Proposed emission control strategies for reducing ozone and particulate matter are evaluated better when air quality and health effects models are used together. The Community Multiscale Air Quality (CMAQ) model is the US Environmental Protection Agency’s model for determining public policy and forecasting air quality. CMAQ was used to forecast air quality changes due to several emission control strategies that could be implemented between 2008 and 2030 for the South Coast Air Basin that includes Los Angeles. The Environmental Benefits Mapping and Analysis Program—Community Edition (BenMAP-CE) was used to estimate health and economic impacts of the different emission control strategies based on CMAQ simulations. BenMAP-CE is a computer program based on epidemiologic studies that link human health and air quality. This modeling approach is better for determining optimum public policy than approaches that only examine concentration changes.
Introduction
Ground-level ozone (O3) and particulate matter (PM) are air pollutants that adversely affect human health.1234–5 Exposure to air pollutants is of particular concern in urban areas because of the highly dense populations exposed to air pollution and the high number of emission sources. Ground-level ozone and PM cause respiratory and cardiovascular health problems. Short-term respiratory impacts range from reduced lung function, coughing, and throat irritation to asthma attacks.6 Longer-term impacts of air pollution include chronic obstructive pulmonary disease (COPD) and cardiovascular health disease. Particles smaller than 2.5 µm (PM2.5) have the most significant health effects. PM2.5 can be inhaled very deeply into the lungs, enter the bloodstream, and lodge in the heart and brain where they can cause the formation of plaques.7
Ground-level ozone not only affects human health but also damages crops, forests, and structural materials.8 Extensive research91011121314–15 has demonstrated the associations between exposure to common air pollutants (PM and O3) and ill-health end points of respiratory and cardiovascular diseases (morbidity and mortality).
Particles that are emitted directly from sources are known as primary PM. Particulate matter emitted from combustion sources includes soot, black carbon, and elemental carbon. Organic compounds found in primary PM are likely to consist of minimally oxygenated organic compounds.1617–18 Other primary PM includes windblown crustal material, resuspended particles from tire and brake wear and fugitive dust from industrial and agricultural activities. Primary PM particles tend to be larger than particles formed in the atmosphere by atmospheric chemistry. Larger particles with effective diameters from 2.5 to 10 µm are designated coarse PM and typically account for most of the total mass of the aerosol measurements known as PM10. PM10 is less of a health problem than PM2.5 because it does not travel as deeply into the lungs, and coarse PM particles settle out relatively rapidly from the atmosphere.
Emissions of nitrogen oxides (NOx = NO + NO2) and volatile organic compounds (VOC) play an important role in air pollution chemistry. The combustion of fossil fuels by automobiles, electric power plants, industrial boilers, refineries, and other sources including chemical plants and painting facilities emits NOx and VOC.19 Strategies designed to improve air quality focus on reducing NOx and VOC emissions.
In the lower troposphere ozone and a significant fraction of atmospheric particles are formed by chemical reactions of NOx and VOC. Ozone and PM formed in the atmosphere are known as secondary pollutants. Secondary PM particles typically are PM2.5.20 There is much overlap between the atmospheric chemistry that produces O3 and PM2.5, and therefore, planned air pollution control strategies for reducing O3 may affect PM2.5 concentrations and vice versa. Formation rates and concentrations of O3 and secondary PM2.5 are highly nonlinear functions of NOx and VOC concentrations and their VOC/NOx ratio.2122–23 For example, reductions in NOx concentrations can lead to higher or lower O3 concentrations depending on specific atmospheric conditions. Hydroxyl radical (HO) reactions with nitrogen dioxide (NO2), sulfur dioxide (SO2), and VOC are major sources of gas-phase production of nitrate, sulfate, and secondary organic aerosols (SOA). The concentrations of HO are highly nonlinear functions of NOx and VOC concentrations.24 Another important factor in determining inorganic secondary aerosol concentrations is ammonia concentration. The extent of the reactions that produce aerosol particles consisting of NH4HSO4, (NH4)2SO4 and NH4NO3 depends on the relative concentrations of HNO3, H2SO4 and NH3, relative humidity, and temperature.252627–28
Meteorology is another very important factor that affects concentrations of O3 and PM.23 Summer days that are warm, clear, and with calm winds are associated with stagnant high-pressure conditions. On these days, NOx and VOC concentrations increase in a polluted urban atmosphere and the resulting photochemistry increases O3 concentrations. However, cool fall days are associated with low mixing heights and cooler temperatures that promote the condensation of ammonium nitrate particles from gas-phase ammonia and nitric acid. These conditions may cause dangerously high concentrations of PM.5,2627–28 High O3 and PM days are associated with health effects that are observed particularly in sensitive populations.
