Article pubs.acs.org/est
International Airport Impacts to Air Quality: Size and Related Properties of Large Increases in Ultrafine Particle Number Concentrations N. Hudda and S. A. Fruin*,† Keck School of Medicine, Environmental Health Division, University of Southern California, Los Angeles, California 90033, United States S Supporting Information *
ABSTRACT: We measured particle size distributions and spatial patterns of particle number (PN) and particle surface area concentrations downwind from the Los Angeles International Airport (LAX) where large increases (over local background) in PN concentrations routinely extended 18 km downwind. These elevations were mostly comprised of ultrafine particles smaller than 40 nm. For a given downwind distance, the greatest increases in PN concentrations, along with the smallest mean sizes, were detected at locations under the landing jet trajectories. The smaller size of particles in the impacted area, as compared to the ambient urban aerosol, increased calculated lung deposition fractions to 0.7−0.8 from 0.5−0.7. A diffusion charging instrument (DiSCMini), that simulates alveolar lung deposition, measured a fivefold increase in alveolar-lung deposited surface area concentrations 2−3 km downwind from the airport (over local background), decreasing steadily to a twofold increase 18 km downwind. These ratios (elevated lung-deposited surface area over background) were lower than the corresponding ratios for elevated PN concentrations, which decreased from tenfold to twofold over the same distance, but the spatial patterns of elevated concentrations were similar. It appears that PN concentration can serve as a nonlinear proxy for lung deposited surface area downwind of major airports.
■
INTRODUCTION Aircraft emit large numbers of particles, and recent studies1,2 have demonstrated that aviation emissions can increase particle number (PN) concentrations for long distances downwind from runways. In Los Angeles, aviation emissions caused a twofold increase in PN concentrations over local background concentrations as far as 18 km downwind during the usual daytime westerly sea breeze.1 These increased PN concentrations extended downwind and eastward over 60 km2 to produce a PN concentration increase equivalent to that resulting from one-quarter to one-half of all freeways in Los Angeles County on an area-weighted basis.1 Hudda et al.1 also inferred from the alignment of locations of maximum concentrations with the landing jet trajectory and noise contours that landing jets were the primary contributor to this impact; this is also suggested by models that include the effect of enhanced downward mixing due to the descending motion of jet vortices.3,4 In a subsequent study, Keuken et al.2 reported a high degree of correlation (r2 of 0.77) between PN concentrations and air traffic intensity 7 km downwind of Schiphol Airport in The Netherlands where PN concentrations were elevated by a factor of 3. They also detected a 20% increase in PN concentrations (on the order of few thousand particles cm−3) at a site 40 km from the airport when prevailing winds were from the direction of the airport. © XXXX American Chemical Society
Besides potentially large increases in PN concentrations downwind of large airports, differences in particle size (as well as composition) are important from exposure and health perspectives. However, until recently, previous particle size distribution measurements of aviation emissions were only conducted within a few hundred meters of airports or were measured in individually intercepted plumes, primarily from takeoffs or idling.5−8 In these studies, higher emission indexes for PN emissions are typically reported for low-thrust conditions like idling than for high-thrust conditions like takeoffs.5,6,8 Particle size distributions are reported to be bimodal, but mode or geometric mean sizes vary in different studies due to differences in engine type and thrust, ambient dilution conditions, and experimental methods.5−8 Nonetheless, a primary mode smaller than 30 nm is commonly reported. The most distant size distribution measurements from an airport were recently made by Keuken et al.2 who reported that nucleation mode particles (10−20 nm) still dominated the size distributions at a monitoring location 3 km from the airport boundary. Such large numbers of very small particles can be Received: October 29, 2015 Revised: February 9, 2016 Accepted: February 19, 2016
A
DOI: 10.1021/acs.est.5b05313 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology expected to increase the lung deposition compared to the larger size distributions typical of aged urban aerosols. Although the most appropriate metric for submicron particle toxicity is unsettled, there is evidence that lung deposited surface area may be a more appropriate toxicity metric than number or mass.10,11 For example, Buonanno et al.12 reported that alveolar-lung deposited surface area daily dose correlated strongly with exhaled nitric oxide (R2 = 0.9, n = 16), a marker of lung airway inflammation, in asthmatic children. Although a number-based metric is dominated by the more numerous ultrafine particles, owing to their small size they contribute relatively less to total surface area and little to particle matter mass. If surface area is the most appropriate health metric, then characterizing the particle surface area is important because the relative contribution of the aviation emissions impact on lung deposited surface area will differ significantly from that of PN concentration. For this reason, we undertook this study to simultaneously quantify the extent of the large-area impacts from aviation in terms of PN, as well as particle size and lung deposited surface, to facilitate future health-related investigations. This also allowed estimating the extent of enhanced lung deposition due to reduced particle size. We present the first spatially resolved, size-related measurements of aviation ultrafine particle emissions at large distances of up to 18 km from a major airport.
