxidation, which can serve as aerosol nuclei or add to existing aerosol. Comparatively few published studies have addressed aerosol from O&NG operations in Colorado. The reason may lie in the difficulty of attribution, which is more difficult for PM2.5 relative to ozone, whose formation mechanism is better defined.

Continuous vertical profiles of aerosol and gaseous components were measured during the NACHTT (Nitrogen, Aerosol Composition, and Halogens on a Tall Tower) campaign at the BAO [Brown et al., 2013]. The aerosol mass was dominated by nitrate, which was mostly from sources within the region. Other significant contributions were from organics and sulfate, with sulfate primarily being transported long-range. While the composition of organic gas phase compounds was noted to have a strong O&NG influence, the study conclusions do not specify O&NG influences on the aerosol composition. An investigation on aerosol dependency on circulation patterns in the NCFR found that cyclone conditions promoted the transport of aerosol constituents from the northern Front Range into the DMA, increasing aerosol mass loadings and reducing visibility [Vu et al., 2016]. The circulation pattern would be expected to cause air to become enriched in O&NG emissions while passing over the DJB, before circling back into the DMA. The organic fraction made the largest component of total aerosol. The study, however, did not specify if the high organic aerosol loading was associated to O&NG precursor emissions.

Two publications report on the O&NG contribution to the NCFR aerosol loading in the context of FRAPPE. Dingle et al. [2016] determined extinction enhancements relative to the amount of the combustion tracer carbon monoxide. They found an increase in the extinction coefficient with the aging of air masses that was accompanied by formation of secondary organic aerosol; the extinction was strongest correlated with organic aerosol in O&NG-influenced air, and with nitrate aerosol in O&NG and agriculture emissions. Bahreini et al. [2018] reported a significant contribution of non-combustion organic aerosol. Organic aerosol was on average 40% higher in plumes with a high O&NG influence, and the organic aerosol was dominated by secondary constituents, suggesting that they may be products of O&NG VOCs oxidation. The study concluded that O&NG sector emissions contribute up to 38% to the secondary organic aerosol in the region.

The impact of O&NG emissions on NCFR air quality is exacerbated by very peculiar atmospheric circulation patterns. Johnson and Toth [1982] were the first ones to present an in depth characterization of the daily cycle of mountain-valley winds. At night, cooler air flows from the mountains and down the Platte River valley over the plains (west to east transport). During the day, the air flow reverses, bringing air from over the plains (and O&NG source regions; east to west transport) back to the foothills. Such recycling can continue over several days. During daytime, upslope flow is a prominent flow regime. This circulation is driven by the warming of the easterly slopes of the Rocky Mountains range, causing convective uplifting that is pulling in air from the east. The flow reverses during the night, with cooler air from higher elevations descending the mountains and forcing west to east air transport.

Figure 5 is a partial reproduction of a figure from Evans and Helmig [2017]. These windroses, generated from summer data at BAO, show the very distinct diurnal flow behavior, with flows from north to southeasterly directions dominating during the daytime hours, and south to westerly winds predominant at night. The study analyzed wind data from eight locations along the NCFR. The average diurnal winds were remarkably consistent, demonstrating the importance of this flow regime for the wider NCFR. The transition time between these two flow regimes changes with distance from the mountain slopes, with locations further east experiencing an on average later onset of upslope flow conditions. The diurnal flow regimes are most pronounced during the summer because of the larger solar irradiance that is providing the thermal forcing.

The upslope flow paths are somewhat segregated, such that there is a separation of air masses that are more heavily influenced by O&NG emissions to the north of the DMA, whereas air masses south of North Denver are more strongly influenced by urban emissions [Pfister et al., 2017a]. Air enriched with emissions from urban, traffic, O&NG, and other regional sources can get ‘trapped’ along the mountain slopes during late afternoon. This is reflected by highest ozone levels being observed at monitoring sites nearest to the mountain slopes [Bien and Helmig, 2018] and at the lower elevations in the foothills [Brodin et al., 2010]. In upslope flow, polluted air from the NCFR regularly reaches high elevation zones on the eastern side of the Rocky Mountain Continental Divide [Brodin et al., 2010; Benedict et al., 2019]. On days with particularly strong flow conditions, NCFR pollution can “spillover” into the valleys to the west of the Continental Divide [Pfister et al., 2017b].

Although a classical view of high pollution episodes invokes a stagnant high pressure region, usually over flat terrain, the meteorology of the NCFR leads to more complex circulation regimes. A common regime in the winter occurs with downslope westerly warm winds from the Rocky Mountains flowing over colder air drawn from the east toward a low pressure trough along the foothills, or due to lee-cyclogenesis located over southeast Colorado [Neff, 1997]. During the summer, the ‘Denver Cyclone’ is often observed. These conditions provide a similar opportunity for trapping pollutants near the surface [Wilczak and Glendening, 1988; Wilczak and Christian, 1990; Szoke, 1991]. In this case, the Denver Cyclone occurs nearer the surface with warmer air aloft from the south that originated over the Palmer Divide, a ridge extending to the east and south of Denver. As the air from the east (underlying the warmer air aloft) carries pollutants and precursors from the eastern plains, the air can stagnate as it encounters the topographic barrier to the west. This circulation pattern can cause pollution to circulate and accumulate for several days, leading to increases of secondary pollutants. Vu et al. [2016] demonstrate an up to an 80% increase in aerosol constituents during a cyclone episode during FRAPPE.

The frequency and prominence of high ozone occurrences is correlated with high pressure systems that promote sunny weather, high temperatures, stagnant air circulation, which are conditions that are favorable for photochemical ozone production. Reddy and Pfister [2016] investigated this relationship and proposed a method in which monthly 500-mbar pressure heights were used for correcting the year-to-year variability in the fourth highest MDA8 ozone. Further, these conditions promote cyclic terrain-driven circulations that reduce pollution transport away from sources. The authors recommend correcting annual MDA8 data using monthly 500-mbar pressure heights for reducing weather influences on ozone trends.

Emission inventories have been developed by state and national regulatory agencies in support of air quality modeling and for directing policy development. These bottom-up inventories are based on emissions estimates of facility types and operations with regional/basin-wide scaling using best available facility counts. Evaluation of the bottom-up inventory estimates has mostly been accomplished by university and NOAA scientists through comparing with top-down flux estimates that were developed from aircraft and surface data. Experimental capabilities for basin-wide, top-down flux determinations have improved remarkably over recent years. Emissions have been estimated by determining the enhancement in the basin outflow by aircraft profiling upwind and downwind of production regions, determination of horizontal winds and boundary layer depth [Karion et al., 2013; Karion et al., 2015; Peischl et al., 2015; Peischl et al., 2018].

Most of these studies have pointed out inconsistencies in inventories and a likely underestimation of O&NG inventory surface emissions. Uncertainties of bottom-up inventories can arise from the extrapolation of limited information of facility-scale emissions from venting, flashing, and leakage, and the neglect of differences in practices of operators. Inventories are annual averages with little to no temporal information. Vaughn et al., [2018] demonstrated that this may be part of inventory uncertainties and discrepancy between bottom-up and top-down emission estimates. Further, the lack of temporal information makes source apportionment and model performance evaluation more difficult.

In an extensive evaluation of methane and VOC emissions representation in the Western Regional Air WRAP Phase III inventory [WRAP, 2009], Pétron et al., [2012] concluded that “there are notable inconsistencies between our results and state and national regulatory inventories”. They further stated “Our analysis suggests that the emissions of the species we measured are most likely underestimated in current inventories and that the uncertainties attached to these estimates can be as high as a factor of two”. Results also showed that methane sources from natural gas industries in Colorado were most likely underestimated by at least a factor of two. Besides methane and total VOC, the study also assessed benzene, and concluded that for this species State inventory estimates were too low by at least a factor of five. Levi [2012] commented on the difficulties and high sensitivity of the top-down emissions estimation based on emission ratios of VOC species.

In their assessment of the National Emissions Inventory (NEI) for 2010, Tzompa-Sosa et al. [2017] found that inventory fossil fuel emissions had to be increased by 40% for the Northern Hemisphere to yield agreement with observations, except for the central U.S., including Colorado, where even the 40% increase under-predicted observed mixing ratios in the lower troposphere.

Pfister et al. [2017a], in their modeling of FRAPPE and DISCOVER-AQ data, found that they had to increase O&NG non-ethane emissions by a factor of four over their best inventory estimate for the best match between observations and model output.

The most recent evaluation, based on NOAA aircraft surveying [Peischl et al., 2018], illustrates relatively little improvement in agreement between inventories and top-down emission estimates. The NOAA study determined a DJB-wide ethane flux of 7.0 +/– 1.1 × 10³ kg hr^–1, which translates to 61 +/– 10 × 10⁶ kg (kilotonnes) yr^–1. This ethane flux alone is higher than then current Regional Air Quality Council O&NG non-ethane total VOC bottom-up inventory flux of 56 × 10⁶ kg yr^–1 [Brimmer, 2019]. With ethane constituting approximately 30% of the total O&NG VOC flux in the regional oil and gas emissions [Gilman et al., 2013; Swarthout et al., 2013], the Peischl et al. [2018] ethane flux equates to a total O&NG VOC flux of ≈230 × 10⁶ kg yr^–1 (Table 3). Excluding ethane yields ≈170 × 10⁶ kg yr^–1, which exceeds the non-ethane Regional Air Quality Council (RAQC) total VOC estimate by a factor of ≈three.

Taken together, these available comparison studies highlight the deviations between the bottom-up and top-down emissions estimates. Unfortunately, there is a scarcity of top-down estimates available for this evaluation, and each of these have relatively large uncertainty windows themselves. Nonetheless, these disagreements diminish the confidence in the bottom-up inventories, and air quality modeling that is building on these most likely under-predicted emissions.

Since 2008, Colorado has implemented regulations to reduce VOC emissions from O&NG sources, and methane-specific regulations came into place in 2014 [Ogburn, 2014; CDPHE, 2019c]. State inventories largely build on projected emissions reductions from these measures. However, there are very few data records that allow an evaluation of the important questions, if and how actual O&NG emissions in the DJB have changed over time. There is no published peer-review literature at this time that has addressed this question and presented trend results that would allow assessing basin-wide emission changes with statistically significant certainty. Data from after the methane emissions rule adoption in 2014 would be most helpful in understanding the benefits of that regulation and current emission levels.

Comparing historic with modern observations of VOC data from Boulder, Thompson et al. [2014] stated “An initial look at comparisons with data sets from previous years reveals that ambient levels for oil and gas-related NMHC in Erie, as well as further downwind in Boulder, have not decreased, but appear to have been increasing, despite tightening of emissions standards for the oil and gas industries in 2008.”

CDPHE has been conducting canister air sampling at Platteville since 2011 [CDPHE, 2019b]. However, inconsistencies in the sampling, uncertainties in the analysis protocols [Hood, 2019], siting of the sampling location, and the proximity of the sampling location to abundant nearby well sites make trend determinations and their interpretation for the wider region from these data uncertain.

Ethane column observations conducted from 2010–2015 at the NCAR Foothills Laboratory in Boulder are presented in Franco et al. [2016]. Their best estimate is a rate of increase of 5.0% per year. This rate is above estimated rates for the increase of ethane in the Northern Hemisphere background atmosphere during this time window [Helmig et al., 2016], which implies increasing ethane emissions in the region. However, the uncertainty interval in this result is rather large. Including newer data, extending the record to 2010–2018, did not yield a trend in the atmospheric ethane abundance (J. Hannigan, NCAR, personal communication, April 2019).

