Vertical profiles of aerosols, NO2, and SO2 were retrieved from Multi-Axis Differential Optical Absorption Spectroscopy (MAX-DOAS) measurements at a field site in northern Alberta, Canada, during August and September 2013. The site is approximately 16 km north of two mining operations that are major sources of industrial pollution in the Athabasca oil sands region. Pollution conditions during the study ranged from atmospheric background conditions to heavily polluted with elevated plumes, according to the meteorology. This study aimed to evaluate the performance of the aerosol and trace gas retrievals through comparison with data from a suite of other instruments. Comparisons of aerosol optical depths (AODs) from MAX-DOAS aerosol retrievals, lidar vertical profiles of aerosol extinction, and the AERONET sun photometer indicate good performance by the MAX-DOAS retrievals. These comparisons and modelling of the lidar S ratio highlight the need for
accurate knowledge of the temporal variation in the S ratio when comparing
MAX-DOAS and lidar data. Comparisons of MAX-DOAS NO2 and SO2 retrievals to Pandora spectral sun photometer vertical column densities (VCDs) and active DOAS mixing ratios indicate good performance of the retrievals, except when vertical profiles of pollutants within the boundary layer varied rapidly, temporally, and spatially. Near-surface retrievals tended to overestimate active DOAS mixing ratios. The MAX-DOAS observed elevated pollution plumes not observed by the active DOAS, highlighting one of the instrument's main advantages. Aircraft measurements of SO2 were used to validate retrieved vertical profiles of SO2. Advantages of the MAX-DOAS instrument include increasing sensitivity towards the surface and the ability to simultaneously retrieve vertical profiles of aerosols and trace gases without requiring additional parameters, such as the S ratio. This complex dataset provided a rare opportunity to evaluate the performance of the MAX-DOAS retrievals under varying atmospheric conditions.
Introduction
The Athabasca oil sands operations in Alberta contain significant sources of industrial atmospheric pollutants, such as sulfur dioxide (SO2) and nitrogen dioxide (NO2) (ECCC, 2018b, c). Oil extraction and upgrading activities such as surface mining, acid gas
flaring, and transporting materials in heavy hauler trucks emit aerosols and
trace gas pollutants (Liggio et al., 2016). Pollutant emissions from the
industrial smokestacks result in uplifted profiles with the potential to be
transported farther downwind compared to emissions released at the surface,
particularly for stacks with high-volume flow rates and temperatures that
can rise high in the atmosphere (Zhang et al., 2018). While the Athabasca
oil sands region (AOSR) experiences moderate annual average concentrations
of SO2 relative to all Canadian in situ stations, the short-term
concentrations can be significantly higher than in most Canadian cities
(Government of Canada, 2018). The AOSR contains some of the few monitoring sites in Canada that experience peak 1 h average concentrations of SO2 of greater than 70 ppb (Government of Canada, 2018), which is the new 2020 Canadian Ambient Air Quality Standard for SO2 (Canadian Council of Ministers of the Environment, 2014). SO2 concentrations of up to 131 ppb were also observed by aircraft measurements downwind of an AOSR industrial facility in 2013, approximately midway between the Syncrude Mildred Lake Plant and Fort McKay (Baray et al., 2018). High concentrations of SO2 over short durations are a health concern because negative pulmonary and respiratory effects of inhalation can occur after exposure periods as short as 10 min (Health Canada, 2016; WHO, 2006). Exposure to NO2 at high concentrations over the short term is also associated with significant health impacts (WHO, 2006), and NOx (NO+NO2) is a precursor to tropospheric ozone (O3), acid rain, and fine particulate matter (Seinfeld and Pandis, 2006).
Emissions of NOx and SO2 lead to the formation of nitrate and sulfate aerosols, which constitute a significant fraction of the PM2.5 air mass in urban and industrially impacted regions (Pui et al., 2014). The highest peak and annual average PM2.5 concentrations in Canada in 2016 were observed at two monitoring stations within Fort McMurray with annual averages of over 18 µgm-3 compared to 8 µgm-3 in an industrial area of Toronto, Ontario (Government of Canada, 2018). Exposure to PM2.5 leads to adverse effects on respiratory and cardiovascular systems (WHO, 2006).
In the troposphere, nearly all SO2 is oxidized to H2SO4 aerosol through reactions in the gas and aqueous phases. The hydroxyl (OH)
radical initiates the oxidation route of SO2 in the gas phase,
forming HOSO2 (Holloway and Wayne, 2010). Sulfuric acid
(H2SO4) is formed through further oxidation of HOSO2 and
condenses onto already present aerosols or can nucleate with water vapour
(H2O) and gaseous ammonia (NH3), forming sulfate aerosol
(Kulmala et al., 2004). Aqueous-phase reactions form sulfate aerosol efficiently, with H2O2 and O3 acting as oxidants (Seinfeld and Pandis, 2006). Wet deposition dominates the removal of sulfate aerosol. Therefore, elevated levels of SO2 and NO2 observed over the AOSR region are an environmental concern since atmospheric depositions of sulfur oxides (SOx) and nitrogen oxides (NOx) can lead to freshwater and soil acidification (Psenner, 1994; Zhao et al., 2009). Deposition of nitrogen compound can harm sensitive ecosystems through eutrophication (excessive nutrient richness) of water bodies (Fenn et al., 2015).
High concentrations of SO2 and other pollutants over the AOSR have
prompted measurements using aircraft studies (Baray et al., 2018; Gordon et al., 2015; Liggio et al., 2016, 2019; Simpson et al., 2010), in situ measurements (Amiri et al., 2018; Hsu, 2013; Tokarek et al., 2018), sun photometers (Fioletov et al., 2016), and satellites (McLinden
et al., 2012, 2014, 2016). Long-term monitoring through satellite measurements is an attractive choice due to the large scale of the
operations. However, surface concentrations are difficult to determine
accurately from satellite measurements (Fioletov et al., 2016), and data acquisition is limited to the satellite overpass times. Satellite retrievals in the AOSR region are also complicated by multiple factors: landscapes are complex, emissions can change relatively rapidly, and the winds within the higher boundary layer can quickly disperse pollution emissions. Rapid industrial expansion can also require updating retrieval algorithms (McLinden et al., 2014). Apparent peak concentrations are reduced, and small-scale variability cannot be resolved, due to spatial averaging within the footprint of a pixel that can be large relative to the scale of point source plumes.
SO2, NO2 and aerosol levels in the total column and near-surface can be simultaneously monitored using the Multi-Axis Differential Optical Absorption Spectroscopy (MAX-DOAS) technique
(Hönninger et al., 2004). The elevated levels of SO2 observed in the AOSR increase the ease of MAX-DOAS measurements compared to within most Canadian cities, where SO2 levels are significantly lower. Differential Optical Absorption Spectroscopy (DOAS) is a remote sensing technique that quantifies tropospheric trace gases using light spectra and the unique spectral absorption cross sections of trace gases. Since its introduction by Platt et al. (1979) DOAS has been used to quantify trace gases in the troposphere, including NO2, SO2, OH, BrO, NO3, NH3, ClO, and others. The technique has the advantage of allowing the simultaneous quantification of multiple trace gases (Platt et al., 2008). The MAX-DOAS method measures scattered sunlight spectra at multiple viewing directions and/or elevation angles to allow sensitive quantification of tropospheric pollutants. Spectra measured at elevation angles close to horizon-pointing have a higher sensitivity to ground-level pollutants since the light paths are longer near the surface (Hönninger et al., 2004). Ground-based MAX-DOAS measurements determine tropospheric vertical column densities (VCDs) of trace gases, quantifying total boundary layer pollution loading. VCDs have the advantage of being independent of boundary layer height and are spatially averaged (horizontally) on the order of a few kilometres along the light path.
Ground-based MAX-DOAS data combined with radiative transfer modelling allows
retrieval of vertical profiles of aerosol extinction and trace gases (Frieß et al., 2006; Hönninger et al., 2004; Hönninger and Platt, 2002; Irie et al., 2008; Wagner et al., 2004). The MAX-DOAS technique has been used to retrieve vertical profiles of aerosol extinction (Clémer
et al., 2010; Frieß et al., 2011; Irie et al., 2008, 2015; Li et al., 2010; Zieger et al., 2011), BrO (Frieß et al., 2011; Hönninger and Platt, 2002), HCHO (Heckel et al., 2005; Wagner et al., 2011), SO2 (Tan et al., 2018), and NO2 (Tan et al., 2018; Wagner et al., 2011).
There are few comparisons of vertical profiles of aerosol extinction from
MAX-DOAS to vertical profiles from other instruments in the literature.
MAX-DOAS aerosol extinction profiles have been compared to smoothed
extinction profiles from a sun photometer (Frieß et al., 2011) and aircraft aerosol profiles (Wagner et al., 2011). Near-surface MAX-DOAS retrievals of aerosol extinction have been compared with in situ measurements of aerosols (Zieger et al., 2011). There are also relatively few published comparisons of MAX-DOAS aerosol optical depths (AODs) with lidar AODs (Irie et al., 2008, 2015). Relatively few studies have focused on MAX-DOAS measurements of anthropogenic SO2 (Irie et al., 2011; Jin et al., 2016; Wang et al., 2014, 2017; Wu et al., 2018, 2013). Most studies that present MAX-DOAS vertical profile retrievals compare them to trace gas VCDs or near-surface measurements from in situ or LP-DOAS instruments. Tan et al. (2018) and Wang et al. (2017) compared MAX-DOAS SO2 VCDs to satellite VCDs of trace gases. Tan et al. (2018) and Wagner et al. (2011) compared MAX-DOAS retrievals of vertical profiles of NO2 to satellite VCDs and near-surface NO2
mixing ratios from LP-DOAS, respectively.
In this study, a MAX-DOAS instrument was deployed during a comprehensive air
quality campaign conducted during August and September 2013. Pollution
conditions ranged from background to heavily polluted with a well-mixed
boundary layer to distinctly elevated pollution plumes. Vertical profiles of
aerosols, NO2, and SO2 in the troposphere were retrieved using optimal estimation inverse modelling from the MAX-DOAS measurements. These retrievals allowed characterization of the vertical structure of the
boundary layer. The retrieval used a two-step approach: (1) aerosol extinction profiles are retrieved from measured MAX-DOAS O4 differential slant column densities (dSCDs) and (2) the aerosol extinction
profiles are used as forward model parameters for retrieval of trace gas
profiles from measured trace gas dSCDs.
Our study adds to the current literature by comparing MAX-DOAS aerosol and
trace gas retrievals with data from numerous other instruments deployed
during the campaign. The aerosol retrievals were compared to aerosol
extinction data from a co-located lidar instrument and a nearby sun photometer. Validation of the aerosol retrievals is essential because these
profiles are used as model parameters for the trace gas retrievals. MAX-DOAS
NO2 and SO2 retrievals were compared to mixing ratios from a
co-located active DOAS instrument and tropospheric VCDs of trace gas from a
Pandora sun photometer. In situ measurements of SO2 from an aircraft
allowed comparison of MAX-DOAS vertical profiles of SO2. Evaluation of the retrievals was aided by co-located, near-surface measurements of
particle size distribution and composition and nearby high-resolution
measurements of vertical profiles of wind speed and direction.
The objectives of our study were to (1) determine the factors required to
validate MAX-DOAS aerosol retrievals through comparison with lidar and sun
photometer data, (2) evaluate the performance of the aerosol and trace gas
retrievals through comparison to other datasets, (3) identify conditions that
limit the use of the MAX-DOAS technique, and (4) identify conditions under
which the MAX-DOAS method was advantageous over other instruments.
This complex dataset from comprehensive measurements in the vicinity of oil
sand operations provided a unique opportunity to test the performance of the
MAX-DOAS aerosol and trace gas retrievals.
ExperimentalField sites
The MAX-DOAS instrument was operated at an elevation of ∼10 m above the surface from 14 August to 9 September 2013 at the Fort McKay South
field site (57.149∘ N, 111.642∘ W) north of Fort McMurray, Alberta, concurrent with an Environment and Climate Change Canada (ECCC) intensive measurement campaign (Figs. 1 and S1 in the Supplement). A second site was located 4 km north of Fort McKay South (Oski-Ôtin; 57.184∘ N, 111.640∘ W) in the Fort MacKay community.
Two major sources of aerosols, NO2, SO2, and other pollutants are located south of Fort McKay South: the Syncrude Mildred Lake Plant and the Suncor Millennium Plant, 12 km to the south and 20 km to the south–south-east,
respectively (Fig. 1). The 2013 NPRI-reported emissions of SO2 and
NOx from these facilities were 63 and 14 kilotons (kt) and 14 and 8 kt, respectively (ECCC, 2018a). Relatively small sources of
pollutants are located north of Fort McKay South: Shell Jackpine and Muskeg
River Mines, CNRL Horizon, and Imperial Oil Kearl Mine (Fig. 1). Tables S1
and S2 in the Supplement show the 2013 NPRI emissions of SO2 and NOx from these five facilities. A recent study suggests that total industrial emissions of NOx were underestimated in the NPRI report, particularly for ground sources (Zhang et al., 2018). Since there are NOx sources that are not included in the NPRI
emissions data, the 2010 vehicular
emissions associated with each facility and 2012–2013 annual stack and area
source emissions from Zhang et al. (2018) are also included in Tables S1 and S2.
Location of field sites Fort McKay South and Oski-Ôtin and major industry sources.
Instrumentation
The mini-MAX-DOAS instrument from Hoffmann Messtechnik GmbH measured
scattered sunlight with a viewing azimuth angle of 155∘
south–south-east (SSE) at sequential viewing elevation angles 2, 4, 8,
15, 30 and 90∘ (zenith) above the horizon. The instrument consisted of a sealed metal box containing entrance optics, a UV fibre-coupled spectrograph, and all electronics. The instrument field of view was approximately 0.6∘. Incident light was focused on a cylindrical quartz lens (focal length =40 mm) into a quartz fibre that transmitted the light into the OceanOptics USB2000 spectrograph. The spectrograph detector was a Sony ILX511 linear silicon charge-coupled device (CCD) array (2048 pixels, pixel size 14µm×200µm, signal-to-noise ratio at full signal 250:1). The
spectrograph had a spectral range of 290–433 nm, a 50 µm wide
entrance slit, and a spectral resolution of ∼0.6 nm full width at half maximum (FWHM). The
spectrograph was cooled by a Peltier stage to maintain the selected
temperature (5∘C). Spectrometer data were transferred to a laptop
computer via USB cable. The instrument was controlled using the software
package DOASIS, which allowed automated measurements by JScript programs.
The instrument was mounted on an elevated scaffold approximately 10 m above
ground level (a.g.l.), at approximately the height of the surrounding forest
canopy. Each recorded measurement spectrum was an average of 2000 measured
spectra with an exposure time that varied between 50 and 200 ms,
depending on the ambient light levels.
MAX-DOAS aerosol and trace gas retrieved data were inter-compared with data
from various other instruments deployed during the campaign. Table 1
provides information on these instruments and papers that describe their
operation.
Description and locations of the study instruments.
InstrumentVariables measuredInstitutionTemporalViewing directionField siteReferenceresolutionMini-MAX-DOASVertical profiles of SO2,York20–30 minSSE at multipleFort McKayCurrent paperNO2, aerosol extinctionUniversityelevation anglesSouthActive DOASMixing ratios of SO2,York∼2 minSSE, horizontalFort McKayMcLaren et al. (2010)NO2UniversitySouthPandora sunVCDs of SO2, NO2ECCC∼1 minDirect sun viewingOski-ÔtinFioletov et al. (2016)photometerSun photometerAOD, Angstrom exponentECCC∼3 minDirect sun viewingOski-ÔtinSioris et al. (2017)Ground-based lidarVertical profile ofECCC1 minZenith viewingFort McKayStrawbridge (2013)aerosol extinctionSouthTSI APS 3321PM10-1 size distributionUniversity6 minn/aFort McKayTokarek et al. (2018)of CalgarySouthTSI SMPSPM1 size distributionUniversity6 minn/aFort McKayTokarek et al. (2018)(3081 DMA,of AlbertaSouth3776 CPC)AerodynerBC, NH4+(p), SO42-(p),University of∼1 minn/aFort McKayLee et al. (2019)SP-AMSNO3-(p), Cl-(p),Toronto andSouthorganicsECCCScintec modelVertical profile of windECCC15 minZenith viewingOski-ÔtinGordon et al. (2018)MFAS windRASSand temperatureAirborne ThermoMixing ratios of SO2ECCC1 sn/an/aBaray et al. (2018)Scientific 43iTLE
n/a: not applicable.
An active DOAS instrument located at the same site was used to retrieve
mixing ratios of NO2 and SO2 at 3.5 m a.g.l. The active DOAS light
path was pointed in a south–south-east direction, approximately parallel to
the MAX-DOAS viewing azimuth angle (Fig. 1). Measurements of trace gases
with the active DOAS system have been described previously (McLaren
et al., 2010, 2012; Wojtal et al., 2011), although details have been changed in the
current study. DOAS measurements were made using a modified DOAS 2000
Instrument (TEI Inc.) utilizing a 150 W high-pressure Xe arc lamp and a
coaxial Cassegrain telescope. The outgoing beam traversed the atmosphere for
1.15 km (pathlength =2.3 km) at an average height of 3.5 m a.g.l., where it
impacted a retroreflector array composed of 30 (5.08 cm diameter) hollow corner
cubes mounted on a raiseable tower. The beam traversed through an
exploration line cut (5–10 m wide by 2 km long) in a mature coniferous
forest. Return light was collected with a 2m×600µm UV-transparent fibre-optic cable and spectrometer (Ocean Optics USB2000, Grating no. 10, λ=288–492 nm, 1800 lines mm-1, 2048 element
CCD, 25 µm slit, UV2 upgrade, L2 lens). Integration times of 30–40 ms and 4000 averages gave ≈2 min resolution with detection limits (3σ) of 120 and 170 ppt for NO2 and SO2, respectively. Xenon lamp, Hg calibration, offset, and dark noise spectra were collected for spectral fitting with DOASIS software. A small diffuser was installed in the entrance of the fibre to lower atmospheric turbulence noise (Stutz and Platt, 1993) in addition to using an optical fibre bending mode mixer.
A Pandora spectral sun photometer at Oski-Ôtin measured in direct sun
and zenith sun viewing modes to retrieve total atmospheric column VCDs of
SO2 and NO2 with precisions (1σ) of 4.6×1015 and 0.3×1015 molecules cm-2, respectively (Fioletov et al., 2016). Tropospheric VCDs of NO2 were determined from the Pandora total column VCDs by subtracting stratospheric VCDs modelled using the PRATMO stratospheric photochemical box model (McLinden, 2000). PRATMO was used as described in Adams et al. (2016), except monthly mean OSIRIS ozone profiles (Degenstein et al., 2009) and MODIS surface reflectivities (McLinden et al., 2014) were employed. The Pandora SO2 VCDs presented are assumed to be representative of tropospheric SO2 VCDs since stratospheric SO2 was assumed to be negligible. Pandora trace gas and MAX-DOAS data were both available for inter-comparison for 4 d during the study. SO2 and NO2 mixing ratios were also measured from the air on board a Convair 580 research aircraft (Baray et
al., 2018) using Thermo Scientific 43iTLE and 42i-TL analyzers,
respectively, between 12 August and 7 September 2013, including a spiral
ascent near Fort McKay South (3 September).
AODs at 380 and 340 nm were obtained from Level 2.0 AERONET data, measured by a second sun photometer at Oski-Ôtin.
Aerosol extinction profiles at 532 nm from 0.1 to 12 km a.g.l. were retrieved using a ground-based, zenith-pointing lidar operated at Fort McKay South (Strawbridge, 2013). In this study, the lidar profiles from 0.1 to 4 km were considered in order to match the vertical observation extent of the MAX-DOAS. The lidar has the advantage over sun photometer instruments because it can determine the vertical profile of optical extinction rather than just a column-averaged value but has higher uncertainty when the S ratio is variable (Strawbridge, 2013). Aerosol extinction profiles are retrieved from the measurements of the laser return signal using a chosen S ratio value. The S ratio is the ratio of the volume extinction coefficient to the backscatter coefficient and dictates the signal strength of the received return of the lidar's pulsed laser source (Strawbridge, 2013). Lidar S ratios are known
to be variable but are often estimated given the type of particles expected
in an environment (Irie et al., 2015). The S ratio depends on the shape, size distribution, and chemical composition of the aerosol particles, as well as the relative humidity (Weitkamp, 2005). A constant lidar ratio (“S ratio”) of 25 was used for the lidar retrievals unless otherwise specified. S ratios were modelled using Mie scattering theory and measurements of surface-level
particle composition and size distribution at Fort McKay South for various
times during 23 August to determine temporal variability in the S ratio. Source code for the Mie scattering calculations can be found in Aggarwal et al. (2018).
Ground-level particle composition was measured using an Aerodyne high-resolution soot particle aerosol mass spectrometer (SP-AMS) (Lee et al., 2019). Particle size distributions were measured using Scanning Mobility Particle Sizer (SMPS) (“dry” line mode) and Aerodynamic Particle Sizer (APS) instruments (see supplementary information in Tokarek
et al., 2018, for more details). Particle diameters measured by the SMPS and
by the APS were 0.014–0.74 and 0.5–19.81 µm, respectively. Data
from these instruments were combined to determine particle size
distributions from 0.014 to 19.81 µm, assuming the particles were unit density. Use of “dry” line mode SMPS increased uncertainties in the size distributions because ambient aerosols have more volume than dry aerosols. However, even in the highest relative humidity range, the ambient aerosol had only 30 % more volume compared to the dry aerosol which, assuming spherical particles, only results in a maximum increase in particle diameter of 9 %. The resulting error is expected to be much smaller than other errors such as converting mobility and aerodynamic diameters to optical
diameters.
A radio acoustic meteorological profiler (windRASS, model MFAS, Scintec,
Germany) at Oski-Ôtin measured temperature, wind speed, and wind direction at
10 m intervals from 40 m to up to a maximum altitude of 800 m (Gordon et al., 2018).
MAX-DOAS data analysisMAX-DOAS fitting
Trace gas differential slant column densities (dSCDs) were obtained using the DOAS technique (Platt et al., 2008) with DOASIS software (Institute of Environmental Physics, Heidelberg, Germany). All spectra were corrected for dark current and electronic offset and wavelength that were calibrated using a measurement of a Hg lamp. Table 2 shows the wavelength windows and fit components used to retrieve dSCDs of NO2, SO2, and O4. Cross sections were obtained from the MPI-Mainz UV/VIS Spectral Atlas of Gaseous Molecules of Atmospheric Interest (Keller-Rudek et al., 2013). Examples of spectral retrievals of the gases are shown in Fig. S2. Each
non-zenith measured spectrum was fit against the closest zenith spectrum in
time, also known as the Fraunhofer reference spectrum (FRS). The statistical
error of the O4 dSCDs was <1.1×1042 molecules cm-2. The O4 error for off-axis measurements relative to the FRS are <6 % for angles below 30∘ and <10 % for the 30∘ measurements. The statistical fit errors of the SO2 and NO2 dSCDs were 0.4–1.2×1016 and 0.4–1.6×1015 molecules cm-2, respectively. Uncertainties in the absorption cross sections result in systematic errors in the retrieved dSCDs. The reported uncertainty in the SO2 and NO2 absorption cross sections used is approximately 3 % (Bogumil et al., 2003). The absolute value of the O4 cross section and its dependence on
temperature is uncertain. Some studies suggest that the absolute value of
the cross section may be overestimated by up to 25 %, requiring the use of a scaling factor (Clémer et al., 2010; Wagner et al., 2002, 2009, 2019). However, Frieß et al. (2011) found that the best results for measured O4 dSCDs and the vertical profiles of aerosol extinction retrieved from them were achieved without a scaling factor. Irie et al. (2015) found that a scaling factor of 1.25 resulted in an overestimation of near-surface aerosol extinction coefficients (AECs) but also reduced residuals at high viewing elevation angles. Wagner et al. (2019) found that measured and radiative transfer modelled O4 absorptions showed good agreement on one study day but poor agreement on the second. A scaling factor was not used for the O4 fitting in this study because of the lack of consensus on the need for a scaling factor within the DOAS community and the good agreement between the MAX-DOAS and smoothed lidar AODs for the 23 August data (see Sect. 3.1.2).
Information on MAX-DOAS spectral fitting.
GasFitting windowIncluded in the fitNO2410–435 nmFRS, Ring, Bogumil (2003) NO2 (293 K) and Bogumil (2003) (293 and 243 K) O3,third-order polynomialSO2310.5–324 nmFRS, Ring, Bogumil (2003) SO2 (293 K) and Bogumil (2003) (293 and 223 K) O3,third-order polynomial, Offset FunctionO4350–375 nmFRS, Ring, Hermans (2011) O4 Bogumil (2003) (293 K) NO2, Bogumil (2003) (293 and 223 K) O3,third-order polynomial
The SO2 fitting range was determined based on an experiment using an
SO2 calibration cell from Resonance Ltd. with a slant column density
(SCD) of 2.2×1017 (±10 %) molecules cm-2 placed inside the MAX-DOAS telescope. Scattered solar light spectra were recorded
around solar noon at multiple viewing elevation angles above the horizon,
followed by a 90∘ measurement without the cell (the FRS). For each of
the measured spectra, dSCDs of SO2 were fit in DOASIS by varying the
fitting windows in ∼0.3 nm increments with a range of lower and upper limits of 303–318 and 309–340 nm, respectively. The fit components
are the same as in Table 2. See Sect. S2 in the Supplement for details. The
NO2 and O4 fitting ranges were from McLaren et al. (2010) and Frieß et al. (2011), respectively.
Retrieval of vertical profiles from MAX-DOAS dSCDs using optimal estimation
Aerosol and trace gas profiles were retrieved using a two-step approach: (1) aerosol extinction profiles were retrieved from measured MAX-DOAS O4 dSCDs, and (2) aerosol extinction profiles were used as forward model parameters for retrieval of NO2 and SO2 profiles from dSCDs of NO2 and SO2, respectively. Vertical profiles were determined from dSCDs using retrieval algorithms based on the Rodgers (2000)
optimal estimation technique (Frieß et al., 2011, 2016, 2019). Generally, the desired state of the atmosphere (x) can be estimated from remote sensing measurements (y) using a forward model F:
y=Fx,b+ε,
where ε is the measurement error and b is the vector of model parameters that are assumed to be known and not determined by the modelling, such as aerosol microphysical properties. In this study, the SCIATRAN radiative transfer model was used as the forward model (Rozanov et al., 2005).
The optimal estimation method determined the most probable atmospheric
state, x^, based on a set of measurements, y, and an
a priori state vector, xa. The xa was the best guess of the vertical profile to be retrieved. The x^ was the aerosol extinctions or the trace gas mixing ratios at a series of altitude intervals for the aerosol retrieval and trace gas retrievals, respectively. The y was the O4 dSCDs and the trace gas dSCDs measured at different angles, for the aerosol and trace gas retrievals, respectively. The agreements between measured and modelled dSCDs based on linear regressions for the retrievals are shown in Sect. S8.1 in Tables S10 and S11, as well as typical degrees of free of signal for the aerosol and trace gas retrievals. Note that in our retrievals, y was the dSCDs measured at sequential elevation angles during 20 min periods before 17:00 local time (LT) and during 30 min periods after 17:00 LT. The wavelengths for the optimal estimation retrievals of O4, NO2, and SO2 were 360.8, 422.5, and 318.0 nm, respectively.
The optimal estimation solution x^ is the maximum a posteriori (MAP)
solution, which selects the most probable state from the set of possible
states described by maximizing the probability of x occurring given the observations y (Rodgers, 2000). The MAP solution is found by minimizing the cost function (χ2):
χ2=(y-F(x,b))TSE-1(y-F(x,b))+(x-xa)TSa-1(x-xa),
where Sa and SE are the error covariance matrices associated with the a priori and measurement vectors, respectively (Rodgers, 2000). The retrieval yields important quantities that allow the characterization of the retrieval. These include the weighting function, K=∂y∂x, which quantifies the sensitivity of the measurement towards the atmospheric state, and the averaging kernel matrix, A=∂x^∂x, which quantifies the vertical resolution of the retrieval. A describes the sensitivity of the retrieved profile to changes in the true atmospheric profile. Rows of A are averaging kernels for each altitude interval in the retrieved profile. The full width half maximum (FWHM) of each kernel gives an estimate of the retrieval's vertical resolution at height z. Each averaging kernel ideally peaks at a magnitude of 1.0 at the height of the kernel. However, the peak value of a kernel is generally less than 1.0 due to finite vertical resolution and may peak at a slightly different height, resulting in the smoothing of the true atmospheric profile into the retrieved profile.
Aerosol extinction retrievals
Retrieval of aerosol extinction profiles was non-linear since the aerosol
extinction affects the radiative transfer in the forward model. The input
for the aerosol retrieval was the measurement vector of the O4 dSCDs at different elevation angles and an a priori state vector that decreased
exponentially with altitude with a scale height of 0.6 km and a surface
magnitude of 0.1 km-1. A single a priori profile choice is preferable
for a set of consecutive measurements where information content is
potentially limited since the a priori will always have some impact on the
retrieved profile (Rodgers, 2000). Otherwise, diurnal and
day-to-day trends in the retrieved profiles due to real atmospheric changes
could be indistinguishable from changes due to a variable a priori profile.
Sensitivity studies of the a priori choice were conducted by varying one
a priori parameter while keeping the remainder of settings the same:
Gaussian shape, Boltzmann shape, scale height of 0.3 km, or scale height of
1.2 km. The results of the sensitivity studies are shown in Sect. S8.2.
Our aerosol retrieval used an iterative algorithm based on the
Levenberg–Marquart method (Levenberg, 1944; Marquardt, 1963). For aerosol retrievals, the weighting function K is calculated using the a priori xa and the measurement vector y. The K of each retrieval depended on the state vector and changed depending on the determined aerosol extinction profile. The height resolution of the aerosol extinction vertical profile grid was 100 m with a maximum height of 4 km. A detailed description of the aerosol retrieval algorithm can be found in Frieß et al. (2006).
Trace gas retrievals
The retrievals of NO2 and SO2 vertical profiles were linear
because these weak absorbers do not significantly impact the radiative
transfer. The inputs for the NO2 and SO2 retrievals were the
measurement vectors of the NO2 and SO2 dSCDs at different
elevation angles, respectively, and an a priori state vector that decreased
exponentially (scale height is 0.6 km and surface magnitude is 30 and 10 ppb for SO2 and NO2, respectively). Sensitivity studies of the a priori choice were conducted by varying one a priori parameter while keeping the remainder of settings the same: Gaussian shape, Boltzmann shape, scale height of 0.3 km, or scale height of 1.2 km. The results of the sensitivity studies are shown in Sect. S8.2.
In this linear case, the forward model is independent of the atmospheric
state x, and the weighting function matrix represents the forward
model.
y=Fx,b+ε=Kx+ε
In our retrieval, the box air mass factors (AMFs) that are components of
K were modelled using the radiative transfer model in SCIATRAN
(Deutschmann et al., 2011; Frieß et al., 2010; Rozanov et al., 2005). The aerosol profiles retrieved in step 1 were used to recalculate the box AMFs for each trace gas retrieval since the extinction profiles varied. The height
resolution of the trace gas vertical profile grid was 100 m with a maximum
height of 4 km a.g.l.
Determination of retrieval errors
The retrieval covariance matrix S^ quantifies the
error of the state vector and is the sum of the independent sources of
error: smoothing error Ss, representing the retrieval's limited vertical resolution, and the retrieval noise SM, representing the uncertainty due to errors in the
measurement (S^=SM+Ss). The error covariance matrix produced by the retrieval does not include model parameter errors or forward model errors (Frieß et al., 2006). The error covariance matrix is calculated following Eq. (4):
S^=KTSε-1K+Sa-1-1.SE and Sa are the measurement
and a priori covariance matrices, respectively. In our retrievals,
Sa was determined by setting the relative error of
the a priori to 100 % and the extra-diagonal terms were set to zero. The
SE was the diagonal matrix of errors of the retrieved dSCDs as determined by the DOASIS retrievals.
Conversion of other instruments' data for comparison to MAX-DOAS data
Lidar and AERONET extinction data were converted to the MAX-DOAS aerosol
retrieval wavelength of 361 nm following Eq. (5):
E361nm=Eλ1⋅361nmλ1-∝.
Equation (5) accounts for the dependence of aerosol extinction on wavelength
based on the Angstrom exponent, ∝. AERONET 300–500 and 340–440 nm
Angstrom exponents were used to convert the 532 nm lidar aerosol extinctions
and the 380 and 340 nm AERONET AODs, respectively. The two resulting
AERONET AODs at 361 nm were then averaged. The Angstrom exponent was assumed
to be constant with altitude and representative of both field sites. The
similarity in trends in AODs and trace gas VCDs between the two sites can
indicate when the Angstrom exponent determined from Oski-Ôtin was valid
for both sites.
Due to the limited vertical resolution of the MAX-DOAS measurements,
MAX-DOAS vertical profiles of aerosol extinction and AODs can only be
directly compared to lidar profiles and AODs after smoothing the 361 nm
lidar profiles using the MAX-DOAS averaging kernel matrix, A (Rodgers and Connor, 2003). The lidar AODs referred to in the paper below and shown in plots are the smoothed AODs determined by vertically integrating the smoothed lidar vertical profiles of extinction at 361 nm unless otherwise stated.
The lidar profiles were averaged into the same altitude and temporal
intervals as the MAX-DOAS retrievals and then smoothed using the respective
matrix A following Eq. (6):
xs=xa+AxL-xa,
where xs is the smoothed lidar profile, xa is the MAX-DOAS retrieval a priori profile, and xL is averaged lidar profile. xs represents the (noise-free) vertical profile that the MAX-DOAS retrieval would produce if xL was the true atmospheric profile given the variable sensitivity of the MAX-DOAS retrieval with altitude. The deviation of xs from xL at each altitude depends on xa and the sensitivity of the MAX-DOAS to the atmosphere at that altitude. MAX-DOAS sensitivity to the true atmospheric state decreases with increasing altitude (Frieß et al., 2006) with
typical height resolutions of ∼200 m at lower altitudes, increasing to ∼700 m at higher altitudes. Therefore, the smoothing is generally expected to smooth the true profiles towards lower altitudes. Also, even if the two instruments viewed the same air mass, the retrieved and smoothed profiles are expected to differ at least slightly due to two factors. The first factor is the retrieval noise Gε, which is unknown since the true measurement error ε is unknown. G is the gain matrix, which describes the retrieval's sensitivity to the
measurements. The true smoothed profiles would be described using the
following Eq. (7):
xretrieved=xa+Axtrue-xa+Gε.
The second factor is that lidar vertical profiles are observed straight up,
are measured only above 100 m a.g.l., and are the least sensitive close to the
surface. The 0–100 m a.g.l. extinction in the lidar profiles was assumed to be equal to the average extinction measured between 100 and 200 m a.g.l. but the vertical profiles may have been variable below 150 m a.g.l. Uncertainty in the lidar vertical profiles is greatest in the lowest 150 m a.g.l., introducing uncertainty into the smoothed lidar profiles.
Error bars on the MAX-DOAS AODs, VCDs, and mixing ratios shown in figures
were obtained from the optimal estimation retrieval. Error bars on the lidar
and AERONET AODs are the standard error of the temporally averaged values
since these instruments have a finer time resolution than the MAX-DOAS
retrievals. Error bars on the active DOAS mixing ratios are the root-sum-square errors of the standard error of the averaged values and the average
error reported by the respective DOASIS retrievals. Error bars on the
Pandora VCDs are root-sum-square errors of the standard error of the average
values and the reported instrumental precision. Deming fit linear
regressions were performed using the Monte Carlo method, which included the
errors on the x and y data, with the “linfitxy” function in MATLAB (Browaeys, 2017). The 23 August and 3 September AERONET and Pandora data were also correlated with MAX-DOAS and lidar data by subtracting 30 min from the Oski-Ôtin data to account for the time of air mass transport
between the Fort McKay South and Oski-Ôtin given the wind speeds.
Results and discussion
This paper discusses results from largely cloud-free days when industrial
plumes were observed (23 August and 3, 4, 6, and 7 September) and 1 d with clean conditions (5 September). The 9 d are not discussed due to the presence of clouds most of the day.
Vertical profiles of wind speed and wind direction measured by the windRASS are
shown in Figs. 2 and 3, respectively. A summary of wind conditions and
pollution levels for each day is shown in Table 3.
Vertical profiles of wind speed: 23 August (a), 3 September (b), 4 September (c), 5 September (d), 6 September (e), and 7 September (f).
Vertical profiles of wind direction: 23 August (a), 3 September (b), 4 September (c), 5 September (d), 6 September (e), and 7 September (f).
Daytime wind and pollution conditions during the study
days.
DateWind directionsWind speedsWind shearPollution levels23 AugustMorning: N to ENELowMinimalLow to very highAfternoon: SE to SSW3 SeptemberVariable; mostly SE to SSWLow to moderateSignificantModerate to high4 SeptemberMostly S to SELow to moderateSignificantLow to moderate5 SeptemberWSW to WHighModerateVery low6 SeptemberN to NELow near the surface, high aloftSignificantLow to moderate7 SeptemberMorning: SSEMorning: lowSignificantLow to moderateAfternoon: SW to SSWAfternoon: moderate to high
The 23 August plume exhibited the greatest enhancements in aerosol and trace gas
pollution during the study. Wind directions in the morning were northerly to
east–north-easterly and south-easterly to south–south-westerly in the afternoon. Winds
were of relatively low speed with minimal wind shear. The pollution enhancement
periods were associated with southerly (S) winds, suggesting that air masses
rich in industrial emissions originated from the Syncrude and Suncor mining
areas south of the sites (Fig. 1). The pollution enhancements impacted both
sites (AMS 13 and Oski-Ôtin).
The 3 September plume exhibited moderate pollution levels. Pollution data are only presented from 11:00 LT onwards due to the presence of clouds before this time. Wind directions varied from south-easterly to south–south-westerly with occasional south-westerly to north-westerly winds. Significant wind shear was observed in the vertical profiles of wind. The pollution enhancements impacted both sites.
The 4 September plume exhibited moderate pollution levels. Wind directions were frequently southerly to south-easterly with intermittent periods of south-westerly and north-westerly winds. Significant wind shear was observed: wind directions
tended to rotate clockwise from south–south-easterly near the surface to
north-easterly as altitude increased. The limited afternoon wind data suggest
north-westerly winds. Wind speeds were low to moderate, tending to increase with
altitude. The pollution enhancements impacted both sites.
The 5 September plume exhibited the cleanest conditions and greatest wind speeds during the study. Winds were west–south-westerly to westerly. Both sites were impacted by air masses that passed over boreal forests.
The 6 September plume exhibited low to moderate pollution enhancements in the morning with low-pollution conditions in the afternoon. Winds were northerly to
north-easterly but varied over time and with altitude. Wind speeds tended to
be low at the surface but moderate to large at higher altitudes; significant
wind shear was present. Fort McKay South was impacted by emissions from
facilities north of the sites: Shell Jackpine and Muskeg River Mines, CNRL
Horizon, and Imperial Oil.
The 7 September plume exhibited moderate to low pollution. Wind directions were
south–south-east during the morning and south-west to south–south-west
during the afternoon. Significant wind shear was observed in the lowest 400 m a.g.l. between 09:00 and 11:00 LT and during the afternoon around 16:00 LT. Different air masses may have impacted the two sites.
Inter-comparisons of MAX-DOAS aerosol retrievals with lidar and AERONET data
The AODs from the MAX-DOAS, lidar, and AERONET sun photometer instruments
exhibited similar temporal trends on 23 August and 3, 4, 5, and 7 September (Figs. 4–9a). Note that the measured and optimal estimation modelled O4 dSCDs showed good agreement for all these days with linear regression slopes of 0.99 and R2=0.91–0.98 (Table S10). The MAX-DOAS AODs were statistically different from the lidar and AERONET AODs for approximately half the data, even when the two sites experienced the same air masses. This result is expected based on three factors: (1) the different vertical extents of the atmosphere observed by the instruments, (2) temporal variability in the lidar S ratio, and (3) the limited sensitivity of the MAX-DOAS measurements at higher altitudes. These factors will be discussed below to evaluate the performance of the MAX-DOAS AOD retrievals under various atmospheric conditions.
The 23 August AODs from MAX-DOAS, lidar (S ratio =44sr> 14:30 LT), and AERONET (30 min) (a); AODs from MAX-DOAS, lidar
(S ratio =25 sr), and AERONET (b); MAX-DOAS and Pandora SO2 VCDs (c); MAX-DOAS and Pandora NO2 VCDs (d); MAX-DOAS 0–100 m and active DOAS SO2 mixing ratios (e); MAX-DOAS 0–100 m and active DOAS NO2 mixing ratios (f); MAX-DOAS VCDs and active DOAS mixing ratios of
SO2(g); and MAX-DOAS VCDs and active DOAS mixing ratios of NO2(h). MAX-DOAS error bars were obtained from the optimal estimation retrieval. Error bars on the lidar and AERONET AODs are the standard error of the temporally averaged values. Error bars on the Pandora VCDs and active DOAS mixing ratios are root-sum-square errors of the standard error from temporal averaging and instrumental precision and the DOASIS
retrieval error, respectively. See Sect. 2.3.3 for details.
The 3 September AODs from MAX-DOAS, lidar, and AERONET (a); MAX-DOAS and Pandora SO2 VCDs (b); MAX-DOAS and Pandora NO2 VCDs (c); MAX-DOAS 0–100 m and active DOAS SO2 mixing ratios (d); MAX-DOAS 0–100 m and active DOAS NO2 mixing ratios (e); MAX-DOAS VCDs and active DOAS mixing ratios of SO2(f); and MAX-DOAS VCDs and active DOAS mixing ratios of NO2(g). MAX-DOAS error bars were obtained from the
optimal estimation retrieval. Error bars on the lidar and AERONET AODs are
the standard error of the temporally averaged values. Error bars on the
Pandora VCDs and active DOAS mixing ratios are root-sum-square errors of the
standard error from averaging and instrumental precision and the DOASIS
retrieval error, respectively. See Sect. 2.3.3 for details.
The 4 September AODs from MAX-DOAS, lidar, and AERONET (a);
MAX-DOAS and Pandora SO2 VCDs (b); MAX-DOAS and Pandora NO2 VCDs (c); MAX-DOAS 0–100 m and active DOAS SO2 mixing ratios (d); MAX-DOAS 0–100 m and active DOAS NO2 mixing ratios (e); MAX-DOAS VCDs and active DOAS mixing ratios of SO2(f); and MAX-DOAS VCDs and active DOAS mixing ratios of NO2(g). MAX-DOAS error bars were obtained from the
optimal estimation retrieval. Error bars on the lidar and AERONET AODs are
the standard error of the temporally averaged values. Error bars on the
Pandora VCDs and active DOAS mixing ratios are root-sum-square errors of the
standard error from averaging and instrumental precision and the DOASIS
retrieval error, respectively. See Sect. 2.3.3 for details.
The 5 September AODs from MAX-DOAS, lidar, and AERONET (a); MAX-DOAS SO2 VCDs (b); MAX-DOAS NO2 VCDs (c); MAX-DOAS 0–100 m and active DOAS SO2 mixing ratios (d); MAX-DOAS 0–100 m and active DOAS NO2 mixing ratios (e); MAX-DOAS VCDs and active DOAS mixing ratios of SO2(f); and MAX-DOAS VCDs and active DOAS mixing ratios of NO2(g). MAX-DOAS error bars were obtained from the optimal estimation
retrieval. Error bars on the lidar and AERONET AODs are the standard error
of the temporally averaged values. Error bars on the active DOAS mixing
ratios are root-sum-square errors of the standard error from averaging and
the DOASIS retrieval error. See Sect. 2.3.3 for details.
The 6 September AODs from MAX-DOAS, lidar, and AERONET (a);
MAX-DOAS SO2 VCDs (b); MAX-DOAS NO2 VCDs (c); MAX-DOAS 0–100 m and active DOAS SO2 mixing ratios (d); MAX-DOAS 0–100 m and active DOAS NO2 mixing ratios (e); MAX-DOAS VCDs and active DOAS mixing ratios of SO2(f); and MAX-DOAS VCDs and active DOAS mixing ratios of NO2(g). MAX-DOAS error bars were obtained from the optimal estimation retrieval. Error bars on the lidar and AERONET AODs are the standard error of the temporally averaged values. Error bars on the active DOAS mixing ratios are root-sum-square errors of the standard error from averaging and the DOASIS retrieval error. See Sect. 2.3.3 for details.
The 7 September AODs from MAX-DOAS, lidar, and AERONET (a);
MAX-DOAS and Pandora SO2 VCDs (b); MAX-DOAS and Pandora NO2 VCDs (c); MAX-DOAS 0–100 m and active DOAS SO2 mixing ratios (d); MAX-DOAS 0–100 m and active DOAS NO2 mixing ratios (e); MAX-DOAS VCDs and active DOAS mixing ratios of SO2(f); and MAX-DOAS VCDs and active DOAS mixing ratios of NO2(g). MAX-DOAS error bars were obtained from the
optimal estimation retrieval. Error bars on the lidar and AERONET AODs are
the standard error of the temporally averaged values. Error bars on the
Pandora VCDs and active DOAS mixing ratios are root-sum-square errors of the
standard error from averaging and instrumental precision and the DOASIS
retrieval error, respectively. See Sect. 2.3.3 for details.
Impact of instrumental vertical sensitivity on AOD
AERONET AODs were generally significantly greater than MAX-DOAS and lidar
AODs, except during the greatest pollution events (Figs. 4–9a). During
the low-pollution day of 5 September, AERONET AODs reached a maximum of
0.15±0.00 while maximum MAX-DOAS AODs were 0.08±0.01 (Fig. 4b). On 5 September the MAX-DOAS and AERONET AODs had a slope of linear correlation of 1.04±0.08 (R2=0.77) but had a linear intercept of -0.08±0.01 km-1. This negative intercept can be attributed to
aerosol loading above the boundary layer that was observed by the sun
photometer but not by the MAX-DOAS. This result is expected because the
AERONET sun photometer observed aerosol extinction throughout the entire
column (tropospheric and stratospheric) while the MAX-DOAS and smoothed
lidar profiles observed up to 4 km. Further, the MAX-DOAS retrieved and
smoothed lidar profiles likely only represented aerosol extinction below 2 km because of the exponentially decreasing a priori profiles used and the
decreasing sensitivity of the MAX-DOAS retrieval with increasing altitude
(see Sect. S8 for explanation of the a priori shape and scale height choice). The MAX-DOAS and smoothed lidar AODs are, therefore, expected to be significantly smaller than the AERONET AODs when the aerosol extinction in the boundary layer was “clean” and contributed a small fraction to the total tropospheric extinction (e.g., Fig. 10; 23 August, 09:10 LT). MAX-DOAS AODs were also significantly smaller than AERONET AODs even under moderately polluted conditions when the magnitudes of the aerosol extinction remained enhanced compared to background above the boundary layer. For example, on 4 September the extinctions between the boundary layer top and 4 km could be relatively large, about one-third of the near-surface extinctions (Figs. 10, 15a and S9a), leading to much smaller MAX-DOAS AODs than AERONET AODs (Fig. 6a). Note that increasing the scale height of the 4 September a priori profile from 0.6 to 1.2 km only slightly increased the MAX-DOAS AODs with a linear regression slope of 1.14±1.13 and intercept =0±0 km-1 (Table S21 and Figs. S17, S24). The AODs from both a priori profiles were ∼50 % of the AERONET AOD values, an expected result since a significant contribution to the total AOD was expected to be present >3 km (Table S21, Fig. S9a). Aerosol loading can be non-trivial in the free
troposphere since fine-mode particles can remain in the atmosphere for days (Zhong and Zaveri, 2017). Dust and smoke aerosol types are transported most efficiently to the free troposphere from the boundary layer compared to other types. Once in the free troposphere, aerosols have much longer residence compared to within the boundary layer. Therefore, enhanced loading of dust and smoke aerosols present in the free troposphere in the AOSR may be due to local sources since dust and smoke are emitted by the industrial activities and biomass burning but also could be transported long distances (e.g., originating from fires in British Columbia). Forest fire smoke has been observed above the boundary layer by airborne lidar measurements in AOSR (see Figs. 12 and 13 and discussion of elevated forest fire plumes in Aggarwal et al., 2018). Therefore, the aerosol extinction above the boundary layer in the AOSR could contribute to a significant fraction of the total atmospheric AOD on certain days depending on the emission and transport of aerosols. Globally, Bourgeois et al. (2018) found that the monthly averaged fractional
contribution of the free troposphere extinction to the total AOD obtained
from satellite observations was greater in northern mid-latitudes (25 %)
compared to southern mid-latitudes (13 %), attributed to the larger number of emission sources and convection activity. While monthly average
contribution was 25 %, the contribution likely varies significantly on a
day-to-day basis, particularly in a region such as the AOSR that has
numerous emission sources, both locally and upwind. The lidar measurements
indicate that the contribution of the AOD above 2 km can vary significantly,
as indicated by a comparison of aerosol extinction vertical profiles from
4 September and 23 August (Fig. S9). These results indicate that the ratio of the MAX-DOAS AODs to AERONET AODs depends on the location of the aerosol
extinction within the tropospheric profile. The use of simple linear
regressions to evaluate the performance of MAX-DOAS AOD retrievals using sun
photometer AODs may be appropriate only when the aerosol extinction in the
boundary layer dominates the total tropospheric AOD.
Examples of lidar vertical profiles of aerosol extinction (averaged into MAX-DOAS retrieval height intervals and times) on 23 August and
4 September.
Impact of S ratio variability on lidar AODs
MAX-DOAS AODs significantly exceeded the smoothed lidar AODs during the most
polluted periods (Fig. 4b). This result is unexpected, given that the
instruments' AODs should ideally be equal when the instruments observed the
same air masses. However, the deviation can be explained by variation in the
lidar S ratio. The S ratio of 25 sr (steradian) appears accurate during
relatively clean periods (e.g., 5 September, 23 August morning), but an
underestimation under the industrially polluted conditions of the afternoon
of 23 August (Figs. 4b and 12). Modelled S ratios for 23 August were 21–28 sr during the low-pollution morning and 36–44 sr during the peak pollution enhancement at ∼ 16:50 LT (Table 4). The morning S ratios were calculated using the refractive indices of toluene or kaolinite based on the dominance of organic particles and dust in the region during background
atmospheric conditions (Fig. 11a). The afternoon S ratios were modelled
using the refractive index of sulfate particles based on the significant
enhancement in sulfate particle loading (Fig. 11a). Increased loading of
sulfate particles tends to increase the S ratio. Note that the 16:50 LT
S ratios were greater than the morning S ratios for all refractive indices
because the particle size distribution of the industrial plume (fine-mode
dominated) increased the S ratio. The modelled variability in the S ratio is
supported by lidar measurements in the AOSR in 2018 that allowed determination of temporal and vertical variability in the S ratio (Strawbridge et al., 2018). Measured S ratios ranged from 20 to 60 sr within the boundary layer at Oski-Ôtin in 2018 (Fig. S6).
Modelled lidar S ratios (sr) for selected periods on 23 August
using refractive indices (RIs) of different particles.
Local timeRI ofRI ofRI of sulfatetoluenekaoliniteaerosol09:1021253009:3025283414:1017333814:3018333716:3031323816:5036404417:15364044
Near-surface particle compositions on 23 August (a), 3 September (b), 4 September (c), 5 September (d), 6 September (e), and 7 September (f). Note the different y axis scale for 23 August and that nitrate and refractory black
carbon are shown multiplied by 10.
Based on these results, lidar vertical profiles of aerosol extinction were
retrieved using an S ratio of 44 sr for the extinction below the free
troposphere after 14:30 LT on 23 August. As shown in Fig. 4a, the updated
lidar AODs are in more reasonable agreement with the MAX-DOAS and AERONET
AODs compared to the original lidar AODs shown in Fig. 4b. The linear
regression of the MAX-DOAS and updated lidar AODs has a slope of 1.15±0.02 with an intercept of -0.01±0.02 (R2=0.97) instead of the 2.18±0.03 found for the original lidar AODs (R2=0.97) (Table S3).
Modelling S ratios using particle data measured at the near-surface appears
to be valid during 23 August because the vertical profile was relatively well mixed. A well-mixed boundary layer was indicated by the similarity in
temporal trends between the active DOAS mixing ratios and MAX-DOAS VCDs and
between the AODs and the surface particle loading (Figs. 4g, h and 11a). However, if the distribution of particles in space is inhomogeneous,
this method cannot be used to determine the S ratio of the total boundary
layer.
Results from 3 and 6 September illustrate that near-surface measurements of
particle properties can be invalid for modelling the total column S ratio
due to complex vertical profiles of particles. Despite near-surface
enhancements in sulfate particles (Fig. 11) and SO2 mixing ratios
observed by active DOAS (Fig. 5f), the MAX-DOAS and lidar AODs were very
similar after 11:30–17:00 LT on 3 September (Fig. 5a). The MAX-DOAS and Pandora SO2 VCDs were moderate compared to the enhancements on 23 August, suggesting that the sulfate enhancements were confined mainly near the surface after 11:30 LT. Due to wind shear, the near-surface air (<200 m a.g.l.) was often impacted by industrial pollution from the south, while the air at higher altitudes was impacted by less polluted regions
(north-westerly and northerly winds), particularly around 14:00 LT (Fig. 3b).
Thus, the S ratio of 25 sr was representative of the total boundary layer
after 11:30 LT despite sulfate enhancements at the surface, leading to similar magnitudes of MAX-DOAS and lidar AODs. S ratios modelled using the
near-surface measurements of particles during the afternoon of 3 September would have overestimated the S ratio within the total boundary layer.
Similarly, near-surface measurements of particles would not represent the
total boundary layer on 6 September due to an elevated industrial plume. The
MAX-DOAS NO2 VCDs remained enhanced while the active DOAS mixing ratios rapidly decreased from ∼7 to ∼1 ppb (Fig. 8g). The MAX-DOAS AODs approached the AERONET AODs around noon (Fig. 8a), maximizing around the time that the lidar observed elevated vertical profiles of aerosol extinction. These results suggest that elevated plumes from the industrial facilities to the north of Fort McKay South (Fig. 1) increased the S ratios at higher altitudes. S ratios modelled using the surface data during this time, therefore, would have underestimated the average S ratio within the boundary layer.
These results suggest that the MAX-DOAS retrievals of AODs performed well
when the vertical extent of instrumental viewing and S ratio variability are
considered.
Comparison of MAX-DOAS vertical profiles of aerosol extinction with averaged and smoothed lidar vertical profiles
MAX-DOAS vertical profiles of aerosol extinction are compared to averaged
and smoothed lidar vertical profiles in Figs. 12 to 18.
The 23 August vertical profiles of aerosol extinction (361 nm)
from S ratio of 25 sr: averaged lidar (a), smoothed lidar (b), and MAX-DOAS retrieved measurements (c).
The 23 August vertical profiles of aerosol extinction (361 nm)
from S ratio of 44 sr within the plume > 14:30 LT:
averaged lidar (a), smoothed lidar (b), and MAX-DOAS retrieved measurements (c).
The 3 September vertical profiles of aerosol extinction (361 nm)
from averaged lidar (a), smoothed lidar (b), and MAX-DOAS retrieved measurements (c).
The 4 September vertical profiles of aerosol extinction (361 nm)
from averaged lidar (a), smoothed lidar (b), and MAX-DOAS retrieved measurements (c).
The 5 September vertical profiles of aerosol extinction (361 nm)
from averaged lidar (a), smoothed lidar (b), and MAX-DOAS retrieved measurements (c). Omitted data in the afternoon were measurements of cirrus clouds.
The 6 September vertical profiles of aerosol extinction (361 nm)
from averaged lidar (a), smoothed lidar (b), and MAX-DOAS retrieved measurements (c).
The 7 September vertical profiles of aerosol extinction (361 nm) from averaged lidar (a), smoothed lidar (b), and MAX-DOAS retrieved measurements (c).
Smoothing alters the shape and magnitude of the averaged lidar profiles in
several ways. Smoothing the averaged lidar profiles generally “compresses”
the profiles by vertically attributing extinction at higher altitudes to
lower altitudes (compare a and b in Figs. 12–18). This result is
expected due to the decreasing sensitivity of the MAX-DOAS retrieval with
increasing altitude apparent in the averaging kernels (Fig. S7). The
smoothing also replaces lidar aerosol extinction above ∼1.5 km a.g.l. with the (small) a priori extinction values because the MAX-DOAS measurements have little information content at high altitudes (Fig. S7).
This effect is apparent when comparing the 3 September averaged and smoothed
lidar profiles above 1.5 km a.g.l. (Fig. 14). Profiles that were relatively
uniform within a few hundred metres of the surface can sometimes be smoothed
into apparently elevated profiles because the averaging kernel attributes
much of the extinction from altitudes aloft to one altitude bin closer to
the surface. For example, the averaging kernels for the 4 September 14:10 LT
retrieval for altitudes 0.55 to 1.25 km peak at 0.45 km rather than at their respective height (Figs. S7 and S8). Conversely, the smoothing can
transform vertically narrow and distinctly elevated profiles near the
surface into exponentially decreasing profiles due to the limited vertical
resolution of the retrieval (see the 09:30 LT profile in Fig. 17). Therefore, interpretation of the retrieved MAX-DOAS profiles must account for the effects of smoothing on the true atmospheric profiles.
On 23 August, the MAX-DOAS and smoothed lidar vertical profiles (S ratio =25 sr) exhibited similar temporal trends and vertical enhancements within the boundary layer (Fig. 12). The magnitudes of aerosol extinctions were
consistent between the smoothed lidar and MAX-DOAS vertical profiles in the
morning, supporting the hypothesis that the S ratio of 25 sr was appropriate
for “clean” periods. In contrast, the MAX-DOAS extinctions exceeded the
smoothed lidar extinctions in the afternoon (Fig. 12). Using an S ratio of
44 sr within the afternoon plume (discussed in Sect. 3.1.2) resulted in smoothed lidar profiles consistent with the MAX-DOAS profiles (Fig. 13). While temporal trends and overall magnitudes were similar, MAX-DOAS retrievals tended to exhibit more distinctly elevated profiles than the smoothed lidar profiles. The use of a constant S ratio within the plume may have caused the lidar profiles to appear more vertically uniform than the true profiles since S ratios can maximize where extinction peaks (Fig. S6). Also, the MAX-DOAS viewing geometry observed air masses south of the field site, closer to industrial sources, where the vertical profiles may have been less well mixed. Finally, MAX-DOAS measurement errors can be mapped into the
retrieved profile, leading to uncertainties, but are probably only important
at higher altitudes where the measurements contain little information content. Deviations in the MAX-DOAS profiles from the smoothed lidar
profiles after 17:00 LT can be attributed to reduced light levels and the
longer retrieval time, reducing signal-to-noise ratio and the probability of
the viewed air masses changing significantly within the time required to
capture the measurements for the retrieval, respectively.
For the 3, 4, and 7 September (morning) comparisons, the MAX-DOAS
retrieved profiles generally captured the same temporal and vertical trends
in extinction enhancements as the smoothed lidar profiles, but the lidar
extinctions were smaller than the MAX-DOAS extinctions (Figs. 14, 15, 18). The S ratio of 25 sr probably underestimated the true values given the
presence of sulfate particles (Fig. 11b, c, f) and enhanced SO2
VCDs (Figs. 5b, 6b, 9b). On 5 September the S ratio of 25 sr is expected to be appropriate due to the clean conditions. The magnitudes of the MAX-DOAS extinctions were unexpectedly greater than the smoothed lidar extinctions but were generally equal within error (Fig. 16). The 6 September MAX-DOAS aerosol retrievals appear noisier than the smoothed lidar profiles (Fig. 17). The elevated plumes present in the MAX-DOAS retrievals but not in
the lidar profiles may be related to an increased S ratio due to the impact
of plumes from the north (Fig. 8g). On 7 September the MAX-DOAS aerosol extinction profiles were of greater magnitude and different in vertical profile shape compared to the smoothed lidar profiles after 12:00 LT. The deviation can be attributed to significant wind shear and rapid temporal variation in the wind profiles after 12:00 LT (Figs. 2f, 3f). Aerosol extinction magnitudes varied by up to a factor of 5 within 10 min in the
afternoon (Fig. S11). These conditions violate two assumptions of the
MAX-DOAS retrievals: low horizontal inhomogeneity and that the spectra
measured during the retrieval time observed the same air mass. Although the
MAX-DOAS retrievals of AOD were consistent with the smoothed lidar AODs, the
temporal and vertical resolutions of the MAX-DOAS retrievals were
insufficient to retrieve accurate vertical profile shapes. The afternoon
MAX-DOAS vertical profile retrievals are, therefore, not expected to
represent the true atmospheric state.
Evaluation of MAX-DOAS trace gas retrievalsComparison of MAX-DOAS and pandora trace gas VCDs
The MAX-DOAS and Pandora SO2 and NO2 VCDs exhibited similar
temporal trends over the 4 d of comparison, except during the afternoon of 7 September (Fig. 4c, d; b, c in Figs. 5, 6 and 9).
On 23 August the MAX-DOAS and Pandora VCDs were strongly correlated
(R2>0.80) with linear regression slopes of
1.55±0.07 and 2.20±0.07 for SO2 and NO2 VCDs,
respectively (Table S4). Greater trace gas enhancements were also observed
near the surface at Fort McKay South compared to Oski-Ôtin through
in situ measurements by the Wood Buffalo Environmental Association (WBEA)
(Wood Buffalo Environmental Association, 2019) (Fig. S12) with slopes of the linear regressions of 1.42±0.05 (R2=0.91) and
1.93±0.07 (R2=0.61) for SO2 and NO2, respectively
(Table S2). The strong correlations between the trace gas measurements
between the sites indicate that the same air mass impacted both sites but
that a more central (higher-concentration) portion of the plume impacted
Fort McKay South or that a significant horizontal dilution of the plume occurred
during transport.
The NO2 VCDs had a regression slope greater than that of the
in situ NO2 measurements (Table S2). NOx may have been lost at a faster rate near the surface during transport due to deposition to the surface (e.g., the boreal forests). Transport times between the sites were relatively long (∼30 min) on this day due to low wind speeds below 600 m a.g.l. (Fig. 2). NOx is lost through surface deposition and photochemical conversion to HONO and HNO3 (Finlayson-Pitts et al., 2003; Wojtal et al., 2011). HONO might be subsequently released as NO and OH, but the HNO3 loss will be virtually permanent.
8NO2+NO2+H2O→HNO3(aq.surface)+HONO(g)9HONO+hν→NO+OH
On 4 September the MAX-DOAS and Pandora VCDs exhibited similar temporal trends and were often equal within error (Fig. 6b, c). The slopes of the linear
correlations of the SO2 and NO2 VCDs were 1.10±0.33
(R2=0.51) and 0.95±0.07 (R2=0.85), respectively (Table S6). The greater variability in SO2 between the sites compared to NO2 is consistent with the in situ data between Fort McKay South
and the WBEA Bertha Ganter site (Fort McKay North; 57.189428, -111.640583) with R2=0.7 and 0.92 for SO2 and NO2, respectively (Table S6) (Wood Buffalo Environmental Association, 2019). SO2 plumes are more localized in the AOSR, originating mostly from large industrial stacks and fewer sources compared to NO2 (Zhang et al., 2018) (Tables S1, S2). Note that when MAX-DOAS SO2 VCDs were
significantly greater than Pandora SO2 VCDs around noon, the SO2 mixing ratios at Fort McKay South were approximately double those at the Bertha Ganter (Fort McKay) site (Fig. S12). These results suggest that the MAX-DOAS performed well in retrieving accurate VCDs of SO2 despite the weaker linear correlation with the Pandora VCDs. The two sites appear to have largely experienced the same air masses within a small temporal period (<20 min), due to higher wind speeds relative to 23 August and 3 September (Fig. 2a, c). Higher wind speeds likely reduced the maximum
enhancements in trace gas VCDs compared to 23 August and 3 September due to greater dispersion. Wind shear on 4 September (Fig. 3c) may also have transported only certain altitudes of the elevated plumes from south of Fort McKay South to the sites. In contrast, wind shear on 23 August was limited within 500 m a.g.l. (Fig. 3a).
On 3 and 7 September, the MAX-DOAS and Pandora VCDs demonstrated weak linear
correlations (R2<0.2) (Tables S5 and S9).
The 3 September VCD correlations are inconclusive due to the limited number of data points and relatively little variation in the Pandora VCDs. The
MAX-DOAS VCDs tended to be higher than the Pandora VCDs. Unlike on 23 August,
an examination of the in situ data between sites is not helpful due to the
significant wind shear on 3 September and the presence of elevated plumes. Based on the good agreement between the MAX-DOAS and Pandora VCDs on 4 September with similar VCD magnitudes, the apparent overestimation could be due to different air masses experienced by the two sites.
On 7 September the MAX-DOAS were similar to the Pandora trace gas VCDs before
13:30 LT but were much larger after (Fig. 9b, c). The deviation of the
MAX-DOAS VCDs is an expected result given the rapid spatial and temporal
variation in the wind profiles (discussed in Sect. 3.1.3). Errors of the trace gas retrievals can be expected to be even greater than the aerosol retrieval errors because the retrieved aerosol profiles were used as forward model parameters in the trace gas retrieval. The afternoon MAX-DOAS trace gas
retrievals on 7 September are not expected to represent the true atmospheric
state. Note that the strength of correlation between the measured and
optimal estimation modelled dSCDs of the trace gases was weaker for the
retrievals from 7 September (slopes =0.9 and R2=0.67–0.74) compared to 23 August (slopes =0.99 and R2=0.97–0.98) and 4 September (slopes =0.97–9.88 and R2=0.91–0.98) (Table S11). These statistical results are consistent with the good performance of the trace gas retrievals on 23 August and 4 September compared to relatively poor performance on 7 September.
Inter-comparisons of the Pandora and MAX-DOAS VCDs show that the MAX-DOAS
retrievals of trace gas VCDs performed well under low to moderate
wind speeds and when vertical profiles of pollution were relatively constant
within the retrieval period.
Comparison of MAX-DOAS 0–100 m retrieval with active DOAS mixing ratios
The 0–100 m a.g.l. MAX-DOAS trace gas retrievals are shown with the
active DOAS mixing ratios in Figs. 4e, f and 5–9d, e. The
MAX-DOAS retrievals generally captured the active DOAS temporal trends but
tended to overestimate the magnitudes. The MAX-DOAS retrieval yields an
estimate of the average concentration within the 0–100 m layer, which is
larger than the surface value in case of uplifted layers. Therefore, in situ
near-ground instruments, such as active DOAS, are required when accurate
surface mixing ratios are required.
The MAX-DOAS retrievals were most consistent with the active DOAS
measurement during the late afternoon of 23 August (Fig. 4e, f). SO2
was at its highest levels and assumed to be relatively well mixed within the
boundary layer based on the similarity in the temporal trends in SO2
VCDs and surface mixing ratios (Fig. 4g) and the uniformity of the lidar
vertical profiles <1 km a.g.l. (Fig. 13a). The mixing ratios
were equal within error during the morning and after 14:00 LT with some
differences in the early afternoon that may be due to the different viewing
geometry. On days other than 23 August, the uncertainty in the surface
retrieval is often too high for reliable comparison when the near-surface
when SO2 and NO2 were <20 and <10 ppb,
respectively. Overall, the MAX-DOAS retrievals of 0–1000 m performed well,
considering the frequently complex vertical profiles observed during the
study.
Temporal trends of MAX-DOAS Trace gas VCDs and active DOAS mixing ratios
Active DOAS mixing ratios are shown with MAX-DOAS VCDs in Fig. 4g, h and
in Figs. 5–9f, g. The VCDs and mixing ratios exhibited similar temporal
trends on 23 August and 4–6 September (Figs. 4g, h; 6 and 7f, g) but not
on 3 and 7 September (f, g in Figs. 5 and 9). The similar temporal trends
in VCDs and mixing ratios observed on 23 August are consistent with the limited vertical wind shear and low to moderate wind speeds, as discussed
previously. In contrast, the ratio of VCDs to mixing ratios sometimes varied
even during short periods on 4 and 6 September. If the boundary layer is
well mixed, the active DOAS mixing ratios and MAX-DOAS VCDs are expected to
have similar temporal trends during short periods since the boundary layer
is expected to be effectively constant. On 4 September, the temporal trends were
very similar until ∼ 13:30 LT, when the rapid decrease in trace
gas mixing ratios was not reflected in the VCDs (Fig. 6f, g), indicating
elevated pollution plumes that are apparent in the lidar measurements (Fig. S7a). These observations are a testament to the ability of MAX-DOAS to
observe elevated pollution plumes not detectable at the surface. The
differences in the short-term trends in VCDs and mixing ratios are
consistent with the wind profile data around 13:30 LT on 4 September, which
indicates westerly to northwesterly wind directions <300 m a.g.l.
that are expected to result in relatively clean air near the surface (Fig. 3c). Although measurements of the wind profiles above ∼250 m a.g.l. were unavailable, southerly winds aloft are suggested by the trace gas VCDs remaining enhanced until ∼ 15:00 LT. While significant
enhancements of trace gas near the surface tend to contribute to enhanced
VCDs, the opposite may not always occur: elevated plumes that cause enhanced
VCDs may not result in large surface mixing ratios (Fioletov et al., 2016).
The observations in this study indicate that elevated enhancements may also
result from vertical wind shear. Techniques for estimating emissions from
industrial facilities must account for the possibility that different
vertical portions of plumes can be transported in different directions. Such
complex pollution conditions require pollution monitoring techniques such as
MAX-DOAS that can detect elevated pollution plumes. In addition to being
able to observe elevated plumes that are under-sampled by in situ, ground
instruments, MAX-DOAS can be used to estimate emissions when deployed using
the mobile MAX-DOAS technique (Davis et al., 2019).
MAX-DOAS retrievals of vertical profiles of SO2 and NO2
MAX-DOAS retrievals of vertical profiles of SO2 and NO2 are shown in Fig. 19. Unlike the aerosol profiles, co-located measurements of the
trace gas vertical profiles were generally not available. The magnitude and
vertical location of the pollution were both highly dependent on wind direction
and wind shear. The greatest trace gas enhancements occurred under
south–south-easterly wind directions (Figs. 3 and 19) where pollution
originated from the greatest sources of SO2 and NO2 to the south (Fig. 1; Tables S1, S2). The MAX-DOAS retrievals performed well in terms of the profile shapes expected based on the wind profiles or evidence of elevated plumes. For example, trace gas pollutants in the MAX-DOAS
retrievals were confined largely to <200 m on the mornings of 4 and 7 September (Fig. 19c, f) as expected from the wind shear (Fig. 3). The elevated profiles of SO2 on 3 September before noon and during the
afternoon of 4 September are consistent with the results discussed previously.
MAX-DOAS vertical profiles of SO2 (left column)
and NO2 (right column): 23 August (a), 3 September (b), 4 September (c), 5 September (d), 6 September (e), and 7 September (f). Note the different colour scale maximum for 23 August and 3 September.
Aircraft measurements of trace gases on 3 September allow some comparison of the MAX-DOAS retrieved profiles. A vertical profile of SO2 measured during an aircraft spiral ascent at ∼ 14:27 LT in the vicinity of Fort
McKay South (Fig. 20) is consistent in magnitude and shape with the MAX-DOAS
retrieved vertical profile for 11:00–11:20 LT (Fig. 20). The MAX-DOAS 11:10 LT profile was used for comparison because it appears to have observed the same plume as the aircraft spiral. Although these two profiles cannot be directly compared due to the differences in time and vertical resolutions, the aircraft profile indicates that the magnitudes and elevated shape of the
MAX-DOAS profiles of SO2 are reasonable. The elevated SO2 plumes measured by the aircraft and MAX-DOAS could have originated from upgrader stacks at either the Syncrude or Suncor facilities south of Fort McKay South. The aircraft also passed over Fort McKay South at 16:32 LT, measuring 30 ppb of SO2 and 5 ppb of NO2 at 395 m a.g.l. The MAX-DOAS retrieval for 16:20–16:40 LT had maximum SO2 values of 57(±19) ppb at 350 m
and maximum NO2 values of 10(±5) ppb at 650 m. Note that the
active DOAS measured 20(±0.1) ppb of SO2 and 4.3(±0.1) ppb of NO2 near the surface. These measurements,
therefore, suggest that elevated plumes were present and that the MAX-DOAS
retrieved magnitudes are reasonable.
The 3 September vertical profiles of SO2 (ppb) from an aircraft spiral measurement (14:26–14:28 LT) and MAX-DOAS retrieved SO2 vertical profile (11:10 LT). Aircraft spiral is shown in a
Google Earth plot (b).
Advantages of MAX-DOAS
MAX-DOAS has an advantage over the zenith lidar technique in detecting
aerosol extinction since lidar retrievals cannot detect close to the surface
due to challenges with signal overlap (Zieger et al., 2011). Quantifying aerosol extinction from lidar measurements also requires additional knowledge (i.e., the S ratio) (Wagner et al., 2004), as has been highlighted in this paper. The advantage of the MAX-DOAS over the sun photometer (in direct sun viewing mode) is the ability to determine vertical profiles of pollutants versus only total columns. The MAX-DOAS is complementary to active DOAS and other point source measurements when pollution within the boundary layer is vertically inhomogeneous (see Sect. 3.2.3). While surface level local measurements of pollutants are often important for applications such as health exposure studies, they may fail to provide the full picture of the total boundary layer pollution. Such in situ measurements provide highly localized information with little information about elevated plumes that may mix down to the surface downwind. MAX-DOAS allows remote sensing of air masses over longer path-lengths, even if plumes are elevated. The MAX-DOAS method is advantageous over satellite measurements when plumes are localized and can provide more information on near-surface trends.
Limitations of the inter-comparisons in this study
A limitation to validating the MAX-DOAS AODs against lidar and sun
photometer data was the different viewing geometry and slightly different
locations. Also, Angstrom exponents used to convert the lidar extinctions to
the MAX-DOAS retrieval wavelength would ideally be measured at Fort McKay
South. Application of a single S ratio modelled from particle measurements
from the near-surface to the entire lidar vertical profile can introduce
errors since the S ratio may vary vertically (see Sect. 3.1.2 and Fig. S6). The S ratio can be significantly non-uniform with altitude when the vertical
profile is composed of layers of anthropogenic (urban, biomass burning)
and/or biogenic aerosols or mixtures of the two. Even if a layer is well mixed,
the lidar ratio can change with height if the vertical profile of relative
humidity is non-uniform (Weitkamp, 2005).
The MAX-DOAS trace gas VCDs should ideally be compared with a co-located
Pandora instrument given the possibility of horizontal inhomogeneity between
the sites. Validation of the MAX-DOAS 0–100 m retrieval using the
active DOAS mixing ratios was complicated by the lowest viewing elevation
angle observing 5 m above the active DOAS light path. The MAX-DOAS
“surface” retrieved values are only expected to be equal to the
active DOAS values when the air masses were well mixed within 0–100 m a.g.l. A more thorough validation of the MAX-DOAS near-surface retrievals could be achieved with trace gas measurements at multiple heights within 100 m a.g.l. from a tall tower.
Summary
In this study, data from a diverse range of instruments have allowed an
expansive characterization of the MAX-DOAS retrievals of aerosol extinction,
NO2 and SO2. The retrievals performed well when capturing the
aerosol loading within the boundary layer. The exception was under
conditions of rapid variation in the vertical profiles of pollutants during
the retrieval period. The ratio of the MAX-DOAS to sun photometer AODs
depended on the vertical location of the aerosol extinction within the
atmospheric column. Direct inter-comparisons of AODs between instruments
must account for the relative spatial extents observed. The comparison of
MAX-DOAS and lidar data combined with S ratio modelling indicated that
accurate S ratio values are essential to retrieve accurate profiles of
aerosol extinction from lidar measurements when particle composition or size
distribution varies significantly temporally or spatially. Direct comparison
of MAX-DOAS and lidar AODs should be made with caution when knowledge of the
S ratio value(s) is limited. S ratios can be estimated from measurements of
particle size distribution and composition using Mie scattering modelling.
However, near-surface measurements of particles should only be used to model
S ratios when the boundary layer is well mixed. Lidar extinction profiles
should ideally be determined using a technique that accounts for the
vertical and temporal variation in the S ratio, as in Strawbridge et al. (2018). When the S ratio variability was accounted for, the results of this study indicate that the MAX-DOAS retrievals of aerosol extinction performed well compared to the smoothed lidar results.
Comparisons of averaged and smoothed lidar profiles of aerosol extinction
indicated that the vertical sensitivity of the MAX-DOAS retrievals smoothed
the true atmospheric profiles towards the surface. This smoothing can
transform vertical profiles that are relatively uniform within the boundary
layer into apparently elevated profiles and vice versa. This shape change
depends on the location of extinction within the true vertical profile and
the averaging kernel matrix of the retrieval. Interpretation of the shape of
the MAX-DOAS vertical profiles must account for the instrument's sensitivity
to the true vertical profile (i.e., the averaging kernel matrix).
MAX-DOAS retrievals of NO2 and SO2 VCDs performed well in
comparison to the Pandora VCDs. The exception was when the aerosol
retrievals were inaccurate due to rapidly varying vertical profiles. This
was an expected result since the aerosol retrievals are used as forward
model parameters in the trace gas retrieval. The MAX-DOAS trace gas
retrievals within 0–100 m a.g.l. captured the temporal trends observed by
the active DOAS measurements, but the MAX-DOAS mixing ratios were
statistically greater than the active DOAS values, particularly when
SO2 and NO2 were <20 and <10 ppb, respectively. Differences between the instruments' values can be attributed
to variability in the trace gas profiles within 150 m a.g.l. The MAX-DOAS
observed elevated enhancements of pollution undetected by ground-based
techniques such as the active DOAS, which is perhaps its greatest asset. Pollution
enhancements at surface level did not always coincide with total boundary
layer enhancements and vice versa, due to elevated plumes and/or
significant wind shear. The MAX-DOAS vertical profiles of trace
gases were consistent with the profiles expected based on the wind direction
and shear conditions. Aircraft measurements of SO2 near Fort McKay
South on 3 September indicated that the magnitudes and elevated shape of the
retrievals were reasonable.
A major advantage of the MAX-DOAS technique is the ability to simultaneously
retrieve total columns and vertical profiles of trace gases and aerosol
extinction in the boundary layer and the lower troposphere from spectral
measurements without requiring knowledge of the aerosol size distributions
or compositions. These advantages are important in industrial regions where
the vertical profiles of pollutants vary temporally and spatially and
in situ monitoring can under-sample plumes. In the AOSR and similar
industrial regions, a full understanding of the air quality conditions
requires instruments, such as MAX-DOAS, capable of observing the total
boundary layer on a horizontal scale of a few kilometres, in addition to
traditional in situ instruments.
List of acronyms used in this paper
AcronymExpansiona.g.l.Above ground levelAERONETAerosol Robotic NetworkAODAerosol optical depthAOSRAthabasca oil sands regionAPSAerodynamic particle sizerBrOBromine oxideCCDCharge-coupled deviceClOChlorine oxideDOASDifferential optical absorption spectroscopydSCDDifferential slant column densityECCCEnvironment and Climate Change CanadaFRSFraunhofer reference spectrumHCHOFormaldehydeLidarLight detection and rangingMAX-DOASMulti-Axis Differential Optical Absorption SpectroscopyNH3AmmoniaNO2Nitrogen dioxideNO3Nitrate radicalNOxNitrogen oxides (NO2+NO)O4TetraoxygenOHHydroxyl radicalPM2.5Particulate matter with a diameter <2.5µmppbParts per billionpptParts per trillionRIRefractive indexSMPSScanning mobility particle sizerSO2Sulfur dioxideSP-AMSSoot particle aerosol mass spectrometerUVUltravioletVCDVertical column densityWBEAWood Buffalo Environmental Association
Data availability
All data used in this study have been published in a
publicly available data portal maintained by Environment Canada. The
MAX-DOAS, active DOAS, SPAMS, and SMPS data can be found at the following
location: ECCC Data (2016a). The sun photometer data and WindRASS data can be
found at the following location: ECCC Data (2017). The aircraft SO2
dataset can be found at the following location: ECCC Data (2016b). All
datasets are available in *.csv format.
The supplement related to this article is available online at: https://doi.org/10.5194/amt-13-1129-2020-supplement.
Author contributions
ZYWD developed and designed the MAX-DOAS study, performed the investigation, data analysis, and data visualization, and wrote the manuscript and made modifications of the same with contribution from all co-authors. UF supervised and validated the MAX-DOAS data analysis. KBS designed and conducted the ground-based lidar study, including the concept, design, investigation, and data analysis. MA developed the airborne lidar study concept and design and conducted the airborne lidar study investigation and analysis as well as the S ratio modelling. SB analyzed and visualized the SO2 flight data. EGS designed and conducted the SMPS investigation and data analysis. AL conducted the active DOAS investigation and performed the active DOAS data analysis. VEF developed and designed the Pandora study and conducted the investigation and data analysis. IA developed and designed the AERONET AOD study and conducted the investigation and data analysis. CAM conducted the PRATMO modelling for Pandora data analysis and provided the Pandora data. JW supervised the airborne lidar study. MDW and AKYL developed and designed the SP-AMS study concept and conducted the investigation. JB performed the project administration and supervised all studies in the field. JO supervised the SMPS study and the APS data analysis. JO'B developed and designed the SO2 flight data study and conducted the investigation. RS developed and designed the windRASS study concept and conducted the investigation and data analysis. HDO supervised the APS study. CM supervised the APS study. RM supervised the MAX-DOAS study and designed, developed, and supervised the active DOAS study.
Competing interests
The authors declare that they have no conflict of interest.
Acknowledgements
Funding for this study was provided by Environment and Climate
Change Canada and the Canada-Alberta Oil Sands Monitoring program. Zoë
Davis, Akshay Lobo and Sabour Baray acknowledge the financial support
provided by the Natural Sciences and Engineering Research Council of Canada
(NSERC) Collaborative Research and Training Experience Program (CREATE)
Integrating Atmospheric Chemistry and Physics from Earth to Space
(IACPES).
Financial support
This research has been supported by the Natural Sciences and Engineering Research Council of Canada (grant nos. RGPIN-2018-05898 and 398061-2011) and Environment Canada (grant no. 583007).
Review statement
This paper was edited by Robyn Schofield and reviewed by two anonymous referees.
ReferencesAdams, C., Normand, E. N., McLinden, C. A., Bourassa, A. E., Lloyd, N. D., Degenstein, D. A., Krotkov, N. A., Belmonte Rivas, M., Boersma, K. F., and Eskes, H.: Limb–nadir matching using non-coincident NO2 observations: proof of concept and the OMI-minus-OSIRIS prototype product, Atmos. Meas. Tech., 9, 4103–4122, 10.5194/amt-9-4103-2016, 2016.Aggarwal, M., Whiteway, J., Seabrook, J., Gray, L., Strawbridge, K., Liu, P., O'Brien, J., Li, S.-M., and McLaren, R.: Airborne lidar measurements of aerosol and ozone above the Canadian oil sands region, Atmos. Meas. Tech., 11, 3829–3849, 10.5194/amt-11-3829-2018, 2018.Amiri, N., Ghahremaninezhad, R., Rempillo, O., Tokarek, T. W., Odame-Ankrah, C. A., Osthoff, H. D., and Norman, A.-L.: Stable sulfur isotope measurements to trace the fate of SO2 in the Athabasca oil sands region, Atmos. Chem. Phys., 18, 7757–7780, 10.5194/acp-18-7757-2018, 2018.Baray, S., Darlington, A., Gordon, M., Hayden, K. L., Leithead, A., Li, S.-M., Liu, P. S. K., Mittermeier, R. L., Moussa, S. G., O'Brien, J., Staebler, R., Wolde, M., Worthy, D., and McLaren, R.: Quantification of methane sources in the Athabasca Oil Sands Region of Alberta by aircraft mass balance, Atmos. Chem. Phys., 18, 7361–7378, 10.5194/acp-18-7361-2018, 2018.Bogumil, K., Orphal, J., Homann, T., Voigt, S., Spietz, P., Fleischmann, O. C., Vogel, A., Hartmann, M., Bovensmann, H., Frerick, J., and Burrows, J. P.: Measurements of molecular absorption spectra with the SCIAMACHY pre-flight model: Instrument characterization and reference data for atmospheric remote sensing in the 230–2380 nm region, J. Photoch. Photobio. A, 157, 167–184, 10.1016/S1010-6030(03)00062-5, 2003.Bogumil, K., Orphal, J., Homann, T., Voigt, S., Spietz, P., Fleischmann, O.
C., Vogel, A., Hartmann, M., Kromminga, H., Bovensmann, H., Frerick, J., and
Burrows, J. P.: Measurements of molecular absorption spectra with the
SCIAMACHY pre-flight model: instrument characterization and reference data
for atmospheric remote-sensing in the 230–2380 nm region, J. Photoch.
Photobio. A, 157, 167–184, 10.1016/S1010-6030(03)00062-5, 2003.Bourgeois, Q., Ekman, A. M. L., Renard, J.-B., Krejci, R., Devasthale, A., Bender, F. A.-M., Riipinen, I., Berthet, G., and Tackett, J. L.: How much of the global aerosol optical depth is found in the boundary layer and free troposphere?, Atmos. Chem. Phys., 18, 7709–7720, 10.5194/acp-18-7709-2018, 2018.Browaeys, J.: Linear fit with both uncertainties in x and in y, MATLAB Central File Exchange, available at:
https://www.mathworks.com/matlabcentral/fileexchange/45711 (last access: 26 February 2019), 2017.Canadian Council of Ministers of the Environment: Sulphur Dioxide, Sulphur
Dioxide, available at:
https://www.ccme.ca/en/resources/air/air/sulphur-dioxide.html (last access: 28 May 2019), 2014.Clémer, K., Van Roozendael, M., Fayt, C., Hendrick, F., Hermans, C., Pinardi, G., Spurr, R., Wang, P., and De Mazière, M.: Multiple wavelength retrieval of tropospheric aerosol optical properties from MAXDOAS measurements in Beijing, Atmos. Meas. Tech., 3, 863–878, 10.5194/amt-3-863-2010, 2010.Davis, Z. Y. W., Baray, S., McLinden, C. A., Khanbabakhani, A., Fujs, W., Csukat, C., Debosz, J., and McLaren, R.: Estimation of NOx and SO2 emissions from Sarnia, Ontario, using a mobile MAX-DOAS (Multi-AXis Differential Optical Absorption Spectroscopy) and a NOx analyzer, Atmos. Chem. Phys., 19, 13871–13889, 10.5194/acp-19-13871-2019, 2019.Degenstein, D. A., Bourassa, A. E., Roth, C. Z., and Llewellyn, E. J.: Limb scatter ozone retrieval from 10 to 60 km using a multiplicative algebraic reconstruction technique, Atmos. Chem. Phys., 9, 6521–6529, 10.5194/acp-9-6521-2009, 2009.Deutschmann, T., Beirle, S., Frieß, U., Grzegorski, M., Kern, C., Kritten, L., Platt, U., Prados-Roman, C., Pukite, J., Wagner, T., Werner, B.,
and Pfeilsticker, K.: The Monte Carlo atmospheric radiative transfer model
McArtim: Introduction and validation of Jacobians and 3D features, J. Quant.
Spectrosc. Ra., 112, 1119–1137, 10.1016/j.jqsrt.2010.12.009, 2011.ECCC: Environment and Climate Change Canada – NPRI Data Search, NPRI Data
Search, available at:
https://pollution-waste.canada.ca/national-release-inventory/archives/index.cfm?lang=en
(last access: 24 April 2019), 2018a.ECCC: Environment and Climate Change Canada – NPRI Data Search – Facility
Search Results, Facil. Search Results – Nitrogen Oxides, available at:
https://pollution-waste.canada.ca/national-release-inventory/archives/index.cfm?do=results&lang=en&opt_facility_name=&opt_facility=&opt_npri_id=&opt_report_year=205 (last access: 24 April 2019), 2018b.ECCC: Environment and Climate Change Canada, NPRI Data Search, Facility
Search Results – Sulphur Dioxide, available at:
https://pollution-waste.canada.ca/national-release-inventory/archives/index.cfm?do=results&lang=en&opt_facility_name=&opt_facility=&opt_npri_id=&opt_report_year=2017&opt_chemical_type=ALL&opt_industry=&opt_cas_name=7446-09-5&opt_cas_num=&opt_location_type=ALL&opt_province=&opt_postal_code=&opt_community=&opt_csic=&opt_csi2=&15 (last access: 24 April 2019), 2018c.ECCC Data: Pollutant Transformation, Summer 2013 Ground-based Intensive
Multi Parameters – Fort McKay, Oil Sands Region, available at:
http://donnees.ec.gc.ca/data/air/monitor/ambient-air-quality-oil-sands-region/pollutant-transformation-summer-2013-ground-based-intensive-multi-parameters-fort-mckay-oil-sands-region/?lang=en
(last access: 4 February 2020), 2016a.ECCC Data: Pollutant Transformation, Summer 2013 Aircraft Intensive Multi Parameters, Oil Sands Region, available at: https://open.canada.ca/data/en/dataset/0bd78f8f-c4da-49c3-9bfd-751e6f300567 (last access: 3 February 2020), 2016b.ECCC Data: Pollutant Transformation, Ground-based Pollutant Monitoring Multi
Parameters – Validated Data – Fort McKay, Oil Sands Region, available at:
http://donnees.ec.gc.ca/data/air/monitor/ambient-air-quality-oil-sands-region/pollutant-transformation-ground-based-pollutant-monitoring-multi-parameters-validated-data-fort-mckay-oil-sands-region/?lang=en
(last access: 4 February 2020), 2017.Fenn, M. E., Bytnerowicz, A., Schilling, S. L., and Ross, C. S.: Atmospheric
deposition of nitrogen, sulfur and base cations in jack pine stands in the
Athabasca Oil Sands Region, Alberta, Canada, Environ. Pollut., 196,
497–510, 10.1016/j.envpol.2014.08.023, 2015.Finlayson-Pitts, B. J., Wingen, L. M., Sumner, A. L., Syomin, D., and
Ramazan, K. A.: The heterogeneous hydrolysis of NO2 in laboratory systems
and in outdoor and indoor atmospheres: An integrated mechanism, Phys. Chem.
Chem. Phys., 5, 223–242, 10.1039/b208564j, 2003.Fioletov, V. E., McLinden, C. A., Cede, A., Davies, J., Mihele, C., Netcheva, S., Li, S.-M., and O'Brien, J.: Sulfur dioxide (SO2) vertical column density measurements by Pandora spectrometer over the Canadian oil sands, Atmos. Meas. Tech., 9, 2961–2976, 10.5194/amt-9-2961-2016, 2016.Frieß, U., Monks, P. S., Remedios, J. J., Rozanov, A., Sinreich, R.,
Wagner, T., and Platt, U.: MAX-DOAS O4 measurements: A new technique to derive information on atmospheric aerosols: 2. Modeling studies, J. Geophys. Res., 111, D14203, 10.1029/2005JD006618, 2006.Frieß, U., Deutschmann, T., Gilfedder, B. S., Weller, R., and Platt, U.: Iodine monoxide in the Antarctic snowpack, Atmos. Chem. Phys., 10, 2439–2456, 10.5194/acp-10-2439-2010, 2010.Frieß, U., Sihler, H., Sander, R., Poehler, D., Yilmaz, S., and Platt,
U.: The vertical distribution of BrO and aerosols in the Arctic:
Measurements by active and passive differential optical absorption
spectroscopy, J. Geophys. Res., 116, D00R04,
10.1029/2011JD015938, 2011.Frieß, U., Klein Baltink, H., Beirle, S., Clémer, K., Hendrick, F., Henzing, B., Irie, H., de Leeuw, G., Li, A., Moerman, M. M., van Roozendael, M., Shaiganfar, R., Wagner, T., Wang, Y., Xie, P., Yilmaz, S., and Zieger, P.: Intercomparison of aerosol extinction profiles retrieved from MAX-DOAS measurements, Atmos. Meas. Tech., 9, 3205–3222, 10.5194/amt-9-3205-2016, 2016.Frieß, U., Beirle, S., Alvarado Bonilla, L., Bösch, T., Friedrich, M. M., Hendrick, F., Piters, A., Richter, A., van Roozendael, M., Rozanov, V. V., Spinei, E., Tirpitz, J.-L., Vlemmix, T., Wagner, T., and Wang, Y.: Intercomparison of MAX-DOAS vertical profile retrieval algorithms: studies using synthetic data, Atmos. Meas. Tech., 12, 2155–2181, 10.5194/amt-12-2155-2019, 2019.Gordon, M., Li, S.-M., Staebler, R., Darlington, A., Hayden, K., O'Brien, J., and Wolde, M.: Determining air pollutant emission rates based on mass balance using airborne measurement data over the Alberta oil sands operations, Atmos. Meas. Tech., 8, 3745–3765, 10.5194/amt-8-3745-2015, 2015.Gordon, M., Makar, P. A., Staebler, R. M., Zhang, J., Akingunola, A., Gong, W., and Li, S.-M.: A comparison of plume rise algorithms to stack plume measurements in the Athabasca oil sands, Atmos. Chem. Phys., 18, 14695–14714, 10.5194/acp-18-14695-2018, 2018.Government of Canada: Canadian Environmental Sustainability Indicators –
Canada.ca, available at:
https://indicators-map.canada.ca/App/CESI_ICDE?keys=AirAmbient_AvgPM&GoCTemplateCulture=en-CA
(last access: 6 February 2019), 2018.Health Canada: Human Health Risk Assessment for Sulphur Dioxide, available at:
http://publications.gc.ca/collections/collection_2016/sc-hc/H144-29-2016-eng.pdf (last access: 1 January 2019), 2016.Hermans, C.: BIRA-IASB Spectroscopy Lab, available at: http://spectrolab.aeronomie.be/index.htm (last access: 27 February 2020), 2011.Heckel, A., Richter, A., Tarsu, T., Wittrock, F., Hak, C., Pundt, I., Junkermann, W., and Burrows, J. P.: MAX-DOAS measurements of formaldehyde in the Po-Valley, Atmos. Chem. Phys., 5, 909–918, 10.5194/acp-5-909-2005,
2005.
Holloway, A. M. and Wayne, R.: Atmospheric Chemistry, RSC Publishing,
Cambridge, UK, 2010.Hönninger, G. and Platt, U.: Observations of BrO and its vertical
distribution during surface ozone depletion at Alert, Atmos. Environ.,
36, 2481–2489, 10.1016/S1352-2310(02)00104-8, 2002.Hönninger, G., von Friedeburg, C., and Platt, U.: Multi axis differential optical absorption spectroscopy (MAX-DOAS), Atmos. Chem. Phys., 4, 231–254, 10.5194/acp-4-231-2004, 2004.Hsu, Y.-M.: Trends in Passively-Measured Ozone, Nitrogen Dioxide and Sulfur
Dioxide Concentrations in the Athabasca Oil Sands Region of Alberta, Canada,
Aerosol Air Qual. Res., 13, 1448–1463, 10.4209/aaqr.2012.08.0224,
2013.Irie, H., Kanaya, Y., Akimoto, H., Iwabuchi, H., Shimizu, A., and Aoki, K.: First retrieval of tropospheric aerosol profiles using MAX-DOAS and comparison with lidar and sky radiometer measurements, Atmos. Chem. Phys., 8, 341–350, 10.5194/acp-8-341-2008, 2008.Irie, H., Takashima, H., Kanaya, Y., Boersma, K. F., Gast, L., Wittrock, F., Brunner, D., Zhou, Y., and Van Roozendael, M.: Eight-component retrievals from ground-based MAX-DOAS observations, Atmos. Meas. Tech., 4, 1027–1044, 10.5194/amt-4-1027-2011, 2011.Irie, H., Nakayama, T., Shimizu, A., Yamazaki, A., Nagai, T., Uchiyama, A., Zaizen, Y., Kagamitani, S., and Matsumi, Y.: Evaluation of MAX-DOAS aerosol retrievals by coincident observations using CRDS, lidar, and sky radiometer in Tsukuba, Japan, Atmos. Meas. Tech., 8, 2775–2788, 10.5194/amt-8-2775-2015, 2015.Jin, J., Ma, J., Lin, W., Zhao, H., Shaiganfar, R., Beirle, S., and Wagner,
T.: MAX-DOAS measurements and satellite validation of tropospheric NO2 and SO2 vertical column densities at a rural site of North China, Atmos. Environ., 133, 12–25, 10.1016/j.atmosenv.2016.03.031, 2016.Keller-Rudek, H., Moortgat, G. K., Sander, R., and Sörensen, R.: The MPI-Mainz UV/VIS Spectral Atlas of Gaseous Molecules of Atmospheric Interest, Earth Syst. Sci. Data, 5, 365–373, 10.5194/essd-5-365-2013, 2013.Kulmala, M., Vehkamäki, H., Petäjä, T., Dal Maso, M., Lauri, A.,
Kerminen, V.-M., Birmili, W., and McMurry, P. H.: Formation and growth rates
of ultrafine atmospheric particles: a review of observations, J. Aerosol
Sci., 35, 143–176, 10.1016/j.jaerosci.2003.10.003, 2004.Lee, A. K. Y., Adam, M. G., Liggio, J., Li, S.-M., Li, K., Willis, M. D., Abbatt, J. P. D., Tokarek, T. W., Odame-Ankrah, C. A., Osthoff, H. D., Strawbridge, K., and Brook, J. R.: A large contribution of anthropogenic organo-nitrates to secondary organic aerosol in the Alberta oil sands, Atmos. Chem. Phys., 19, 12209–12219, 10.5194/acp-19-12209-2019, 2019.Levenberg, K.: A method for the solution of certain non-linear problems in
least squares, Q. Appl. Math., 2, 164–168, 10.1090/qam/10666, 1944.Li, X., Brauers, T., Shao, M., Garland, R. M., Wagner, T., Deutschmann, T., and Wahner, A.: MAX-DOAS measurements in southern China: retrieval of aerosol extinctions and validation using ground-based in-situ data, Atmos. Chem. Phys., 10, 2079–2089, 10.5194/acp-10-2079-2010, 2010.Liggio, J., Li, S.-M., Hayden, K., Taha, Y. M., Stroud, C., Darlington, A.,
Drollette, B. D., Gordon, M., Lee, P., Liu, P., Leithead, A., Moussa, S. G.,
Wang, D., O'Brien, J., Mittermeier, R. L., Brook, J. R., Lu, G., Staebler,
R. M., Han, Y., Tokarek, T. W., Osthoff, H. D., Makar, P. A., Zhang, J.,
Plata, D. L. and Gentner, D. R.: Oil sands operations as a large source of
secondary organic aerosols, Nature, 534, 91–94, 10.1038/nature17646, 2016.Liggio, J., Li, S.-M., Staebler, R. M., Hayden, K., Darlington, A.,
Mittermeier, R. L., O'Brien, J., McLaren, R., Wolde, M., Worthy, D. and
Vogel, F.: Measured Canadian oil sands CO2 emissions are higher than
estimates made using internationally recommended methods, Nat. Commun.,
10, 1863, 10.1038/s41467-019-09714-9, 2019.
Marquardt, D. W.: An algorithm for lesat squares estimation of non-linear
parameters, J. Soc. Ind. Appl. Math., 11, 431–441, 1963.McLaren, R., Wojtal, P., Majonis, D., McCourt, J., Halla, J. D., and Brook, J.: NO3 radical measurements in a polluted marine environment: links to ozone formation, Atmos. Chem. Phys., 10, 4187–4206, 10.5194/acp-10-4187-2010, 2010.McLaren, R., Wojtal, P., Halla, J. D., Mihele, C. and Brook, J. R.: A survey
of NO2:SO2 emission ratios measured in marine vessel plumes in the Strait of Georgia, Atmos. Environ., 46, 655–658, 10.1016/j.atmosenv.2011.10.044, 2012.McLinden, C. A.: Stratospheric ozone in 3-D models: A simple chemistry and
the cross-tropopause flux, J. Geophys. Res., 105, 14653–14665,
10.1029/2000JD900124, 2000.McLinden, C. A., Fioletov, V., Boersma, K. F., Krotkov, N., Sioris, C. E.,
Veefkind, J. P., and Yang, K.: Air quality over the Canadian oil sands: A
first assessment using satellite observations, Geophys. Res. Lett., 39,
L04804, 10.1029/2011GL050273, 2012.McLinden, C. A., Fioletov, V., Boersma, K. F., Kharol, S. K., Krotkov, N., Lamsal, L., Makar, P. A., Martin, R. V., Veefkind, J. P., and Yang, K.: Improved satellite retrievals of NO2 and SO2 over the Canadian oil sands and comparisons with surface measurements, Atmos. Chem. Phys., 14, 3637–3656, 10.5194/acp-14-3637-2014, 2014.McLinden, C. A., Fioletov, V., Krotkov, N. A., Li, C., Boersma, K. F. and
Adams, C.: A Decade of Change in NO2 and SO2 over the Canadian Oil Sands As Seen from Space, Environ. Sci. Technol., 50, 331–337,
10.1021/acs.est.5b04985, 2016.Platt, U., Perner, D., and Patz, H.: Simultaneous Measurement of Atmospheric
CH2O, O3, and NO2 by Differential Optical-Absorption, J. Geophys. Res., 84, 6329–6335, 10.1029/JC084iC10p06329, 1979.
latt, U. and Stutz, J.: Differential optical absorption spectroscopy: Principles and applications, Springer Verlag, Berlin, Germany, 2008.Psenner, R.: Environmental impacts on freshwaters: acidifaction as a global
problem, Sci. Total Environ., 143, 53–61, 10.1016/0048-9697(94)90532-0, 1994.Pui, D. Y. H., Chen, S.-C., and Zuo, Z.: PM2.5 in China: Measurements,
sources, visibility and health effects, and mitigation, Particuology, 13,
1–26, 10.1016/j.partic.2013.11.001, 2014.
Rodgers, C. D.: Inverse methods for atmospheric sounding: Theory and
practice, World Scientific, Singapore, 2000.Rodgers, C. D. and Connor, B. J.: Intercomparison of remote sounding
instruments, J. Geophys. Res., 108, 4116, 10.1029/2002JD002299, 2003.Rozanov, A., Bovensmann, H., Bracher, A., Hrechanyy, S., Rozanov, V.,
Sinnhuber, M., Stroh, F., and Burrows, J. P.: NO2 and BrO vertical profile retrieval from SCIAMACHY limb measurements: Sensitivity studies, in: Atmospheric Remote Sensing: Earth's Surface, Troposphere, Stratosphere and Mesosphere – I, edited by: Burrows, J. P. and Eichmann, K. U., Elsevier Science Ltd, Oxford, 36, 846–854, 2005.
Seinfeld, J. H. and Pandis, S. N.: Atmospheric Chemistry and Physics: From
Air Pollution to Climate Change, John Wiley & Sons, New York, 2006.Simpson, I. J., Blake, N. J., Barletta, B., Diskin, G. S., Fuelberg, H. E., Gorham, K., Huey, L. G., Meinardi, S., Rowland, F. S., Vay, S. A., Weinheimer, A. J., Yang, M., and Blake, D. R.: Characterization of trace gases measured over Alberta oil sands mining operations: 76 speciated C2–C10 volatile organic compounds (VOCs), CO2, CH4, CO, NO, NO2, NOy, O3 and SO2, Atmos. Chem. Phys., 10, 11931–11954, 10.5194/acp-10-11931-2010, 2010.Sioris, C. E., Abboud, I., Fioletov, V. E., and McLinden, C. A.: AEROCAN, the
Canadian sub-network of AERONET: Aerosol monitoring and air quality
applications, Atmos. Environ., 167, 444–457, 10.1016/j.atmosenv.2017.08.044, 2017.Strawbridge, K. B.: Developing a portable, autonomous aerosol backscatter lidar for network or remote operations, Atmos. Meas. Tech., 6, 801–816, 10.5194/amt-6-801-2013, 2013.Strawbridge, K. B., Travis, M. S., Firanski, B. J., Brook, J. R., Staebler, R., and Leblanc, T.: A fully autonomous ozone, aerosol and nighttime water vapor lidar: a synergistic approach to profiling the atmosphere in the Canadian oil sands region, Atmos. Meas. Tech., 11, 6735–6759, 10.5194/amt-11-6735-2018, 2018.Stutz, J. and Platt, U.: Problems in Using Diode-Arrays for Open Path Doas
Measurements of Atmospheric Species, in: Proc. SPIE 1715, Optical Methods in Atmospheric Chemistry, edited by: Schiff H. I. and Platt, U., International Society for Optics and Photonics, SPIE, 329–340, 10.1117/12.140192, 1993.Tan, W., Liu, C., Wang, S., Xing, C., Su, W., Zhang, C., Xia, C., Liu, H., Cai, Z., and Liu, J.: Tropospheric NO2, SO2, and HCHO over the East China Sea, using ship-based MAX-DOAS observations and comparison with OMI and OMPS satellite data, Atmos. Chem. Phys., 18, 15387–15402, 10.5194/acp-18-15387-2018, 2018.Tokarek, T. W., Odame-Ankrah, C. A., Huo, J. A., McLaren, R., Lee, A. K. Y., Adam, M. G., Willis, M. D., Abbatt, J. P. D., Mihele, C., Darlington, A., Mittermeier, R. L., Strawbridge, K., Hayden, K. L., Olfert, J. S., Schnitzler, E. G., Brownsey, D. K., Assad, F. V., Wentworth, G. R., Tevlin, A. G., Worthy, D. E. J., Li, S.-M., Liggio, J., Brook, J. R., and Osthoff, H. D.: Principal component analysis of summertime ground site measurements in the Athabasca oil sands with a focus on analytically unresolved intermediate-volatility organic compounds, Atmos. Chem. Phys., 18, 17819–17841, 10.5194/acp-18-17819-2018, 2018.Wagner, T., von Friedeburg, C., Wenig, M. O., Otten, C., and Platt, U.:
UV-visible observations of atmospheric O-4 absorptions using direct moonlight and zenith-scattered sunlight for clear-sky and cloudy sky conditions, J. Geophys. Res., 107, 4424, 10.1029/2001JD001026, 2002.Wagner, T., Dix, B., von Friedeburg, C., Frieß, U., Sanghavi, S.,
Sinreich, R., and Platt, U.: MAX-DOAS O4 measurements: A new technique to derive information on atmospheric aerosols – Principles and information content, J. Geophys. Res., 109, D22205,
10.1029/2004JD004904, 2004.Wagner, T., Deutschmann, T., and Platt, U.: Determination of aerosol properties from MAX-DOAS observations of the Ring effect, Atmos. Meas. Tech., 2, 495–512, 10.5194/amt-2-495-2009, 2009.Wagner, T., Beirle, S., Brauers, T., Deutschmann, T., Frieß, U., Hak, C., Halla, J. D., Heue, K. P., Junkermann, W., Li, X., Platt, U., and Pundt-Gruber, I.: Inversion of tropospheric profiles of aerosol extinction and HCHO and NO2 mixing ratios from MAX-DOAS observations in Milano during the summer of 2003 and comparison with independent data sets, Atmos. Meas. Tech., 4, 2685–2715, 10.5194/amt-4-2685-2011, 2011.Wagner, T., Beirle, S., Benavent, N., Bösch, T., Chan, K. L., Donner, S., Dörner, S., Fayt, C., Frieß, U., García-Nieto, D., Gielen, C., González-Bartolome, D., Gomez, L., Hendrick, F., Henzing, B., Jin, J. L., Lampel, J., Ma, J., Mies, K., Navarro, M., Peters, E., Pinardi, G., Puentedura, O., Puḳīte, J., Remmers, J., Richter, A., Saiz-Lopez, A., Shaiganfar, R., Sihler, H., Van Roozendael, M., Wang, Y., and Yela, M.: Is a scaling factor required to obtain closure between measured and modelled atmospheric O4 absorptions? An assessment of uncertainties of measurements and radiative transfer simulations for 2 selected days during the MAD-CAT campaign, Atmos. Meas. Tech., 12, 2745–2817, 10.5194/amt-12-2745-2019, 2019.Wang, T., Hendrick, F., Wang, P., Tang, G., Clémer, K., Yu, H., Fayt, C., Hermans, C., Gielen, C., Müller, J.-F., Pinardi, G., Theys, N., Brenot, H., and Van Roozendael, M.: Evaluation of tropospheric SO2 retrieved from MAX-DOAS measurements in Xianghe, China, Atmos. Chem. Phys., 14, 11149–11164, 10.5194/acp-14-11149-2014, 2014.Wang, Y., Beirle, S., Lampel, J., Koukouli, M., De Smedt, I., Theys, N., Li, A., Wu, D., Xie, P., Liu, C., Van Roozendael, M., Stavrakou, T., Müller, J.-F., and Wagner, T.: Validation of OMI, GOME-2A and GOME-2B tropospheric NO2, SO2 and HCHO products using MAX-DOAS observations from 2011 to 2014 in Wuxi, China: investigation of the effects of priori profiles and aerosols on the satellite products, Atmos. Chem. Phys., 17, 5007–5033, 10.5194/acp-17-5007-2017, 2017.Weitkamp, C.: Lidar: range-resolved optical remote sensing of the
atmosphere, Springer Science+Business Media, New York, 2005.WHO: WHO Air quality guidelines for particulate matter, ozone, nitrogen
dioxide and sulfur dioxide, Global update 2005, Summary of Risk Assessment, available at:
https://apps.who.int/iris/bitstream/handle/10665/69477/WHO_SDE_PHE_OEH_06.02_eng.pdf?sequence=1 (last access: 18 April 2019), 2006.Wojtal, P., Halla, J. D., and McLaren, R.: Pseudo steady states of HONO measured in the nocturnal marine boundary layer: a conceptual model for HONO formation on aqueous surfaces, Atmos. Chem. Phys., 11, 3243–3261, 10.5194/acp-11-3243-2011, 2011.Wood Buffalo Environmental Association: Historical Environmental Monitoring
Data, available at:
https://wbea.org/historical-monitoring-data/, last access: 29 April 2019.Wu, F., Xie, P., Li, A., Mou, F., Chen, H., Zhu, Y., Zhu, T., Liu, J., and Liu, W.: Investigations of temporal and spatial distribution of precursors SO2 and NO2 vertical columns in the North China Plain using mobile DOAS, Atmos. Chem. Phys., 18, 1535–1554, 10.5194/acp-18-1535-2018, 2018.Wu, F. C., Xie, P. H., Li, A., Chan, K. L., Hartl, A., Wang, Y., Si, F. Q., Zeng, Y., Qin, M., Xu, J., Liu, J. G., Liu, W. Q., and Wenig, M.: Observations of SO2 and NO2 by mobile DOAS in the Guangzhou eastern area during the Asian Games 2010, Atmos. Meas. Tech., 6, 2277–2292, 10.5194/amt-6-2277-2013, 2013.
Zhang, J., Moran, M. D., Zheng, Q., Makar, P. A., Baratzadeh, P., Marson, G., Liu, P., and Li, S.-M.: Emissions preparation and analysis for multiscale air quality modeling over the Athabasca Oil Sands Region of Alberta, Canada, Atmos. Chem. Phys., 18, 10459–10481, 10.5194/acp-18-10459-2018,
2018.Zhao, Y., Duan, L., Xing, J., Larssen, T., Nielsen, C. P., and Hao, J.: Soil
Acidification in China: Is Controlling SO2 Emissions Enough?, Environ. Sci. Technol., 43, 8021–8026, 10.1021/es901430n, 2009.Zhong, S. and Zaveri, R.: Atmospheric Aerosols, in: International
Encyclopedia of Geography: People, the Earth, Environment and Technology, edited by: Richardson, D., Castree, N., Goodchild, M. F., Kobayashi, A., Liu, W., and Marston, R. A., John Wiley and Sons, Hoboken, New Jersey, 1–5, 10.1002/9781118786352.wbieg0304, 2017.Zieger, P., Weingartner, E., Henzing, J., Moerman, M., de Leeuw, G., Mikkilä, J., Ehn, M., Petäjä, T., Clémer, K., van Roozendael, M., Yilmaz, S., Frieß, U., Irie, H., Wagner, T., Shaiganfar, R., Beirle, S., Apituley, A., Wilson, K., and Baltensperger, U.: Comparison of ambient aerosol extinction coefficients obtained from in-situ, MAX-DOAS and LIDAR measurements at Cabauw, Atmos. Chem. Phys., 11, 2603–2624, 10.5194/acp-11-2603-2011, 2011.