Figure 1. Map of Los Angeles International Airport (LAX) with an example of the typical mobile monitoring route and stationary sampling locations. The inset shows the wind rose for 2014.
■
using the NanoScan SMPS at locations shown in Figure 1; varying subsets of these locations were monitored on each run. These locations were chosen to be sufficiently distant from major arterial roadways or intersections in order to minimize the impact of vehicular emissions and provide stable PN concentrations for the duration of each scan. The vehicle engine was shut-off during stationary monitoring to avoid selfsampling and for mobile measurements the vehicle was driven in “green” mode that switched the power to electric during idling. Both mobile and stationary measurements of mean (numberaverage) particle size were conducted with a Diffusion Size Classifier DiSCMini (Matter Aerosol, Wohlen, Switzerland) during summer 2014 monitoring. The DiSCMini is a unipolar diffusional charging-based classifier that measures current from collected charged particles in two stages: a diffusion screen stage that measures current while preferentially collecting the smallest particles (d50 is 40 nm), and a filter stage that collects current from the remaining charged particles. The resulting ratio of these two currents is used to estimate a numberaveraged particle size. The instrument’s range is 10−300 nm and accuracy is ±30%.15 During mobile monitoring, the size distribution stability was inadequate for accurate measurements so only stationary measurements of size distribution are reported. Of the sizerelated mobile measurements, only DiSCMini based mean (number-average) size are reported. Stationary comparisons showed that although the mean size from the two instruments are highly correlated (r2 > 0.94), the DiSCMini reported relatively smaller sizes than the NanoScan SMPS when mean particle size was smaller than 35 nm. This observation has been reported by other studies.16−19 It should be noted that the agreement between the two instruments depends on the shape of the size distribution.16−19 (Detailed comparisons are presented in the Supporting Information, Figure S1 and Table S1). For this study, the DiSCMini was factory calibrated to an aerosol geometric standard deviation (GSD) of 1.7. Deviations from this GSD were unknown outside of stationary NanoSMPS measurements, so we reported DiSCMini mean
MATERIALS AND METHODS LAX Airport. The LAX complex extends about 4.5 km east to west and about 2.5 km north to south, and is located beside the Pacific Ocean in the western part of Los Angeles. The prevalent wind direction at LAX is WSW (252°).13 Almost 100% of the takeoffs are to the west over the Pacific Ocean. During 0630−0000 h, flights land into the westerly winds as they approach LAX from the east over residential areas14 on two sets of runways that are aligned to 263° (from true north). During overocean operation 0000−0630 h, flights land over the Pacific Ocean, but such operations typically account for only 4% of daily arrivals. A map of LAX, its surroundings, and the rose diagram for wind direction measured by National Weather Service at LAX during 2014 is shown in Figure 1. Mobile Monitoring and Instruments. To measure downwind impacts from this large source, an instrumented gasoline-electric hybrid vehicle1 was driven in multiple kilometer long, north−south transects in the area east of LAX, beginning from the blast fence and continuing to downwind distances up to 18 km. A monitoring route example is shown in Figure 1. Measurements were conducted over eight days in summer 2014 and four days in winter 2013, subject to instrument availability; all runs have been reported. Table 1 provides details of monitoring dates and times. For additional winter data see Hudda et al.1 Throughout the campaign, PN concentrations were monitored with a Condensation Particle Counter (CPC, Model 3007, TSI Inc., MN). It measured particle concentrations in the range 0−105 particles cm−3 with ±20% accuracy and has a minimum detectable size (d50) of 10 nm. All PN concentrations reported in this study were based on these CPC measurements. Stationary measurements of particle size distributions were conducted with a portable TSI NanoScan Scanning Mobility Particle Sizer (SMPS) 3910. The NanoScan SMPS was operated at 60 s scanning mode and reported 13 channels in the 10−420 nm size range. All size distribution data presented in this study were based on stationary monitoring conducted B
DOI: 10.1021/acs.est.5b05313 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology Table 1. Monitoring Dates, Times, Meteorological Conditions, and Baseline PN Concentrations wind direction (%)
