Atmos Chem Phys 21 12227ndash12241 2021httpsdoiorg105194acp-21-12227-2021copy Author(s) 2021 This work is distributed underthe Creative Commons Attribution 40 License

Measurement report Saccharide composition in atmospheric fineparticulate matter during spring at the remote sites of southwestChina and estimates of source contributionsZhenzhen Wang1 Di Wu1 Zhuoyu Li1 Xiaona Shang1 Qing Li1 Xiang Li1 Renjie Chen2 Haidong Kan2Huiling Ouyang3 Xu Tang3 and Jianmin Chen13

1Shanghai Key Laboratory of Atmospheric Particle Pollution and Prevention (LAP3) Department of Environmental Scienceamp Engineering Fudan Tyndall Centre Fudan University Shanghai 200438 China2Key Lab of Public Health Safety of the Ministry of Education NHC Key Laboratory of Health Technology AssessmentSchool of Public Health Fudan University Shanghai 200032 China3IRDR International Center of Excellence on Risk Interconnectivity and Governance on WeatherClimate Extremes Impactand Public Health Institute of Atmospheric Sciences Fudan University Shanghai 200438 China

Correspondence Jianmin Chen (jmchenfudaneducn)

Received 29 January 2021 ndash Discussion started 29 March 2021Revised 2 July 2021 ndash Accepted 2 July 2021 ndash Published 16 August 2021

Abstract Based on source-specific saccharide tracers thecharacteristics of biomass burning (BB) and biogenic emis-sions of saccharides were investigated in three rural sites atLincang which is 65 covered with forest in the south-west border of China The total saccharides accounted for84plusmn 27 of organic carbon (OC) and 16plusmn 06 ofPM25 The measured anhydrosugars accounted for 485 oftotal saccharides among which levoglucosan was the mostdominant species The high level of levoglucosan was bothattributed to the local BB activities and biomass combustionsmoke transported from the neighboring regions of South-east Asia (Myanmar) and the northern Indian subcontinentThe measured mono- or disaccharides and sugar alcohols ac-counted for 249plusmn 83 and 266plusmn 99 of the total sac-charides respectively and both proved to be mostly emit-ted by direct biogenic volatilization from plant material orsurface soils rather than byproducts of polysaccharide break-down during BB processes Five sources of saccharides wereresolved by non-negative matrix factorization (NMF) anal-ysis including BB soil microbiota plant senescence air-borne pollen and plant detritus with contributions of 340 160 210 237 and 53 respectively The resultsprovide information on the magnitude of levoglucosan andcontributions of BB as well as the characteristic of biogenicsaccharides at the remote sites of southwest China which

can be further applied to regional source apportionment mod-els and global climate models

1 Introduction

Biomass burning (BB) and biogenic aerosols are thought toplay important roles in air quality human health and cli-mate through direct or indirect effects (Jacobson et al 2000Christner et al 2008 Poumlschl et al 2010 Despreacutes et al2012 Chen et al 2017 Tang et al 2019) Atmospheric sac-charide components have been extensively reported to orig-inate from natural or anthropogenic biomass burning (BB)suspended soil or dust and primary biological aerosol par-ticles (PBAPs) eg fungal and fern spores pollens algaefungi bacteria plant debris and biogenic secondary organicaerosol (SOA) (eg Rogge et al 1993 Graham et al 2003Jaenicke 2005 Medeiros et al 2006 Elbert et al 2007 Fuet al 2013) As one of the major classes of water-soluble or-ganic compounds saccharides in atmospheric aerosols havebeen detected over urban areas forests mountains and re-mote marine regions (Pashynska et al 2002 Yttri et al2007 Fu et al 2009 Burshtein et al 2011 Jia and Fraser2011 Chen et al 2013 Pietrogrande et al 2014 Li et al2016a b) It has been reported that saccharides account for

Published by Copernicus Publications on behalf of the European Geosciences Union

12228 Z Wang et al Saccharide composition in atmospheric fine particulate matter

13ndash26 of the total organic compound mass identified incontinental aerosols and up to 63 of oceanic aerosols (Si-moneit et al 2004)

Levoglucosan and related anhydrosugar isomers (man-nosan and galactosan) produced from pyrolysis of cellu-lose and hemicellulose are considered to be relatively sta-ble in the atmosphere (Schkolnik et al 2005 Puxbaum etal 2007) thus have been recognized as specific molec-ular markers for BB source emissions (Simoneit et al1999 Fraser and Lakshmanan 2000 Sullivan et al 2014)However some studies have challenged this knowledge andproved that levoglucosan alone was unsuitable to be a dis-tinct marker for BB in various regions and periods This isbecause there is evidence that levoglucosan was also emit-ted from non-BB sources (Wu et al 2021) such as coalburning (Rybicki et al 2020 Yan et al 2018) open wasteburning (Kalogridis et al 2018) incense burning (Tsai et al2010) and food cooking (Reyes-Villegas et al 2018) It wasreported that the levoglucosan emission contribution of BBsources ranged from 213 to 959 (Wu et al 2021) Cur-rent studies in China have reported values of 26ndash2891 and116ndash18031 ng mminus3 respectively over Beijing and Wangduin summer (Yan et al 2019) 24ndash10641 ng mminus3 over Shang-hai all year round (Xiao et al 2018) 156ndash4729 ng mminus3

over Guangzhou (Zhang et al 2010) 211ndash915 ng mminus3 overHong Kong (Sang et al 2011) 602ndash4819 ng mminus3 overXirsquoan (Yang et al 2012) 360ndash18209 ng mminus3 over Chengdu(Yang et al 2012) and 101ndash3834 ng gminus1 dry weight in cry-oconites over the Tibetan Plateau (Li et al 2019) In northChina the high concentration of levoglucosan was a seri-ous problem due to the drastic enhancement of coal and BBfor house heating in winter and autumn (Zhang et al 2008Zhu et al 2016) The BB pollution might be exacerbatedunder unfavorable meteorological conditions such as in theChengdu Plain (Chen and Xie 2014) In general BB witha notable contribution to organic carbon (OC) was an im-portant source of fine particulate matter in China (Zhanget al 2008 Cheng et al 2013 Chen et al 2017) Con-trols on BB could be an effective method to reduce pollutantemissions Recently a study reported that total levoglucosanemissions in China exhibited a clear decreasing trend from2014 (1457 Gg) to 2018 (809 Gg) (Wu et al 2021) sug-gesting BB activities might be reduced in China

Saccharide compounds including a variety of primary sac-charides (monosaccharides and disaccharides) and sugar al-cohols (reduced sugars) have been measured to estimate thecontribution of biogenic aerosols including fungi virusesbacteria pollen and plant and animal debris (Simoneit etal 2004 Jaenicke et al 2007) For instance arabitol andmannitol have been proposed as biomarkers for airborne fun-gal spores (Bauer et al 2008 Zhang et al 2010 Holdenet al 2011 Liang et al 2013a b) because both of themcan function as storage or transport carbohydrates to regu-late intracellular osmotic pressure (Bauer et al 2008) Glu-cose and sucrose are thought to originate from natural bio-

genic detritus including numerous microorganisms plantsand animals (Simoneit et al 2004 Tominaga et al 2011)As the oxidation products of isoprene methyltetrols (includ-ing 2-methylthreitol and 2-methylerythritol) have been sug-gested as tracers of isoprene-derived SOA (Claeys et al2004 Kleindienst et al 2007 Ding et al 2016) In a pre-vious study the contributions of fungal spores to OC wereestimated to be 141plusmn 105 and 73plusmn 33 respectivelyat the rural and urban sites of Beijing (Liang et al 2013b)Airborne pollen and fungal spores contributed 12 ndash22 to the total OC in ambient aerosols collected in Toronto(Womiloju et al 2003) Jaenicke (2005) found that PBAPscould account for 20 ndash30 of the total atmospheric PM(gt 02 mm) from Lake Baikal (Russia) and Mainz (Ger-many) However studies on quantifying the abovementionedbiogenic aerosol contributions to ambient aerosol are inade-quate

Lincang located on the southwest border of China is a tra-ditional agricultural area of Yunnan Province where plant-ing large areas of tea sugar cane rubber macadamia nutsetc is common It is the largest production base of black teaand macadamia nuts in China Referring to the official web-site of Lincang Municipal Peoplersquos Government the forestcoverage of Lincang reaches 65 It has a wide variety ofplant species and has six nature reserves covering an areaof sim 222 000 ha accounting for 856 of the total area Asa residential area for ethnic minorities Lincang has uniqueculture humanity and living habits The proportion of housesthat use wood burning for cooking is very high in villages inthe vicinity as well as in a large area of Southeast Asia andforest fires frequently occur in this area especially in the dryseason (MarchndashApril) These facts imply that there are abun-dant biogenic aerosols in Lincang and BB pollution may bean essential potential source of air pollution However littleinformation on the magnitude of biogenic and BB tracers inthis area is available The contributions of biogenic aerosoland BB as well as BB types are poorly understood

In this study the sampling was conducted from 8 Marchto 9 April 2019 at three mountaintop sites of Lincang whichis an ideal site for investigating the BB emission characteris-tics BB tracers including anhydrosugars and K+ as well asbiogenic aerosol tracers (primary saccharides and sugar alco-hols) were measured to gain the information on source andcontributions of BB and biogenic emissions to PM25 overthe rural Lincang This study would be useful and valuablefor providing reliable information on sources and magnitudesof saccharides involving rural BB and biological emissions inChina

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12229

2 Experimental section

21 Aerosol sampling

PM25 samples were simultaneously collected on threemountaintop sites in Lincang Datian (2411 N 10013 E1960 m asl) Dashu (2412 N 10011 E 1840 m asl)and Yakoutian (2412 N 10009 E 1220 m asl) whichare located sim 300 km west of Kunming (the capital of Yun-nan Province in China) and sim 120 km east from the Burmaborder (shown in Fig S1 in the Supplement) These sitesare surrounded by massive mountains and scattered villageswithout obvious nearby traffic or major industrial emissionsEach sampling was performed over a 235 h period everyday and was collected on quartz by high-volume air sam-plers (Thermo) equipped with a size-selective inlet to samplePM25 at a flow rate of 113 m3 minminus1 Altogether 91 sam-ples were collected

Quartz filters (Whatman 2032times 2540 cm) were pre-baked at 550 C for 4 h in a muffle furnace to remove organicmaterial and they were then stored in prebaked aluminumfoils The samples were stored at about minus20 C in a refriger-ator until analysis Field blanks were collected by mountingfilters in the sampler without air flow to replicate the envi-ronmental exposure The data reported were corrected by theblanks at the sampling sites

22 Measurements

The concentrations of OC and elemental carbon (EC) weremeasured using a multiwavelength carbon analyzer (DRImodel 2015 Aerosol Inc USA) Typically a 058 cm2

punch of the filter was placed on a boat inside the thermaldesorption chamber of the analyzer and then stepwise heat-ing was applied Carbon fractions were obtained followingthe Interagency Monitoring of PROtected Visual Environ-ments (IMPROVE-A) thermaloptical reflectance (TOR) pro-tocol (Chow et al 2007) Replicate analyses were conductedonce every 10 samples A blank sample was also analyzedand used to correct the sample results

A punch (47 cm2) of each quartz filter was ultrason-ically extracted with 100 mL deionized water (resistiv-ity= 182 M cmminus1) for 40 min The aqueous extracts werefiltrated through syringe filters (PTFE 022 microm) to removeinsoluble materials Ion chromatography (Metrohm Switzer-land) coupled with Metrosep C6-150 and A6-150 columnswas used to detect water-soluble ions (Clminus NOminus3 POminus4 SO2minus

4 Na+ NH+4 K+ Mg2+ and Ca2+) with a detectionlimit (DL) range of 0001ndash0002 microg mminus3

Five saccharide alcohols (glycerol erythritol inositol ara-bitol and mannitol) and five primary saccharides (fructoseglucose mannose sucrose and trehalose) together withthree anhydrosugars (levoglucosan mannosan and galac-tosan) were quantified by an improved high-performanceanion-exchange chromatograph coupled with a pulsed am-

perometric detector (Engling et al 2006 Caseiro et al2007 Zhang et al 2013) This method developed by En-gling et al (2006) was validated to be a powerful method forthe detection of carbohydrates without derivatization tech-niques and it has been successfully applied to the atmo-spheric tracers (eg Zhang et al 2010 Holden et al 2011Liang et al 2013a b Li et al 2016a b Kalogridis et al2018 Yan et al 2018) The separation of the saccharides wasperformed on an ion chromatograph (Metrohm Switzerland)equipped with a Metrosep Carb 4ndash250 analytical column anda guard column The aqueous eluent of sodium hydroxideand sodium acetate was pumped by a dual-pump module at aflow rate of 04 mL minminus1 The low concentration of 50 mMsodium hydroxide and 10 mM sodium acetate (eluent A) wasapplied to pump 1 while the high concentration of 250 mMsodium hydroxide and 50 mM sodium acetate (eluent B) wasapplied to pump 2 The gradient generator was set as 0ndash10 min 100 of eluent A 10ndash20 min 50 of eluent A and50 of eluent B 20ndash50 min 100 of eluent B and 50ndash60 min 100 of eluent A for equilibration The extractionefficiency of this analytical method was determined to bebetter than 90 based on analysis of quartz filters spikedwith known amounts of mannitol The method DL of the re-ferred carbohydrate compounds was 0005ndash001 mg Lminus1 Allcarbohydrate species were below detection limits in the fieldblanks

23 Other data

The meteorological parameters including temperature (T )relative humidity (RH) solar irradiation (W mminus2) and rain-fall (mm) were obtained from the Physical Sciences Labora-tory of NOAA (httpspslnoaagov last access 8 Septem-ber 2020) The temporal changes in meteorological variablesover the observation sites during the sampling periods areshown in Fig S2

