Intensive aerosol properties of boreal and regional biomass burning aerosol at Mt. Bachelor Observatory: larger and black carbon (BC)-dominant particles transported from Siberian wildﬁres

. We characterize the aerosol physical and optical properties of 13 transported biomass burning (BB) events. BB events included long-range inﬂuence from ﬁres in Alaskan and Siberian boreal forests transported to Mt. Bachelor Observatory (MBO) in the free troposphere (FT) over 8–14 + d and regional wildﬁres in northern California and southwestern Oregon transported to MBO in the boundary layer (BL) over 10 h to 3 d. Intensive aerosol optical properties and normalized enhancement ratios for BB events were derived from measured aerosol light scattering coefﬁcients ( σ scat ), aerosol light-absorbing coefﬁcients ( σ abs ), ﬁne particulate matter (PM 1 ), and carbon monoxide (CO) measurements made from July to September 2019, with particle size distribution collected from August to September. The observations showed that the Siberian BB events had a lower scattering Ångström exponent (SAE), a higher mass scattering efﬁciency (MSE; (cid:49)σ scat /(cid:49) PM 1 ), and a bimodal aerosol size distribution with a higher geometric mean diameter ( D g ). We hypothesize that the larger particles and associated scattering properties were due to the transport of ﬁne dust alongside smoke in addition to contributions from condensation of secondary aerosol, coagulation of smaller particles, and aqueous-phase processing during transport. Alaskan and Siberian boreal forest BB plumes were transported long distances in the FT and characterized by lower absorption Ångström exponent (AAE) values indicative of black carbon (BC) dominance in the radiative budget. Signiﬁcantly elevated AAE values were only observed for BB events with < 1 d transport, which suggests strong production of brown carbon (BrC) in these plumes but limited radiative forcing impacts outside of the immediate region.


Introduction
Biomass burning (BB) is a major source of atmospheric aerosols (Bond et al., 2013;Andreae and Merlet, 2001) and significantly impacts public health and regional air quality up to thousands of kilometers from the source (Jaffe et al., 2020;Boucher et al., 2013).BB aerosols also impact global climate by scattering or absorbing solar radiation, acting as cloud condensation nuclei, and altering cloud albedo (Boucher et al., 2013;Spracklen et al., 2011;Pierce et al., 2007).How-ever, there are large uncertainties in BB aerosol formation, evolution, and radiative properties that limit our understanding of their climate impacts (Bond et al., 2013;Boucher et al., 2013Boucher et al., , 2020;;Bellouin et al., 2020).BB emissions and their associated impacts are likely to increase globally due to hot and dry conditions resulting from climate change, particularly in the western USA (Westerling, 2016;Liu et al., 2014) and sub-Arctic boreal forests of North America and Russia (Flannigan et al., 2009;Stocks et al., 1998).
BB particles are predominantly organic carbon (OC) and black carbon (BC), with some inorganic material (Vakkari et al., 2014;Reid et al., 2005;Zhou et al., 2017).BC is the most significant contributor to the absorption properties of BB particles (Healy et al., 2015;Bond et al., 2013), but OC also contributes to BB particle absorption in the form of brown carbon (BrC) (Andreae and Gelencsér, 2006).BC absorbs nearly uniformly in the range of 0.4 to 1 µm, resulting in an absorption Ångström exponent (AAE) of ∼ 1. High levels of BrC in BB increase AAE values due to its preferential absorption at lower wavelengths of sunlight (Andreae and Gelencsér, 2006).Atmospheric BrC is produced by incomplete combustion (Kirchstetter and Thatcher, 2012;Desyaterik et al., 2013;Lack and Langridge, 2013;Mohr et al., 2013) and secondary formation (Li et al., 2020;Nguyen et al., 2013;Updyke et al., 2012;Laskin et al., 2015).Saleh (2020) describes four classes of brown carbon, with widely varying optical properties, and points out that the BrC properties may depend on the measurement method.Photobleaching reduces the absorption of BrC and therefore significantly impacts the global radiation budget, but there are large uncertainties in this process (Liu et al., 2020a).In contrast, atmospheric BC is chemically inert and primarily produced from flaming combustion (Healy et al., 2015;Bond et al., 2013).In situ aerosol optical measurements are essential to improved modeling of the contributions of BrC and BC-to-BB aerosol absorption properties (Brown et al., 2021).
As plumes age, BB particles grow from their initial diameter (30-100 nm) (Hosseini et al., 2010;Levin et al., 2010) and undergo chemical and physical changes (Carrico et al., 2016;Reid et al., 2005;Vakkari et al., 2014).Changes in particle size are primarily the result of coagulation and the condensation of secondary organic aerosol (SOA) onto existing particles (Reid et al., 2005).However, the condensation of SOA is counterbalanced by loss due to evaporation and oxidation of primary organic aerosol during plume dilution (Collier et al., 2016;Zhou et al., 2017;Garofalo et al., 2019;May et al., 2013).The net condensation/evaporation effect in BB plumes can lead to an increase in PM mass relative to CO due to SOA production (Hobbs et al., 2003;Yokelson et al., 2009;Vakkari et al., 2014), while others have observed limited or no net mass increase (Akagi et al., 2012;Jolleys et al., 2015;Garofalo et al., 2019).Even when normalized PM mass concentrations do not change with age, particle diameter can shift through coagulation and particle-vapor mass transfer (Kleinman et al., 2020).Mie theory predicts that mass scattering efficiency (MSE; σ scat /PM) will increase as the average particle diameter grows toward the measurement wavelength (e.g., 300-700 nm) (Seinfeld and Pandis, 2006).Therefore, understanding the balance of aerosol condensation, evaporation, and removal processes is critical to understanding the scattering properties of aged BB particles.
Airborne mineral dust particles, which can be uplifted into the free troposphere alongside smoke by intense fire-related winds (Wagner et al., 2018), may also impact the radiative forcing of BB emissions.Dust particles are typically larger in size than smoke particles and can produce a net positive or negative radiative forcing depending on surface properties, particle size distribution, and composition (Balkanski et al., 2007;Durant et al., 2009;Russell et al., 2002).However, the climate properties of dust mixed with smoke are so far poorly understood.Dust emissions from fires are most frequently observed in arid regions (Chalbot et al., 2013;Nisantzi et al., 2014;Li et al., 2021) and can be transported over thousands of kilometers (Ansmann et al., 2009;Clements et al., 2008;Baars et al., 2011).Recent observations of lofted dust from fires in the coniferous forests of the western USA (Maudlin et al., 2015;Schlosser et al., 2017;Creamean et al., 2016) and Russia (Popovicheva et al., 2014) suggest that fires in non-arid regions may also emit dust.Despite in situ evidence, fires are not considered to be a source of airborne mineral dust in climate or aerosol models.
In this study, we describe an overview of the intensive optical properties and normalized enhancement ratios of submicron aerosols (PM 1 ) observed during the summer of 2019 at Mt. Bachelor Observatory (MBO), a remote, high-altitude site in the Pacific Northwest USA.Thirteen BB events were observed, including smoke from nine regional fires (northern California and Oregon, transported < 1-3 d) and four boreal forest fires (Alaska and Siberia, transported 8-10+ d).Aerosol optical and physical properties of these events were explored with respect to source location, emission characteristics, and transport time.The goals of this work are to elucidate how source and transport characteristics of BB events influence their climate impact through aerosol size distributions and the associated scattering properties as well as contributions of BrC and BC to absorption properties.

