Spatial variations in the partitioning of nitrate between gas phase nitric acid (HNO3) and particulate ammonium nitrate were observed using airborne measurements of trace gas mixing ratios, particle size distributions, and particle composition. During the Intercontinental Transport and Chemical Transformation experiment in April and May 2002 the NOAA WP-3 aircraft flew up to 8 km altitude on 11 research flights from Monterey, California. The formation of semivolatile aerosols was studied by examining the enhancement of fine-particulate ammonium nitrate and depletion of gas-phase HNO3 over the San Joaquin Valley, Los Angeles Basin, and Mojave Desert. Gas-phase particle precursors, HNO3 and ammonia (NH3), were converted to particulate ammonium nitrate at higher altitudes within the boundary layer. These particle layers were a consequence of lower ambient temperatures that caused a reduction of the dissociation constant for ammonium nitrate aerosol so that gas phase HNO3 was depleted and particle mass was formed. The resulting vertical gradients in particulate matter and HNO3 were observed in well-mixed boundary layers where other directly emitted trace gases (CO) and secondary pollutants (O3) exhibited no vertical gradients. Hence the equilibrium between the gas and particle phases occurred faster than boundary layer mixing times and chemical rather than meteorological effects were responsible for the layers of enhanced particulate matter aloft. Coincident HNO3 depletion and ammonium nitrate formation was also observed downwind from regions characterized by large agricultural NH3 emissions in the Los Angeles Basin and San Joaquin Valley.
 The presence of aerosols in the troposphere can affect air quality by threatening human health and lowering visibility. Consequently, the development of pollution control strategies aimed at improving local and regional air quality by reducing airborne particulate matter require an understanding of the processes that govern particle formation and growth. In many regions of the United States with substantial air pollution problems, secondary generation of fine-aerosol nitrates accounts for a large fraction of the particle mass and visibility reduction. In southern California, which has long been affected by pollution that exceeds state and federal air quality standards, the conditions and sources that lead to high levels of particulate ammonium nitrate have been studied extensively [e.g., Russell et al., 1983; Russell and Cass, 1986; Chow et al., 1994; Kleeman et al., 1999].
 Ammonium nitrate aerosol is formed from the reaction of gas phase ammonia (NH3) and nitric acid (HNO3). This secondary particle formation typically results in accumulation mode fine particles with diameters less than 1 μm. Elevated ammonium nitrate levels are the consequence of anthropogenic emissions of NH3 and NOx (NOx = NO + NO2). In sunlit urban areas, HNO3 is primarily formed from the oxidation of NOx emitted from tailpipes and other combustion sources (for example, power generation). NH3 emission sources include animal waste in addition to tailpipe emissions and other area sources. Ambient temperature and relative humidity (RH) affect ammonium nitrate concentrations by determining the partitioning of ammonium and nitrate between the gas and particulate phase. When ammonium nitrate aerosol is formed, the product of the partial pressures of gas phase HNO3 and NH3 is given by the dissociation constant for ammonium nitrate [Stelson and Seinfeld, 1982; Mozurkewich, 1993]. The equilibrium between gas and particle phases of ammonium nitrate has been examined using models and measurements [e.g., Meng and Seinfeld, 1996; Dassios and Pandis, 1999; Moya et al., 2002; Zhang et al., 2003]. Since the deposition velocity of particulate ammonium nitrate is slower than that for gas phase HNO3, the lifetime of nitrate in the atmosphere, and hence its transport, is affected by the partitioning between the gas and particle phases.
 Understanding the spatial distribution of tropospheric aerosols is important for developing effective pollution control strategies. Spatial inhomogenieties in pollutant concentrations can lead to uncertainties in models that assume concentrations are uniform inside a model grid volume [Collins et al., 2000]. Rapid variations in meteorological conditions and trace gas mixing ratios that occur in the atmosphere typically cannot be captured by filter-based measurements of particles and gas-phase precursors, which require sampling times of several hours [Solomon et al., 1992; Chow et al., 1994]. Additionally, measurements that average over several hours may have limited utility for the study of the equilibrium between the gas and particle phases [Zhang et al., 2003]. Fast-response measurements are valuable for accurately quantifying temporal and spatial variations in pollutant concentrations to provide a better understanding of the processes that determine their abundance. Additionally, compared to integrated filter techniques, the in situ measurements reported here are less prone to sampling artifacts associated with measurements of semivolatile particles like ammonium nitrate [Weber et al., 2003].
 The aircraft-based measurements reported here facilitate the study of processes that occur rapidly in the atmosphere and permit a high-resolution determination of both horizontal and vertical gradients of the measured species. Although layers of pollutants over southern California have been identified previously using aircraft observations [e.g., Blumenthal et al., 1978; McElroy and Smith, 1986; Collins et al., 2000] and numerical simulations [Lu and Turco, 1995], only the meteorological effects responsible for these layers have been studied. Simultaneous in situ measurements of particle size distributions, particle composition, and gas-phase precursor species yield new insights into the formation of layers of pollutants. The relationships between trace gases and particulate matter reported here show that gas-to-particle conversion processes cause spatial variability of pollutants and layers of fine particles to exist within the boundary layer. Gas-to-particle conversion is defined here to be secondary particle formation caused by changes in ambient temperature, relative humidity, or NH3 concentration that drives the ammonium nitrate equilibrium from the gas phase to the particulate phase. Characterizing these elevated layers of pollution is important for fully understanding pollutant transport, the causes of visibility reduction, and the radiative impact of tropospheric aerosols [Charlson et al., 1992].
