Ambient Nonmethane Hydrocarbon Levels Along Colorado ’ s Northern Front Range: Acute and Chronic Health Risks

: Oil and gas (O&G) facilities emit air pollutants that are potentially a major health risk for nearby populations. We characterized prenatal through adult health risks for acute (1 h) and chronic (30 year) residential inhalation exposure scenarios to nonmethane hydrocarbons (NMHCs) for these populations. We used ambient air sample results to estimate and compare risks for four residential scenarios. We found that air pollutant concentrations increased with proximity to an O&G facility, as did health risks. Acute hazard indices for neurological (18), hematological (15), and developmental (15) health e ﬀ ects indicate that populations living within 152 m of an O&G facility could experience these health e ﬀ ects from inhalation exposures to benzene and alkanes. Lifetime excess cancer risks exceeded 1 in a million for all scenarios. The cancer risk estimate of 8.3 per 10 000 for populations living within 152 m of an O&G facility exceeded the United States Environmental Protection Agency ’ s 1 in 10 000 upper threshold. These ﬁ ndings indicate that state and federal regulatory policies may not be protective of health for populations residing near O&G facilities. risk assessment results can be used for informing policies and studies aimed at reducing and understanding health e ﬀ ects associated with air pollutants emitted from O&G facilities.


■ INTRODUCTION
Horizontal drilling and high-volume hydraulic fracturing have resulted in a dramatic increase in the number of oil and gas (O&G) wells located on a single pad. 1,2 It is now common for O&G well sites to contain 20 to 40 wells, related infrastructure, and tank batteries to store and/or pipelines to transport petroleum products and exploration and production (E&P) waste. 3 Additional equipment and facilities, such as gathering lines, compressor stations, and E&P waste disposal sites may also be located in areas of intensive O&G development. 4 In the Denver Julesberg Basin (DJB) on the Colorado Northern Front Range (CNFR), the O&G industry is rapidly expanding at the same time that housing construction is increasing to accommodate a rapidly growing population. 5 As a result, 19% of the population (∼356 000 people) in the DJB live within 1600 m of an active O&G well site. 5 Between 2000 and 2012, the number of people living within 1600 m of an O&G well site grew almost 3 times faster than the population living further away. 5 Colorado mandated regulatory exclusion zones around residential structures in which the drilling of O&G wells is discouraged are referred to as setback distances. Colorado setback distances were historically as short as 150 feet (46 m) and are currently at 500 feet (152 m). 6 Additionally, setback distances of 1000 feet (305 m) apply to high occupancy buildings serving 50 or more people (e.g., schools and hospitals) as well as operating child care centers for 5 or more children. 6 While the setback distances are intended to protect the general public's safety and welfare from environmental and nuisance impacts resulting from O&G develop-ment, they are not intended to address potential human health impacts associated with O&G development air emissions. 7 Air pollution is one of the major potential health risks for populations living near O&G sites. 4,8 O&G sites directly emit nonmethane hydrocarbons (NMHCs) into the air, 9−11 and several studies have identified O&G facilities as major contributors to ambient NMHC levels along the CNFR. 12−17 Some of these NMHCs such as benzene, toluene, ethylbenzene, and xylenes (BTEX) are defined as hazardous air pollutants. 18 Because of higher atmospheric stability at night, nighttime emissions do not disperse as much as during the daytime, and average nighttime benzene levels are approximately twice daytime levels. 15 Recent Colorado studies observed that infants with congenital heart defects and children diagnosed with leukemia are more likely to live in the densest areas of O&G wells. 19,20 Studies in Pennsylvania and Texas have observed associations between proximity to O&G wells and fetal death, low birthweight, preterm birth, asthma, fatigue, migraines, and chronic rhinosinusitis. 21−25 The few previous human health risk assessments conducted in areas with O&G development have used results from ambient air samples to predict the risk for both noncancer and cancer health effects in the surrounding population. 8,26,27 Risks for noncancer health effects have been expressed as semi-quantitative hazard indices (HI), and cancer risks have expressed as risk in excess of the baseline lifetime cancer risk for Americans of 44 per 100 (lifetime excess cancer risk). 28 The previous assessments indicate the potential for short-term respiratory, neurological, hematological, and developmental effects and elevated estimated lifetime excess cancer risk for populations living within approximately 800 m of O&G well pads, 8 while the potential for chronic noncancer health effects and lifetime excess cancer risk are lower for populations living further from O&G sites. 8,26,29,30 One important shortcoming of previous risk assessments is that they did not consider shortterm and repeated nighttime peak exposures. They did not explicitly address childhood exposures or incorporate findings from the most recent studies on health effects associated with ambient benzene exposure. Additionally, data sets supporting most of the previous risk assessments were not sufficient for assessment of short-term exposures to air pollution O&G facilities or the variance in health risks with differing setback distances from O&G facilities.
