Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.683909
Title: Pattern recognition and an OH reactivity study of urban air pollution
Author: Purkayastha, Kanan K.
ISNI:       0000 0004 5919 0697
Awarding Body: University of Bristol
Current Institution: University of Bristol
Date of Award: 2015
Availability of Full Text:
Access from EThOS:
Abstract:
Pattern recognition focuses on the recognition of pattern and regularities in data. In urban air pollution studies such as the study of the relationship between air pollutant concentration and meteorology data or atmospheric photochemistry, such as hydroxyl molecule loss in the atmosphere, pattern recognition is pertinent. This thesis presents various case studies where emission potential to atmosphere, concentration pattern in an air basin, extreme event pattern on a local or regional scale, patterns of the relationship between meteorological parameters and chemical parameter and atmospheric chemical loss processes have been explored. The thesis presents some new approaches to studying urban air pollution. The first study is a landfill gas emission potential study, where the composite signal has been analysed. The study proposes a new Monte Carlo permutation scheme to capture the composite landfill gas emission potential to the atmosphere. It also compares the new method with the traditional Fisher g statistics method. The second study reviews the existing methods for identifying extreme events of particulate matter on a local and regional scale and develops a new method based upon medcouple calculations of dispersion of data. This study also informs the relationship between local meteorology and extreme events using a new mathematical approach. The third study involves different pattern recognition techniques such as self organising maps, hierarchical and non-hierarchical clustering techniques, the receiver operating characte11stic curve method and a multivariate statistical approach to explore the pattern of air pollution on a regional scale. The fourth study involves applying soft modelling approaches such as block similarity measures, partial least square and principal component regression techniques to explore the relationship between meteorology and urban air pollution. This study applies a new kind of instrument known as an AQ mesh. So appropriate calibration approaches for studying these data have been followed. The fifth study is the measurement of total OH reactivity in a subtropical climatic region in the autumn season. Non-methane volatile organic carbon and different trace species contribution to OH reactivity have been assessed. This study involves the development of an instrument to generate the OH radical in a laboratory environment and kinetic study of the OH radical with volatile organic carbons and trace species in ambient air. An attempt has also been made in this study to calculate the potential "Criegee intermediate" by the ozonolysis of identified alkenes in the ambient air.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.683909  DOI: Not available
Share: