Agricultural operations across the United States are encountering difficulties complying with current air pollution regulations for particulate matter (PM). PM is currently regulated in terms of particle diameters less than or equal to a nominal 10 μm (PM10); however, current legislation is underway to regulate PM with diameters less than or equal to a nominal 2.5 μm (PM2.5). The goals of this research were to determine the biases and uncertainties associated with current PM10 and PM2.5 sampling methods and to determine the extent to which these errors may impact the determination of cotton gin emission factors. Ideally, PM samplers would produce an accurate measure of the pollutant indicator; for instance, a PM10 sampler would produce an accurate measure of PM less than or equal to 10 μm. However, samplers are not perfect and errors are introduced because of the established tolerances associated with sampler performance characteristics and the interaction of particle size and sampler performance characteristics. Results of this research indicated that a source emitting PM characterized by a mass median diameter (MMD) of 20 μm and a geometric standard deviation (GSD) of 1.5 could be forced to comply with a 3.2 and 14 times more stringent regulation of PM10 and PM2.5, respectively, than a source emitting PM characterized by a MMD of 10 μm and a GSD of 1.5. These estimates are based on both sources emitting the same concentrations of true PM or concentrations corresponding to the particle diameters less than the size of interest. Various methods were used to estimate the true PM10 and PM2.5 emission factors associated with cotton gin exhausts and the extent to which the sampler errors impacted the PM regulation. Results from this research indicated that current cotton gin emission factors could be over-estimated by about 40%. This over-estimation is a consequence of the relatively large PM associated with cotton gin exhausts. These PM sampling errors are contributing to the misappropriation of source emissions in State Implementation Plans, essentially forcing Air Pollution Regulatory Agencies to require additional controls on sources that may be incorrectly classified has high emitters.

Knowledge gaps related to mass transfer processes involved in passive air sampling of semivolatile organic compounds and factors potentially influencing passive sampling rates (PSRs) were addressed with controlled laboratory experiments, mass transfer modeling, and a field sampling campaign. The observed non-uniform SVOC distributions within porous passive sampling media (PSMs) contradict an assumption in an earlier passive air sampling theory and proved the existence of a kinetic resistance on the PSM side. This resistance can affect PSRs as revealed by a new PAS model which is based on fundamental laws of mass transfer in air and porous media. By considering mass transfer processes within the PSM, the model is able to explain the large variations of field calibrated PSRs with temperature and between SVOC species and the two-stage uptake process, which cannot be addressed by the earlier PAS theory. Because the PSM side kinetic resistance invalidates the assumption that depuration compounds added to the PSM prior to deployment are subject to the same kinetic resistance as the sampled SVOCs, PSRs derived from the loss rates of depuration compounds can differ from the actual PSRs of the sampled SVOCs. Using such PSRs could thus introduce additional uncertainty to PAS-derived air concentrations. Experiments using XAD-resin and silica-gel filled mesh cylinder as PSMs for the uptake of SVOCs and water vapor respectively revealed that sorbent in the inner portion of the PSM does not take part in chemical uptake; PSRs are thus proportional to the interfacial transfer area but not the amount of the sorbent. Accordingly, thinner PSM can be used to reduce the amount of sorbent while keeping or even increasing the PSRs. Optimized designs of PASs could be tested time efficiently using the gravimetrical approach based on water vapor uptake by silica gel.