The goal of this analysis is to characterize prenatal through adult noncancer and cancer health risks from both short-term (acute) and long-term (chronic) residential exposures to NMHCs measured in CNFR O&G development areas and how health risks vary with proximity to O&G facilities.   31 California's OEHHA guidance addresses developmental outcomes not fully covered in standard United States Environmental Protection Agency's (USEPA) risk guidance. 32,33 Specifically, OEHHA guidance has incorporated recent research findings on the developmental toxicity of ambient level benzene into their toxicity factors as well as a lifespan beginning in the third trimester of pregnancy into their cancer risk assessment. 31,34,35 Data Sources. We characterized risks for residential populations based on proximity to the nearest O&G facility, as recorded in the Colorado Oil and Gas Information System. 36 All samples included in this risk assessment were collected at plausible residential locations (i.e., at a distance greater that Colorado's historic 150 foot (46 m) setback distance from the nearest O&G facility) using NMHC measurement results from one of three CNFR studies conducted in the summer of 2014 ( Figure 1, Figure 2, Supporting Information): Study 1: 1-min canister samples (hereafter 1-min samples) and continuous air monitoring (1-s time resolution readings every minute: 60 consecutive minutes averaged to represent 1 h of exposure, hereafter 1-h samples) from the 2014 DISCOVER-AQ (Deriving Information on Surface conditions from Column and VERtically resolved observations relevant to Air Quality) field campaign. 37 The 1-min samples collected during the DISCOVER-AQ study targeted O&G, power generation, agricultural facilities, and high vehicle traffic areas. In situ methane, carbon dioxide, and carbon monoxide measurements were used to identify plume and background locations. In addition to residential plausibility-based setback distances, we reviewed Google Earth satellite images dated between June 2 and October 6, 2014 to determine if the sample location was appropriate for evaluating residential exposures. On the basis of these reviews, we excluded results from seven 1-min samples that had been collected at locations adjacent to large O&G processing plants or propane tank facilities because these locations were not plausible residential locations (Supporting Information Table S1). The 1-h samples were collected at a fixed site located in a rural area at 229 m from the nearest O&G facility.
Study 2: 72−96-h integrated canister samples (hereafter 72− 96-h samples) collected for the 2014 Boulder County air quality monitoring study. 38 All sites were located in residential areas. The site in western Boulder was selected as a reference site (i.e., with minimal O&G development influence) and was located on the grounds of the Boulder County Public Health offices near a busy intersection. Results from this location are subjected to urban and traffic influences and thus provide a representative urban signature. Twelve summa canister samples were collected at the Boulder site. The sites in Eastern Boulder County, along the border between Boulder and Weld County, were selected to assess the geographical gradients of NMHCs resulting from dense O&G development. The Eastern Boulder County sites were public facilities (a school, park, church, and fire station) located in residential areas, and other obvious sources of NMHCs were avoided. A total of 47 summa canister samples were collected from 4 sites in Eastern Boulder County (11−12 samples at each site). Study 3: 41 3-h integrated canister samples (hereafter 3-h samples) collected by the Colorado Department of Public Health and Environment (CDPHE) at a single site in a residential area of Platteville, CO located 247 m from the nearest O&G facility. 39 Data Assessment. Field measurement results were grouped by exposure scenario and sample type, as indicated in Figure 2. Because the underlying measurements were not normally distributed, all data were log transformed prior to statistical analysis. We evaluated the remaining measurements from the 1-min samples as well as the 1-h, 3-h, and 72−96-h samples for outliers using Q-Q plots. 39 No results were removed based on the outlier analysis. For results below the limit of detection, we substituted the limit of detection for statistical evaluations and calculations of mean concentrations. 40 Supporting Information Tables S2−S4 contain summary statistics and limits of detection for NMHCs included in the risk assessment.
We used analysis of variance (ANOVA) to compare means of specific NMHC concentrations between the four scenarios described below. We evaluated differences between mean NMHC concentrations with post hoc Tukey's studentized range (HSD) tests. Results were considered statistically different at an α of 0.05. All statistical analyses were conducted using SAS software version 9.3 (Cary, NC). 40 Exposure Assessment. Acute and chronic exposure estimates were developed for four scenarios based on Colorado regulatory setback distances between O&G wells and residential facilities of 152 m 6 and literature reference points of 1600 m: 5,41 (1) No O&G facilities within 1600 m. The nearest O&G facility from each sample location included well pads, tank batteries, E&P waste disposal sites, gathering lines, and processing facilities.
To estimate acute and chronic noncancer health hazards and cancer risks, we considered short and long-term exposures to NMHCs. For acute exposure, we used the maximum measured concentration of each NMHC from the 1-min samples to estimate the maximum 1-h ambient air concentration for all scenarios and compounds, except for benzene, toluene, and C9 aromatics in the 152−610 m scenario. For the 152−610 m acute scenario, we used maximum 1-h sample results for benzene, toluene, and C9 aromatics that were available only for this scenario. For chronic hazards and cancer risks, timeweighted average (TWA) mean (1-min and 3-h samples) over 24 h and mean (72−96-h samples) ambient NMHC concentrations were used to represent the average concentration and calculate a daily intake dose for each NMHC according to OEHHA Guidance (Supporting Information). 31 We calculated a TWA mean for the 1-min and 3-h samples because these samples were mostly collected in the daytime and do not represent nighttime concentrations. Continuous sampling results at the Platteville location indicate that the average mean benzene concentration from 19:00 in the evening to 7:00 the following morning is 2.34 times higher than between 7:00 and 19:00 h. 15 On the basis of these observations, eq 1 was used to calculate the TWA mean: Toxicity Assessment and Risk Characterization. For noncarcinogens, we expressed inhalation toxicity health-based factors as a reference concentration (RfC) in units of μg/m 3 (Supporting Information Table S5). We used OEHHA chronic reference exposure levels (REL) to evaluate long-term exposures of 8 or more years (OEHHA 2015). 31 If an OEHHA chronic REL was not available, we used the USEPA's risk screening level (RSL) for ambient air. 42 The OEHHA chronic RELs and EPA chronic RSLs are applicable to 24-h per day exposures over 10 to ≥12% of a 70-year lifespan (i.e., 7 to 8 years of exposure). 31,34,42 We used OEHHA acute RELs to evaluate acute 1-h exposures. 31 If an acute OEHHA REL was not available, we used the ATSDR's acute MRLs. 43 If RELs, RSL, or MRL were not available, RfCs were obtained from (in order of preference) EPA's Integrated Risk Information System (IRIS) subchronic RfCs, 44 EPA's subchronic Provisional Peer-Reviewed Toxicity Values (PPRTV), 45 or Health Effects Assessment Summary Tables. 46 We used surrogate RfCs according to EPA guidance for C 5 to C 8 alkanes and C 9 aromatic hydrocarbons, which do not have a chemical-specific toxicity value. 45 We derived semiquantitative noncancer hazard quotients (HQs), defined as the ratio between the estimated exposure concentration and RfC. 31,32 We summed HQs for specific NMHCs to calculate the HI. We also separated the HQs specific to neurological, respiratory, hematological, and developmental effects and calculated a separate end pointspecific HI for each of health effect. HQ's and HI's > 1 indicate that the estimated exposure exceeds the threshold exposure and the possibility of adverse health effects. 32 Because the HI is semiquantitative metric, it does not imply a multiplier. For example, an HI of 4 is not twice the risk of an HI of 2. For carcinogens, we expressed inhalation toxicity measurements as OEHHA inhalation cancer potency factors (CPFs) summarized in units of (mg/kg-day) −1 . 