Air quality modeling systems that combine meteorology, emission source modeling, and atmospheric chemistry are used to predict the concentrations of air pollutants that include ozone and PM, and more recently they have been used with new software to assess health effects.22,29 The research presented in this article focuses on the health effects of planned emission reductions in the South Coast Air Basin (SoCAB; Los Angeles and its surroundings). The emission reductions will affect ozone and secondary PM concentrations. Potential health benefits resulting from different emission reduction scenarios are estimated.
Air quality and health effects
Epidemiologic studies have shown that there are a large number of adverse health effects associated with ozone pollution and PM (Table 1).
Table 1.
Epidemiologic studies that have been used to estimate health effects used to develop the Environmental Benefits Mapping and Analysis Program—Community Edition used in this project.
The table shows several epidemiologic studies that were used to develop the Environmental Benefits Mapping and Analysis Program—Community Edition, version 1.1 (BenMAP-CE) used in this study.38,39 BenMAP-CE is a computer application that is widely used by federal, state, and local agencies to assess the relative benefits of air quality policies. It is a powerful open source program that can be enhanced or customized to the needs of specific users.
Health impact functions that quantitatively relate health effects to exposure to PM2.5 and ozone were developed from epidemiologic studies. The developers of BenMAP-CE assumed that the natural logarithm of the incidence of a health effect is linearly related to the concentration of an air pollutant (Figure 1). The figure shows a hypothetical relationship between the concentration of an air pollutant and the incidence of some adverse health effect that was derived from epidemiologic studies.38 The slope, β, of the linear regression line relates the pollutant concentration to the incidence of an adverse health effect.
Figure 1.
A hypothetical relationship between the concentration of an air pollutant and the incidence of an adverse health effect.
The slope, β, is used in a health impact function to make a health effect estimate, ΔY40:
The term Δ[X] is an air quality change and it is the difference between a chosen baseline air pollutant concentration and a concentration at some future year. In this study, PM and ozone concentrations during selected episodes that occurred during the year 2008 were used as the baseline concentrations, and therefore, 2008 will be referred as the base year in this article. The selected future year in this study is 2030. The same set of episodes with the 2008 meteorological conditions but with differing emission inventory scenarios for the year 2030 were modeled as discussed below. The air quality change, Δ[X], is the differences between the simulated future year and base year concentrations.38
The exposed population, Pop, is the number of people in a region under consideration, and in this study, it was the SoCAB. BenMAP-CE has estimated population data sets from the year 2000 to 2040 that are derived from US Census data. BenMAP-CE can also estimate population data for different grid definitions (such as city and state) based on spatially weighted averages of original census data.41
The health baseline incidence rate, Yo, is the average number of people who die (or have another adverse health effect) in the population under consideration over a given period of time. One example of a health incidence rate is the probability that a person will die in a given year.38
BenMAP-CE calculates the economic value of a change in air quality from the health effect estimate, ΔY.38 The economic value estimate, V, is the product of the health effect, ΔY, and the benefit value or cost, C, of the health effect. The health effect estimate and economic value will vary from place to place across a region such as the SoCAB due to spatial variations in air pollutant concentrations and population:
The economic values for avoided mortalities are estimated according to the value of a statistical life based on 26 studies that included all ages.38 The economic value for avoided emergency room visits and hospital admissions are calculated based on the cost of illness. Cost of illness includes the cost of medical treatment and the value of lost wages. The costs of medical treatment are estimated from the average length of a hospital stay for illnesses associated with poor air quality and average hospital daily charges. Wages lost are estimated based on the average length of a hospital stay per illness and the average daily wage, regardless of whether or not a particular individual is working.39 The economic values associated with acute respiratory symptoms are estimated based on an individual’s willingness to pay to avoid illness.
Methods
CMAQ model
The Community Multiscale Air Quality (CMAQ) model is an air quality model developed by the US Environmental Protection Agency (EPA) to simulate the formation of air pollutants, their concentrations, and distributions over spatial scales that range from urban to continental.42,43 CMAQ takes meteorological fields, chemical initial and boundary conditions, and emissions as input and uses these to simulate time-dependent, 3-dimensional distributions of air pollutants such as O3 and PM (Figure 2).
