Thesis (Ph. D.)--University of Texas at Austin, 2007. / Vita. Includes bibliographical references.

Environmental factors can have a significant impact on an individual's health and well-being, and a primary characteristic of environments is air quality. Air sensing equipment is available to the public, but it is often expensive,stationary, or unusable for persons without technical expertise. The goal of this project is to develop an inexpensive and portable sensor module for public use. The system is capable of measuring temperature in Celsius and Fahrenheit, heat index, relative humidity, and carbon dioxide concentration. The sensor module, referred to as the "sniffer," consists of a printed circuit board that interconnects a carbon dioxide sensor, a temperature/humidity sensor, an Arduino microcontroller, and a Bluetooth module. The sniffer is small enough to be worn as a pendant or a belt attachment, and it is rugged enough to consistently collect and transmit data to a user's smartphone throughout their workday. The accompanying smartphone app uses Bluetooth and GPS hardware to collect data and affix samples with a time stamp and GPS coordinates. The accumulated sensor data is saved to a file on the user's phone, which is then examined on a standard computer.

Background: Air pollution has become a serious environmental issue. It is responsible for hundreds of fatalities each year and it poses a serious threat to human health and environment. It leads to global warming, greenhouse effect and it also causes respiratory problems like asthma, lung cancer etc. It is important to predict the quality of the air to regulate air pollution. Air quality index (AQI) is a measure of air quality which describes the level of air pollution. Machine learning algorithms can help in predicting the AQI. Linear regression, LASSO regression, ridge regression, and SVR algorithms were used to forecast the AQI.  Objectives: The main objective of the thesis is to build and train a models using machine learning algorithms and find out the most accurate model in predicting the AQI.  Methods: Literature Review and Experimentation were chosen as methods to answer the research questions. There are a number of research papers written on prediction of AQI and literature review helped us a lot in research and references. Experimentation is also used to find out the most accurate machine learning model in predicting the air quality. In the experimentation phase, four machine learning algorithms were trained with air quality data to create predictive models for fore- casting AQI.  Results: Algorithms like Logistic Regression, Ridge Regression, LASSO Regression, and SVR are selected through literature review. Upon experimentation and training the algorithm with "Air Quality Data in India (2015-2020)" data set has showed that Ridge regression has the least MAE and RMSE and the highest R- square, which shows that it has the highest performance in predicting the AQI.  Conclusions: Four models are built by training with machine learning algorithms like Logistic Regression, Ridge Regression, LASSO Regression, and SVR and "Air Quality Data in India (2015-2020)" data set. After experimentation, it was found that Ridge Regression and LASSO regression has the better performance in the prediction of AQI.

published_or_final_version / Environmental Management / Master / Master of Science in Environmental Management