 The measurements of gas-to-particle conversion discussed here were obtained from the NOAA WP-3 aircraft during the Intercontinental Transport and Chemical Transformation experiment during April and May of 2002 (ITCT 2K2). Data were obtained up to 8 km altitude on 11 research flights from Monterey, California and on transit flights to and from Monterey. All flights were conducted during daylight, and the results presented here use data collected during the afternoon under clear skies. Although during much of the study the aircraft flew in the free troposphere to examine the long-range transport of pollutants from Asia to the western United States, the emission and transport of California pollution was also examined. Most flights included descents into the boundary layer over California, and two flights focused on measuring pollution within the boundary layer in the vicinity of urban and agricultural areas. This study uses data obtained during measurements within the boundary layer in the vicinity of urban areas to investigate gas-to-particle conversion.
 Fast-response instruments are necessary for achieving high spatial resolution from an aircraft platform. For the nominal aircraft speed, a spatial resolution of 100 m was obtained from instruments with 1-s time resolution. Meteorological parameters important for understanding the trace gas and particle measurements included 1 Hz measurements of aircraft position, ambient temperature, relative humidity, dew point, wind speed, and wind direction.
 Particle size distributions were determined using optical particle counters [Jonsson et al., 1995] and condensation nucleus counters [Brock et al., 2000]. Particles were sampled through electrically conductive tubing under laminar conditions in a low turbulence inlet [Brock et al., 2003]. Dry particle number concentration, surface area concentration, and volume concentration were determined with 1 s time resolution for particles with diameters between 0.005 μm and 7 μm. The analysis presented here primarily uses the dry particle volume concentration (μm3 cm−3), reported at standard temperature and pressure, for fine particles with diameters between 0.15 μm and 1 μm and for coarse particles with diameters between 1 μm and 7 μm diameter. During the 2.9 s inlet residence time, the volatilization of solid or deliquesced ammonium nitrate was estimated to be low (<10% by mass) based on chemical modeling and empirical data [Dassios and Pandis, 1999; Bergin et al., 1997]. In the polluted boundary layer studied here, fine-particle volume concentration was measured with a precision better than 16% and a bias less than 36%. Precision was calculated using a Monte Carlo simulation of random errors from counting statistics and flow measurement uncertainties through the data processing algorithm following Brock et al. . Biases are based on these simulations and on measured changes in the particle diameter reported by the optical particle counter instruments for real refractive indices varying from 1.42 to 1.58. Particle losses in flow splits, bends and straight tubing in the laminar sampling system were calculated [Baron and Willeke, 2001] and applied to the data. These corrections were negligible for particles with diameters <1 μm.
 Converting all measurements to the same mixing ratio units facilitates quantitative comparisons of particle volume concentrations with gas phase and bulk particle composition measurements. For a known particle density and molar mass, fine-particle mixing ratios can be determined from fine-particle volume concentrations. If the fine-particle composition were ammonium nitrate (density = 1.7 g cm−3, molar mass = 80 g mole−1), then fine-particle mixing ratios in parts per billion by volume (ppbv) would be calculated as 0.48 times fine-particle volume. This assumption of particle composition is tested below (section 3.1) by comparing the relationships between fine particles, HNO3, and bulk fine-particle composition to those expected from ammonium nitrate formation from the conversion of gas phase HNO3.
 One-second average HNO3 mixing ratios were measured using a chemical ionization mass spectrometer (CIMS) [Neuman et al., 2002]. Ambient air was sampled through a heated (50°C) Teflon inlet that extended 35 cm perpendicular to the aircraft fuselage. The flow rate through the inlet was approximately 8 standard liters per minute, and the residence time in the inlet and tubing that delivered the sampled air to the instrument was <50 ms. Particulate nitrate was not detected by the CIMS. Since the inlet was perpendicular to the airflow, some particles were excluded from the sampling line. Furthermore, the ion-molecule reaction used to measure HNO3 is sensitive only to gas-phase HNO3, and the residence time and temperature in the inlet were sufficiently low to prevent volatilizing HNO3-containing particles. Evidence for this particle discrimination is shown in the results section, where the CIMS measured low levels of HNO3 in the presence of high levels of particulate nitrate.
 In-flight standard addition calibrations and measurements of instrumental background signals were performed regularly to determine the accuracy, stability and sensitivity of the instrument. Calibrations were achieved by admitting 2.1 ppbv of HNO3 from a permeation source into the inlet. In between flights, the permeation rate from the temperature and flow-controlled HNO3 calibration system was measured continuously using a UV optical absorption system [Neuman et al., 2003]. The accuracy was determined to be ±15%, using the measured calibration stability throughout the study and the uncertainty in determining the output from the HNO3 calibration source. The instrument background was measured approximately once every 20 min by diverting the ambient air through a sodium bicarbonate impregnated nylon wool scrubber. Differences between consecutive background measurements were used to determine that an additional inaccuracy of ±30 parts per trillion by volume (pptv) was imposed by the background instability. Averaging the data from a section of flight does not reduce the accuracy or the uncertainty caused by the changing instrument background. The instrument precision, which was limited by counting statistics on the background signal, was ±25 pptv (1σ) for 1 s measurements and ±9 pptv for 10 s averages of the data.
 One-second average measurements of carbon monoxide (CO), ozone (O3), NO, NO2, and NOy are used to characterize the gas-to-particle conversion processes studied here. CO mixing ratios were measured with an uncertainty estimated to be 2.5% using a vacuum ultraviolet fluorescence spectrometer [Holloway et al., 2000]. O3 was measured using NO chemiluminescence and calibrated using an UV absorption instrument with an uncertainty of ±(0.4 ppbv + 2%). NO was measured by ozone-induced chemiluminescence with an uncertainty of ±(10 pptv + 5%). NO2 was measured with an uncertainty of ±(30 pptv + 10%) using the difference between the NO measurement and a measurement of NO preceded by a NO2 photolysis cell [Ryerson et al., 2000]. NOy (=NO + NO2 + HNO3 + PAN + …) was measured with an uncertainty of ±(15 pptv + 10%) [Ryerson et al., 1999] using catalytic reduction on a 300°C gold converter followed by NO chemiluminsence detection.