CMAQ (version 4.7) was used to make the O3 and PM simulations for this project. The aerosol module simulates both PM2.5 and PM10. It simulates primary emissions of elemental and organic carbon, dust, and unspecified species. PM10 particles include windblown dust and marine particles such as sea salt. It simulates secondary PM including sulfate, nitrate, ammonium, and SOA that are formed from anthropogenic and biogenic emissions. CMAQ’s aerosol module was developed from the Regional Particulate Model (RPM),44 and the RPM was derived from the Regional Acid Deposition Model, version 2 (RADM2).45 The aerosol module in CMAQ represents the aerosol size distribution using a mode approach. The mode approach uses a superposition of 3 lognormal subdistributions (modes) to represent the aerosol size distribution.46
CMAQ includes important physical processes such as coagulation. Differential equations for the number and species mass conservation are represented by analytical solutions to increase numerical accuracy. A very serious weakness in CMAQ aerosol module (and most other similar air quality models) is that it calculates PM concentrations somewhat independently of CMAQ gas-phase chemistry module. This is inconsistent with real atmospheric processes and it is a limitation on the accuracy of this and similar air quality modeling research.
Typical EPA-State Implementation Plan protocols for the development O3 control strategies involve the selection and modeling of short-term episodes with high O3 concentrations. The selected episodes have representative meteorology and emissions and have high-quality measurements. An analogous protocol was followed for this study of PM. Two episodes with high PM2.5 concentrations, representative meteorology and representative emissions, were selected for CMAQ modeling. The selected episodes had good-quality PM2.5 data to allow the simulations to be properly evaluated. The first episode began on Friday, September 12 and ended Monday, September 15, 2008, and the second episode began on Tuesday, November 11 and ended Monday, November 24, 2008. The selection of episodes for the base year is discussed in more detail elsewhere.47,48
Base year modeling and sensitivity simulations
CMAQ model was used to make simulations of the selected high PM episodes for the base year. The South Coast Air Quality Management District (SCAQMD) provided the meteorological simulation data and emissions inventory required for CMAQ simulations.49 The SCAQMD profile files for initialization were used.47,48 The modeling protocol included also running CMAQ with a 4 km × 4 km spatial grid resolution; the gas-phase chemistry mechanism was SAPRCC99 (with about 400 species of VOC) and the aerosol module with the 5-aerosol module with saprcc99_ae5_aq. A total of 18 days were simulated and analyzed (September 12-15, 2008 and November 11-24, 2008). A couple of extra days were simulated before each episode, and these extra days were regarded as “spin-up time” to fully initialize CMAQ.
Simulations were made to estimate the effect of changes in the VOC emission inventory on secondary PM2.5 for the base year and the effect of 6 NOx emission reduction scenarios on PM2.5 for the future year (Table 2). The same meteorology that occurred during the selected high PM episodes for the base year was used to make additional CMAQ simulations with different emissions’ inventory scenarios for the future year 2030. The target future year based on current SCAQMD emission reduction plans is scenario R′1N′1. The target reduction between 2008 and 2030 emissions for NOx is 61%, and for VOC, the reduction is 32%. The target NOx emissions for 2030 were multiplied by factors of 2.5, 1.75, 1.0, 0.75, 0.5, and 0.3 to create cases 1R′2.5N′, 1R′1.75N′, 1R′1N′, 1R′.75N′, 1R′.5N′, and 1R′.3N′, respectively. The VOC emissions were kept at the same levels as the target reduction in 32% for 2030.
Table 2.
2008 and 2030 model simulations with baseline total VOC and NOx emissions and sensitivity cases with varying adjustments to baseline emissions.
The Environmental Benefits Mapping and Analysis Program—Community Edition
BenMAP-CE (version 1.1) was used in this study to calculate health effect estimates and their economic value due to concentration changes in ozone and PM. As discussed above, epidemiologic studies were used to derive health impact functions and these functions are implemented in BenMAP-CE program. A user specifies air quality concentrations for a base case and concentrations that are expected after an emission control strategy is implemented. Concentration differences are calculated between the base year and a 2030 future year control strategy scenario. BenMAP-CE uses the concentration changes along with population data and health impact functions to estimate the health impact due to the pollutant concentration changes.
Three major steps are required for BenMAP-CE to estimate the health and economic benefits associated with changes in air quality (Figure 3). The first step is to create 2-dimensional air quality surfaces for a base case and a control strategy case. During the second step, BenMAP-CE is used to estimate the health impact due to the change in air quality between the 2 cases. In the third step, BenMAP-CE is used to estimate the economic value of the health impact.38,48,50 The base year case (1R1N) was used as baseline for all analysis made to evaluate the health and economic benefits of the 6 future year NOx and VOC emission control scenarios.
Figure 3.
Steps and the input data files required to run BenMAP-CE (Environmental Benefits Mapping and Analysis Program—Community Edition).38
Air quality surfaces were created for both O3 and PM2.5 based on the difference between the base year and the future year pollutant concentrations for the SoCAB. Air quality surfaces were created for each of the 6 future year simulations; a set of 6 surfaces for PM2.5 and 6 for O3 was created. These air quality surfaces were then used to make health impact estimates for the end points involved with premature death, respiratory disease, and cardiovascular disease (Table 1). The impacts due to O3 and PM2.5 exposure, were examined separately in this project, and the impacts included avoided mortalities, emergency room visits due to asthma, hospital admissions (due to both respiratory and cardiovascular causes), and acute respiratory symptoms.