Sampling site
Measurements were conducted at MBO, located near the summit of Mt.Bachelor (43.981 • N, 121.691 • W; 2764 m a.s.l.), from 1 July to 10 September 2019.These observations were part of the larger Fire Influence on Regional to Global Environments and Air Quality (FIREX-AQ) experiment (Liao et al., 2021;Xu et al., 2021;Decker et al., 2021;Wiggins et al., 2021;Makar et al., 2021), which included extensive observations at MBO (Farley et al., 2022).Due to its remote location and limited anthropogenic influence, MBO is an ideal site for measurements of wildfire plumes ranging from locally emitted to long-range transport events (Laing et al., 2016;Baylon et al., 2017;Wigder et al., 2013).The atmospheric conditions during this study were typical for a clean background location as the PM 1 concentrations were relatively low (avg.±1σ ) (2.8±3.8 µg m −3 ), consistent with periods without wildfire influence observed at MBO in 2013 (.8 ± 2.8 µg sm −3 ) (Zhou et al., 2019).
Boundary-layer (BL) dynamics play an important role in the diurnal variation of aerosol composition at MBO.During daytime, upslope flow can mix in BL air to the sampling site, while at night the site is influenced by free troposphere (FT) air masses.These regimes can be differentiated based on ambient water vapor (WV) concentrations.WV is calculated as the difference between the WV (g kg −1 ) at the time CO peaked during the event and the WV previously found to demarcate the BL and FT (July-August: 5.23 g kg −1 , September: 4.60 g kg −1 ) (Wigder et al., 2013;Baylon et al., 2015;Zhang and Jaffe, 2017).Typically, local smoke plumes will be seen in upslope BL air with a positive WV, whereas distant smoke (Alaska, Siberia) will be transported in the FT with a negative WV.