Knowledge gaps related to mass transfer processes involved in passive air sampling of semivolatile organic compounds and factors potentially influencing passive sampling rates (PSRs) were addressed with controlled laboratory experiments, mass transfer modeling, and a field sampling campaign. The observed non-uniform SVOC distributions within porous passive sampling media (PSMs) contradict an assumption in an earlier passive air sampling theory and proved the existence of a kinetic resistance on the PSM side. This resistance can affect PSRs as revealed by a new PAS model which is based on fundamental laws of mass transfer in air and porous media. By considering mass transfer processes within the PSM, the model is able to explain the large variations of field calibrated PSRs with temperature and between SVOC species and the two-stage uptake process, which cannot be addressed by the earlier PAS theory. Because the PSM side kinetic resistance invalidates the assumption that depuration compounds added to the PSM prior to deployment are subject to the same kinetic resistance as the sampled SVOCs, PSRs derived from the loss rates of depuration compounds can differ from the actual PSRs of the sampled SVOCs. Using such PSRs could thus introduce additional uncertainty to PAS-derived air concentrations. Experiments using XAD-resin and silica-gel filled mesh cylinder as PSMs for the uptake of SVOCs and water vapor respectively revealed that sorbent in the inner portion of the PSM does not take part in chemical uptake; PSRs are thus proportional to the interfacial transfer area but not the amount of the sorbent. Accordingly, thinner PSM can be used to reduce the amount of sorbent while keeping or even increasing the PSRs. Optimized designs of PASs could be tested time efficiently using the gravimetrical approach based on water vapor uptake by silica gel.

Time series involving count data are present in a wide variety of applications. In many applications, the observed counts are usually small and dependent. Failure to take these facts into account can lead to misleading inferences and may detect false relationships. To tackle such issues, a Poisson parameter-driven model is assumed for the time series at hand. This model can account for the time dependence between observations through introducing an autoregressive latent process. In this thesis, we consider Bayesian approaches for estimating the Poisson parameter-driven model. The main challenge is that the likelihood function for the observed counts involves a high dimensional integral after integrating out the latent variables. The main contributions of this thesis are threefold. First, I develop a new single-move (SM) Markov chain Monte Carlo (MCMC) method to sample the latent variables one by one. Second, I adopt the idea of the particle Gibbs sampler (PGS) method \citep{andrieu} into our model setting and compare its performance with the SM method. Third, I consider Bayesian composite likelihood methods and compare three different adjustment methods with the unadjusted method and the SM method. The comparisons provide a practical guide to what method to use. We conduct simulation studies to compare the latter two methods with the SM method. We conclude that the SM method outperforms the PGS method for small sample size, while they perform almost the same for large sample size. However, the SM method is much faster than the PGS method. The adjusted Bayesian composite methods provide closer results to the SM than the unadjusted one. The PGS and the selected adjustment method from simulation studies are compared with the SM method via a real data example. Similar results are obtained: first, the PGS method provides results very close to those of the SM method. Second, the adjusted composite likelihood methods provide closer results to the SM than the unadjusted one.

In this thesis, we discuss some topics in random matrix theory which have applications to probability, statistics and quantum information theory. In Chapter 2, by relying on the spectral properties of an associated adjacency matrix, we find the distribution of the maximum of a Dyck path and show that it has the same distribution function as the unsigned Brownian excursion which was first derived in 1976 by Kennedy. We obtain a large and moderate deviation principle for the law of the maximum of a random Dyck path. Our result extends the results of Chung, Kennedy and Khorunzhiy and Marckert. In Chapter 3, we discuss a method of sampling called the Gibbs-slice sampler. This method is based on Neal's slice sampling combined with Gibbs sampling. In Chapter 4, we discuss several examples which have applications in physics and quantum information theory.