NOAA conducted sampling of VOCs, with up to daily resolution, from 2007–2016 at 300 m height from the BAO tower. Data for the O&NG VOC tracer propane collected during midday to afternoon hours, when boundary layer mixing is most progressed, do not show statistically significant changes, indicating stable total emissions of O&NG VOCs during this 9-year time window [Oltmans et al., 2020].

Lastly, the methane flux estimates listed in Table 2, covering observations between 2008–2015, do not show any changes in the total methane flux that are outside of the uncertainty windows of the individual observations. Assuming that the VOC/methane ratio has remained constant, these methane flux determinations do not suggest changes in basin-wide VOC emissions. Considering the large increase in natural gas production during this time period, a reduction in the fraction of emitted methane (relative to the produced quantity of natural gas) and VOCs appears probable [Peischl et al., 2018].

Available VOC emissions estimates, differentiated by ethane, benzene, and total VOC, are provided in Table 3. The latest estimate in each category is based on relative observations of VOC/methane at the Boulder Reservoir during 2017–2018, scaled to the Peischl et al. [2018] year 2015 methane flux estimate. Therefore, these two data sets are linked to each other.

In summary, at this time, there do not appear to be observational records that allow deducing, with statistical significant certainty, if and how methane and/or VOC emissions may have changed in the DJB over the past 15 years. There is no convincing evidence for an overall decrease in VOC emissions at this time. Certainly, none of these data show increases that scale with the DJB O&NG production increase (e.g. 3.5–6.5 times for natural gas and oil, respectively, for 2010–2018). Therefore, it appears likely that relative emissions rates have declined, likely due to the implementation of stricter emission controls. However, the growth of the number of operations has probably counteracted those relative emissions reductions, resulting in overall basin-wide stable total emissions.

Air quality impacts from O&NG emissions arise from acute, chronic, and carcinogenic effects of primary emissions, particularly of BTEX VOC in close proximity to operations, and emissions of NO_x and particulates from equipment and on-site power generation. These exposures are of concern for residents living within a few hundred meters to kilometers of O&NG operations. According to the survey of McKenzie et al. [2016], in 2012, ≈56,000 citizens lived within a radius of 1000 feet of O&NG operations in Colorado. These populations are at greatest risks for these exposures. Secondary products that are formed via photochemical processing of emissions during transport are another concern. Here, the pollutants of importance are ozone and PM2.5. These species are transported across a wide spatial scale in the NCFR, thereby affecting a much larger population. In excess of 3.5 million people live in the NCFR ozone NAA. Approximately half of the NAA (mostly the northern part) is moderately to heavily influenced by O&NG emissions. This part of the NCFR is where O&NG emissions have the greatest impact on ozone and exceedances of the NAAQS. Atmospheric levels of particulates are relatively modest in the NCFR, with particulate air quality thresholds being exceeded only occasionally, for instance during wildfire plume transport events and wintertime inversion conditions. Nonetheless, health impacts from particulates originating from O&NG sources are estimated to be similar as for ozone [Fann et al., 2018]. However, ozone is currently the much more recognized regional pollutant.

Emissions of most primary air pollutants continue downwards trends in most of the United States. This also applies to surface ozone; implementation of pollution control measures has resulted in declining surface ozone across wide geographical scales in developed North American and European countries [Fleming et al., 2018]. For instance, the compilation of ozone trends shown in Figure 6 provides a nice testimony for decreases in surface ozone across the U.S. These downward trends are particularly remarkable in light of the population growth, increase in energy demand and production, and climate change, which is driving higher ozone production rates from the increase of ozone precursors and faster reaction rates in a warmer climate. Assessments in the magnitude of this effect vary by study. This ozone ‘climate penalty’ potentially can be rather significant, with some estimates predicting an up to 3–6 ppb increase in surface ozone per degree of temperature increase [Rasmussen et al., 2012].

