Browsing by Author "Collett, Jeffrey L., Jr., author"
Now showing 1 - 12 of 12
Results Per Page
Sort Options
Item Open Access Aerosol black carbon measurements in Fort Collins, Colorado(Colorado State University. Libraries, 1999-03) Calame, Lorraine, author; Collett, Jeffrey L., Jr., authorAerosol black carbon concentrations were measured in Fort Collins, a small city in northern Colorado. A Magee Scientific aethalometer was used to obtain real time data. In 1997 a gradual build-up in monthly BC concentrations over the late spring and summer was observed to peak in August and then decline. The number of days featuring high concentrations (> 2000 ng/m3 for 30 minutes) was also observed to rise and fall with an August peak. The monthly trends in aerosol black carbon concentrations suggest some seasonality. Observations of aerosol black carbon concentrations made during this study reveal monthly mean concentrations that vary from 487 to 1004 ng/m3. The BC concentration timelines showed there are usually two daily peaks in the data. These peaks usually occur between 0600 - 0800 and 1600 - 2100, hours associated with peak traffic. Concentrations of BC were found to correlate reasonably well with CO concentrations over time scales of up to a week. The covariance in our data set suggests they were emitted from a common set of combustion sources. The results of this study indicate there is little relationship between measured BC concentrations and long range transport patterns, as back-trajectory results indicated that long range transport of air masses from particular regions does not influence whether BC concentrations in Ft. Collins are high or low. The observed BC concentrations tended to decrease with increasing wind speed which is consistent with emissions from a local source that would tend to be diluted at higher wind speeds. On shorter timescales of minutes to hours, the study revealed the importance of individual sources on BC concentrations. Passages of diesel locomotives near the site were shown to increase BC concentrations on many occasions when local winds were favorable to transporting locomotive emissions to the aerosol sample site. The impact of this source, however, was fairly limited in its duration, with concentration spikes typically lasting only about 10 minutes. Train passage also exerted an indirect effect on BC concentrations, perhaps associated with a period of increased BC emissions by vehicles idled at train crossings. Fires were also shown to influence BC concentrations during the study. The large fires associated with the July 1997 flood and smaller fires associated with a CSU student riot were observed to increase BC concentrations. The number of fires during the study period, however, was very low. Impacts of two prescribed bums along the Front Range on Ft. Collins BC concentrations were observed to be fairly minor. Correlation of BC concentrations with "total" particle number concentrations (as measured by a Condensation Nucleus Counter), and number concentrations of particles with diameters greater than 0.3 µm (as measured by a Climet Optical Particle Counter) were extremely weak on long time scales and often weak on time scales as short as a day. This reflected the diverse sources of the aerosols. Some of these were probably primary combustion aerosol (and therefore are likely to correlate with BC), but others were the secondary aerosol fraction of CN or particles larger than 0.3 µm that are not expected to correlate strongly with BC concentrations. BC concentrations were independent of the diurnal pressure patterns, but were affected by synoptic pressure patterns. Frontal passages were marked by low BC concentrations. BC concentrations were not found to be depressed on summer days featuring precipitation, as the duration of summer thunderstorms is typically too short to scavenge enough aerosol particles to significantly influence the daily average BC concentration. In addition, primary aerosol particles that are generated locally by vehicle emissions would be replenished rapidly in the atmosphere following precipitation. Absorption coefficients were derived from the aethalometer and the difference between transmissometer and nephelometer data. The monthly mean values ranged from 0.005 - 0.009 km-' for the aethalometer, and 0.036 - 0.038 km-1 for the transmissometer and nephelometer. The aethalometer derived absorption coefficients are on the order of 10 - 20% of the absorption coefficients determined from the paired nephelometer and transmissometer measurements, and only 6 - 11% of the total extinction measured by the transmissometer.Item Open Access Application of a tracer technique to study sulfur dioxide oxidation in cloud drops as a function of drop size(Colorado State University. Libraries, 2000-08) Reilly, Jill E., author; Collett, Jeffrey L., Jr., authorItem Open Access Characterization of carbonaceous aerosol during the Big Bend Regional Aerosol and Visibility Observational study(Colorado State University. Libraries, 2001-12) Brown, Steven G., author; Herckes, Pierre, author; Kreidenweis, Sonia M., author; Collett, Jeffrey L., Jr., authorThe Big Bend Regional Aerosol and Visibility Observational (BRAVO) study was a four month field campaign (July-October 1999) to investigate aerosol particle properties, sources, and impacts on regional visibility in Big Bend National Park, Texas. Daily PM2.5 aerosol samples were collected on pre-fired quartz fiber filters for detailed molecular analysis of the aerosol organic carbon fraction. Aerosol black carbon concentrations during BRAVO were measured with an aethalometer. The molecular characterization of the organic carbon fraction of aerosol present during the BRAVO study was performed using gas chromatography - mass spectroscopy (GC-MS). Organic carbon concentrations on individual days were too low for a detailed analysis by GC-MS. Therefore, multi-day composite samples, selected based on common air mass trajectories and temporal proximity, were extracted and analyzed for numerous compounds, including n-alkanes, polycyclic aromatic hydrocarbons (PAH), and alkanoic acids. Low alkane Carbon Preference Indices (CPIs) during July through September reflect similar concentrations of n-alkanes containing odd and even numbers of carbon atoms and indicate that anthropogenic emissions were important contributors to carbonaceous aerosol during this period, when air masses generally were advected from the east over Texas and Mexico. In October, CPIs increased, reflecting increased influence of odd carbon numbered alkanes and suggesting a predominant biogenic aerosol influence with air masses arriving from the north and the south. Plant wax contributions to odd carbon number alkanes (C25-C33) were estimated to range between 26% and 78%, with the highest contributions occurring in October with air masses arriving from the north and south. Periods with transport from eastern Texas and northeastern Mexico had much smaller plant wax contributions. Alkanoic acids were the most abundant compound class, with CPIs that were high throughout the study. The high acid CPI suggests that the alkanoic acids may be largely biogenic in origin, a finding consistent with other studies. Caution is required in interpreting the acid CPI, however, as alkanoic acids can also be formed as secondary products of atmospheric reactions. Polycyclic aromatic hydrocarbons (P AH) were usually not found in abundance, suggesting that upwind combustion emissions were not important contributors to carbonaceous aerosol or that P AH were removed by reaction or deposition in transit. Higher P AH concentrations during one period indicated a more significant contribution from fresh combustion emissions. Molecular source tracer (hopanes for vehicle emissions, levoglucosan for wood combustion, cholesterol for meat cooking) concentrations were generally not detected. Based on analytical detection limits for these species, it was estimated that wood smoke contributed no more than 1% of the total Organic Carbon (OC) present, vehicle exhaust contributed no more than 4%, and smoke from meat cooking contributed less than 13%. The presence of other wood smoke tracer molecules, however, suggests a possibly greater influence from wood combustion and possible chemical instability of levoglucosan during multi-day transport in an acidic atmosphere. Several observations suggest that secondary production contributed significantly to BRAVO carbonaceous aerosol. Examination of ratios of aerosol organic carbon to elemental carbon indicates that secondary organic aerosol may have contributed between 45% and 90% of the total BRAVO aerosol organic carbon. High ratios of saturated/unsaturated C18 acids, an abundance of nonanoic acid, and high concentrations of 6,10,14 trimethylpentadecan-2-one (an indicator of secondary aerosol production from vegetation emissions) all support the conclusion that secondary aerosol formation was important in the region. Total black carbon (BC) concentrations ranged from below detection limit (71 ng/m3) to 267 ng/m3, averaging 129 ng/m3. Fine (< 1 μm) aerosol BC concentrations averaged 114 ng/m3, and comprised 89% of the total BC. BC concentrations correlated reasonably well with aerosol sulfate concentrations, suggesting similar source regions for these species.Item Open Access Chemical heterogeneity among cloud drop populations and its influence on aerosol processing in winter clouds(Colorado State University. Libraries, 1998-12) Xu, Gang, author; Collett, Jeffrey L., Jr., authorDrop-size resolved measurements of winter cloud composition in the Rocky mountains of northern Colorado revealed significant variations of cloud drop pH, ion (SO/, N03-, NH/, Ca2+) concentrations, and concentrations of trace metal catalysts (Fe and Mn) with drop size. The observed chemical heterogeneity across the cloud drop size spectrum was used to evaluate its influence on rates of in-cloud sulfate production and scavenging of major inorganic aerosol components by precipitation. The findings indicates that the size-dependent chemical composition of clouds tends to enhance aqueous sulfate production rates and to reduce the efficiency with which accumulation mode aerosol species are scavenged by precipitation. Simultaneous measurements of snow chemical composition and the degree of cloud drop capture by snow crystals (riming) revealed a significant positive correlation between snow composition and the extent of ice crystal riming in two of four cases studied. In the remaining cases it was found that the variations of snow composition were significantly correlated with air mass changes. Measurements of the size spectra of cloud drops attached on the surface of snow crystals showed that 10-17 µm and >17 µm cloud drop size fractions contributed significantly to accreted drop mass; little accreted mass was associated with drops smaller than 10 µm diameter.Item Open Access Cloud chemical heterogeneity and its influence on aqueous sulfur (IV) oxidation(Colorado State University. Libraries, 1997-06) Rao, Xin, author; Collett, Jeffrey L., Jr., authorDifferences in chemical composition among cloud and fog drops of diverse sizes were investigated at several locations across the United States. Chemical species including acidity, sulfur (IV), hydrogen peroxide, formaldehyde, hydroxymethanesulfonate (HMS) and trace metals iron and manganese were measured. The study examined coastal stratus and stratocumulus clouds in southern California and northern Oregon, frontal and orographic clouds at Mt. Mitchell, North Carolina and Whiteface Mountain, New York, and radiation fogs in California's San Joaquin Valley. Samples were collected with three cloud samplers capable of partitioning the cloud drop size spectrum into two or three independent drop size fractions. Measurements of pH variations within natural cloud drop populations reveal that small drops are often more acidic than large drops. Differences between small and large cloud drop acidities as large as two pH units were observed, although differences were generally below one pH unit. The chemical heterogeneity can significantly enhance oxidation of sulfur dioxide to sulfate within clouds, relative to oxidation rates predicted from the average cloudwater composition. Trace metal concentrations were found to vary with drop size in clouds and fogs sampled at a variety of U.S. locations. Significantly higher concentrations of total iron and manganese were found in large drops in clouds sampled at Mt. Mitchell, North Carolina, and along the southern California coast, while small drops were often enriched in concentrations of iron and manganese in fogs sampled in California's San Joaquin Valley and coastal clouds sampled at Angora Pk., Oregon. Iron speciation measurements in San Joaquin Valley fogs revealed that dissolved iron in small fog drops was present almost entirely as Fe (III). The observed size dependence of trace metal concentrations in cloud and fog drops is expected to influence in-cloud S(IV) oxidation rates as well. Effects of chemical heterogeneity on overall in-cloud S(IV) oxidation rates will largely depend on contributions of the different oxidation paths. Errors in predicting sulfur oxidation rates based on average cloud drop compositions are smallest when abundant hydrogen peroxide is present, for example in the summer clouds at La Jolla Pk., California and at Mt. Mitchell, North Carolina. About 84 percent of the samples are calculated to experience little enhancement in S(IV) oxidation, due to the dominance of the02 path. Approximately 9 percent of the samples are calculated to experience oxidation rate enhancement between I0 and 30%, while 