In this study, the spatial region used was the SoCAB, and the population was the predicted population of the SoCAB region in the year 2030. During each BenMAP-CE run, the program was set to do 10 000 Monte Carlo simulations when estimating health outcomes. Finally, the economic value of the health impacts for each of the 6 future year sensitivity cases was calculated for both O3 and PM2.5 exposure.
Results
CMAQ model base year simulations and evaluation
Simulations of the 2 high PM episodes were evaluated together by comparing simulated PM2.5 and other calculated components with SoCAB measurements data. The details of the evaluation are provided elsewhere.47,48 The simulated concentrations of PM2.5 were low as shown by the normalized mean error (NME) and the normalized mean bias at 6 PM2.5 measurement sites in the SoCAB (Figure 4). The 2008 base year simulations all displayed negative mean bias. Eder and Yu51 suggest that acceptable model performance for PM simulations is a NME within 50%. However, there is no generally accepted statistic for acceptable model performance for PM modeling.
Figure 4.
Normalized mean error (upper line) and the normalized mean bias (lower line) are shown for 6 PM2.5 measurement sites in the South Coast Air Basin (Los Angeles [CELA], Long Beach [LBCH], Anaheim [ANAH], Azusa [AZUS], Fontana [FONT], and Rubidoux [RUBI]) and all 6 sites combined.
The NME was within 50% for 4 of the 6 stations with Fontana being close. However, CMAQ simulations for Azusa did not meet this NME standard. Considering all stations, the normalized error was 47.25% and that was within the suggested standard.51
CMAQ model future year simulations
Average 8-hour O3 plots were made for the fall aerosol season for the future year emission cases (Figure 5). For these cases, lower NOx emissions led to greater the O3 concentrations across the SoCAB. The highest reduction in 2030 NOx emissions from the base year led to the highest concentrations of O3 across the basin. This is probably due to the relatively low rate of nitrogen dioxide photolysis during the fall. Because this is the source of O3, a lower production rate increases the importance of O3 titration by NO. If there is limited NO available to destroy O3, there will be an increase in O3 concentrations with lower NOx emissions.
Figure 5.
Contour plots showing 8-hour average O3 concentration due to adjustments in 2030 NOx emissions over Southern California during the aerosol season. The red box outlines the future year target emission scenario. (A) 1R′N′2.5, (B) 1R′N′1.75, (C) 1R′N′1, (D) 1R′N′0.75, (E) 1R′N′0.50, and (F) 1R′N′0.30.
CMAQ simulated 24-hour average PM2.5 concentrations over the SoCAB for the future year on November 20, 2030 during the Fall-Winter Aerosol Season are presented (Figure 6). Plot C is the simulation with the projected 2030 NOx emission inventory and over the basin shows the highest 24-hour average PM2.5 concentrations. Plots A and B show simulations with the projected 2030 NOx emission inventory multiplied by factors of 2.5 and 1.75, respectively. Between plots A, B, and C, the simulated 24-hour average PM2.5 concentrations decrease with increasing emissions. Plots D, E, and F show simulations with the projected 2030 NOx emission inventory multiplied by factors of 0.75, 0.50, and 0.30, respectively. Over the sequence between C and F, the simulated 24-hour average PM2.5 concentrations decrease with decreasing NOx emissions. Therefore, the projected 2030 NOx emission inventory is the least effective in reducing 24-hour average PM2.5 concentrations than any of the other NOx emission inventory scenarios simulated.
Figure 6.
Community Multiscale Air Quality–simulated 24-hour average PM2.5 concentrations over the South Coast Air Basin for the future year on November 20, 2030 during the Fall-Winter Aerosol Season. The plot outlined in red shows the plot associated with the emission reductions in VOC and NOx planned for the year 2030. The other plots are simulations with the NOx emissions multiplied by the factors given about the plots and to the right of the N′ labels. The VOC emissions were not varied. (A) 1R′N′2.5, (B), 1R′N′1.75, (C) 1R′N′1, (D) 1R′N′75, (E) 1R′N′0.5, and (F) 1R′N′0.3. VOC indicates volatile organic compounds.
The mass concentrations of ammonium nitrate show the greatest sensitivity to changes in VOC and NOx emissions. Simulated component compounds at Anaheim are shown as an example (Figure 7). Ammonium nitrate mass concentrations increase as VOC emissions are increased for the base year. For the future year, the ammonium nitrate mass concentration is greatest for the target 2030 NOx emission case N′1. For cases N′2.5 and N′1.75, the ammonium nitrate mass concentrations are less than the N′1 case and the ammonium nitrate mass concentrations decrease with increasing NOx emissions. Cases N′.75, N′.50, and N′.3 all have lower NOx emissions than N′1, and for these cases, there are lower ammonium nitrate concentrations as NOx emissions are reduced.