CO, aerosol, and meteorology
Details of the CO, aerosol, and meteorology measurements at MBO employed in the current study have been previously described in detail (Laing et al., 2016(Laing et al., , 2020;;Baylon et al., 2017) and are thus only briefly described here.Basic meteorology measurements include temperature, humidity, and wind speed (Ambrose et al., 2011).CO measurements were made using a Picarro G2302 cavity ring-down spectrometer.Calibrations were performed every 8 h using three different National Oceanographic and Atmospheric Administration (NOAA) calibration gas standards, which are referenced to the World Meteorological Organization (WMO) mole fraction calibration scale (Gratz et al., 2015).Total CO uncertainty based on the precision of calibrations over the campaign was 3 %.
Dry (relative humidity -RH -less than 35 %) aerosol scattering and absorption coefficients, aerosol number size distribution (30-600 nm), and particle mass were measured during the 2019 summer campaign in 5 min averages.An inline 1 µ m impactor was located upline of the aerosol instruments.All particle measurements were corrected to standard temperature and pressure (STP; T = 273.15K, P = 101.325kPa) and are reported this way throughout the paper.Aerosol light-scattering coefficients (σ scat ) were measured by an integrating nephelometer (model 3563, TSI Inc., Shoreview, MN) at 450 nm (blue), 550 nm (green), and 700 nm (red).Data reduction and uncertainty analysis for the scattering data are outlined by Anderson and Ogren (1998).
We measured aerosol light-absorbing coefficients (σ abs ) with a 3λ tricolor absorption photometer (TAP, Brechtel Inc., Hayward, CA) at wavelengths 467 nm (blue), 528 nm (green), and 660 nm (red) (Laing et al., 2020).Unless otherwise stated, σ scat and σ abs values represent measurements taken at 550 and 528 nm, respectively.The absorption coefficients were corrected using the filter-loading and aerosolscattering correction factors derived by Virkkula (2010).Uncertainty calculations were based on those used in a previous study at MBO for measurements with a 3λ Particle Soot Absorption Photometer (PSAP) (Fischer et al., 2010).Com-bining sources of uncertainties (Anderson et al., 1999;Bond et al., 1999;Virkkula et al., 2005) yielded total uncertainties for σ abs of 30 %-40 % during BB events.Combined scattering uncertainties yielded lower total uncertainties for σ scat of 15 %-20 % during BB events.However, the relative uncertainties between events for absorption and scattering are much lower (< 10 %).This was estimated by calculating the relative standard deviation of the absorption and scattering measurements over a 60 min time window with no wildfire influence at MBO.
The power-law relationship between scattering and wavelength was used to adjust the 550 nm σ scat measurement to 528 nm using Eq.(1) (Laing et al., 2016): where λ is the wavelength and SAE is the scattering Ångström exponent calculated with the 450-550 nm pair.Absorption Ångström exponent (AAE) values were calculated for the σ abs pair of 467 and 660 nm using Eq. ( 2 Uncertainties for the intensive aerosol optical properties SAE and AAE values (Table S1 in the Supplement) were calculated by propagating the uncertainties from the measurements used in the respective calculations using addition in quadrature (Fischer et al., 2010).
We measured the 5 min-averaged dry aerosol number size distribution with a TSI 3938 Scanning Mobility Particle Sizer (SMPS).The SMPS system consisted of a TSI 3082 electrostatic classifier with a TSI 3081 differential mobility analyzer (DMA) and a TSI 3787 water-based condensation particle counter.Geometric mean diameter (D g ) was calculated from lognormal fits of event-averaged size distributions performed with the standard fitting algorithm of Igor Pro analysis software (fit parameters: Dry particle mass under 1 µm (PM 1 ) was measured with an optical particle counter (OPC, model 1.109, Grimm Technologies, Douglasville, GA).The OPC measures the particle size distribution for particles with aerodynamic diameters from 0.25 to 32 µm in 31 size bins.We note that this size ranges misses a significant fraction of particle mass for smaller particles.For the biomass burning plumes observed in this study, approximately 42 % of the PM 1 mass is accounted for by particles with diameters < 250 µm as determined by the SMPS data.The OPC PM 2.5 mass concentrations are determined via the manufacturer calibration against standard filter methods (Grimm and Eatough, 2009).The OPC is a USA EPA Federal Equivalent Method for measuring PM 2.5 mass concentrations.August and September OPC https://doi.org/10.5194/acp-23-2747-2023Atmos.Chem.Phys., 23, 2747-2764, 2023 data were determined to be artificially biased high, which coincided with the installation of a thermodenuder in the aerosol sampling line on 31 July 2019.During this time, the OPC mass concentrations were adjusted to match the SMPS mass concentrations.

Enhancement ratio calculation
As in prior studies of BB events at MBO (Laing et al., 2016;Briggs et al., 2016), normalized enhancement ratios ( Y / X, NERs) of σ scat / CO, σ abs / CO, and PM 1 / CO were calculated from the slope of the reduced major axis (RMA) regression of Y plotted against X.Intensive aerosol optical properties mass scattering and mass absorption efficiencies (MSEs and MAEs) were calculated as the NERs of σ scat / PM 1 and σ abs / PM 1 , respectively, at 550 nm for σ scat and 528 nm for σ abs .Single scattering albedo (ω) was calculated as the RMA regression of scattering and total extinction (scattering + absorption) using values at 528 nm.In all cases the enhancements are large compared to the background, thus avoiding the problems described by Briggs et al. (2016) for small enhancements above the background.As in prior studies (Laing et al., 2016), uncertainties for the NER calculations were determined from the uncertainties in the extensive properties used in calculating the NERs and the uncertainty of the RMA regression using addition in quadrature.
Precision uncertainty and total uncertainty were calculated as described by Anderson et al. (1999) for all values derived from optical measurements (Table S1).Precision uncertainty is the uncertainty associated with noise and instrument drift.Total uncertainty includes precision uncertainty, the uncertainty associated with the corrections we applied to the data, and the uncertainty associated with the calibration method.Precision uncertainty is best used to compare the individual BB events seen at MBO in this study, whereas total uncertainty is more appropriate to consider when comparing the measurements presented in this study with data collected using other measurement methods.