Pesticide runoff from agricultural applications poses a potential threat to water quality in the world heritage listed Great Barrier Reef (GBR) and sensitive monitoring tools are needed to facilitate effective monitoring of these pollutants. This thesis has worked to advance passive sampling tools for monitoring trace organic pollutants and their potential impacts on the GBR. The suitability of several available passive sampling tools for detecting trace concentrations of target pesticide analytes was investigated in a laboratory calibration chamber before experiments were conducted to further understanding of the response of Chemcatcher passive samplers in environmental conditions likely to be experienced in the GBR. Passive samplers were deployed in a survey of pesticides in the GBR environment and extracts were applied in bioassays to investigate their application for predicting mixture toxicity to GBR biota. When employed and interpreted appropriately, passive sampling tools have been shown to provide for sensitive and reproducible detection of organic pollutants in relatively pristine environments. While considerable research has been conducted into the performance and theory of analyte uptake by a range of passive sampling devices, several key knowledge gaps existed and were addressed in this study. The applicability of the performance reference compound (PRC) concept as an in situ calibration method for passive samplers using Empore Disk sampling phases (Chemcatchers) to monitor polar compounds was investigated. This experiment showed that while uptake of pesticides was linear and reproducible, PRC loss was not linear, meaning that the dissipation rates of these PRCs cannot be used to estimate the effect of field exposure conditions on uptake rates. An alternative in situ calibration technique using PRC loaded polydimethylsiloxane (PDMS) disks deployed alongside the Chemcatchers as a surrogate calibration phase has been tested and shows promise for future applications. Pesticide concentrations in waters flowing to the GBR have been shown to undergo dramatic fluctuations over short time periods and the potential for these conditions to limit the integrative period of sampling was investigated by simulating a changing concentration event in a calibration chamber. The ability for samplers to predict average concentrations was dependant on the deployment configuration (with or without membrane) used and the period of deployment relative to the changing concentration event. Passive samplers were employed in a survey of pesticides in GBR waters during a wet and dry season at river mouths, two nearshore regions and an offshore region. The nearshore marine environment was shown to be contaminated with pesticides in both the dry and wet seasons (average water concentrations of 1.3-3.8 ng L-1 and 2.2-6.4 ng L-1, respectively), while no pesticides were detected further offshore. Continuous monitoring of two rivers over 13 months showed waters flowing to the GBR were contaminated with herbicides (diuron, atrazine, hexazinone) year round, with highest average concentrations present during summer months (350 ng L-1). The use of passive samplers has enabled identification of insecticides in GBR waters which have not been reported in the literature previously. Extracts from passive samplers deployed at three sites in the GBR were applied to bioassays targeting integral life stages or functions of coral reef biota: scleractinian coral larvae, sea urchin larvae, a marine diatom and marine bacteria. The results demonstrate the utility of pairing passive sampling with bioassays and reveal that mixtures of organic pollutants in the GBR have the potential to cause detrimental effects to coral reef biota. This research outlines an approach that reduces one of the levels of simplification of risk assessment of pollutants to ecosystems by incorporating mixtures of chemicals present in the environment. The use of passive sampler extracts in toxicity testing allows pollutant mixtures to be assessed at a range of enrichment factors and, with the inclusion of biota from the ecosystem of concern, improves the relevance of results for predicting real world effects. The findings of this thesis are intended to be used to improve the application of passive sampling tools for routine monitoring to provide managers with understanding of the pesticides present, the potential effects of those pollutant mixtures and feedback on the efficacy of implemented land management practices in halting and reversing the impacts of pesticides on the GBR.

Thesis (M. En.)--Miami University, Institute of Environmental Sciences, 2004. / Title from first page of PDF document. Includes bibliographical references (p. 33).

In this thesis, we discuss some topics in random matrix theory which have applications to probability, statistics and quantum information theory. In Chapter 2, by relying on the spectral properties of an associated adjacency matrix, we find the distribution of the maximum of a Dyck path and show that it has the same distribution function as the unsigned Brownian excursion which was first derived in 1976 by Kennedy. We obtain a large and moderate deviation principle for the law of the maximum of a random Dyck path. Our result extends the results of Chung, Kennedy and Khorunzhiy and Marckert. In Chapter 3, we discuss a method of sampling called the Gibbs-slice sampler. This method is based on Neal's slice sampling combined with Gibbs sampling. In Chapter 4, we discuss several examples which have applications in physics and quantum information theory.