Figure 7.
Stack plot of simulated particulate matter mass concentrations by component for the 2008 base year and the future year at Anaheim. The components are primary organic matter (primary OM), anthropogenic secondary organic aerosol (anthropogenic SOA), biogenic secondary organic aerosol (biogenic SOA), elemental carbon (EC), ammonium nitrate (AmNit), ammonium sulfate (AmSul), unidentified PM2.5 (unid PM2.5), and coarse mass.
Comparison of the PM2.5 components shows that the greatest change between the base year (BASE) and the target future year NOx emissions scenario (N′1) was in the ammonium nitrate mass concentration (Figure 7). The ammonium nitrate mass concentration increased in the future year.
Health and economic benefits associated with the response of PM2.5 and O3 to adjustments in VOC and NOx emission between the base and future year
Health burden for the base year scenario
The concentration of PM2.5 and O3 from the base case (1R1N) CMAQ simulation was used as the baseline air quality case for the SoCAB. The health and economic burdens associated with the 2008 baseline PM2.5 and O3 concentrations in all 4 counties in the SoCAB were estimated. According to the State of California’s Department of Public Health, there were 11 189, 13 137, 16 179, and 55 108 mortalities, respectively, in San Bernardino, Riverside, Orange, and Los Angeles County in 2008. Mortalities caused from accidents and intentional self-harm or suicide were not included in the total deaths presented here because these deaths are not associated with poor air quality.
Using BenMAP-CE, we estimate at the 95% confidence interval (CI) that in 2008 exposure to PM2.5 was responsible for 131 (CI: 89-173), 434 (CI: 296-569), 409 (CI: 278-539), and 781 (CI: 596-1031) deaths, respectively, in San Bernardino, Riverside, Orange, and Los Angeles County in 2008 (Table 3). Exposure to O3 was responsible for 301 (CI: 208-392) deaths in San Bernardino, 287 (CI: 199-375) deaths in Riverside County, 346 (CI: 234-450) deaths in Orange County, and 1356 (CI: 938-1765) deaths in Los Angeles County (Table 4). These results taken together indicate that O3 and PM2.5 exposure in 2008 was responsible for at least 3.86%, 5.48%, 4.67%, and 3.88% of the total 2008 deaths in San Bernardino, Riverside, Orange, and Los Angeles County, respectively.
Table 3.
Estimated PM2.5-related health burden for the base year scenario.
Table 4.
Estimated O3-related health burden for the base year scenario.
Avoided premature mortalities
The health impacts resulting from the change in PM2.5 concentrations between the base year and the future year with the target 2030 VOC and NOx emissions (1R′N′1 case) resulted in a total of 750 (CI: 508-990) avoided mortalities (aged 30-99 years) across the SoCAB. In this reduction scenario, 53 (CI: 36-70) avoided mortality cases were estimated in Los Angeles, 161 (CI: 109-212) cases in San Bernardino, 390 (CI: 264-513) in Riverside, and 146 (CI: 99-193) in Orange County.
The target future year scenario saved the fewest lives among all the 6 control scenarios assessed (Figure 8). An 88% reduction in NOx emissions from 2008 to the future year (1R′N′0.3 case) led to the greatest reduction in PM2.5 concentrations (Figure 6). The basin-wide decrease in PM2.5 concentrations due to the 88% reduction in NOx emissions increased the total lives saved from 750 (CI: 508-990) to 1244 (CI: 846-1635). The number lives saved in this control strategy would increase by 17% in Los Angeles, 35% in San Bernardino, 107% in Riverside, and 8% in Orange County.
Figure 8.
Estimated avoided PM2.5-related future year mortalities for the 6 volatile organic compounds and NOx emission reduction scenarios in the South Coast Air Basin region.
The high percentage increase in the avoided deaths in Riverside may be caused by the shift in spatial distribution of PM2.5 particles across the SoCAB. The decrease in 2030 PM2.5 concentrations based on NOx reductions in 2% (1R′N′2.5), 31% (1R′N′1.75), 70% (1R′N′0.75), and 81% (1R′N′0.5) resulted in total avoided mortalities of 817 (CI: 554-1076), 697 (CI: 472-919), 883 (CI: 599-1165), and 1074 (CI: 729-1415), respectively, across the basin (Figure 8).