 The inorganic bulk composition of fine particles (diameter < 1 μm, selected using an impactor) was obtained using a particle into liquid sampler (PILS) with analysis by ion chromatography [Orsini et al., 2003; Weber et al., 2003]. Ambient air was sampled through the low turbulence inlet (described above) and transported to the instrument at 30 l/min through sections of copper and rubber tube (Reynolds number = 1800). At the instrument rack, the flow was split between two detectors. Calculations (discussed above) indicate that the inlet and sampling losses were negligible for sub-micron particles, and no corrections were applied to the measurements. Once every 4 min, measurements of particulate calcium, magnesium, potassium, chloride, sodium, sulfate, ammonium (NH4+), nitrite, and nitrate (NO3−) were obtained. Each measurement was an average over either a 2 min or 2 min 40s sampling interval. For the NO3− ions discussed here, the 3 σ detection limit was 0.002 μg m−3, and the uncertainty was estimated to be approximately ±20%. The detection limit was lower than previously reported [Orsini et al., 2003] as a consequence of more sensitive calibrations and the use of new types of chromatography columns that improved the instrument background signals. The ion balance indicated an excess of cations, even in pristine regions where a neutral aerosol was expected, suggesting that the NH4+ measurement was high. Since the cause of this discrepancy is unknown, the NH4+ data reported here are used mainly for investigating relative changes in concentration. The particle mass concentrations from the PILS are converted to mixing ratios here using the measured ambient air pressure and temperature and the molar mass of the measured species. The composition of single particles was determined using a laser ionization mass spectrometer instrument [Murphy and Thomson, 1997]. The instrument alternated every several minutes between obtaining positive and negative mass spectra from individual laser-ionized particles with diameters from 0.2 to >2 μm.
 Gas-to-particle conversion was observed when the aircraft flew within the planetary boundary layer over California. HNO3 depletion and enhancements in particle mass were observed when the aircraft encountered lower temperatures at higher elevations within the boundary layer (during altitude profiles) and also when air was sampled downwind from large NH3 sources. The observations of gas-to-particle conversion (Figure 1) in these two different environments were obtained during flights that sampled air over the San Joaquin Valley (8 May and 17 May), Mojave Desert (13 May and 25 April), and Los Angeles Basin (13 May).
3.1. Altitude Profiles
 On several flights, the aircraft performed spiral ascents and descents (altitude profiles) over California and the Pacific Ocean to determine boundary layer heights and examine vertical gradients in tropospheric composition. Climb rates and descent rates were typically 5–10 m/s, so 1-s measurements give a vertical resolution of ≤10 m. During altitude profiles, the vertical resolution of the PILS measurements was limited by the several-minute sampling period to be approximately 1 km, which was not sufficient to fully capture the observed vertical gradients. Hence, when the aircraft encountered rapidly changing ambient conditions during altitude profiles, the bulk aerosol ionic composition measurements are used to indicate qualitatively regions where particulate NO3− and NH4+ were enhanced. Frequently the boundary layer was well mixed, as indicated by CO and O3 mixing ratios that were independent of altitude within the boundary layer. In contrast, HNO3 and fine-particle mixing ratios were often strongly dependent on altitude, where HNO3 was depleted and fine particles were enhanced at higher altitudes within the boundary layer.
 Gas-to-particle conversion occurred in the upper portion of the boundary layer over the San Joaquin Valley. Meteorological parameters, trace gas mixing ratios, and particle mixing ratios are shown in Figure 2 when the aircraft descended from the free troposphere into the boundary layer near Visalia California on 17 May 2002 (Figure 1). The boundary layer height was indicated by the weak (<1°C) inversion in the temperature profile between 1.9 and 2.3 km (Figure 2a). Below the temperature inversion, the temperature lapse rate was nearly adiabatic at −8.9°C/km. The relative humidity (RH) increased with altitude within the boundary layer, but remained below the deliquescence RH [Stelson and Seinfeld, 1982; Mozurkewich, 1993], which varied from 62% to 68%. When the aircraft descended from the free troposphere into the planetary boundary layer, there was a sharp change in trace gas mixing ratios. Above the temperature inversion in the free troposphere, measurements of CO and O3 between 2.3 and 4 km altitude averaged 107 ppbv and 43 ppbv respectively. Below the temperature inversion, the mixing ratios of CO and O3 were nearly constant and were approximately twice as high as the values in the free troposphere, as shown in Figure 2b. O3 measurements (not shown) discussed here were obtained during the ascent from the San Joaquin Valley, since the descent data were compromised by an instrument malfunction.
 HNO3 and fine-particle mixing ratios, in contrast to CO and O3, exhibited large changes not only between the free troposphere and boundary layer but also within and below the temperature inversion (Figure 2c). HNO3 was depleted and fine particles were enhanced at higher altitudes within the boundary layer. During this time, the fine particles were dominated by the accumulation mode with diameters between 0.2 and 0.6 μm, and fine-particle volume was approximately half the coarse-particle volume. Coarse-particle composition was not measured here. Bulk particle composition measurements were not obtained in the upper part of the boundary layer during this altitude profile (Figure 2c), so that NO3− measurements were not coincident with the observed HNO3 depletion and fine-particle enhancement. However, bulk particle composition measurements during other portions of the altitude profile showed that the inorganic fine-particle ionic composition was primarily NH4+, NO3−, and sulfate.