The overall goals of this doctoral dissertation are to provide knowledge of workers' exposure to nanomaterials and to assist in the development of standard methods to measure personal exposure to nanomaterials in workplace environments. To achieve the first goal, a field study investigated airborne particles generated from the weighing of bulk carbon nanotubes (CNTs) and the manual sanding of epoxy test samples reinforced with CNTs. This study also evaluated the effectiveness of three local exhaust ventilation (LEV) conditions (no LEV, custom fume hood and biosafety cabinet) for control of exposure to particles generated during sanding of CNT-epoxy nanocomposites. Particle number and respirable mass concentrations were measured with direct-read instruments, and particle morphology was determined by electron microscopy. Sanding of CNT-epoxy nanocomposites released respirable size airborne particles with protruding CNTs very different in morphology from bulk CNTs that tended to remain in clusters (>1µm). Respirable mass concentrations in the operator's breathing zone were significantly greater when sanding took place in the custom hood (p <0.0001) compared to the other LEV conditions. This study found that workers' exposure was to particles containing protruding CNTs rather than to bulk CNT particles. Particular attention should be placed in the design and selection of hoods to minimize exposure. Two laboratory studies were conducted to realize the second goal. Collection efficiency of submicrometer particles was evaluated for nylon mesh screens with three pore sizes (60, 100 and 180 µm) at three flow rates (2.5, 4, and 6 Lpm). Single-fiber efficiency of nylon mesh screens was then calculated and compared to a theoretical estimation expression. The effects of particle morphology on collection efficiency were also experimentally measured. The collection efficiency of the screens was found to vary by less than 4% regardless of particle morphology. Single-fiber efficiency of the screens calculated from experimental data was in good agreement with that estimated from theory for particles between 40 and 150 nm but deviated from theory for particles outside of this range. New coefficients for the single-fiber efficiency model were identified that minimized the sum of square error (SSE) between the experimental values and those estimated with the model. Compared to the original theory, the SSE calculated using the modified theory was at least threefold lower for all screens and flow rates. Since nylon fibers produce no significant spectral interference when ashed for spectrometric examination, the ability to accurately estimate collection efficiency of submicrometer particles makes nylon mesh screens an attractive collection substrate for nanoparticles. In the third study, laboratory experiments were conducted to develop a novel nanoparticle respiratory deposition (NRD) sampler that selectively collects nanoparticles in a worker's breathing zone apart from larger particles. The NRD sampler consists of a respirable cyclone fitted with an impactor and a diffusion stage containing eight nylon- mesh screens. A sampling criterion for nano-particulate matter (NPM) was developed and set as the target for the collection efficiency of the NRD sampler. The sampler operates at 2.5 Lpm and fits on a worker's lapel. The cut-off diameter of the impactor was experimentally measured to be 300 nm with a sharpness of 1.53. Loading at typical workplace levels was found to have no significant effect (2-way ANOVA, p=0.257) on the performance of the impactor. The effective deposition of particles onto the diffusion stage was found to match the NPM criterion, showing that a sample collected with the NRD sampler represents the concentration of nanoparticles deposited in the human respiratory system.

Porous polyurethane foam was evaluated as a potential replacement substrate for the eight nylon meshes currently being used in the diffusion stage of the Nanoparticle Respiratory Deposition (NRD) Sampler. A semi-empirical particle deposition model and preliminary tests were used to select the dimensions of foam substrate needed to match the collection of the NRD sampler at recommended sampling conditions. The foam substrate consisted of a cylinder nominally 25-mm diameter by 40 mm in depth, housed in a conductive plastic cassette cowl (internal diameter of 23 mm) compatible with the existing NRD sampler. Pristine foam was evaluated for metals content via acid-assisted microwave digestion and inductively coupled plasma-optical emissions spectroscopy (ICP-OES) chemical analysis. Foam collection efficiency was evaluated using salt (NaCl) and metal fume test aerosols in independent tests. Foam collection efficiency was compared to the nanoparticulate matter (NPM) criterion (established to reflect the total deposition in the human respiratory system for particles smaller than 300 nm) and theoretical modeling. The collection efficiency of NaCl particles was similar to the NPM criterion (R2 = 0.98) and the model underestimated the experimental efficiency (R2 = 0.38). Increased collection efficiency of metal fume was observed for particles larger than 70 nm presumably due to increased interception effects of fractal shaped particles. The pressure drop across the pristine foam was 1/12th that of the nylon meshes. Foam and nylon meshes were loaded with metal fume particles to evaluate performance under simulated field conditions. Changes in collection efficiency and pressure drop were used as measures of performance. Foam had substantially lower changes in collection efficiency and pressure drop with ~ 19 mg metal fume loaded compared to the nylon meshes with ~ 3 mg metal fume loaded.