31 The lifetime excess cancer risk for each carcinogenic NMHC was derived per OEHHA guidance (Supporting Information). 31 We summed individual lifetime excess cancer risks for each NMHC to estimate cumulative lifetime excess risk. Risks are expressed as excess cancers over a lifetime per 1 million population based on exposure over 30 years, per OEHHA guidance, 34 which is consistent with the USEPA reasonable maximum exposure. 32 To estimate the population cancer risk for residential scenarios of less than 1600 m, we adjusted the cumulative lifetime excess cancer risk by subtracting the risk for populations with no O&G facilities within 1600 m of their home from the risk for populations living in closer proximity to facilities. We derived 2014 population estimates by adjusting the DJB 2012 population estimates at specific distances from   O&G wells for 2 years of population growth based on the average annual rate of growth between 2000 and 2012. 5 We then estimated the population risk for each scenario by multiplying the adjusted cumulative lifetime excess cancer risk by the estimated 2014 population estimate. 31

■ RESULTS AND DISCUSSION
For the 1-min samples (Figure 3), mean ambient BTEX and total alkane concentrations increased as the distance of the sample collection from the nearest O&G facility decreased (p < 0.001). The mean ambient benzene, toluene, ethylbenzene, total xylene and total alkane concentrations from the 1-min samples collected within 152 m of the nearest O&G facility were 41, 34, 35, 32, and 86 times higher, respectively, than the mean from 1-min samples collected further than 1600 m from the nearest O&G facility (p < 0.05, Figure 3). Supporting Information Table S6 presents the ANOVA results comparing selected NMHC mean concentrations as a function of distance to O&G facility.
For the 3-, 72-, and 96-h samples (Figure 3b and Table S6b), the TWA mean concentration from the 3-h samples collected in Platteville (247 m from nearest O&G facility) was compared to the mean concentration from the 72-and 96-h samples collected in Boulder (>1600 m from nearest O&G facility) and Eastern Boulder County (448−625 m from nearest O&G facility). Mean ambient benzene, ethylbenzene, and total xylene concentrations did not vary significantly between locations, although the lowest mean concentrations occurred in samples collected in Boulder. Mean ambient toluene concentrations in Platteville and Eastern Boulder County were twice the mean concentration in Boulder (p < 0.05). Mean ambient total alkane concentrations in Platteville were 1.3 and 1.8 times greater than those in Eastern Boulder County and Boulder (p < 0.05) (Figure 3b).
Noncancer Hazards. Figure 4a presents acute and chronic noncancer HQ and HI estimates based on maximum 1-min and 1-h sample results, and TWA mean 1-min sample results, respectively (Supporting Information Tables S7). Acute and chronic HQs and HIs increased with decreasing distance to the nearest O&G facility. Acute HIs for nervous system (18), blood system (15), and developmental effects (15) were >1 for residents living within 152 m of an O&G facility, and benzene and alkanes contributed more than 80 and 13%, respectively, to these end point specific HIs. Acute respiratory HIs were <1 for all scenarios. Chronic HIs for blood system (4) and developmental effects (4) were >1 for residents living within 152 m of an O&G facility, and benzene contributed to 84% of the HI (Supporting Information Tables S7). Acute and chronic HIs for all effects were <1 at distances greater than 152 m from O&G facilities.  Table S8). Total chronic HIs were highest in Eastern Boulder County, where the total HI was 1.1, followed by Boulder, and then Platteville. Nervous system, respiratory system, blood system, and developmental HIs were <1 for all locations.
The chronic HIs based on 72−96-h samples are 2−10 times greater than those reported for residential exposures to NMHC in O&G areas in previous risk assessments 8,26,27 primarily because the OEHHA chronic REL for benzene (3 μg/m 3 ) is 10 times less than the USEPA's chronic RfC (30 μg/m 3 ). The OEHHA chronic benzene REL 34 considers several studies published after USEPA's 2002 benzene assessment, 47 which found increased efficiency of benzene metabolism at low doses, 48−51 decreased peripheral blood cell counts at low doses (800−1860 μg/m 3 ) with no apparent threshold, 52−54 and large population variation in the response of metabolic enzymes involved in benzene activation and detoxification. 55 Lifetime Excess Cancer Risk Estimates. Figure 5 presents lifetime excess cancer risks based on daily inhalation intake dose estimates calculated from 1-min, 3-h, and 72−96-h sample results (Supporting Information Tables S9 and S10). All cumulative lifetime excess cancer risks exceeded USEPA's de minimus benchmark of 1 in a million 58 with benzene representing more than 95% of the total risk estimate for all scenarios. The cumulative lifetime excess cancer risk increased with decreasing distance to the nearest O&G facility. For  Table S9). The cumulative lifetime excess cancer risk was higher in Eastern Boulder County and Platteville than in Boulder and reached the USEPA's upper bound risk level of 1 in 10 000 in Platteville (nearest O&G facility 247 m) based on the 3-, 72-, and 96-h sample results (Supporting Information  Table S10). For similar scenarios, the cumulative lifetime excess cancer risks based on mean 72−96-h sample results are greater than risks based on TWA mean 1-min and 3-h sample results.
These lifetime excess cancer risk estimates are 10−100 times greater than those reported in previous risk assessments in O&G development areas that used USEPA guidance. 8,26,27 This is partly because the OEHHA inhalation benzene CPF (0.1 (mg/kg-day) −1 is 4 times higher than USEPA's benzene slope factor (0.027 (mg/kg-day) −1 . 31,44 The OEHHA approach addresses methodological shortcomings in the derivation of USEPA's current slope factor, which was calculated with a linear extrapolation model that assumes excess risk is proportional to the lifetime average exposure, is the same for all ages, and does not explicitly address the impact of episodic exposure peaks. 57 OEHHA's inhalation CPF was calculated using a weighted cumulative exposure/relative risk procedure that assumes with continuous exposure, age-specific cancer incidence continues to increase as a power function of the elapsed time since the initial exposure. 35 Additionally, OEHHA includes prenatal exposures in the calculation of lifetime excess cancer risk based on recent studies indicating increased susceptibility to benzene in early life. 35 Even using USEPA's current slope factor, which would reduce the lifetime excess cancer risk from benzene for residents living within 152 m of an O&G facility to 2.2 in 10 000, our results remain above USEPA's 1 in 10 000 upper bound for remedial action.