Avoided hospital admissions: respiratory and cardiovascular causes
The decrease in PM2.5 concentrations between the base year and the future year with the targeted 2030 VOC and NOx emissions may lead to a total of 172 (CI: 99-245) avoided hospital admissions due to respiratory causes across the SoCAB (Figure 9). The most avoided respiratory hospital admissions due to PM2.5 were in Riverside County (84 avoided cases, CI: 48-119), and the fewest avoided respiratory cases were in Los Angeles County (15 cases, CI: 9-21). When NOx emissions are reduced by 88% from the base year, the total number of avoided respiratory cases increased from 172 (CI: 99-245) to 282 (CI: 163-400). The number of avoided cases also increased in Los Angeles (17 cases, CI: 10-25), San Bernardino (49 cases, CI: 28-70), Riverside (175 cases, CI: 101-248), and Orange County (40 cases, CI: 23-57).
Figure 9.
Estimated avoided PM2.5-related future year hospital admissions due to respiratory causes for the 6 volatile organic compounds and NOx emission reduction scenarios in the South Coast Air Basin region.
There were fewer avoided hospital admissions (78 cases, CI: 30-127) due to cardiovascular causes than respiratory causes (Figures 9 and 10) for the future year with the targeted 2030 VOC and NOx emissions. Los Angeles had only 6 avoided cardiovascular cases (CI: 2-10). In San Bernardino, Riverside, and Orange County, there were 15 (CI: 6-25), 41 (CI: 16-66), and 16 (CI: 6-26) avoided cases, respectively (Figure 10). The most avoided hospital admissions for cardiovascular causes were estimated when 88% reductions in NOx emissions were made across the SoCAB.
Figure 10.
Estimated avoided PM2.5-related future year hospital admissions due to cardiovascular causes for the 6 volatile organic compounds and NOx emission reduction scenarios in the South Coast Air Basin region.
There were more avoided emergency room visits for asthma than hospital admissions for both respiratory and cardiovascular causes due to NOx and VOC emission control in the future year (Figure 11). A total of 322 (CI: 85-555) asthma cases may be avoided if the targeted 2030 emissions are met. But similar to respiratory and cardiovascular causes, the most avoided asthma cases were in Riverside County (158 cases, CI: 42-270) and the fewest number of asthma cases were in Los Angeles County (25 cases, CI: 7-44). There was a 16% increase in avoided asthma cases in Los Angeles with an 88% reduction in NOx emissions. The 88% reduction in NOx emissions scenario also increases the number of avoided asthma cases by 35.5%, 108.2%, and 6.1% in San Bernardino, Riverside, and Orange County, respectively.
Figure 11.
Estimated avoided PM2.5-related future year emergency room visits due to asthma for the 6 volatile organic compounds and NOx emission reduction scenarios in the South Coast Air Basin region.
Overall, Los Angeles benefited the least for avoided asthma cases from all emission control scenarios. This may be because of the spatial distribution of particles over SoCAB. The heavy pollution from this highly populated city in the SoCAB (10.02 million, US 2013 Census data)52 causes the PM2.5 particles to be highly concentrated over Los Angeles County and less affected by NOx emission reductions for all of the control scenarios.
Ozone-related health impacts based on NOx emissions reduction
Avoided premature mortalities due to ozone
The targeted reduction in VOC and NOx emissions across the basin between the base year and the future year resulted in 204 (CI: 140-260) total avoided O3-related deaths (Figure 12). Exposures to high concentrations of PM2.5 particles generally cause more premortalities than O3.53,54 The changes in O3 concentrations for this scenario were beneficial in Los Angeles (160 avoided deaths, CI: 110-210) and San Bernardino County (45 avoided deaths, CI: 31-60). However, in Riverside and Orange County, the changes in O3 were not beneficial. It was estimated that 2030 base O3 concentrations will cause one additional death in Riverside County and that no lives were saved in Orange County. Significant reductions in NOx concentrations increased O3 concentrations during the winter (Figure 5). This trend was also captured in BenMAP-CE analysis. For the 88% reduction in NOx case, the number of deaths decreased by 81% in Los Angeles and 87% in San Bernardino County from the base year. The increase in O3 concentrations due to the highest NOx emission reduction scenario resulted in an increase in the number of deaths at Riverside (34 deaths, CI: 23-45) and San Bernardino (26 deaths, CI: 18-34) (Figure 12).