Several studies have pointed out a decline of ozone precursor emissions from other source categories (non-O&NG) in the NCFR. Several publications have noted reductions in DMA NO_x, based on CDPHE NO_x surface monitoring data [Bishop and Stedman, 2008; Cooper et al., 2012; Abeleira and Farmer, 2017; Bien and Helmig, 2018]. DMA and NCFR declining NO_x trends are further confirmed by satellite imaging [Witman et al., 2014; Lamsal et al., 2015], and indirectly inferred from the diurnal ozone behavior [Bien and Helmig, 2018]. Because of the distribution of measurement sites, these analyses mostly reflect NO_x emissions in the DMA and not the entire NCFR, and emission reductions that have been achieved from automobiles and power generation plants.

Trends in VOCs are more difficult to assess. In downtown Denver, there is clear evidence that automobile-associated VOC emissions have been declining [Bishop and Stedman, 2008]. Currently, there are no other publications that have reported DMA or NCFR VOC trend analyses, and there is no peer-reviewed research that supports the State agency’s conclusion of “a dramatic decline in ambient levels of oil and gas related VOCs” (Supplemental Materials). Taken together, findings from these Colorado NO_x and VOC studies from non-O&NG sources mirror the national trend.

Large year-to-year variations in surface ozone causes trend analyses to be sensitive to the chosen time window. Trend behavior can differ substantially for different ozone metrics, i.e. summer versus annual ozone, different percentile values in the ozone distribution, and the MDA8 or the Design Value (the 3-year running mean of the 4^th highest annual MDA8).

Reddy and Pfister [2016] corrected the dependence of high summer ozone occurrences on the predominance of high pressure weather conditions in their investigation of the 1995–2013 NCFR ozone record. They report that these corrected, deemed more robust time series analyses, showed “…a general increase for the Front Range [MDA8] since 2004, broken only by the recession of late 2008”.

Lower ozone percentile values, reflecting mostly nighttime ozone, have clearly increased since 2000, most likely due to a weakening nighttime ozone sink from reaction with NO [Bien and Helmig, 2018]. The increasing low percentile/nighttime ozone values are possibly contributing to the observed increases in mean and median ozone. During 2000–2015, 10 out of 11 DMA/NCRF sites displayed a positive rate of change, with four out of those being statistically significant trends (p-value < 0.05) [Bien and Helmig, 2018]. Trends in the high percentile ozone values that are most relevant for health effects and regulatory considerations are more inconsistent. 2000–2014 Design Value time series plots for the DMA/NCFR sites Chatfield, Rock Flats North, South Boulder Creek, Fort Collins West, and the National Renewable Energy Lab in Golden [Bien and Helmig, 2018], suggest a behavior of gradually decreasing values; however, linear regression analyses do not result in statistically significant trends. For the 2000–2015 window, considering a total of 11 DMA/NCFR sites, and 28 linear regressions for summer ozone 95^th percentile values, MDA8, and Design Value trend analyses, 9 slope results were positive and 18 were negative. The two times higher negative values count may suggest a predominance of declining ozone behavior. However, the only statistically significant trend results (three) were all positive, indicating increasing ozone. Inclusion of 2016–2018 data in the ozone trend analysis indicates a steadily declining regional Design Value in the last seven years (Supplemental Materials).

The DMA/NCFR ozone behavior deviates from that of most other regions in the U.S. This is most evident in the summer daytime average ozone trends (Figure 6). While this ozone metric has clearly (at many sites with statistical significance) been heading downwards across the U.S., increasing values were determined for most sites in the DMA/NCFR. Persisting elevated ozone conditions were evident during 2018; ozone data collected by CDPHE in the NCRF were higher than in any of the previous five years, with a season maximum of 89 ppb and 32 exceedance days of the 8-hour 70 ppb NAAQS at the Boulder Reservoir site alone (see Figure 7 for the Boulder Reservoir July 2018 ozone record; additional exceedance days for the ozone NAAQS were recorded in all other months from May–September).