by Shu Keung Choi. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1996. / Includes bibliographical references (leaves 145-150). / ABSTRACT --- p.i / ACKNOWLEDGMENTS --- p.iii / TABLE OF CONTENTS --- p.iv / LIST OF FIGURES --- p.ix / LIST OF PLATES --- p.xi / LIST OF TABLES --- p.xii / Chapter CHAPTER 1. --- INTRODUCTION --- p.1 / Chapter 1.1 --- Concerns on Current Air Quality Modeling Process --- p.2 / Chapter 1.2 --- Objective --- p.2 / Chapter 1.3 --- Rationale --- p.4 / Chapter 1.4 --- System Overview --- p.4 / Chapter 1.5 --- Questions --- p.6 / Chapter 1.6 --- Thesis Organization --- p.6 / Chapter CHAPTER 2. --- LITERATURE REVIEW --- p.8 / Chapter 2.1 --- Introduction --- p.8 / Chapter 2.2 --- Geographic Information System --- p.8 / Chapter 2.2.1 --- Data Assimilation --- p.9 / Chapter 2.2.1.1 --- Data Representation --- p.9 / Chapter 2.2.1.2 --- Data Format --- p.10 / Chapter 2.2.1.3 --- Data Alignment --- p.11 / Chapter 2.2.2 --- Modeling Support --- p.11 / Chapter 2.3 --- Environmental Modeling --- p.12 / Chapter 2.4 --- Integration of GIS and Environmental Modeling --- p.15 / Chapter 2.4.1 --- The Need for Integration --- p.15 / Chapter 2.4.2 --- Forms of Integration --- p.17 / Chapter 2.5 --- Air Quality Modeling --- p.20 / Chapter 2.5.1 --- Classes of Models --- p.21 / Chapter 2.5.1.1 --- Classification by Spatial Scale --- p.21 / Chapter 2.5.1.2 --- Classification by Modeling Approach --- p.22 / Chapter 2.6 --- Gaussian Plume Models --- p.24 / Chapter 2.6.1 --- Formulation --- p.24 / Chapter 2.6.2 --- Determination of σy and σz --- p.25 / Chapter 2.6.3 --- The Stability Classification --- p.26 / Chapter 2.6.4 --- Estimation of σy and σz --- p.27 / Chapter 2.6.5 --- Assumptions in the Gaussian Model --- p.30 / Chapter 2.7 --- Air Quality Model Evaluation --- p.31 / Chapter 2.7.1 --- Model Uncertainties --- p.31 / Chapter 2.7.1.1 --- Inherent Uncertainty --- p.31 / Chapter 2.7.1.2 --- Reducible Uncertainty Errors --- p.33 / Chapter 2.7.1.2.1 --- Meteorological Data Errors --- p.33 / Chapter 2.7.1.2.2 --- Emission Data Errors --- p.34 / Chapter 2.7.1.2.3 --- Model Errors --- p.34 / Chapter 2.7.2 --- Operational Performance Evaluation --- p.36 / Chapter 2.7.2.1 --- Woods Hole Performance Measures --- p.36 / Chapter 2.7.2.2 --- Fractional Bias and Fractional Scatter --- p.38 / Chapter 2.7.2.3 --- Measuring the Normalized Ratios --- p.39 / Chapter 2.7.2.4 --- Combination of Statistical Measures --- p.40 / Chapter 2.8 --- Visualization --- p.43 / Chapter 2.8.1 --- Visualization Software Framework --- p.43 / Chapter 2.8.2 --- GIS and Visualization --- p.46 / Chapter 2.9 --- Conclusion --- p.47 / Chapter CHAPTER 3. --- SYSTEM DESIGN --- p.48 / Chapter 3.1 --- System Overview --- p.48 / Chapter 3.2 --- Supported Models --- p.50 / Chapter 3.3 --- System Software Platforms --- p.51 / Chapter 3.3.1 --- ARC/INFO --- p.52 / Chapter 3.3.1.1 --- Overview --- p.52 / Chapter 3.3.1.2 --- The Role in the System --- p.53 / Chapter 3.3.2 --- Advanced Visualization System (AVS) --- p.54 / Chapter 3.3.2.1 --- Overview --- p.54 / Chapter 3.3.2.2 --- The Role in the System --- p.54 / Chapter 3.4 --- System Requirements and Specification --- p.56 / Chapter 3.4.1 --- Notation --- p.56 / Chapter 3.4.2 --- Data Preprocessing --- p.57 / Chapter 3.4.3 --- Data Postprocessing --- p.63 / Chapter 3.4.4 --- Model Performance Evaluation --- p.68 / Chapter 3.4.5 --- Visualization --- p.74 / Chapter 3.4.5.1 --- Reading ARC/INFO Data --- p.74 / Chapter 3.4.5.2 --- Applying Visualization Techniques --- p.77 / Chapter 3.4.5.2.1 --- Surface Mesh --- p.77 / Chapter 3.4.5.2.2 --- Multi-window Approach --- p.79 / Chapter 3.5 --- Data File Format --- p.85 / Chapter CHAPTER 4. --- A TEST CASE --- p.92 / Chapter 4.1 --- Introduction --- p.92 / Chapter 4.2 --- Test Case Components --- p.92 / Chapter 4.2.1 --- Study Area --- p.92 / Chapter 4.2.2 --- Source Data --- p.93 / Chapter 4.2.3 --- Air Quality Model - MPTER --- p.93 / Chapter 4.2.4 --- Meteorological Data Preprocessor - RAMMET --- p.95 / Chapter 4.3 --- Executing the Test Case --- p.95 / Chapter 4.3.1 --- Main Menu --- p.95 / Chapter 4.3.2 --- Viewing the study area --- p.96 / Chapter 4.3.3 --- Data Preprocessing --- p.96 / Chapter 4.3.3.1 --- Define Data Mapper --- p.98 / Chapter 4.3.3.2 --- Execute Data Preprocessor --- p.101 / Chapter 4.3.3.3 --- Meteorological Data Preprocessing --- p.102 / Chapter 4.3.3.4 --- Model Input File Editing --- p.103 / Chapter 4.3.3.5 --- Discussions --- p.105 / Chapter 4.3.4 --- Model Execution --- p.107 / Chapter 4.3.5 --- Data Postprocessing --- p.107 / Chapter 4.3.5.1 --- Import Model Result to GIS --- p.108 / Chapter 4.3.5.2 --- Iso-line of Concentration Map --- p.108 / Chapter 4.3.5.3 --- Discussions --- p.109 / Chapter 4.3.6 --- Model Performance Evaluation --- p.112 / Chapter 4.3.6.1 --- Program Execution --- p.113 / Chapter 4.3.6.2 --- Discussions --- p.113 / Chapter 4.3.7 --- Visualization --- p.116 / Chapter 4.3.7.1 --- Surface Mesh --- p.116 / Chapter 4.3.7.2 --- Multi-window Approach for 4D Data set --- p.117 / Chapter 4.3.7.2.1 --- Overview --- p.117 / Chapter 4.3.7.2.2 --- Overall Controls and Relations between Viewers --- p.121 / Chapter 4.3.7.2.3 --- Independent Controls within Each Viewers --- p.122 / Chapter 4.3.7.2.4 --- "The x,y,z-volume Viewer" --- p.123 / Chapter 4.3.7.2.5 --- "x,y,t-volume in ViewerZ" --- p.128 / Chapter 4.3.7.2.6 --- Other Viewers --- p.132 / Chapter 4.3.7.3 --- Discussions --- p.134 / Chapter 4.4 --- Conclusion --- p.137 / Chapter CHAPTER 5. --- CONCLUSION --- p.138 / Chapter 5.1 --- System Design Summary --- p.138 / Chapter 5.2 --- Summary of the Functions --- p.139 / Chapter 5.2.1 --- Data Preprocessing --- p.139 / Chapter 5.2.2 --- Data Postprocessing --- p.140 / Chapter 5.2.3 --- Model Evaluation --- p.140 / Chapter 5.2.4 --- Visualization --- p.141 / Chapter 5.3 --- Further Research --- p.143 / BIBLIOGRAPHY --- p.145