Avoided hospital admissions due to respiratory causes: asthma-related emergency room visits and acute respiratory symptoms due to ozone
The targeted reduction in VOC and NOx emissions across the basin between the base year and the future year resulted in produced changes in ozone concentrations that were estimated to avoid 659 (CI: 196-1114) total hospital admissions due to respiratory causes across the SoCAB (Figure 13). Los Angeles and San Bernardino benefited from the changes in O3 concentrations with 541 (CI: 160-914) and 122 (CI: 36-206) avoided respiratory cases. The future year O3 concentrations led to a disbenefit for Riverside with BenMAP-CE predicting that there would be 4 (CI: 1-7) more cases of hospital admissions. In Orange County, the number of respiratory cases was not predicted to change. Decreasing NOx emissions to 88% of the base year led to a reduction in the number of hospital admissions by 79.6% in Los Angeles and 84.4% in San Bernardino County. At this NOx reduction level, the O3 concentration at Riverside increased the number of respiratory cases from 4 to 91. There was also an increase in the number of respiratory cases in Orange County (0-81 cases, CI: 24-140).
Figure 13.
Estimated avoided O3-related future year hospital admissions due to respiratory causes for the 6 volatile organic compounds and NOx emission reduction scenarios in the South Coast Air Basin region.
Similar trends were also observed for avoided emergency room visits due to asthma (Figure 14). For the future year target emissions case, the resulting O3 concentrations led to 80 (CI: −16 to 176) avoided asthma cases in the SoCAB. These O3 concentrations were beneficial in Los Angeles (60 cases avoided, CI: −12 to 131) and San Bernardino (21 cases avoided, CI: −4 to 45). However, in Riverside and Orange County, no avoided asthma cases were estimated. If 2008 NOx emissions were reduced by 88% in 2030, this results in an estimated 109 (CI: −21 to 237) avoided asthma cases in Los Angeles and 38 (CI: −7 to 84) avoided cases in San Bernardino (Figure 14). In accordance with the other health impacts, the O3 change for the 88% reduction scenario led to an increase in the number of asthma case at Riverside (11 more cases, CI: 2 to −24) and Orange County (6 more cases, CI: 1 to −14).
Figure 14.
Estimated avoided O3-related future year emergency room visits for the 6 volatile organic compounds and NOx emission reduction scenarios in the South Coast Air Basin region.
For acute respiratory symptoms, the O3 concentrations for future year target emissions case resulted in 184 705 (CI: 76 412-291 896) avoided cases in Los Angeles, 53 134 (CI: 21 984-83 962) avoided cases in San Bernardino and 129 (CI: 55-199) avoided cases in Orange County (Figure 15). In Riverside County, the future year O3 concentrations were estimated to increase the number of cases of acute respiratory symptoms (1520 more cases, CI: 599-2491). The O3 concentrations from the 88% NOx emission reduction case increased the number of acute respiratory symptoms to 39 697 (CI: 16 163-63 741) in Riverside and 29 604 (CI: 12 167-47 091) in Orange County. In Los Angeles and San Bernardino County, the numbers of acute respiratory symptoms decreased by 80% and 85%, respectively.
Total monetized health benefits for NOx and VOC emission scenarios in the SoCAB region
The total health costs related to PM2.5 in 2008 are estimated to be US $13.06 billion (in 2010 constant dollars) (Figure 16). This base year total health cost was estimated from the sum of monetized health impacts associated with the number of mortalities, acute respiratory symptoms, hospital admissions due to respiratory causes, and emergency room visits due to asthma in all counties in the SoCAB. In the future year scenario with the target VOC and NOx emissions, the number of deaths from PM2.5 exposure resulted in the highest economic burden. In this future year scenario, the decrease in PM2.5 concentrations led to a total avoided cost of US $6.6 billion (in 2010 constant dollars) (Figure 16). However, the resulting economic benefits from this control strategy were significantly lower than other 2030 reduction scenarios evaluated in this study. The trend in total economic benefit of each control scenario analyzed is shown (Figure 16). For the 88% NOx reduction strategy, there was an increase in 66% in the total avoided cost from the target emission future year scenario. This means that significantly greater NOx reductions from 2008 emission levels might save the SoCAB region billions of dollars. For the 70% and 81% NOx emission reduction cases, there was an increase in the total avoided costs by 17.7% and 42.3% across the basin.
Figure 16.
Total monetized health benefits for all avoided PM2.5-related health impacts for the 6 volatile organic compounds and NOx emission reduction scenarios in the South Coast Air Basin region.
For O3, the economic benefits of future year NOx reductions were lower than PM2.5. This is consistent with previous studies that suggest that PM2.5 exposure can lead to more mortalities than O3.53,54 The O3 concentrations at the 2030 target emission case resulted in a total economic benefit of US $1.83 billion (in 2010 constant dollars). Even though there was an NOx disbenefit for some counties in the basin (lower NOx emissions increased O3 concentrations), the total economic benefit increased by 54% from the 2030 target emission case to the 88% NOx emission case (Figure 17). The increase in avoided cases in Los Angeles and San Bernardino County at the NOx reduction level may have caused the overall increase in the economic benefit across the basin. The economic benefit at the 2%, 31%, and 81% NOx reduction strategies results in an increase in the avoided costs by 99%, 40%, and 12% over the future year case with the target VOC and NOx emissions, respectively. However, at the 75% NOx emission reduction, the basin will lose 2.2% in avoided O3-related medical costs compared with the 2030 target emission case.
