THESIS
2018
xvi, 207 pages : color illustrations, color maps ; 30 cm
Abstract
The main objectives of this thesis are to: (1) develop an approach for determining the
minimum sample size required for estimating statistics of interest for pollutant measurements
taken at a given microenvironment with a specific level of precision, and (2) using PM
2.5 as an
example, quantify the exposure concentrations in various microenvironments at specified
precision levels.
Field sampling campaigns were conducted to quantify PM
2.5 exposure concentrations in
several typical transport microenvironments (TMEs), including Mass Transit Railway (MTR),
minibus, double-decker bus, and tramcar (on-road). Linear regression equations between PM
2.5
concentrations measured by a DustTrak aerosol monitor and by a USEPA benchmark method
were used to bias correct PM
2.5 measurements. The c...[
Read more ]
The main objectives of this thesis are to: (1) develop an approach for determining the
minimum sample size required for estimating statistics of interest for pollutant measurements
taken at a given microenvironment with a specific level of precision, and (2) using PM
2.5 as an
example, quantify the exposure concentrations in various microenvironments at specified
precision levels.
Field sampling campaigns were conducted to quantify PM
2.5 exposure concentrations in
several typical transport microenvironments (TMEs), including Mass Transit Railway (MTR),
minibus, double-decker bus, and tramcar (on-road). Linear regression equations between PM
2.5
concentrations measured by a DustTrak aerosol monitor and by a USEPA benchmark method
were used to bias correct PM
2.5 measurements. The calibrated data was analyzed to understand
the infiltration of outdoor pollutants into the different transport microenvironments. Key
findings include: (1) ambient air was a major contributor to microenvironmental PM
2.5
concentrations, with impact ranging from 50% to nearly 100%; and (2) the in-cabin spatial
variability of PM
2.5 concentration was observed for MTR mainly due to the change of
ventilation conditions.
A bootstrap-based approach was also used for estimating the minimum sample size and
the sample error statistics (e.g., mean and linear regression slope) of a PM
2.5 data set given a
prescribed level of precision. Simulation studies were conducted using both synthetic as well
as measured microenvironmental data. The results show that factors including internal
variability in the sample data, the averaging intervals, and linear regression R
2
between paired
data sets would impact the minimum sample size required to estimate the regression slopes
given specific level of precision. Under our measurement conditions, 3 days of measurements
were sufficient for estimating the infiltration factor of PM
2.5 at school classrooms to within 10%
uncertainty, while around 10 and 5 days were needed to capture the spatial variations in on-road PM
2.5 concentrations in winter and summer with coefficients of determination greater than
0.6, respectively.
The results should be helpful for more accurately estimate microenvironmental PM
2.5
exposures and can be used to improve the study design of sampling campaigns to quantify
infiltration of ambient PM
2.5 into urban microenvironments.
Post a Comment