 The strong variations apparent in HNO3 and particle volume were a consequence of gas-to-particle conversion. At the higher altitudes and lower temperatures, the ammonium nitrate equilibrium was driven from the gas phase to the particulate phase, such that gas-phase HNO3 was removed from the air mass and particle mass was enhanced. The connection between HNO3 depletion and particle mass growth is indicated by the strong anticorrelation between HNO3 and fine particles shown in Figure 3. The linear least squares fit to the fine particle versus HNO3 mixing ratio had a correlation coefficient of −0.92 and a slope of −1.12 (Table 1, row 7). If the fine-particle enhancement were a consequence of ammonium nitrate formation from the conversion of gas phase HNO3 and NH3, then each mole of HNO3 removed would result in 1 mole of ammonium nitrate formed, and the fine particle:HNO3 slope would be −1. Within the combined measurement uncertainty, the measured slope is consistent with the assumption that the fine-particle enhancements were caused by ammonium nitrate formation. Mixing processes can also affect the measured slope. Mixing often produced a positive correlation between HNO3 and fine particles (as seen at higher and lower altitudes in Figure 3), since both are secondary pollutants with common sources from polluted urban areas and low values in remote areas and the free troposphere. The preservation of the strong anticorrelation between fine particles and HNO3 and the correlation slope of approximately −1 indicates that the gas-to-particle conversion of NH3 and HNO3 to ammonium nitrate occurred rapidly compared to the mixing of different air masses within the boundary layer.
Table 1. Location, Time, and Altitude During Periods When HNO3 and Fine-Particle Mixing Ratios Were Anticorrelateda
Location (Latitude, Longitude)
Fine Particle, ppbv
Fine Particle:HNO3 Correlation
The range of RH, temperature, HNO3, and fine particles during these periods are given in columns 4–7. The average ±1 standard deviation of the CO mixing ratios are also given (column 8). The fine particle:HNO3 slopes are from a linear least squares fit with 95% confidence intervals. Time is in LST.
Mojave Desert (34.5, 117.0)
25 April, 10:50–11:10 AM
176 ± 8
−1.1 ± 0.2
San Joaquin Valley (37.1, 120.4)
8 May, 3:35–3:40 PM
161 ± 7
−0.9 ± 0.1
Rubidoux (34.0, 117.3)
13 May, 12:15–12:23 PM
305 ± 63
−1.58 ± 0.06
Redlands (34.1, 117.2)
13 May, 12:34–12:45 PM
292 ± 59
−2.5 ± 0.2
San Bernardino Mountains (34.1, 116.9)
13 May, 2:49–2:54 PM
215 ± 51
−10 ± 1
Mojave Desert (33.9, 116.5)
13 May, 3:02–3:14 PM
175 ± 13
−0.37 ± 0.03
San Joaquin Valley (36.4, 119.1)
17 May, 4:20–4:40 PM
210 ± 14
−1.12 ± 0.09
 The conversion of gas phase HNO3 to fine-particle mass was observed on several other altitude profiles in other locations, when ambient conditions differed considerably from the San Joaquin Valley (Table 1). Similar vertical gradients in HNO3 mixing ratios and particle concentrations were observed in altitude profiles over the Mojave Desert east of Los Angeles. On 13 May, a spiral profile was performed east of Palm Springs. From 0.2 km to the top of the boundary layer at 3.6 km (Figure 4), HNO3 decreased dramatically and fine particles increased at higher altitudes. Ammonium nitrate formation was indicated by enhancements of fine-particulate NO3− and NH4+ of ∼0.8 ppbv that was coincident with the HNO3 depletion and increase in fine particles. During this time, the sulfate concentration was unchanged at approximately 0.2 ppbv. The fine-particle volume was approximately 15% greater than the coarse-particle volume, and fine particles with diameters between 0.2–0.6 μm dominated the fine-particle mass.
 The boundary layer height was indicated by the weak (<1°C) inversion in the temperature profile that began at approximately 3.6 km (Figure 4a). The RH increased with altitude within the boundary layer, but remained well below the deliquescence RH, which increased from ∼58% at 0.2 km (34°C ambient temperature) to 75% at 4 km (1°C ambient temperature). Below the temperature inversion, the temperature lapse rate was approximately adiabatic at −9.7°C/km. When the aircraft entered the free troposphere, there was a sharp change in trace gas mixing ratios. In the free troposphere above 3.9 km altitude, O3 mixing ratios were 42 ± 2 ppbv, and CO mixing ratios were 89 ± 4 ppbv. Between 0.2–3.6 km, O3 and CO were nearly constant and were approximately twice as high as the values in the free troposphere, as shown in Figure 4b.
 From 0.2–3.6 km above the Mojave Desert, large vertical gradients in fine-particle mass and HNO3 were caused by gas-to-particle conversion. Neither O3 nor CO mixing ratios (Figure 4b) exhibited the large vertical variations that were apparent in the HNO3 and fine-particle mixing ratios (Figure 4c). HNO3 and fine particles were anticorrelated between 1.9 and 3.5 km, whereas they were positively correlated at lower altitudes in the boundary layer and at higher altitudes in the free troposphere. In this example, HNO3 and fine particles were less strongly anticorrelated (r = −0.78), and the linear least squares fit to the fine particle to HNO3 relationship had a slope of −0.37. In this case, mixing of differing air masses may have reduced both the anticorrelation and the slope of the fine particle to HNO3 relationship. NO3− was enhanced when HNO3 was depleted, though the 0.8 ppbv NO3− enhancement was much smaller than the ∼2 ppbv depletion of HNO3. Since coarse-particle composition was not measured, it is not possible to exclude the possibility that some nitrate growth occurred in the coarse mode. The gas-to-particle conversion over the desert occurred at considerably lower RH, higher altitude, and lower ambient temperature (Table 1, row 6) compared to the San Joaquin Valley (Table 1, row 7). Similarly, gas-to-particle conversion during the altitude profile over the Mojave Desert on April 25 (Table 1, row 1) occurred at low temperatures, although the RH was much greater than the example described above. The conversion of gas phase HNO3 to ammonium nitrate was observed over a range of locations, ambient temperatures, and relative humidities.
