MEASUREMENTS OF GROUND LEVEL AIR POLLUTION IN CENTRAL ONTARIO BY FOURIER TRANSFORM INFRARED SPECTROSCOPY

A Thesis Submitted to the Committee on Graduate Studies in Partial Fulfillment of the Requirements for the Degree of Master of Science in the Faculty of Arts and Science

TRENT UNIVERSITY Peterborough, Ontario, Canada

© Denise Louise McMaster 2001

Watershed Ecosystems M.Sc. Program May 2002

ABSTRACT Measurements of Ground Level Air Pollution in Central Ontario by Fourier Transform Infrared Spectroscopy DENISE LOUISE MCMASTER

Ground level concentrations of carbon dioxide (CO2 ), carbon monoxide (CO), methane (CH4), and nitrous oxide (N2 O) were investigated in Peterborough, Ontario using Fourier transform infrared spectroscopy. Ground level concentrations of ozone (O3) and nitrogen oxides (NOx ) have also been investigated. The seasonal variation and the influence of wind direction on the ground level concentrations of these pollutants has been examined and provided insight to the factors affecting pollutant concentration and sources of air pollution in central Ontario. The source of O3 pollution has also been investigated and indicated that O3 pollution in Peterborough was mainly a result of photochemical activity and that Peterborough was influenced by long-range transport of pollutants from source regions upwind. This result was confirmed by the investigation of the diurnal cycles of O3 and NOx . Finally, the relationship between lower tropospheric O3 and surface O3 concentrations was examined. The results suggested that lower troposphere O3 concentrations were significantly greater than surface O3 levels during the morning, but lower tropospheric O3 did not influence surface morning O3 concentrations.

ACKNOWLEDGEMENTS

First and foremost, I wish to thank my supervisor, Wayne Evans, for his support, patience and trust throughout this project and for his enlightening conversations. I also wish to thank Eldon Puckrin, whose time, guidance and expertise was invaluable. Erin Bell also deserves recognition for her involvement in organizing committee meetings and the partnership between Enbridge Consumers Gas and myself, and for her great advice. I would like to thank Ken Fowler for his help in the experimental design of this project and Kerry Humphery, for helping to take measurements during the summertime. I would also like to express my appreciation to Tim Adamson and Tom Hutchinson, for their interest in this project and their direction. Enbridge Consumers Gas and NSERC also deserves recognition for their financial support and the Ontario Ministry of the Environment for providing their ozone and nitrogen oxide measurements. I would like to express my gratitude to The Department of Environment, Health and Safety at Enbridge Consumers Gas, for opening my eyes to the industrial and corporate world and who made me feel welcome throughout my internship. Finally, I would not have been able to continue my education without the love, support and understanding of my parents, George and Karen, and my future husband, Jason. Thank you for always believing in me and for always being there when I needed you.

TABLE OF CONTENTS ABSTRACT ACKNOWLEDGMENTS TABLE OF CONTENTS LIST OF FIGURES LIST OF TABLES LIST OF ABBREVIATIONS

ii iii iv vii xii xv

1. INTRODUCTION 1.1 Outline 1.2 Objectives and Species Studied 1.3 Pollution Meteorology 1.3.1 The Influence of Wind Direction on the Concentrations of Ground Level Air Pollutants 1.4 The Source of Ozone Pollution in Peterborough and Central Ontario 1.4.1 Stratospheric-Tropospheric Ozone Exchange 1.4.2 Determining the Source of Ozone Pollution in Peterborough and Central Ontario 1.5 Seasonal Variations of Pollutant Concentrations 1.6 Diurnal Variations of Ozone and Nitrogen Oxides 1.7 Lower Tropospheric Ozone Concentrations

1 1 1 4

2. THEORY AND INSTRUMENTATION 2.1 Introduction 2.2 Principles of Infrared Spectroscopy 2.2.1 Absorption of Infrared Radiation by Molecules 2.2.2 Thermal Emission Infrared Spectroscopy 2.2.3 The Bouguer-Beer-Lambert Absorption Law 2.3 Fourier Transform Infrared Spectroscopy 2.3.1 The Michelson Interferometer and Optical System 2.3.2 Fourier Transformation 2.3.3 Resolution 2.3.4 Signal-to-Noise Ratio 2.3.5 Infrared Detectors 2.3.5.1 Detector Linearity 2.4 TEI Ozone Analyzer – Principle of Operation 2.5 TEI NO-NO2-NOx Analyzer – Principle of Operation