date
time
planetary boundary layer height (m)a
12/18/2013 12/19/2013 12/20/2013 12/23/2013 7/14−15/2014 7/15/2014 7/16/2014 07/17−18/2014 7/18/2014 07/19−20/2014 7/20−21/2014 7/21/2014
1700−2042 1713−2048 1630−2007 1506−1948 2130−0130 1415−1928 0641−1558 1548−0144 0908−1730 1716−0300 1853−0036 0510−1347
246 395 159 179 107 362 376 296 551 171 155 269
temp (°C)
E
SE
S
SW
W
NW
N
NE
14.5 12.4 13.9 16.3 21.2 21.4 22.3 20.5 22.4 19.8 20.2 21.9
0 30 2 0 6 0 0 0 0 0 0 3
2 15 2 0 16 0 0 0 0 0 0 9
29 0 5 2 4 0 0 0 0 0 1 15
47 2 67 8 17 25 8 3 15 9 40 15
21 7 21 76 52 74 85 97 85 74 59 57
0 0 1 6 2 1 6 0 0 15 0 0
0 0 0 6 1 0 0 0 0 2 0 0
0 45 0 1 2 0 0 0 0 0 0 0
wind speed (m s−1)
PNC baseline outside the impacted areas (particles cm−3)b
ratio of maximum impact to unimpacted baseline 10 km downwind
± ± ± ± ± ± ± ± ± ± ± ±
10 000 17500 15000 10 000 10 000 10 000 7500 7500 10 000 7500 7500 10 000
6 n/a 4 11 3.5 4 5.3 6 4.6 5.3 6 4.5
2.9 2.2 1.9 2.5 1.8 5.6 5.1 5.6 6.0 2.8 4.3 4.4
1.1 0.8 1.4 1.3 0.6 1.3 1.8 1.5 1.1 1.2 1.3 2.3
a Average of NOAA HYSPLIT GDAS based forecast for coordinates 34.94°N and 118.41°W over the sampling duration. bThe average baseline PN concentrations in the un-impacted areas away from local traffic sources are reported to nearest 2500 particles cm−3.
GPS time and location (Garmin GPSMAP 76CSC) were assigned to the middle of each 31 s interval after data were also aligned with the fastest instrument to account for differences in response times. For parameters that were expected to decrease due to the impact of LAX-related emissions, such as particle size, we similarly used the rolling 95th percentile value over a 31 s period. Particle size distributions measured at stationary sites were used to estimate total deposition fraction in lung by number (TDFN). TDFN is defined as the average of size-specific deposition fractions (DFN) weighted by the number concentrations for that size; in our case TDFN was the average of DFN for each midpoint mobility diameter of the 13 discrete bins reported by the NanoSMPS weighted by total number concentrations in those bins. Size-specific DFN values can be derived using ICRP human respiratory tract deposition model21 based on aerodynamic diameter and is commonly reported for the “reference worker” case. We chose to calculate TDFN corresponding to our measured electrical mobility diameters instead of converting to aerodynamic diameters. Electrical mobility based DFN values were derived from the following two experimental studies using human subjects: (1) Rissler et al.,22 who developed a parametrized size-dependent DFN model based on total respiratory tract deposition results in 10 healthy subjects exposed to nearly hydrophobic fresh diesel engine exhaust and (2) Löndahl et al.,23 who measured lung deposition in 9 healthy subjects exposed to curbside aerosol that had 10−30% hygroscopic fraction. Löndahl et al.23 also note that experimentally determined DFN for ultrafine particles are higher than the ICRP model21 (likely due to the limited number of studies that inform the ICRP model in that range and/or intersubject variability); this has also been reported by other studies.24 The DFN values used in this study are reported in Figure S4 and Table S2 of the SI. In general, DFN increases with decreasing size in the 10−200 nm range, that is, for a hydrophobic spherical particle, the minimum DFN value of ∼0.2 occurs at about 200 nm and increases to ∼0.4 by 100 nm and to ∼0.8 by 20 nm.20 However, DFN values also depend on other physical and chemical properties of the aerosol, such as hygroscopicity, which affects particle growth in the high humidity conditions of the lung, and DFN is also influenced
size as measured but limited our discussion to ranges of size when interpreting DiSCMini data. Despite these limitations, the DiSCMini provided good temporal resolution (1 s) for detecting relative changes in mean particle size with good spatial resolution during mobile monitoring. The DiSCMini also reported PN concentrations (instrument measurement range: 103−106 particles cm−3), but we relied on the CPC for PN concentrations. For comparison, linear regression correlation coefficients between PN concentrations reported by CPC and the DiSCMini are presented in Figure S2(a−c) (SI). The r2 was above 0.98, and DiSCMini data were within the manufacturer specified ±20% accuracy compared to a CPC.15 However, we observed that the agreement was size dependent. For example, the ratio of 1 min concentrations reported by CPC to that by DiSCMini was 0.98 ± 0.11 when mean size was ≥30 nm and 0.70 ± 0.25 when mean size was