Particles reduce visibility, change climate, and affect human health. In 1997, the National Ambient Air Quality Standard (NAAQS) for PM2.5 (particles less than 2.5 mm) was promulgated. The annual mean PM2.5 mass concentrations in Atlanta, Georgia exceed the standard, and control is needed. The first goal of this study is to develop the control strategies of PM2.5 in Atlanta, Georgia. Based on the statistical analysis of measured data, from 22% to 40% of emission reductions are required to meet the NAAQS at 95% CI. The estimated control levels can be tested using the Community Multiscale Air Quality (CMAQ) model to better assess if the proposed levels will achieve sufficient reduction in PM2.5. The second goal of this study is to analyze various uncertainties residing in CMAQ. For the model to be used in such applications with confidence, it needs to be evaluated. The model performance is calculated by the relative agreement between volume-averaged predictions and point measurements. Up to 14% of the model error for PM2.5 mass is due to the different spatial scales of the two values. CMAQ predicts PM2.5 mass concentrations reasonably well, but CMAQ significantly underestimates PM2.5 number concentrations. Causes of the underestimation include that assumed inaccurate particle density and particle size of the primary emissions in CMAQ, in addition to the expression of the particle size with three lognormal distributions. Also, the strength and limitations of CMAQ in performing PM2.5 source apportionment are compared with those of the Chemical Mass Balance with Molecular Markers. Finally, the accuracy of emissions, one of the important inputs of CMAQ, is evaluated by the inverse modeling. Results show that base level emissions for CO and SO2 sources are relatively accurate, whereas NH3, NOx, PEC and PMFINE emissions are overestimated. The emission adjustment for POA and VOC emissions is significantly different among regions.

Two long-standing air quality challenges in the United States are the control of tropospheric ozone and particulate matter, both of which are responsible for widespread damage to human health and the environment. This thesis presents three modeling applications in support of policy development and regulatory actions for control of these pollutants in the eastern United States, taking advantage of recent advancements in sensitivity techniques in a regional Eulerian photochemical air quality model. A broad evaluation of regional atmospheric pollution and transboundary air quality management, including the international scale, and an analysis of successful transboundary management efforts are also presented. The first modeling application is an evaluation of local and interstate impacts on ozone and fine particulate matter (PM2.5) from ground-level and elevated nitrogen oxide plus nitrogen dioxide and from sulfur dioxide emissions from individual states. This analysis identifies states responsible for a significant amount of regional secondary pollution, and states which do not have independent control over much of their pollution concentrations. An average of approximately 77% of each state s ozone and PM2.5 concentrations that are sensitive to the emissions evaluated are found to be formed from emissions from other states. The second application is an assessment of impacts from emissions from a single power-plant on resulting regional ozone concentrations. Three sensitivity techniques and two 3D photochemical models are applied. Ozone increases greater than 0.5 ppbv are found over eight states downwind from the power-plant. The third application supports the extension of a body of research aimed at advancing understanding of the ozone formation potential, or reactivity , of VOCs for use in regional-scale, rather than urban-scale, regulations. Air quality impacts of VOCs emissions from solvent use and manufacture are presented, scientific barriers to accounting for reactivity in regulations are discussed, current and upcoming regulatory applications are described, and results from a regional scale evaluation of reactivity quantification are presented.

