UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An investigation of ozone distribution downwind of Greater Vancouver, British Columbia using a novel… O’Kane, Stephen M. 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1997-0185.pdf [ 6.4MB ]
Metadata
JSON: 831-1.0052592.json
JSON-LD: 831-1.0052592-ld.json
RDF/XML (Pretty): 831-1.0052592-rdf.xml
RDF/JSON: 831-1.0052592-rdf.json
Turtle: 831-1.0052592-turtle.txt
N-Triples: 831-1.0052592-rdf-ntriples.txt
Original Record: 831-1.0052592-source.json
Full Text
831-1.0052592-fulltext.txt
Citation
831-1.0052592.ris

Full Text

AN INVESTIGATION OF OZONE DISTRIBUTION DOWNWIND OF GREATER VANCOUVER, BRITISH COLUMBIA USING A NOVEL AIRCRAFT MEASUREMENT SYSTEM By Stephen M . O'Kane B.Sc. (Atmospheric Science) University of British Columbia, Vancouver, B.C., 1993 A T H E S I S S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S A T M O S P H E R I C S C I E N C E P R O G R A M M E We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1997 © Stephen M . O'Kane, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of BritisrkjZolumbia Vancouver, Canada Department DE-6 (2/88) ABSTRACT A fast response chemiluminescent ozone sonde originally intended for use as a balloon borne instrument, was deployed on a single engine light aircraft to determine mean ozone concentrations in the boundary layer downwind of a major urban centre. Vancouver is a coastal city at the mouth of a major river delta and valley walled in by mountains to the north and east. Hence, the pollutant distribution during anti-cyclonic conditions is influenced by the thermally induced flows of the sea/land breeze and mountain/valley wind systems. This study focused on ozone concentration in a steep walled, glacial valley near the city, where ozone concentrations had not been previously monitored but were suspected to be high. Measurements confirmed this hypothesis and the mass budget of ozone for the valley was calculated for a period shortly after sunrise when the rate of boundary layer growth was at its maximum. The high levels of ozone found in the valley were partly attributed to residual layer storage of ozone above the nocturnal stable layer. This overnight storage of ozone during sustained anti-cyclonic conditions suggests that the Pitt River valley and possibly other tributary valleys represent an important net sink for ozone emanating from precursor sources in the Lower Fraser Valley. ii Table of Contents Abstract n List of Tables _ v List of Figures V 1 Chapter 1: Introduction 1 1.1 Introduction . 1 1.2 Air Pollution Meteorology. 4 1.3 Aircraft Measurements 13 1.4 Objectives 14 Chapter 2: Measurements and Method of Analysis 16 2.1 Introduction_ . 16 2.2 Tethersonde Measurements 16 2.3 Aircraft Measurements 17 2.4 Passive Samplers. 19 2.5 Ancillary Data '. 21 2.6 Mass Budget Technique 22 2.7 Summary_ 25 Chapter 3: A Novel Airborne Survey Method 26 3.1 Introduction 26 3.2 System Design _27 3.3 Ozone Sonde Calibration and Sensitivity 28 3.4 Measurements and Determination of Position 33 3.5 Flight Lengths and Stationarity of the Sample 33 3.6 Validation of the Aircraft Measurements 35 Chapter 4: The Spatial Distribution of Ozone Across the Lower Fraser Valley and the Pitt River Valley 39 4.1 Introduction 39 4.2 Study Period 39 4.3 July 22, 1994 - High Ozone Concentrations 41 4.4 July 27,1994 - Moderate Ozone Concentrations 46 4.5 August 2, 1994 - Moderate Ozone Concentrations 49 4.6 August 13, 1994 - Moderate Ozone Concentrations 53 4.7 August 30, 1994 - Low Ozone Concentrations • 56 iii 4.8 Passive Samplers . 59 4.9 Ozone Gradients in the Pitt River and Fraser Valleys - A Comparison_60 Chapter 5: Processes Contributing to the Observed Spatial Distribution of Ozone 63 5.1 Introduction . 63 5.2 Potential Processes Contributing to Elevated Ozone Concentration in the Pitt River Valley 64 5.3 13 August, 1994: A Case Study 69 5.4 The Mass Budget of Ozone for the Pitt River Valley 13 August, 1994, Early Morning 76 5.5 Entrainment and Residual Layers, A Conceptual Model 83 Chapter 6: Conclusions 85 6.1 Objectives 85 6.2 Future Research 89 Bibliography, 91 iv List of Tables 1. Measurement Schedule and Location List of Figures 1. Map of the area under study 2 2. Mixed layer and residual layer formation as present in Garratt (1992) pp.146 8 3. The injection of pollutants into the inversion layer by slope flows, from Lu and Turco (1994) 9 4. Comparison of passive sampler measured average ozone concentration to G V R D monitor measured ozone concentration 21 5. The Mass budget of an atmospheric constituent Q, in a layer with top at height Zj and bottom at z^ 24 6. Configuration of the GFAS, OS-B-2 ozone sonde 28 7. Air intake design for the aircraft ozone sonde 29 8. Calibration curve and equation of the OS-B-2 ozone sonde 30 9. Temperature sensitivity curve and equation for the OS-B-2 ozone sonde 32 10. Validation of the aircraft measurement technique. Comparison of 5 minute average surface measurements to 20 second average aircraft measurements of ozone concentration 37 11. Comparison of ozone measurements made using a tethered balloon and a potassium iodide solution based ozone sonde to measurements using a light aircraft and chemiluminescent sonde 38 12. Maximum hourly ozone concentration measured at the G V R D surface monitor station 16 ; 40 13. Ozone profile 21 July, 1994. Harris Road 43 14. 500 hPa and surface analysis for 22 July 1994 44 15. Ozone isopleths and aircraft trace for 22 July 1994 45 16. 500 hPa and surface analysis for 27 July 1994 47 vi 17. Ozone isopleths and aircraft trace for 27 July 1994 48 18. 500 hPa and surface analysis for 2 August 1994 50 19. Ozone isopleths and aircraft trace for 2 August 1994 51 20. 500 hPa and surface analysis for 13 August 1994 54 21. Ozone isopleths and aircraft trace for 13 August 1994_ 55 22. 500 hPa and surface analysis for 30 August 1994 57 23. Ozone isopleths and aircraft trace for 30 August 1994 58 24. Average ozone measured by passive sampler 59 25. Ozone gradients in the Pitt River and Lower Fraser Valley 62 26. Profiles of atmospheric variables measured 13 August 1994, 0903 PST 71 27. Ozone traces measured by aircraft 13 August 1994 73 28. Mean hourly concentrations 13 August 1994 75 29. Profiles of atmospheric variables measured 13 August 1994, 1348 PST 81 vii C H A P T E R 1: I N T R O D U C T I O N 1.1 I N T R O D U C T I O N The latter half of the twentieth century has seen air pollution meteorology become an increasingly important field of study in the atmospheric sciences. As the world's population and cities have grown exponentially in the past two hundred years, so too has the problem of air pollution. Many large cities such as Mexico City, Shanghai, Athens, and Los Angeles have frequent air pollution episodes that are detrimental to human health, particularly of the very old, very young, or infirm. Extreme air pollution episodes usually occur when the local meteorology restricts dispersion and is conducive to the formation of secondary pollutants such as ozone and fine particulates (Stern et al, 1973). Poor air quality is not restricted to the largest metropolitan centres around the world. Smaller less industrialized cities can experience meteorological conditions which restrict dispersion and lead to air quality conditions that can affect human health and the local ecology. Greater Vancouver, British Columbia is a metropolitan area on the Pacific Northwest Coast of North America (Figure 1). The urban area is situated at the mouth of the Lower Fraser Valley (hereafter referred to as the LFV) in an area of marked relief. Terrain to the north of the city rises from sea level to heights in excess of 1800 m and there are a number of tributary valleys that flow out of the mountainous areas into the L F V . This area has a complex thermally driven circulation which is evident during anti-cyclonic conditions, and helps create episodes of poor air quality (Steyn and McKendry, 1988). Under stagnant synoptic conditions, the transport and dispersion of pollutants emanating from Greater Vancouver is strongly influenced by the thermally forced flows 1 Figure 1. Map of the area under study. Top map shows the Lower Fraser Valley and its tributary the Pitt River valley, and the urban areas. Bottom map shows the position of the G V R D surface ozone monitors, the passive sampler locations (marked with PS), the tethered balloon site, and the flight paths for the aircraft. Ozone gradients were calculated northward and eastward from Pitt Meadows Airport. 2 of the sea/land breeze and the mountain/valley winds (Steyn and McKendry, 1988). To a lesser extent, the urban/rural circulation of Greater Vancouver may also affect the transport and dispersion of pollutants (Oke,1987). Ground level ozone is a pollutant of particular concern in Greater Vancouver (GVRD, 1994). Ozone is a secondary pollutant formed from the photo-disassociation of oxides of nitrogen and volatile organic compounds. In Greater Vancouver, the major source of ozone precursors is motor vehicles (Steyn et al, 1992). When vehicle emissions are exposed to strong solar radiation and warm temperatures a number of complex chemical reactions take place to form ozone (Gtisten, 1986). The synoptic conditions that are most conducive to forming ozone in the L F V are also those that are conducive to forming a convectively dominant boundary layer and allow thermally induced circulations to strengthen, and dominate the synoptic wind systems. The L F V experiences these synoptic conditions usually during the summer months (McKendry, 1993). The Greater Vancouver Regional District (GVRD), and the B.C. Ministry of Environment, operate an air quality monitoring network in the L F V with 23 surface stations measuring ozone and other pollutants (Figure 1). Maximum ozone concentrations are measured during the summer months under anti-cyclonic conditions (GVRD, 1994). Through the 1980's, the monitoring network recorded an exceedance of the Federal Objective 1 hour Maximum Acceptable ozone concentration of 82 ppb, an average of 160 times a year. The distribution of ground-level pollutant monitoring stations are usually sparse in rural areas, as in the case of Greater Vancouver and the L F V . Often monitoring networks are biased toward emission sources rather than secondary pollutant pathways. Therefore, the distribution of pollutants are not accurately resolved by standard monitoring networks, 3 making it difficult to assess impacts on visibility, human health or the local ecology. This is the case with the G V R D monitoring network of the L F V , as can be seen in Figure 1. There are few stations downwind of Greater Vancouver in the eastern L F V and no stations to the north-east of the city in the tributary valleys of the L F V . An ozone measurement program in the L F V that provides data on ozone concentration in the more remote parts of the area is necessary to: • determine impacts on ecology, visibility and human health downwind of Greater Vancouver, particularly in the undeveloped areas of the tributary valleys of the L F V ; • determine the mass budget of ozone, particularly sources and sinks, and; • validate numerical models. 1.2 A I R P O L L U T I O N M E T E O R O L O G Y 1.2.1 Synoptic Air quality conditions are determined by emission factors and the state of the atmosphere. Although atmospheric conditions are partly controlled by processes at scales larger than synoptic, generally this is the largest scale examined when investigating regional air pollution meteorology (Stern et al, 1973). The synoptic conditions that produce air pollution problems differ from one region to the next and from one pollutant type to the next. For example, the synoptic conditions that are conducive to elevated ozone conditions may not produce the worst case conditions for particulate matter. Mid-winter high pressure systems in North America can restrict dispersion of pollutants when winds are calm and boundary layers are shallow, but will not be conducive to ozone formation as temperatures and solar radiation angles are low. Although the specific 4 synoptic conditions that cause poor air quality episodes vary, the restriction of dispersion of pollutants is at the root of any air pollution problem. Synoptic conditions that inhibit dispersion are often associated with high pressure systems and slack surface pressure gradients (Stern et al, 1973). An upper level ridge will cause subsidence which will suppress boundary layer heights, and for the case of summertime ozone problems, produce clear skies, maximising temperatures and solar radiation inputs. A slack surface pressure gradient promotes low wind speeds and little mechanically produced turbulence, thereby restricting the transport and dispersion of pollutants. In mid-latitude North America, elevated ozone levels downwind of major urban centres are generally well correlated with summertime high pressure systems (McKendry, 1994; Comrie and Yarnal, 1992). An upper level ridge combined with a slack surface pressure gradient, or a stagnating surface anti-cyclone can produce conditions favourable to ozone formation, especially i f temperatures are high and these conditions persist for a few days. Slack surface pressure gradients and high temperatures are also associated with local, thermally produced circulation systems that can further worsen air quality by trapping pollutants in a closed circulation (Atkinson, 1981; Sturman, 1987). 1.2.2 Mesoscale Circulations During summertime anti-cyclonic conditions, mesoscale meteorology is influenced by thermally induced circulations such as the sea/land breeze, and mountain/valley winds (Atkinson, 1981; Sturman, 1987; Steyn and McKendry, 1988). These circulations have a significant impact on air pollution as they can transport, and 5 dilute or concentrate pollutants emanating from a city. The differential heating of the land and sea causes a sea/land breeze system to develop which can trap pollutants in a closed circulation. A strong onshore flow can develop when local, surface pressure gradients are modified by daytime heating of the land and is usually associated with a return flow aloft and a stratified lower atmosphere, (Simpson, 1994; Steyn and Faulkner, 1986; Lalas et al., 1983; Atkinson, 1981). At night the circulation reverses and a weak land breeze can be observed. The diurnal reversal of the circulation can result in pollutants being transported back and forth across a coastal area with little dispersion (Steyn, 1995). The daytime sea breeze is of a larger scale than the nocturnal land breeze, in strength, depth and area affected. In some areas, the sea breeze can penetrate over 50 km inland at greater wind speeds and through a deeper boundary layer than the land breeze which is barely measurable except very near the coast (Atkinson, 1981). Thus, pollutants emanating from a coastal urban area are forced inland during the day and can remain in one area over night or even begin to move back to the source region in the land breeze. If synoptic conditions persist for a few days then the sea/land breeze has the potential to concentrate pollutants inland of coastal cities. Differential heating of slopes and valleys can also cause other thermal circulations that are most strongly developed during anti-cyclonic conditions (Atkinson, 1981; Sturman, 1987). In coastal complex terrain, there is potential for such mountain/valley winds to accentuate the sea/land breeze. As both circulations are a direct result of solar heating of the surface, the diurnal cycles of both are quite similar. A night-time land breeze can be associated with mountain winds and the circulation is reversed in the morning when solar heating of the land changes local pressure gradients. As with the 6 diurnal reversal of the sea/land breeze the daytime valley winds are stronger and blow through a deeper boundary layer than the nocturnal mountain winds (Atkinson, 1981; Sturman, 1987). Thus, daytime advection into mountain valleys near a coast can be of greater magnitude than nocturnal flow out of the valleys. The thermal circulations described above are well documented and have been observed in coastal and mountainous locations around the world (Atkinson, 1981). For large coastal cities such as Athens, Los Angeles, Tokyo, and Greater Vancouver thermal circulations may contribute to ground level ozone problems during anti-cyclonic conditions (Steyn, 1996; Chang et al, 1989; Lalas et al, 1983; McElroy and Smith, 1993; Steyn and McKendry, 1988). Interaction of thermal mesoscale flows with boundary layer evolution and decay can produce variations in the vertical distribution of pollutants. This is discussed in the next section. 