Long term trend in the concentrations of selected VOCS and their meteorological dependences at two sites in the UK

Al-Rushoud, Dhoha (2020). Long term trend in the concentrations of selected VOCS and their meteorological dependences at two sites in the UK. University of Birmingham. M.Phil.

AlRushoud2020MPhil.pdf
Text - Accepted Version
Restricted to Repository staff only until 1 August 2030.
Available under License All rights reserved.
Download (10MB) | Request a copy

Abstract

Volatile organic compounds (VOCs) are emitted from different sources, such as traffic related sources, solvents, combustion, gasoline vapour, and biogenic sources. The objective of the current study is detecting and quantifying changes in times series of hourly data of selected VOCs concentrations in urban and suburban sites in London using a statistical approach and a random forest model. A series of data analysis are applied to examine daily, seasonal and long term variations in the selected species with time and location. A machine learning algorithm - random forest model- is used to investigate the effect of meteorological and emission changes on the concentrations and long term trends in the concentrations of the selected compounds. The first site is London Marylebone Road, a complex urban environment with a heavy traffic density and the studied period is from 1998 to 2017. The second site is London Eltham, a suburban environment with less traffic and more vegetation areas and the studied period is from 1994 to 2017. The studied pollutants include BETX group (benzene, ethylbenzene, toluene, o-xylene, (m+p) xylene), ethyne, ethane, propane and isoprene. In general, the concentration of all compounds showed a marked seasonal variations with highest concentrations in winter and lowest average concentrations in summer months except for isoprene. Long term trend analysis showed that all compounds decreased during the studied periods except ethane at London Eltham site which has increased from 2004 to 2017. The random forest models performed well for all species in predicting observed concentrations with R2 values ranging between 0.70 and 0.86 for all species at both sites. Comparisons between annual average measured and normalised data indicated the large effect of weather conditions in reducing or increasing emissions concentrations at both urban and suburban sites. For example; the annual average concentration of benzene in 2016 was 1.33 µg/m3 and 1.26 µg/m3 before and after weather normalisation, respectively. This means that meteorological conditions increased the levels by 5.26 % in 2016. However, in 2017 benzene concentrations where 1 µg/m3 and 1.2 µg/m3 before and after weather normalisation, respectively. This means that meteorological conditions helped in reducing benzene levels by 20%. The meteorologically normalised time series showed that there were a remarkable decrease in benzene and toluene in the year of implementing the European Union emissions standards for passenger cars in 2000 (Euro III) at both sites.

Type of Work:

Thesis (Masters by Research > M.Phil.)

Award Type:

Masters by Research > M.Phil.

Supervisor(s):