Overall Strengths and Limitations. We assessed acute and chronic health risks from air pollution associated with O&G operations using data collected in close proximity to O&G facilities and realistic residential scenarios tied to regulatory setback distances and literature reference points. This approach allowed us to incorporate proximity, spatial variability, and temporally relevant sampling durations into our exposure scenarios. The consistent application of exposure and toxicity parameters for all four scenarios allows for the comparison of hazards and risks between the scenarios. We found increasing (1) hematological and developmental HIs and (2) cumulative lifetime excess cancer risks with decreasing distance to the nearest O&G facility. These results are consistent with findings from observational epidemiological studies that indicate an increased likelihood of adverse birth outcomes and childhood acute lymphocytic leukemia with increasing proximity to O&G wells. [19][20][21]23,24 The weight of evidence increasingly suggests that plausible outcomes to explore in future epidemiological studies of exposure to O&G-related pollutants include: neural tube defects, 58 changes in blood cell and platelet counts and aberrant nucleic acid methylation patterns, 59−61 and increased levels of 8-hydroxydeoxygunaosine, a biomarker of short-term nucleic acid damage. 62−64 However, the uncertainties in our risk assessment are substantial, and the results are best suited for scoping policy and future studies. Some of our assumptions are inherent in the risk assessment process (Table 1), while others are more specific to this study. Exposure to benzene had the largest contribution our predicted health risks. While there is extensive evidence that occupational benzene exposure is linked to leukemia, the evidence in nonoccupational populations is less robust. Nonetheless, the body of literature suggesting that exposures to ambient levels of benzene are associated with incidence of childhood leukemia is increasing. Additionally, recent studies that were included in the derivation of the RELs for toluene in this risk assessment suggest that low dose toluene exposure can alter fetal and adult testosterone levels. 65−67 Reductions in testosterone and mRNA 3B-hydroxysteroid dehydrogenase levels were observed in male fetal rats at low toluene exposures. 65 Using the lowest observed adverse effect level of 3400 μg/m 3 from that study results in chronic and acute toxicity values of approximately 300 and 30 000 times lower, respectively, than current toxicity values, resulting in a corresponding increase of acute and chronic toluene HQs > 1 for the <610 m exposure scenarios.
Ideally, chronic HIs and cancer risk are estimated from a large number of air samples that represent 24-h exposure. In the summer of 2014, only the 72-to 96-h samples collected in Boulder and Eastern Boulder County captured full 24-h exposure periods. Because 24-h data was not available to estimate chronic HIs and cancer risks for populations living in close proximity to O&G facilities (i.e., 305 m), 5 we used 3-h and 1-min measurements as they were the best available to estimate 24-h exposures for this scenario. However, the 1-min and 3-h samples were collected mostly during the daytime and do not represent the contribution of what are likely higher nighttime ambient NMHC concentrations, 15 even after the TWA adjustment. This is likely the reason our chronic cumulative HIs and cancer risks based on mean 72−96-h sample results are greater than those based on TWA mean 1min and 3-h sample results for similar scenarios. This also indicates that the HIs and cancer risks calculated from the 3-h and 1-min samples are likely not overestimated because they are an empirical estimate of local short-term concentrations of these compounds.
Our findings are based on ambient air samples collected in the summer of 2014 that may not capture temporal variations in NMHC concentrations associated with O&G activities. For example, NMHC concentrations likely differ by season and will vary in the future as O&G emission control technology evolves. Existing studies suggest that winter levels of these pollutants are higher because longer nights and cold daytime temperatures keep the atmospheric boundary layer lower than that in summer and thus increase NMHC concentrations near the surface. 12 Multiple uncertainty factors applied in derivation of the reference concentrations. 4 The lifetime excess cancer risk assumed that residents spend 72−85% of their time at home over a 30 year period. 5 Reference concentrations that were mostly derived from occupational studies on adults or animal toxicity studies may not adequately represent the current understanding of developmental and reproductive effects. Finally, exposure to other air pollutants, drinking water contaminants, and nonchemical stressors (e.g., noise) associated with upstream O&G operations could further contribute to the health risks estimated in this study. For example, alkanes emitted from O&G operations contribute to approximately 20% of summertime photochemical ozone production along the CNFR, 69 and each 10 ppb increase in ozone may result in an 2% increase in mortality. 70 Ozone levels in several CNFR cities exceed National Ambient Air Quality standards. 71 To better understand health risks from air pollution originating from O&G operations, systematic ongoing sampling of NMHCs (especially BTEX), source tracers, and other air pollutants such as formaldehyde, acetaldehyde, and polycyclic aromatic hydrocarbons 72,73 associated with O&G activities is needed. Future research should focus on providing measured and modeled exposure estimates of key risk drivers (e.g., BTEX) for populations living near O&G operations. 74

■ POLICY IMPLICATIONS
Our results indicate that State regulatory setback distances (the minimum distance an O&G wellhead may be located from a home) and reverse setback distances (the minimum distance a home may be located from an O&G wellhead) and related municipal codes may not protect nearby residents from health effects resulting from air pollutants emitted from these facilities. Setback distances between homes and other types of O&G facilities (e.g., tank batteries, waste disposal sites, gathering lines, compressor stations, etc.) have not been specified, 6 and very few municipal codes regulate the siting of homes near existing O&G well sites. 5,75,76 We found that Colorado populations within 152 m of an O&G facility are more likely to experience neurological, hematological, and developmental health effects from acute inhalation exposures to benzene and alkanes. We also estimated cumulative lifetime excess cancer risks for populations living within 610 m of an O&G facility exceed USEPA's upper threshold of 1 in 10 000.