3.2. Los Angeles Basin
 Gas-to-particle conversion was also observed at lower altitudes in the boundary layer, especially in air masses located downwind from large NH3 emissions. The flight in the Los Angeles Basin on 13 May 2002 examined both urban and agricultural source regions of pollutants and locations at the periphery of the basin where pollutants were likely to be transported [McElroy and Smith, 1986]. In the Los Angeles Basin, a sea breeze and coastal temperature inversion often determine the distribution of pollutants. A strong temperature inversion (∼7°C) [Schultz and Warner, 1982] caused by the cooling of air over the Pacific Ocean often exists near the coast, and an afternoon sea breeze regularly carries pollutants emitted in the urban area eastward. As air is advected inland, it is warmed such that the temperature inversion weakens and the boundary layer grows. The transport of pollutants from the Los Angeles Basin is often inhibited by the temperature inversion and the surrounding San Gabriel and San Bernardino Mountains. When the mixed layer stabilizes in the late afternoon, layers of pollutants can remain aloft above the boundary layer and within the temperature inversion. During sea breeze conditions, ventilation occurs when an air mass is lifted over the mountains at the eastern (downwind) end of the basin [Lu and Turco, 1995].
 Typical sea breeze conditions existed during the study period. The aircraft entered the Los Angeles basin at approximately 12:00 PM local standard time (LST) and departed at approximately 4:00 PM LST. The prevailing winds were from the west at 5 ± 2 m/s. The aircraft flew at an altitude of approximately 700 m over the valley regions and up to 4 km during spiral altitude profiles and over the San Gabriel and San Bernardino Mountains. The flight track and locations discussed in the text are shown in Figure 5. During the flight, HNO3, NOx, NOy, O3, and CO mixing ratios were elevated downwind of the Los Angeles urban area (the loop in the flight track at 34.0 N, −118.0 W) where mixing ratios reached 15, 47, 67, 93, and 700 ppbv, respectively. On this clear late spring afternoon with steady winds from the west, a small fraction of the emitted NOx was oxidized, and most of the NOx oxidation produced HNO3. The highest O3 (113 ppbv) and HNO3 (17 ppbv) levels were observed downwind from the Los Angeles urban area near Ontario. While slightly more photochemical processing of emissions had occurred at this inland location compared to locations closer to the urban center, the ratios of CO/NOy and NOx/NOy did not vary dramatically across the basin. Throughout the basin, the NOx:NOy correlation slope was approximately 0.7, the HNO3:NOy correlation slope was approximately 0.2, and O3 mixing ratios were generally 80 ± 10 ppbv. The ratios of these trace gases were relatively constant across the basin because they were all strongly influenced by the same tailpipe area source.
 In contrast to the trace gases, particle concentrations exhibited substantial variability within the Los Angeles Basin. Figure 6 shows fine-particle mixing ratio versus CO for the flight track shown in Figure 5. In the western part of the Los Angeles Basin, fine particles and CO were positively correlated, whereas fine particles and CO were often uncorrelated in the eastern part of the Los Angeles Basin. This indicates that many fine particles in the eastern Los Angeles Basin had a source different from the source of CO. Additionally, NO3− was especially enhanced in the eastern part of the basin, and the highest NO3− was 5.0 ppbv (12.7 μg m−3) inland near Rubidoux. Nitrate was the most abundant of the measured water-soluble inorganic fine-aerosol species, and enhancements in NH4+ were coincident with the NO3− enhancements. Fine-sulfate aerosol was always less than 0.2 ppbv (1 μg m−3) over the Los Angeles Basin and much less than NO3− or NH4+ in the polluted areas.
 The gas-to-particle conversion processes that caused the elevated particle layers observed during the altitude profiles also contributed to the particle mass enhancements in the eastern Los Angeles Basin. The circles along the flight track shown in Figure 5 indicate NO3− measurements and are sized according to the fraction of nitrate in the particle phase relative to the total nitrate [NO3−/(NO3− + HNO3)]. In the western part of the Los Angeles Basin, most of the total nitrate was in the gas phase as HNO3, whereas most of the total nitrate was in the particle phase as NO3− in many areas in the eastern part of the Basin. Downwind of the regions of agricultural NH3 emissions near Rubidoux in the eastern part of the Los Angeles Basin [Russell and Cass, 1986], ammonium nitrate was formed from the reaction of gas phase HNO3 with NH3.
 The effects of NH3 emission upon the conversion of gas phase HNO3 to particulate ammonium nitrate in the eastern Los Angeles Basin is evident in the segment of flight shown in Figure 7. Substantial enhancements in fine particles, NH4+, and NO3−, and the depletion of HNO3, occurred after the aircraft flew downwind of the livestock operations located in the Rubidoux area. O3 was nearly constant at 80 ± 2 ppbv, while CO varied considerably between 150–450 ppbv. Neither CO nor O3 were correlated with the enhancements in particles or the depletion of HNO3. Fine-particulate magnesium, chloride, nitrite, potassium, and sodium were <0.05 μg m−3, and calcium was measured to be 0.15 μg m−3, indicating a negligible contribution of mineral and sea-salt particles to the fine-particle mass. Most of the fine-particle mass was in the accumulation mode with diameters between 0.2 and 0.6 μm. Coarse-particle volume was approximately 20% larger than the fine-particle volume.