7 percent of the samples are calculated to experience oxidation rate enhancement of30% or more. Effects of chemical heterogeneity on enhancements in sulfur oxidation rates are likely to be strong when (I) hydrogen peroxide concentrations are low, for example the radiation fog in California's San Joaquin Valley, where the calculated enhancement factors range from 1.10 to 1.65, or (2) the droplet pH is high enough to support rapid S(IV) oxidation by ozone and metal-catalyzed S(IV) autooxidation, for example in relatively pristine environments like Angora Pk., Oregon, where the calculated enhancement factors range from 1.02 to 2.0. We expect real clouds to contain more than two chemically distinct drop populations. A wide distribution of drop compositions can support even faster sulfur oxidation rates.Item Open Access Design and testing of a new aircraft-based cloud water sampling system(Colorado State University. Libraries, 2002-12) Straub, Derek J., author; Collett, Jeffrey L., Jr., authorExperimental studies of cloud processing mechanisms necessitate the collection of representative samples of cloud water for chemical analysis. In order to provide samples from clouds that are inaccessible from ground-based sampling stations, a new aircraft-based cloud water collection system has been developed . The objective of the design process was to produce an automated collector that can acquire well-characterized cloud water samples and is portable between multiple research aircraft. Issues such as cloud drop shatter and re-entrainment, structural integrity, system size and weight, material compatibility with the anticipated chemical analyses, and ease of use during field operation w re all considered during the design process. The new cloud water collection system utilizes an axial-flow cyclone to centrifugally separate cloud drops from the air stream. Up to seven individual samples can be stored over the course of a single research flight. An analysis of the axial-flow cyclone was performed with a finite volume based computational fluid dynamics (CFD) code. Solutions were obtained for air flow patterns and cloud drop trajectories. The predicted continuous phase (air) velocity field indicates that the axial-flow cyclone generates a strong rotational ow field with a tangential velocity of 85 ms-'. Based on simulations of cloud drop trajectories, centrifugal force in the rotational flow field is sufficient to quickly move entrained cloud drops to the wall of the axial-flow cyclone duct where they can be removed for storage. Collection efficiency as a function of drop size was ascertained and the 50% cut diameter was determined to be approximately 8 microns. An experimental laboratory calibration involving monodisperse fluorescein-tagged drops verified the numerical modeling results. The system was deployed during the Dynamics an Chemistry of Marine Stratocumulus, Phase II (DYCOM -II) field project in July 2001. The DYCOMS-II campaign served as a testing and evaluation program for the system as well as an opportunity to study the chemical composition of stratocumulus clouds in the remote marine environment. Over the course of the project, 50 samples were obtained during seven nighttime and two daytime flights. Sample pH was measured on-site after each flight. Peroxide, formaldehyde, S(IV), trace metals and major ions (Cr, NO3-, so/-, Na+, NH/, K+, ca2+, and Mg2+) were preserved on site and analyzed after the field campaign. The analyses were used to characterize the composition of the sampled clouds and to investigate cloud processing mechanisms, including the potential for rapid aqueous phase oxidation of S(IV) to sulfate.Item Open Access Design, construction and evaluation of the CSU optical fog detector(Colorado State University. Libraries, 2001-07) Emert, Scott E., author; Collett, Jeffrey L., Jr., authorThe goal of this project was to develop an inexpensive cloud/fog detector that could be used to automate sampling equipment at remote (unmanned) cloud/fog research sites. A secondary objective was to test the ability of this sensor to measure/track trends in fog/cloud liquid water content (LWC). This characteristic is important because LWC is a significant indicator of a cloud's ability to process aerosols and gases and changes in LWC often correspond to changes in fog/cloud solute concentration. The following actions were taken to help realize these objectives. An evaluation of the use of commercially available optical components for fog detection has been performed. The research reinforced the need to