There are convincing arguments that support the conclusion that the deviation in the Colorado ozone behavior with the national trend is caused by emissions from the O&NG sector, both from O&NG signatures seen in elevated ozone episodes [Cheadle et al., 2017; Oltmans et al., 2019] and from photochemical modeling [Pfister et al., 2017a]. As already pointed out above, biogenic VOC emissions have a relatively minor contribution to regional ozone production; elevated ozone episodes are primarily associated with elevated anthropogenic VOCs [Cheadle et al., 2017; Zaragoza et al., 2017; Lindaas et al., 2019]. The effect of O&NG emissions on ozone production in the NCFR is exacerbated by the dominant summertime air circulation patterns that tend to transport pollution-enriched air from the DJB towards the foothills. Continuing ozone production in these accumulated air masses causes peak ozone values along the westerly parts of the plains stretching from Highlands, along Golden, Boulder, Longmont to Fort Collins and westwards a few miles into the mountain slopes. Boulder County is particularly vulnerable, being the closest and most directly downwind located area of the DJB. This conclusion was stated in the NCAR FRAPPE summary report: “On average, oil and gas emissions show a stronger influence in the northern part of the NFRMA and the northern foothills, while mobile emissions dominate farther south and in the southern foothills. Both sectors contribute, on average, 30–40% each to total NFRMA ozone production on high ozone days.”

Peer-reviewed literature is consistent in emphasizing that NCFR ozone exceedances are caused by the locally produced ozone that is added to the ozone background that is transported into the State. For Denver, this background is up to 14 ppb higher in comparison to other U.S. cities [EPA, 2008], which lowers the amount of ozone that can be added locally to reach exceedance of the standard. This margin is smaller than for other U.S. NAA, making meeting the standard more challenging. However, the local ozone production is mostly within the control of the State. Meeting the standard is calling for a concerted and aggressive effort in curbing regional ozone precursor emissions.

O&NG emissions are impacting air quality in the NCFR in multiple ways and at several scales. Exposures in close proximity arise from primary emissions. Current assessments indicate that the most concerning health impacts are from aromatic VOCs (BTEX), and for citizens living within a 1000 feet radius of wells and O&NG operations. On the order of 3.5 million Colorado residents live in the NCFR ozone NAA. Despite efforts to reduce ozone precursor emissions, and gains made in certain important emission sectors, including transportation and electrical power generation, the region is still subjected to an abundance of elevated ozone occurrences and exceedances of the NAAQS every year. This calls for concerted efforts for better characterizing emissions and air quality impacts of O&NG emissions and for emissions regulation. Specific recommendations are:

• – The lack of long-term NO_x monitoring within the DJB hampers the assessment of the contribution of O&NG emissions to regional NO_x. NO_x monitoring should be implemented at key locations upwind, within, and downwind of the DJB. More research is needed to better define NO_x point emissions from O&NG facilities. Remote sensing tools and data should be included in the evaluation of O&NG NO_x sources and emission trends.

• – Very little research has been done on evaluating and quantifying the contribution of O&NG emissions to atmospheric particulates. The prospect of 25–49 premature annual human deaths in Colorado from exposure to particles caused by O&NG emissions under current industry growth scenarios by 2025 [Fann et al. 2018] should motivate a concerted effort to investigate and better define particulates pollution, and to regulate particulates and secondary aerosol precursor emissions from the industry.

• – VOC data, mostly from occasional and campaign-type observations, as well as the CDPHE monitoring at Platteville, clearly show a strong contribution from O&NG operations on total VOCs and the ozone-producing VOC reactivity in the region. VOC monitoring is crucial for assessing O&NG air quality impacts. The current distribution of monitoring sites has a number of shortcomings for evaluating and monitoring changes of O&NG emissions. VOC monitoring is needed near operations to assess facility emissions and exposure risks of nearby residencies. This monitoring needs to be expanded to activities such as flowback, liquid unloading, and wastewater separation, which appear to be associated with high emissions and which have been mostly neglected or been underrepresented in previous assessments. In order to capture the high variability of these emission, this monitoring should be at high time resolution, ideally in real time. VOC monitoring needs to be tailored for characterizing emission trends, representative for a wide regional footprint. This can, for instance, be achieved by sampling at elevated sites or/and from inlets high above the surface, and best during mid-day to afternoon hours, when chances to sample mixed boundary layer air are highest. This monitoring would be most promising if it is conducted continuously, and at highest possible accuracy. Continuous, concurrent, and coordinated monitoring at strategically selected sites upwind and downwind of the DJB would allow assessing changes in basin-wide emissions.