12 12 12 13 15 16 19

3. METHODOLOGY 3.1 Introduction 3.2 Ground Level Long Path Air Pollution Measurements

29 29 29

4 6 6 7 8 9 9

20 21 22 23 25 25 26 26

3.2.1 3.2.2 3.2.3 3.2.4

3.3

3.4

Measurement Parameters and Activities Experimental Setup Modeling Ground Level Air Pollutants Error Analysis for Ground Level Air Pollution Measurements 3.2.4.1 Nitrous Oxide Cross Section Analysis 3.2.4.2 Carbon Dioxide Analysis Thermal Emission Measurements 3.3.1 Experimental Technique 3.3.2 Modeling Lower Tropospheric Ozone Concentrations 3.3.2.1 Atmospheric Profiles 3.3.2.2 Zenith Angle 3.3.2.3 Effect of Cloud Height 3.3.2.4 Effect of Water Vapour 3.3.2.5 Effect of Pressure Profile 3.3.2.6 Effect of Temperature Profile 3.3.2.7 Effect of Ozone Above Clouds 3.3.2.8 Total Uncertainty Associated with Thermal Emission of Ozone Measurements Statistical Analysis

4. RESULTS AND DISCUSSION OF GROUND LEVEL LONG PATH AIR POLLUTION MEASUREMENTS 4.1 Introduction 4.2 Seasonal Variations of Ground Level Air Pollutant Concentrations 4.2.1 Monthly Variations of Air Pollutant Concentrations 4.2.1.1 Monthly Variation of Carbon Dioxide 4.2.1.2 Monthly Variation of Carbon Monoxide 4.2.1.3 Monthly Variation of Methane 4.2.1.4 Monthly Variation of Nitrous Oxide 4.2.1.5 Monthly Variation of Ozone 4.2.1.6 Monthly Variation of Nitric Oxide and Nitrogen Dioxide 4.2.2 Seasonal Variations of Air Pollutant Concentrations 4.3 The Influence of Wind Direction on the Concentrations of Ground Level Air Pollutants 4.3.1 Carbon Dioxide 4.3.2 Carbon Monoxide 4.3.3 Methane 4.3.4 Nitrous Oxide 4.3.5 Ozone 4.3.6 Oxides of Nitrogen

30 31 32 37 37 38 38 39 40 42 46 47 47 47 48 48

49 49

51 51 54 54 54 55 62 63 64 65 66 70 71 71 72 73 74 75

4.4

4.5

Source of Ozone Pollution in Peterborough and Central Ontario 4.4.1 Relationship between Ozone and Carbon Monoxide 4.4.1.1 Spring 4.4.1.2 Summer 4.4.1.3 Fall 4.4.1.4 Winter 4.4.2 Relationship between Ozone and Oxides of Nitrogen 4.4.3 Relationship between Oxides of Nitrogen and Carbon Monoxide 4.4.4 Relationship between Ozone and Temperature Diurnal Variations of Ozone and Oxides of Nitrogen 4.5.1 Ozone 4.5.2 Oxides of Nitrogen

86

86 86 88 89 89 90 96 96 105 105 106

5. RESULTS AND DISCUSSION OF LOWER TROPOSPHERIC OZONE CONCENTRATIONS 5.1 Introduction 5.2 Measurements of Lower Tropospheric Ozone 5.3 Comparison between Lower Tropospheric Ozone Concentrations and Ground Level Ozone Concentrations

114

6. SUMMARY AND CONCLUSIONS 6.1 Seasonal Variations 6.2 Source of Ozone Pollution in Peterborough and Central Ontario 6.3 Influence of Wind Direction 6.4 Diurnal Variations of Ozone and Oxides of Nitrogen 6.5 Lower Tropospheric Ozone Concentrations 6.6 Final Conclusions 6.7 Future Research