Conclusions
Epidemiologic studies have shown that there are a large number of adverse health effects associated with ozone and PM. Adverse health impacts of air pollution include premature death, respiratory disease, and cardiovascular disease. CMAQ and BenMAP-CE are useful tools to assess the health impacts of proposed air quality improvement strategies.
CMAQ was used to simulate air quality of the SoCAB for selected fall season episodes. The selected episodes occurred during the year 2008 and therefore it was considered to be the base year. These episodes with the same 2008 meteorological conditions but with differing emission inventory reduction scenarios for the year 2030 were simulated and therefore the year 2030 was regarded as the future year. BenMAP-CE was used to assess health impacts of O3 and PM2.5 during the base year and to evaluate the health benefits of reducing O3 and PM2.5 concentrations between the base and future years.
It was found that for the base year, exposure to PM2.5 could be responsible for 131, 434, 409, and 781 deaths in San Bernardino, Riverside, Orange, and Los Angeles County, respectively. The economic benefits analysis estimated the total cost of 2008 PM2.5 health-related impacts across the SoCAB to be US $13.06 billion (in 2010 constant dollars).
During the base year exposure to O3 and PM2.5 could be responsible for a combined total of 432, 721, 755, and 2137 deaths, and these correspond to 3.9%, 5.5%, 4.7%, and 3.9% of deaths in San Bernardino, Riverside, Orange, and Los Angeles County, respectively. Exposure to PM2.5 may cause 30%, 60%, 54%, and 37% of the total deaths attributed to O3 and PM2.5, respectively.
The effect of NOx emission reductions on health impacts due to O3 exposure across the SoCAB was more complicated than PM2.5 during the fall due to lower temperatures and lower photolysis. Reductions in NOx emissions increased O3 concentrations in some areas of the SoCAB, whereas reducing them in others. The reductions in NOx could lead to increases in adverse health outcomes during the fall season. In Riverside County, the future year O3 concentrations were estimated to increase the number of cases of acute respiratory symptoms yielding 1520 more cases in the fall, although further reductions in NOx emissions led to even more cases.
The current target emission control strategy for the SoCAB is not the most effective scenario for reducing health impacts resulting from PM2.5 exposure of the strategies evaluated for reducing harmful air quality outcomes. A strategy with lower or greater levels of NOx emission reductions would be better for reducing PM2.5 concentrations. Reductions in NOx emissions for the future year had the greatest effect on PM2.5 concentrations and health in Riverside County, whereas Los Angeles County was the least affected.
Acknowledgements
The authors gratefully acknowledge the South Coast Air Quality Management District for providing their CMAQ modeling input and output files for their use.
REFERENCES
Notes
[1] Financial disclosure The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Support for this research was provided by the Truck and Engine Manufactures Association (EMA). The Coordinating Research Council (CRC) under separate projects (A-101 and A-109) supported the air quality modeling that was the starting point for the research reported here. Funding was also provided by the NOAA Center for Atmospheric Science—Meteorology (NCAS-M) which is funded by the National Atmospheric Administration/Educational Partnership Program under Cooperative Agreement #NA16SEC4810006. Neither the CRC, EMA, NOAA, members of CRC, EMA, and NOAA, UTEP nor any person acting on their behalf (1) makes any warranty, express or implied, with respect to the use of any information, apparatus, method, or process disclosed in this report; (2) assumes any liabilities with respect to use of, inability to use, or damages resulting from the use or inability to use, any information, apparatus, method, or process disclosed in this research publication; and (3) this publication reflects the independent work and conclusions of its authors and has not been endorsed, accepted or otherwise approved by any of the funders of the underlying work.
[2] Conflicts of interest The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
[3] DRS performed the air quality modeling and led the health effects modeling, ES developed health effects modeling protocols used in this project, RAP preformed air quality modeling, RF was a co-leader of the project and was responsible for the air quality modeling, DEC preformed data analysis and model evaluation, and WRS conceived the project, was a co-leader, and wrote the final manuscript.
[4] As a requirement of publication, authors have provided to the publisher signed confirmation of compliance with legal and ethical obligations including but not limited to the following: authorship and contributorship, conflicts of interest, privacy, and confidentiality. No human or animal subjects were involved. The authors have read and confirmed their agreement with the ICMJE authorship and conflict of interest criteria. The authors have also confirmed that this article is unique and not under consideration or published in any other publication, and that they have permission from rights holders to reproduce any copyrighted material. There are no additional disclosures to make.