BB event identification
The summer of 2019 was a relatively low-activity fire season in the Pacific Northwest, while higher temperatures in the Arctic and sub-Arctic contributed to extensive wildfires across boreal forests in Alaska and Siberia.We identified 13 BB events from July to September 2019, which ranged from 2 to 44 h in duration (Fig. 1).These were split when discernable plumes were separated by a multi-hour drop below the BB event criteria of σ scat > 20 Mm −1 and CO > 110 ppbv.BB event criteria was chosen based on our previous work (Wigder et al., 2013).Only 8 % of the 1 h averages at MBO met our criteria for a BB event.Most observed BB events exhibited average σ scat and CO in the ranges of 20-40 Mm −1 and 130-170 ppbv, respectively.For comparison, 51 % of the 5 min averages met a higher BB event criterion (σ scat > 20 Mm −1 and CO > 150 ppbv) during the high fire year of 2015, where most event average σ scat and CO ranged from 100 to 300 Mm −1 and from 200 to 400 ppbv, respectively (Laing et al., 2016).Continuing the approach from Laing et al. (2016), we use the term "event", not "plume", because of the long duration of some of the events and the fact that some BB events observed in 2019 were influenced by emissions from multiple fires.A comparison of CO, σ scat , σ abs , and PM 1 for the BB event and background periods is presented in Fig. S1 in the Supplement.

Source and transport of BB events
NOAA HYSPLIT air-mass back trajectories and NASA MODIS Aqua and Terra images of smoke were used to identify the origin of each event and estimate the transport time.Two BB events (5-6 July), arriving at MBO with a negative WV (-1.53 g k −1 ) (Table 1) indicative of FT transport, were determined to originate from a complex of lightninginitiated fires in boreal forests of the interior of Alaska (AK).The fires started from 22 June to 2 July, and the smoke was transported to MBO over the course of 8-10 d (Fig. S2).Results from a FIREX-AQ model study by Makar et al. (2021) previously suggested wildfires in AK impacted aerosol optical depth (AOD) in North America during this period.
Three BB events with the shortest transport time to MBO observed in this study (∼ 10-15 h) (Fig. S3) originated from the Milepost 97 fire southeast of MBO near Canyonville, OR, which was initiated on 24 July and burned 5300 ha or 5.3 × 10 7 m 2 .Of the southwestern Oregon (SW OR) events, two exhibited a positive WV (0.15, 1.47 g k −1 ) indicative of BL transport typically observed for local fires and one exhibited a negative WV (−0.70 g k −1 ) indicative of FT arrival (Table 1).Six BB events that had multiple fires active along the back trajectory were categorized as "Regional" in Table 1.The August regional events may have been influenced by emissions from forest fires in Washington, Oregon, and/or California, with transport times estimated to be 1-3 d (Figs.S4 and S5).One August regional event had a WV near zero, indicative of mixed BL and FT influence.The September regional BB events originated from multiple forest fires directly south of MBO in Oregon and California, with transport times estimated to be < 12-48 h (Fig. S6).September regional events were characterized by a positive WV (0.85-4.2 g k −1 ) consistent with prior observations of BL transport of regional BB events to MBO (Wigder et al., 2013).
Fires in the boreal forest of Siberia near Lake Baikal were initiated by lightning in July 2019 and continued through August, burning over 3 million ha with smoke transported eastward towards Alaska and the continental USA.Ten-day air-mass back trajectories from MBO for the peak 12 and    (Johnson et al., 2021).Therefore, Siberian boreal forest fires were determined to be the origin of the 12 and 17 August BB events, with the transport time estimated to be > 14 d (Fig. S7).The negative WV of the Siberian events (−1.43, −2.82 g k −1 ) was indicative of long-range transport to MBO occurring in the FT, which is consistent with established transport pathways of Siberian BB events (Laing et al., 2016).

Overview of intensive aerosol properties of BB aerosol at MBO
Table 1 provides an overview of NERs and intensive optical properties of 13 BB events of Siberian, Alaskan, SW OR, and regional origin observed at MBO during the summer of 2019.Observed BB event ranges of σ scat / CO (0.33 to 0.91 Mm −1 ppbv −1 ), σ abs / CO (0.02 to 0.09 Mm −1 ppbv −1 ), and PM 1 / CO (0.05 to 0.26 µg m −3 ppbv −1 ) were comparable to prior summer measurements of transported BB aerosol (Laing et al., 2016;Baylon et al., 2017;Wigder et al., 2013).The range of BB event ω (σ scat /(σ scat + σ abs )) observed in summer 2019 (0.88-0.95) was lower, (i.e., more absorbing) than previously observed in summer 2015 at MBO (0.95-0.98).The observed range of event-averaged AAE (0.97 to 2.55) was also lower, indicative of greater BC than previous observations in summer 2015 (2.3 to 4.12).The precision and total uncertainties of intensive parameters derived from extensive measurements are provided for these events in Table S1.

Intensive aerosol properties of BB aerosol at MBO
We observed substantial differences in the intensive physical and optical aerosol properties of Siberian, Alaskan, SW OR, and regional BB events.The average (±1σ ) PM 1 / CO of SW OR (0.25 ± 0.09 µg m −3 ppbv −1 ) events transported < 13 h and regional events (0.17 ± 0.10 µg m −3 ppbv −1 ) transported 1-3 d were comparable to the PM 1 / CO emission factor ratio (0.16±0.11 µg m −3 ppbv −1 ) for fresh BB in temperate regions (Wigder et al., 2013;Akagi et al., 2012).Therefore, they can be characterized as relatively balanced in PM production and loss.In contrast, Siberian events were characterized by the lowest average PM 1 / CO (0.045 ± 0.007 µg m −3 ppbv −1 ).This result suggests that net PM 1 loss through deposition, evaporation, and/or cloud processing was greater than secondary aerosol production during > 14 d transport in the FT from Siberia.The concurrent measurement of decreasing OA/ CO with increased transport time in Farley et al. (2022) demonstrates that net OA loss through evaporation and deposition was greater than growth through secondary processing.Thus, the trend of PM 1 / CO follows the OA/ CO, which can support the observation of PM 1 loss in this study as OA is the dominant component of PM 1 .Wigder et al. (2013) and Weiss-Penzias et al. (2006) previously reported a similar association of decreased PM 1 / CO with increased transport > 750 km for wildfire smoke arriving at MBO in 2004-2011.However, measurements of elevated PM 1 / CO for the 8-10 d transported Alaska BB events (0.30 ± 0.07 µg m −3 ppbv −1 ) demonstrate that long-range transport in the FT is not always associated with greater PM 1 loss.Similarly, there was no significant difference in the PM 1 / CO of summer 2015 Siberian and regional BB events, which Laing et al. (2016) attributed to dry FT transport with limited precipitation.Observations at MBO thus indicate that long-range transport in the FT can yield a wide range of PM 1 / CO for BB events depending on removal characteristics.
BB aerosols lofted to the FT are typically less likely to be removed by dry deposition and can have a longer atmospheric lifetime of up to 40 d, but wet deposition can enhance removal and therefore reduce lifetime (Bond et al., 2013).Analysis of the impact of wet deposition of Canadian BB plumes demonstrates that it is the dominant mechanism for BC removal from the atmosphere and consequently determines lifetime and atmospheric burden (Franklin et al., 2014;Taylor et al., 2014).Notably, the 2019 Siberian events were transported over the Arctic, where inefficient longrange transport in the summer is attributed to meteorological conditions enhancing wet deposition (Mori et al., 2020).Raut et al. (2017) determined that a large fraction of BC particles in Siberian plumes transported to the Arctic region is mixed with sufficient water-soluble compounds to become CCNactive and is scavenged by large-scale precipitation and wet convective updrafts.They found the role of dry deposition to be minor and limited to the lower troposphere.We used precipitation rates from NOAA HYSPLIT ensemble back trajectories for the AK (Fig. S2, 10 d back-trajectory) and Siberian (Fig. S7, 14 d back-trajectory) BB events as a measure of wet deposition during long-range transport.Average precipitation rates for Siberian events (0.0038 mm h −1 ) were more than double Alaskan events (0.0016 mm h −1 ) (Fig. S8).Increased wet deposition during trans-Arctic FT transport thus likely contributed to the reduced PM 1 / CO in Siberian BB events observed at MBO.Indeed, Farley et al. (2022) concluded that the larger accumulation mode (700 nm D va ) in the Siberian BB event PM 1 mass size distribution was the result of increased aqueous-phase cloud processing during transport.
Examination of intensive aerosol absorption properties suggests that the Siberian BB aerosols were generally more absorptive and BC-dominant than the other BB events observed.The average MAE of the Siberian BB events (0.40 m 2 g −1 ) is similar to the regional events (0.43 m 2 g −1 ) but higher than the AK and SW OR events (0.29 and 0.24 m 2 g −1 , respectively).The greater MAEs in Siberian events are consistent with higher BC/PM 1 ratios in Siberian events than the other cases identified by Farley et al. (2022) via the concurrent SP-AMS measurements at MBO. Possible reasons for the high mass fractions of BC in Siberian aerosols were wet deposition and enhanced oxidative aging during long-range transport (Farley et al., 2022), which remove BC at a lower rate than the non-BC materials (Liu et al., 2020b).Another possible reason for the higher MAE in Siberian events is greater influence from flaming fires.Indeed, the Siberian events, despite wet deposition and prolonged atmospheric aging, showed higher normalized enhancement ratios of BC ( BC/ CO) than the other cases (Farley et al., 2022).Further information on the relative contributions of BC and BrC to BB absorptivity to the transported BB events observed at MBO can be inferred from the intensive aerosol optical property AAE (Figs. 2 and 3).Consistent with SP-AMS measurements, the long-range transported Siberia BB events exhibited low AAE values (1.29) indicative of BC dominance.Alaskan BB events exhibited similar low AAE values (1.27) indicative of BC dominance but were not characterized by enhanced MAE.Prior observations of Siberian events at MBO with similarly enhanced MAE and lower AAE were hypothesized to originate from hotter, more flaming portions of the fires (Laing et al., 2016)  Flaming fires are characterized by enhanced BC emissions and pyroconvective energy needed to loft the plume high into the atmosphere, where it can undergo long-range transport.ARCTAS-A aircraft measurements in Alaska reported a much larger BC/CO ratio in Siberian fire plumes than North American fire plumes (Kondo et al., 2011), providing further evidence that the emissions from flaming portions comprise the majority of BB transported from Siberian boreal forest fires to North America.An alternative explanation for low-AAE Siberian and Alaskan BB events is BrC removal by photolysis and oxidation during week-long transport (Laing et al., 2016;Dasari et al., 2019).Analysis of 10 years of aerosol properties at MBO by Zhang and Jaffe (2017) found Asian long-range transport wildfires with lower AAEs (1.45±0.02and 1.54±0.39)than regional BB events (1.81 ± 0.59).Similarly, we found a general decrease in AAE with increased transport time (Fig. 2).Regional events with 1-3 d transport had a very slightly elevated AAE (mean of 1.36), indicative of slightly more BrC mixed with BC.The highest AAE (mean of 1.88) indicated that stronger BrC was observed for the short-range transported (10-15 h) SW OR events.The highest values of AAE for the SW OR event were associated with the highest PM 1 concentration (Fig. S9), likely due to the presence of nearby, fresh, and concentrated plumes.These results are consistent with the short half life (∼ 9 h) of BrC and prior measurements of AAE declining as BrC decays over the course of hours, with little remaining after days of atmospheric transport (Forrister et al., 2015;Hems et al., 2021).A recent laboratory study also found that the imaginary part of BrC could be half decayed in a few hours, in line with the loss of its absorptivity after transport (Liu et al., 2021).
Furthermore, the laboratory work by Cappa et al. (2020) and field observations by Wu et al. (2021) suggest that the evolution of AAE and BrC absorptivity with photochemical aging is dependent on the burn conditions and the initial properties of emitted particles.Aerosols from more flaming fires show an initial enhancement of AAE and BrC absorptivity followed by a gradual decrease at longer aging times, while more smoldering aerosols are suggested to experience continuous decrease upon aging.Thus, the highest AAE of the SW OR events could also be related to burning conditions and the initial enhancement of AAE during the short-range transport.
The highest scattering ω (0.94 ± 0.01) was observed in Siberian plumes.The average ω values for the Alaska (0.93± 0.01) and SW OR (0.93 ± 0.01) events were elevated compared to the ω for regional events (0.90 ± 0.01).The relatively high ω for Siberian events is thus consistent with our finding of a significantly elevated MSE ( σ scat / PM 1 ) (8.76 ± 2.15 m 2 g −1 ).Satellite observations and 3D atmospheric modeling of 2016 Siberian BB suggested that an observed increase in ω with transport time was likely due to atmospheric processing of SOA, producing an increase in the mass scattering efficiency of BB aerosol (Konovalov et al., 2021).However, the MSE of long-range transported Siberian smoke plumes arriving at MBO in 2015 exhibited no difference from regional events (Laing et al., 2016), and MSE values of aged BB plumes are typically < 6 m 2 g −1 (Hand and Malm, 2007;McMeeking, 2005).No substantial differences were observed in the average MSE values for AK, SW OR, and regional events (3.65±0.08,3.40±0.42,3.97± 0.56 m 2 g −1 ).According to Mie theory, the uniquely enhanced scattering efficiency of the Siberian BB events sug-gests an increase in particle size towards the measurement wavelength (Seinfeld and Pandis, 2006).
To better understand the enhanced scattering observed in Siberian events, we related the variation in the MSE values that we observed to aerosol size distributions measured by the SMPS.We found the SMPS geometric mean diameter (D g ) to be correlated with MSE across all 2019 BB events (R 2 = 0.93) (Fig. 4), which is consistent with Mie theory and previous ambient observations of aged BB (Lowenthal and Kumar, 2004;Laing et al., 2016) and laboratory studies (McMeeking, 2005).However, MSE in 2019 was ∼ 40 % higher than 2015 at a similar D g (Laing et al., 2016).It was suggested that larger size distributions in 2015 BB events may be the result of more concentrated BB plumes with slower particle evaporation rates, leading to greater net PM and SOA accumulation during aging (Hodshire et al., 2019).This conclusion was supported by the correlation of 2015 BB event D g with CO, σ scat , and PM 1 , which can be thought of as surrogates for plume concentration.However, D g was not found to be correlated with CO, σ scat , or PM 1 in 2019, with the largest particles observed in the more dilute Siberian event.Plume concentration does not appear to be the determining factor in the particle size of 2019 BB events.
SAE is an intensive aerosol optical property that can be used as a measure of particle size, with small values indicating large sizes (Kleinman et al., 2020).All SAE values here (Figs. 2 and 3) are expected to be > 1, because only fine particles (PM 1 ) were measured.August-September BB event SAEs were negatively correlated with D g (R 2 = 0.85) (Fig. 4), generally supporting its use as a measure of particle size.SW OR and regional events exhibited elevated SAE values characteristic of the smaller size distributions of fresh to moderately aged BB plumes (May et al., 2014;Levin et al., 2010).In contrast, both long-range transported AK and Siberian events exhibited lower SAE values, suggesting that long-range transport may produce a shift in BB aerosol to larger particle sizes (Jung et al., 2012).However, particle morphology, composition, and coatings can also alter SAE values.Thin coatings on BC particles may yield lower SAE values (Zhang et al., 2020).
Multiple field observations have observed D g and MSE to increase as a function of photochemical age (e.g., Akagi et al., 2012;Carrico et al., 2016;Kleinman et al., 2020).An increase in MSE with age suggests a rearrangement of particle mass that favors large-diameter efficient scatters at the expense of small inefficient scatters.This could be the result of a transfer of mass between the gas and particulate phases (i.e., condensation) and amongst particles (i.e., coagulation) (Kleinman et al., 2020).We explore the relationship between SMPS number size distributions and SP-AMS aerosol composition measurements in Sect.3.5 to examine the potential influence of aerosol growth processes during transport of Siberian BB events.While secondary aerosol growth and/or coagulation during transport present plausible explanations for larger particles, we also consider the unexpected contribution of dust mixed with smoke.SAE was previously used by Zhang and Jaffe (2016) to identify larger particle sizes in long-range transport of industrial pollution from Asia, which was suggested to be due to mixing with mineral dust sources.In addition, prior measurements in spring have identified an association of dust mixed with smoke in long-range transported plumes from Asia with a lower SAE at MBO (Fischer et al., 2010).Elevated AAE alongside low SAE (Fig. 3) can be indicative of dust aerosols with enhanced absorption at short wavelengths (Bergstrom et al., 2007;Titos et al., 2017).However, the distribution and average AAE of the Siberian BB events were similar to the Alaskan BB events.Wagner et al. (2018) illustrated via high-resolution eddy simulation that the energy released by wildfires leads to a significant increase in near-surface wind speed and enhanced dust uplift potential.Based on this model, we hypothesize that the enhanced pyroconvective energy of Siberian fires that injects BC into the FT (Laing et al., 2016)   side smoke into the FT, where it was transported to MBO.Enhancements in much larger coarse-mode particles are typically used to identify the influence of dust (Lee and Cho, 2007), but recent measurements in the western USA found evidence of fine-mode dust mixed with smoke (Maudlin et al., 2015;Schlosser et al., 2017;Jahn et al., 2021).Therefore, we propose that dust mixed with smoke may have contributed to the enhanced larger particle size distribution in the Siberian BB events.In the following sections we present satellite data showing the transport of dust alongside smoke from Siberian boreal forest fires (Sect.3.4) and SMPS size distributions to examine the potential contributions of dust to Siberian BB events (Sect.3.5).