120 120 121

7. REFERENCES

128

APPENDIX I. Summary of Descriptive Statistics for Compounds Measured Separated into Wind Sector

136

114 114 114

122 123 124 125 125

LIST OF FIGURES

Figure 1.1 Generalized synoptic weather pattern over southern Ontario. This type of event can lead to the accumulation of air pollutants and provide favourable conditions for the production of photochemical smog (MOE, 1998)

Page 5

Figure 1.2 Characteristic diurnal cycle of ozone (oxidant), nitric oxide and nitrogen dioxide during a smoggy day in an urban area (Botkin and Keller, 1995).

10

Figure 2.1 The electromagnetic spectrum illustrating the various types of electromagnetic waves. Infrared waves occur between the wavelengths of 1 mm to 7x10-7 m. (Goodacre, 1997).

12

Figure 2.2 Infrared spectrum of the mid-infrared region between 600 to 4000 cm-1

17

Figure 2.3 Planck blackbody radiance curves for temperatures of 200 K, 18 300 K and 400 K. Also illustrated is a thermal emission spectrum of the atmosphere between the region of 600 to 3000 cm-1 Figure 2.4 The principles of FTIR spectrometry shown for infrared absorption spectroscopy

19

Figure 2.5 A schematic diagram of a Michelson interferometer. The primary components of the interferometer include a fixed mirror, a moving mirror and a beamsplitter (Weisstein)

20

Figure 2.6 A typical interferogram without apodization. The large amplitude in the middle of the interferogram is known as the centerburst (Weisstein).

21

Figure 2.7 Schematic diagram of the TEI Model 49C Ozone Analyzer (TEI Model 49C, 1998).

27

Figure 2.8 Schematic diagram of the TEI NO-NO2-NOx analyzer (TEI Model 42C, 1999).

28

Figure 3.1 Map of southwestern and central Ontario

30

Figure 3.2 Instrumental setup used for ambient ground level measurements of carbon monoxide, carbon dioxide, methane and nitrous oxide using a Bomem FTS

32

Figure 3.3 Transmission spectrum of carbon monoxide taken using long-path absorption spectroscopy. The figure illustrates a measured transmission and two simulated spectra of carbon monoxide with concentrations of 0 ppbv and 178.8 ppbv. The absorption peak at 2099.1 cm-1 was used to determine the carbon monoxide concentration in the atmospheric path since it is free from the interference of other gases.

33

Figure 3.4 Transmission spectrum of carbon dioxide taken using long34 path absorption spectroscopy. The figure illustrates a measured transmission and two simulated spectra of carbon dioxide with concentrations of 0 ppmv and 337 ppmv. The absorption peak at 2056.5 cm-1 was used to determine the carbon dioxide concentration in the atmospheric path since it is free from the interference of other gases. Figure 3.5 Transmission spectrum of methane taken using long-path absorption spectroscopy. The figure illustrates a measured transmission and two simulated spectra of methane with concentrations of 0 ppmv and 1.724 ppmv. The absorption peak at 3038.5 cm-1 was used to determine the methane concentration in the atmospheric path since it is free from the interference of other gases.

35

Figure 3.6 Transmission spectrum of nitrous oxide taken using longpath absorption spectroscopy. The figure illustrates a measured transmission and two simulated spectra of nitrous oxide with concentrations of 0 ppbv and 290.2 ppbv. The absorption peak at 2201.5 cm-1 was used to determine the nitrous oxide concentration in the atmospheric path since it is free from the interference of other gases.

36

Figure 3.7 Transmission spectrum of nitrous oxide taken with a 5 cm gas cell. The figure illustrates a measured transmission of 181.3 mbar of nitrous oxide and two simulated spectra of 181.3 mbar and 197.9 mbar nitrous oxide.