1.2.3 Boundary Layer Effects Air quality is strongly affected by boundary layer processes. Well capped shallow boundary layers can trap pollutants in the lowest level of the atmosphere restricting dispersion and increasing concentrations (Stern et al, 1973). Under stagnating anti-cyclonic conditions, the diurnal cycle of growth and decay of the boundary layer can store pollutants aloft away from surface sinks, concentrating pollutants over one area. Air pollution can remain aloft over night in layers formed from the growth and decay of the boundary layer. Stull (1988) and Garratt (1992) describe residual mixed layers formed when nocturnal surface inversions undercut the developed mixed layer which is capped 7 by a subsidence inversion (Figure 2). Pollutants trapped in the residual layer are entrained into the growing mixed layer of the following day, adding to the concentration of new emissions. z t f ' ' ~ ' • : \: " ^ f v ' ~- ' ~~ .- Time •Risc;!A./:. B Set Figure 2. Representation of mixed layer and residual layer formation throughout the diurnal period over land under clear skies, after Garratt (1992) pp.146 Lu and Turco (1994) also describe a type of elevated residual layer formed from the injection of pollutants into the inversion layer by topographic flows and the undercutting of the mixed layer by the sea breeze (Figure 3). Similar residual layers of ozone were observed by McElroy and Smith (1992). McElroy and Smith (1992) also attributed the residual layers to the undercutting of the mixed layer by the sea breeze. These elevated layers of pollution also have the potential to be entrained into the growing mixed layer of the following day. 8 Free atmosphere Figure 3. An idealised view of a coupled sea breeze and mountain slope circulation in the vicinity of a high coastal mountain range, showing the injection of pollutants into the inversion layer by slope flows, from L u and Turco (1994). In the summer of 1990, a large scale air pollution study in complex terrain over the Swiss midlands, POLLUMET, investigated the dynamics of photo-chemical smog . In that study Wanner et al (1993) found that the concentrations of ozone within the boundary layer were spatially variable and were determined not only by emission factors but by the complex, thermally driven flows of the area. They found that mountain/valley winds dominated local wind patterns during ozone episodes and the alpine valleys experienced high levels of ozone concentration that peaked in late afternoon when up slope flows were at a maximum. At night, clean, katabatic flows were found to emanate from these valleys. Warmer et al (1993) stated that ozone destruction in the boundary layer was primarily at night and confined to the lowest 120 m of the atmosphere, leaving elevated residual layers of ozone above the downslope flows. They attributed ozone destruction to surface deposition and chemical scavenging of ozone by NO x . 9 Entrainment of residual layers of ozone into evolving mixed layers was investigated during P O L L U M E T by Neuet al (1994), who found that residual layer ozone could contribute up to 70 % of the ozone in the mixed layer formed the following day. They also described a transilient turbulence model in which ozone is mixed down into the growing mixed layer of the morning. The residual ozone layers described in the studies by Luand Turco (1994), Warmer et al (1993), McElroy and Smith (1992), and Neu et al (1994) are similar to those described by Garret (1992) and Stull (1988) but are more closely tied to the dynamics of the thermal flows than simply the remnant mixed layer that can be observed after the formation of the nocturnal inversion. 1.2.4 P R E V I O U S R E S E A R C H IN T H E L O W E R F R A S E R V A L L E Y There have been a number of studies conducted in the L F V concerning ozone. Pry or et al (1995) investigated the synoptic meteorology associated with ground level ozone, while McKendry (1993) used a statistical technique to identify the synoptic patterns of the surface and 500 hPa geopotential height level, associated with high ground level ozone concentrations. Taylor (1992) investigated the relationship between surface pressure gradients and 850 hPa level temperatures and ozone concentrations. The meteorological variables used to forecast elevated ozone concentrations in the L F V were identified by Taylor (1991), and Steyn et al (1990) described the meteorological characteristics that are favourable to elevated ozone concentrations. The synoptic pattern that is associated with the highest mean ozone concentrations in the L F V is a low level, thermal trough near the south-west B.C. coast together with an 10 upper level ridge centred over the middle of B.C. just east of the L F V (Pryor et al, 1995; McKendry, 1994). This pattern produces strong subsidence, weak surface pressure gradients, high temperatures and clear skies. These conditions are conducive to the formation of high ozone concentrations as dispersion is restricted due to a well capped mixed layer and light winds, and photo-chemical production of ozone is maximised with strong radiation and high temperatures. Ozone concentrations are greatest when these conditions persist for a few days, suggesting that high ozone concentrations are linked to concentrations in the mixed layer of the previous day (McKendry, 1994). The synoptic conditions conducive to ozone formation in the L F V are also linked to the development of the sea breeze of the area (Steyn and Faulkner, 1986). Consequently, these are also the right conditions for the development of valley winds. The simultaneous diurnal change in direction of these two types of circulation in the L F V results in no true sea breeze front being observed (Steyn and McKendry, 1988). The flow moves onshore and upslope at the same time and thus the entire air mass of the L F V begins to move at about the same time. Thus the observed daytime winds during summertime high pressure systems on the coast of the L F V are westerly to south-westerly. Farther inland near the middle of the L F V the winds are more westerly and along the northern edge of the L F V the flow is more southerly due to differential heating of the mountain sides and valley floor. The movement and dispersion of pollutants by the thermal flows of the L F V , determine mixed layer concentrations and therefore must be examined closely when predicting the maximum daily ozone concentrations. One area of particular interest in this study was a tributary of the L F V , the Pitt River valley, which is north-east of Greater Vancouver (Figure 1). Historically, the area 11 at the mouth of the Pitt River valley has experienced high levels of ozone during conditions conducive to thermally induced flows (GVRD, 1994; Steyn et al,1990; Evans et al, 1992). Furthermore, a numerical modelling study by Miao, (1993) found that air parcels emanating over the metropolitan area of Greater Vancouver under synoptic conditions conducive to ozone formation, can be transported up the Pitt River valley. However, the ozone concentrations within the Pitt River valley itself have never been measured. The most comprehensive and extensive study of air pollutants in the L F V ever conducted was Pacific '93 which took place in July and August 1993. The month long study was conducted co-operatively by a number of organisations and is to be described in detail in a special issue of Atmospheric Environment. Findings from Pacific '93 of particular interest in the context of this study were: • elevated and residual ozone layers over the central L F V were found by tethered balloon. The characteristic profiles were attributed to nocturnal surface deposition and the formation of the nocturnal surface inversion (Thomson et al, 1995; McKendry et al, 1997); • elevated levels of atmospheric aerosols over the L F V , found by ground based and aircraft fitted lidar systems, (McKendry et al, 1997); • clean and relatively ozone free, katabatic, nocturnal flows out of a tributary valley of the L F V , the Pitt River valley (Banta et al, 1996) and; • aircraft flights revealed high pollutant concentrations in the Pitt River valley on one occasion. 12 In summary, ad hoc modelling and observations suggest that the tributary valleys of the L F V , particularly the Pitt River valley, may play an important role in the mass budget of ozone in the region. 1.3 AIRCRAFT MEASUREMENTS Aircraft have been commonly used to measure ozone within the troposphere (Beekmann et al, 1995; Van Valin et al,1994; Kondo et al, 1987; Lenschow et al, 1981). However, studies such as these generally use dedicated research aircraft with large, multi-instrument systems using UV-photometers or chemiluminescent ozone analysers. Lenschow et al (1981) used a Queen Air aircraft equipped with a chemiluminescent ozone sensor and air motion, temperature, and humidity sensors for measurements within the boundary layer, while Van Valin et al (1994) used a King Air aircraft equipped with an UV-photometer as well as SO2 and NO-NO x analysers along with air motion, temperature, pressure, and humidity sensors when they compared aircraft and surface, air quality measurements. These systems require pressure and flow rate corrections and can be prohibitively expensive. Furthermore, the speed and size of the required aircraft impose constraints on the terrain and altitudes over which the aircraft can operate. Recently, the development of cheap, portable, dry chemiluminescent ozone sondes for balloon deployment offers the potential for cheap airborne systems aboard light aircraft. In this study an inexpensive and robust, dry chemiluminescent ozone sensor originally intended for use as a balloon borne, disposable instrument was fitted to a light aircraft. The design of the instrument package allowed for quick placement aboard a rental aircraft 13 with no modifications to the plane and was a fraction of the cost of an ozone sensor of the same chemiluminescent type but manufactured for use on dedicated research aircraft. In the studies by Van Valin et al (1994) and Lenschow et al (1981) the concentration of ozone measured by aircraft in the boundary layer was higher than surface measurements. This was the same result as Galbally (1971), who documented the increase of ozone with height up to the top of the boundary layer and attributed this vertical pattern to the surface deposition of ozone. A comparison of surface ozone measurements to measurements aloft by tethered balloon also found the same result (Harrison etal, 1978). 1.4 OBJECTIVES Recent research suggests the tributary valleys of the L F V play an important role in the air pollution meteorology of the L F V and may constitute an important component of the mass budget of ozone. Given uncertainties regarding ozone concentrations in the tributary valleys and the processes contributing to them, as well as the limitations imposed by the distribution of surface monitors, the primary objectives of this thesis are: • develop a method of measuring mean ozone concentrations with a light aircraft and a dry, chemiluminescent ozone sonde; • determine the spatial distribution of summertime ozone concentrations within the L F V and its tributary the Pitt River valley, and; • investigate the processes that lead to high ozone concentrations in the Pitt River valley by using a mass budget technique. The next chapter will describe the methods used to reach these goals. The measurement programme and the mass budget technique used for the investigation of 14 ozone concentration change in the Pitt River valley are also explained. Chapter three is a detailed description of the aircraft instrument package including the calibration technique, the determination of position of the observations, the assumptions needed when measuring from a moving platform, and the validation of the aircraft measurements. Chapter four is a summary of the results of the measurement program. The spatial distribution of ozone in the L F V and the Pitt River valley under a variety of synoptic conditions is presented as well as some of the observed differences in ozone concentration between the two valleys. Chapter five is an investigation of the processes that could lead to high ozone concentration in the Pitt River valley. The mass budget of ozone in the Pitt River valley is calculated for the early morning of 13 August 1994. The mass budget is calculated for the time period when mixed layer growth is at a maximum to determine the relative importance of photo-chemical production, advection, and entrainment of ozone in the mixed layer. In the last chapter the measurement program is assessed and the results associated with each of the objectives of this study are summarised. 15 CHAPTER 2: MEASUREMENTS AND METHOD OF ANALYSIS 2.1 INTRODUCTION To investigate the horizontal and vertical distribution of ozone in the L F V and its tributary, the Pitt River valley, a large data set comprising both conventional and novel measurement techniques was collected. Permanent surface ozone monitors and mobile tethered balloons equipped with ozone and meteorological sensors provided reliable data for much of the area under study. However, as a large portion of this area is remote, inaccessible, or not covered by the existing G V R D surface ozone monitoring network, two additional measurement strategies were adopted. Firstly, a light aircraft was fitted with an ozone sonde, originally intended for use as a balloon borne instrument, and used for aerial survey. Secondly, small chemical targets were deployed as passive ozone samplers. With the addition of weather station data, a detailed description of the three dimensional ozone distribution under a variety of meteorological conditions and some of the processes involved in creating this distribution was possible. In the Pitt River valley in particular, the processes involved in the change of ozone concentration could be investigated through a calculation of the mass budget of ozone. 2.2 TETHERSONDE MEASUREMENTS Vertical soundings of ozone, pressure, temperature, wind speed and direction, and relative humidity to an approximate height of 1000 m A G L via tethered balloon were made in the Pitt Valley at Little Goose Island in Pitt Lake, (49° 26' 15" N , 122° 31' 20" W, Figure 1 ). An Atmospheric Instrumentation Research Inc. tethersonde (TS-3A-SPH) with ozone sonde (OZ-3A-T) and a 5 m 3 helium filled kytoon was used for these 16 measurements. The soundings provided data on valley circulations and the vertical distribution of ozone needed for mass budget calculations and parameterizations. The ozone sonde used for balloon flights uses the neutral potassium iodide method to measure oxidants, (McKendry et al, 1997). This sonde was calibrated against a Teco 49 U V photometric ozone analyser and generator which was calibrated against a G V R D ozone monitor. The Teco analyser was accurate to less than + 1 ppb. This allowed direct comparison to ozone measured by the G V R D surface stations. Tethersonde operations coincided with aircraft flights as much as possible to provide both the vertical variations in ozone as well as an independent check on aircraft measurements. 2.3 A I R C R A F T M E A S U R E M E N T S An inexpensive, yet reliable and compact, system was developed to measure ozone and temperature from a single engine light aircraft. Using a chemiluminescent ozone sonde manufactured for use as a disposable, balloon borne instrument, and a thermistor probe, a mobile instrument package was designed for rapid deployment on a rented Cessna 175 or 155. The details of design and calibration of the aircraft instrument package are presented in the next chapter. The light aircraft was used for the measurement of mean ozone concentration and temperature within the boundary layer of the area under study. Most importantly, the aircraft measured ozone in the Pitt River valley and eastern L F V where there are few ozone measurements available and concentrations are expected to be high (Steyn et al, 1990; G V R D 1993; Oke and Hay, 1994; Evans et al, 1992). Proposed flight paths for the study dates of 22, 27 July and 2, 13, 30 August, 1994 are shown in Figure 1. Actual flight 17 paths were flown as close to the original flight plan as possible when air traffic control allowed. Occasional deviations from proposed flight paths were made to avoid aircraft or other potential hazards. Table 1 lists all the days when the aircraft flew, the flight path taken, when tethersonde measurements were made, and the surface ozone monitors which were directly under the aircraft's path. D A T E F L I G H T P A T H T A K E N T E T H E R E D B A L L O O N O P E R A T I O N S | S U R F A C E S T A T I O N S j F O R C O M P A R I S O N 0 8 / 2 2 / 9 4 O 3 - P I T T N / A j 1 5 , 1 6 0 8 / 2 7 / 9 4 O 3 - P I T T O 3 - F R A S E R E A S T O 3 - F R A S E R W E S T N / A j 1 2 , 1 5 , 1 6 , 2 8 0 9 / 0 2 / 9 4 O 3 - P I T T O 3 - F R A S E R E A S T O 3 - F R A S E R W E S T L I T T L E G O O S E I S L A N D j 1 2 , 1 5 , 1 6 , 2 8 0 9 / 1 3 / 9 4 O 3 - P I T T ( 4 F L I G H T S ) L I T T L E G O O S E I S L A N D j 1 5 , 1 6 0 9 / 3 0 / 9 4 O 3 - P I T T ( 2 F L I G H T S ) L I T T L E G O O S E I S L A N D 1 1 5 , 1 6 Table 1. Measurement schedule and locations. Refer to Figures 1 and 2 for locations of flight paths, balloon operations, and surface stations. Although the aircraft flew for up to six hours per day, continuous monitoring was restricted to thirty minute flight legs. This allowed for an assumption of stationarity of the air sample for each measurement time period. The following chapter provides more detail on the spatial and temporal restrictions when measuring from a moving platform. 