• – VOCs emitted from O&NG sources constitute the majority of the OH reactivity in the DJB north of the DMA. These emissions contribute to a temporal and locally variable ozone production. Summertime elevated ozone occurrences show a high correlation to transport from O&NG extraction regions and atmospheric O&NG influences. Due to the ozone production dynamics and air circulation patterns, the daytime peak maximum ozone values are often observed along the NCFR foothills, tens of kilometers downwind of the O&NG emissions source regions, and thereby impacting communities outside of the production regions. These downwind air quality impacts from O&NG industries should be a strong consideration in the design of monitoring networks and decision-making on regulating existing and new O&NG development in the region.

• – Several independent measurements near O&NG operations have shown spikes with highly elevated concentrations of BTEX compounds that exceed health risks thresholds for nearby residents. Highest concentrations have been reported downwind of disposal facilities, rather than from well pads. Available data are mostly from short episodic measurements. This clearly demonstrates that characterization of BTEX emissions warrants more attention. This needs to include continuous monitoring and consideration of the diverse types of O&NG facilities. Thus far, health assessments have predominantly focused on well pads. Further research is needed on incorporating these other emissions sources given the growing body of literature showing their significant emissions and resulting elevated downwind concentrations.

• – Bottom-up inventories have large uncertainties, neglect temporal variation, and consistently appear to be lower than top-down emissions determinations. The increase in well sites, the size and number of wells per pad, changes in operational practices, and new regulations make bottom-up emission inventories an ever changing challenge. Inconsistencies between national and state inventories persist. There appears to have been little progress in improving agreement between bottom-up inventory estimates and top-down estimates during the past decade. Experimental tools for aircraft basin-wide top-down emissions determination have improved remarkably during the past five years. A concerted effort building on these capabilities by regularly (e.g. monthly) light aircraft profiling, could, within a short time frame, yield significant improvements of the basin-wide total emissions characterization.

• – Assessments of ozone contribution from O&NG emissions will have high uncertainty, and will under-predict the true ozone production as long as they rely on underestimated O&NG inventory emissions. Ozone impact studies need to be revisited with consideration of the most realistic NO_x and VOC emissions from O&NG industry sources.

• – Ozone pollution in the NCFR is well within the range where ecosystem impacts and production yield losses in agriculture are predicted. Given the size of the agricultural industry, and from available literature on ozone effects on crops, it is expected that the economic loss to the State’s farming industry from the O&NG-contributed ozone may be quite significant. A quantification of the actual revenue loss is needed for evaluating these adverse economic impacts of O&NG industry emissions.

• – There has been a remarkable growth in the number of peer-review studies on air quality impacts from O&NG emissions. Consideration of the findings from these resources, and closer communication and collaboration between state regulators and academic and federal researchers will likely be beneficial for directing Colorado’s O&NG policy development, guiding policy implementation, and for monitoring and assessing policy effectiveness.

Hélène Angot, William Neff, and Lisa McKenzie, all from the University of Colorado, provided text corrections and proofreading. The detailed comments from four anonymous reviewers and the handling editor were invaluable for the improvement of the manuscript. CDPHE provided monitoring data and network information. Publication fees for this article were covered by the University of Colorado Boulder Libraries Open Access Fund.

This research was in part funded by a contract from Earthworks, Durango, Colorado.

DH has no competing interests relating to the content of this publication. DH is one of Elementa Editors-in-Chief. He was not involved in the review process of this manuscript.

DH directed the study, reviewed the literature, and prepared the manuscript.