39

Figure 3.8 A thermal emission spectrum of a cloud between 600 and 2000 cm-1 measured at the surface on August 20 1999 with a MCT detector. The region between 980 to 1100 cm-1 is due to ozone emission

42

Figure 3.9 Curve A represents an expanded view of a thermal emission measurement taken on August 20 1999 at 8:30 a.m. for the ozone region between 990 to 1070 cm-1. The cloud altitude

43

was 2.1 km and the temperature of the cloud base was estimated to be 273.75 K. Also shown is a FASCOD3 simulation in the absence of ozone which is represented by Curve B. The resulting emission due to ozone is seen by Curve C which is simply the subtraction of Curve B from Curve A. Figure 3.10 Curve A represents a FASCOD3 simulation of the thermal emission of ozone beneath a 273.75 K blackbody at an altitude of 2.13 km with a constant mixing ratio of 30 ppbv. Curve B represents the same FASCOD3 simulation without any ozone present. Subtraction of Curve B from Curve A yields Curve C

44

Figure 3.11 A comparison of the observed and simulated ozone emission beneath a cloud at an altitude of 2.13 km where Curves A and B represent simulations with constant mixing ratios of 61 ppbv and 30 ppbv respectively. Curve C is the measured ozone emission. Curve A shows good agreement with the measured ozone emission, seen by Curve D which represents the difference between Curve A and Curve C

45

Figure 3.12 A thermal emission spectrum of the atmosphere beneath a cloud measured at the surface on August 20 1999 with an InSb detector. The region between 2400 to 2500 cm-1 was used to determine the cloud temperature since it is free from water absorption.

46

Figure 4.1 Comparison of monthly variations of carbon dioxide (A), 56 carbon monoxide (B), methane (C), nitrous oxide (D), ozone (E), nitric oxide (F) and nitrogen dioxide (G). The points represent the arithmetic mean concentration observed in Peterborough Ontario from May 1999 to March 2001 for each species. Error bars represent the standard deviation of the mean. Figure 4.2 Relationship between fall and winter time carbon dioxide concentrations and carbon monoxide concentrations to determine whether anthropogenic emissions contribute to the increase in carbon dioxide during these months

62

Figure 4.3 Summary of the seasonal variations of carbon dioxide (A), carbon monoxide (B), methane (C), nitrous oxide (D), ozone (E), nitric oxide (F) and nitrogen dioxide (G) for spring, summer and fall 1999, spring, summer and fall 2000, and

68

winter 2001. Figure 4.4 Comparison between oxides of nitrogen levels measured at Trent University and at a MOE monitoring station located 10 km away in downtown Peterborough

70

Figure 4.5 Polar scatter plots of carbon dioxide (A), carbon monoxide (B), methane (C), nitrous oxide (D), ozone (E), nitric oxide (F) and nitrogen dioxide (G) concentrations as a function of wind direction in Peterborough Ontario. Radial axis represents pollutant concentration and the angluar axis represents wind direction

77

Figure 4.6 Mean concentration of carbon dioxide (A), carbon monoxide (B), methane (C), nitrous oxide (D), ozone (E), nitric oxide (F) and nitrogen dioxide (G) as a function of wind sector and season

81

Figure 4.7 Scatter plots and linear regression fits of ozone with carbon 92 monoxide concentration for spring 1999 (A), spring 2000 (B), summer 1999 (C), summer 2000 (D), fall 1999 (E), fall 2000 (F) and winter 2001 (G). Data has been separated into north and south wind sectors and regression lines represent data from the north wind sector (red line), the south wind sector (blue line) and for both wind sectors (solid black line). Table 4.14 displays the results of the regression analysis Figure 4.8 Scatter plot of ozone concentrations with relative humidity for 94 spring 1999 (A) and spring 2000 (B). Data has been separated into north and south wind sectors and regression lines represent data from each wind sector and for both wind sectors combined Figure 4.9 Scatter plot of ozone concentrations with relative humidity for 95 winter 2001. Data has been separated into north and south wind sectors and regression lines represent data from each wind sector and for both wind sectors combined Figure 4.10 Scatter plots and linear regression fits of ozone with nitrogen oxides concentration for spring 2000 (A), summer 2000 (B) and winter 2001 (C). Data has been separated into north and south wind sectors and regression lines represent data from the north wind sector (red line), the south wind sector (blue line) and for both wind sectors (solid black line). Table 4.17 displays the results of the regression analysis..