18 On all study dates, aircraft operations were completed by 1600 PST, when ozone concentrations were near peak levels over much of the study area. In the eastern L F V and northward into the Pitt River valley, ozone concentrations usually peak later in the day so that daily maximum ozone concentrations were usually missed (GVRD, 1994). Flights later in the day were restricted due to logistical problems, but a well resolved representation of the early afternoon spatial distribution of ozone could be made. 2.4 P A S S I V E S A M P L E R S To augment the data set, three passive ozone samplers were deployed in the Pitt River valley, and one was deployed atop Mount Seymour over-looking the L F V (Figure 2). The Pitt River valley is close to Greater Vancouver yet still fairly remote and inaccessible, except by boat. There are no surface ozone monitors within the valley to compare with aircraft data as in the L F V , and no power source or roads for any future permanent stations. The passive samplers are small filters coated with a nitrite salt, (NO2"), onto which ozone is diffused. They require no power source or special platforms with which to operate, and provide only an integrated ozone concentration for the period of deployment. Thus, direct comparison with instantaneous aircraft data is not feasible but comparison to average ozone concentration and gradients can be made. The passive samplers were originally developed as personal ozone monitors to be worn by subjects to measure total exposure to the gas (Koutrakis et al,1993). They are ozone specific and there is no interference from other oxidants common in urban plumes (i.e. NO2). The samplers use a nitrite salt (NO2"), which is oxidised by ozone and forms a nitrate (NO3") according to the reaction: 19 N02 + 0 3 -> M? 3~ + 02 1 Every mole of NO3" formed is then equal to one mole of ozone absorbed (Koutrakis et al, 1993). To determine the average ozone concentration over the sampling period, the amount of NO3" produced is divided by the sampling duration and sampling rate. The sampling rate is based on Fick's Law of diffusion and was validated by Koutrakis et al (1993). The filters were deployed for at least 24 hours then collected and sealed. The filters are then analysed for the nitrate ion, formed by the reaction described by equation 1, by ion chromatography. Thus, the passive samplers provide a cheap and easy method of measuring mean ozone concentration change in areas where accessibility is limited. In this study, the passive samplers were deployed in the Pitt River valley along the valley bottom to determine ozone gradients in the surface layer. One passive sampler was placed beside the G V R D ozone monitor, station 16, at the mouth of the Pitt River valley for validation (Figure 1) and two more were deployed farther north up the valley. A comparison between the Pitt Meadows monitor and the collocated sampler is shown in Figure 4. For integration periods ranging from 24 to 144 hours, the passive sampler values were consistently lower than values integrated from hourly measurements from the monitor. However, the relative change in ozone concentrations found using the two methods did agree. The approach was deemed to be useful for relative ozone measurements along the Pitt River valley as r2 between the monitor data and the passive sampler data was calculated to be 0.76. 20 Comparison of G V R D Measured Ozone Average and Passive Sampler Measured Ozone Average 30 n < < < < < < < < < < < < < < < • I I I I I I I I I I I I | | ~ ~ Date ~ ~ ™ ^ ^ ™ 0 4 Figure 4. Comparison of passive sampler measured ozone concentration to G V R D monitor measured ozone concentration averaged over equivalent time periods. 2.5 A N C I L L A R Y D A T A The current surface based atmospheric chemistry monitoring network operated by the G V R D provided data for much of the L F V (Figure 1). However, these monitoring stations are not optimally located for determination of the spatial distribution of ozone, in part because the network was designed to measure multiple pollutants and has evolved over many years and in a number of municipalities, each with its own air quality and financial priorities. Consequently, monitoring stations tend to reflect the distribution of population and primary pollutants rather than the downwind distribution of secondary pollutants such as ozone. There are no monitoring stations within the Pitt River valley itself and few stations in the eastern part of the L F V downwind of Greater Vancouver, where air quality can be extremely poor (Steyn et al,1990; G V R D , 1993; Evans et al,1992). The surface monitors do provide a continuous data set and some stations have 21 continuous records for ozone extending back to 1978. Each station measures a variety of pollutants depending on its location, however ozone is measured at all stations in Figure 1. Crucial to this study was a surface ozone monitor located near the mouth of the Pitt River valley at the Pitt Meadows airport (station 16, Figure 2 ). This served as a comparison for airborne measurements into the Pitt River valley and as a base station for the passive sampler measurements. Data from an automated weather station operated by the Atmospheric Environment Program of the Canadian government at the Pitt Meadows airport was also used. In addition, meteorological data from weather stations at Vancouver International Airport and Abbotsford airport augmented the data set. 2.6 M A S S B U D G E T T E C H N I Q U E The observations of ozone by aircraft, tethersonde, surface monitor, and passive sampler allowed a detailed description of pollutant distribution in the L F V and the Pitt River valley to be made. Furthermore, these data permit an examination of the dominant processes contributing to the diurnal cycle of ozone within the tributary valley mixed layer. Using a mass budget method similar to that used by Lenschowet al (1981), the relative importance of advection, photo-chemical production, and entrainment of ozone into the boundary layer of the Pitt River valley can be determined. Equation 2 describes the change in concentration of an atmospheric constituent, assuming horizontal homogeneity for the flux terms and negligible vertical velocity. In the Pitt River valley lateral ozone gradients, (y-axis), can be assumed to be near zero when the change in ozone within the surface layer of the valley walls is ignored. In a convectively dominant 22 mixed layer under clear skies, the mean vertical velocity is approximately zero (Stull, 1988). dQ -dQ dQ'w' dt dx dz * s (2) (/) (//) (iii) (iv) Here, Q is an atmospheric constituent, u is the mean wind aligned along the x axis, w' the fluctuation of vertical velocity, and Qs is the internal sources and sinks of the constituent Q. The rate of change of the constituent (term /) is balanced by horizontal advection (term ii) vertical flux divergence (term iii) and production or loss of Q (term iv) within the volume (Lenschow, 1981). For ozone, Qs is the net photo-chemical production or chemical destruction of ozone. The rate of destruction or production of O 3 is dependent on the ratio of NO2/NO (Guicherit & van Dop,1970). An air mass can be assumed to be "chemically aged," when all chemical scavengers of ozone have been transformed (i.e. all NOx has been used for photo-chemical production of ozone and no NO left for destruction). With a chemically aged air mass and little solar radiation term iv can be neglected. The vertical flux divergence can then be calculated without any turbulence measurements, knowing only mean ozone gradients and wind speeds. An alternative to the first order approximation of the vertical flux divergence at the top of the mixed layer, described above, is to utilise a parameterization of the flux divergence through the entire mixed layer. The vertical flux divergence over the entire layer is the flux of ozone at the top of the layer less the flux of ozone to the surface divided by the depth of the layer (Equation 3). These quantities are represented schematically in Figure 5. 23 t Figure 5. The Mass budget of an atmospheric constituent Q, in a layer with top at height Zj and bottom at z 0 . Fluxes at the top and bottom are parameterized as shown, and described below. Equation 4 parameterizes the flux at the top of the mixed layer using a first order slab model (Lenschow et al, 1981; Stull, 1988), and the flux to the surface is parameterized using a surface resistance (R s) to ozone uptake and ozone concentration at the bottom of the layer (Galbally and Roy, 1981). The entrainment velocity we is given in equation 5 and is simply the growth rate of the boundary layer in the absence of any subsidence (Stull, 1976). dQW ( 0 3 V ) ( . - ( 0 3 V ) c dz Zj -z (3) (GYO,. -(GYw')0 =-weAO, - i ? ; ' a (4) dz{ 24 The parameterization of entrainment velocity is only valid for a convectively dominant mixed layer and should not be used in cases where mechanical production of turbulence is dominant. This parameterization should determine the correct sign of the flux divergence if not the absolute magnitude. Once the direction of the vertical flux divergence at the top of the layer is known, it then can be determined whether the growing mixed layer is entraining ozone rich air from above. 2.7 SUMMARY In this chapter, a measurement strategy was described that permits a detailed examination of the three dimensional distribution of ozone in the L F V and its surrounding environment. Both the airborne and passive sampling strategies described are novel and represent a cheap and efficient means of extending spatial monitoring in the region. The airborne data also permit the application of a simple mass budget approach to elucidate processes contributing to the observed distribution of ozone. 25 CHAPTER 3: A NOVEL AIRBORNE SURVEY METHOD 3.1 I N T R O D U C T I O N Accurate representation of the distribution of mean ozone concentrations downwind of a major urban centre is difficult to achieve without an extensive surface monitoring network,, remote sensing or aircraft measurements. In the absence of a dense surface network, aircraft permit direct measurements of ozone over a wide area. In contrast, remote sensing offers only an indirect measurement of ozone through correlation with other atmospheric constituents, or an estimation of the mean concentration in an atmospheric column (Stephens, 1984). However, along with other scientific considerations, prohibitive costs have traditionally constrained aircraft use for such applications. Furthermore, any modification to an aircraft must be inspected and approved before it is allowed to fly. Consequently, an instrument package that can be kept on board with sensors that require no exterior braces but simply hang out the side of the cabin, is a very desirable attribute in any airborne measurement system. One focus of this study was the development and application of a cheap, reusable ozone measurement system capable of rapid deployment on a rental, light aircraft. This chapter describes the system designed to measure mean ozone concentration and temperature from a single engine Cessna, the required calibration and measurement corrections, the method of fixing a location to the data, the validation of the measurement technique and the problems involved in measuring characteristics of a turbulent field from a moving platform. 26 3.2 SYSTEM DESIGN The instrument package designed for use in a light aircraft, (a Cessna 175 or 155), used a Campbell Scientific 107 thermistor probe and a Gesellschaft fur Angewandte Systemtechnik (GFAS), OS-B-2 chemiluminescent ozone sonde designed for balloon borne deployment into the stratosphere (Figure 6). The sonde is a dry surface chemiluminescent type which employs a chemically reactive target that emits light when exposed to ozone. The light is sensed by a side window photo-multiplier tube to be converted to a voltage for output. The amount of light emitted by the chemical reaction is proportional to the absolute concentration of ozone in the air sample (Speuser et al, 1989). An air intake was designed and placed on the exterior of the plane, while the sonde itself was placed in the aircraft cabin attached to a Campbell Scientific CR10 datalogger, power supply and laptop computer. The thermistor was simply taped to the wing strut and attached to the data logger in the cabin. There was a second thermistor inside the aircraft cabin to measure the air temperature as it passed through the sonde. To measure mean ozone concentration accurately from the aircraft, a well mixed, uncontaminated sample must be supplied to the sonde inside the aircraft cabin. This meant an intake made of chemically unreactive material had to be designed. Figure 7 shows the teflon intake design which was fitted and taped outside the port side window directly in the propeller wash, ensuring a well mixed, immediate air sample. The port side must be used as contamination from the aircraft's exhaust is possible on the starboard side. A pressure bleed off for the intake was included to allow the chemiluminescent sensor to operate at a near constant flow rate. The sonde is equipped with a pump to ensure a constant flow rate of 60 1/min over the chemiluminescent target. 27 Exceeding this flow rate does not affect the sensitivity of the sonde as it is independent of flow rates above a critical value (Sahand et al, 1989). Wind tunnel tests show that at flight speeds, flow rates through the teflon tubing to the sonde are well above the critical threshold for correct operation of the ozone sonde (Zawar-Reza, personal communication). Light trap Photomultiplier Gherailurninescencc target* Figure 6. Configuration of the GFAS, OS-B-2 ozone sonde (from GFAS, Ozone Sonde OS-B-2 Operation Manual). 3.3 OZONE SONDE CALIBRATION AND SENSITIVITY The OS-B-2 sonde was calibrated against the Teco 49 ozone analyser, as was the potassium iodide sonde used for balloon profiles. In laboratory tests, the OS-B-2 sonde was accurate to a standard error of 4.5 ppb, over the range 0-120 ppb with a 95% 28 Pressure Bleed Off Figure 6. Air intake design for the aircraft ozone sonde. Top photograph shows the placement of the intake on the aircraft. Bottom diagram displays the intake design. 