 Ammonium nitrate formation is demonstrated by comparing the magnitude of the HNO3 depletion to the NO3− and NH4+ enhancement. If ammonium nitrate were formed from HNO3 depletion, the molar enhancements of NO3− and NH4+ would be equivalent to the reduction of gas phase HNO3. When HNO3 values decreased from 6 ppbv to 1.5 ppbv near Rubidoux, the NO3− enhancement was 4.8 ppbv. Within the combined uncertainties of the measurements, the HNO3 depletion was equivalent to the enhancement in fine-particulate NO3−. The NH4+ molar enhancements were 17 % larger than the NO3− enhancements (Figure 7). Within the 20% inaccuracy of the bulk particle composition measurements, these results are consistent with fine-particulate ammonium nitrate formation. Although coarse-particle composition was not measured here, the bulk fine-particle composition measurements demonstrated that the conversion of HNO3 to ammonium nitrate occurred mostly on fine particles here.
 Particle growth is studied further by examining the relationship between fine particles and HNO3 (Figure 8). When NO3− was enhanced, the linear least squares fit to the fine particle versus HNO3 data had a correlation coefficient of −0.92 and a slope of −1.58. The measured enhancement in fine particles was greater than can be accounted for solely from ammonium nitrate formation. Similarly, in the eastern Los Angeles Basin near Redlands (Table 1, row 4), the correlation slope between fine particles and HNO3 was significantly steeper than −1. Secondary organic aerosol formation can contribute to the fine-particle mixing ratio to produce this difference. Organic material has been shown to account for a substantial fraction of fine-particle mass in this area [e.g., Chow et al., 1994].
 Although gas phase NH3 was not measured during this study, NH3 concentration can be calculated when ammonium nitrate particles were detected. During the measurements near Rubidoux (Table 1, row 3), the RH was considerably lower than the 60% deliquescence relative humidity. Using the measured gas phase HNO3 mixing ratio (1 ppbv) that existed with the ammonium nitrate aerosol and the dissociation constant of 60 ppbv2 at 28°C, gas phase NH3 is calculated to be approximately 60 ppbv here, consistent with previous reports of large NH3 emissions and concentrations in this region of the Los Angeles Basin [Russell and Cass, 1986].
 Gas-to-particle conversion was observed over a wide range of ambient conditions at several locations in the Los Angeles Basin. For example, at higher altitudes and lower temperatures, nearly all HNO3 was removed from an air mass when the aircraft flew over the San Bernardino Mountains from Cajon Pass to Palm Springs (Table 1, row 5). HNO3 decreased from 2 ppbv to below 20 pptv. HNO3 levels were below instrument detection limits, and over 99% of the HNO3 was removed from this air mass. High NO3− levels (>2 ppbv) coincident with low HNO3 levels (<20 pptv) demonstrated that the HNO3 instrument discriminated against aerosol NO3− by >99%. The variability in trace gas and particle mixing ratios in this region make a quantitative analysis of the observed gas-to-particle conversion less precise. The HNO3 depletion varied from 2–5 ppbv, and the NO3− enhancement varied from 1.7–3.4 ppbv, in qualitative agreement with the measured HNO3 depletion. Although the RH was high compared to the low altitude measurements near Rubidoux, it was still well below the 74% deliquescence relative humidity. The dissociation constant at these lower temperatures was 0.1 ppbv2, and nearly all NO3− was in the particle phase. The fine-particle volume was approximately twice as large as the coarse-particle volume, so that fine particles accounted for most of the particle mass. The anticorrelation between fine particles and HNO3 had a slope of −10. This suggests that the fine particles measured here include substantial contributions from other species, and are less likely to be pure ammonium nitrate particles compared to the particles sampled near Rubidoux.
3.3. San Joaquin Valley
 Spatial variability in gas-to-particle conversion was observed on 8 May 2002 during level flight at 0.55 km in the boundary layer over the Pacific Ocean, San Francisco Bay area, and northern San Joaquin Valley (Figure 1). In the San Joaquin Valley, winds were from 20° west of north at 6 ± 1 m s−1. The measurements over the San Joaquin Valley were enhanced in aerosol NO3− and NH4+ and fine particles compared to the measurements over the ocean and San Francisco. Over the ocean NO3− was <0.07 ppbv (0.2 μg m−3), over San Francisco NO3− was <0.12 ppbv (0.3 μg m−3), while NO3− levels exceeded 3.3 ppbv (1.3 μg m−3) over the San Joaquin Valley from Stockton to Merced. Sulfate levels over the San Joaquin Valley and San Francisco urban area were ∼0.1 ppbv (0.4 μg m−3) and were comparable to the levels measured over the Los Angeles Basin. HNO3 levels were higher over San Francisco (1.5–4 ppbv) than over the San Joaquin Valley (1–2 ppbv). Over San Francisco, the fine- and coarse-particle volumes were approximately equivalent, whereas the coarse-particle volume was approximately four times the fine-particle volume over the San Joaquin Valley. These results are consistent with previous measurements in the summertime in rural areas in the San Joaquin Valley, where coarse particles containing crustal materials have been shown to sometimes account for the majority of the particle mass [Chow et al., 1996].
 Although the HNO3, NO3−, and fine-particle mixing ratios were approximately an order of magnitude lower over the San Joaquin Valley than measured over the Los Angeles Basin, and fine particles did not account for the majority of the particle mass, the same gas-to-particle conversion process was observed. When fine-particulate NO3− was substantially enhanced, HNO3 was anticorrelated with fine particles with a slope of −0.9 (Table 1, row 2), and NH4+ was also enhanced. This result supports the assumption that the enhancement in fine particles over the San Joaquin Valley was from ammonium nitrate. Furthermore, gas-to-particle conversion was not observed over the Pacific Ocean or San Francisco, but only over the San Joaquin Valley, which is influenced by agricultural NH3 emissions.
