UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Ground-based LiDAR and air quality observations on Grouse Mountain, British Columbia during the summer… Pomeroy, Carrington 2019

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


24-ubc_2019_november_pomeroy_carrington.pdf [ 3.8MB ]
JSON: 24-1.0384581.json
JSON-LD: 24-1.0384581-ld.json
RDF/XML (Pretty): 24-1.0384581-rdf.xml
RDF/JSON: 24-1.0384581-rdf.json
Turtle: 24-1.0384581-turtle.txt
N-Triples: 24-1.0384581-rdf-ntriples.txt
Original Record: 24-1.0384581-source.json
Full Text

Full Text

Ground-based LiDAR and air qualityobservations on Grouse Mountain, BritishColumbia during the summer of 2018byCarrington PomeroyB.Sc., Carleton University, 2017A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMaster of ScienceinTHE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES(Geography)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)October 2019©Carrington Pomeroy, 2019The following individuals certify that they have read, and recommend to the Faculty ofGraduate and Postdoctoral Studies for acceptance, a thesis entitled:Ground Based LiDAR and Air Quality Observations on Grouse Mountain, British ColumbiaDuring the Summer of 2018submitted by: Carrington Pomeroy in partial fulfillment of the requirementsfor the degree of Master of Sciencein GeographyExamining Committee:Ian McKendry, GeographySupervisorPaul Cottle, GeographySupervisory Committee MemberBrett Eaton, GeographyAdditional ExamineriiAbstractWidespread and persistent summer multi-day episodes characterized by dense layers of wild-fire smoke emanating from western wildfires have increased in frequency in recent years acrosswestern Canada. These events often occur under otherwise clear sky anti-cyclonic weatherconditions and have significant impacts on surface temperatures, surface radiation and en-ergy budgets. Here, we present previously undocumented mountain-top, wildfire influencedparticulate matter concentrations and compare them to those recorded in the valley. Thedistribution of particulate matter both temporally and spatially is presented as well. The fo-cus of this observational study is in the vicinity of Grouse Mountain, near Vancouver, BritishColumbia. Observations are made using a GRIMM 1.108 Dustcheck mini-mass-spectrometer,a Dylos DC1100 Pro air quality monitor, a mini micropulse LiDAR (light detection andranging) and vertical sounding using mini sondes (WINDSOND). The Hybrid Single Parti-cle Langrangian Integrated Trajectory (HYSPLIT) air pollution modelling software is usedto track parcels of wildfire smoke. Results show enhanced mountain-top particulate matterconcentrations with many instances displaying higher concentrations on Grouse than in thevalley, most commonly under anti cyclonic conditions. Evidence of a mountain boundarylayer in the presence of smoke is presented, as well as signs of suppressed convective ventingand more stable vertical profiles, likely due to the radiative effects of smoke.iiiLay SummaryThe interaction between wildfire smoke and mountain top environments has been sparselyaddressed in the literature. This study examines the effects of wildfire smoke on particulatematter concentrations and distribution on Vancouver’s Grouse Mountain during the summerof 2018. The difference between air quality measurements taken in Vancouver and on topof Grouse during wildfire events is of particular interest. Results show that when wildfireplumes arrive from the northern, mountainous regions, particulate matter concentrationson Grouse increased well before those in Vancouver. The atmosphere is also shown to bemore stable in the presence of smoke and mountain flow processes are shown to still function,albeit with less effect than normal. These results begin to fill knowledge gaps in air pollutionand weather modelling as well as air quality monitoring and advisory.ivPrefaceThe research topic and study design was developed by Carrington Pomeroy and Ian McK-endry. Quality assurance and quality control for LiDAR data was performed by Paul Cottle.Data from Windsond balloon launches was collected by Carrington Pomeroy, Madison Fer-rara and Dylan Weyell. The final manuscript was prepared by Carrington Pomeroy withediting by Ian McKendry, Paul Cottle and Allan Betram. Figures 4.5 and 4.6 appear in apaper that has been submitted for publication:Ferrara, M, Pomeroy, C, McKendry IG, Stull, R and Strawbridge, K. 2019. Suppressionof Mountain Convective Boundary Layer ”Handover” Processes by Persistent WildfireSmoke over Southwestern British ColumbiavContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiLay Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vContents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiList of Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.1.1 Dispersal of Aerosols . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.1.2 Known Responses to Aerosols . . . . . . . . . . . . . . . . . . . . . . 41.1.3 Potential Impacts in Mountainous Regions . . . . . . . . . . . . . . . 51.2 Research Objectives and Thesis Structure . . . . . . . . . . . . . . . . . . . 82 Experimental Methods and Analysis . . . . . . . . . . . . . . . . . . . . . . 102.1 Description of Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102.1.1 Geography of the Lower Fraser Valley . . . . . . . . . . . . . . . . . . 102.1.2 Climate of the Lower Fraser Valley . . . . . . . . . . . . . . . . . . . 112.1.3 Mesoscale Circulations within the Lower Fraser Valley . . . . . . . . 122.2 Instrumentation and Data Analysis . . . . . . . . . . . . . . . . . . . . . . . 142.2.1 GRIMM 1.108 Aerosol Spectrometer . . . . . . . . . . . . . . . . . . 142.2.2 MetroVancouver’s Instrument Network . . . . . . . . . . . . . . . . . 15vi2.2.3 Dylos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172.2.4 LiDAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.2.5 Data Analysis Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 233 Non-Smoke Particulate Matter Concentrations and Distribution . . . . . 243.1 PM2.5 Spatial Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253.2 LiDAR Imagery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273.3 GRIMM PM2.5 Concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . 313.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334 Smoke Event Particulate Matter Concentrations and Distribution . . . . 354.1 PM2.5 Spatial Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364.2 LiDAR Imagery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384.3 GRIMM PM2.5 Concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . 434.4 Case Study on August Fire Event . . . . . . . . . . . . . . . . . . . . . . . . 454.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565.1 Key Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565.2 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.3 Future Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60viiList of Tables2.1 Summary of Field Data Collection . . . . . . . . . . . . . . . . . . . . . . . . 23viiiList of Figures1.1 A Hybrid Single Particle Langrangian Integrated Trajectory (HYSPLIT) modeldemonstrating short range smoke transport. Adapted from McKendry et al.(2011). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.2 A HYSPLIT model demonstrating long range smoke transport. Adapted fromTakahama et al. (2011). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.3 Daily PM2.5 readings taken at Vancouver International Airport in 2017 (Source:MetroVancouver) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.4 The different mountain airflow processes. E represents entrainment; MV,mountain venting; AV, advective venting; and MCV, mountain-cloud venting.Vectors show directions of airflow while c(z) and (z) represent vertical profilesof pollutant concentration and potential temperature respectively. Adaptedfrom De Wekker and Kossmann (2015) . . . . . . . . . . . . . . . . . . . . . 61.5 A visual representation of the different atmospheres that arise in a mountainenvironment. Adapted from De Wekker and Kossmann (2015) . . . . . . . . 72.1 Map of locations mentioned in text. Source: Google Maps . . . . . . . . . . 112.2 Depiction of airflow and boundary layer processes over urbanized complexcoastal terrain. Adapted from (McKendry and Lundgren, 2000) . . . . . . . 132.3 Pictures of the GRIMM and its location on Grouse Mountain . . . . . . . . 152.4 Linear regression done on the readings taken from the collocated GRIM andMetroVancouver FEM instruments. . . . . . . . . . . . . . . . . . . . . . . . 162.5 Pictures of the Dylos and the manner in which it was carried by the surveyor. 182.6 The path followed during the walking surveys and different locations men-tioned in the text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.7 The location of the LiDAR with respect to Grouse Mountain. . . . . . . . . 212.8 An overview of a windsonde launch. Retrieved from www.windsond.com2019/08/15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223.1 Figure3.1: Relative PM2.5 concentrations observed around Grouse Mountain’speak area on July 14th, 2018. . . . . . . . . . . . . . . . . . . . . . . . . . . 26ix3.2 Averaged relative PM2.5 concentrations observed around Grouse Mountain’speak area for the nine surveys done on clear days during the summer of 2018. 273.3 Ground based, upward facing LiDAR measurements taken from near the baseof Grouse Mountain, from June 16th to 21st, 2018. . . . . . . . . . . . . . . 283.4 Ground based, upward facing LiDAR measurements taken from near the baseof Grouse Mountain, from July 23rd to 29th, 2018. The arrow in this figurepoints to what are potentially remnants of wildfire emissions from Siberia andAlaska. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293.5 Incoming radiation, ozone, PM, temperature and windspeed measurementsrecorded at Mahon Park, North Vancouver during the ”photochemical smogevent”. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303.6 300h HYSPLIT backwards dispersion trajectories run from Grouse MountainJuly 26th . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313.7 PM2.5 concentrations observed on Grouse Mountain and at North Vancouver’sMahon Park on July 17th and 23rd, 2018. . . . . . . . . . . . . . . . . . . . 323.8 PM2.5 concentrations observed on Grouse Mountain and at North Vancouver’sMahon Park on July 28th and 29th, 2018, during the “photochemical smogevent”.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334.1 NOAA’s HMS Imagery(left) and MODIS TERRA satellite imagery (right) onAugust 13th 2018 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354.2 Sample survey of relative PM2.5 concentrations observed around Grouse Moun-tain’s peak area in the presence of wildfire smoke on August 17th, 2018. . . . 374.3 Averaged relative PM2.5 concentrations observed around Grouse Mountain’speak area for the nine surveys done on smoke days during the summer of 2018. 384.4 Ground based, upward facing LiDAR measurements taken from near the baseof Grouse Mountain, from August 13th to 19th, 2018. red arrow points to themain plume of wildfire smoke that affected the LFV in 2018. The red dashedline shows the elevation of Grouse Mountain. The white brackets are used toshow smaller plumes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394.5 Average vertical profiles of potential temperature taken from the base ofGrouse Mountain. Profiles from clear days are shown in blue while smokedays are shown in black. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414.6 Vertical profiles superimposed on the earlier LiDAR imagery . . . . . . . . . 42x4.7 HYSPLIT backwards dispersion trajectories that were run. The origin of theback trajectories is Grouse Mountain and the models were run for 48 hoursstarting at 12pm on August 13th, 14th and 15th . . . . . . . . . . . . . . . . 424.8 HYSPLIT backwards dispersion trajectories that were run. The origin of theback trajectories is Grouse Mountain and the models were run for 48 hoursstarting at 12pm on August 16th, 17th and 18th . . . . . . . . . . . . . . . . 434.9 PM2.5 concentrations observed on Grouse Mountain and at North Vancouver’sMahon Park on August 13th during the day, the afternoon of August 15thand August 16th, 2018. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444.10 Aerosol optical depth values measured at Saturna Island, just southwest ofVancouver over the course of August 2018. . . . . . . . . . . . . . . . . . . . 464.11 View of Cypress and Grouse Mountain from Jericho Beach on a clear day(left) and August 20th (right). . . . . . . . . . . . . . . . . . . . . . . . . . . 474.12 Ground based, upward facing LiDAR measurements taken from near the baseof Grouse Mountain, from August 13th to 25th, 2018. The red arrow pointsto the main plume of wildfire smoke that affected the LFV in 2018. The reddashed line shows the elevation of Grouse Mountain. . . . . . . . . . . . . . 474.13 Ozone, PM, incoming radiation, temperature and windspeed measurementsrecorded at Mahon Park, North Vancouver during 2018’s main fire event. . . 494.14 HYSPLIT backwards dispersion trajectories that were run using EDAS 40kmmeteorological inputs. The origin of the back trajectories is Grouse Mountainand the models were run for 300 hours starting at 12am on August 19th. . . 504.15 A map showing the composite mean pressure using isobars along the westcoast of North America for the week of the smoke event. . . . . . . . . . . . 514.16 A map of all active fires in BC on August 18th 2018. Flames represent firesthat present a threat to public safety, red dots are fires that started withinthe past 24h and orange dots are fires that do not represent threats to publicsafety. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524.17 PM2.5 concentrations observed on Grouse Mountain and at North Vancouver’sMahon Park on August 18th, August 19th and the morning of August 20th,2018. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534.18 PM2.5 concentrations observed at North Vancouver’s Mahon Park from August19th to August 26th . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54xiList of SymbolsSymbol Definition UnitsPM2.5 particulate matter with a diameter < 2.5µm µ g m−3PM10 particulate matter with a diameter < 10µm µ g m−3S(z, λ) signal received at the detector counts/µ sE0 energy of laser pulse JA usable area of the receiver km2z height kmγ full system efficiency constant counts−µs∗µJO(z) overlap correction function -βpi volume backscatter coefficient km2sr−1α extinction coefficient -SB background signal counts µs−1d(S) deadtime correction factor -a(z) afterpulse correction factor m counts/µsxiiAcknowledgementsThis thesis would not have been possible without the support of many people. I would firstlike to express sincere thanks and gratitude to Ian McKendry, for his unwavering support,patience and guidance throughout my time at UBC. His hands-off approach allowed me togrow as a researcher and his thoughtful, optimistic demeanor meant that no problem seemedtoo big.I would also like to thank my committee members, Paul Cottle and Allan Bertram forproviding support for this project as well as honest and useful advice. As well as DylanWeyell for the time and effort he brought to this project.This project would not have been possible if not for the generosity of the Weyell familyand the management of Grouse Mountain. Having a safe location to store equipment madethis project significantly easier. Accordingly, I would like to thank the Weyells, Erik Bowkettand the entire staff of Grouse Mountain for being so accommodating.I would like to extend thanks to the Natural Sciences and Engineering Research Council,UBC Geography and Diana L. Belhouse (by way of the Henry C. Belhouse scholarship) forproviding funding this project and many others.Although they may not think they helped; my friends in Vancouver, particularly Stefanand Claire, made this experience unquestionably easier and more enjoyable. Similarly, thesupport and advice I’ve gotten from Kieran Jones throughout this process and most of myacademic career can’t go unmentioned, his ability to listen and talk through things has beeninvaluable.As always, family has played a large role in this pursuit. Thank you to my David andDiane Bond for their keen interest in my education and their company during holidays outwest. Thank you to my parents, John Pomeroy and Lee Anne Johnston, for their unwaveringlove and support in my pursuits. I cannot thank them enough for the opportunities theyhave provided me with.Last but not least, I would like to thank my partner, Isabelle MacLean, for her patience,her belief in me and for providing unmatched inspiration.xiiiChapter 1IntroductionOver the past few decades, wildfires have been increasing in severity and frequency in westernNorth America due to a gradual change in climate and increased human involvement(McClureand Jaffe, 2018). The gradual change in climate has resulted in longer, hotter and driersummers as well as an expansion of the latitudinal range of the mountain pine beetle, all ofwhich result in an increase in a forest’s susceptibility to fire (Carroll et al., 2004; McClureand Jaffe, 2018). The increase in human activity in forests has also had this effect due tosuppression of past fires resulting in larger present fires and increased exposure to fire (Fuscoet al., 2018). As this trend is projected to continue (IPCC , 2014), understanding the impactsof wildfires is becoming increasingly important.Along with the undeniable destruction caused by wildfires, the smoke emitted from theburning materials can have important impacts. Air quality downwind of wildfires can quicklydeteriorate due to the small solid particles and other compounds that are generated eitherdirectly from the source or indirectly through the reaction between sunlight and the emittedorganic compounds (Logan et al., 2013). These small solid particles are called aerosols ormore specifically, Particulate Matter (PM) and are of particular interest in this study. Thesmaller size ranges (PM2.5) are capable of increasing risk of respiratory disease, particularlyin populations susceptible to cardio-pulmonary ailments, as well as affecting regional climateand biogeochemical processes by scattering and absorbing solar radiation.Wildfire induced changes to air quality have been well documented in different environ-ments, with urban areas being of particular interest. Yet, due to the complicated interactionsmountains have with airflow, little is known about the quality of air and the concentrationsof aerosols in mountain environments. Similarly, little is known regarding the effects ofmountains on the distribution of aerosols in their surrounding regions.In 2018, British Columbia experienced a second consecutive exceptionally active wildfireseason that led to several air quality advisory warnings in the City of Vancouver. There were1multiple extended periods with recorded PM2.5 values that exceeded the Canadian AmbientAir Quality Standards(CAAQS) of 28 µg/m3. These values, recorded by a network of sensorsthroughout the Vancouver area, did not, however, account for one significant area.Grouse Mountain, hereafter referred to as ’Grouse’, is one of the most popular summertourist destinations in Vancouver, British Columbia. Dubbed ”The Peak of Vancouver”, itis located 15 minutes north of Vancouver’s downtown core and is easily accessible for visitorsby car or public transit(Grouse Mountain, 2014). It offers restaurants, outdoor activitiesand, at 1200m above sea level, a beautiful view of the city. Every summer, 150000 peoplehike up Grouse’s famous ”Grouse Grind” trail (Grouse Mountain, 2018) while others preferto take the famous ”skyride” cable car. With daily visits approaching the thousands, Grousewas named Vancouver’s top tourist destination in 2014 (Grouse Mountain, 2014).As there is a limited resident population on Grouse, it is understandable that there is noroutine fixed monitoring as is the case in the urbanised Fraser Valley. The second half ofthis chapter will outline literature that collectively suggests that air quality measurementstaken at Grouse could be of great interest.1.1 Background1.1.1 Dispersal of AerosolsSurface air quality can be degraded by aerosols both close to and far from the source.The path aerosols take through the atmosphere depends on many variables including localand regional synoptic conditions(McKendry and Lundgren, 2000), topography and globalcirculation patterns. Additionally, the chemical, optical and physical properties of aerosolscan change as they travel(McKendry et al., 2011).Pollutant dispersal across the surface of the Earth has long been a focus of air qualityresearch. Recently, numerous methods to simulate and predict the transport and dispersionof pollutants have been developed(McKendry and Lundgren, 2000). Developments in bothground-based and satellite-borne remote sensing technologies such as LiDAR (light detectionand ranging), have led to significant advances in the understanding of aerosol distribution.Using these technologies facilitates the tracing of emission transport and dispersion overvarious scales. (Amiridis et al., 2009; McKendry et al., 2011).Under stable conditions, plumes are confined to a shallow atmospheric boundary layer,leading to enhanced aerosol concentrations in the local region. Figure 1.1, shows a case ofmedium range smoke transport. Here, smoke traveled to Vancouver from Northern Californiawhile being contained to lower elevations by an elevated inversion (McKendry et al., 2011).2Figure 1.1: A Hybrid Single Particle Langrangian Integrated Trajectory (HYSPLIT) modeldemonstrating short range smoke transport. Adapted from McKendry et al. (2011).When in proximity to mountains or under unstable conditions, aerosols can be lifted tohigher elevations. Once in the free troposphere, large-scale circulation systems can transportthem hundreds to thousands of kilometers from the source and impact air quality in thesedownwind regions (Figure 1.2)(Kaulfus et al., 2017). For example, in 2012 Teakles et al.(2017) and Cottle et al. (2014) documented the transport of smoke from Siberian wildfiresto the Pacific Northwest region. Similarly Takahama et al. (2011) observed transpacifictransport of smoke from a volcanic eruption in Russia.3Figure 1.2: A HYSPLIT model demonstrating long range smoke transport. Adapted fromTakahama et al. (2011).1.1.2 Known Responses to AerosolsUpon arrival in different ecological settings, aerosols can have various impacts both directlyand indirectly, depending on the nature o f the environment.The radiative forcing impacts of aerosols, both direct and indirect, demonstrate significantvariability in space and time. Often, when particles are in sufficient concentrations in theatmosphere, the earth absorbs less radiation and the surface experiences cooling (Amiridiset al., 2012). This was observed in 2010 during a wildfire near Boulder, Colorado where theground beneath the plume was cooled by 2-5oC(Liu et al., 2014).Similarly, Feingold et al. (2005) found that aerosols in a daytime convective boundarylayer would warm the top of smoke/aerosol layers and affect cloud formation. In this case,the warming due to the smoke absorbing solar radiation, coupled with the simultaneouscooling of the ground and the lower atmosphere, would make the atmosphere more stableand, therefore, suppress cloud development. With less cloud formation, there is less waterinput from precipitation, leading to potential drought like conditions(Tosca et al., 2010).The net change in air temperature depends on the magnitude of the absorption and theresulting change in sensible heat flux on the ground surface(Liu et al., 2014). Additionally, a4Figure 1.3: Daily PM2.5 readings taken at Vancouver International Airport in 2017 (Source:MetroVancouver)change in atmospheric thermal structure due to the direct radiative forcing of particles willfurther change regional circulation(Booth et al., 2012).PM can be broken down into two different categories based on size: particles with a diam-eter less than 10 micrometers (PM10) and particles with a diameter less than 2.5 micrometers(PM2.5). PM2.5 is of primary concern as particles with a diameter of 5 micrometers or lesscan penetrate the lower respiratory tract and irritate the alveoli and bronchioles(Fowler ,2003). Symptoms of over-inhalation of PM include respiratory problems such as shortnessof breath, coughing, wheezing, respiratory tract inflammation and discomfort when breath-ing(Fowler , 2003). The CAAQS of 28 µgm−3 has been eclipsed regularly in the presence ofsmoke. (Figure 1.3) shows PM2.5 readings taken at Vancouver International Airport duringa wildfire event in September 2017. This period had a max 24h PM2.5 reading of 32.4 ug/m3.1.1.3 Potential Impacts in Mountainous RegionsAlmost half of the Earth’s land surface is covered by complex, uneven terrain(De Wekkerand Kossmann, 2015). Due to the variety in terrain shape, interactions with airflow are quite5complex. Similarly, due to the size of features and the complexity of the earth’s systems,their effects can be felt in far away places (De Wekker and Kossmann, 2015).Mountain Flow ProcessesThere are multiple different processes by which mountains can affect the dispersal of aerosolsand they can be separated into active and passive effects (Kossmann and Sturman, 2003;Teixeira et al., 2016). Active effects include thermally driven wind systems such as slopeflows and mountain venting that are generated by horizontal temperature and pressure dif-ferences(Zardi and Whiteman, 2013). Slope flows are caused by daily radiative warmingof mountain slopes. A warm slope surface generates winds in the upslope direction. Whenaerosols are present in the system, this process forces them up to higher elevations(De Wekkerand Kossmann, 2015). This also results in a weak return flow at higher elevations in orderto maintain continuity, resulting in polluted air being injected horizontally into elevated in-version layers. Mountain venting, also called the chimney effect, occurs when strong slopeflows and updrafts are forced through an overhead inversion. When this occurs, pollutantscan be injected into the free troposphere and possibly form elevated pollutant layers (Figure1.4) (McKendry and Lundgren, 2000; De Wekker and Kossmann, 2015).Figure 1.4: The different mountain airflow processes. E represents entrainment; MV, moun-tain venting; AV, advective venting; and MCV, mountain-cloud venting. Vectors show direc-tions of airflow while c(z) and (z) represent vertical profiles of pollutant concentration andpotential temperature respectively. Adapted from De Wekker and Kossmann (2015)6Passive effects involve momentum exchange between the surface and the atmosphere,and occur when a flow is modified by the presence of mountains(Richner and Ha¨chler ,2013). Examples include flow blocking, flow channeling, and lee waves. However, theseprocesses are not the sole determinant of how mountains affect airflow. Considerable day-to-day variability has been found in the trends of aerosol and meteorological variables at theWhistler Observatory (seen below in Figure 2.1), reflecting the impact of frequently changingsynoptic-scale weather conditions(Gallagher et al., 2011, 2012).Mountainous Boundary Layer EffectsPlanetary Boundary Layer (PBL) heights and the related vertical transport and mixing pro-cesses have recently gained attention for their role in explaining uncertainties in estimating airpollution and greenhouse gas budgets. The PBL is usually defined as the atmospheric layerthat interacts directly with the Earth’s surface. Over mountainous terrain, the atmosphericstructure becomes much more complicated(De Wekker and Kossmann, 2015). There arefour generally acknowledged separate atmospheres that pertain to mountain environments,the free, mountain, valley and slope atmospheres. (Figure 1.5).Figure 1.5: A visual representation of the different atmospheres that arise in a mountainenvironment. Adapted from De Wekker and Kossmann (2015)Relevant to pollutants, light synoptic-scale winds allow thermal flows to drive the verticaltransport of PBL air to the mountaintop level(Gallagher et al., 2011), elevating pollutants toa layer near the mountain peak(Henne et al., 2005). Understanding of such elevated layersof pollution is vital as they represent a potential sink for pollutants from the PBL, therebyeffectively ventilating the PBL and influencing tropospheric chemistry. These layers also7have the potential to be mixed down to ground and contribute to pollutant concentrationsthat influence human and vegetation health(McKendry and Lundgren, 2000).Aerosol DistributionOwing to the complex interactions mountains have with airflow, they could potentially have alarge impact on how aerosols are distributed. Similarly, aerosols could govern how mountainflow processes proceed. For example, as mountain environments are exposed to wildfiresmoke, both mechanical and thermotopographic processes could be affected. An inversioncaused by elevated aerosols could prevent most of the air from being injected into the freetroposphere and cause it to be recirculated downwards. Similarly, reduced radiation receivedat the surface would slow down slope flow processes. This could potentially change howmountain circulations evolve during smoke events.Urban Mountaintop ChemistryOther areas of interest include mountaintop level aerosol concentrations, specifically on welltraveled mountains such as Grouse. Of particular interest is mountaintop PM2.5 concen-trations during wildfire events; McKendry et al. (2011) measured CO and O3 mountaintopvalues at the Whistler Atmospheric Chemistry station during the 2008 smoke events but sofar, PM2.5 measurements are lacking. Considering there was a noticeable increase in bothCO and O3, it is likely that PM2.5 levels would increase as well.1.2 Research Objectives and Thesis StructureBased on the existing literature, there is a high likelihood that, due to thermo-topographicprocesses during wildfire events, air quality in affected mountain areas will differ greatly fromthose at lower elevation valley locations. Consequently, over the course of this dissertation Iaim to answer the following questions:• How does air quality on Grouse Mountain compare to other areas around Vancouver,and how does this change during a wildfire event? Does Grouse Mountain experiencehigh concentrations of PM during smoke events and are these concentrations highenough to deleteriously impact human health?• How do Vancouver’s local mountains affect local and regional aerosol distribution? Dothe mountains aid in the removal of PM from the valley? Does this process changeunder smoke conditions?8This study will generate new information on the concentrations of aerosols on an urbanmountaintop environment, specifically during a wildfire event. This is relevant to townssituated in mountains near urban centres such as Park City, Utah, and Lake Tahoe, Califor-nia. Similarly, the results will have implications for mountains that act as summer touristdestinations such as Whistler-Blackcomb; Yosemite, California; Mt Hood, Oregon and; ofcourse, Grouse Mountain.These results will also advance knowledge on the effects of mountains adjacent to ur-banised areas on aerosol distribution. This will help improve accuracy when modeling aerosoldistribution from sources such as forest fires, factories or dust events. This has importantimplications for public health analysts and weather predictions.With extreme events, such as the 2010 dust event and 2017 and 2018 smoke events,occurring more frequently, data derived from these events will be extremely useful whenpredicting dust or smoke distribution within a mountain PBL.The following chapters are designed around these two questions. Chapter 2 presents thestudy location and its associated climate and geography. The methods of data collectionand analysis are also presented, including the methods used for instrument calibration.Chapter 3 concentrates on the background air quality found on Grouse and how it com-pares to Vancouver proper. The spatial variability of aerosols in non-smoke conditions ispresented, sources and sinks are proposed, and concentrations are compared to an instru-ment network in lower elevation Vancouver.Chapter 4 addresses the changes that occur at Grouse and corresponding nearby areasunder the influence of wildfire emissions. The topics covered in Chapter 3 are addressedagain and a case study examining synoptic conditions during these events is presented.Chapter 5 provides a summary of the findings of the preceding two chapters, as wellas placing the work within a real-world context and suggesting possible future avenues ofresearch.9Chapter 2Experimental Methods and Analysis2.1 Description of Study Area2.1.1 Geography of the Lower Fraser ValleyThe Lower Fraser Valley (LFV) in Southwestern British Columbia, includes Vancouver, oneof Canada’s largest cities, and a variety of complex terrain (Figure 2.1). This landscapeincludes features such as coastlines, agricultural land, mountains, and urban centres. TheNorth Shore Mountains, including Grouse Mountain, Cypress Mountain, and Mount Sey-mour, form the northern boundary of the valley while the North Cascade Mountains comprisethe barrier to the south. The city of Vancouver is situated on the western edge of the valleyand is adjacent to the Strait of Georgia, a large body of water separating Vancouver’s LowerMainland and Vancouver Island. To the east, the valley is dominated by agricultural landand smaller urban centres.10Figure 2.1: Map of locations mentioned in text. Source: Google Maps2.1.2 Climate of the Lower Fraser ValleyThe climate of the region is primarily controlled by seasonal shifts in the position and strengthof the jet stream. In late fall and winter, a strong westerly jet stream results in the frequentpassage of frontal systems over the region that often bring rain to lower elevations and snowto the mountains (Oke and Hay , 1994).The presence of the ocean to the west also has a large impact on the LFV’s climate. Thislarge body of water moderates local temperatures and acts as the source of the majority ofthe precipitation that falls in this region (Mass , 2009). This temperature moderation resultsin very mild winters with temperatures rarely falling below 0oC. However, cold temperaturesdo occur when cold continental air is pushed over the Rocky Mountains (to the North-East)and undercuts a marine air-mass (Mass , 2009). Most of the annual precipitation falls duringwinter months, in part defining the LFV as a Mediterranean climate.11In the summer (the present study period), the northward migration of the jet streamresults in persistent anti-cyclonic conditions in this area. An upper-level ridge is common overthe LFV during summer months and leads to stable, fair weather conditions near the surface.Summers are typically dry in Southwestern British Columbia compared to winters with onlyabout 10% of annual precipitation occurring between June - August (EnvironmentCanada).Weather observations taken at the Vancouver International Airport (YVR) between 1981 -2010 show an average temperature of 18 oC for July and August and 14.9oC for September.The station at YVR is about 2 km from the coast and a few meters above sea-level and istherefore affected greatly by the ocean temperature moderation. Consequently, areas furtherinland from the coast experience a slightly higher average temperature during summers andhigh elevation areas experience slightly lower average temperatures.2.1.3 Mesoscale Circulations within the Lower Fraser ValleyMesoscale air circulation patterns are commonly the result of pressure gradients caused bydifferential heating of a landscape and often arise in areas of complex terrain. In Vancouver,the ocean and mountains result in the formation of two different mesoscale circulations, aland/sea breeze and mountain-valley breezes. They are ubiquitous throughout the summerand can play a significant role in the distribution of pollutants when present (McKendry andLundgren, 2000).Land/sea breeze circulations are common in the presence of a coastal environment. Wa-ter’s relatively large thermal inertia, compared to land, results in the land heating muchmore rapidly than the ocean during the day. When larger scale flow is relatively weak, ahorizontal pressure gradient can form perpendicular to the coast during the day, bringingmarine air from the cold ocean to the warm valley (Figure 2.2 number 2). These processesbecome particularly pronounced during Vancouver’s summers due to the light winds and so-lar forcing that arise from the common, clear anticyclonic conditions. Sea-breezes of around3m/s in this region have been observed extending to Abbottsford about 60% of the timewhen the breeze is present (Oke and Hay , 1994). At night, the direction of the pressuregradient switches due to the land cooling much more rapidly than the ocean. This results ina near-surface wind flowing from the valley to the coast (Figure 2.2 number 1). This breezeis normally weaker than the sea-breeze, with an average of 2 m/s (Oke and Hay , 1994).Local scale pressure gradients are also responsible for the formation of mountain-valleybreezes. As sloped terrain is heated during the day, the near-surface air is heated fasterthan the adjacent air. Again, this results in a horizontal pressure gradient that pushes windtoward the slope, where it is then angled upwards by the physical presence of the slope. This12flow is referred to as upslope or ”anabatic” wind and as discussed in the previous chapter, isthe dominant driver of mountain venting(Figure 2.2 number 5). Moreover, to compensate forthis upslope flow, air over the valley subsides and brings air from higher elevations towardsthe surface (Rampanelli and Zardi , 2004), potentially lowering pollutants into the valley(Figure 2.2 number 6). During the night, the direction of the pressure gradient switchesresulting in ‘katabatic’ or down-slope flow, reversing the circulation (Figure 2.2 number 1).Figure 2.2: Depiction of airflow and boundary layer processes over urbanized complex coastalterrain. Adapted from (McKendry and Lundgren, 2000)132.2 Instrumentation and Data AnalysisTo examine aerosol concentrations and distributions at Grouse’s peak area, several instru-ments were deployed. As this was the first study of this nature to be done at Grouse, solelythe size and concentrations of particles were observed, not the composition. A GRIMM 1.108Aerosol Spectrometer (https://www.wmo-gaw-wcc-aerosol-physics.org/files/opc-grimm-model–1.108-and-1.109.pdf) was used to observe mountain top concentrations and a Dylos DC 1700Laser Particle Counter (http://www.dylosproducts.com/dc1700.html) was used to conductwalking surveys to examine aerosol spatial variability on the summit. This section willprovide a more detailed description of the instruments used along with the data analysismethods used in subsequent chapters.2.2.1 GRIMM 1.108 Aerosol SpectrometerThe GRIMM 1.108 Aerosol Spectrometer (Figure 2.3 ) is a portable (24 x 12 x 6cm) massspectrometer. It is capable of measuring particles in the range of 0.3 to 20 micrometers indiameter and can provide PM data in either particle counts (particles per volume of air) ormass concentrations (mass per meter cubed of air). Further, the mass concentrations can beoutput as either ”environmental” which outputs PM10, PM2.5 and PM1 or as 16 particle sizeswithin the range mentioned above. For this study, we chose to use the latter as it providedthe opportunity to examine potential sources of PM.Measurements can be obtained every six seconds when connected to a computer via anRS-232C cable interface, or once every minute when run on its own. The GRIMM was notpermanently connected to the computer for this field study.The GRIMM instrument samples at a constant 1.21 l/min rate using an isokinetic pump.Particles then enter the sample cell that contains a laser diode beam and a photo diodedetector. The entering particles subsequently disrupt and scatter the path of the laser beamat angles proportional to their size. The photo diode detector recognizes these aberrationsand determines particle size using the magnitude of their scattering angle, and sends a signalto a pulse height analyser. There, the mass concentration is estimated using a fixed densityestimate, the fixed pump volume and the particle size.The GRIMM was placed in a location out of reach of foot traffic. Similarly, it washoused in a Stevenson Screen to avoid solar radiation and wind altering the sample rate andto protect it from the elements. The screen’s angled slats minimize the interior windspeedwhile still permitting airflow, and the roof and white color eliminate the possibility of airflowinduced by a forced temperature gradient.14Figure 2.3: Pictures of the GRIMM and its location on Grouse MountainCalibrationPreceding this study, the GRIMM was collocated with Metro Vancouver’s sensors that wereupgraded in 2013 to meet the U.S. Environmental Protection Agency’s PM2.5 Federal Equiva-lent Method (FEM) (Metro Vancouver , 2010) at Vancouver Airport for three days. Through-out the summer and following the study, the GRIMM was brought down to run next to MetroVancouver’s FEM sensor for 30 to 60 minutes to ensure that the GRIMM was still reportingaccurately. A linear regression was performed on over 200 observations taken using the twoinstruments when collocated, the resulting R2 value was 0.9917 (Figure 2.4)2.2.2 MetroVancouver’s Instrument NetworkThe LFV Air Quality Monitoring Network includes 29 air quality monitoring stations lo-cated from Horseshoe Bay in West Vancouver to Hope.(Metro Vancouver , 2010) Air qualityand weather data from all but one station are collected automatically on a continuous ba-sis, transmitted to Metro Vancouver’s Head Office in Burnaby, and stored in an electronicdatabase. The station used most commonly in this study is found at Mahon Park in NorthVancouver and is located approximately 7km south of Grouse’s summit (Figure 2.1)and thusprovides reliable information about the The federal CAAQS 24-hour PM2.5 standard of 28µg/m3 was implemented in 2015. British Columbia’s objective is 25 µg/m3 (Metro Vancou-ver , 2010). A standard or objective is achieved if ambient concentrations are at or lowerthan the stated objective concentrations(Metro Vancouver , 2010).15Figure 2.4: Linear regression done on the readings taken from the collocated GRIM andMetroVancouver FEM instruments.162.2.3 DylosA low-cost Dylos DC1700 Pro air quality monitor was used to determine aerosol distributionon the summit area. Although there are studies (Semple et al., 2015; Steinle et al., 2015;Manikonda et al., 2016)) describing the use of the Dylos as an effective instrument formeasuring PM2.5, this portion of the study focused on relative differences/magnitudes inaerosol concentrations and not exact values, calibration of the Dylos with MetroVancouver’sFEM instruments was not undertaken.The Dylos is a small, very portable laser particle counter that contains a small, very quietfan that channels air through the measurement chamber. Like the Grimm, measurementsare taken using a laser diode and a photo diode detector. Unlike the Grimm, it only logsparticles in two size classes: 0.5-2.5 µm ”small” and 2.5 µm ”large”. In this study, theseparticle counts are then corrected using the conversion developed by Steinle et al. (2015)seen in eq 2.1. In their study, which examined calibration methods for the Dylos and theviability of using it as low-cost PM2.5 sensor, determined that this linear equation was foundto result in the best calibration in an urban outdoor environment.PM2.5 = 4.75 + 2.8 × 10−5 × Dylos Small count (2.1)On a full battery charge, the Dylos runs for approximately 6 hours. The built-in memorycan store approximately one week of data when sampling continuously. This means that theDylos can be used for multiple walking surveys before being brought to a PC where the datacan then be downloaded as a text file for further analysis. This transfer requires a 9pin serialcable or USB-to-COM adapter. To carry the Dylos and to ensure consistent sampling, itwas placed in the outer pocket of a small hiking backpack and secured with zip-ties, similarto the method used in Steinle et al. (2015). Since the the Dylos is not water-proof, yet needsto be exposed to ambient air, it could only be deployed in dry conditions.The chosen survey path (see Figure 2.6) accounted for all the main features of Grouse’ssummit. Beginning at the exit of Grouse’s gondola, it passes the main lodge and followsthe route suggested for Grouse Mountain tourists. The densely developed main area, theopen plain, the bear habitat and the peak area/chairlift are all accounted for. The loop wasapproximately 1.5 kilometers and was retraced in each instance to increase accuracy.17Figure 2.5: Pictures of the Dylos and the manner in which it was carried by the surveyor.GarminA Global Positioning System (GPS) receiver was used in combination with the Dylos torelate observed particle concentrations to time and location. The Garmin Edge 500 GPSwas selected for this study because of its small form factor (4 × 7 × 2 cm), low weight(60g) and the ease of data transfer from the device to the computer in usable formats. TheGarmin recorded date, time, altitude, longitude and latitude approximately every 2 seconds,depending on signal quality.The path followed during mobile surveys using the Dylos monitor is shown in Figure 2.6.Each loop was performed mid morning on non-rainy days and took approximately 25 minutes.This path followed the peak’s walking trail, passing all of the area’s main attractions. Uponcompletion of a loop, a subsequent loop was performed in reverse to reduce the risk ofcontamination by external factors as well as to exclude the potential for temporal changesin concentration during the survey. Areas that were left out of the survey path included themountain bike trails and rental facility as well as the Grouse Mountain staff office.18Figure 2.6: The path followed during the walking surveys and different locations mentionedin the text2.2.4 LiDARRemote sensing of aerosols is often used to monitor the location, transport and elevationof aerosol layers. McKendry et al. (2010),McKendry et al. (2011), Cottle et al. (2014) andTeakles et al. (2017) have used various forms of LiDAR to study wildfire plumes in BritishColumbia. Similarly, Akaoka et al. (2017) used LiDAR to examine emissions of coal carryingtrains in South Delta, BC and Cottle et al. (2013) used it to document dust transport fromAsia. For this project, a mini micro-pulse LIDAR (mMPL) was set up two kilometers fromthe base of Grouse Mountain (Figure 2.6) for the duration of the summer and was used toexamine the vertical structure of pollutants above and around Grouse Mountain.19The mMPL is a cost-effective and portable alternative to most LiDAR systems. It con-tinuously emits thousands of low-energy pulses at 532 nm and integrates them into a singleprofile. Each pulse has an energy of 4µJ which is several orders of magnitude less thanlarger, non-portable systems. To make up for the weaker signal, the mMPL has a widerbeam diameter of 7.62 cm and a much higher pulse frequency of 4 KHz. Owing to this,the system is rated eye-safe at 3.5 m and is therefore less restricted by safety regulations.The mMPL is also housed in a steel, air-conditioned enclosure that allows it to be run inunfavorable conditions.Due to incomplete overlap of the emitted beam and the receiver’s field of view, the first150 m above the LiDAR cannot be profiled. However, the mMPL can obtain a profile thefollowing 12 km overhead. The time of integration can be set anywhere between the range1 s - 60 min and the vertical resolution can be set to 5, 15, 30, or 75 m. For this study, anaveraging time of 5 min and vertical resolution of 30 m was chosen.The main output of the mMPL in this study is Normalized Relative Backscatter (NRB).NRB is a range corrected product that is derived from the standard LiDAR equation (2.2).S(z, λ) = Eo(λ)Az2γO(z)βpi(z, λ) exp[−2∫ z0α (z′, λ) dz′]+ SB (2.2)Where S(z, λ) is the signal received at the detector (counts/µs), E0 is the energy of laserpulse (J), A is usable area of the receiver (km2), z is height (km), γ is the full systemefficiency constant ( counts−µs∗µJ ). O(z) is the incomplete overlap correction function (unitless),βpi(z, λ) is volume backscatter coefficient (km2sr−1), α is the extinction coefficient [unitless],and SB is the background signal (counts µs−1). The signal in eq 2.2 is then corrected in eq2.3 to arrive at NRB.NRB(z) =S(z) ∗ d(S) − a(z) − SB(z)E0 ∗O(z) ∗ z2 (2.3)Where d(S) represents the deadtime correction factor and a(z) is the afterpulse correctionfactor.20Figure 2.7: The location of the LiDAR with respect to Grouse Mountain.Dispersion ModellingIn the interest of determining the source of the aerosols affecting Grouse, trajectory modelingwas utilized. Lagrangian particle dispersion models are often used to track emissions ordetermine their sources (Figure 1.2). The National Oceanic and Atmospheric Association(NOAA)’s Hybrid Single Particle Langrangian Integrated Trajectory (HYSPLIT) model hasbeen used to great effect in western North America (McKendry et al., 2011) and was used inthis study to determine the origin of air parcels on Grouse for the entire summer. The highresolution EDAS 40km meteorology dataset was used to create back trajectories startingfrom Grouse Peak at 0, 100 and 500m above ground level.WindsondsWindsonds (http://windsond.com/) consist of a miniature host of sensors that are deployedfrom the ground upwards through the atmosphere by way of a helium filled balloon. TheWindsond measures air temperature, relative humidity, pressure, wind speed, wind direction,as well as altitude and GPS location.21Figure 2.8: An overview of a windsonde launch. Retrieved from www.windsond.com2019/08/15This allows the user to create an atmospheric profile of the variables of interest. Themaximum altitude the Windsond can reach is dependent on the initial volume of the balloonas well as air pressure at upper-levels. In this case, the sensor was placed in a 12 g Styrofoamcup and attached to a 30-litre helium balloon. The Windsond collects data at three-secondintervals and generally has an ascent rate between one to two meters per second, resultingin an average vertical resolution of about three to four meters. In this study, the ascent ratewas much slower, likely due to insufficient inflation of the balloons. These balloons reachedat least two km above ground level (AGL) for this study. The temperature sensor has anaccuracy of 0.3°C.222.2.5 Data Analysis MethodsThis thesis is an observational study using data from in-situ measurements. Spatial andtemporal variations in PM concentrations on Grouse Mountain are analyzed to better un-derstand the variations in PM levels in this area. The focus of this study is to understandhow PM concentrations on Grouse compare to those in Vancouver proper, especially duringwildfire smoke events. Therefore, the following chapters will compare the differences be-tween values observed by Metro Vancouver’s air quality sensor network and those on top ofGrouse during both clear and smoke conditions. Further, the recorded concentrations willbe compared to the National and Provincial Air Quality Standards.Table 2.1: Summary of Field Data CollectionClear Days Smoke DaysDylos Surveys 9 7Grimm Deployments (days) 13 6Windsond Flights (days) 13 20Lidar Deployment June 26th - September 7th23Chapter 3Non-Smoke Particulate MatterConcentrations and DistributionIn this chapter, spatial and temporal variations in PM on Grouse Mountain under non-smokeconditions during the summer of 2018 are investigated. For the purposes of this analysis,any day that NOAA’s satellite observation based Hazard Mapping System (HMS) Fire andSmoke Product imagery (https://www.ospo.noaa.gov/Products/land/hms.html) displayed asmoke plume that covered the LFV constituted a smoke day and all others were considered”non-smoke” days. Accordingly, even though the last week of July registered very high PMvalues, those days were considered non-smoke days because there were no plumes detectedby satellite imagery in the Vancouver area. Analysis is performed using data collected fromthe GRIMM, Dylos and LiDAR on non-smoke days. Observations are compared to the thoseregistered at Metro Vancouver’s Mahon Park station and then averaged over a 24h periodwhen possible, to be compared to the CAAQS.In terms of concentrations, based on existing literature (Monn et al., 1997),we expectto see relatively low concentrations at Grouse’s peak under non-smoke conditions with verylittle spatial and temporal variability. Although Grouse’s helicopter will affect the spatialvariation on some surveys, it is expected that these effects will be too rare to impact theoverall study. However, the chairlift that takes visitors to Grouse’s actual peak may well havelower concentrations due to its separation from dust kicked up on the established walkingpaths.As Grouse’s peak is situated above the Lower Fraser Valley’s (including Vancouver)boundary layer, pollutants from the city and valley are not expected to mix high enoughto influence concentrations. However, mountain venting processes such as those describedin the introduction, could lead to enhanced PM values around the summit (De Wekker andKossmann, 2015). This is particularly possible during the cloudless summer days that are24typical in Vancouver’s climate. With enhanced heating of mountain slopes, the potential formountain venting is increased. On a smaller temporal and spatial scale, potential sourcesof PM such as the main lodge and the helicopter pad, exist at the peak and could lead tospikes in concentration.3.1 PM2.5 Spatial VariabilityExample results from these loops are contained in Figure 3.1. Each circle represents amoment when both the GPS and the Dylos recorded a measurement simultaneously. Thesize and color of each circle represent the relative magnitude of the value recorded. Asmentioned earlier, we cannot be completely confident in the absolute reading of the Dylos butwe will use them to approximate relative concentration magnitudes and we will be comparingreadings to each other to get an idea of spatial variability. Immediately noticeable amongstthese figures are small spikes in the otherwise consistent concentration values. These spikesappear near the main lodge and bear refuge in most (but not all) of the survey results. Theseareas are arguably the busiest areas of the summit and thus, these readings are consistentwith expectations.25Figure 3.1: Figure3.1: Relative PM2.5 concentrations observed around Grouse Mountain’speak area on July 14th, 2018.Figure 3.2 is a hexbox summary plot of the nine clear day surveys done during the summerof 2018. Each hexagon in this plot takes the average of every measurement found withinits range. Since the path was retraced in every survey, each hexagon should be the averageof approximately 18 measurements. Due to variation in the GPS’s recording frequency,this number could vary. The resulting figure illustrates the average of all the concentrationpaths. It is shown here that there is a consistent difference between the bear refuge area andthe rest of the mountain path. The rest of the peak area demonstrates consistently lowerconcentrations, especially in the area accessed by the peak chairlift. These local maxima arelikely due to dust kicked up by local foot traffic. However, in the subsequent sections, weshow that although the bear refuge and main lodge show high concentrations relative to therest of the peak area, the concentrations are still not high enough to be of concern.26Over the nine surveys, the standard deviation was 4.1 µgm−3 (90% of the mean of 4.7) andthe max recorded was 16 µgm−3. Since the surveys were taken at the same time every dayunder similar synoptic conditions, there was likely little effect due to differences in boundarylayer growth. The observed mean was very similar to the average reported by the Grimmfor the entire summer and was lower than the 7.4 µgm−3reported by MetroVancouver atMahon Park. Interestingly, the maximum value was higher than the 10.4 µgm−3 maximumobserved at Mahon Park. Since the highest value was recorded near the helicopter pad, thiswas likely observed during a helicopter takeoff.Figure 3.2: Averaged relative PM2.5 concentrations observed around Grouse Mountain’s peakarea for the nine surveys done on clear days during the summer of 2018.3.2 LiDAR ImageryGiven that aerosols emanating from the surface are normally contained within the boundarylayer, and that LiDAR backscatter values give a relative representation of aerosol concen-trations, the LiDAR backscatter values can be used here to approximate boundary layergrowth and height (Hennemuth and Lammert , 2006; Manninen et al., 2018; De ArrudaMoreira et al., 2019).27Similarly, due to the LiDAR’s proximity to the base of Grouse, the overlying boundarylayer is affected by Grouse’s elevation and is therefore likely much higher than that observedover the Lower Fraser Valley. Because of this, it is possible to examine the extent to whichaerosols are mixed as far up as Grouse’s peak elevation. Examples of the resulting imageryare found in Figure 3.3 and Figure 3.4.These figures display the change in aerosol backscatter in the atmosphere directly abovethe LiDAR over time. As the LiDAR was set up below our area of interest, there is a redline accompanying the graph that represents Grouse peak’s elevation.In Figure 3.3, the first three days were clear and demonstrated consistent diurnal bound-ary layer growth to above 3km (shown by the height of the green columns).Figure 3.3: Ground based, upward facing LiDAR measurements taken from near the base ofGrouse Mountain, from June 16th to 21st, 2018.The aerosol backscatter values in Figure 3.3 show evidence of consistent diurnal convectiveboundary layer growth under clear sky, anti-cyclonic conditions. This is shown by the relativeabsence of aerosol backscatter at 500-2000m of elevation during the night and an increasein concentrations at those heights during the day. Interestingly, on June 19th, it looks asthough there are two layers that have formed, one that extends to 1200m in height (A)and one that extends to almost 4000m in height (B). The former is likely evidence of theclassic convective boundary layer found over flat ground (Stull , 1988) and the latter couldpotentially be the mountain boundary layer that was mentioned earlier and first proposedby De Wekker and Kossmann (2015).28Most importantly, Figure 3.3 show that the convective boundary layer grows to over1000m, or at least there is evidence of increased aerosol concentrations at that height, demon-strating that aerosols generated in Vancouver proper are capable of being mixed upwards toGrouse’s peak elevation.Under prolonged hot, anti-cyclonic conditions, with very little nighttime cooling, as seenJuly 25-30 2018 (Figure3.4) the concentrations at 1000m persist for the entire duration,with no nighttime decline. This is likely due to the prolonged warming, but also due to thelow winds associated with high pressure conditions. Figure 3.5 displays the abnormally highdaytime and nighttime temperatures of that week, along with the relatively low wind speeds.Similarly, the smooth sinusoidal incoming radiation values are indicative of a prolonged highpressure system with little to no clouds. With little wind to expel anthropogenic aerosols,they accumulate in the surrounding area. Owing to this, under these conditions, aboveaverage aerosol concentrations are observed at Grouse’s peak (Figure 3.8) and even up to2000m. It is important to note, however, that under these conditions, air quality is stillworse at the surface (shown by the yellow coloring).Figure 3.4: Ground based, upward facing LiDAR measurements taken from near the base ofGrouse Mountain, from July 23rd to 29th, 2018. The arrow in this figure points to what arepotentially remnants of wildfire emissions from Siberia and Alaska.29Figure 3.5: Incoming radiation, ozone, PM, temperature and windspeed measurementsrecorded at Mahon Park, North Vancouver during the ”photochemical smog event”.30Of note, the high concentrations observed in this weeklong event were reported by CBC(2018) to have been affected by a local bog fire and remnants of wildfire emissions fromAlaska and Siberia, the latter explaining the long descending line of aerosols shown by thearrow in Figure 3.4. Figure 3.6illustrates a Hybrid Single Particle Lagrangian IntegratedTrajectory (HYSPLIT) model run for those days, showing the origins and paths of differentair parcels that reached Grouse during this period. This model was run using GDAS 0.5degree meteorological inputs. The origin of the back trajectories is Grouse Mountain andthe models were run for 300 hours. Back-trajectories were calculated to determine where airparcels at 1000m, 3500m and 4000m above sea level originated. These heights where chosento correspond with the top of Grouse, the lower extent of the aerosols seen descending towardsthe LiDAR on July 26th in Figure 3.4 and the upper extent of the same mass of aerosols,respectively. The two higher trajectories are shown to have passed through Russia and allpaths are shown passing through Alaska, further substantiating the wildfire provenance ofthe event.Figure 3.6: 300h HYSPLIT backwards dispersion trajectories run from Grouse MountainJuly 26th3.3 GRIMM PM2.5 ConcentrationsFigure 3.7 shows the minute averaged concentrations recorded by the GRIMM on GrouseMountain on two ”non-smoke” days. To reiterate, as seen in Figure 2.2the observationsrecorded by the GRIMM were found to be nearly identical to those recorded by MetroVan-couver’s FEM instruments when co-located. Also included on the figures are a line indicatingthe minute averaged concentrations taken simultaneously by MetroVancouver’s instruments.Again, it is important to reiterate, that due to concerns regarding the effect of moisture on31the GRIMM’s internal mechanisms, this instrument was only used on low humidity days.Figure 3.7: PM2.5 concentrations observed on Grouse Mountain and at North Vancouver’sMahon Park on July 17th and 23rd, 2018.On clear days, the concentrations recorded by the GRIMM were always lower than thoserecorded at Mahon Park. Under these conditions, the average observed hour-long PM2.5concentration was 4.8µgm−3 with a maximum of 7.4 µgm−3 whereas the average seen atMahon Park was 7.6µgm−3 with a max of 12.3µgm−3. The many small spikes seen in therecordings taken at Grouse are potentially due to the helicopter that would take off and landa few hundred meters from the GRIMM instrument. Neither location registered an hour-longperiod where the average measurement was close to the one hour CAAQS of 28µgm−3 underclear conditions.Further, likely due to the decline of the boundary layer and the lack of tourist activity,values recorded at Grouse tend to drop at night and rise gradually throughout the day. Thisagrees with the conclusions drawn from the LiDAR plots in the previous section. The average”non-smoke” day minute concentration from 6pm to 6am was found to be 3.25µgm−3whilethe average between 6am and 6pm was 6.41µgm−3. In the urban area, at Mahon Park, therewas an increase during the day as well, but it was much less pronounced. It changed from anaverage of 6.9µgm−3 to 8.3µgm−3 during the day. Not included in the results above was the”photochemical smog event” that occurred under the persistent high-pressure system fromJuly 25-30. This event provided the only instance where PM concentrations were higher onGrouse during ”non-smoke” conditions.32Figure 3.8: PM2.5 concentrations observed on Grouse Mountain and at North Vancouver’sMahon Park on July 28th and 29th, 2018, during the “photochemical smog event”..Over this period, the recorded PM2.5 concentrations at each location were substantiallycloser relative to a normal clear day. This can be see in Figure 3.8and in the recordedaverage minute concentrations for this period. Measurements on Grouse showed an averageof 13.4µgm−3 and Mahon Park reported an average of 14.1µgm−3 over this event.Similarly, as opposed to the other clear days, PM2.5 concentrations revealed very littlediurnal variability during this event. The nighttime averages at Mahon Park and Grouse peakwere 13.9µgm−3 and 13.2 µgm−3 respectively while the daytime averages were 14.3µgm−3and 13.6 µgm−3 respectively.These observations align well with the LiDAR imagery seen in the previous section thatshows consistent aerosol distribution with height during this event. This in turn, strengthensthe case for the prolonged high-pressure system and associated high nighttime temperaturesleading to a persistent convective boundary layer even through the night.However, even in this week of poor air quality, neither location recorded an hour thataveraged PM2.5 over 28 µgm−3. It should be noted that, Mahon Park came close with amaximum hour long average of 22.1 µgm−3 while Grouse peak had a max of 19.6 µgm−3.3.4 ConclusionIn this chapter, the spatial and temporal variations in PM on Grouse Mountain under non-smoke conditions were analyzed. Using the GRIMM, Dylos and GPS it was shown thatunder normal, clear sky conditions, Grouse’s peak area has very little spatial and temporalvariability in PM concentrations. This was further corroborated using ground based LiDARobservations that show very little variation in PM concentrations at Grouse’s elevation onclear days. As predicted, the concentrations observed on these days were very low, were33lower than those recorded in the valley at Mahon Park and show no risk of being harmful tovisitor’s health.During 2018’s photochemical smog event, however, the prolonged high-pressure condi-tions caused PM to build up in the lower atmosphere (a ”smog” event). This resulted in anincrease in PM levels and a decrease in the difference between concentration levels on Grouseand those in the valley. Again, the LiDAR imagery reinforced this point by displaying higher,more uniform PM concentrations with height in the lower atmosphere.Finally, variations in boundary layer growth around Grouse were also observed. Underclear sky conditions, evidence of the mountain boundary layer proposed by De Wekker andKossmann (2015) was observed and is shown in Figure 3.4. This is an interesting finding interms of the meteorology of the area and further research could be very beneficial. Boundarylayer depth plays a key role in weather modelling and air pollution movement.34Chapter 4Smoke Event Particulate MatterConcentrations and DistributionThis chapter will investigate the effect wildfire smoke has on PM concentration and distri-bution around Grouse Mountain. NOAA’s HMS imagery was used again here to determinewhether smoke was present in the LFV. All days that displayed smoke in the LFV wereconsidered ”smoke days”. Figure 4.1 shows an example of what constitutes a smoke day inthis study. In the image on the left, obtained from NOAA’s HMS Imagery, the LFV (circledin blue) is covered in smoke. On the right, NOAA’s MODIS TERRA satellite shows a plumeof smoke covering the LFVFigure 4.1: NOAA’s HMS Imagery(left) and MODIS TERRA satellite imagery (right) onAugust 13th 201835This analysis is performed using the data collected from the GRIMM, Dylos and LiDARon smoke days. Data was collected using the GRIMM and Dylos on seven different occasionswhile the LiDAR was run for the entire duration of the fire season. Unfortunately, towardsthe end of the summer’s main fire event, the GRIMM filter became clogged with PM, andthe subsequent measurements were unusable. The usable observations are compared to thoseregistered at Metro Vancouver’s Mahon Park station and then averaged over a 24h period,when possible, to compare to the CAAQS.Due to the limited number of studies of this nature, information regarding wildfire smokePM concentrations around mountains is limited. However, with an understanding of generalmountain flow processes, and results from similar studies over flat terrain, general effects canbe proposed.In studies by (Peshev et al., 2017) and (Ansmann et al., 2018), it is shown that, over flatground, the most concentrated portion of a smoke plume is often found toward the top of theboundary layer and far from the surface. In Chapter 1, basic mountain flow processes areoutlined and show that mountains facilitate the transport of air parcels from lower elevationsto higher ones. It is possible that PM concentrations at the top of Grouse Mountain willbe higher than those in the LFV. If Grouse is contained solely within a mountain boundarylayer, the concentrations will not be substantially higher, but if the peak is affected by theLFV’s boundary layer, concentrations will likely be much higher.4.1 PM2.5 Spatial VariabilityThe path followed during these surveys is the same as that is shown in Figure 3.1. Figure4.2 shows a sample survey taken under smoke conditions, on August 17th, when smoke fromcentral British Columbia was starting to arrive in the LFV.Relative to Figure 3.2, the concentrations recorded in the presence of smoke are larger inmagnitude but follow a similar spatial distribution. The highest recorded values on this dayare located around the bear refuge and the area’s main attractions. The difference in valuesbetween the smoke and non smoke days will be covered later in this chapter.36Figure 4.2: Sample survey of relative PM2.5 concentrations observed around Grouse Moun-tain’s peak area in the presence of wildfire smoke on August 17th, 2018.Figure 4.3 is a hexbox summary plot of the seven smoke day surveys done during thesummer of 2018. Each hexagon in this plot takes the average of every measurement foundwithin its range. Since the path was retraced in every survey, each hexagon represents theaverage of approximately 14 measurements. The resulting figure illustrates the average ofall the concentration paths.The main difference between the figure produced in the presence of smoke and the oneproduced in clear conditions is the magnitude of the observations. The maximum valuereported in one of the hexagons in absence of smoke was over seven whereas in the presenceof smoke, the highest value was over 80 µgm−3. Later in this chapter, the absolute valuesobserved on Grouse Mountain will be reported and discussed.37Because of the difference in magnitude of the values in Figure 4.2 and Figure 3.3, thisdistribution appears to be more uniform. However, the magnitude of the relative differencebetween locations is similar to those observed on clear days with a standard deviation of28 µgm−3 (90% of the mean). Again, the observations in the area at the top of the peakchairlift (the northernmost point) are still lower, meaning that even in the presence of wildfiresmoke, separation from human activities is still an important factor or that the smoke isbeing contained in this area for another reason. As the surveys were taken at the same timeevery day under similar synoptic conditions, there was likely little effect due to differencesin boundary layer growth.Figure 4.3: Averaged relative PM2.5 concentrations observed around Grouse Mountain’s peakarea for the nine surveys done on smoke days during the summer of 2018.4.2 LiDAR ImageryAs in the previous chapter, LiDAR imagery is used to approximate aerosol concentrations inthe atmosphere. In the case of wildfire emissions, the vertical extent to which the smoke isspread and how that changes over time is of interest, as are the concentrations and how theycompare to the smoke free days. An example of the resulting imagery is found in Figure 4.4.Figure 4.4 displays the week of August 13th 2018. This week began with light smoke in38the Valley from earlier fires. Smoke persisted for most of the week (shown by the consistentcolors in the imagery) before the large smoke plume arrived in the Valley (shown by thearrow). This smoke plume resulted in the LFV having some of the worst air quality in theworld over the weekend of the 19th and the following days.Aersol backscatter values in Figure 4.4 show very little evidence of diurnal variability.When compared to (Figure 3.3), the backscatter values remain very consistent over timeand do not drop over night. The vertical extent of the aerosols did not seem to changeover the week, with high concentrations not extending past 2000m. For most of the week,highest backscatter (and likely highest concentrations) appeared to be below the elevationof Grouse’s peak (shown by the red dashed line) with the brighter green colors containedmostly between that height and the surface. However, towards the end of the 18th and earlyon the 19th the values at 1000m were much higher than those on the surface. This will bediscussed in more detail later.Figure 4.4: Ground based, upward facing LiDAR measurements taken from near the baseof Grouse Mountain, from August 13th to 19th, 2018. red arrow points to the main plumeof wildfire smoke that affected the LFV in 2018. The red dashed line shows the elevation ofGrouse Mountain. The white brackets are used to show smaller plumes.As for clear days (Chapter 3), there is a defined boundary layer limit at around 1000mof elevation. It is shown by the green colored portion of the plot that is contained below the1000m line. There is also evidence of a second boundary layer that extends to around 2000mand is shown by the blue colored portion of the plot. This second boundary layer is likely moreevidence of the mountain boundary layer as described by De Wekker and Kossmann (2015).This is especially interesting because this boundary layer has been previously undocumentedin the presence of smoke. This could simply be an artifact of smoke’s vertical variability inthe atmosphere or that mountain flow processes continue to function, at least in part, in the39presence of wildfire smoke. The latter implies that the reduced radiative impact of smokewould not completely shut down active flow processes. A better understanding of this isrequired as both phenomena would affect how smoke dispersion and weather, in general, aremodeled.Contrary to what was seen in the photochemical smog event, with smoke in the atmo-sphere, there was little to no enhanced boundary layer growth. Throughout this event, thelower boundary layer extends close to 1000m, which is similar to what is seen in Figure 3.3,suggesting that, in this particular case, the reduced radiative input at the surface resulted inless convective turbulence and therefore less boundary layer growth (Talukdar et al., 2019).Potential temperature and its lapse rate is often used to determine the stability of thelower atmosphere. Figure 4.5 displays the average of vertical profiles of potential temperaturetaken from the base of Grouse Mountain. Profiles from 13 sondes launched on clear days areshown in blue while the 20 sondes launched on smoke days are shown in black. This showsthat on smoke days potential temperature on average increases more with height than onclear days (this is indicative of a more stable environment). These conditions are much lesssuited to boundary layer growth and are likely to reduce the venting effects of the mountainsand lead to a build up of particles in the atmosphere.40Figure 4.5: Average vertical profiles of potential temperature taken from the base of GrouseMountain. Profiles from clear days are shown in blue while smoke days are shown in black.Figure 4.6 is another LiDAR plot to help illustrate this phenomenon. This shows verticalprofiles taken during smoke days in 2018 superimposed on the contemporaneous LiDARimagery. Here, sudden changes in potential temperature at or near the top of the smokeplumes are evidence of very stable, capping inversions. These inversions occur at or within200m of Grouse’s peak height (shown by the red line), meaning that they would likelysuppress venting mechanisms normally associated with the mountain.Figure 4.7 and Figure 4.8 depict HYSPLIT trajectories that were run using EDAS 40kmmeteorological inputs. The origin of the back trajectories is Grouse Mountain and the modelswere run for 48 hours starting at 12pm on August 13th , 14th and 15th (Figure 4.7) and16th, 17th and 18th (Figure 4.8). Notably, the days that appear to have smoke at Grouse’selevation in Figure 4.4 (Aug 13, 14 and 16) have HYSPLIT trajectories pass through thenorthern mountains, whereas the days where there is no smoke at Grouse’s elevation, haveHYSPLIT trajectories that come off the water.41Figure 4.6: Vertical profiles superimposed on the earlier LiDAR imageryFigure 4.7: HYSPLIT backwards dispersion trajectories that were run. The origin of theback trajectories is Grouse Mountain and the models were run for 48 hours starting at 12pmon August 13th, 14th and 15th42Figure 4.8: HYSPLIT backwards dispersion trajectories that were run. The origin of theback trajectories is Grouse Mountain and the models were run for 48 hours starting at 12pmon August 16th, 17th and 18th4.3 GRIMM PM2.5 ConcentrationsFigure 4.9 shows a subset of the minute averaged concentrations recorded by the GRIMMon Grouse Mountain on smoke days. Also included on each figure is a line indicating theconcentrations taken simultaneously by MetroVancouver’s FEM instruments.43Figure 4.9: PM2.5 concentrations observed on Grouse Mountain and at North Vancouver’sMahon Park on August 13th during the day, the afternoon of August 15th and August 16th,2018.In the presence of smoke, PM2.5 concentrations recorded by the GRIMM were oftenhigher than those recorded at Mahon Park. In Figure 4.9, the HYSPLIT model runs showthat these smoke plumes passed from north or northeast of the city before arriving at GrouseMountain. It is likely then, that these plumes were contained in the mountain boundary layerbefore arriving in the LFV and subsequently entrained downwards into the valley’s boundarylayer. This would result in the GRIMM recording the higher concentrations first at Grouse’speak, and then Mahon Park recording these observations later on once the smoke arrivedlower in the valley. This is supported by the LiDAR imagery that shows, for these days,smoke plumes that begin above the altitude of Grouse Mountain when the concentrationsare higher at that elevation and are then shown to be affecting lower elevations when theconcentrations are similar between the two elevations.(shown by the white brackets on Figure4.4. As this happens, since the stability has likely increased (Wang and Christopher , 2006)and active flow processes have been reduced (Calvo et al., 2010), less venting of smoke islikely, meaning that the air quality will worsen near the surface rather than what happenedduring the photochemical smog event.44Correspondingly, on the 16th, when the GRIMM and Mahon Park’s readings lined up,the air that was shown to have arrived at Grouse that day did not come from the North orNortheast. Rather, it came from the Southwest and therefore was likely not carrying smokeor the smoke had already descended into the valley boundary layer.This implies that, depending on the synoptic conditions, air quality on Grouse will worsenbefore the air quality in the LFV. Under these conditions, the average observed hour-longPM2.5 concentration was 56.3µgm−3 with a maximum of 180.2µgm−3, whereas the averageseen at Mahon Park was 32.7µgm−3 with a max of 114. µgm−3. Both locations registeredmultiple 24-hour long periods that exceeded the one-hour CAAQS of 28µgm−3.Similar to observations during the photochemical smog event, lack of diurnal variabilityin the boundary layer resulted in the concentrations recorded at Grouse remaining constantthroughout the day and night. This agrees with the conclusions drawn from the LiDAR plotsin the previous section. The average ”non-smoke” day minute concentration from 6pm to6am was found to be 58.1µg/m3 while the average between 6am and 6pm was 54.5µgm−3. Inthe urban area, at Mahon Park, observations on these days also exhibited very little diurnalvariability. Concentrations changed from an average of 30.9µgm−3 at night to 34.5µgm−3during the day.4.4 Case Study on August Fire EventBetween August 13th and August 25th, wildfire emissions were prevalent in Vancouver andthe LFV. Aerosol Optical Depth (AOD) values, which are a measure of light attenuationthrough the atmosphere and under clear skies, average around 0.1 (Mckendry , 2018) butduring the smoke event, they were regularly over two and surpassed four on many occasions.Figure 4.10 shows the AOD values recorded just southwest of Vancouver for the month ofAugust 2018. The first ten days of the month recorded normal AOD values for this areawhile August 11th to 25th display values consistently over 1. The different colored linesdepict the different aerosol optical depths for each wavelength. i.e AOD 500 is the aerosoloptical depth of light with a wavelength of 500nm.45Figure 4.10: Aerosol optical depth values measured at Saturna Island, just southwest ofVancouver over the course of August 2018.46Figure 4.11: View of Cypress and Grouse Mountain from Jericho Beach on a clear day (left)and August 20th (right).LiDAR imagery in Figure 4.12 displays the vertical extent of aerosols over the course ofthis wildfire smoke event. As mentioned earlier, the red arrow points to the main plume ofsmoke that affected Vancouver. This plume is also the most interesting from this study’sperspective because it very clearly begins above the height of Grouse mountain and the,from the perspective of the Lidar, descends downwards over the course of two days. It is alsopossible that this was due to the inherent vertical variability of air masses. This observationhas served to inform the basis of this case study designed to look at all possible observationsto better understand what happened.Figure 4.12: Ground based, upward facing LiDAR measurements taken from near the base ofGrouse Mountain, from August 13th to 25th, 2018. The red arrow points to the main plumeof wildfire smoke that affected the LFV in 2018. The red dashed line shows the elevation ofGrouse Mountain.47Figure 4.13 displays ozone, PM, windspeed, incoming radiation and temperature overthe course of this event. When compared to readings from the photochemical smog event inFigure 3.6 the relative difference in PM and Ozone is substantial. In the presence of wildfiresmoke, the average PM reading was 48.3µgm−3 and the average Ozone reading was 15.3 ppb.While during the photochemical smog event, the average PM reading was 11.4µgm−3 andthe average Ozone reading was 25.6ppb.Also of note are the abnormally low surface temperature and incoming radiation values,especially when compared to Figure 3.6. This meant that the LFV was not receiving asmuch incoming radiation as normal, nor as much as was forecast. Therefore, most thermo-topographic processes were likely not functioning as they normally do (Mckendry , 2018).Conceivably, this could have affected mountain flow processes as radiation plays a key rolein the active processes. This will be touched on later.48Figure 4.13: Ozone, PM, incoming radiation, temperature and windspeed measurementsrecorded at Mahon Park, North Vancouver during 2018’s main fire event.Figure 4.14 depicts a HYSPLIT dispersion model run for those days, demonstratingthe origins and paths of different air parcels that reached Grouse during this period. Thedispersion model was run backwards to see where air parcels at 500m, 1000m and 2500moriginated. These heights were chosen to correspond with midway up Grouse, and the lowerand higher extents of aerosols seen descending towards the LiDAR on August 18th in Figure4.11, respectively.The largest plume, whose path is best shown by the green line in Figure 4.14, is shownto be part of an air parcel that passed east over the mountains to the north of Vancouvertowards interior BC where most of the summer’s fires occurred (shown in Figure 4.16).49Following this, the parcel turned west and approached Vancouver from the northeast via theCoast Mountains. This meant that the parcel needed to be similar to or above the heightof the mountains when it arrived at Grouse, which is what is shown to have happened inFigure 4.11.Figure 4.14: HYSPLIT backwards dispersion trajectories that were run using EDAS 40kmmeteorological inputs. The origin of the back trajectories is Grouse Mountain and the modelswere run for 300 hours starting at 12am on August 19th.50Figure 4.15: A map showing the composite mean pressure using isobars along the west coastof North America for the week of the smoke event.51Figure 4.16: A map of all active fires in BC on August 18th 2018. Flames represent firesthat present a threat to public safety, red dots are fires that started within the past 24h andorange dots are fires that do not represent threats to public safety.Average PM2.5 concentration observed at Vancouver International Airport for this periodwas 42µgm−3, which included seven out of the 12 days registering averages over the CAAQS.This includes the 20th, 21st and 22nd, all of which recorded 24h averages of over 89µgm−3.These enhanced concentrations were not limited to the Valley; readings on Grouse Mountainwere as high, if not higher than those recorded in Vancouver, as seen in Figure 4.17.52On the morning of August 18th, PM2.5 levels on Grouse Mountain were similar to thosefound in Mahon Park. However, towards the end of the day, concentrations began to increaserapidly as the smoke plume, show by the red arrow in Figure 4.11, drifted in. At Mahon Park,observed PM2.5 concentrations remained consistent with what was seen earlier in the day asthe LFV had not yet begun to be exposed to this plume. This reflects what is seen in Figure4.4 where, on August 18th, the plume existed at or above 1000m and did not touch lowerelevations. The following Figure 4.17 shows how concentrations at Mahon Park changedover the course of the following week in response to this smoke plume. Unfortunately, onthe morning of August 19th, the increased concentrations clogged the GRIMM’s filter andmeasurements were no longer reliable.Figure 4.17: PM2.5 concentrations observed on Grouse Mountain and at North Vancouver’sMahon Park on August 18th, August 19th and the morning of August 20th, 2018.Figure 4.18 depicts the course of PM2.5 concentrations in the week following the largesmoke plume’s arrival in the LFV. Concentrations spiked towards the end of August 19th,and held at levels above 100µgm−3 for most of the day. This is important because this spikeoccurred half a day after the major spike observed on Grouse Mountain. This reflects what isshown above in Figure 4.11 where the major smoke plume (shown by the red arrow) residesat an altitude close to that of Grouse’s peak on August 18th/early 19th before smoke movesinto the valley throughout the 19th and early on the 20th.This further substantiates the idea that, depending on the synoptic conditions, for a shortperiod of time, Grouse’s peak will see much higher PM concentrations than the valley below.Similarly, it can serve as a warning or an indicator of imminent PM levels in the LFV.53Figure 4.18: PM2.5 concentrations observed at North Vancouver’s Mahon Park from August19th to August 26th4.5 ConclusionsIn this chapter, the spatial and temporal variations in PM on Grouse Mountain in thepresence of smoke were analyzed. Using the GRIMM, Dylos and GPS it was shown thatin the presence of smoke Grouse’s peak area still has very little spatial variability in PMconcentrations, similar to what was shown under clear sky conditions.Temporal variations in concentrations were more prominent in the presence of smoke.Concentrations observed using the LiDAR, the GRIMM and the Dylos varied considerablybetween days with smoke present. However, contrary to what was seen in under clear skyconditions, there was very little diurnal fluctuation in concentrations. Instead of rising overthe course of the day and declining overnight, concentrations would regularly experience arapid increase and then a very slow decline.At times during this event, concentrations on Grouse Mountain surpassed 100 µgm−3and there were four 24h periods where the average concentration was over the CAAQS andthus very harmful to human health. Considering the demand of most activities on GrouseMountain, the staff may consider closing on days with forecasted poor air quality. It may bewise to install their own air quality sensors for the safety of both the visitors and themselves.54More interesting was that Grouse often registered high values well before the LFV did.Specifically, when smoke arrived from the north, northeast or northwest, concentrations onGrouse would spike around half a day earlier than the concentrations in the valley. Followingthis, the smoke would descend into the valley and concentrations would become more uniformwith height. This was corroborated using the LiDAR imagery. Knowing this, further researchinto this matter may be of interest for Grouse’s staff. If the early, elevated concentrationscontinue to be a regular occurrence, it would be in their and their visitor’s best interest forthem to monitor and track the progress of BC wildfire plumes.Radiosonde vertical profiles taken during this event demonstrated greater atmosphericstability within smoke layers (during otherwise clear sky conditions) when compared tonormal clear sky conditions. This increase in stability could conceivably suppress the area’sregular mountain venting mechanisms and is potentially responsible in part for the lessdefined mountain boundary layer in Figure 4.11. This relationship has not been well studiedyet, and is an area that requires more research. With reduced venting mechanisms that arenot accounted for in models, air quality in the LFV is likely to be worse than predicted inthe presence of smoke.Finally, similar to under non-smoke conditions, evidence of the mountain boundary layerproposed by (De Wekker and Kossmann, 2015) was observed and is shown in Figure 4.11.This is interesting in the context of the LFV’s meteorology and because boundary layerdevelopment around mountains in the presence of smoke has been very sparsely studied.Further research on this topic, with aid from more radiosonde flights, could advance under-standing of these processes in the interest of improving weather modelling as well as theaccuracy of air quality advisories.55Chapter 5Conclusion5.1 Key FindingsThe main objective of this project was to improve understanding of the spatial and temporalvariations in PM concentrations on Grouse Mountain in the presence and absence of wildfireemissions. The effects of wildfire emissions on air quality was of primary interest, but theirinteractions with mountain flow processes were also examined. This chapter will summarizethe outcomes of this project and provide suggestions for future research.At Grouse’s peak area, the walking air quality surveys demonstrated very little spatialvariability in PM. On both clear and smoke days, PM concentrations were mostly uniformacross the summit. As seen in Figure 3.3 and Figure 4.2, there was some variation along thechosen path with the area near the bear habitat having slightly higher readings.On clear days, over nine surveys, the spatial standard deviation was 4µgm−3 (90% of themean) and the max recorded was four times that (16µgm−3). On smoke days, the absolutevalues were much higher (max > 100µgm−3) but the percent difference across the summitwas similar with a standard deviation of 28µgm−3 (90% of the mean).In contrast, the temporal variation of PM on Grouse Mountain was noteworthy, both inthe presence of smoke and not. PM levels exhibited both diurnal and longer-term variability,depending on different factors. The discrepancy between levels observed on Grouse and lowerin the valley at Mahon Park also varied substantially over the summer.Under non-smoke conditions, the diurnal variation in the boundary layer and anthro-pogenic activity drove the temporal variation in PM concentrations. The average of levelsobserved from 6pm to 6am was found to be 3.25µgm−3 while the average between 6am and6pm was 6.41µgm−3. Evidence of the effect of the boundary layer is shown in Figure 4.4.There, aerosol concentrations are shown to rise throughout the atmosphere with time.The photochemical smog event that occurred during the week of July 23rd brought a56complete change in the temporal variability of PM. Instead of exhibiting a diurnal pattern,concentrations remained relatively constant throughout every 24-hour period. The nighttimeaverage at Grouse peak was 13.2µgm−3 while the daytime average was 13.6µgm−3. This too,is shown by LiDAR imagery in Figure 3.4 where concentrations at Grouse peak’s elevationremain relatively similar over time.Overall, in the presence of smoke, the average observed hour-long PM concentration onGrouse Peak was 56.3µgm−3 with a maximum of 180.2µgm−3, higher than the average seenat Mahon Park which was 32.7µgm−3 with a max of 114.3µgm−3, demonstrating the needfor long term monitoring at Grouse Mountain.In the presence of smoke, diurnal variations were subdued in comparison to non-smokeconditions, but longer-term variations were enhanced. As seen in Figure 4.4, concentrationsat Grouse’s elevation did not diminish overnight. However, PM levels varied significantlyover the course of the week shown in Figure 4.4 and Figure 4.11. Using the LiDAR plots,the GRIMM observations, a composite map and the HYSPLIT model runs, it was shownthat these variations depend on the synoptic conditions and time since first exposure. Manyof the major plumes that affected Grouse arrived via the North, Northwest or Northeast(Figure 4.8 and Figure 4.9) in the presence of anti cyclonic conditions (Figure 4.15. In thesecases, air quality on Grouse degraded much earlier than the air quality in the city. Later,the smoke would entrain downwards into the valley and air quality values would on Grouseand in the valley would become more uniform.Finally, local circulation processes were also affected by the presence of smoke. In Figure4.5, the average potential temperature lapse rate in the presence of smoke is shown to be muchlower than the corresponding non smoke lapse rate; indicating a more stable atmosphere.This, coupled with less solar radiative input at the surface likely led to reduced mountainventing processes and accordingly, a less defined mountain boundary layer in Figure 4.4(smoke) when compared to Figure 3.4 (no smoke). The reduced venting processes wouldalso likely reduce cloud formation and aid in the persitance of smoke in the valley, creatinga feedback loop that would be extremely important to consider in weather forecasting in thepresence of smoke.5.2 LimitationsAs an observational study, this project does come with limitations. Although it forms thebasis for other, more thorough studies, it is difficult to draw concrete conclusions from it,owing to the lack of statistical rigour. Many more events would need to be sampled todraw definitive conclusions. The intermittent nature of the GRIMM observations at Grouse57Mountain and its failure halfway through the largest smoke event also limited the results ofthis study. The installation of a more permanent, reliable instrument would have greatlyincreased the validity of the outcomes.A lack of a full calibration for the Dylos instrument also limited the potential use of theresults as having more