The Air Quality Act (AQA) Act No.39 of 2004 promulgated in 2004 follows the outdated Atmospheric Pollution Prevention Act (APPA) (Act No.45 of 1965). The legislative approach shifted from a source- based, end of pipe, command and control, guideline principle to ambient air quality management and improvement of compliance to standards through a consultative process. The AQA’s management framework incorporates a co-operative and integrated approach with government, communities and polluters to look at the holistic management of ambient air quality and the identified roles and responsibilities for all stakeholders. The AQA branched from the National Environmental Management Act (NEMA) 107 of 1998, which is the first piece of legislation formalizing the principles of the Integrated Pollution Waste Management (IPWM) Policy published in 2000 and the Bill of Rights. Government and Industry have a role to play in the implementation of the AQA. Government’s role covers the management and enforcement aspects, whilst industries’ role includes the management of air emissions and compliance reporting to improve the overall ambient air quality. The AQA’s industrial requirements range from compliance and reporting by ensuring emission licenses are in place, compliance with standards set by different spheres of government and the management of these emissions. The management of these requirements includes understanding the legislation, its implications and the provision of other financial, human and technological resources. Industry needs to consider the impacts of these legislative changes and how it may impact business as a whole. The aim of this study is to analyze the industrial perspectives of the AQA and its implementation through the use of a questionnaire. Open-ended questionnaires were administered to a total of forty industrial companies in the chemical, petrochemical, energy and mining sectors in the Gauteng, North West and Durban industrial areas. Industries were identified as those which have scheduled process certificates or companies that will be impacted by the impending changes as a result of the AQA. The overall outcome of the industrial responses revealed poor general knowledge of the principles, purpose and the reasons for the transition from APPA to AQA. Few industries had insight into the type of challenges they may face from the AQA’s listed control measures and the control measures that would apply to their particular industry. There is a general concern surrounding the government’s lack of support and the essential enforcement that is required to ensure ambient air quality compliance. These challenges and recommendations are discussed in the thesis. / Thesis (M.Sc.)-University of KwaZulu-Natal, Westville, 2009.

No description available.

Road Deposited Sediment (RDS) is an important pathway of pollution material in the urban environment. Traditional particulate matter (PM) monitoring methods are typically expensive and time consuming. To date, urban sediment studies have not fully explored the application of mineral magnetic technologies as an alternative to characterise RDS or, perhaps more importantly, their use as particle size proxy. Therefore, this study addresses these issues by determining the extent of any linkages between magnetic properties and the physio-chemical concentrations of RDS. Investigations have focussed on a spatial temporal study (2008-10) of RDS from the City of Wolverhampton (n = 546) and a similar ‘snap-shot’ study of eight selected town and cities across the UK (n = 306), plus a comparison investigation linked to regional monitoring of air sampling units (ASU) (n = 208). A suite of analytical approaches, namely mineral magnetism, laser diffraction, X-Ray Fluorescence spectroscopy (XRF), Scanning Electron Microscopy (SEM) and Loss on Ignition (LOI), were employed to characterize sample properties. Data interrogation identified mainly weak correlations exist between most mineral magnetic parameters and particle size classes (i.e. sand, silt and clay) and respiratory health-related size classes (i.e. PM10, PM2.5 and PM1.0). The few strongest correlations (p <0.001) were found between mineral magnetic concentration and <PM10 parameters. In Wolverhampton this occurred for samples collected during the spring months (March and May), indicating possible seasonal influences on RDS dynamics and sources. Elsewhere in the UK, and at ASU stations, results revealed mainly limited or insignificant (p >0.05) correlations exist between mineral magnetic parameters and particle size. However, for some locations (most notably, London and Scunthorpe), results exhibit signatures perceived to be associated with environmental factors. Detailed multivariate Factor Analysis plots and Geographical Information System (GIS) images have been used to explore these findings further. These illustrate RDS properties of road types (arterial and residential) display significantly different characteristics, with raised mineral magnetic concentrations for arterial roads, compared to lesser concentrations for residential roads, which corresponds to traffic flow data. This is supported by SEM analyses that reveal elevated concentrations of iron oxide spheres in samples collected from arterial roads, which are indicative of inputs from anthropogenic combustion sources. Contextualising these findings within the framework of existing knowledge, a conceptual approach has been presented that explores factors (i.e. sampling area, topography, land use, sediment source and potential mixing), which influence the reliability of using mineral magnetic techniques a particle size proxy. This demonstrates that any increase in the complexity of these factors (sampling area dynamics) can be used to predict the likelihood of being able to employ mineral magnetic measurements as a proxy. To surmise the work overall, despite mineral magnetic technologies offering an inexpensive and rapid means of analysing RDS, its use as a proxy measure for particulate matter appears to be limited by a series of site-related factors but the technique seems to offer valuable insights for pollution source studies.