have an inexpensive cloud/fog detector that could be used to automate sampling equipment at remote (unmanned) cloud/fog research sites. No such instrument is currently available commercially. Requirements for components of the CSU Optical Fog Detector (OFD) were defined. Important factors included transmitter wavelength and modulation characteristics, detector sensitivity, and component stability/durability over a range of environmental conditions. Readily available commercial components were utilized to ensure the sensor could be built economically. Laboratory tests in a glove box filled with artificially generated fog proved that optical components purchased from Banner Engineering were capable of monitoring changes in fog liquid water content (L WC) when operated in a light attenuation mode. After an initial calibration, the signal from the CSU OFD was found to correlate strongly with LWC measured by a Gerber Scientific Particulate Volume Monitor (PVM-100). Theoretical calculations of attenuation of 880 run light passing through a population of fog drops were completed. The results indicated extinction decreases as the drops are shifted to larger sizes (with a fixed LWC and lognormal distribution breadth). Accordingly, the response of the CSU OFD is expected to vary with mean fog/cloud drop size. Numerous fog detector design configurations were tested and the current attenuation design of the CSU optical fog detector was deemed successful in that it provides, at a minimum, an inexpensive switch capable of automating remote fog sensing equipment. It also provides useful information concerning fog LWC. Two calibrated OFD's were compared to PVM LWC measurements during initial field tests of orographic clouds at Storm Peak Laboratory (SPL) in Steamboat Springs, Colorado. The combined results from both OFD's overall time periods yield a regression equation of LWCofd = 0.99 * LWCpvm with a correlation coefficient of 0.92. Tests performed in the absence of fog on top of our laboratory in Fort Collins provided a measure of OFD baseline noise. Analysis of the observed noise yielded a minimum detection limit of 4.4 mg m·3 for the OFD and a comparable value (5.6 mg m·3) for the PVM. The OFD was incorporated in several automated fog sampling systems deployed in California's San Joaquin Valley as part of the California Regional Particulate Air Quality Study (CRP AQS). The OFD performed well as a fog detector and provided some insight into fog LWC. LWC measurements by a PVM and a co-located OFD showed good correlation (R2 = 0.91) and only modest bias (LWCofd = 1.16 LWCpvm) during an extended radiation fog episode.Item Open Access Heterogeneous fog chemistry and S(IV) oxidation in the San Joaquin valley(Colorado State University. Libraries, 1997-10) Hoag, Katherine J., author; Collett, Jeffrey L., Jr., authorChemical and microphysical properties of winter radiation fogs present in the San Joaquin Valley (SJV) of California were measured during December 1995 and January 1996 as a part of the 1995 Integrated Monitoring Study (IMS95). The purpose of the study was to characterize winter particulate air quality problems often observed in this area. Stagnant conditions in the winter often lead to the formation of persistent and dense radiation fogs. These fogs act as processors of particulate matter through particle scavenging and removal in sedimenting fog drops and through particle production via chemical reactions that occur in the aqueous phase. The goal of the fog study portion of IMS95 was to gain more information about the fog drop chemistry as a function of drop size in an effort to more accurately determine the role of fogs in aerosol processing in this region. The major species measured in fogwater during this project were ammonium, nitrate, sulfate, acetate, formate and formaldehyde. The fogwater was also alkaline (median pH = 6.5) compared to a pH of 5.6 expected for atmospheric water in equilibrium with carbon dioxide. Droplet pH affects the amount of SO2 that is absorbed by the droplets and influences the relative importance of different S(IV) oxidation pathways. S(IV) oxidation by ozone was determined to be of primary importance (dominating 88 percent of the time in the southern SJV) because at the high pH values observed it becomes faster than S(IV) oxidation by hydrogen peroxide, which is more commonly investigated, and trace metal catalyzed autooxidation. High fog pH and abundant formaldehyde also led to significant formation of the S(IV) aldehyde complex hydroxymethanesulfonate (HMS). Size resolved fog drop collection revealed differences in fogwater chemical composition as a function of droplet size. In general, smaller drops were more acidic and more concentrated in ionic species than large drops. Calculations have been performed to determine the effect this chemical heterogeneity has on the roles fogs have in aerosol processing. Due to the nonlinearity of the ozone S(IV) oxidation rate, the acidity variations with drop size tend to enhance the oxidation. The observed S(IV) oxidation rates were 1.0 to 7.8 times faster than rates calculated using average fogwater acidity. Variations in other solute concentrations with drop size could affect the removal rates of these species from the air by deposition. Calculations assuming settling as the dominant deposition process estimate that using average fogwater composition of ionic species overestimates their removal rates by as much as a factor of 3.5.Item Open Access How does acidification of drops due to aqueous phase acid production limit aerosol formation in fog drops?(Colorado State University. Libraries, 1998-12) Collett, Jeffrey L., Jr., author; Hoag, Katherine J., author; Pandis, Spyros N., authorItem Open Access Influence of drop size-dependent fog chemistry on aerosol production and deposition in San Joaquin Valley fogs(Colorado State University. Libraries, 1998-10) Collett, Jeffrey L., Jr., author; Hoag, Katherine J., author; Pandis, Spyros N., authorItem Open Access Numerical and experimental performance evaluation of two multi-stage cloud collectors(Colorado State University. Libraries, 1999-01) Straub, Derek J., author; Collett, Jeffrey L., Jr., authorAn evaluation of the collection characteristics of two new multi-stage cascade inertial impactors designed for size-resolved cloud drop collection has been performed. The FROSTY supercooled cloud collector is intended for the collection of supercooled cloud drops in a winter environment in three independent size fractions with stage 50% cut diameters of 15 μm, 10 μm, and 4 μm . The CSU 5-Stage cloud collector is designed for sampling warm clouds in five distinct fractions on five stages that have desired 50% cut diameters of 30, 25, 15 , 10, and 4 μm. Two approaches were selected for the evaluation of the FROSTY and CSU 5-Stage cloud collectors. Numerical simulations provided a visualization of the air flow patterns and drop trajectories through the collectors while experimental laboratory calibrations provided a quantitative analysis of true collection performance. For each of these methods, 50% cut diameters, efficiency curves, and wall losses for each stage of the FROSTY and CSU 5-Stage collectors were derived. The experimental calibration work indicated that distinct fractions of cloudwater are collected in each stage of the FROSTY and CSU 5-Stage collectors. At laboratory conditions, the experimentally determined 50% cut diameters for the three stages of the FROSTY supercooled cloud collector were 19, 11.5, and 5 μm. Drop losses to the interstage wall surfaces in the FROSTY collector peaked at approximately 35% for 16 μm drops and were lower for larger and smaller drop sizes. For operation at design conditions of 3000 m elevation and -4° C, the 50% cut diameters are expected to decrease to 17, 10.5, and 4.5 μm. The experimentally determined 50% cut diameters, measured at laboratory conditions, for the CSU 5-Stage cloud collector were 25.5, 29, 17.5, 10.5, and 4.5 μm for stages 1 through 5, respectively. Wall losses tended to be higher than those for the FROSTY cloud collector across the drop size range under consideration. Losses peaked at nearly 45% for drops between 10 and 18 μm in diameter and decreased to about 20% at the largest and smallest drop sizes. 50% cut diameters are expected to remain essentially unchanged for CSU 5-Stage collector operation at sea level design conditions. Numerical modeling of the air flow patterns as well as drop trajectories through the FROSTY and CSU 5-Stage cloud collectors was performed with the commercially available Computational Fluid Dynamics (CFO) software package FLUENT, from Fluent, Inc. FLUENT offered two alternatives for the calculation of drop trajectories. Trajectory simulations based on the average continuous phase (air) velocity field as well as trajectory simulations which included the effects of statistically derived turbulent velocity fluctuations on drop motion were performed. Drop collection patterns based on these types of trajectory calculations were used to generate collection efficiency curves. Comparisons were made between the