29 confidence level. Figure 8 shows the calibration curve of the ozone sonde for a typical operational flight. During pre-flight calibration, the ozone sonde was exposed to a range of 3 concentrations produced by the GFAS ozone generator supplied with the ozone sonde. A regression was then developed and applied to the data collected during the subsequent flight. OS-B-2 Ozone Sonde Calibration 13 August, 1994 -I 1 1 1 1 1 1 1 1 1 1 \ 1 0 10 20 30 40 50 60 70 80 90 100 110 120 Ozone [ppb] (from Teco 4A) Figure 8. Calibration curve and equation of the OS-B-2 ozone sonde. This allowed direct comparison between surface, balloon and aircraft measurements of ozone. The OS-B-2 sonde had a response time of less than 1 second while the Teco 49 analyser was much slower with a response time of about 30 seconds depending on the length of sampling tube attached to it. Calibration of the chemiluminescent ozone sonde was dependent on a constant supply of ozone from a generator supplied by the manufacturer. The ozone generator used ambient air to supply an ozonic sample to the sonde and therefore would deliver a 30 range of concentrations depending on the fluctuation of ozone concentrations that existed in the ambient air. Calibrations in the field during high ozone episodes and convective conditions would then have much higher standard errors due to the natural fluctuation of ozone concentration and the different response times of the sonde and Teco 49 ozone analyser. Therefore, an ozone scrubber on the inlet line of the ozone generator would produce more constant results when calibrating in the field and laboratory. The chemiluminescent target sensitivity is dependent on temperature, pressure, and large humidity changes. For temperature corrections, a thermistor was needed to measure the temperature of the air as it flowed through the sonde. Temperature measurements were only needed when the air temperature exceeded 30°C as the sonde was equipped with a heater which kept the target at a constant temperature. Figure 9 shows the temperature sensitivity of the chemical target. The sensitivity of the sonde to pressure and large humidity changes was of little consequence when flying at a constant altitude within the cloud free troposphere. However, a correction due to the difference in air pressure between calibration and measurement was required. Lower measurement pressures meant increased sensitivity of the sonde, and the calibration curve had to be adjusted using equation 6, which was supplied by the manufacturer. Fp=(\-0.63-\og(pmeasl pcal) (6) The measurement pressure, p m e a s , was needed to adjust the sonde's raw data and was calculated using the aircraft's altimeter (which works using pressure measurements and the hydrostatic equation). Extreme changes in humidity (as are found when moving 31 Temperature Dependence of OS-B-2 Ozone Sonde 12 y = 0.1871x + 7.0737 u. -30 -20 -10 0 10 20 Temperature f C ) Figure 9. Temperature sensitivity curve and equation for the OS-B-2 ozone sonde (from GFAS, OS-B-2 manual). from the troposphere to stratosphere) affect the sensitivity, but in this study the maximum change in humidity experienced was when flying in or out of the boundary layer. This was deemed to have a negligible effect. Equation 7 describes the corrections needed for large humidity changes, but was not used for the boundary layer measurements made in this study. where PH^0 is the partial pressure due to water vapour. This correction factor was also supplied by the manufacturer. Fh = 1 + 0.4(1 - exp(-0.238 • PHiQ)) + 0.01 \P, (7) 32 3.4 M E A S U R E M E N T S A N D D E T E R M I N A T I O N OF POSITION To determine position when sampling from the aircraft, a simple yet effective and inexpensive method was used. Prominent landmarks along the predetermined flight paths were chosen, and as the aircraft passed over these landmarks the corresponding time was recorded. Secondary flight paths and landmarks were also selected for use when air traffic control forced last minute changes. This provided a fix for ozone concentration at discrete points and a straight line at a constant altitude between the points was assumed to accurately describe the aircraft's path. A conservative estimate of error of 3s in marking the time, equates to a distance error of +150m (at 180 km/hr), for the horizontal location of the ozone measurement. This was deemed to be a reasonable error for the measurement of mean concentrations near the top of the boundary layer. Altitude was kept constant for each flight leg and an error of +50m in the vertical was estimated for the location of the ozone measurements. 3.5 F L I G H T L E N G T H S A N D S T A T I O N A R I T Y OF T H E S A M P L E Measurements from a moving platform, not following the mean flow, are in neither the Lagrangian or Eulerian frame of reference, unless the movement is so fast as to make the data collection instantaneous. Taylor's hypothesis of "frozen turbulence" must then be evoked. Taylor's hypothesis states that experimental one-dimensional spectra can be obtained by moving a probe through a turbulent field so rapidly that the change in the velocity field, u(t), during the measurement time is negligible. If the traversing speed, U , of the probe is large enough then the time, t, of measurement can be substituted by x/U. Therefore, u(t) can be specified as u(x/U) i f u/U « 1 , (Tennekes and 33 Lumley,1972). This hypothesis is often used when measuring turbulent fields moving over a fixed point on the surface, but the assumption is actually more closely approached when flying at an order of magnitude faster than the mean wind speed. Thus, when measuring horizontal gradients, the measurement time x/U, must be less than the time required for the gradients to change due to advection. Length scales should also be considered when determining gradients and turbulent variability from a data set. Length scales needed for determining turbulent variability are proportional to boundary layer heights (Lenschow and Stankov, 1986). Measurement lengths of 10 to 100 times the boundary layer height are required to measure variances to a 10% accuracy, and this length scale can then be used along with the true airspeed to determine the time scale needed. Fully developed boundary layer heights in the L F V and Pitt valley vary spatially and temporally. However, 800-900 m is a typical height of the convectively dominant, fully developed boundary layer found near the central L F V (Steyn and Oke, 1982). Mixed layer heights are generally less within the Pitt River valley. A 90 km flight leg would then provide a long enough sample to determine the turbulent variability and a more than desired accuracy for the measurement of mean ozone gradients. At 180 km/hr, a time scale of 30 minutes is a reasonable upper limit on the length of flights. The 90 km length scale for determining gradients was not always attainable in this study. The urban plume of Greater Vancouver may not reach 90 km downwind on some days and the extreme relief found in the L F V often restricted flight lengths. For the measurement of only mean gradients and not the turbulent variability, the length scale can be significantly reduced and a standard error cited as proof of accuracy. Using the standard advective time scale of T - L / U, where L is the length of the measurement path 34 and U is mean wind speed, a lower limit on the time required for gradients to change due to advection, and hence the time needed between measurements is estimated. Thus, for a 25 km measurement length and a mean wind speed of 5 m/s in the Pitt valley, successive flights must be made at least 1.38 hours apart to capture changes in gradients due to advection. At a measurement speed of 180 km/hr the time taken to measure a 25 km gradient is only 8 minutes and 20 seconds. Therefore if the flight length is 30 minutes or less and flights are made at 2 hour intervals then the time scales important in measuring mean concentration and gradients are easily met. 3.6 V A L I D A T I O N OF T H E A I R C R A F T M E A S U R E M E N T S The first successful aircraft measurements, using the configuration described above, were made 22 July, 1994 when ozone concentrations in the L F V were high. The OS-B-2 sonde was operational for over two hours on this day and over seven hours on subsequent measurement days and showed no drift in its sensitivity. Exposure to an ozone rich air sample (either from ambient air or from an ozone generator) kept the sensitivity of the sonde stable during flights. When measurements were to be made throughout the day, and when time allowed, the sonde was removed from the aircraft between flights and placed on an ozone generator to preserve its sensitivity. One method of validating aircraft ozone measurements is to compare the airborne measurements with surface monitor measurements. In the absence of any other surface measurements, five minute surface averages of ozone were compared to aircraft measurements. Assuming that ozone is well mixed in the boundary layer, a five minute 35 average, surface ozone measurement should be comparable to a 20 second average of ozone measured by aircraft directly over the surface monitor. This is deemed to be a reasonable assumption based on tethersonde profiles of ozone from Pitt Meadows (McKendry et al, 1997). Results from 11 flights on 5 different days are shown in Figure 10. The aircraft consistently measured slightly higher values of ozone concentration than the surface monitors, except for 2 August, 1994 when stations 16, 15, and 27 recorded higher concentrations than at 450m A G L . These three exceptions occurred when ozone concentrations were low and could be considered as background levels. Surface concentrations should be slightly lower due to the removal of ozone through dry deposition and are not equivalent to concentrations throughout the depth of the boundary layer (Van Dop and Guicherit,1980). This difference between surface and aircraft measurements of ozone concentration in the boundary layer was also found by Van Valin et al (1994) when testing trace gas instruments aboard a King Air research aircraft. The absolute percentage differences in concentration between the surface and 450m A G L when normalised by the concentration at 450m A G L were consistently less than 15% except in four instances. These four instances were: 13 August, 1994 at -0600,-0800, and -1000 PST, the aircraft data compared with surface station 16; and 2 August, 1994, aircraft data compared with surface station 15. The three flights of the morning of 13, August were made before convective vertical mixing was effective enough to raise surface concentrations to levels comparable to the mixed layer while the 36 Comparison of Aircraft and Surface Ozone Data 90 s o N O 80 1 a a 70 1 2 60 y 50 1 1 40 t 30 20 1 10 1 ^ 22/0.7-JS' • 22/07-'' -16+ 22/07-15-' % 22/07-16 13/08-16 -0600PST + 27/07'f6 ^ 27/.07-15 • 27/07-12 13/08-16 • 27/07-28 13/08-16 -1000 PST^ 02/0&-12' -0800 P^T 02^-i»4<3ti?6i.16 02/0^45 # 02/08-16 02/08-27 • 02/08-15 + + + + 10 20 30 40 50 60 Ozone Measured at Surface (ppb) 70 80 90 Figure 10. Validation of the aircraft measurement technique. Comparison of 5 minute average surface measurements to 20 second average aircraft measurements of ozone concentration is shown here. Each point is identified by a numerical code which states the day and month of flight and surface station used for comparison. low concentrations of 2 August meant small differences in concentration resulted in large relative differences. A near constant relation between ozone measured at the surface and in the mixed layer was not expected in this study particularly as measurements were made over a variety of surface types and meteorological conditions. It could be concluded that the ozone concentrations measured by aircraft are the result of the combined surface effects of the L F V . The relation between surface and mean mixed layer concentrations in the L F V observed was interesting and perhaps deserves further investigation. 3 7 Further evidence for the validity of the measurement technique comes from the several occasions when the aircraft flew directly past the tethered balloon. Figure 11 compares measurements made by the potassium iodide sonde on the balloon and airborne measurements from the OS-B-2 sonde. Data points displayed are 20 second averages calculated when the plane flew within 100m of the kytoon. Except for one data point, all measurements made via both methods are within 3.5 ppb and provide confidence in the reliability of the aircraft's instrument package. The one extreme data point in Figure 11, could have been the result of contaminated iodide solution in the tethersonde. Aircraft Ozone Sonde Verification 0 10 20 30 40 50 60 70 80 Tethered Balloon Measurements of Ozone (ppb) Figure 11. Comparison of ozone measurements made using a tethered balloon and a potassium iodide solution based ozone sonde to measurements using a light aircraft and chemiluminescent sonde. 38 C H A P T E R 4: T H E S P A T I A L DISTRIBUTION OF O Z O N E A C R O S S T H E L O W E R F R A S E R V A L L E Y A N D T H E PITT R I V E R V A L L E Y 4.1 I N T R O D U C T I O N The spatial distribution of ozone in the L F V was resolved using data from G V R D surface ozone monitors, kytoon and aircraft mounted ozone sondes, and passive ozone samplers deployed in the Pitt River valley and atop Mount Seymour,. This chapter summarises the results of the measurement program of the summer of 1994 and identifies the basic characteristics of ozone distribution over the L F V . Measurements were made under a variety of synoptic conditions when ozone concentrations were expected to be moderate to high. 4.2 STUDY P E R I O D The summer of 1994 was representative of average air quality conditions for the G V R D monitoring area. There were three days in which a station episode occurred, resulting in eight stations reporting average hourly concentrations above 82 ppb (GVRD, 1994). A station episode occurs when one or more monitoring stations exceed the Provincial Objective Level B, which is defined as a one hour average ozone concentration in excess of 82 ppb (after CSC, 1985). These three dates were 21, 22, and 23 July 1994. Air quality was poor, with respect to ozone, on many other days as can be seen from Figure 12, a plot of maximum ozone concentrations measured near the mouth of the Pitt River valley, at station 16. The Provincial Objective Level A , maximum desirable ozone concentration was exceeded thirteen times at station 16, while the Level B, maximum acceptable, was exceeded twice. The measurement days for the aircraft survey of ozone were planned to coincide with elevated surface ozone levels, so air 39 quality forecasting played an integral part in conducting a successful measurement program. Maximum Hourly Average Ozone Station 16 July & August 1994 Federal Objective Level B Maximum Acceptable Federal Objective Level A Maximum Desirable 15/07/94 29/07/94 12/08/94 26/08/94 Date Figure 12. Maximum hourly ozone concentration measured at the G V R D surface monitor at Pitt Meadows Airport for each day between 1 July 1993 and 31 August 1993. After a successful first deployment of the airborne GFAS, OS-B-2 chemiluminescent ozone sonde on 22 July 1994, further flights were made when elevated ozone concentrations were expected. Using the atmospheric characteristics of an ozone episode set out by Steyn et al (1990) and Taylor (1991) and the synoptic types identified by McKendry (1994), ozone concentrations were expected to be high on 27 July and 13 August, 1994. Ozone was also expected to be above background concentrations on 2 and 30 August, 1994. These five dates in the summer of 1994 represent the entire measurement program for the aircraft survey of ozone in the L F V . Due to logistical problems, tethered balloon flights did not coincide with all aircraft measurements. 100 90 + 80 ^ 70 I 60 "ST 50 I 40 O 30 20 10 0 01/07/94 40 Permanent G V R D surface stations provided ozone data for comparison and analysis with aircraft observations on all measurement days. The following sections summarise the ozone data collected on the study dates. For each study day, available ozone data from aircraft observations, G V R D surface monitors, passive samplers and tethered balloons were used to construct ozone isopleth maps at 450 m above ground level, over the L F V . Each map represents a "snapshot" of ozone concentrations at 450 m. For the assumption of stationarity of the measured aircraft data to hold, the isopleths can only represent the distribution of ozone during the flight time. A l l measurement flight lengths were performed in less than 30 minutes as outlined in Section 3.5, thus Taylor's hypothesis is used for both the stationary surface monitors and the moving aircraft measurement platform. G V R D surface monitor data was used to estimate the ozone concentration at 450 m if aircraft data or tethered balloon data was not available. Ozone concentrations at 450 m were estimated using surface data by employing a similar relation as that described by Figure 10. For each measurement day, the average difference between averaged aircraft measured data and averaged surface data was used to account for the vertical separation between the surface and 450 m above ground level. Passive sampler data was not used for the estimation of absolute concentrations only for determination and conformation of relative horizontal ozone gradients. 4.3 J U L Y 22,1994 - High Ozone Concentrations The first day of the measurement program followed a two day persistent ozone "episode" (i.e. the occurrence of three or more consecutive 1-hourly average ozone 41 concentrations greater than 82 ppb measured at a surface station in the L F V on at least two of three consecutive calendar days (after CSC, 1985)). Boundary layer ozone concentrations in the L F V on 21 July, 1994, were in excess of 90 ppb and well mixed below 600m. The sharp vertical gradient in ozone expected with a strong capping layer is easily identified in Figure 13. The top of the well mixed layer of the previous day, 21 July 1994, is clearly evident in the profile of ozone measured to a height of approximately 1000 m at the mouth of the Pitt River Valley. Synoptic conditions on 22 July 1994, were still favourable to the development of elevated ozone concentrations as an upper level ridge (Figure 14a) was positioned over western B.C. producing strong subsidence and suppressed mixed layer heights (McKendry, 1994). The upper level ridge of 22 July, 1994 was accompanied by a thermal trough in south-central B.C., providing weak synoptic flow and a convectively dominated boundary layer (Figure 14b). Skies were clear over the entire area, while temperatures at Vancouver International Airport and Abbotsford exceeded 27°C and 32°C respectively, excellent conditions for the photo-chemical production of ozone. The weak synoptic flow and high temperatures produced well developed thermal flows allowing the sea breeze and valley winds to concentrate pollutants to the east and north-east of the metropolitan area. Winds at Vancouver International Airport were WSW at 2.8-4.2 m s"1 by 1300 PST and southerly at 0-1.4 m s"1 at Pitt Meadows Airport. Surface stations 28 and 12 directly east of Vancouver had 1 hour average ozone concentrations that exceeded 80 ppb and stations 16 and 9, to the north-east, had 1 hour average concentrations in excess of 62 ppb. 42 Only one flight was made on July 22, following a north-south route across the easterly moving urban plume, then up into the Pitt River valley flying with the thermally forced southerly flow (Figure 15). The ozone distribution across the entire study area derived from aircraft and surface station data is shown in Figure 15a, while Figure 15b is the actual trace of ozone measured by the aircraft. Ozone concentrations measured by the 1000 -r 900 -55 60 65 70 75 80 85 90 95 OZONE (PPB) Figure 13. 21 July 1994 1605 PST, ozone profile from Harris Road, Pitt Meadows near the mouth of the Pitt River valley. 43 Figure 14. (a)500 hPa geopotential height chart for 22 July 1994 at 0500 PST, (b) Surface pressure chart for 22 July 1994 at 1100 PST. 44 Fraser and Pitt River Valley Ozone Concentrations Measured at 450m A G L 22 July, 1994 Boundary Bay: : East Surrey Pitt Meadows . S: Shore Pitt Lake N . Shore Pitt Lake Position Figure 15. (a)Ozone isopleths for 22 July 1994 at -1315 PST determined from aircraft, surface monitoring and passive sampler data, dotted path represents the flight path for the ozone trace shown in (b). 45 aircraft were near constant across the central part of the L F V and then rose dramatically in the Pitt River valley up to a maximum of 110 ppb at the northern end of the flight leg, 5 km past the north end of Pitt Lake. It is possible that concentrations were higher further north in the Pitt River valley. The map of ozone isopleths illustrates the sharp gradient of ozone expected on the downwind side of an urban centre, and an additional gradient within the Pitt River valley, while concentrations across much of the L F V are much less variable. The trace of ozone measured by aircraft confirms the gradient and shows the extremely high levels of ozone found in the valley. 4.4 J U L Y 27,1994 - Moderate Ozone Concentrations An upper level ridge had been expected to develop on 27 July, 1994, but was pushed east over Alberta and eastern B.C. as a deep low off the coast of B.C. caused a south-westerly upper level flow over the L F V (Figure 16a). Consequently, subsidence was weak, although the surface pressure gradients were slack and a weak thermal trough was evident in central and eastern B.C. (Figure 16b). July 27 did not exhibit the synoptic patterns associated with a typical ozone episode, but temperatures were high in the L F V and sea breeze and valley winds were evident during the day. The maximum temperatures at Vancouver International Airport and Abbotsford were 26°C and 29°C respectively. Winds were south-west at 4.2-6.6 m s"1 by 1300 PST at Vancouver International Airport forcing pollutants from Vancouver to the north-east as opposed to the east. The surface ozone monitor 16 at Pitt Meadows, north-east of Greater Vancouver, had a maximum hourly average of 65 ppb, while the maxima reached at monitors 12 and 28 in the eastern L F V were 63 ppb and 57 ppb respectively. 46 Figure 16. (a)500 hPa geopotential height chart for 27 July 1994 at 0500 PST, (b) Surface pressure chart for 22 July 1994 at 1100 PST. 47 Fraser Valley Ozone Concentrations Measured at 450rn A G L 27 July, 1994 70 P i t t Meadows Hancy Whoanock p . . Matsqui Island Chilliwack Figure 17. (a)Ozone isopleths for 27 July 1994 at -1315 PST, determined from aircraft, surface monitoring and passive sampler data, dotted path represents the flight path for the ozone trace shown in (b). 48 The ozone concentrations across the entire study area are shown in Figure 17a. The strong ozone gradient at the eastern edge of Greater Vancouver and the high concentrations to the north-east of the city into the Pitt River valley are consistent with the 22 July distribution, and ozone isopleths for episodes described by Steyn et al (1990) and Evans et al (1992). Over much of the broad, flat central part of the L F V , ozone concentrations showed little variation apart from a slight maximum over Chilliwack where the L F V narrows and temperatures were slightly higher. A trace of ozone measured by aircraft along the path shown in Figure 17a shows a weak horizontal gradient in ozone concentrations within the L F V (Figure 17b) and is consistent with the surface observations. 4.5 A U G U S T 2,1994 - Moderate Ozone Concentrations A n upper level low still dominated the synoptic pattern 6 days after the previous flight producing a south-west flow over the study area (Figure 18a). The deep low began to fill and the cold front of July 27 passed through the region on the 31 flushing the valley of pollutants. There was negligible subsidence from the slight ridge over the eastern part of the province, and the surface pressure gradients were weak (Figure 18b). Skies were clear and temperatures were reasonably high (Vancouver Airport reached 25°C and Abbotsford airport exceeded 30°C), but ozone concentrations were low at all but one monitoring station. The winds at Vancouver International Airport were east at 2.2 ms"1 at 0800 PST then WSW at 4.2 ms"1 at 1400 PST. At Pitt Meadows Airport and Abbotsford Airport, winds were calm and variable at 0800 PST and south at 1.1-2.2 ms"1 at 1400 PST. At the Pitt Meadows ozone monitor, station 16, the maximum hourly 49 a) Figure 18. (a)500 hPa geopotential height chart for 2 August 1994 at 0500 PST, (b) Surface pressure chart for 2 August 1994 at 1100 PST. 50 Fraser Valley OzoneConcentrations Measured at 450m A(JL 2 August, 1994 50 A JB : Mrigley Wkmnock Matsqui Island j? Boundary Bay Cloverdale Airstrip Position Chilliwack Figure 19. (ajOzone isopleths for 2 August 1994 at -1215 PST, determined from aircraft, ' surface monitoring, tethered balloon and passive sampler data, dotted path represents the flight path for the ozone trace shown in (b). 51 average ozone concentration reached only 43 ppb and only 48 ppb at station 28, Abbotsford. Chilliwack, station 12, recorded the highest levels of ozone (65 ppb) very late in the day (1700 PST), but this was the only surface monitor in the network to exceed even 55 ppb. The map of ozone isopleths for August 2 (Figure 19a) and the trace of ozone measured by aircraft (Figure 19b) again show little variation in concentration over the central L F V , but unlike previous days, the strong ozone gradient on the eastern edge of Greater Vancouver is not evident. Instead, concentrations appear highest along the northern edge of the L F V and into the Pitt River valley. This likely reflects the strong southerly component to the flow and lack of sea breeze development on this day. Another flight north into the Pitt River valley made later in the day coincided with tethered balloon flights to provide data for proof of the measurement accuracy of the sonde. This was displayed in an earlier chapter (Figure 11). Ozone concentrations were about 5 ppb higher in the Pitt River valley compared to the eastern L F V . The trace of ozone presented in Figure 21 was measured along an east-west path outlined in Figure 20. In this case, even under synoptic conditions that produced low to moderate ozone concentrations, up valley gradients of ozone in both the Pitt River valley and the L F V were still evident. This suggests that background ozone concentrations for the L F V are lower than 40-45 ppb and photochemistry associated with anthropogenic sources was active on 2 August. 52 4.6 A U G U S T 13,1994 - Moderate Ozone Concentrations Four flights into the Pitt valley were made on this day in anticipation of reasonably high ozone concentrations. A series of flights were needed to calculate the ozone mass budget of the valley for the first few hours after sunrise. A detailed description of measurements and analysis of the early morning data is given in the following chapter. A compilation of surface, balloon and aircraft data for early afternoon is presented here. The upper level geopotential height chart of August 13 reveals a weak ridge centred over eastern B.C. (Figure 20a). Surface weather analysis shows very weak synoptic forcing over the study area and a slight thermal trough just south-east of the L F V (Figure 20b). An ozone episode was not expected on this day, but the warm temperatures, 26°C at Vancouver airport and 28°C at Abbotsford, and clear skies meant elevated ozone concentrations were possible. Winds were southerly at 2.8-5.6 m s"1 at Vancouver Airport and southerly at 1.4-2.8 m s"1 at Pitt Meadows, while winds were west at 2.8-4.2 m s"1 at Abbotsford at 1300 PST. The thermal forcing of the slopes of the Pitt River valley and the northern edge of the Fraser valley was strong enough to oppose the synoptic forcing, while the small temperature difference between coastal Vancouver and inland Abbotsford meant the sea breeze had less effect (Steyn and Faulkner, 1986). This resulted in higher hourly ozone concentrations to the north-east of Greater Vancouver. Stations 16 and 9 reached 61 ppb and 58 ppb respectively while stations to the east reached 56 ppb at station 28 and a high of 62 ppb at station 12. As on 2 August, Chilliwack (station 12) recorded the highest ozone concentrations in the network and late in the afternoon. 53 Figure 20. (a)500 hPa geopotential height chart for 13 August 1994 at 0500 PST, (b) Surface pressure chart for 13 August 1994 at 1100 PST. 54 70S 5 0 + a. • 8 30 20.;* 10 N —• jl318:Q*J° Pitt River Valley Ozone Concentrations Measured at 450m A G L 13 August, 1994 A 1 3 2 2 : 4 7 ;::1325:56 B S. Shore Pitt Lake • C Lake Elbow Position N - Shore Pitt Lake Figure 21. (a)Ozone isopleths for 13 August 1994 at -1330 PST, determined from aircraft, surface monitoring, tethered balloon and passive sampler data, dotted path represents the flight path for the ozone trace shown in (b). 55 accompanying the map is a trace of ozone measured by aircraft up the Pitt River valley along the route marked in Figure 21a. On this day at this time, the up valley gradient of ozone was slightly negative. That is, ozone concentrations decreased slightly with distance up the valley. Concentrations exceeded 55 ppb at the north end of Pitt Lake. 4.7 A U G U S T 30,1994 - Low Ozone Concentrations Towards the end of the northern, mid-latitude summer, the potential for elevated ozone concentrations is reduced. Flights made this day, as well as surface data, revealed ozone levels only slightly higher than what could be considered background levels. The 500 hPa height chart (Figure 22a) shows a reasonably strong ridge aligned in a south-west to north-east position over B.C.. However, there is no thermal trough evident in the surface chart, and a weak frontal system is just over the B.C. coast (Figure 22b). Temperatures were lower than any other measurement day, 26°C at Abbotsford and only 22°C at Vancouver airport. By 1300 PST winds were light at 1.4-2.8 m s"1 and from the west-south-west across the entire study area. Flights were made late in the afternoon to allow the photo-chemical production of ozone to reach its full potential. Although concentrations were low in the L F V , the compilation of aircraft, balloon and surface data into Figure 23a revealed a now familiar pattern of ozone distribution. Concentrations were highest in the Pitt valley and the extreme eastern part of the L F V and the spatial variation in mean concentration across the central L F V was small. The maximum gradient in ozone concentration was again found at the eastern edge of the city as the winds on most clear sunny days in this area tend to be westerly to south-westerly. The trace of ozone measured along a flight path into the Pitt River valley (Figure 23b) 56 - 1024 1020 Figure 22. (a)500 hPa geopotential height chart for 30 August 1994 at 0500 PST, (b) Surface pressure chart for 30 August 1994 at 1100 PST. 57 Pitt River Valley ©zone Concentrations Measured at 450m A G L :30^ugusi!,*1994 45 , • - — Pitt Meadows S. Shore Pitt Lake Position . Lake Elbow. N,Shore Pitt Lake Figure 23. (a)Ozone isopleths for 30 August 1994 at -1445 PST, determined from aircraft, surface monitoring, tethered balloon and passive sampler data, dotted path represents the flight path for the ozone trace shown in (b). 58 shows an up valley gradient of ozone, although concentrations reached only 43 ppb. 4.8 PASSIVE S A M P L E R S The passive ozone samplers were deployed in the Pitt River valley and atop Mount Seymour (shown in Figure 1) from 3 August to 5 September. Each sampler was left for a period of not more than six days and not less than 24 hours. The average ozone concentrations for the period of deployment at the four sites are plotted in Figure 24. 35 i 30 ^ 25 a 3 20 § 15 N O 10 5 0 ao 3 oo 3 Mount Seymour North Shore Pitt Lake Goose Island Pitt Meadows oo ON 00 3 oo 3 oo 3 00 3 < 1 00 3 < 1 ON a. 00 Figure 24. Average ozone concentration measured by passive sampler, integrated over the time of deployment, at four sites in the study area, refer to Figure 1 for locations. Average concentrations were quite low, as expected, as the samplers were measuring ozone within the surface layer. Over night surface ozone concentrations usually fall to zero which also significantly reduces the total amount of ozone absorbed by the nitrate salt. Contamination of the samplers by condensation may also have happened which would have hindered the reaction described by equation 1. The highest 59 measured average ozone concentrations by passive sampler were at the Mount Seymour site, some 1500 m above the L F V . This is consistent with other studies at remote elevated sites which show small diurnal variations in ozone concentrations due to the effects of slope flows bringing a steady supply of polluted air from elevated residual layers at night and the valley bottoms during daytime (Puxbaum et al, 1991). This results in high integrated concentrations over the period of deployment. Figure 24 also indicates a positive horizontal gradient of ozone into the Pitt River valley. Integrated concentrations systematically increase with distance up the valley on all days including 13 August 1994, indicating that the ozone gradients measured by the aircraft in the mixed layer also exist in the surface layer. 4.9 O Z O N E GRADIENTS IN T H E PITT RIVER AND F R A S E R V A L L E Y S - A C O M P A R I S O N The measurements presented above were made during varying synoptic weather conditions. Only one measurement day, July 22 was during a synoptic situation expected to produce poor air quality and high ozone concentrations in the L F V (McKendry, 1994). Yet with these varying conditions there were some consistent patterns in the spatial distribution of ozone. Of these patterns, already outlined above, one notable and expected pattern in ozone concentrations was the increase of ozone with distance, up the Fraser and Pitt River valleys. In all cases, ozone increased with distance downwind from Greater Vancouver. This is expected as ozone is a secondary pollutant requiring sufficient time and the correct ratio of precursor species for formation. However, the 60 ozone gradient in both these valleys was not equal. Ozone gradients in the tributary Pitt River valley were consistently greater than or equal to those in the Fraser valley. Figure 25 compares the ozone gradients calculated for both valleys along an equal distance (30 km), measured from a point just east of the position of the strongest ozone gradients on the eastern edge of the city (see Figures 15, 17, 19, 21, and 23 for location of gradients). Valley gradients were calculated from the same point northward into the Pitt valley and eastward into the Fraser valley. Ozone gradients were determined from the maps of the previous section, which are compilations of aircraft, surface monitor, and tethered balloon data. The choice of position for the calculation of gradients assumes a general wind pattern of westerly to south-westerly across Greater Vancouver diverging to separate southerly and westerly flows near Pitt Meadows (Miao, 1993). Calculation of gradients from a point within the city would not change the relative difference in the Pitt and Fraser valley gradients assuming trajectories that would carry pollutants east across the city then diverging northward and eastward. The daytime mixed layer ozone gradient in the Pitt River valley is consistently higher or equivalent to the gradient in the Fraser valley suggesting that this tributary valley may be an important net sink for ozone and its precursors emanating from Vancouver. In the following chapter, the processes contributing to this pattern are explored in detail. 61 Comparison of Pitt River Valley and the Lower Fraser Valley Ozone Gradients s .22 •3 s-O s o , , o £ « '— 1 > 1-> n nnn£ 0.0005 0.0004 0.0003 0.0002 0.0001 01)01 22 July 1994 30 August 1994 27 July 1994 _ 0 . t l o n . • 2 August 1994 -0.0001 4-•13 August 1994 0.00015 0.0002 00025 Lower Fraser Valley Ozone Gradient [ppb/m] Figure 25. Ozone gradients in the Pitt River and Lower Fraser Valleys. The solid line represents a 1:1 ratio of valley gradients. Note the Pitt River valley gradients are consistently higher than the Lower Fraser Valley ozone gradients. 62 CHAPTER 5: PROCESSES CONTRIBUTING TO OBSERVED SPATIAL DISTRIBUTION OF OZONE 5.1 I N T R O D U C T I O N The aircraft survey of ozone across the L F V and its tributary, the Pitt River valley, proved to be a success. A new measurement technique was developed, and the spatial distribution of ozone in the L F V and Pitt River valley was determined for a number of days under a variety of weather conditions. Some recurrent features of the distribution were found even under conditions that produced only slightly elevated ozone levels. Just east of the city, horizontal gradients of ozone were steep and ozone concentrations across the broad, flat L F V were relatively constant. Of more interest were the consistent horizontal gradients and high levels of ozone found in the tributary valley. In the Pitt River valley, ozone concentrations were consistently higher than or equivalent to those measured in regions of the L F V at comparable distances downwind of Greater Vancouver. This supports the hypothesis that the Pitt River valley plays an important role in the ventilation of the L F V and may be a net sink for ozone emanating from Greater Vancouver. In this chapter, the processes involved in the change of ozone concentration in the Pitt River valley will be investigated using a mass budget technique and then a conceptual model presented to try and explain the results. Four flights into the Pitt River valley were made on 13 August, 1994, and this day will be used as a case study for the calculation of the ozone mass budget. 63 5.2 P O T E N T I A L P R O C E S S E S C O N T R I B U T I N G T O E L E V A T E D O Z O N E C O N C E N T R A T I O N S I N T H E P I T T R I V E R V A L L E Y The high levels of ozone found in the Pitt River valley cannot be explained by photo-chemical production alone. Although this process accounts for almost all the ozone found downwind of Greater Vancouver, it does not explain the difference in concentrations found in the Pitt River valley and the L F V . Trajectory distance and time downwind from the major precursor source region is equivalent for both valleys while air temperatures and hours of direct sunlight are generally less in the Pitt River valley. Cooler temperatures and less direct radiation should limit the photo-chemical production of ozone in the Pitt River valley. Processes that could explain high ozone concentrations in the tributary valley are: • advection; the daytime anabatic flow in the valley could be lighter than the sea breeze dominated flow of the L F V resulting in net convergence in the boundary layer of the Pitt River valley. • boundary layer entrainment; less turbulent kinetic energy in the Pitt River valley would result in less entrainment of the clean, free atmosphere above the mixed layer. • reduced surface deposition of ozone; fresh water can have up to ten times the resistance to ozone uptake as soil or grassland. Polluted air parcels moving northward into the Pitt River valley must pass over a 25 km long, fresh water lake covering the southern end of the valley. • multi-day ozone storage; ozone from the previous day can be stored in layers above the clean, nocturnal katabatic flow and mixed into the growing mixed layer of the next morning. 64 Each of these process will be discussed in turn, with special attention given to the mechanism of multi-day storage of ozone through a calculation of the early morning ozone budget in the Pitt River Valley. 5.2.1 Advection Wind data collected by tethersonde in the Pitt River Valley indicates that convergence in the boundary layer of the Pitt River Valley is not evident. The tethersonde located on Little Goose Island in Pitt Lake consistently measured higher wind speeds than at Pitt Meadows airport near the mouth of the valley, making the mouth of the valley an area of net divergence. Advection is of greater magnitude in the Pitt River Valley than in the L F V , and this long glacial valley should disperse pollutants well, as air masses advected into the valley can move over 200 km up valley passing numerous smaller tributaries into which the air can disperse. Dry deposition of ozone onto the heavily forested slopes of the valleys is a major removal process for mixed layer ozone and should clean the air in the Pitt River valley (Banta et al, 1996). The high wind speeds coupled with the rough surface of the coniferous forest that covers most of the Pitt River Valley, generates turbulence and air is mixed downward to the surface. When the highly reactive ozone comes into contact with the surface, a chemical reaction takes place and ozone is destroyed (Galbally and Roy, 1980). Dry deposition of ozone onto the slopes of the Pitt River valley is most effective just as night falls when drainage flows cause ozone rich air to move downslope. The nocturnal katabatic winds in the valley move ozone rich air from the lower part of the boundary layer down the slopes of the valley and ozone is deposited as the air is mixed along the surface. This results in a relatively ozone free nocturnal jet flowing down and out of the valley. During Pacific '93, Banta et al (1996) 65 found that the nocturnal katabatic flows emanating from the Pitt River valley were essentially free of ozone and this was attributed to dry deposition along the slopes of the valley. An area of divergence near the mouth of the Pitt River Valley and strong advection within the valley do not support the hypothesis of higher ozone concentrations in the Pitt River Valley than in the L F V . 5.2.2 Boundary Layer Entrainment The mixed layer height of the Pitt River Valley has been observed to be less than the mixed layer height of the L F V , implying that there is less turbulent kinetic energy in the lower atmosphere of the Pitt River valley (McKendry et al, 1996). If the entrainment velocity above the tributary valley is less than the entrainment velocity above the L F V , and the free atmosphere above the mixed layer is relatively free of ozone then entrainment could be a factor in the observed differences in ozone concentration. Using the definition of entrainment velocity described by equation 5 (Stull, 1976), the entrainment velocity is equal to the growth rate of the mixed layer. From observations during ozone flights, the fully developed mixed layer of the L F V is noticeably thicker than the Pitt River valley suggesting a faster growth rate and hence a larger entrainment velocity. The entrainment velocity of the Pitt River valley for 13 August 1994 was calculated using tethersonde data and was estimated at 0.016 m s*1. In a study modelling the mixed layer over suburban Vancouver, Steyn's (1990) model results suggest an entrainment velocity of 0.026 m s"1 during anti-cyclonic conditions. Thus the difference in entrainment velocities in the Pitt River and L F V could help to explain the observed 66 differences in ozone concentrations, assuming that there is less ozone in the free atmosphere above the mixed layer. 5.2.3 Surface Deposition Ozone concentration in the mixed layer is also dependent on surface processes. Reduced surface deposition over water can account for high near surface ozone concentrations and could result in a negative vertical flux divergence of ozone within the Pitt River Valley. There is a 25 km long fresh water lake that covers the entire valley floor near the southern end of the Pitt River valley and the surface resistance of water to the uptake of ozone can be up to ten times greater than that of grassland (Galbally and Roy, 1980). In terms of a mass budget of ozone, a higher surface resistance would create a decreased vertical flux divergence across the mixed layer. The surface would still remain a net sink for ozone but the difference in ozone concentration between air parcels moving over water and grassland could explain the difference in concentrations observed between the Pitt River Valley and the L F V . For a single day from Pacific '93, McKendry (1996; personal communication) calculated a 3 ppb hr"1 difference in net photochemical production of ozone in a vertical column advected over the 25 km stretch of Pitt Lake compared to air advected over grassland in the L F V . This may explain some of the observed difference in gradients between the Pitt River valley and the L F V (shown in Figure 25). However, in estimating the effect of overwater trajectories several significant assumptions were made (eg. the boundary layer is treated as a slab and deposition losses are assumed to be mixed right through the layer) and important processes were neglected. For example the effects of a 67 strong stable layer over the lake surface and deposition along the slopes of the valley were neglected. Further research is required to estimate the true effect of the lake surface on ozone concentrations within the boundary layer over Pitt Lake. 5.2.4 Multi-Day Storage Pollutants are advected into the valley during the day and clean air advected out of the valley in a shallow jet at night (Banta et al, 1995). However, not all ozone in the mixed layer of the valley may be chemically scavenged or removed through surface processes. Ozone could be stored in elevated residual layers above the nocturnal katabatic flows. If this is the case then the stored ozone could be detected in the early morning as the mixed layer begins to evolve. Stored ozone would be entrained into the growing mixed layer of the morning and a negative vertical flux divergence at the top of the layer would be evident. The Pitt River Valley horizontal circulations are intimately tied to vertical mixing, especially during early morning hours when the mixed layer begins to evolve. When the flow in the valley begins to move up valley after sunrise, the mixed layer begins to evolve and air is entrained into the growing mixed layer. Vertical mixing within the valley could also involve elevated layers of ozone from convective chimneys out of the boundary layer or the injection of pollutants aloft from mountain slope flows above the boundary layer, in addition to residual mixed layers from the previous day (McKendry et al,1996; Evans et al,1992). The valley circulations may involve the entrainment of these layers of pollutants from above, producing a negative vertical flux divergence across the top of the mixed layer. Using a simple mass budget 68 technique will not resolve the evolution of these elevated layers but will provide evidence of their existence. 5.3 13 A U G U S T , 1994: A C A S E S T U D Y In the previous sections, several hypotheses have been presented to account for the observed high concentrations of ozone generally encountered in the Pitt River valley. It is possible that all may contribute to the observed ozone distribution. It is also possible that the ozone gradients calculated from aircraft data are only apparent horizontal gradients. If the aircraft is flying through a stratified mixed layer where the isosurfaces are slightly sloped, then the gradients calculated would be the result of the aircraft flying through the sloping ozone layers. This is unlikely as the diurnal change in advection in the valley would prevent the formation of sloped isosurfaces. During early mornings, it is likely the differences in vertical flux divergence between the Pitt River valley and the L F V are important. In this section, intensive aircraft observations from 13 August, 1994 are exploited to investigate the role of multi-day storage and early morning vertical flux divergence in contributing to the observed patterns. The 13 of August, 1994 turned out to be the best day of the month to use as a case study for the calculation of the ozone mass budget of the Pitt River valley. Surface monitors in the L F V , and particularly at Pitt Meadows airport (station 16, Figure 1) recorded the highest hourly mean concentrations of ozone for the month of August 1994. Ozone concentrations on the previous day were the second highest for the month of August. Therefore, i f residual ozone layers are a factor in the high concentrations found in the Pitt River valley they should be evident on the morning of the 13 August. The 69 sea/land breeze and mountain/valley wind systems were also well developed on the 13th as the surface pressure gradient was weak and temperatures high, (see previous chapter and Figure 21b). Winds at Vancouver International Airport were east at 2.8-4.2 m/s at 0800 PST and south-west at 4.2-5.6 m/s by 1300 PST. In the Pitt River Valley the tethersonde measured a down valley flow at ~1.4 m/s at 0600 PST, (the tethersonde was not measuring profiles at this time but was stationary at approximately 15 m above the ground) switching to an up valley flow at -0900 PST and reaching a maximum near 4.2 m/s by 1400 PST. Aircraft measurements in the valley began at 0600 PST and tethersonde operations were underway by 0900 PST providing a wealth of data for analysis. The first early morning profiles of specific humidity, potential temperature, ozone, and wind speed and direction measured at Little Goose Island (see Figure 1) by tethered balloon on 13 August are shown in Figure 26. The mixed layer height is not distinguishable from these profiles. Plots of specific humidity and potential temperature usually provide a good estimate on the mixed layer height, as large vertical gradients of humidity and potential temperature are expected at the top of the mixed layer. Here, all five profiles were used to estimate the elevation where the mixed layer began to evolve. Later in the day, the top of the mixed layer was easily distinguishable by identifying the level where vertical gradients change significantly. At 0900 PST, the layer near the top of the mixed layer where entrainment and vertical fluxes are strongest was estimated to be between 400m and 500m A G L and the aircraft flew through this layer at 450m A G L . Note the high concentrations of ozone 70 1000 900 800 700 600 £ j 500 a < "S 400 300 200 100 0 0 0 360 Wind rection o Wind Speed (m/s) Aircraft Measurement Height 450m Ozone (PPb) I .Specific Humidi ty (g/kg) Potential Temp. (K) Figure 26. Profiles of specific humidity, potential temperature, wind speed and direction, and ozone measured at Little Goose Island, 0903 PST, 13 August 1994. Profiles are averaged measurements taken during ascent and descent of the tethersonde. 71 above 450m and the corresponding wind directions at this level. The wind direction in this layer of elevated ozone concentrations, had a down valley component, (wind directions near 360° are down valley) suggesting that this was ozone from the previous days mixed layer and was de-coupled from the flow in the growing mixed layer of the morning. Aircraft measurements were made at 450m, near the top of the evolving, early morning, Pitt River Valley mixed layer, at approximately 0600, 0800, 1000, and 1300 PST. The traces of ozone measured by the aircraft are displayed in Figure 27. Traces of ozone at 0600 and 0800 PST show no marked change in concentration at the north end of Pitt Lake some 35 km into the valley. There is a difference in the horizontal gradients of ozone between 0600 and 0800 PST. Ozone gradients change sign from a down valley gradient to an up valley gradient over this two hour time period. The measured trace at 1000 PST displays an up valley gradient, an increase in ozone concentration over the north end of the lake and more variability in the ozone signal. By 1000 PST temperatures in the Pitt River valley at 450m A G L had increased by about 5°C since the first flight into the valley. The heating of the valley and subsequent convective activity probably account for the variability in the trace of ozone measured at 1000 PST. The last ozone flight into the valley at 1300 PST found significantly higher concentrations then previous flights. Photo-chemical production of ozone and advection into the valley had strengthened and could account for the higher ozone concentration. The change in ozone gradients from 0600 to 0800 PST is a much more puzzling aspect as advection was from north to south and photo-chemical production of ozone was 72 negligible as the valley was still shaded from direct solar radiation by the high valley walls. Mean hourly ozone concentrations measured at the G V R D monitoring station near the mouth of the Pitt River valley (station 16) are plotted with ozone concentrations measured from the aircraft in Figure 28. The ozone concentration measured by aircraft is a 15 minute average concentration taken when the aircraft reversed its course and flew back down the valley. The aircraft made a number of 360° turns before flying back down valley. These manoeuvres took approximately 10 minutes to complete and provides an average ozone concentration at 450m A G L at the north end of Pitt Lake. This gives some insight into the change in ozone gradients in the Pitt River valley between 0600 and 0800 PST. Ozone concentrations measured near the surface at station 16 increased rapidly between 0600 and 0800 PST, then levelled off for an hour and then continued to rise at a slower rate until 1600 PST. The initial rise in concentrations can be attributed to convective down mixing of ozone due to warming after sunrise. The rise in concentrations is limited by the amount of ozone available for down mixing and further increases are seen only when solar radiation and temperatures are high enough for photo-chemical production of ozone. This occurred at 0900 PST at station 16, which is situated in a broad flat area of the L F V , not shaded by extreme relief as in the Pitt River Valley. The rise in concentrations after 0900 PST was constant and must be due to advection and photo-chemical production. In comparison the change in mean concentration within the Pitt River valley is much less between 0600 and 1000 PST. However, gradients of ozone within the valley reverse even when opposed by advection. This leads into the hypothesis 74 that vertical, turbulent down mixing within the valley caused the change in gradients. After 1000 PST ozone concentrations in the valley rose at the same rate as at station 16 in the L F V , most likely due to the similar rates of photo-chemical production. The processes involved in the increase of ozone in the Pitt River valley before photo-chemical production are significant and must be investigated more closely. The complete mass budget of ozone in the valley for the early morning of the 13 August, 1994 is calculated in the next section to determine if observed ozone gradients are caused by entrainment of ozone from elevated layers of ozone. Figure 28. Mean hourly ozone concentrations measured near the mouth of the Pitt River valley at the surface and at the north end of Pitt Lake at between 50 and 600 metres some 40 km up the valley, 13 August 1994. 75 5.4 T H E M A S S B U D G E T OF O Z O N E F O R T H E PITT R I V E R V A L L E Y 13 A U G U S T 1994, E A R L Y M O R N I N G If overnight storage of pollutants in the valley is a significant factor contributing to observed ozone gradients, this will be discovered through a calculation of the mass budget of ozone. However, some assumptions are needed to simplify equation 2. Previous research into air quality in the Pitt Valley has shown that air that has been advected into the valley with high ozone concentrations is depleted in NO, (McKendry et al, 1994), leaving few ozone scavenging species. The air mass of the Pitt Valley during elevated ozone episodes can become "chemically aged," after travelling up the valley during the day and then being trapped within the valley over night leaving few scavenging species for the chemical destruction of ozone by next morning. Further evidence of a chemically aged air mass comes from visual observations of the pollutant plume during measurement flights. Pollutants in the mixed layer of the Pitt River valley often appeared white as opposed to the brown mixed layer in the L F V near the mouth of the Pitt River valley. The white colour of pollutants in the tributary valley is due to effective scattering of light by the small secondary fine aerosols associated with a chemically aged air mass. The larger particles and NO2 found nearer the city are less effective at scattering light and are better absorbers of radiation making them appear brown (Wallace and Hobbs, 1977). Photo-chemical production of ozone is likely negligible in the Pitt River valley until the sun can rise high enough to illuminate the steep-walled valley. Temperatures are also quite low before 1000 PST which would also limit photo-chemical production of ozone. For the early morning then, it can be assumed 76 that the internal sources or sinks for ozone are negligible and equation 2 reduces to equation 8. dO, -dO, dO'w' at ox oz Using equation 8, the mass budget of a slab of air near the top of the boundary layer, along the centre of the valley, just after sunrise can be calculated. The following budget calculations are only valid for the layer the gradients are measured in, away from the surface processes of the valley walls, when photo-chemical production of ozone is negligible. Aircraft observations provide data for the calculation of ozone gradients, (x axis), and the rate of change of ozone concentration, and tethered balloon data provides an estimate of u . The vertical flux divergence can now be estimated without any turbulence measurements, providing insight into the vertical turbulent transport of ozone in the valley. The time derivative in equation 8 is estimated from the difference in the mean values between each flight and the horizontal gradient is estimated from a least squares fit of ozone data measured along the valley (x axis) see Figure 27. The 23 km gradient that was calculated from measurements made at 0600 and 0800 PST had a standard error of 1.4 ppb, and the gradient calculated from the measurements made at 1000 PST had a standard error of 2.2 ppb. The acceptable standard error implies that a sufficiently long stretch was used to estimate ozone gradients. The error associated with the estimation of the gradient is inversely proportional to the length of the time series flown, to the 3/2 power (Lenschow, 1970; Lenschow, 1980). A longer flight leg could therefore reduce the error, providing the flight is not so long as to be beyond the urban plume. In the case 77 study presented here, gradients were estimated from a 23 km section of observations over water and each flight was at least 2 hours apart. Balloon operations were undertaken near the middle of the section used for the calculation of gradients. Data from aircraft and tethered balloon flights for the morning of August 13 are used in this case study. The mean ozone concentrations at the north end of Pitt Lake at 450m are plotted in Figure 28. From this figure the rate of change of ozone concentration was calculated. 0600 PST 0 3 = 35.77 ppb = 70.82 jug m"3 0800 PST 0 3 = 37.79 ppb = 74.82 jug m"3 Between the first two flights d03 (74.82 - 70.82)//gm-3 _ dt 7200s Advection of ozone for this time was; 0.55 ng m s = 1 ppb hr" d03 '.—— = - 2 m s dx ( -7.92 ju gm"3 A 23,000m = 0.69 ng m"3 s"1 = 1.25 ppb hr"1 where the negative sign for u denotes a down valley wind direction, and the negative sign in the numerator of the gradient implies less ozone with distance up the valley, a down valley gradient. With the above assumptions, advection must be balanced by a negative vertical flux divergence. By residual, from equation 7 the vertical flux divergence was; d 03W . . . . . , . — r — =-0.55ng m"3 s"1 -0.69ng m"3 s"1 = -1.24 ng m"3 s"1 = -2.25 ppb hr"1 oz With a down valley wind and ozone gradient, advection of ozone can be described as clean air flowing out of the valley, yet ozone levels remain constant or increased slightly at the north end of the lake. This implies that the negative vertical flux 78 divergence maintained or increased ozone concentrations at 450m A G L (ie. entrainment of ozone rich air from above, into the mixed layer). The down mixing of ozone into the boundary layer and the advection of "clean air" out of the valley .work to reverse the ozone gradient and by 0800 advection reversed in sign and could be characterised as "dirty" air flowing down the valley. By 0900 solar heating of the valley walls begins to reverse the thermal circulation to an anabatic flow and again the advection term changes sign. Now advection can be characterised as "clean air" flowing up valley. The time rate of change of ozone at the north end of the lake for the period 0800-1000 was also calculated from Figure 28. 0800 PST 0 3 = 37.79 ppb = 74.82 jug m3 1000PST 0 3 = 41.82 ppb = 82.80 jugm3 d02 (82.80-74.82)//gm' 3 _3 —— = = l . l ngm s =2ppbhr dt 7200s 5 v v and advection was calculated to be: d03 —— = lms ox ^8.91//gm' 3 ^ 23,000m = 0.39 ng i n 3 s"1 = 0.71 ppb hr" Thus, the vertical flux divergence term is approximately: W = - L l n g m " 3 s"' -0.39ng rn"3 s'1 = -1.49 ng m"3 s"1 = -2.71 ppb hr"1 o z It must be noted that by the end of this time period, photo-chemical production cannot be ignored, and the vertical flux divergence is probably over estimated. However, it is still evident that entrainment of ozone rich air from above is still important. 79 The last flight on the thirteenth was made at 1300 PST and by this time photo-chemical production of ozone and advection became the main processes by which concentrations in the valley increased. In Figure 28, the slope of both series is equivalent for the period 1000-1300 PST, indicating the increase of ozone at both sites is via the same process. The boundary layer in the valley had grown over two hundred meters and advection was weak due to the weak gradient in ozone. It is impossible to estimate the vertical flux divergence term by the above method, but it is expected to be quite small as the boundary layer had grown to a height such that it had encompassed or entrained any residual layer that was evident earlier in the day. Figure 29 shows profiles of atmospheric variables at approximately the same time as the last aircraft flight of 13 August 1994. At 13:48 PST, the measured profiles showed a more familiar mixed layer pattern. The top of the mixed layer was between 600m and 700m A G L and any elevated ozone layers that were evident earlier in the day are gone. Strong vertical gradients in potential temperature and ozone indicate a mixed layer height close to 600m while the profile of specific humidity suggests a mixed layer height closer to 700m. Ozone, and potential temperature were near constant with height in the mixed layer with sharp vertical gradients at the top of the layer. At this time, there is no evidence of elevated ozone levels above the mixed layer as the early morning layer of ozone had been entrained into the mixed layer. Wind direction was constant at 180°, an up valley flow, throughout the entire measured profile and wind speeds were higher than the early morning especially in the lowest part of the atmosphere where the wind speed exceeded 7 ms"1. A faster flow through a thicker layer than in the morning is now advecting pollutants into the valley. 80 a < 1000 900 800 700 600 500 .2? 400 300 200 100 0 Speed (m/s) Wind Direction o Ozone 1 (ppb) Specific Humidity /kg) | 1 Potential 297 298 299 300 301 3 0 2 ™ Figure 29. Profiles of specific humidity, potential temperature, wind speed and direction, and ozone measured at Little Goose Island, 1348 PST, 13 August 1994. Profiles are averaged measurements taken during ascent and descent of the kytoon. 81 If entrainment of ozone is indeed the process which is accounting for the high ozone concentrations in the early morning boundary layer, then there must be a positive vertical gradient of ozone as there is no surface source of ozone. From Figure 26, a profile of ozone taken the morning of the 13th at Little Goose Island, a layer of ozone above the mixed layer is clearly evident. If the flux of ozone at the top of the boundary layer is parameterized using equation 3 (Lenschow, 1981) and the destruction of ozone at the surface is parameterized using equation 5 (Galbally and Roy, 1980) then it is possible to estimate the vertical flux divergence across the entire boundary layer. Using the early morning profiles at Little Goose Island and entrainment velocities calculated from equation 4 (Stull, 1976) and the surface resistance of water, R s _ 1 , to be lOOOs/m (Galbaly, 1981) the vertical flux divergence for the time period 0900-1045 PST is: - w A a - r ' o , = -0.19 ng m"3 s'1 = -0.35 ppb hr"1 100m , where: we = ——— = 0.16 m s 6300s A0 3 =5ppb = 9.9ngm- 3 0 3 = 26 ppb = 51.48 ng m"3 zi = 550 m This is a lower value than the calculations made by residual (-1.24 and -1.49 ng m"3 s"1). However, calculations made by parameterizing the fluxes at the top and bottom of the boundary layer represent the flux divergence over the entire layer, where the possibility of vertical ozone gradients within the mixed layer could reduce the magnitude of the flux divergence. The parameterization also does not take into account the atmospheric 82 resistance to surface deposition from the shallow stable layer over Pitt lake's cold surface. Most importantly, the two methods agree on the sign of the flux divergence, implying an entrainment of ozone rich air into the early morning growing mixed layer of the Pitt River valley. 5.5 ENTRAINMENT AND RESIDUAL LAYERS, A CONCEPTUAL MODEL The vertical down mixing of ozone rich air into the growing mixed layer of the Pitt River valley is evidence of residual layer storage of the type described by Neu et al (1994) and observed by Wanner et al (1992). Air is advected into the valley by day as the valley is warmed and up slope flows develop (Sturman, 1987). The mixed layer of the valley also grows and by the middle to late afternoon the fully developed layer reaches a height of approximately 700-800m. Within this layer advection and photo-chemical production of ozone raise ozone concentrations to levels comparable to those found in the L F V east of Vancouver. Advection of ozone rich and photo-chemically active air into the valley is through the entire boundary layer which is much deeper than at night and even the early morning. There is considerable warming along the valley walls and the air moves up slope depositing ozone, (Broder & Gygax, 1985). Ozone in the middle of the valley, above the surface layer and some distance away from the valley walls is essentially excluded from this removal process. After sunset, the valley cools and the shallow, stable surface layer evident throughout the day grows, reducing mixing and de-coupling the mixed layer from the surface. Along the valley walls, cool air moves down slope mixing along the surface and deposits ozone resulting in a cool, relatively ozone free, low level nocturnal jet along the valley axis (Banta et al, 1996). The source of air 83 for this nocturnal jet is the down slope flow caused by surface cooling and can originate from well above the previous days mixed layer height as the valley walls rise to heights in excess of 1800m. Overnight, the light and variable mountain winds continue, however the flow is confined to a much shallower layer than the fully developed mixed layer and at lower wind speeds than those observed at mid to late afternoon. Above the stable nocturnal layer the previous days mixed layer is still evident as a relatively ozone rich residual layer. Mixing at the top of this layer with the free atmosphere works to erode the capping inversion and disperse pollutants trapped in this layer. Under stable synoptic conditions, a strong anti-cyclone or ridge, the subsidence from synoptic forcing keeps this mixing to a minimum and the residual mixed layer can remain essentially intact overnight. By morning, the residual mixed layer is still evident and high levels of ozone are still present above the stable nocturnal layer. Below, the new mixed layer grows as turbulent kinetic energy increases with surface heating by solar radiation. As the new mixed layer grows, mixing occurs along the top of this layer and ozone rich air is entrained from the residual layer of the previous day. Results from 13 August 1994, suggest that this process may occur in the early morning, giving rise to an along valley gradient of ozone. During the day newly formed pollutants and precursors are advected up the Pitt River valley increasing ozone concentrations as photochemical acitivity continues with distance from the source of pollutants. Under these conditions, it is likely that differences in surface deposition between trajectories in the L F V and Pitt River valleys contribute to the observed high concentrations in the Pitt River valley. Thus the Pitt River valley can be considered a net sink for ozone emanating from Greater Vancouver due to strong daytime advection and overnight storage of ozone. 84 CHAPTER 6: CONCLUSIONS 6.1 OBJECTIVES The primary objectives of this study were to: • develop a method of measuring mean ozone concentrations with a light aircraft and a chemiluminescent ozone sonde, • determine the spatial distribution of summertime ozone concentrations within the Lower Fraser valley and its tributary the Pitt River valley, and • investigate the processes that lead to high ozone concentrations in the Pitt River valley. The main results associated with each of these goal are summarised below. Finally, recommendations for future research are presented. 6.1.1 Alternative Monitoring Techniques The GFAS, OS-B-2, dry, chemiluminescent ozone sonde was intended for use as a balloon borne disposable instrument. It proved to be a reliable and highly successful instrument for measuring mean ozone concentrations from an aircraft. A simple Teflon intake that included a pressure bleed off, was fitted to the port side of a single engine, light aircraft directly in the propeller wash, and attached to the OS-B-2 sonde inside the planes cabin. The pressure from the propeller wash of the aircraft was sufficient to ensure a constant supply of air to the sonde and did not affect the ozone concentration in the air sample. Verification of the aircraft measurement technique came from comparison to measurements made by electro-chemical ozone sondes attached to tethered balloons and comparisons to mean ozone concentrations measured by permanent surface ozone monitors. The aircraft measured mean ozone concentrations within 3.5 ppb of ozone 85 concentrations measured via tethered balloon, when the aircraft was within 100 m of the kytoon. Comparison to surface ozone concentrations showed the aircraft consistently measured higher concentrations at 450m A G L and that the ozone concentration at that height was always within 15% of the surface concentration, after 1000 PST. After 1000 PST the boundary layer of the L F V is sufficiently well mixed for this relation to hold true. Although the aircraft flew over a variety of surface types and mean ozone concentration levels at 450m A G L ranged from 40 ppb to 110 ppb the surface concentrations were representative of concentrations throughout the mixed layer. As the measurements were made within the convective boundary layer of the L F V , ozone was well mixed vertically and surface concentrations were slightly lower than at 450m A G L due to dry deposition. To help with the verification of the aircraft measurements and determining surface layer ozone concentrations, passive ozone samplers were incorporated into the measurement program. The passive samplers were designed for use as personal ozone monitors so determining their applicability in measuring average ozone concentrations was a secondary objective of this study. The passive samplers systematically underestimated the average ozone concentration but were able to confirm along valley gradients observed by the aircraft. 6.1.2 The Spatial Distribution of Ozone The spatial distribution of ozone concentrations at 450m A G L was determined for a variety of summertime weather conditions. The results of this study are in agreement with previous studies by Pry or et al (1995) and Evans et al (1992) which investigated the 86 surface distribution of ozone across the LFV. The highest ozone concentrations in the L F V were found downwind of Greater Vancouver in the eastern half of the L F V and to the north-east of Greater Vancouver into the Pitt River valley. For the first time, ozone gradients in the Pitt River valley itself were determined and were found to be greater that or equivalent to gradients in the LFV. The greatest gradients in ozone concentrations were found just east of the urban area approximately 20 km from the city of Vancouver. Ozone concentrations over the eastern half of the broad, flat L F V varied little with only weak horizontal gradients. Horizontal ozone gradients in the Pitt River valley were greater than in the eastern L F V especially when mean concentrations were high and had been for more than one day. When conditions are conducive to high ozone concentrations, the up-valley flow in the Pitt River valley is well developed, and pollution is advected into the valley. If these conditions persist in the valley during a multi-day ozone episode then it appears that mean concentrations can be influenced by the recirculation of pollutants within the valley. 6.1.3 Processes Affecting Concentrations in the Pitt River Valley The hypothesis of multi-day ozone storage in the Pitt River valley was tested using a mass budget technique. The early morning mass budget of ozone in the Pitt River valley was calculated for a layer of air at 450m A G L in the centre of the valley. The results of this analysis showed that the early morning vertical flux divergence at this level is negative. Thus ozone that is stored above the growing mixed layer of the morning is entrained down into the layer resulting in high mean ozone concentrations in the fully developed mixed layer of the afternoon. A first order slab model of ozone entrainment 87 was also used to confirm the mass budget results and the direction of the vertical flux divergence was the same. The slab model parameterization produced a negative flux divergence although the magnitude of the divergence was less than the mass budget results. The slab model calculates the flux divergence across the entire mixed layer while the mass budget technique is only applicable at the level of gradient measurement. Both methods did show that vertical down mixing of ozone into the growing mixed layer of the Pitt River Valley has a significant impact on the level of mean concentrations in the valley. The ozone that was found to be entrained into the growing mixed layer was likely from previous days mixed layers. Residual mixed layers in the Pitt River Valley result from the difference in anabatic and katabatic flows. The daytime anabatic flow in the valley is through a deeper layer and at greater wind speeds than the nocturnal katabatic flow. The anabatic flow in the valley is accentuated by the sea breeze flow which results in strong advection into the valley. At night the katabatic flow is confined to a shallower layer and the air above this layer can be left undisturbed. Entrainment of pollution out of the residual mixed layer is kept to a minimum from subsidence forcing from an upper level ridge that is generally associated with elevated ozone concentrations. When previous days ozone concentrations are high, entrainment of residual mixed layer ozone into the growing new mixed layer is of greater magnitude. This model of residual layer ozone storage in the Pitt River valley shows how the valley can be considered a net sink for ozone emanating from the LFV. 88 6.2 F U T U R E R E S E A R C H This study was the first attempt to measure mean ozone concentrations in the boundary layer of the Pitt River valley. The data that was gathered offered an opportunity to investigate some of the processes affecting pollutant distribution within this tributary valley. The extent to which other large tributary valleys of the L F V (in particular the valleys including Stave and Harrison Lakes) affect the distribution of pollutants in similar ways to the Pitt River valley remain uncertain. The mass budget of ozone in these tributary valleys should be investigated and compared to the mass budget calculations presented here. In the present study an assessment of the mass budget was made for only one moderate ozone episode, and using observations from one level and one tethersonde system. A more detailed mass budget calculation could be made in the Pitt River and other tributary valleys by utilising multiple tethered balloons and aircraft observations at multiple levels in and above the mixed layer. The change with respect to height in the vertical flux divergence of ozone could then be determined and thus reveal the relative importance of entrainment and surface deposition in determining mean ozone concentrations in the mixed layer. An investigation of the entrainment of ozone into the mixed layer could also be achieved via numerical techniques. Using balloon profile data as model input, a transilience turbulence model would more accurately predict ozone fluxes into or out of the mixed layer of the valley. Numerical modelling of the flows and photochemistry of the Pitt River valley would also be informative. In particular, this would help determine what effect mixing of chemically aged air has on ozone concentrations in the mixed layer. 89 Finally, an investigation into the impacts of high pollutant concentrations (including ozone) on the tributary valleys is needed. These valleys of the L F V represent a significant, economic, recreational and ecological resource. The high concentrations observed in this study suggest that the ecological aesthetics (particularly visibility) and economic value of these remote areas is being severely compromised. 90 B I B L I O G R A P H Y Atkinson, B.W., 1981: Meso-Scale Atmospheric Circulations, Academic Press, Toronto. Banta, R .M. , P.B. Shepson, J.W. Bottenheim, K. Anlauf, H.A. Wiebe, A.J . Gallant, T. Biesenthal, L.D. Olivier, C.J. Zhu, D.G.Steyn, and I.G. McKendry, 1995: Nocturnal Cleansing Flows in a Tributary Valley, (in press, accepted 28 March 1996), Atmospheric Environment. Broder, B. , and H.A. Gygax, 1985: The Influence of Locally Induced Wind Systems on the Effectiveness of Nocturnal Dry Deposition of Ozone, Atmospheric Environment, Vol. 19, No. 10, 1627-1637. Concord Scientific Corporation (CSC), 1985: Vancouver Oxidants Study, Air Quality Analysis Study Update 1982-1984. Prepared for Environment Canada, Environmental Protection Service. Chang, Y.S. , G.R. Carmichael, H . Kurita and H. Ueda, 1989: The Transport and Formation of Photochemical Oxidants in Central Japan, Atmospheric Environment, Vol. 23(2), 363-393. Evans, C , B. Martin, F. Froude and B. Thomson 1992: Ozone and Meteorological Profdes in the Lower Fraser Valley of British Columbia, Atmospheric Environment Service, Pacific Region. Galbally, I.E., and C R . Roy, 1980: Destruction of Ozone at the Earth's Surface. Q. J. Roy. Met. Soc, Vol. 106, 599-620. Galbally, I.E., 1971: Ozone Profiles and Ozone Fluxes in the Atmospheric Surface Layer. Quart. J. R. Met. Soc, Vol. 97, 18-29. 91 Garratt, J.R., 1992: The Atmospheric Boundary Layer. Cambridge University Press, New York. Greater Vancouver Regional District (GVRD), 1993: Ambient Air Quality Annual Report Greater Vancouver Regional District (GVRD), 1994: Ambient Air Quality Annual Report Guicherit, R. and H. Van Dop, 1970: Photochemical Production of Ozone in Western Europe (1971-1975) and Its Relation to Meteorology. Atmospheric Environment, . Vol. 11, 145-155. Giisten, H., 1986: Formation,Transport and Control of Photo-chemical Smog. Air Pollution, O. Hutzinger, Ed., The Handbook of Atmospheric Chemistry, Vol. 4A, 53-106. Harrison, R .M. , C D . Holman, H.A. McCartney, and J.F.R. Mcllveen, 1978: Nocturnal Depletion of Photochemical Ozone at a Rural Site. Atmospheric Environment, Vol. 12, 2021-2026. Koutrakis, P., J .M. Wolfson, A . Bunyaviroch, S.E. Froehlich, K. Hirano and J.D. Mulik, 1993: Measurement of Ambient Ozone Using a Nitrite-Coated Filter, Analytical Chemistry, Vol. 65, 209-214. Lenschow, D.H., and B.B. Stankov, 1986: Length Scales in the Convective Boundary Layer. J. of the Atmospheric Sciences, Vol. 43, No. 12, 1198-1210. Lenschow, D.H., R. Pearson and B.B. Stankov, 1981: Estimating the Ozone Budget in the Boundary Layer by Use of Aircraft Measurements of Ozone Eddy Flux and Mean Concentration. J. of Geophysical Res., Vol. 86, No. C8, 7291-7297. Lenschow, D.H., J.C. Wyngaard, and W.T. Pennell, 1980: Mean-Field Second-Moment 92 Budgets in a Baroclinic, Convective Boundary Layer., J. of the Atmospheric Sciences, Vol.37, 1313-1326. Lenschow, D.H., 1970: Airplane Measurements of Planetary Boundary Layer Structure. Journal of Applied Meteorology, Vol. 9, 874-884. Lalas, D.P., D.N. Asimakopoulos, D.G. Deligiorgi and C.G. Helmis, 1983: Sea Breeze Circulation and Photo-Chemical Pollution in Athens, Greece., Atmospheric Environment, Vol. 17, 1621-1632. Lu, R., and R.P. Turco, 1994: Air Pollutant Transport in a Coastal Environment. Parti: Two-Dimensional Simulations of Sea-Breeze and Mountain Effects. J. of the Atmospheric Sciences, Vol. 51, No. 15, 2285-2308. McElroy, J.L. and T.B. Smith 1992: Creation and Fate of Ozone Layers Aloft in Southern California., Atmospheric Environment, Vol. 27A, No. 12, 1917-1929. McKendry, I.G., D.G. Steyn, J. Lundgren, R . M . Hoff, W. Strapp, K . Anlauf, F. Froude, J.B. Martin, R . M . Banta, and L.D. Olivier 1997: Elevated Ozone Layers and Vertical Down Mixing over the Lower Fraser Valley, B.C., Atmospheric Environment. McKendry, I.G., 1994: Synoptic Summertime Ground-Level Ozone Concentrations at Vancouver, British Columbia. J. Applied Meteorology. Vol.33(5), 627-641. McKendry, I.G., D.G. Steyn, R . M . Banta, and R . M . Hoff, 1994: Pacific '93: Elevated Pollutant Layers over the Lower Fraser Valley. American Geophysical Union, Fall Meeting, San Francisco, Cal. Dec. 5-9. Miao, Y. , 1993: Mesometeorological Modelling and Trajectory Studies During an Air 93 Pollution Episode in the Lower Fraser Valley, British Columbia, Canada. The University of British Columbia, M.Sc. Thesis. Neu, U . , T. Kiinzle and H . Wanner, 1994: On the Releation Between Ozone Storage in the Residual Layer and Daily Variation in Near-Surface Ozone Concentration-A Case Study. Boundary-Layer Meteorology, Vo l . 69, 221-247. Oke, T.R. and J. Hayl994: The Climate of Vancouver., B.C. Geographical Series, Second Edition, No. 50 Oke, T.R., 1987: Boundary Layer Climates., Methuen and Co., London and New York, p. 297. Olivier, L.D. , R . M . Banta, R . M . Hardesty, 1994: Wind Flow in the Fraser Valley as Measured by a Pulsed CO2 Doppler LIDAR. Applications of Air Pollution Meteorology with A W M A , January, 1994. Pryor, S.C., I.G. McKendry, and D.G. Steyn, 1995: Synoptic-Scale Meteorological Variability and Surface Ozone Concentrations in Vancouver, B.C. , Journal of Applied Meteorology, Vo l . 34, No. 8, 1824-1833. Puxbaum, H. , K. Gabler, S. Smidt, F. Glattes, 1991: A One-Year Record of Ozone Profiles in an Alpine Valley (Zillertal/Tyrol, Austria, 600-2000m a.s.L). Atmospheric Environment, Vol . 25A, No. 9, 1759-1765. Simpson, J.E., 1994: Sea Breeze and Local Winds. Cambridge University Press, Cambridge, 234p. Speuser, W., S. Sahand, and U . Schurath, 1989: A Novel Fast Response Chemiluminescence Sonde for Routine Soundings of Stratospheric Ozone up to 94 1.5 mb. Ozone in the Atmosphere., R.D. Bojlov and P. Fabien Eds., Deepak Publishing. Stern, A.C. , H.C. Wohlers, R.W. Boubel, W.P. Lowry, 1973: Fundamentals of Air Pollution. Academic Press, New York. Stephens, G.L., 1984: The Parameterization of Radiation for Numerical Weather Prediction and Climate Models. Monthly Weather Review., Vol. 112, 826-867. Steyn, D.G., 1996: Air Pollution in Coastal Cities. Air Pollution Meteorology and its Application (XI). H . Van Dop and G. Kallos, Eds., in Press., Plenum, New York. Steyn, D.G., M . Bovis, M . North and O. Slaymaker, 1992: The Biophysical Environment Today. Vancouver and Its Region. G. Wynn and T.R. Oke, Eds., U B C Press, University of British Columbia, Vancouver, B.C., Canada. Steyn, D.G., 1990: An Advective Mixed-Layer Model for Heat and Moisture Incorporating and Analytic Expression for Moisture Entrainment. Boundary-Layer Meteorology, Vol. 53, 21 -31. Steyn, D.G., A .C . Roberge and C. Jackson, 1990: Anatomy of an Extended Air Pollution Episode in British Columbia's Lower Fraser Valley. Report prepared for the Waste Management Branch, British Columbia Ministy of Environment Steyn, D.G. and D.A. Faulkner, 1986: The Climatology of Sea-Breezes in the Lower Fraser Valley, B.C., Climatological Bulletin, Vol. 20(3), pp. 21-39. Steyn, D.G. and T.R. Oke, 1982: The Depth of the Daytime Mixed Layer at two coastal Sites: A Model and Its Validation. Boundary-Layer Meteorology, Vol. 24 pp. 161-180. 95 Stull, R.B., 1988: An Introduction to Boundary Layer Meteorology. Kluwer Academic Publishers, The Netherlands, pp. 461. Stull, R.B., 1976: The Energetics of Entrainment Across a Density Interface. J. of the Atmospheric Science, Vol . 33, 1260-1267. Sturman, A.P. , 1987: Thermal Influences on Airflow in Mountainous Terrain., Progress in Physical Geography II. pp 183-206 Taylor, B. , 1992: The Relationship Between Ground-Level Ozone Concentratios, Surface Pressure Gradients and 850 mb Temperatures in the Lower Fraser Valley of British Columbia. Atmospheric Issues and Services Branch, Atmospheric Environment Service, Report PAES-92. Taylor, B. , 1991: Forecasting Ground-Level Ozone in Vancouver and the Lower Fraser Valley of British Columbia. Atmospheric Issues and Services Branch, Atmospheric Environment Service, Report PAES-92. Tennekes, H. , and J.L. Lumley, 1972: A First Course in Turbulence. MIT Press Thomson, R.B., C.E. Evans, F.Froude and B. Martin, 1995: Ozone and Meteorological Profiles in the Lower Fraser Valley, B.C. 29th Annual CMOS Congress, May 1995. Van Dop, H. , and R. Guicherit, 1980: The Vertical Distribution of Ozone in the Atmospheric Boundary Layer. Q. J. Roy. Met. Soc, Vol . 106, 599-620. Van Valin, C.C. et al, 1994: The Compatibility Between Aircraft and Ground-Based Air Quality Measurements. J. of Geophysical Res., Vol . 99, No. D l , 1043-1057. Wallace, J.M., and P.V. Hobbs, 1977: Atmospheric Science An Introductory Survey. 96 Academic Press, New York. Wanner, H. , T. Kunzle, U . Neu, B. Ihly, G. Baumbach and B. Steisslinger, 1993: On the Dynamics of Photochemical Smog over the Swiss Middleland-Results of the First P O L L U M E T Field Experiment., Meteorology and Atmospheric Physics, Vol. 51, 117-138. 97 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0052592/manifest

Comment

Related Items