Sources of air pollutants other than O&G facilities (e.g., non O&G related traffic) likely partially contributed to the health risks for all exposure scenarios. Nonetheless, our results indicate that air pollutants from O&G facilities increasingly contribute to the health risks as the distance from the nearest O&G facility decreases. For the more than 380 000 people along the CNFR estimated to be living within 1600 m of an O&G well in 2014, we estimate an additional 17 to 27 cases of cancer over a lifetime (70 years). We estimate that more than 50% of these additional cancers may occur in the population living within 152 m of an O&G facility (Supporting Information  78 Our results provide further justification for implementation of these regulations because such regulations could also reduce emissions of hazardous air pollutants such as benzene that drive our risk estimates. However, our risk results include both older and smaller O&G facilities, and other studies indicate that older and smaller O&G facilities could emit significant levels of air pollutants that are important contributors to health risk. Because Colorado Regulations 6 and 7 are less stringent for older wells, controls on small emitters (<6 tons VOCs per year), and temporary frac tanks, 78 they may not be sufficiently protective for residents living near these facilities. The NSPS OOOO is applicable only to wells built or modified starting in 2015; the current United States administration is seeking delay of implementation, and the O&G industry is challenging the federal regulation in court. 77,79 A study in the Marcellus shale found that older well sites generally had much high production normalized methane emission rates than newer multiwell sites because of a range of issues, including a lack of maintenance. 80 Small emitters outside the nonattainment area for zone are required under Regulation 7 to conduct monthly audio, visual, and olfactory (AVO) inspections, and they are currently required to conduct only one inspection with an approved instrument monitoring method (e.g., an infrared [IR]camera capable of detecting hydrocarbon and VOC emissions) over the lifetime of the facility. 81 Additionally, the regulations exempt storage tanks from emission inspections and small emitters from emission's management plans. 81 A recent analysis conducted with an infrared camera in Boulder County, CO detected gas leaks at 65% of 145 inspected O&G sites, most of which were small emitters developed prior to 2011. Ninety-two percent of the detected gas leaks were at storage tanks, separators, or wellheads, and at least 31% involved pneumatic devices or equipment associated with them. 82 Several studies have implicated storage tanks, thief hatches, and pneumatic devices as major sources of VOC emissions from O&G facilities. 12,15,83−85 The Boulder County analysis found that only 2.5% of AVO inspections identified a gas leak, compared to 66% of inspections using an IR camera, indicating that in this program monthly AVO inspections are not effective in detecting VOC emissions from leaking equipment. 82 Intermittent and continuous bleed pneumatic devices may be a significant source of emissions even when operating properly. 86 Zero-bleed pneumatic devices could also significantly reduce emissions. 86 Ultimately, this means that the large number of old, low producing, and insufficiently inspected O&G operations could be a significant source of air pollutant emissions and related health risks.
While the magnitude of air pollutant emissions may vary by region, these results also have implications for policies in other O&G regions where homes are located near O&G facilities 75 This study provides further evidence that populations living nearest to O&G facilities bear the greatest risk of acute and chronic health risk from exposures to NMHC air pollutants emitted from upstream O&G facilities. Therefore, this analysis supports and highlights the importance of policies aimed at reducing or eliminating air emissions from O&G equipment and facilities, particularly those near homes, and effective monitoring of emissions from these facilities.