101

Figure 4.11 Scatter plots and linear regression fits of carbon monoxide with nitrogen oxides concentration for spring 2000 (A), summer 2000 (B) and winter 2001 (C). Data has been separated into north and south wind sectors and regression lines represent data from the north wind sector (red line), the south wind sector (blue line) and for both wind sectors (solid black line). Table 4.18 displays the results of the regression analysis

102

Figure 4.12 Scatter plots and linear regression fits of ozone 103 concentration with temperature for spring 1999 (A), summer 1999 (B), spring 2000 (C), summer 2000 (D), fall 1999 (E), fall 2000 (F) and winter 2001 (G). Table 4.19 displays the results of the regression analysis. Figure 4.13 Diurnal variation of ozone for April to August 2000, October to December 2000, and January to March 2001. Points represent the mean hourly concentration of ozone taken during the respective months. Error bars represent single standard deviation

108

Figure 4.14 Diurnal variation of nitric oxide for April to August 2000, December 2000, and January to March 2001. Points represent the mean hourly concentration of nitric oxide taken during the respective months. Error bars represent single standard deviation

110

Figure 4.15 Diurnal variation of nitrogen dioxide for April to August 2000, December 2000, and January to March 2001. Points represent the mean hourly concentration of nitrogen dioxide taken during the respective months. Error bars represent single standard deviation

112

Figure 5.1 Box and whisker plot of lower tropospheric ozone concentrations and surface ozone concentrations.

118

LIST OF TABLES

Table 1.1 Estimated sources and sinks of carbon dioxide, carbon monoxide, methane, nitrous oxide, ozone and oxides of nitrogen (NOx =NO+NO2)

Page 3

Table 3.1 FTS measurement parameters for long-path air pollution monitoring

31

Table 3.2 FTS measurement parameters for measuring the thermal emission of ozone beneath clouds

40

Table 4.1 Descriptive statistics of the compounds measured at Trent University in Peterborough Ontario from spring 1999 to winter 2001

52

Table 4.2 Observed temperature and precipitation data reported by Environment Canada for central Ontario for 1999, 2000 and winter of 2001.

53

Table 4.3 Descriptive statistics for compounds measured in Peterborough Ontario for 1999. Data is separated into months of year

58

Table 4.4 Descriptive statistics for compounds measured in Peterborough Ontario for 2000. Data is separated into months of year.

59

Table 4.5 Descriptive statistics for compounds measured in Peterborough Ontario for 2000. Data is separated into months of year

61

Table 4.6 Results of the Kruskal-Wallis ANOVA applied to the seasonal differences of carbon dioxide, carbon monoxide, methane, nitrous oxide, ozone and nitrogen oxides

67

Table 4.7 Results of the Kruskall-Wallis ANOVA by ranks test for wind 75 sector and the concentration of carbon dioxide, carbon monoxide, methane, nitrous oxide, ozone and nitrogen oxides separated by season Table 4.8 Multiple comparison of wind sector and carbon dioxide concentration for fall 1999, spring and summer 2000 and winter 2001

83

Table 4.9 Multiple comparison of wind sector and carbon monoxide concentration for spring and summer 1999, spring, summer and fall 2000 and winter 2001

83

Table 4.10 Multiple comparison of wind sector and methane concentration for spring and summer 1999, spring, summer and fall 2000 and winter 2001

84

Table 4.11 Multiple comparison of wind sector and nitrous oxide concentration for spring and summer 1999, spring, summer and fall 2000 and winter 2001

84

Table 4.12 Multiple comparison of wind sector and ozone concentration for spring and summer 1999, spring, summer and fall 2000 and winter 2001

85

Table 4.13 Multiple comparison of wind sector and concentrations of nitrogen oxides for spring and summer 2000 and winter 2001.

85

Table 4.14 Spearman correlation coefficients for ozone and carbon monoxide measured in Peterborough in 1999 to 2001. Correlation coefficients were found for all wind directions and then separated into north and south wind direction. Highlighted R values indicate that relationship was significant (p