3.4. Single Particle Measurements
 Single particle measurements from a laser ionization mass spectrometer (PALMS) are used to examine the abundance of organic aerosol relative to ammonium nitrate during the measurements in the boundary layer over California. The ion signal at mass 30 corresponded to NO+ resulting from the laser ionization of particles containing NH4+ or NO3− [Murphy and Thomson, 1997]. Laser ionization of organic particulate matter caused a signal in positive mass spectra at mass 12 from ionized carbon atoms (C+). Relative contributions from organic matter and ammonium nitrate are assessed qualitatively by examining the fraction of the total ion current from C+ and NO+, respectively. During most of the flight over the Los Angeles Basin on 13 May the signal from C+ accounted for a larger fraction of the ion signal than the signal from NO+. Only downwind from regions of large NH3 emissions was the ammonium nitrate signal (NO+) dominant over the organic aerosol signal (C+). Additionally, in the eastern Los Angeles Basin where ammonium nitrate formation was large, the absolute C+ signal was enhanced relative to the western Los Angeles Basin, where most of the nitrate was in the gas phase. This enhancement of the absolute C+ signal downwind from NH3 emissions is consistent with the formation of organic aerosols from organic acids (formic and acetic acids, for example).
 Large variations in aerosol composition apparent in the single particle measurements coincided with the large variations in HNO3 and fine-particle mixing ratios. Figure 9 shows the average fraction of the total ion current from NO+ (ammonium nitrate) and C+ (organics) as a function of fine-particle mixing ratio for periods when gas-to-particle conversion was observed (Table 1). During these periods when HNO3 was depleted and NO3−, NH4+, and fine-particle mass were enhanced, the average fraction of the total ion current from NO+ was positively correlated with the fine-particle mixing ratio (Figure 9a). At the highest fine-particle values (and lowest HNO3), NO+ accounted for the majority (∼80%) of the ion current, consistent with particles that contained primarily ammonium nitrate. The relative contribution from organics to the ion signal (Figure 9b) decreased at high fine-particle mixing ratios, but it did not drop to zero. Organic material always contributed to the particle mass. At the lowest fine-particle mixing ratios, when HNO3 depletion was small, the signal from particulate organics dominated that from ammonium nitrate (Figure 9b). These results are consistent with the presence of organic aerosols that could cause fine particle:HNO3 slopes to be steeper than −1. However, the range of measured slopes (Table 1) is not explained by the differences in particle composition between locations measured by the PALMS instrument.
 These results agree qualitatively with previous studies that have shown substantial contributions from secondary organic aerosols in addition to ammonium nitrate to the fine-particle mass near Riverside California in the eastern part of the Los Angeles Basin [Chow et al., 1994; Kleeman et al., 1999; Liu et al., 2000]. Although the particle mass measured here in the spring is lower than was reported in summer, fall, and winter measurements cited above, the particle composition appears to be consistent.
 The particle composition over the San Joaquin Valley on 17 May differed from the particle composition observed over the Mojave Desert and Los Angeles Basin. Over the San Joaquin Valley, the fraction of the ion current from C+ was larger and that from NO+ was smaller, indicating that organic carbon accounted for a larger fraction of the particle mass. Differences in particle size distributions between the measurements in the San Joaquin Valley and other locations were also apparent. Coarse particles accounted for the largest fraction of the particle mass over the San Joaquin Valley (discussed above). These results agree with previous measurements that have shown that large contributions to particle mass come from both secondary organic aerosol and coarse particles containing crustal materials in the San Joaquin Valley [Chow et al., 1996].
 The spatial and vertical variations in the observed depletion of HNO3 and enhancement of particle mass reveal several important features in the chemical processes that determine the distribution of pollutants in California. In previous studies of pollutant layers over California [Blumenthal et al., 1978; McElroy and Smith, 1986; Lu and Turco, 1995; Collins et al., 2000], meteorological effects involving the coupling of the sea breeze with diurnal changes in the stabilization of the mixed layer height were found to cause pollution layers within and above the temperature inversion at the top of the boundary layer. The layers of particles reported here occurred within the boundary layer and were apparent only in species associated with the particle formation. The layers observed here reveal the importance of gas-to-particle conversion processes in addition to meteorology for explaining the spatial variability of gas and particulate phase pollutants in this region. Variations in ammonium nitrate abundance that correspond to changes in the dissociation constant with diurnal and seasonal temperature fluctuations have been previously studied [e.g., Chow et al., 1994; Solomon et al., 1992; Weber et al., 2003]. The temperature dependence of the partitioning of ammonium nitrate between the gas and particle phases, which caused ammonium nitrate enhancements during colder periods of the day and colder seasons in those studies, also cause the vertical variations in ammonium nitrate observed here. Since the processes observed here occurred within and below weak temperature inversions over a wide range of locations and conditions, they are expected to occur regularly in these regions within the planetary boundary layer.