In order to characterize the origin and transport pathwayof the air masses to the sampling sites 72 h back-trajectoriesof the aerosol were calculated using the HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) modeldeveloped by NOAAARL (Draxler and Hess 1998) viathe NOAA ARL READY website (httpreadyarlnoaagovHYSPLITphp last access 6 September 2020) with an end-point height of 1500 m To investigate the influence of BBemissions fire pixel counts were obtained from ModerateResolution Imaging Spectroradiometer (MODIS) observa-tions on NASA satellites (httpsearthdatanasagov last ac-cess 3 September 2020)

24 Statistical analysis

A Pearsonrsquos correlation test was performed using the statis-tical product and service solutions software for the datasetcontaining ambient concentrations of the measured saccha-rides inorganic ions and solar irradiation Non-negative ma-

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12228 Z Wang et al Saccharide composition in atmospheric fine particulate matter

13ndash26 of the total organic compound mass identified incontinental aerosols and up to 63 of oceanic aerosols (Si-moneit et al 2004)

Levoglucosan and related anhydrosugar isomers (man-nosan and galactosan) produced from pyrolysis of cellu-lose and hemicellulose are considered to be relatively sta-ble in the atmosphere (Schkolnik et al 2005 Puxbaum etal 2007) thus have been recognized as specific molec-ular markers for BB source emissions (Simoneit et al1999 Fraser and Lakshmanan 2000 Sullivan et al 2014)However some studies have challenged this knowledge andproved that levoglucosan alone was unsuitable to be a dis-tinct marker for BB in various regions and periods This isbecause there is evidence that levoglucosan was also emit-ted from non-BB sources (Wu et al 2021) such as coalburning (Rybicki et al 2020 Yan et al 2018) open wasteburning (Kalogridis et al 2018) incense burning (Tsai et al2010) and food cooking (Reyes-Villegas et al 2018) It wasreported that the levoglucosan emission contribution of BBsources ranged from 213 to 959 (Wu et al 2021) Cur-rent studies in China have reported values of 26ndash2891 and116ndash18031 ng mminus3 respectively over Beijing and Wangduin summer (Yan et al 2019) 24ndash10641 ng mminus3 over Shang-hai all year round (Xiao et al 2018) 156ndash4729 ng mminus3

over Guangzhou (Zhang et al 2010) 211ndash915 ng mminus3 overHong Kong (Sang et al 2011) 602ndash4819 ng mminus3 overXirsquoan (Yang et al 2012) 360ndash18209 ng mminus3 over Chengdu(Yang et al 2012) and 101ndash3834 ng gminus1 dry weight in cry-oconites over the Tibetan Plateau (Li et al 2019) In northChina the high concentration of levoglucosan was a seri-ous problem due to the drastic enhancement of coal and BBfor house heating in winter and autumn (Zhang et al 2008Zhu et al 2016) The BB pollution might be exacerbatedunder unfavorable meteorological conditions such as in theChengdu Plain (Chen and Xie 2014) In general BB witha notable contribution to organic carbon (OC) was an im-portant source of fine particulate matter in China (Zhanget al 2008 Cheng et al 2013 Chen et al 2017) Con-trols on BB could be an effective method to reduce pollutantemissions Recently a study reported that total levoglucosanemissions in China exhibited a clear decreasing trend from2014 (1457 Gg) to 2018 (809 Gg) (Wu et al 2021) sug-gesting BB activities might be reduced in China

Saccharide compounds including a variety of primary sac-charides (monosaccharides and disaccharides) and sugar al-cohols (reduced sugars) have been measured to estimate thecontribution of biogenic aerosols including fungi virusesbacteria pollen and plant and animal debris (Simoneit etal 2004 Jaenicke et al 2007) For instance arabitol andmannitol have been proposed as biomarkers for airborne fun-gal spores (Bauer et al 2008 Zhang et al 2010 Holdenet al 2011 Liang et al 2013a b) because both of themcan function as storage or transport carbohydrates to regu-late intracellular osmotic pressure (Bauer et al 2008) Glu-cose and sucrose are thought to originate from natural bio-

genic detritus including numerous microorganisms plantsand animals (Simoneit et al 2004 Tominaga et al 2011)As the oxidation products of isoprene methyltetrols (includ-ing 2-methylthreitol and 2-methylerythritol) have been sug-gested as tracers of isoprene-derived SOA (Claeys et al2004 Kleindienst et al 2007 Ding et al 2016) In a pre-vious study the contributions of fungal spores to OC wereestimated to be 141plusmn 105 and 73plusmn 33 respectivelyat the rural and urban sites of Beijing (Liang et al 2013b)Airborne pollen and fungal spores contributed 12 ndash22 to the total OC in ambient aerosols collected in Toronto(Womiloju et al 2003) Jaenicke (2005) found that PBAPscould account for 20 ndash30 of the total atmospheric PM(gt 02 mm) from Lake Baikal (Russia) and Mainz (Ger-many) However studies on quantifying the abovementionedbiogenic aerosol contributions to ambient aerosol are inade-quate

Lincang located on the southwest border of China is a tra-ditional agricultural area of Yunnan Province where plant-ing large areas of tea sugar cane rubber macadamia nutsetc is common It is the largest production base of black teaand macadamia nuts in China Referring to the official web-site of Lincang Municipal Peoplersquos Government the forestcoverage of Lincang reaches 65 It has a wide variety ofplant species and has six nature reserves covering an areaof sim 222 000 ha accounting for 856 of the total area Asa residential area for ethnic minorities Lincang has uniqueculture humanity and living habits The proportion of housesthat use wood burning for cooking is very high in villages inthe vicinity as well as in a large area of Southeast Asia andforest fires frequently occur in this area especially in the dryseason (MarchndashApril) These facts imply that there are abun-dant biogenic aerosols in Lincang and BB pollution may bean essential potential source of air pollution However littleinformation on the magnitude of biogenic and BB tracers inthis area is available The contributions of biogenic aerosoland BB as well as BB types are poorly understood

In this study the sampling was conducted from 8 Marchto 9 April 2019 at three mountaintop sites of Lincang whichis an ideal site for investigating the BB emission characteris-tics BB tracers including anhydrosugars and K+ as well asbiogenic aerosol tracers (primary saccharides and sugar alco-hols) were measured to gain the information on source andcontributions of BB and biogenic emissions to PM25 overthe rural Lincang This study would be useful and valuablefor providing reliable information on sources and magnitudesof saccharides involving rural BB and biological emissions inChina

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12229

2 Experimental section

21 Aerosol sampling

PM25 samples were simultaneously collected on threemountaintop sites in Lincang Datian (2411 N 10013 E1960 m asl) Dashu (2412 N 10011 E 1840 m asl)and Yakoutian (2412 N 10009 E 1220 m asl) whichare located sim 300 km west of Kunming (the capital of Yun-nan Province in China) and sim 120 km east from the Burmaborder (shown in Fig S1 in the Supplement) These sitesare surrounded by massive mountains and scattered villageswithout obvious nearby traffic or major industrial emissionsEach sampling was performed over a 235 h period everyday and was collected on quartz by high-volume air sam-plers (Thermo) equipped with a size-selective inlet to samplePM25 at a flow rate of 113 m3 minminus1 Altogether 91 sam-ples were collected

Quartz filters (Whatman 2032times 2540 cm) were pre-baked at 550 C for 4 h in a muffle furnace to remove organicmaterial and they were then stored in prebaked aluminumfoils The samples were stored at about minus20 C in a refriger-ator until analysis Field blanks were collected by mountingfilters in the sampler without air flow to replicate the envi-ronmental exposure The data reported were corrected by theblanks at the sampling sites

22 Measurements

The concentrations of OC and elemental carbon (EC) weremeasured using a multiwavelength carbon analyzer (DRImodel 2015 Aerosol Inc USA) Typically a 058 cm2

punch of the filter was placed on a boat inside the thermaldesorption chamber of the analyzer and then stepwise heat-ing was applied Carbon fractions were obtained followingthe Interagency Monitoring of PROtected Visual Environ-ments (IMPROVE-A) thermaloptical reflectance (TOR) pro-tocol (Chow et al 2007) Replicate analyses were conductedonce every 10 samples A blank sample was also analyzedand used to correct the sample results

A punch (47 cm2) of each quartz filter was ultrason-ically extracted with 100 mL deionized water (resistiv-ity= 182 M cmminus1) for 40 min The aqueous extracts werefiltrated through syringe filters (PTFE 022 microm) to removeinsoluble materials Ion chromatography (Metrohm Switzer-land) coupled with Metrosep C6-150 and A6-150 columnswas used to detect water-soluble ions (Clminus NOminus3 POminus4 SO2minus

4 Na+ NH+4 K+ Mg2+ and Ca2+) with a detectionlimit (DL) range of 0001ndash0002 microg mminus3

Five saccharide alcohols (glycerol erythritol inositol ara-bitol and mannitol) and five primary saccharides (fructoseglucose mannose sucrose and trehalose) together withthree anhydrosugars (levoglucosan mannosan and galac-tosan) were quantified by an improved high-performanceanion-exchange chromatograph coupled with a pulsed am-

perometric detector (Engling et al 2006 Caseiro et al2007 Zhang et al 2013) This method developed by En-gling et al (2006) was validated to be a powerful method forthe detection of carbohydrates without derivatization tech-niques and it has been successfully applied to the atmo-spheric tracers (eg Zhang et al 2010 Holden et al 2011Liang et al 2013a b Li et al 2016a b Kalogridis et al2018 Yan et al 2018) The separation of the saccharides wasperformed on an ion chromatograph (Metrohm Switzerland)equipped with a Metrosep Carb 4ndash250 analytical column anda guard column The aqueous eluent of sodium hydroxideand sodium acetate was pumped by a dual-pump module at aflow rate of 04 mL minminus1 The low concentration of 50 mMsodium hydroxide and 10 mM sodium acetate (eluent A) wasapplied to pump 1 while the high concentration of 250 mMsodium hydroxide and 50 mM sodium acetate (eluent B) wasapplied to pump 2 The gradient generator was set as 0ndash10 min 100 of eluent A 10ndash20 min 50 of eluent A and50 of eluent B 20ndash50 min 100 of eluent B and 50ndash60 min 100 of eluent A for equilibration The extractionefficiency of this analytical method was determined to bebetter than 90 based on analysis of quartz filters spikedwith known amounts of mannitol The method DL of the re-ferred carbohydrate compounds was 0005ndash001 mg Lminus1 Allcarbohydrate species were below detection limits in the fieldblanks

23 Other data

The meteorological parameters including temperature (T )relative humidity (RH) solar irradiation (W mminus2) and rain-fall (mm) were obtained from the Physical Sciences Labora-tory of NOAA (httpspslnoaagov last access 8 Septem-ber 2020) The temporal changes in meteorological variablesover the observation sites during the sampling periods areshown in Fig S2

In order to characterize the origin and transport pathwayof the air masses to the sampling sites 72 h back-trajectoriesof the aerosol were calculated using the HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) modeldeveloped by NOAAARL (Draxler and Hess 1998) viathe NOAA ARL READY website (httpreadyarlnoaagovHYSPLITphp last access 6 September 2020) with an end-point height of 1500 m To investigate the influence of BBemissions fire pixel counts were obtained from ModerateResolution Imaging Spectroradiometer (MODIS) observa-tions on NASA satellites (httpsearthdatanasagov last ac-cess 3 September 2020)

24 Statistical analysis

A Pearsonrsquos correlation test was performed using the statis-tical product and service solutions software for the datasetcontaining ambient concentrations of the measured saccha-rides inorganic ions and solar irradiation Non-negative ma-

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12229

2 Experimental section

21 Aerosol sampling

PM25 samples were simultaneously collected on threemountaintop sites in Lincang Datian (2411 N 10013 E1960 m asl) Dashu (2412 N 10011 E 1840 m asl)and Yakoutian (2412 N 10009 E 1220 m asl) whichare located sim 300 km west of Kunming (the capital of Yun-nan Province in China) and sim 120 km east from the Burmaborder (shown in Fig S1 in the Supplement) These sitesare surrounded by massive mountains and scattered villageswithout obvious nearby traffic or major industrial emissionsEach sampling was performed over a 235 h period everyday and was collected on quartz by high-volume air sam-plers (Thermo) equipped with a size-selective inlet to samplePM25 at a flow rate of 113 m3 minminus1 Altogether 91 sam-ples were collected

Quartz filters (Whatman 2032times 2540 cm) were pre-baked at 550 C for 4 h in a muffle furnace to remove organicmaterial and they were then stored in prebaked aluminumfoils The samples were stored at about minus20 C in a refriger-ator until analysis Field blanks were collected by mountingfilters in the sampler without air flow to replicate the envi-ronmental exposure The data reported were corrected by theblanks at the sampling sites

22 Measurements

The concentrations of OC and elemental carbon (EC) weremeasured using a multiwavelength carbon analyzer (DRImodel 2015 Aerosol Inc USA) Typically a 058 cm2

punch of the filter was placed on a boat inside the thermaldesorption chamber of the analyzer and then stepwise heat-ing was applied Carbon fractions were obtained followingthe Interagency Monitoring of PROtected Visual Environ-ments (IMPROVE-A) thermaloptical reflectance (TOR) pro-tocol (Chow et al 2007) Replicate analyses were conductedonce every 10 samples A blank sample was also analyzedand used to correct the sample results

A punch (47 cm2) of each quartz filter was ultrason-ically extracted with 100 mL deionized water (resistiv-ity= 182 M cmminus1) for 40 min The aqueous extracts werefiltrated through syringe filters (PTFE 022 microm) to removeinsoluble materials Ion chromatography (Metrohm Switzer-land) coupled with Metrosep C6-150 and A6-150 columnswas used to detect water-soluble ions (Clminus NOminus3 POminus4 SO2minus