numerically predicted collection efficiency curves and efficiency curves established through experimental calibration. These comparisons indicated that the inclusion of turbulent fluctuation effects on drop motion provided better agreement with experimental observations than trajectories based only on average flow field velocities. However, the use of velocity fluctuations defined by default parameters also produced unrealistic losses to wall surfaces for small drop sizes. The parameters controlling turb lent velocity fluctuation effects on drop motion were examined in an effort to provide better agreement between the numerical and experimental results. Despite this shortcoming, numerically derived 50% cut diameters and overall collection efficiency curve shapes, for drop trajectories including turbulent velocity fluctuations, agreed reasonably well with experimental observations in most cases.Item Open Access The ionic composition of aerosol in Big Bend National Park(Colorado State University. Libraries, 2002-07) Lee, Taehyoung, author; Collett, Jeffrey L., Jr., author; Cooperative Institute for Research in the Atmosphere (Fort Collins, Colo.), publisherThe chemical compositions of PM2.5 and size-resolved aerosol particles were measured from July to October, 1999 in Big Bend National Park, Texas, during the Big Bend Regional Aerosol and Visibility Observational (BRAVO) Study. Daily PM2.5 samples were collected using a URG cyclone/annular denuder/filter pack sampling system consisting of a PM2.5 cyclone inlet, two coated annular denuders in series (for nitric acid and ammonia), and a filter pack. Aerosol particles collected on a Teflon filter were analyzed for major ions and a backup nylon filter was used to capture nitric acid volatilized from the collected particles. A Micro Orifice Uniform Deposition Impactor (MOUDI) was used to collect 24 hr size-resolved aerosol particle samples in 9 size categories (D50 = 18, 10, 5.6, 3.2, 1.8, 1.0, 0.56, 0.32 and 0.18 μm). 41 MOUDI sample days were selected for analysis of the ionic chemical composition as a function of particle size. PM2.5 and size-resolved aerosol concentrations of C1-, SO42-, NO3-, Na+, NH4+, K+, Mg2+, and Ca2+ were obtained through ion chromatographic analysis of the filter and impactor samples. Aerosol acidity was measured on-site in the daily PM2.s filter samples. The composition of the BRAVO PM2.5 aerosol was dominated by sulfate and ammonium. Daily average sulfate and ammonium concentrations were strongly correlated (R2=0.97). The ratio of ammonium to sulfate averaged 1.54 with standard deviation of 0.30. This ratio is consistent with the direct pH measurements of aerosol acidity. The highest concentrations of sulfate were observed from August to October, reaching as high as 8.5 μg/m3. Back-trajectories suggested long-range transport from regions along the Texas/Mexico border and east Texas was associated with peak sulfate concentrations in the park. The particle composition as a function of size obtained from the MOUDI samples suggests that most of the particulate nitrate is associated with coarse mode particles in the range of 4 - 5 μm diameter. Aerosol nitrate concentrations were correlated with the sum of aerosol Na+ and Ca2+ concentrations (R2 = 0.70 and 0.60 for MOUDI and URG, respectively), demonstrating the importance of sea salt and soil dust particles in providing non-acidic surfaces for the condensation of nitric acid. The MOUDI samples indicate that nitrate and sulfate are separated into supermicron (mode diameter 4 - 5 μm) and submicron (mode diameter 0.4 - 0.5 μm) particles, respectively. The MOUDI samples show that a 1 μm size cut would have provided a better division between the fine mode and the coarse mode aerosol during the BRAVO study. Comparison of ISORROPIA and SCAPE2 thermodynamic model predictions of solid phase sulfate species shows reasonable agreement between the models, although ISORROPIA sometime predicts higher concentrations of some species. ISORROPIA often predicts the presence of solid phase Na2SO4, while SCAPE2 seldom does. The difference between solid phase sulfate concentrations predicted by the two models largely reflects differences in predicted aerosol water content. Both models reasonably predict the observed phase partitioning of N(-III) but poorly predict the observed phase partitioning of N(V). The underprediction of aerosol nitrate by these bulk aerosol models reflects the fact that the PM2.5 aerosol is externally mixed, containing acidic submicron sulfate particles and supermicron nitrate particles.