 At the inland locations studied here, the meteorological conditions observed in the afternoon were typical. The elevated boundary layer heights, weak temperature inversions, and adiabatic temperature lapse rates characterized a boundary layer where pollutants are rapidly vertically mixed [Schultz and Warner, 1982]. The presence of large variations in particle mass and condensable gases, but not in directly emitted (e.g., CO) or other photochemically produced (e.g., O3) species, demonstrated that the gas-to-particle conversion process occurred faster than boundary layer mixing times. The time for air to cycle between the bottom and top of the boundary layer is estimated to be approximately 10–20 min, calculated using typical vertical kinematic heat fluxes and measured ambient temperature and boundary layer heights [Stull, 1988]. The timescales for fine-particulate ammonium nitrate formation is expected to be rapid (∼minutes) compared to other time-dependent changes in the atmosphere [Meng and Seinfeld, 1996; Dassios and Pandis, 1999]. Our observations confirm the rapid equilibration time for fine ammonium nitrate aerosols and illustrate that a boundary layer that is well mixed with respect to some emitted and secondary pollutants may contain substantial variations in condensable gases and fine particles. The observation that essentially all HNO3 was removed from an air mass at high altitudes indicates that there is not excess HNO3 for particle formation at these locations. Although NH3 was not measured here, the vertical profiles of fine particles and HNO3 indicate that NH3 (when not in substantial excess) would be depleted at higher elevations in a manner similar to HNO3.
 The conversion of HNO3 to particulate matter at higher elevations is important to the distribution and transport of nitrate in the atmosphere. Gas phase HNO3 is lost rapidly to dry and wet deposition, while the deposition velocity for fine particles is less than one-tenth as fast [Warneck, 1988]. Since the gas-to-particle conversion processes were enhanced at higher elevations where the colder temperatures reduced the dissociation constant for ammonium nitrate aerosol by nearly two orders of magnitude compared to the surface, these processes may be important when air masses are lofted over terrain during transport. For example, when an air mass is transported over mountains surrounding the Los Angeles Basin, it is likely to be enhanced in particulate NO3− relative to gas phase HNO3. The conversion of gas phase HNO3 to the particle phase may increase the lifetime, and hence the transport, of NO3− through the boundary layer. Furthermore, the elevated layers of ammonium nitrate reported here may not be captured by ground-based observations. Hence it may be important to determine the spatial extent of these layers to accurately determine the climatic effects of anthropogenic aerosols.
 The formation of ammonium nitrate particulate matter was observed at lower altitudes only in the San Joaquin Valley and the Los Angeles Basin, downwind from regions where NH3 emissions from livestock activities are known to be large [Russell and Cass, 1986; Kleeman et al., 1999; Chow et al., 1996]. In flights within the boundary layer over the Sacramento Valley in northern California, the San Francisco Bay and Monterey Bay areas, and other altitude profiles over northern California and the north eastern Pacific Ocean, HNO3 depletion associated with ammonium nitrate enhancement was not observed. In the western part of the Los Angeles Basin, where NH3 emissions are lower than in the eastern part of the basin, most of the nitrate was in the gas phase. At lower altitudes and warmer temperatures, the gas-particle processes that cause the depletion of HNO3 were apparent only in regions affected by large agricultural NH3 emissions. The partitioning of nitrate into the particle phase downwind of NH3 emissions observed here is in agreement with previous studies that used multiple ground-based monitoring stations throughout the Los Angeles Basin to examine the composition and spatial distribution of aerosol [Solomon et al., 1992; Chow et al., 1994; Kleeman et al., 1999].
 The loss of gas-phase HNO3 to particles is also important for understanding O3 photochemistry. The efficiency of O3 production from NOx photochemistry has been described by the ratio of O3 to NOy-NOx, where the quantity NOy-NOx represents the products of NOx oxidation [Trainer et al., 1993]. If NOy measurements include only gas phase species, then these gas-to-particle conversion processes that remove HNO3 from the gas phase would lower the measured NOx oxidation products and spuriously increase the calculated O3 production efficiency.
 Using an aircraft platform with fast response measurements of trace gases, particle size distributions, and particle composition, spatial variations in trace gas mixing ratios and particle mass have been studied. These aircraft-based measurements demonstrate that there exists an important vertical component to the spatial variability of ammonium nitrate aerosol. In altitude profiles over the Los Angeles Basin, San Joaquin Valley, and Mojave Desert, CO and O3 mixing ratios were nearly constant within the boundary layer, whereas HNO3 was depleted and fine particles were enhanced at higher altitudes. Layers of particles observed at higher altitudes within the boundary layer were caused by the temperature dependence to the partitioning of ammonium nitrate between the gas and particle phases. At higher altitudes within the boundary layer, where ambient temperatures were lower than at the surface, the dissociation constant for ammonium nitrate was reduced, nitrate was driven into the particle phase, and gas phase HNO3 was depleted. The relationship between fine particle and HNO3 mixing ratios demonstrated that fine particle enhancements were entirely accounted for by the conversion of NH3 and HNO3 to ammonium nitrate in many altitude profiles. The vertical gradients in the observed HNO3 and particle mixing ratios, but not in other trace gases, indicated that the equilibration time for ammonium nitrate aerosol formation was faster than boundary layer mixing times.
 The formation of fine-particulate ammonium nitrate and coincident depletion of gas phase HNO3 was also observed in regions influenced by agricultural NH3 emissions. In the eastern Los Angeles Basin downwind from livestock facilities, most nitrates were in the particle phase, whereas gas-phase HNO3 was substantially greater than fine-particulate nitrate upwind from these NH3 sources. Excess HNO3 was not present at higher elevations downwind from regions with large NH3 emissions, as indicated by measurements in an air mass that was completely depleted of HNO3 but rich in NO3−. When HNO3 depletion was observed in the Los Angeles Basin, fine-particulate nitrate accounted for all the HNO3 depletion, but ammonium nitrate did not entirely account for the enhancement in fine-particle mixing ratio. Single particle measurements demonstrated that organic aerosols also contributed to the enhancements in fine-particle mass in these regions characterized by large NH3 emissions.
 We thank the NOAA WP-3 flight and support crew and are especially grateful for their extraordinary efforts in achieving the flight in the Los Angeles Basin. This work was performed while JBN held a National Research Council Research Associateship Award at the NOAA Aeronomy Lab.