4 Na+ NH+4 K+ Mg2+ and Ca2+) with a detectionlimit (DL) range of 0001ndash0002 microg mminus3

Five saccharide alcohols (glycerol erythritol inositol ara-bitol and mannitol) and five primary saccharides (fructoseglucose mannose sucrose and trehalose) together withthree anhydrosugars (levoglucosan mannosan and galac-tosan) were quantified by an improved high-performanceanion-exchange chromatograph coupled with a pulsed am-

perometric detector (Engling et al 2006 Caseiro et al2007 Zhang et al 2013) This method developed by En-gling et al (2006) was validated to be a powerful method forthe detection of carbohydrates without derivatization tech-niques and it has been successfully applied to the atmo-spheric tracers (eg Zhang et al 2010 Holden et al 2011Liang et al 2013a b Li et al 2016a b Kalogridis et al2018 Yan et al 2018) The separation of the saccharides wasperformed on an ion chromatograph (Metrohm Switzerland)equipped with a Metrosep Carb 4ndash250 analytical column anda guard column The aqueous eluent of sodium hydroxideand sodium acetate was pumped by a dual-pump module at aflow rate of 04 mL minminus1 The low concentration of 50 mMsodium hydroxide and 10 mM sodium acetate (eluent A) wasapplied to pump 1 while the high concentration of 250 mMsodium hydroxide and 50 mM sodium acetate (eluent B) wasapplied to pump 2 The gradient generator was set as 0ndash10 min 100 of eluent A 10ndash20 min 50 of eluent A and50 of eluent B 20ndash50 min 100 of eluent B and 50ndash60 min 100 of eluent A for equilibration The extractionefficiency of this analytical method was determined to bebetter than 90 based on analysis of quartz filters spikedwith known amounts of mannitol The method DL of the re-ferred carbohydrate compounds was 0005ndash001 mg Lminus1 Allcarbohydrate species were below detection limits in the fieldblanks

23 Other data

The meteorological parameters including temperature (T )relative humidity (RH) solar irradiation (W mminus2) and rain-fall (mm) were obtained from the Physical Sciences Labora-tory of NOAA (httpspslnoaagov last access 8 Septem-ber 2020) The temporal changes in meteorological variablesover the observation sites during the sampling periods areshown in Fig S2

In order to characterize the origin and transport pathwayof the air masses to the sampling sites 72 h back-trajectoriesof the aerosol were calculated using the HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) modeldeveloped by NOAAARL (Draxler and Hess 1998) viathe NOAA ARL READY website (httpreadyarlnoaagovHYSPLITphp last access 6 September 2020) with an end-point height of 1500 m To investigate the influence of BBemissions fire pixel counts were obtained from ModerateResolution Imaging Spectroradiometer (MODIS) observa-tions on NASA satellites (httpsearthdatanasagov last ac-cess 3 September 2020)

24 Statistical analysis

A Pearsonrsquos correlation test was performed using the statis-tical product and service solutions software for the datasetcontaining ambient concentrations of the measured saccha-rides inorganic ions and solar irradiation Non-negative ma-

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12230 Z Wang et al Saccharide composition in atmospheric fine particulate matter

trix factorization (NMF) analysis was utilized to resolvepotential emission source and estimate their contributionsto atmospheric saccharides NMF introduced by Lee andSeung (1999) was similar to positive matrix factorization(PMF) Both methods find two matrices (termed the con-tribution matrix of W and the source profile matrix of H)to reproduce the input data matrix (V) using the factoriza-tion approach (V=WH) as a positive constraint (Wge 0 andHge 0) However PMF forces the negative factors to be pos-itive but the NMF method only retains non-negative factorsNMF minimizes the conventional least-squares error andthe generalized KullbackndashLeibler divergence (Shang et al2018) Therefore the results obtained from NMF are moreresponsive to the original characteristics of input dataset andfewer factors will be extracted (Zhang et al 2019) Half ofthe DL was used for the value below the detection limit Inthis study galactosan mannose and inositol were excludedbecause their concentration in most samples was below theDL Concentrations of the other 10 saccharide species for atotal of 91 samples were subjected to NMF analysis The un-certainties in NMF analysis were estimated as 03 plus theanalytical detection limit according to the method by Xie etal (1999) The constant 03 corresponding to the log (geo-metric standard deviation) was calculated from the normal-ized concentrations for all measured species and it was usedto represent the variation of measurements

3 Results and discussion

31 Saccharide concentration and composition

The temporal variations of PM25 mass OC EC and vari-ous saccharides measured in all samples are shown in Fig 1A statistical summary of all the data is listed in Table S1During the sampling periods the PM25 mass concentra-tions ranged between 137 and 878 microg mminus3 with an aver-age value of 418 microg mminus3 The concentrations of OC andEC respectively were in the ranges of 25ndash224 and 03ndash43 microg mminus3 with average values of 84 and 17 microg mminus3 OCaccounted for 199plusmn37 of total PM25 mass The ambientconcentrations of the total saccharides varied between 2445and 12916 ng mminus3 with an average value of 6384 ng mminus3The total saccharides quantified in PM25 accounted for84plusmn 27 (range 38 ndash206 ) of the OC and accountedfor 16plusmn 06 (range 06 ndash30 ) of the PM25 Figure 2presents the average concentration levels of 12 measured sac-charide compounds categorized into anhydrosugar mono- ordisaccharide and sugar alcohol as well as the relative con-tribution of these saccharides for all samples The values foreach site are shown in Fig S3

311 Anhydrosugars

The average concentrations of levoglucosan and mannosanwere 2877 and 316 ng mminus3 respectively with respectiveranges of 956ndash7147 and 0ndash1347 ng mminus3 for all 91 sam-ples Galactosan was detected only in six samples with arange of 25ndash55 ng mminus3 The anhydrosugars accounted for501 of the total measured saccharides Levoglucosan wasthe most dominant species among all the saccharides Theaverage levoglucosan concentration in this study was com-parable to the value at urban Beijing collected in spring of2012 (above 200 ng mminus3) (Liang et al 2016) and at urbanXirsquoan collected in winter of 2015 (2685 ng mminus3) (Wang etal 2018) It was higher than the value at the rural Teng-chong mountain site (1938 ng mminus3) (Sang et al 2013) aturban Shanghai collected in spring of 2012 (660 ng mminus3) (Liet al 2016a) and at urban Hong Kong collected in spring of2004 (360 ng mminus3) (Sang et al 2011) as well as at urbanBeijing collected in summer of 2013 (494 ng mminus3) (Yan etal 2019) but it was lower than that at a rural site in Xirsquoan(093 mg mminus3) (Zhu et al 2017) and at a rural site in easterncentral India (2258 ng mminus3) (Nirmalkar et al 2015) Duringthe observation period several instances of elevated levoglu-cosan occurred peaking on 8 16 23 March and 1 April Itwas thought that the ambient levoglucosan was primarily at-tributed to domestic biomass fuel burning the high levoglu-cosan emission on these peak days might be from open BBevents

Regression analyses of levoglucosan and the other two an-hydrosugars (mannosan galactosan) are shown in Fig 3aLevoglucosan was highly correlated with mannosan andgalactosan with coefficients of determination (R) of 081(P lt 001) and 089 (P lt 001) respectively indicatingsimilar combustion sources of them The ratios of levoglu-cosan to mannosan (LM) and mannosan to galactosan(MG) had been employed to identify the specific types ofBB although these ratios were quite variable (Fabbri et al2009) Previous studies suggested that LM ratios for theburning of softwood were 3ndash10 hardwood ratios were 15ndash25 and those from crop residues were often above 40 (Chenget al 2013 Zhu et al 2015 Kang et al 2018) The averageLM and MG ratios were statistically reported as 326 and12 for crop residue combustion 40 and 39 for softwoodcombustion and 215 and 15 for hardwood combustion re-spectively (Sang et al 2013 Shen et al 2018) In this studythe ratios of LM and MG ranged from 47 to 161 (aver-age 97 n= 91) and from 39 to 61 (average 48 n= 6)respectively crudely indicating major contribution from soft-wood burning The samples collected from 31 March to1 April and from 8 to 10 March respectively had consid-erably lower and higher concentrations of mannosan thanpredicted by the levoglucosanndashmannosan regression model(Fig 3a) The results suggested that BB aerosols collectedfrom 31 March to 1 April (LM= 1152plusmn 134) and from8 to 10 March (LM= 657plusmn 053) originated from differ-

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12231

Figure 1 Temporal variations of OC EC PM25 and total sugars at the three sites during the sampling periods

Figure 2 The absolute concentration (bar chart) and the relativecontribution (pie chart) of various saccharide compounds during thesampling periods

ent types of BB compared with the remaining sampling pe-riods (LM= 934plusmn 120) Therefore the high levoglucosanemission from 31 March to 1 April and from 8 to 10 Marchmight be from different open BB events possibly an openagricultural waste burning event or a forest fire while the BBof most sampling days originated from biomass fuel for do-mestic cooking and heating It is worth noting that the peakdays from 31 March to 1 April (LM= 1152plusmn 134) nearedthe Qingming Festival One possibility for BB events is thatpeople burned joss paper as a sacrifice to their ancestors ac-cording to Chinese tradition

Both anhydrosugars and water-soluble potassium (K+)have been widely utilized as source tracers of BB emissions(eg Puxbaum et al 2007 Wang et al 2007 Zhang et al2008 Engling et al 2011) The daily variations of concen-trations of levoglucosan and K+ are shown in Fig S4 theregression analysis of K+ and three anhydrosugars is shownin Fig 3b The K+ concentration was weakly correlated withlevoglucosan mannosan and galactosan with R values of

033 028 and 074 respectively This could be explainedby the additional emissions of K+ from soil and seawaterSince Lincang is far from the coast sea salt could be negligi-ble Because of the inhomogeneity of crustal K+ associatedwith soil types it was difficult to fully account for crustalK+ contributions from soil (Harrison et al 2012 Chenget al 2013) The ratio of levoglucosan to K+ (LK+) wasalso used to track possible sources of BB in the previousstudies The ratios of LK+ strongly depended on BB pro-cesses namely smoldering and flaming Studies suggestedthat relatively high LK+ ratios were obtained from smolder-ing combustion at low temperatures compared with flamingcombustion (Schkolnik et al 2005 Lee et al 2010) Pre-vious results showed the emissions from the combustion ofcrop residuals such as rice straw wheat straw and corn strawexhibited comparable LK+ ratios typically below 10 Theaverage LK+ ratio in this study was 048plusmn 020 which ishigher than the ratio for wheat straw (010plusmn 000) and cornstraw (021plusmn 008) but is lower than the ratio for Asian ricestraw (062plusmn 032) (Cheng et al 2013) In this study higherLK+ ratios were observed during 8ndash10 March (120plusmn 019)compared to those from 31 March to 1 April (040plusmn 013)which suggested that the open fire event from 8 to 10 Marchwas more possibly due to smoldering combustion of residuesat low temperatures

Figure 3c and d show the scatterplots and regression anal-yses of K+ versus PM25 OC and EC and levoglucosanversus PM25 OC and EC respectively Linear regressionof K+ on PM25 OC and EC resulted in R values of 064063 and 062 respectively which were generally higherthan those of levoglucosan on PM25 OC and EC with R

values of 040 054 and 048 respectively This shows thatK+ is more highly correlated with PM25 OC and EC whichcould be explained by either the photo-oxidative decay oflevoglucosan (Hennigan et al 2010) or different types ofBB processes (Schkolnik et al 2005 Lee et al 2010) Evenso the results suggest that BB imposed a great impact on

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12232 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 3 (a) Regression analyses of levoglucosan versus the other two anhydrosugars (b) K+ versus the other three anhydrosugars (c) lev-oglucosan versus PM25 OC and EC and (d) K+ versus PM25 OC and EC

fine aerosols The ratio of levoglucosan to PM25 (LPM25)was also helpful in distinguishing the contributions of dif-ferent levoglucosan sources (Wu et al 2021) The ratio ofLPM25 in this study was 00041ndash00162 (average 00072)indicating that the levoglucosan emission in the areas mightmainly come from wood (001ndash009) and crop straw (0001ndash0008) not excluding incense burning (0001ndash0007) ritualitem burning (0004ndash0086) and meat cooking (0005ndash006)However certainly it was not from coal burning (00001ndash0001) or waste incineration (00022)

An empirical ratio of levoglucosan to OC (82 ) calcu-lated from main types of Chinese cereal straw (rice wheatand corn) based on combustion chamber experiments (Zhanget al 2007) was used to estimate the BB-derived OCThe average mass concentration of BB-derived OC was35344 ng mminus3 whilst the contributions of BB to OC was413 with a large range of 191 ndash739 The contri-butions were higher than those previously reported such as65 ndash11 in Hong Kong (Sang et al 2011) 18 ndash38 inBeijing (Zhang et al 2008) 189 ndash454 over southeast-ern Tibetan Plateau (Sang et al 2013) and 264 ndash302 in Xirsquoan (Zhang et al 2014) The large range of 191 ndash739 revealed that the daily contribution of BB variedgreatly suggesting that an open BB event or forest fire hap-pened occasionally The contribution apportionment of pri-mary BB might be underestimated due to the degradationof levoglucosan during atmospheric aging of BB-influencedair mass after long-range transport (Hennigan et al 2010

Mochida et al 2010 Lai et al 2014) Moreover Wu etal (2021) have reported that the total levoglucosan emissionexhibited a clear decreasing trend in China However it wasnoteworthy that the average concentration of levoglucosan(2877 ng mminus3) and the BB contributions to OC (413 )at Lincang mountain site were both higher than the valuesof 1918 ng mminus3 and 284 at Tengchong mountain site in2004 in spring (Sang et al 2013) This result suggests thatthere is no significant reduction in BB emissions in southwestYunnan Province

312 Mono- or disaccharides

The total concentrations of five mono- or disaccharidesincluding glucose fructose mannose sucrose and tre-halose were in the range of 252ndash3737 ng mminus3 (average1589 ng mminus3) which contributed 249plusmn 83 of the totalmeasured saccharides The average values of glucose fruc-tose mannose sucrose and trehalose were 312 246 27864 and 138 ng mminus3 respectively Sucrose was the domi-nant species in the mono- or disaccharides group The resultsagreed with those of previous studies (Yttri et al 2007 Jia etal 2010 Fu et al 2012) which found that sucrose was oneof the dominant species in fine aerosols in spring Rupturedpollen may be an important source of sucrose especially inthe spring blossom season (Yttri et al 2007 Fu et al 2012Miyazaki et al 2012) In spring and early summer farmlandtilling after wheat harvest causes an enhanced exposure of

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12233

soil containing wheat roots to the air which is beneficial forthe release of sucrose stored in roots (Medeiros et al 2006)thus resulting in a sharply increased sucrose concentration

It was reported that sugars such as glucose sucrose andfructose could be emitted from developing leaves (Grahamet al 2003) Glucose could be released from both soils andplant material (eg pollen fruits and their fragments) (Gra-ham et al 2003 Simoneit et al 2004 Fu et al 2012) Glu-cose and sucrose were rich in biologically active surface soils(Rogge et al 2007) In this study positive correlations werefound between sucrose and glucose (R = 052) (Table S2)suggesting a similar origin of glucose and sucrose in thisstudy Glucose and fructose have also been identified as a mi-nor product of cellulose pyrolysis because they were foundto be enriched in BB emissions (Nolte et al 2001) andthey correlated well with K+ (Graham et al 2002) and lev-oglucosan (Kang et al 2018) Herein no significant correla-tion was found between K+ levoglucosan and these mono-or disaccharides Therefore the detected glucose fructoseand sucrose might mostly be emitted by direct volatilizationfrom plant material or surface soils rather than as productsof polysaccharide breakdown during BB processes The highabundance of sucrose as well as glucose and fructose wasresponsible for biogenic aerosols associated with developingleaves and flowers as well as surface soil suspension

Trehalose as a stress protectant of various microorganismsand plants (Medeiros et al 2006 Jia and Fraser 2011) wasfound to be abundant in the fine-mode soil and has it beenproposed as a marker compound for fugitive dust from bio-logically active surface soils (Simoneit et al 2004 Medeiroset al 2006 Rogge et al 2007 Fu et al 2012) A previousstudy found a positive correlation between trehalose and cal-cium (Nishikawa et al 2000) In this study there was nosignificant correlation between trehalose and calcium Be-sides mannose has been reported to be one of the majormonosaccharide components in phytoplankton which origi-nated from marine biological fragments (Tanoue and Handa1987) Mannose was detected in only a few samples and pre-sented with low concentrations in this study

313 Sugar alcohols

Five sugar alcohol compounds including glycerol threitolmannitol arabitol and inositol were detected in PM25These reduced sugars are often reported to be related toplant senescence and decay by microorganisms (Simoneitet al 2004 Tsai et al 2013) they are produced by fungilichens soil biota and algae (Elbert et al 2007 Bauer et al2008) The average concentration of total sugar alcohols was1599 ng mminus3 with a range of 531ndash2540 ng mminus3 which ac-counted for 251plusmn 99 of the total measured saccharidesGlycerol has been widely found in soil biota (Simoneit et al2004) Previous studies suggested that the source of glyc-erol was not specific to biological emissions but the biomasscombustion might increase atmospheric glycerol concentra-

tions (Jia et al 2010 Graham et al 2002 Wang et al 2011)Herein glycerol was the second most abundant saccharidewith an average concentration of 1237 ng mminus3 accountingfor 51 ndash446 (average 226 ) of the total measuredsaccharides

Mannitol and arabitol have been proposed as tracersfor airborne fungal spores (Elbert et al 2007 Bauer etal 2008 Zhang et al 2010 Burshtein et al 2011)Mannitol and arabitol were detected with a concentra-tion range of 00ndash386 ng mminus3 (147 ng mminus3) and 00ndash211 ng mminus3 (58 ng mminus3) respectively The average concen-trations of mannitol and arabitol were comparable to those(average 113 and 91 ng mminus3) reported in the Beijing springaerosols (Liang et al 2013b) but were lower than those (av-erage 219 and 843 ng mminus3) in the Mediterranean summeraerosols (Burshtein et al 2011) and (30 and 24 ng mminus3) atHyytiaumllauml Finland in summer (Yttri et al 2011) Poor corre-lations (r = 038) were found among mannitol and arabitolin this study Nevertheless a positive correlation was foundbetween trehalose and mannitol (r = 079 P lt 005) (Ta-ble S2)

In the previous studies the total measured mannitol hasbeen measured and used for estimating the contribution offungal spores to organic carbon (Elbert et al 2007 Baueret al 2008 Zhang et al 2010) A factor of mannitol perspore (049plusmn 020 pg) was used to calculate the number con-centrations of fungal spores (Liang et al 2013a) and thenthe carbon content of fungal spores could be calculated us-ing a conversion factor of 13 pg C per spore obtained earlieras the average carbon content of spores from nine airbornefungal species with an uncertainty of 20 (Bauer et al2008) The diagnostic tracer ratio of mannitol to OC was cal-culated to be 00377 according to this research (Bauer et al2008 Liang et al 2013a) and then used to estimate the con-tribution of fungal spores to OC The contribution of fungalspores might be underestimated because previous results hadindicated that mannitol and arabitol were mainly associatedwith the coarse PM fraction (Samakeacute et al 2019) The av-erage mannitol concentration was 147plusmn 112 ng mminus3 duringthe observation period The average spore-derived OC wascalculated to be 3903 ng C mminus3 which contributed to 49 of the total OC

Claeys et al (2004) firstly identified two diastereoisomeric2-methyltetrols as oxidation products of isoprene in the Ama-zonian rain forest aerosols Henceforward 2-methyltetrolshave been used as tracers for isoprene-derived SOA (Lianget al 2012 Fu et al 2016 Yan et al 2019) In the previousstudies erythritol was often quantified as a surrogate of 2-methyltetrols (2-methylthreitol and 2-methylerythritol) dueto a lack of standards (Claeys et al 2004 Ding et al 20132016) In this study the concentration range of erythritol was04ndash198 ng mminus3 (average 111 ng mminus3) The values of in-ositol ranged from 00 to 228 ng mminus3 with average valuesof 58 ng mminus3 Moreover the sugar alcohols not only origi-nate from biological emissions but also from BB (Wan and

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12234 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Yu 2007 Jia et al 2010) Different levels of glycerol ara-bitol mannitol erythritol and inositol in fine particles havebeen found during the burning of crop residues and fallenleaves as well as indoor biofuel usage for heating and cook-ing (Graham et al 2002 Burshtein et al 2011 Wang et al2011 Yang et al 2012 Kang et al 2018) In this study onlyinositol correlated with levoglucosan (R = 042) suggestingthat inositol might be linked to biomass combustion sourcesHence the primary source of sugar alcohols associated withfine particles was biogenic aerosols at the observation sites

32 Sources and transport

Since the distinct concentration of the studied compoundswas due to different emission sources arising from differ-ent wind directions the 72 h backward trajectories for thesamples at the Dashu site (2412 N 10011 E) and the spa-tial distribution of the fire spots (8 March to 9 April 2019)were calculated to understand the source of saccharides inaerosol (Fig 4) The analysis of air-mass backward trajec-tories suggested that the air mass over Lincang was almostfrom the westerlies during the sampling periods and could beseparated into two episodes of remote western source above2000 m and local western source below 2000 m as shownin wine red and green lines Specifically 516 of air-massbackward trajectories were generally above 2000 m whereas484 of them were below 2000 m

The average concentrations of saccharide compounds aswell as the contributions for the episodes above and below2000 m are shown in Fig 5 The average concentrations oflevoglucosan and mannosan for the above 2000 m samples(3274 and 356 ng mminus3) were higher than those for the below2000 m samples (2503 and 273 ng mminus3) The anhydrosugarsaccounted for 492 and 369 of total saccharides respec-tively for the above and below 2000 m samples This impliedthat the levoglucosan at the observation site was attributedto both the local BB activities and BB smoke transportedfrom the neighboring regions of Southeast Asia (Myanmar)and the northern Indian subcontinent The southwest windfrom the Indian Ocean prevailed at Lincang all year roundIn spring the southwest wind was often affected by the low-temperature downhill wind blowing from the snow-coveredHengduan Mountains The weather frequently alternated be-tween hot and cold with unstable air pressure and strongwind Therefore the lower air could be diluted by the rel-atively clean cold air over the plateau The upper air mainlycame from the westerlies These results were in agreementwith the fact that residents across Southeast Asia use woodas an energy source to cook and generate heat

For glucose fructose and sucrose concentrations were alittle higher in the below 2000 m samples (averages 335264 and 1062 ng mminus3) than those in the above 2000 m sam-ples (averages 292 229 and 678 ng mminus3) This impliedthat biogenic aerosols (such as ruptured pollen) carrying sug-ars could cover long distance which was supported by pre-

Figure 4 Spatial distribution of the fire spots observed by MODISas well as the corresponding 72 h backward air-mass trajectory clus-ters arriving at 1500 m above ground level (agl) during the sam-pling periods for the collected samples The backward trajectorieswere separated into two episodes of remote western source above2000 m and local western source below 2000 m as shown in winered and green lines

vious studies which have observed long-range atmospherictransport of fine pollen from the Asian continent to the re-mote island Chichijima under the influence of the westerlies(Rousseau et al 2008) Although the pollen grains are usu-ally coarse with various shapes and hard shells this results ina relatively short retention time in the atmosphere Thereforeit could be concluded that in addition to the local pollen theconcentration of sucrose in Lincang was also influenced bythe transport of airborne pollen derived from South Asia

33 Source apportionment of saccharides

Based on the compositional data of saccharides and key rep-resentative markers for difference sources five factors asso-ciated with the emission sources of saccharides were finallyresolved by NMF As shown in Fig 6a factor 1 was char-acterized by high levels of levoglucosan (718 ) and man-nosan (787 ) suggesting the source of BB (Simoneit et al1999 Nolte et al 2001) Factor 2 was characterized by tre-halose (999 ) and mannitol (1000 ) and was enrichedin the other saccharide components ie arabitol (441 )glucose (296 ) erythritol (182 ) glycerol (178 ) lev-oglucosan (147 ) and sucrose (86 ) These saccharidecompounds had all been detected in the suspended soil par-

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12235

Figure 5 Average concentrations and contributions of saccharide compounds for the aerosol samples separated above and below 2000 m

ticles and associated microbiota (eg fungi bacteria and al-gae) (Simoneit et al 2004 Rogge et al 2007) A recentstudy found that leaves were a major source of saccharide-associated microbial taxa in a rural area of France (Samakeacuteet al 2020) Hence this factor was attributed to soil and leafmicrobiota Factor 3 had high levels of glycerol (714 ) anderythritol (582 ) and showed loadings of glucose (128 )and fructose (118 ) Kang et al (2018) reported that glyc-erol and erythritol presented large amounts in winter and au-tumn when vegetation is decomposed This factor was at-tributed to plant senescence and decay by microorganismsFactor 4 exhibited a predominance of sucrose (787 ) andshowed loadings of glucose (172 ) and arabitol (118 )This factor was regarded as the source of airborne pollen be-cause pollen was the reproductive unit of plants and containsthese saccharides and saccharide alcohols as nutritional com-ponents (Bieleski 1995 Miguel et al 2006 Fu et al 2012)Factor 5 characterized by the dominance of fructose (882 )was resolved and it was enriched in glucose (382 ) andarabitol (212 ) thus it could be regarded as the source ofplant detritus

The pie chart in Fig 6b shows the contribution of eachsource to total saccharides BB of factor 1 (340 ) wasfound as the dominant contributor to total saccharides Fac-tors 2ndash5 could all be associated with a biogenic source ac-counting for a total contribution of 660 The sources ofsoil microbiota (factor 2) plant senescence (factor 3) air-borne pollen (factor 4) and plant detritus (factor 5) respec-tively contributed 160 210 237 and 53 to totalsaccharides During the sampling periods daily variations inthe proportion of the five factors are shown in Fig S5 Fac-tor 2 soil microbiota emissions could be associated with soilreclamation and cultivation of farming periods whereas fac-tor 3 (plant senescence) and factor 5 (plant detritus) couldbe associated with the harvesting of vegetation or crops

During the observation period of a month along with theweather warming as sunshine enhanced humans left two ob-vious traces of cultivated soil from 9 to 17 March and from27 March to 8 April and a trace of vegetation or crop harvestfrom 17 to 30 March The stronger pollen discharge occurredin March probably due to the flowering of certain plants TheBB emissions peaked on 9 and 16 March and 1 April weremore prone to be open waste burnings

Since there is still some uncertainty in the factor appor-tionment the proportion of sources are only relative anduncertain It is still difficult to distinguish the contributionsof microorganisms and plants to biogenic aerosols Further-more all the above speculations about farming and harvest-ing periods are based on only 1 month of observation andlong-term observations are needed to obtain more accurateand effective information

4 Conclusion

With the help of various atmospheric saccharides this studypresents the characteristics of BB and various biogenic emis-sions to ambient aerosol in the rural sites of southwest ChinaLevoglucosan was the most dominant species among all thesaccharides with a concentration of 2877 ng mminus3 The ratiosof levoglucosan OC were 19 ndash89 (average 37 )BB contributed to 191 ndash739 of OC (average 413 )The results indicated that domestic biomass fuel burningopen BB events possibly open agricultural waste burningforest fire or sacrificial activity are significant during thespring in this area The total concentrations of five mono-or disaccharides and five sugar alcohols respectively con-tributed 249plusmn 83 and 266plusmn 99 to the total measuredsaccharides Based on the regression analysis these mono-or disaccharides and sugar alcohols were mostly emitted by

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12236 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 6 (a) Factor profile obtained by NMF analysis (b) Source contribution of the five factors to the total saccharides in PM25 samples

direct biogenic volatilization from plant material or surfacesoils rather than BB processes The sampling sites sufferedfrom both local emissions and BB via long-range transportfrom Southeast Asia (Myanmar Bangladesh) and the north-ern Indian subcontinent Five sources of saccharides wereresolved by NMF analysis including BB (34 ) soil mi-crobiota (160 ) plant senescence (210 ) airborne pollen(237 ) and plant detritus (53 ) at rural Lincang in spring

The data indicated that biofuel and open BB activities inrural southwest China and neighboring regions could havea significant impact on ambient aerosol levels In additionto the residential biofuel usage field burning of agriculturalresidues fallen leaves and forest fire were non-negligibleSome new technical measures of biomass resource utilizationare urgently needed to improve the open-burning emissionscenario in rural areas along with strict prohibition policy ofBB Meanwhile the characteristic analysis of saccharides inthe region can serve as a valuable reference for future stud-ies to evaluate temporal variations of biomass combustionand biogenic emission during modeling predictions and pol-icy making

Code availability At present the URLs of the database in our labare still being established The code used in the non-negative ma-trix factorization (NMF) analysis is available upon request from thecorresponding authors

Data availability The meteorological parameters were obtainedfrom the Physical Sciences Laboratory of NOAA (httpspslnoaagovdata last access 8 September 2020 NOAA 2020) 72 h back-trajectories of the aerosol were calculated via the NOAA ARLREADY website (httpreadyarlnoaagovHYSPLITphp last ac-cess 6 September 2020 ARL 2020) Fire pixel counts wereobtained from Moderate Resolution Imaging Spectroradiometer(MODIS) observations on NASA satellites (httpsearthdatanasagov last access 3 September 2020 NASA 2020) The three web-sites have open-access data At present the URLs of the databasein our lab are still being established All data described in this studyare available upon request from the corresponding authors

Supplement The locations of the sampling sites are shown inFig S1 Temporal variations of RH temperature solar irradiationand rainfall are shown in Fig S2 Average concentrations of saccha-ride compounds and the contribution of them for the Datian Dashuand Yakoutian samples are shown in Fig S3 Daily variation on av-erage concentrations of levoglucosan and K+ (panel a) arabitol andmannitol (panel b) PM25 Ca2+ and trehalose (panel c) at the threesites throughout the sampling period are shown in Fig S4 Figure S5shows daily variations on the proportion of the five factors to the to-tal saccharides in PM25 sampled at the three sites during the sam-pling periods Table S1 lists the concentrations of the carbonaceouscomponents and soluble inorganic ions in PM25 during the sam-pling periods of spring 2019 Correlation matrix for the dataset ofthe determined saccharide compounds in PM25 samples is shownin Table S2 The supplement related to this article is available onlineat httpsdoiorg105194acp-21-12227-2021-supplement

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12231

Figure 1 Temporal variations of OC EC PM25 and total sugars at the three sites during the sampling periods

Figure 2 The absolute concentration (bar chart) and the relativecontribution (pie chart) of various saccharide compounds during thesampling periods

ent types of BB compared with the remaining sampling pe-riods (LM= 934plusmn 120) Therefore the high levoglucosanemission from 31 March to 1 April and from 8 to 10 Marchmight be from different open BB events possibly an openagricultural waste burning event or a forest fire while the BBof most sampling days originated from biomass fuel for do-mestic cooking and heating It is worth noting that the peakdays from 31 March to 1 April (LM= 1152plusmn 134) nearedthe Qingming Festival One possibility for BB events is thatpeople burned joss paper as a sacrifice to their ancestors ac-cording to Chinese tradition

Both anhydrosugars and water-soluble potassium (K+)have been widely utilized as source tracers of BB emissions(eg Puxbaum et al 2007 Wang et al 2007 Zhang et al2008 Engling et al 2011) The daily variations of concen-trations of levoglucosan and K+ are shown in Fig S4 theregression analysis of K+ and three anhydrosugars is shownin Fig 3b The K+ concentration was weakly correlated withlevoglucosan mannosan and galactosan with R values of

033 028 and 074 respectively This could be explainedby the additional emissions of K+ from soil and seawaterSince Lincang is far from the coast sea salt could be negligi-ble Because of the inhomogeneity of crustal K+ associatedwith soil types it was difficult to fully account for crustalK+ contributions from soil (Harrison et al 2012 Chenget al 2013) The ratio of levoglucosan to K+ (LK+) wasalso used to track possible sources of BB in the previousstudies The ratios of LK+ strongly depended on BB pro-cesses namely smoldering and flaming Studies suggestedthat relatively high LK+ ratios were obtained from smolder-ing combustion at low temperatures compared with flamingcombustion (Schkolnik et al 2005 Lee et al 2010) Pre-vious results showed the emissions from the combustion ofcrop residuals such as rice straw wheat straw and corn strawexhibited comparable LK+ ratios typically below 10 Theaverage LK+ ratio in this study was 048plusmn 020 which ishigher than the ratio for wheat straw (010plusmn 000) and cornstraw (021plusmn 008) but is lower than the ratio for Asian ricestraw (062plusmn 032) (Cheng et al 2013) In this study higherLK+ ratios were observed during 8ndash10 March (120plusmn 019)compared to those from 31 March to 1 April (040plusmn 013)which suggested that the open fire event from 8 to 10 Marchwas more possibly due to smoldering combustion of residuesat low temperatures

Figure 3c and d show the scatterplots and regression anal-yses of K+ versus PM25 OC and EC and levoglucosanversus PM25 OC and EC respectively Linear regressionof K+ on PM25 OC and EC resulted in R values of 064063 and 062 respectively which were generally higherthan those of levoglucosan on PM25 OC and EC with R

values of 040 054 and 048 respectively This shows thatK+ is more highly correlated with PM25 OC and EC whichcould be explained by either the photo-oxidative decay oflevoglucosan (Hennigan et al 2010) or different types ofBB processes (Schkolnik et al 2005 Lee et al 2010) Evenso the results suggest that BB imposed a great impact on

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12232 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 3 (a) Regression analyses of levoglucosan versus the other two anhydrosugars (b) K+ versus the other three anhydrosugars (c) lev-oglucosan versus PM25 OC and EC and (d) K+ versus PM25 OC and EC

fine aerosols The ratio of levoglucosan to PM25 (LPM25)was also helpful in distinguishing the contributions of dif-ferent levoglucosan sources (Wu et al 2021) The ratio ofLPM25 in this study was 00041ndash00162 (average 00072)indicating that the levoglucosan emission in the areas mightmainly come from wood (001ndash009) and crop straw (0001ndash0008) not excluding incense burning (0001ndash0007) ritualitem burning (0004ndash0086) and meat cooking (0005ndash006)However certainly it was not from coal burning (00001ndash0001) or waste incineration (00022)

An empirical ratio of levoglucosan to OC (82 ) calcu-lated from main types of Chinese cereal straw (rice wheatand corn) based on combustion chamber experiments (Zhanget al 2007) was used to estimate the BB-derived OCThe average mass concentration of BB-derived OC was35344 ng mminus3 whilst the contributions of BB to OC was413 with a large range of 191 ndash739 The contri-butions were higher than those previously reported such as65 ndash11 in Hong Kong (Sang et al 2011) 18 ndash38 inBeijing (Zhang et al 2008) 189 ndash454 over southeast-ern Tibetan Plateau (Sang et al 2013) and 264 ndash302 in Xirsquoan (Zhang et al 2014) The large range of 191 ndash739 revealed that the daily contribution of BB variedgreatly suggesting that an open BB event or forest fire hap-pened occasionally The contribution apportionment of pri-mary BB might be underestimated due to the degradationof levoglucosan during atmospheric aging of BB-influencedair mass after long-range transport (Hennigan et al 2010

Mochida et al 2010 Lai et al 2014) Moreover Wu etal (2021) have reported that the total levoglucosan emissionexhibited a clear decreasing trend in China However it wasnoteworthy that the average concentration of levoglucosan(2877 ng mminus3) and the BB contributions to OC (413 )at Lincang mountain site were both higher than the valuesof 1918 ng mminus3 and 284 at Tengchong mountain site in2004 in spring (Sang et al 2013) This result suggests thatthere is no significant reduction in BB emissions in southwestYunnan Province

312 Mono- or disaccharides

The total concentrations of five mono- or disaccharidesincluding glucose fructose mannose sucrose and tre-halose were in the range of 252ndash3737 ng mminus3 (average1589 ng mminus3) which contributed 249plusmn 83 of the totalmeasured saccharides The average values of glucose fruc-tose mannose sucrose and trehalose were 312 246 27864 and 138 ng mminus3 respectively Sucrose was the domi-nant species in the mono- or disaccharides group The resultsagreed with those of previous studies (Yttri et al 2007 Jia etal 2010 Fu et al 2012) which found that sucrose was oneof the dominant species in fine aerosols in spring Rupturedpollen may be an important source of sucrose especially inthe spring blossom season (Yttri et al 2007 Fu et al 2012Miyazaki et al 2012) In spring and early summer farmlandtilling after wheat harvest causes an enhanced exposure of

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12233

soil containing wheat roots to the air which is beneficial forthe release of sucrose stored in roots (Medeiros et al 2006)thus resulting in a sharply increased sucrose concentration

It was reported that sugars such as glucose sucrose andfructose could be emitted from developing leaves (Grahamet al 2003) Glucose could be released from both soils andplant material (eg pollen fruits and their fragments) (Gra-ham et al 2003 Simoneit et al 2004 Fu et al 2012) Glu-cose and sucrose were rich in biologically active surface soils(Rogge et al 2007) In this study positive correlations werefound between sucrose and glucose (R = 052) (Table S2)suggesting a similar origin of glucose and sucrose in thisstudy Glucose and fructose have also been identified as a mi-nor product of cellulose pyrolysis because they were foundto be enriched in BB emissions (Nolte et al 2001) andthey correlated well with K+ (Graham et al 2002) and lev-oglucosan (Kang et al 2018) Herein no significant correla-tion was found between K+ levoglucosan and these mono-or disaccharides Therefore the detected glucose fructoseand sucrose might mostly be emitted by direct volatilizationfrom plant material or surface soils rather than as productsof polysaccharide breakdown during BB processes The highabundance of sucrose as well as glucose and fructose wasresponsible for biogenic aerosols associated with developingleaves and flowers as well as surface soil suspension

Trehalose as a stress protectant of various microorganismsand plants (Medeiros et al 2006 Jia and Fraser 2011) wasfound to be abundant in the fine-mode soil and has it beenproposed as a marker compound for fugitive dust from bio-logically active surface soils (Simoneit et al 2004 Medeiroset al 2006 Rogge et al 2007 Fu et al 2012) A previousstudy found a positive correlation between trehalose and cal-cium (Nishikawa et al 2000) In this study there was nosignificant correlation between trehalose and calcium Be-sides mannose has been reported to be one of the majormonosaccharide components in phytoplankton which origi-nated from marine biological fragments (Tanoue and Handa1987) Mannose was detected in only a few samples and pre-sented with low concentrations in this study

313 Sugar alcohols

Five sugar alcohol compounds including glycerol threitolmannitol arabitol and inositol were detected in PM25These reduced sugars are often reported to be related toplant senescence and decay by microorganisms (Simoneitet al 2004 Tsai et al 2013) they are produced by fungilichens soil biota and algae (Elbert et al 2007 Bauer et al2008) The average concentration of total sugar alcohols was1599 ng mminus3 with a range of 531ndash2540 ng mminus3 which ac-counted for 251plusmn 99 of the total measured saccharidesGlycerol has been widely found in soil biota (Simoneit et al2004) Previous studies suggested that the source of glyc-erol was not specific to biological emissions but the biomasscombustion might increase atmospheric glycerol concentra-

tions (Jia et al 2010 Graham et al 2002 Wang et al 2011)Herein glycerol was the second most abundant saccharidewith an average concentration of 1237 ng mminus3 accountingfor 51 ndash446 (average 226 ) of the total measuredsaccharides

Mannitol and arabitol have been proposed as tracersfor airborne fungal spores (Elbert et al 2007 Bauer etal 2008 Zhang et al 2010 Burshtein et al 2011)Mannitol and arabitol were detected with a concentra-tion range of 00ndash386 ng mminus3 (147 ng mminus3) and 00ndash211 ng mminus3 (58 ng mminus3) respectively The average concen-trations of mannitol and arabitol were comparable to those(average 113 and 91 ng mminus3) reported in the Beijing springaerosols (Liang et al 2013b) but were lower than those (av-erage 219 and 843 ng mminus3) in the Mediterranean summeraerosols (Burshtein et al 2011) and (30 and 24 ng mminus3) atHyytiaumllauml Finland in summer (Yttri et al 2011) Poor corre-lations (r = 038) were found among mannitol and arabitolin this study Nevertheless a positive correlation was foundbetween trehalose and mannitol (r = 079 P lt 005) (Ta-ble S2)

In the previous studies the total measured mannitol hasbeen measured and used for estimating the contribution offungal spores to organic carbon (Elbert et al 2007 Baueret al 2008 Zhang et al 2010) A factor of mannitol perspore (049plusmn 020 pg) was used to calculate the number con-centrations of fungal spores (Liang et al 2013a) and thenthe carbon content of fungal spores could be calculated us-ing a conversion factor of 13 pg C per spore obtained earlieras the average carbon content of spores from nine airbornefungal species with an uncertainty of 20 (Bauer et al2008) The diagnostic tracer ratio of mannitol to OC was cal-culated to be 00377 according to this research (Bauer et al2008 Liang et al 2013a) and then used to estimate the con-tribution of fungal spores to OC The contribution of fungalspores might be underestimated because previous results hadindicated that mannitol and arabitol were mainly associatedwith the coarse PM fraction (Samakeacute et al 2019) The av-erage mannitol concentration was 147plusmn 112 ng mminus3 duringthe observation period The average spore-derived OC wascalculated to be 3903 ng C mminus3 which contributed to 49 of the total OC

Claeys et al (2004) firstly identified two diastereoisomeric2-methyltetrols as oxidation products of isoprene in the Ama-zonian rain forest aerosols Henceforward 2-methyltetrolshave been used as tracers for isoprene-derived SOA (Lianget al 2012 Fu et al 2016 Yan et al 2019) In the previousstudies erythritol was often quantified as a surrogate of 2-methyltetrols (2-methylthreitol and 2-methylerythritol) dueto a lack of standards (Claeys et al 2004 Ding et al 20132016) In this study the concentration range of erythritol was04ndash198 ng mminus3 (average 111 ng mminus3) The values of in-ositol ranged from 00 to 228 ng mminus3 with average valuesof 58 ng mminus3 Moreover the sugar alcohols not only origi-nate from biological emissions but also from BB (Wan and

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12234 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Yu 2007 Jia et al 2010) Different levels of glycerol ara-bitol mannitol erythritol and inositol in fine particles havebeen found during the burning of crop residues and fallenleaves as well as indoor biofuel usage for heating and cook-ing (Graham et al 2002 Burshtein et al 2011 Wang et al2011 Yang et al 2012 Kang et al 2018) In this study onlyinositol correlated with levoglucosan (R = 042) suggestingthat inositol might be linked to biomass combustion sourcesHence the primary source of sugar alcohols associated withfine particles was biogenic aerosols at the observation sites

32 Sources and transport

Since the distinct concentration of the studied compoundswas due to different emission sources arising from differ-ent wind directions the 72 h backward trajectories for thesamples at the Dashu site (2412 N 10011 E) and the spa-tial distribution of the fire spots (8 March to 9 April 2019)were calculated to understand the source of saccharides inaerosol (Fig 4) The analysis of air-mass backward trajec-tories suggested that the air mass over Lincang was almostfrom the westerlies during the sampling periods and could beseparated into two episodes of remote western source above2000 m and local western source below 2000 m as shownin wine red and green lines Specifically 516 of air-massbackward trajectories were generally above 2000 m whereas484 of them were below 2000 m

The average concentrations of saccharide compounds aswell as the contributions for the episodes above and below2000 m are shown in Fig 5 The average concentrations oflevoglucosan and mannosan for the above 2000 m samples(3274 and 356 ng mminus3) were higher than those for the below2000 m samples (2503 and 273 ng mminus3) The anhydrosugarsaccounted for 492 and 369 of total saccharides respec-tively for the above and below 2000 m samples This impliedthat the levoglucosan at the observation site was attributedto both the local BB activities and BB smoke transportedfrom the neighboring regions of Southeast Asia (Myanmar)and the northern Indian subcontinent The southwest windfrom the Indian Ocean prevailed at Lincang all year roundIn spring the southwest wind was often affected by the low-temperature downhill wind blowing from the snow-coveredHengduan Mountains The weather frequently alternated be-tween hot and cold with unstable air pressure and strongwind Therefore the lower air could be diluted by the rel-atively clean cold air over the plateau The upper air mainlycame from the westerlies These results were in agreementwith the fact that residents across Southeast Asia use woodas an energy source to cook and generate heat

For glucose fructose and sucrose concentrations were alittle higher in the below 2000 m samples (averages 335264 and 1062 ng mminus3) than those in the above 2000 m sam-ples (averages 292 229 and 678 ng mminus3) This impliedthat biogenic aerosols (such as ruptured pollen) carrying sug-ars could cover long distance which was supported by pre-

Figure 4 Spatial distribution of the fire spots observed by MODISas well as the corresponding 72 h backward air-mass trajectory clus-ters arriving at 1500 m above ground level (agl) during the sam-pling periods for the collected samples The backward trajectorieswere separated into two episodes of remote western source above2000 m and local western source below 2000 m as shown in winered and green lines

vious studies which have observed long-range atmospherictransport of fine pollen from the Asian continent to the re-mote island Chichijima under the influence of the westerlies(Rousseau et al 2008) Although the pollen grains are usu-ally coarse with various shapes and hard shells this results ina relatively short retention time in the atmosphere Thereforeit could be concluded that in addition to the local pollen theconcentration of sucrose in Lincang was also influenced bythe transport of airborne pollen derived from South Asia

33 Source apportionment of saccharides

Based on the compositional data of saccharides and key rep-resentative markers for difference sources five factors asso-ciated with the emission sources of saccharides were finallyresolved by NMF As shown in Fig 6a factor 1 was char-acterized by high levels of levoglucosan (718 ) and man-nosan (787 ) suggesting the source of BB (Simoneit et al1999 Nolte et al 2001) Factor 2 was characterized by tre-halose (999 ) and mannitol (1000 ) and was enrichedin the other saccharide components ie arabitol (441 )glucose (296 ) erythritol (182 ) glycerol (178 ) lev-oglucosan (147 ) and sucrose (86 ) These saccharidecompounds had all been detected in the suspended soil par-

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12235

Figure 5 Average concentrations and contributions of saccharide compounds for the aerosol samples separated above and below 2000 m

ticles and associated microbiota (eg fungi bacteria and al-gae) (Simoneit et al 2004 Rogge et al 2007) A recentstudy found that leaves were a major source of saccharide-associated microbial taxa in a rural area of France (Samakeacuteet al 2020) Hence this factor was attributed to soil and leafmicrobiota Factor 3 had high levels of glycerol (714 ) anderythritol (582 ) and showed loadings of glucose (128 )and fructose (118 ) Kang et al (2018) reported that glyc-erol and erythritol presented large amounts in winter and au-tumn when vegetation is decomposed This factor was at-tributed to plant senescence and decay by microorganismsFactor 4 exhibited a predominance of sucrose (787 ) andshowed loadings of glucose (172 ) and arabitol (118 )This factor was regarded as the source of airborne pollen be-cause pollen was the reproductive unit of plants and containsthese saccharides and saccharide alcohols as nutritional com-ponents (Bieleski 1995 Miguel et al 2006 Fu et al 2012)Factor 5 characterized by the dominance of fructose (882 )was resolved and it was enriched in glucose (382 ) andarabitol (212 ) thus it could be regarded as the source ofplant detritus

The pie chart in Fig 6b shows the contribution of eachsource to total saccharides BB of factor 1 (340 ) wasfound as the dominant contributor to total saccharides Fac-tors 2ndash5 could all be associated with a biogenic source ac-counting for a total contribution of 660 The sources ofsoil microbiota (factor 2) plant senescence (factor 3) air-borne pollen (factor 4) and plant detritus (factor 5) respec-tively contributed 160 210 237 and 53 to totalsaccharides During the sampling periods daily variations inthe proportion of the five factors are shown in Fig S5 Fac-tor 2 soil microbiota emissions could be associated with soilreclamation and cultivation of farming periods whereas fac-tor 3 (plant senescence) and factor 5 (plant detritus) couldbe associated with the harvesting of vegetation or crops

During the observation period of a month along with theweather warming as sunshine enhanced humans left two ob-vious traces of cultivated soil from 9 to 17 March and from27 March to 8 April and a trace of vegetation or crop harvestfrom 17 to 30 March The stronger pollen discharge occurredin March probably due to the flowering of certain plants TheBB emissions peaked on 9 and 16 March and 1 April weremore prone to be open waste burnings

Since there is still some uncertainty in the factor appor-tionment the proportion of sources are only relative anduncertain It is still difficult to distinguish the contributionsof microorganisms and plants to biogenic aerosols Further-more all the above speculations about farming and harvest-ing periods are based on only 1 month of observation andlong-term observations are needed to obtain more accurateand effective information

4 Conclusion

With the help of various atmospheric saccharides this studypresents the characteristics of BB and various biogenic emis-sions to ambient aerosol in the rural sites of southwest ChinaLevoglucosan was the most dominant species among all thesaccharides with a concentration of 2877 ng mminus3 The ratiosof levoglucosan OC were 19 ndash89 (average 37 )BB contributed to 191 ndash739 of OC (average 413 )The results indicated that domestic biomass fuel burningopen BB events possibly open agricultural waste burningforest fire or sacrificial activity are significant during thespring in this area The total concentrations of five mono-or disaccharides and five sugar alcohols respectively con-tributed 249plusmn 83 and 266plusmn 99 to the total measuredsaccharides Based on the regression analysis these mono-or disaccharides and sugar alcohols were mostly emitted by

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12236 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 6 (a) Factor profile obtained by NMF analysis (b) Source contribution of the five factors to the total saccharides in PM25 samples

direct biogenic volatilization from plant material or surfacesoils rather than BB processes The sampling sites sufferedfrom both local emissions and BB via long-range transportfrom Southeast Asia (Myanmar Bangladesh) and the north-ern Indian subcontinent Five sources of saccharides wereresolved by NMF analysis including BB (34 ) soil mi-crobiota (160 ) plant senescence (210 ) airborne pollen(237 ) and plant detritus (53 ) at rural Lincang in spring

The data indicated that biofuel and open BB activities inrural southwest China and neighboring regions could havea significant impact on ambient aerosol levels In additionto the residential biofuel usage field burning of agriculturalresidues fallen leaves and forest fire were non-negligibleSome new technical measures of biomass resource utilizationare urgently needed to improve the open-burning emissionscenario in rural areas along with strict prohibition policy ofBB Meanwhile the characteristic analysis of saccharides inthe region can serve as a valuable reference for future stud-ies to evaluate temporal variations of biomass combustionand biogenic emission during modeling predictions and pol-icy making

Code availability At present the URLs of the database in our labare still being established The code used in the non-negative ma-trix factorization (NMF) analysis is available upon request from thecorresponding authors

Data availability The meteorological parameters were obtainedfrom the Physical Sciences Laboratory of NOAA (httpspslnoaagovdata last access 8 September 2020 NOAA 2020) 72 h back-trajectories of the aerosol were calculated via the NOAA ARLREADY website (httpreadyarlnoaagovHYSPLITphp last ac-cess 6 September 2020 ARL 2020) Fire pixel counts wereobtained from Moderate Resolution Imaging Spectroradiometer(MODIS) observations on NASA satellites (httpsearthdatanasagov last access 3 September 2020 NASA 2020) The three web-sites have open-access data At present the URLs of the databasein our lab are still being established All data described in this studyare available upon request from the corresponding authors

Supplement The locations of the sampling sites are shown inFig S1 Temporal variations of RH temperature solar irradiationand rainfall are shown in Fig S2 Average concentrations of saccha-ride compounds and the contribution of them for the Datian Dashuand Yakoutian samples are shown in Fig S3 Daily variation on av-erage concentrations of levoglucosan and K+ (panel a) arabitol andmannitol (panel b) PM25 Ca2+ and trehalose (panel c) at the threesites throughout the sampling period are shown in Fig S4 Figure S5shows daily variations on the proportion of the five factors to the to-tal saccharides in PM25 sampled at the three sites during the sam-pling periods Table S1 lists the concentrations of the carbonaceouscomponents and soluble inorganic ions in PM25 during the sam-pling periods of spring 2019 Correlation matrix for the dataset ofthe determined saccharide compounds in PM25 samples is shownin Table S2 The supplement related to this article is available onlineat httpsdoiorg105194acp-21-12227-2021-supplement

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

12232 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 3 (a) Regression analyses of levoglucosan versus the other two anhydrosugars (b) K+ versus the other three anhydrosugars (c) lev-oglucosan versus PM25 OC and EC and (d) K+ versus PM25 OC and EC

fine aerosols The ratio of levoglucosan to PM25 (LPM25)was also helpful in distinguishing the contributions of dif-ferent levoglucosan sources (Wu et al 2021) The ratio ofLPM25 in this study was 00041ndash00162 (average 00072)indicating that the levoglucosan emission in the areas mightmainly come from wood (001ndash009) and crop straw (0001ndash0008) not excluding incense burning (0001ndash0007) ritualitem burning (0004ndash0086) and meat cooking (0005ndash006)However certainly it was not from coal burning (00001ndash0001) or waste incineration (00022)

An empirical ratio of levoglucosan to OC (82 ) calcu-lated from main types of Chinese cereal straw (rice wheatand corn) based on combustion chamber experiments (Zhanget al 2007) was used to estimate the BB-derived OCThe average mass concentration of BB-derived OC was35344 ng mminus3 whilst the contributions of BB to OC was413 with a large range of 191 ndash739 The contri-butions were higher than those previously reported such as65 ndash11 in Hong Kong (Sang et al 2011) 18 ndash38 inBeijing (Zhang et al 2008) 189 ndash454 over southeast-ern Tibetan Plateau (Sang et al 2013) and 264 ndash302 in Xirsquoan (Zhang et al 2014) The large range of 191 ndash739 revealed that the daily contribution of BB variedgreatly suggesting that an open BB event or forest fire hap-pened occasionally The contribution apportionment of pri-mary BB might be underestimated due to the degradationof levoglucosan during atmospheric aging of BB-influencedair mass after long-range transport (Hennigan et al 2010

Mochida et al 2010 Lai et al 2014) Moreover Wu etal (2021) have reported that the total levoglucosan emissionexhibited a clear decreasing trend in China However it wasnoteworthy that the average concentration of levoglucosan(2877 ng mminus3) and the BB contributions to OC (413 )at Lincang mountain site were both higher than the valuesof 1918 ng mminus3 and 284 at Tengchong mountain site in2004 in spring (Sang et al 2013) This result suggests thatthere is no significant reduction in BB emissions in southwestYunnan Province

312 Mono- or disaccharides

The total concentrations of five mono- or disaccharidesincluding glucose fructose mannose sucrose and tre-halose were in the range of 252ndash3737 ng mminus3 (average1589 ng mminus3) which contributed 249plusmn 83 of the totalmeasured saccharides The average values of glucose fruc-tose mannose sucrose and trehalose were 312 246 27864 and 138 ng mminus3 respectively Sucrose was the domi-nant species in the mono- or disaccharides group The resultsagreed with those of previous studies (Yttri et al 2007 Jia etal 2010 Fu et al 2012) which found that sucrose was oneof the dominant species in fine aerosols in spring Rupturedpollen may be an important source of sucrose especially inthe spring blossom season (Yttri et al 2007 Fu et al 2012Miyazaki et al 2012) In spring and early summer farmlandtilling after wheat harvest causes an enhanced exposure of

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12233

soil containing wheat roots to the air which is beneficial forthe release of sucrose stored in roots (Medeiros et al 2006)thus resulting in a sharply increased sucrose concentration

It was reported that sugars such as glucose sucrose andfructose could be emitted from developing leaves (Grahamet al 2003) Glucose could be released from both soils andplant material (eg pollen fruits and their fragments) (Gra-ham et al 2003 Simoneit et al 2004 Fu et al 2012) Glu-cose and sucrose were rich in biologically active surface soils(Rogge et al 2007) In this study positive correlations werefound between sucrose and glucose (R = 052) (Table S2)suggesting a similar origin of glucose and sucrose in thisstudy Glucose and fructose have also been identified as a mi-nor product of cellulose pyrolysis because they were foundto be enriched in BB emissions (Nolte et al 2001) andthey correlated well with K+ (Graham et al 2002) and lev-oglucosan (Kang et al 2018) Herein no significant correla-tion was found between K+ levoglucosan and these mono-or disaccharides Therefore the detected glucose fructoseand sucrose might mostly be emitted by direct volatilizationfrom plant material or surface soils rather than as productsof polysaccharide breakdown during BB processes The highabundance of sucrose as well as glucose and fructose wasresponsible for biogenic aerosols associated with developingleaves and flowers as well as surface soil suspension

Trehalose as a stress protectant of various microorganismsand plants (Medeiros et al 2006 Jia and Fraser 2011) wasfound to be abundant in the fine-mode soil and has it beenproposed as a marker compound for fugitive dust from bio-logically active surface soils (Simoneit et al 2004 Medeiroset al 2006 Rogge et al 2007 Fu et al 2012) A previousstudy found a positive correlation between trehalose and cal-cium (Nishikawa et al 2000) In this study there was nosignificant correlation between trehalose and calcium Be-sides mannose has been reported to be one of the majormonosaccharide components in phytoplankton which origi-nated from marine biological fragments (Tanoue and Handa1987) Mannose was detected in only a few samples and pre-sented with low concentrations in this study

313 Sugar alcohols

Five sugar alcohol compounds including glycerol threitolmannitol arabitol and inositol were detected in PM25These reduced sugars are often reported to be related toplant senescence and decay by microorganisms (Simoneitet al 2004 Tsai et al 2013) they are produced by fungilichens soil biota and algae (Elbert et al 2007 Bauer et al2008) The average concentration of total sugar alcohols was1599 ng mminus3 with a range of 531ndash2540 ng mminus3 which ac-counted for 251plusmn 99 of the total measured saccharidesGlycerol has been widely found in soil biota (Simoneit et al2004) Previous studies suggested that the source of glyc-erol was not specific to biological emissions but the biomasscombustion might increase atmospheric glycerol concentra-

tions (Jia et al 2010 Graham et al 2002 Wang et al 2011)Herein glycerol was the second most abundant saccharidewith an average concentration of 1237 ng mminus3 accountingfor 51 ndash446 (average 226 ) of the total measuredsaccharides

Mannitol and arabitol have been proposed as tracersfor airborne fungal spores (Elbert et al 2007 Bauer etal 2008 Zhang et al 2010 Burshtein et al 2011)Mannitol and arabitol were detected with a concentra-tion range of 00ndash386 ng mminus3 (147 ng mminus3) and 00ndash211 ng mminus3 (58 ng mminus3) respectively The average concen-trations of mannitol and arabitol were comparable to those(average 113 and 91 ng mminus3) reported in the Beijing springaerosols (Liang et al 2013b) but were lower than those (av-erage 219 and 843 ng mminus3) in the Mediterranean summeraerosols (Burshtein et al 2011) and (30 and 24 ng mminus3) atHyytiaumllauml Finland in summer (Yttri et al 2011) Poor corre-lations (r = 038) were found among mannitol and arabitolin this study Nevertheless a positive correlation was foundbetween trehalose and mannitol (r = 079 P lt 005) (Ta-ble S2)

In the previous studies the total measured mannitol hasbeen measured and used for estimating the contribution offungal spores to organic carbon (Elbert et al 2007 Baueret al 2008 Zhang et al 2010) A factor of mannitol perspore (049plusmn 020 pg) was used to calculate the number con-centrations of fungal spores (Liang et al 2013a) and thenthe carbon content of fungal spores could be calculated us-ing a conversion factor of 13 pg C per spore obtained earlieras the average carbon content of spores from nine airbornefungal species with an uncertainty of 20 (Bauer et al2008) The diagnostic tracer ratio of mannitol to OC was cal-culated to be 00377 according to this research (Bauer et al2008 Liang et al 2013a) and then used to estimate the con-tribution of fungal spores to OC The contribution of fungalspores might be underestimated because previous results hadindicated that mannitol and arabitol were mainly associatedwith the coarse PM fraction (Samakeacute et al 2019) The av-erage mannitol concentration was 147plusmn 112 ng mminus3 duringthe observation period The average spore-derived OC wascalculated to be 3903 ng C mminus3 which contributed to 49 of the total OC

Claeys et al (2004) firstly identified two diastereoisomeric2-methyltetrols as oxidation products of isoprene in the Ama-zonian rain forest aerosols Henceforward 2-methyltetrolshave been used as tracers for isoprene-derived SOA (Lianget al 2012 Fu et al 2016 Yan et al 2019) In the previousstudies erythritol was often quantified as a surrogate of 2-methyltetrols (2-methylthreitol and 2-methylerythritol) dueto a lack of standards (Claeys et al 2004 Ding et al 20132016) In this study the concentration range of erythritol was04ndash198 ng mminus3 (average 111 ng mminus3) The values of in-ositol ranged from 00 to 228 ng mminus3 with average valuesof 58 ng mminus3 Moreover the sugar alcohols not only origi-nate from biological emissions but also from BB (Wan and

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12234 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Yu 2007 Jia et al 2010) Different levels of glycerol ara-bitol mannitol erythritol and inositol in fine particles havebeen found during the burning of crop residues and fallenleaves as well as indoor biofuel usage for heating and cook-ing (Graham et al 2002 Burshtein et al 2011 Wang et al2011 Yang et al 2012 Kang et al 2018) In this study onlyinositol correlated with levoglucosan (R = 042) suggestingthat inositol might be linked to biomass combustion sourcesHence the primary source of sugar alcohols associated withfine particles was biogenic aerosols at the observation sites

32 Sources and transport

Since the distinct concentration of the studied compoundswas due to different emission sources arising from differ-ent wind directions the 72 h backward trajectories for thesamples at the Dashu site (2412 N 10011 E) and the spa-tial distribution of the fire spots (8 March to 9 April 2019)were calculated to understand the source of saccharides inaerosol (Fig 4) The analysis of air-mass backward trajec-tories suggested that the air mass over Lincang was almostfrom the westerlies during the sampling periods and could beseparated into two episodes of remote western source above2000 m and local western source below 2000 m as shownin wine red and green lines Specifically 516 of air-massbackward trajectories were generally above 2000 m whereas484 of them were below 2000 m

The average concentrations of saccharide compounds aswell as the contributions for the episodes above and below2000 m are shown in Fig 5 The average concentrations oflevoglucosan and mannosan for the above 2000 m samples(3274 and 356 ng mminus3) were higher than those for the below2000 m samples (2503 and 273 ng mminus3) The anhydrosugarsaccounted for 492 and 369 of total saccharides respec-tively for the above and below 2000 m samples This impliedthat the levoglucosan at the observation site was attributedto both the local BB activities and BB smoke transportedfrom the neighboring regions of Southeast Asia (Myanmar)and the northern Indian subcontinent The southwest windfrom the Indian Ocean prevailed at Lincang all year roundIn spring the southwest wind was often affected by the low-temperature downhill wind blowing from the snow-coveredHengduan Mountains The weather frequently alternated be-tween hot and cold with unstable air pressure and strongwind Therefore the lower air could be diluted by the rel-atively clean cold air over the plateau The upper air mainlycame from the westerlies These results were in agreementwith the fact that residents across Southeast Asia use woodas an energy source to cook and generate heat

For glucose fructose and sucrose concentrations were alittle higher in the below 2000 m samples (averages 335264 and 1062 ng mminus3) than those in the above 2000 m sam-ples (averages 292 229 and 678 ng mminus3) This impliedthat biogenic aerosols (such as ruptured pollen) carrying sug-ars could cover long distance which was supported by pre-

Figure 4 Spatial distribution of the fire spots observed by MODISas well as the corresponding 72 h backward air-mass trajectory clus-ters arriving at 1500 m above ground level (agl) during the sam-pling periods for the collected samples The backward trajectorieswere separated into two episodes of remote western source above2000 m and local western source below 2000 m as shown in winered and green lines

vious studies which have observed long-range atmospherictransport of fine pollen from the Asian continent to the re-mote island Chichijima under the influence of the westerlies(Rousseau et al 2008) Although the pollen grains are usu-ally coarse with various shapes and hard shells this results ina relatively short retention time in the atmosphere Thereforeit could be concluded that in addition to the local pollen theconcentration of sucrose in Lincang was also influenced bythe transport of airborne pollen derived from South Asia

33 Source apportionment of saccharides

Based on the compositional data of saccharides and key rep-resentative markers for difference sources five factors asso-ciated with the emission sources of saccharides were finallyresolved by NMF As shown in Fig 6a factor 1 was char-acterized by high levels of levoglucosan (718 ) and man-nosan (787 ) suggesting the source of BB (Simoneit et al1999 Nolte et al 2001) Factor 2 was characterized by tre-halose (999 ) and mannitol (1000 ) and was enrichedin the other saccharide components ie arabitol (441 )glucose (296 ) erythritol (182 ) glycerol (178 ) lev-oglucosan (147 ) and sucrose (86 ) These saccharidecompounds had all been detected in the suspended soil par-

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12235

Figure 5 Average concentrations and contributions of saccharide compounds for the aerosol samples separated above and below 2000 m

ticles and associated microbiota (eg fungi bacteria and al-gae) (Simoneit et al 2004 Rogge et al 2007) A recentstudy found that leaves were a major source of saccharide-associated microbial taxa in a rural area of France (Samakeacuteet al 2020) Hence this factor was attributed to soil and leafmicrobiota Factor 3 had high levels of glycerol (714 ) anderythritol (582 ) and showed loadings of glucose (128 )and fructose (118 ) Kang et al (2018) reported that glyc-erol and erythritol presented large amounts in winter and au-tumn when vegetation is decomposed This factor was at-tributed to plant senescence and decay by microorganismsFactor 4 exhibited a predominance of sucrose (787 ) andshowed loadings of glucose (172 ) and arabitol (118 )This factor was regarded as the source of airborne pollen be-cause pollen was the reproductive unit of plants and containsthese saccharides and saccharide alcohols as nutritional com-ponents (Bieleski 1995 Miguel et al 2006 Fu et al 2012)Factor 5 characterized by the dominance of fructose (882 )was resolved and it was enriched in glucose (382 ) andarabitol (212 ) thus it could be regarded as the source ofplant detritus

The pie chart in Fig 6b shows the contribution of eachsource to total saccharides BB of factor 1 (340 ) wasfound as the dominant contributor to total saccharides Fac-tors 2ndash5 could all be associated with a biogenic source ac-counting for a total contribution of 660 The sources ofsoil microbiota (factor 2) plant senescence (factor 3) air-borne pollen (factor 4) and plant detritus (factor 5) respec-tively contributed 160 210 237 and 53 to totalsaccharides During the sampling periods daily variations inthe proportion of the five factors are shown in Fig S5 Fac-tor 2 soil microbiota emissions could be associated with soilreclamation and cultivation of farming periods whereas fac-tor 3 (plant senescence) and factor 5 (plant detritus) couldbe associated with the harvesting of vegetation or crops

During the observation period of a month along with theweather warming as sunshine enhanced humans left two ob-vious traces of cultivated soil from 9 to 17 March and from27 March to 8 April and a trace of vegetation or crop harvestfrom 17 to 30 March The stronger pollen discharge occurredin March probably due to the flowering of certain plants TheBB emissions peaked on 9 and 16 March and 1 April weremore prone to be open waste burnings

Since there is still some uncertainty in the factor appor-tionment the proportion of sources are only relative anduncertain It is still difficult to distinguish the contributionsof microorganisms and plants to biogenic aerosols Further-more all the above speculations about farming and harvest-ing periods are based on only 1 month of observation andlong-term observations are needed to obtain more accurateand effective information

4 Conclusion

With the help of various atmospheric saccharides this studypresents the characteristics of BB and various biogenic emis-sions to ambient aerosol in the rural sites of southwest ChinaLevoglucosan was the most dominant species among all thesaccharides with a concentration of 2877 ng mminus3 The ratiosof levoglucosan OC were 19 ndash89 (average 37 )BB contributed to 191 ndash739 of OC (average 413 )The results indicated that domestic biomass fuel burningopen BB events possibly open agricultural waste burningforest fire or sacrificial activity are significant during thespring in this area The total concentrations of five mono-or disaccharides and five sugar alcohols respectively con-tributed 249plusmn 83 and 266plusmn 99 to the total measuredsaccharides Based on the regression analysis these mono-or disaccharides and sugar alcohols were mostly emitted by

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12236 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 6 (a) Factor profile obtained by NMF analysis (b) Source contribution of the five factors to the total saccharides in PM25 samples

direct biogenic volatilization from plant material or surfacesoils rather than BB processes The sampling sites sufferedfrom both local emissions and BB via long-range transportfrom Southeast Asia (Myanmar Bangladesh) and the north-ern Indian subcontinent Five sources of saccharides wereresolved by NMF analysis including BB (34 ) soil mi-crobiota (160 ) plant senescence (210 ) airborne pollen(237 ) and plant detritus (53 ) at rural Lincang in spring

The data indicated that biofuel and open BB activities inrural southwest China and neighboring regions could havea significant impact on ambient aerosol levels In additionto the residential biofuel usage field burning of agriculturalresidues fallen leaves and forest fire were non-negligibleSome new technical measures of biomass resource utilizationare urgently needed to improve the open-burning emissionscenario in rural areas along with strict prohibition policy ofBB Meanwhile the characteristic analysis of saccharides inthe region can serve as a valuable reference for future stud-ies to evaluate temporal variations of biomass combustionand biogenic emission during modeling predictions and pol-icy making

Code availability At present the URLs of the database in our labare still being established The code used in the non-negative ma-trix factorization (NMF) analysis is available upon request from thecorresponding authors