certainty in the absolute results would have allowed more concreteconclusions to be drawn about the difference between areas on the mountain. Another optionwould have been to pass the Dylos by the GRIMM multiple times during the Dylos surveys.Lastly, based on the results, this study would have benefitted from the use of a scanningLidar instead of, or as well as, the stationary Lidar. Point measurements are more of use overflat ground where there are less variables affecting the distribution of aerosols and horizontalmovement is less important (Feingold et al., 2005; Stone et al., 2008). Another option wouldhave been to have more Lidar instruments with at least another to the north of Grouse andone on top of Grouse as well as increased use of radiosondes for a more robust measurementof boundary layer depth.5.3 Future StudiesAs wildfire events are limited to a short period of time each year, these results would benefitfrom further similar studies in the future. Ideally, a long-term air quality monitoring stationwould be set up on Grouse Mountain to advance knowledge in this area and give staff atGrouse Mountain means to make informed air quality related decisions.A quantification of photochemical smog concentrations on Grouse Mountain in the pres-ence and absence of smoke would complement this study as well. In increased concentrations,photochemical smog can be detrimental to human health (Fowler , 2003) and is one of themain pollutants associated with wildfire smoke (McKendry and Lundgren, 2000). In chapter3 it is explained that Vancouver underwent an ”photochemical smog event” in late July andthe effects of this event on mountaintop concentrations could have been documented in asimilar manner to how PM was addressed in this study.Similarly, another long period of LiDAR measurements and Windsond flights would alsobe beneficial in the interest of understanding the proposed boundary layer dynamics andmountain flow processes in the presence of smoke. A comparison of these results withweather models that predict boundary layer height such as the WRF model would also beuseful in the interest of validating and improving these models.Incorporating realistic smoke plumes into weather models such as WRF to compare andcontrast to these events would also be of use to better understand these processes.Finally, similar studies in other locations would be useful. Given the recent propensity58for wildfires in California and the abundance of mountains in the north end of the state,it would be an ideal location, as would most locations in the Pacific Northwest. However,it would also be useful to examine mountainous locations that are less influenced by thepresence of a coastline.59BibliographyAkaoka, K., I. McKendry, J. Saxton, and P. W. Cottle, Impact of coal-carrying trains on par-ticulate matter concentrations in South Delta, British Columbia, Canada, EnvironmentalPollution, 223, 376–383, doi:10.1016/j.envpol.2017.01.034, 2017.Amiridis, V., D. S. Balis, E. Giannakaki, A. Stohl, S. Kazadzis, M. E. Koukouli, and P. Zanis,Optical characteristics of biomass burning aerosols over Southeastern Europe determinedfrom UV-Raman lidar measurements, Atmospheric Chemistry and Physics, 9 (7), 2431–2440, doi:10.5194/acp-9-2431-2009, 2009.Amiridis, V., et al., Impact of the 2009 Attica wild fires on the air quality in urban Athens,Atmospheric Environment, 46, 536–544, doi:10.1016/j.atmosenv.2011.07.056, 2012.Ansmann, A., H. Baars, A. Chudnovsky, I. Mattis, I. Veselovskii, M. Haarig, P. Seifert, R. En-gelmann, and U. Wandinger, Extreme levels of Canadian wildfire smoke in the stratosphereover central Europe on 21-22 August 2017, Atmospheric Chemistry and Physics, 18 (16),11,831–11,845, doi:10.5194/acp-18-11831-2018, 2018.Booth, B. B., N. J. Dunstone, P. R. Halloran, T. Andrews, and N. Bellouin, Aerosols im-plicated as a prime driver of twentieth-century North Atlantic climate variability, Nature,484 (7393), 228–232, doi:10.1038/nature10946, 2012.Calvo, A. I., V. Pont, A. Castro, M. Mallet, C. Palencia, J. C. Roger, P. Dubuisson, andR. Fraile, Radiative forcing of haze during a forest fire in Spain, Journal of GeophysicalResearch Atmospheres, 115 (8), doi:10.1029/2009JD012172, 2010.Carroll, M. S., P. J. Cohn, and K. A. Blatner, Private and tribal forest landowners and firerisk: A two-county case study in Washington State, Canadian Journal of Forest Research,34 (10), 2148–2158, doi:10.1139/X04-085, 2004.CBC, Air quality advisory issued for eastern parts of Metro Vancouver and Fraser Valley —CBC News, 2018.60Cottle, P., K. Strawbridge, I. Mckendry, N. O’Neill, and A. Saha, A pervasive and persis-tent asian dust event over north america during spring 2010: Lidar and sunphotometerobservations, Atmospheric Chemistry and Physics, 13 (9), 4515–4527, doi:10.5194/acp-13-4515-2013, 2013.Cottle, P., K. Strawbridge, and I. McKendry, Long-range transport of Siberian wildfiresmoke to British Columbia: Lidar observations and air quality impacts, AtmosphericEnvironment, 90, 71–77, doi:10.1016/j.atmosenv.2014.03.005, 2014.De Arruda Moreira, G., et al., Analyzing the turbulent planetary boundary layer by remotesensing systems: The Doppler wind lidar, aerosol elastic lidar and microwave radiome-ter, Atmospheric Chemistry and Physics, 19 (2), 1263–1280, doi:10.5194/acp-19-1263-2019,2019.De Wekker, S. F., and M. Kossmann, Convective boundary layer heights over mountain-ous terrain—A review of concepts, Frontiers in Earth Sciences, 3 (December), 1–22, doi:10.3389/feart.2015.00077, 2015.Feingold, G., H. Jiang, and J. Y. Harrington, On smoke suppression of clouds in Amazonia,Geophysical Research Letters, 32 (2), 1–4, doi:10.1029/2004GL021369, 2005.Fowler, C., Human Health Impacts of Forest Fires in the Southern United States: A Litera-ture Review, Journal of Ecological Anthropology, 7 (1), 39–63, doi:10.5038/2162-4593.7.1.3,2003.Fusco, E., B. Bradley, and J. T. Abatzoglou, Human-Related Ignitions Increase the Numberof Large Wildfires across U.S. Ecoregions, Fire, 1 (1), 4, doi:10.3390/fire1010004, 2018.Gallagher, J. P., I. G. Mckendry, A. M. Macdonald, and W. R. Leaitch, Seasonal and diurnalvariations in aerosol concentration on whistler mountain: Boundary layer influence andsynoptic-scale controls, Journal of Applied Meteorology and Climatology, 50 (11), 2210–2222, doi:10.1175/JAMC-D-11-028.1, 2011.Gallagher, J. P., I. G. McKendry, P. W. Cottle, A. M. Macdonald, W. R. Leaitch, andK. Strawbridge, Application of lidar data to assist airmass discrimination at the whistlermountaintop observatory, Journal of Applied Meteorology and Climatology, 51 (10), 1733–1739, doi:10.1175/JAMC-D-12-067.1, 2012.Grouse Mountain, Grouse Mountain Named Top Tourist Destination in Vancouver — GrouseMountain - The Peak of Vancouver, 2014.61Grouse Mountain, Grouse Mountain - The Peak of Vancouver, 2018.Henne, S., M. Furger, and A. S. Pre´voˆt, Climatology of mountain venting-induced elevatedmoisture layers in the lee of the Alps, Journal of Applied Meteorology, 44 (5), 620–633,doi:10.1175/JAM2217.1, 2005.Hennemuth, B., and A. Lammert, Determination of the atmospheric boundary layer heightfrom radiosonde and lidar backscatter, Boundary-Layer Meteorology, 120 (1), 181–200,doi:10.1007/s10546-005-9035-3, 2006.IPCC, Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II andIII to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change,151 pp., 2014.Kaulfus, A. S., U. Nair, D. Jaffe, S. A. Christopher, and S. Goodrick, Biomass Burning SmokeClimatology of the United States: Implications for Particulate Matter Air Quality, En-vironmental Science and Technology, 51 (20), 11,731–11,741, doi:10.1021/acs.est.7b03292,2017.Kossmann, M., and A. P. Sturman, Pressure-driven channeling effects in bentvalleys, Journal of Applied Meteorology, 42 (1), 151–158, doi:10.1175/1520-0450(2003)042¡0151:PDCEIB¿2.0.CO;2, 2003.Liu, Y., S. Goodrick, and W. Heilman, Wildland fire emissions, carbon, and cli-mate: Wildfire-climate interactions, Forest Ecology and Management, 317, 80–96, doi:10.1016/j.foreco.2013.02.020, 2014.Logan, T., B. Xi, and X. Dong, Biomass burning aerosol properties over the Northern GreatPlains during the 2012 warm season, Atmospheric Chemistry and Physics Discussions,13 (12), 32,269–32,289, doi:10.5194/acpd-13-32269-2013, 2013.Manikonda, A., N. Z´ıkova´, P. K. Hopke, and A. R. Ferro, Laboratory assessment of low-costPM monitors, Journal of Aerosol Science, 102, 29–40, doi:10.1016/j.jaerosci.2016.08.010,2016.Manninen, A. J., T. Marke, M. Tuononen, and E. J. O’Connor, Atmospheric BoundaryLayer Classification With Doppler Lidar, Journal of Geophysical Research: Atmospheres,123 (15), 8172–8189, doi:10.1029/2017JD028169, 2018.Mass, C., The weather of the Pacific Northwest, vol. 46, 46–3860–46–3860 pp., University ofWashington Press, doi:10.5860/choice.46-3860, 2009.62McClure, C. D., and D. A. Jaffe, US particulate matter air quality improves except inwildfire-prone areas, Proceedings of the National Academy of Sciences of the United Statesof America, 115 (31), 7901–7906, doi:10.1073/pnas.1804353115, 2018.McKendry, I., et al., Californian forest fire plumes over southwestern British Columbia:Lidar, sunphotometry, and mountaintop chemistry observations, Atmospheric Chemistryand Physics, 11 (2), 465–477, doi:10.5194/acp-11-465-2011, 2011.Mckendry, I. G., UBC GEOB 402: Air Pollution Meteorology, 2018.McKendry, I. G., and J. Lundgren, Tropospheric layering of ozone in regions of urbanizedcomplex and/or coastal terrain: A review, Progress in Physical Geography, 24 (3), 329–354,doi:10.1191/030913300701542660, 2000.McKendry, I. G., J. Gallagher, P. Campuzano Jost, A. Bertram, K. Strawbridge, R. Leaitch,and A. M. MacDonald, Ground-based remote sensing of an elevated forest fire aerosol layerat Whistler, BC: Implications for interpretation of mountaintop chemistry, AtmosphericChemistry and Physics, 10 (23), 11,921–11,930, doi:10.5194/acp-10-11921-2010, 2010.Metro Vancouver, 2009 Lower Fraser Valley Air Quality Report, Tech. Rep. December, 2010.Monn, C., A. Fuchs, D. Ho¨gger, M. Junker, D. Kogelschatz, N. Roth, and H. U. Wanner,Particulate matter less than 10 µm (PM10) and fine particles less than 2.5 µm (PM2.5):Relationships between indoor, outdoor and personal concentrations, Science of the TotalEnvironment, 208 (1-2), 15–21, doi:10.1016/S0048-9697(97)00271-4, 1997.Oke, T. R., and J. E. Hay, The climate of Vancouver, 23, 48 p. pp., Vancouver : Dept. ofGeography, University of British Columbia, 1994.Peshev, Z. Y., A. D. Deleva, T. N. Dreischuh, and D. V. Stoyanov, Lidar measurementsof wildfire smoke aerosols in the atmosphere above Sofia, Bulgaria, in 19th InternationalConference and School on Quantum Electronics: Laser Physics and Applications, vol.10226, p. 102260Y, SPIE, doi:10.1117/12.2263103, 2017.Rampanelli, G., and D. Zardi, A method to determine the capping inversion of the convec-tive boundary layer, Journal of Applied Meteorology, 43 (6), 925–933, doi:10.1175/1520-0450(2004)043¡0925:AMTDTC¿2.0.CO;2, 2004.Richner, H., and P. Ha¨chler, Understanding and Forecasting Alpine Foehn, pp. 219–260,2013.63Semple, S., A. E. Ibrahim, A. Apsley, M. Steiner, and S. Turner, Using a new, Low-Cost airquality sensor to quantify Second-Hand smoke (SHS) levels in homes, Tobacco Control,24 (2), 153–158, doi:10.1136/tobaccocontrol-2013-051188, 2015.Steinle, S., et al., Personal exposure monitoring of PM2.5 in indoor and out-door microenvironments, Science of the Total Environment, 508, 383–394, doi:10.1016/j.scitotenv.2014.12.003, 2015.Stone, E. A., D. C. Snyder, R. J. Sheesley, A. P. Sullivan, R. J. Weber, and J. J. Schauer,Source apportionment of fine organic aerosol in Mexico City during the MILAGRO ex-periment 2006, Atmospheric Chemistry and Physics, 8 (5), 1249–1259, doi:10.5194/acp-8-1249-2008, 2008.Stull, R. B., An introduction to boundary layer meteorology, Kluwer Academic Publishers,Dordrecht, Boston and London, doi:http://dx.doi.org/10.1007/978-94-009-3027-8, 1988.Takahama, S., R. E. Schwartz, L. M. Russell, A. M. MacDonald, S. Sharma, and W. R.Leaitch, Organic functional groups in aerosol particles from burning and non-burning forestemissions at a high-elevation mountain site, vol. 11, 6367–6386 pp., doi:10.5194/acp-11-6367-2011, 2011.Talukdar, S., M. Venkat Ratnam, V. Ravikiran, and R. Chakraborty, Influence of BlackCarbon Aerosol on the Atmospheric Instability, Journal of Geophysical Research: Atmo-spheres, 124 (10), 5539–5554, doi:10.1029/2018JD029611, 2019.Teakles, A. D., et al., Impacts of the July 2012 Siberian fire plume on air quality in the PacificNorthwest, Atmospheric Chemistry and Physics, 17 (4), 2593–2611, doi:10.5194/acp-17-2593-2017, 2017.Teixeira, M. A. C., D. J. Kirshbaum, H. O´lafsson, P. F. Sheridan, and I. Stiperski, Ed-itorial: The atmosphere over mountainous regions, Frontiers in Earth Science, 4, doi:10.3389/feart.2016.00084, 2016.Tosca, M. G., J. T. Randerson, C. S. Zender, M. G. Flanner, and P. J. Rasch, Do biomassburning aerosols intensify drought in equatorial Asia during El NiA˜±o?, AtmosphericChemistry and Physics, 10 (8), 3515–3528, doi:10.5194/acp-10-3515-2010, 2010.Wang, J., and S. A. Christopher, Mesoscale modeling of Central American smoke transportto the United States: 2. Smoke radiative impact on regional surface energy budget andboundary layer evolution, Journal of Geophysical Research Atmospheres, 111 (14), doi:10.1029/2005JD006720, 2006.64Zardi, D., and C. D. Whiteman, Diurnal mountain wind systems. Springer AtmosphericSciences, Mountain Weather Research and Forecasting, pp. 35–119, doi:10.1007/978-94-007-4098-3 2, 2013.65


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items