Satellite identification of dust mixed with smoke in Siberian BB events
The NASA Dust Score satellite product, which is computed from infrared channels of the NASA Aqua/AIRS satellite (DeSouza-Machado et al., 2010), was overlaid onto MODIS images to examine the presence of dust mixed with smoke emitted and transported from identified source fires.Dust Score is a qualitative representation of the presence of dust in the atmosphere, with values ranging from 400 to 500 here.The sensor resolution is 45 km with once-per-day coverage.A significant Dust Score (> 360) was not observed near the origin of the Alaskan, SW OR and regional BB events.In contrast, a large geographic area of dust mixed and transported with smoke is observed in time series (5, 7, 10 August) of Dust Score overlaid on MODIS images of Siberian boreal forest fires (Fig. 5).We used V3.30 aerosol classification products from the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) instrument on the Cloud-Aerosol Lidar Infrared Pathfinder Satellite Observation (CALIPSO) satellite (Winker et al., 2010) to further examine long-range transport of dust mixed with smoke for Alaskan and Siberian BB events.A time series (10, 12, and 17 August) of the NASA CALIPSO aerosol-type satellite cross sections suggests that "dust", "polluted dust", and "smoke" underwent transport in the FT from Siberia to the western USA (Fig. S10).A time series (30 June and 2 and 5 July) of the NASA CALIPSO aerosol-type satellite cross sections shows "polluted continental/smoke" and "smoke", with comparatively minor contributions from "dust" transported from AK to the western USA (Fig. S11).CALIPSO has previously been used in a similar manner to identify the transport of dust with smoke in plumes of Asian origin arriving at MBO in spring (Fischer et al., 2009).However, due to the lack of direct measurement of dust in this study, we cannot conclude with confidence that dust was present in the Siberian BB events.Nevertheless, the CALIPSO result suggests that fires in the Siberian boreal forest, as well as other non-arid regions, should be considered a source of airborne mineral dust.

Aerosol number size distribution and composition
Aerosol number size distribution (30-600 nm) was measured by SMPS for nine BB events as well as interstitial background periods in August and September 2019 (Fig. 6).Geometric diameters (D g ) of aerosol number size distributions are presented in Table 1, with further size distribution parameters detailed in Table S2.The SW OR and regional events exhibited similar unimodal number size distributions, with D g values between 146 and 162 nm.A prominent "tail" consisting of higher-than-expected number concentrations of small-diameter particles (30-90 nm) was observed in the size distribution for events arriving in the BL at MBO, similar to results in Laing et al. (2016).Background periods exhibited a unimodal number size distribution with a smaller D g (101 nm).Siberian events were characterized by bimodal number size distributions, with a larger accumulation-mode peak > 200 nm and a lower ultrafine-mode peak < 100 nm (Fig. 6), which could reflect new particle formation following cloud processing (Taylor et al., 2014).
The larger accumulation modes of Siberian events suggest contributions from fine dust particles identified in the Dust Score and CALIPSO satellite products.Submicron dust particles as small as 300 nm have been observed in Africa (Kaaden et al., 2009;Kandler et al., 2009;Denjean et al., 2016), and Asian dust storms have been found to be correlated with an increase in 200-500 nm particles (Liang et al., 2013).Furthermore, submicron dust particles in the size range measured by SMPS in this study are expected to be transported more efficiently than the coarse dust particles that may have been present in the initial Siberian BB plume.Farley et al. ( 2022) measured a larger mass-weighted size distribution by SP-AMS for Siberian BB events compared to regional BB events.However, without aerosol chemical composition measurements of dust, which are not measured by SP-AMS, we cannot conclude with confidence that dust was present in the Siberian BB events.
Ultrafine particles are typically not the result of longrange transport, as it would be expected that small particles would grow through coagulation relatively quickly.Therefore, the observed < 100 nm modes in long-range transported Siberian BB events are indicative of the influence of entrained background air and/or new particle formation (NPF).A year-long study at the Swiss mountaintop observatory Jungfraujoch suggested summer NPF to be triggered by pre-viously entrained precursors or by boundary-layer injections on the same day (Tröstl et al., 2016).

Conclusions
We characterized the physical and optical properties of 13 aged biomass burning events observed at Mt. Bachelor Observatory in summer 2019.Our conclusions were as follows.
-Low PM 1 / CO values of Siberian BB events are likely the result of enhanced wet deposition in trans-Pacific and trans-Arctic FT transport from Siberia to MBO.
-BB events from Siberian and Alaskan boreal forest fires with multi-day transport in the FT were associated with lower AAE values indicative of BC dominance.The aerosols during these events have lower AAE values compared to regional BB events.We propose two hypotheses for the BC dominance of boreal forest BB events.
a.The short half life of BrC leads to its decays over the course of hours, with little remaining in the plume after a week or more of transport.
b. Siberian events with higher MAE originated from hotter, more flaming portions of the fires.

Figure 1 .
Figure1.Time series (UTC) of the extensive properties CO (ppb), PM 1 (µg m −3 ), scattering (Mm −1 ), and absorption (Mm −1 ) measured at MBO from 1 July to 14 September 2019.BB event periods are highlighted by grey-filled dashed boxes, with source identification and the number of discrete BB events with σ scat,green > 20 Mm −1 and CO > 110 ppbv per period above.The blue, green and red lines in the σ abs and σ scat plots represent measurements in the blue, green and red channels, respectively.

Figure 2 .
Figure 2. Combined violin and box plots of 1 h average SAE and AAE values for July-September 2019 BB events observed at MBO grouped by source (Siberia, Alaska, regional, SW OR).SAE describes particle sizes, with small values indicating large sizes, and AAE describes particle composition with larger values, indicating greater BrC content.Violin plots in red show the rotated Gaussian kernel probability density, with the mean shown as an open diamond, the median shown as a solid diamond, and 1 h average values shown as open circles.Boxes in black represent upper and lower quartiles, with whiskers representing the minimum and maximum.

Figure 3 .
Figure 3. 1 h average SAE and AAE values plotted for all BB events (July-September 2019) observed at MBO, with classification categories from Cappa et al. (2016).SAE describes particle sizes, with small values indicating large sizes, and AAE describes particle composition, with larger values indicating greater BrC content.BB events are colored by the source, with boreal (Alaska, Siberia) shown as triangles in shades of grey, SW OR shown as squares in shades of brown, and regional shown as circles in shades of blue to green.

Figure 4 .
Figure 4. (a) Mass scattering efficiency (MSE) versus geometric mean diameter (D g ) from the SMPS size distribution of the 2019 BB events observed at MBO, with linear regression and correlation coefficients shown.(b) D g and SAE for the August-September 2019 events observed at MBO, with linear regression and correlation coefficients shown.The Siberian BB events are shown with squares and all others as circles.Both relationships are statistically significant with P < 0.05.

Figure 5 .
Figure 5. (a) 1 August, (b) 5 August, and (c) 10 August 2019 NASA MODIS true color satellite images of Siberian boreal forest fires.Fire and thermal anomalies shown in red dots and Dust Score values shown in pixels colored according to the embedded scale.Areas of dust score and visible smoke are outlined in orange boxes.

Figure 6 .
Figure 6.Event-averaged aerosol number size distributions in dN/d log(D p ) measured by SMPS for the SW OR (left axis), regional (left axis), and Siberian (right axis) BB events identified in August-September 2019 at MBO as well as the background average aerosol number size distribution (right axis).

Table 1 .
Overview of intensive plume properties observed in June-September 2019 at MBO. WV was calculated as the difference between the WV at the peak CO of the event and the cutoff value.AAE and SAE values are event averages of hourly data, and all other NERs were calculated as the slope of reduced major axis (RMA) regression of hourly data.MSE is the NER of σ Farley at al. (2022) calculated from 550 nm (green) measurement wavelength data.Geometric diameter (D g ) was calculated from lognormal fits of event-averaged SMPS size distributions.Category values at the bottom of the table are averages.Events with measurements reported byFarley at al. (2022)are identified by an asterisk.