Data availability The meteorological parameters were obtainedfrom the Physical Sciences Laboratory of NOAA (httpspslnoaagovdata last access 8 September 2020 NOAA 2020) 72 h back-trajectories of the aerosol were calculated via the NOAA ARLREADY website (httpreadyarlnoaagovHYSPLITphp last ac-cess 6 September 2020 ARL 2020) Fire pixel counts wereobtained from Moderate Resolution Imaging Spectroradiometer(MODIS) observations on NASA satellites (httpsearthdatanasagov last access 3 September 2020 NASA 2020) The three web-sites have open-access data At present the URLs of the databasein our lab are still being established All data described in this studyare available upon request from the corresponding authors

Supplement The locations of the sampling sites are shown inFig S1 Temporal variations of RH temperature solar irradiationand rainfall are shown in Fig S2 Average concentrations of saccha-ride compounds and the contribution of them for the Datian Dashuand Yakoutian samples are shown in Fig S3 Daily variation on av-erage concentrations of levoglucosan and K+ (panel a) arabitol andmannitol (panel b) PM25 Ca2+ and trehalose (panel c) at the threesites throughout the sampling period are shown in Fig S4 Figure S5shows daily variations on the proportion of the five factors to the to-tal saccharides in PM25 sampled at the three sites during the sam-pling periods Table S1 lists the concentrations of the carbonaceouscomponents and soluble inorganic ions in PM25 during the sam-pling periods of spring 2019 Correlation matrix for the dataset ofthe determined saccharide compounds in PM25 samples is shownin Table S2 The supplement related to this article is available onlineat httpsdoiorg105194acp-21-12227-2021-supplement

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12233

soil containing wheat roots to the air which is beneficial forthe release of sucrose stored in roots (Medeiros et al 2006)thus resulting in a sharply increased sucrose concentration

It was reported that sugars such as glucose sucrose andfructose could be emitted from developing leaves (Grahamet al 2003) Glucose could be released from both soils andplant material (eg pollen fruits and their fragments) (Gra-ham et al 2003 Simoneit et al 2004 Fu et al 2012) Glu-cose and sucrose were rich in biologically active surface soils(Rogge et al 2007) In this study positive correlations werefound between sucrose and glucose (R = 052) (Table S2)suggesting a similar origin of glucose and sucrose in thisstudy Glucose and fructose have also been identified as a mi-nor product of cellulose pyrolysis because they were foundto be enriched in BB emissions (Nolte et al 2001) andthey correlated well with K+ (Graham et al 2002) and lev-oglucosan (Kang et al 2018) Herein no significant correla-tion was found between K+ levoglucosan and these mono-or disaccharides Therefore the detected glucose fructoseand sucrose might mostly be emitted by direct volatilizationfrom plant material or surface soils rather than as productsof polysaccharide breakdown during BB processes The highabundance of sucrose as well as glucose and fructose wasresponsible for biogenic aerosols associated with developingleaves and flowers as well as surface soil suspension

Trehalose as a stress protectant of various microorganismsand plants (Medeiros et al 2006 Jia and Fraser 2011) wasfound to be abundant in the fine-mode soil and has it beenproposed as a marker compound for fugitive dust from bio-logically active surface soils (Simoneit et al 2004 Medeiroset al 2006 Rogge et al 2007 Fu et al 2012) A previousstudy found a positive correlation between trehalose and cal-cium (Nishikawa et al 2000) In this study there was nosignificant correlation between trehalose and calcium Be-sides mannose has been reported to be one of the majormonosaccharide components in phytoplankton which origi-nated from marine biological fragments (Tanoue and Handa1987) Mannose was detected in only a few samples and pre-sented with low concentrations in this study

313 Sugar alcohols

Five sugar alcohol compounds including glycerol threitolmannitol arabitol and inositol were detected in PM25These reduced sugars are often reported to be related toplant senescence and decay by microorganisms (Simoneitet al 2004 Tsai et al 2013) they are produced by fungilichens soil biota and algae (Elbert et al 2007 Bauer et al2008) The average concentration of total sugar alcohols was1599 ng mminus3 with a range of 531ndash2540 ng mminus3 which ac-counted for 251plusmn 99 of the total measured saccharidesGlycerol has been widely found in soil biota (Simoneit et al2004) Previous studies suggested that the source of glyc-erol was not specific to biological emissions but the biomasscombustion might increase atmospheric glycerol concentra-

tions (Jia et al 2010 Graham et al 2002 Wang et al 2011)Herein glycerol was the second most abundant saccharidewith an average concentration of 1237 ng mminus3 accountingfor 51 ndash446 (average 226 ) of the total measuredsaccharides

Mannitol and arabitol have been proposed as tracersfor airborne fungal spores (Elbert et al 2007 Bauer etal 2008 Zhang et al 2010 Burshtein et al 2011)Mannitol and arabitol were detected with a concentra-tion range of 00ndash386 ng mminus3 (147 ng mminus3) and 00ndash211 ng mminus3 (58 ng mminus3) respectively The average concen-trations of mannitol and arabitol were comparable to those(average 113 and 91 ng mminus3) reported in the Beijing springaerosols (Liang et al 2013b) but were lower than those (av-erage 219 and 843 ng mminus3) in the Mediterranean summeraerosols (Burshtein et al 2011) and (30 and 24 ng mminus3) atHyytiaumllauml Finland in summer (Yttri et al 2011) Poor corre-lations (r = 038) were found among mannitol and arabitolin this study Nevertheless a positive correlation was foundbetween trehalose and mannitol (r = 079 P lt 005) (Ta-ble S2)

In the previous studies the total measured mannitol hasbeen measured and used for estimating the contribution offungal spores to organic carbon (Elbert et al 2007 Baueret al 2008 Zhang et al 2010) A factor of mannitol perspore (049plusmn 020 pg) was used to calculate the number con-centrations of fungal spores (Liang et al 2013a) and thenthe carbon content of fungal spores could be calculated us-ing a conversion factor of 13 pg C per spore obtained earlieras the average carbon content of spores from nine airbornefungal species with an uncertainty of 20 (Bauer et al2008) The diagnostic tracer ratio of mannitol to OC was cal-culated to be 00377 according to this research (Bauer et al2008 Liang et al 2013a) and then used to estimate the con-tribution of fungal spores to OC The contribution of fungalspores might be underestimated because previous results hadindicated that mannitol and arabitol were mainly associatedwith the coarse PM fraction (Samakeacute et al 2019) The av-erage mannitol concentration was 147plusmn 112 ng mminus3 duringthe observation period The average spore-derived OC wascalculated to be 3903 ng C mminus3 which contributed to 49 of the total OC

Claeys et al (2004) firstly identified two diastereoisomeric2-methyltetrols as oxidation products of isoprene in the Ama-zonian rain forest aerosols Henceforward 2-methyltetrolshave been used as tracers for isoprene-derived SOA (Lianget al 2012 Fu et al 2016 Yan et al 2019) In the previousstudies erythritol was often quantified as a surrogate of 2-methyltetrols (2-methylthreitol and 2-methylerythritol) dueto a lack of standards (Claeys et al 2004 Ding et al 20132016) In this study the concentration range of erythritol was04ndash198 ng mminus3 (average 111 ng mminus3) The values of in-ositol ranged from 00 to 228 ng mminus3 with average valuesof 58 ng mminus3 Moreover the sugar alcohols not only origi-nate from biological emissions but also from BB (Wan and

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12234 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Yu 2007 Jia et al 2010) Different levels of glycerol ara-bitol mannitol erythritol and inositol in fine particles havebeen found during the burning of crop residues and fallenleaves as well as indoor biofuel usage for heating and cook-ing (Graham et al 2002 Burshtein et al 2011 Wang et al2011 Yang et al 2012 Kang et al 2018) In this study onlyinositol correlated with levoglucosan (R = 042) suggestingthat inositol might be linked to biomass combustion sourcesHence the primary source of sugar alcohols associated withfine particles was biogenic aerosols at the observation sites

32 Sources and transport

Since the distinct concentration of the studied compoundswas due to different emission sources arising from differ-ent wind directions the 72 h backward trajectories for thesamples at the Dashu site (2412 N 10011 E) and the spa-tial distribution of the fire spots (8 March to 9 April 2019)were calculated to understand the source of saccharides inaerosol (Fig 4) The analysis of air-mass backward trajec-tories suggested that the air mass over Lincang was almostfrom the westerlies during the sampling periods and could beseparated into two episodes of remote western source above2000 m and local western source below 2000 m as shownin wine red and green lines Specifically 516 of air-massbackward trajectories were generally above 2000 m whereas484 of them were below 2000 m

The average concentrations of saccharide compounds aswell as the contributions for the episodes above and below2000 m are shown in Fig 5 The average concentrations oflevoglucosan and mannosan for the above 2000 m samples(3274 and 356 ng mminus3) were higher than those for the below2000 m samples (2503 and 273 ng mminus3) The anhydrosugarsaccounted for 492 and 369 of total saccharides respec-tively for the above and below 2000 m samples This impliedthat the levoglucosan at the observation site was attributedto both the local BB activities and BB smoke transportedfrom the neighboring regions of Southeast Asia (Myanmar)and the northern Indian subcontinent The southwest windfrom the Indian Ocean prevailed at Lincang all year roundIn spring the southwest wind was often affected by the low-temperature downhill wind blowing from the snow-coveredHengduan Mountains The weather frequently alternated be-tween hot and cold with unstable air pressure and strongwind Therefore the lower air could be diluted by the rel-atively clean cold air over the plateau The upper air mainlycame from the westerlies These results were in agreementwith the fact that residents across Southeast Asia use woodas an energy source to cook and generate heat

For glucose fructose and sucrose concentrations were alittle higher in the below 2000 m samples (averages 335264 and 1062 ng mminus3) than those in the above 2000 m sam-ples (averages 292 229 and 678 ng mminus3) This impliedthat biogenic aerosols (such as ruptured pollen) carrying sug-ars could cover long distance which was supported by pre-

Figure 4 Spatial distribution of the fire spots observed by MODISas well as the corresponding 72 h backward air-mass trajectory clus-ters arriving at 1500 m above ground level (agl) during the sam-pling periods for the collected samples The backward trajectorieswere separated into two episodes of remote western source above2000 m and local western source below 2000 m as shown in winered and green lines

vious studies which have observed long-range atmospherictransport of fine pollen from the Asian continent to the re-mote island Chichijima under the influence of the westerlies(Rousseau et al 2008) Although the pollen grains are usu-ally coarse with various shapes and hard shells this results ina relatively short retention time in the atmosphere Thereforeit could be concluded that in addition to the local pollen theconcentration of sucrose in Lincang was also influenced bythe transport of airborne pollen derived from South Asia

33 Source apportionment of saccharides

Based on the compositional data of saccharides and key rep-resentative markers for difference sources five factors asso-ciated with the emission sources of saccharides were finallyresolved by NMF As shown in Fig 6a factor 1 was char-acterized by high levels of levoglucosan (718 ) and man-nosan (787 ) suggesting the source of BB (Simoneit et al1999 Nolte et al 2001) Factor 2 was characterized by tre-halose (999 ) and mannitol (1000 ) and was enrichedin the other saccharide components ie arabitol (441 )glucose (296 ) erythritol (182 ) glycerol (178 ) lev-oglucosan (147 ) and sucrose (86 ) These saccharidecompounds had all been detected in the suspended soil par-

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12235

Figure 5 Average concentrations and contributions of saccharide compounds for the aerosol samples separated above and below 2000 m

ticles and associated microbiota (eg fungi bacteria and al-gae) (Simoneit et al 2004 Rogge et al 2007) A recentstudy found that leaves were a major source of saccharide-associated microbial taxa in a rural area of France (Samakeacuteet al 2020) Hence this factor was attributed to soil and leafmicrobiota Factor 3 had high levels of glycerol (714 ) anderythritol (582 ) and showed loadings of glucose (128 )and fructose (118 ) Kang et al (2018) reported that glyc-erol and erythritol presented large amounts in winter and au-tumn when vegetation is decomposed This factor was at-tributed to plant senescence and decay by microorganismsFactor 4 exhibited a predominance of sucrose (787 ) andshowed loadings of glucose (172 ) and arabitol (118 )This factor was regarded as the source of airborne pollen be-cause pollen was the reproductive unit of plants and containsthese saccharides and saccharide alcohols as nutritional com-ponents (Bieleski 1995 Miguel et al 2006 Fu et al 2012)Factor 5 characterized by the dominance of fructose (882 )was resolved and it was enriched in glucose (382 ) andarabitol (212 ) thus it could be regarded as the source ofplant detritus

The pie chart in Fig 6b shows the contribution of eachsource to total saccharides BB of factor 1 (340 ) wasfound as the dominant contributor to total saccharides Fac-tors 2ndash5 could all be associated with a biogenic source ac-counting for a total contribution of 660 The sources ofsoil microbiota (factor 2) plant senescence (factor 3) air-borne pollen (factor 4) and plant detritus (factor 5) respec-tively contributed 160 210 237 and 53 to totalsaccharides During the sampling periods daily variations inthe proportion of the five factors are shown in Fig S5 Fac-tor 2 soil microbiota emissions could be associated with soilreclamation and cultivation of farming periods whereas fac-tor 3 (plant senescence) and factor 5 (plant detritus) couldbe associated with the harvesting of vegetation or crops

During the observation period of a month along with theweather warming as sunshine enhanced humans left two ob-vious traces of cultivated soil from 9 to 17 March and from27 March to 8 April and a trace of vegetation or crop harvestfrom 17 to 30 March The stronger pollen discharge occurredin March probably due to the flowering of certain plants TheBB emissions peaked on 9 and 16 March and 1 April weremore prone to be open waste burnings

Since there is still some uncertainty in the factor appor-tionment the proportion of sources are only relative anduncertain It is still difficult to distinguish the contributionsof microorganisms and plants to biogenic aerosols Further-more all the above speculations about farming and harvest-ing periods are based on only 1 month of observation andlong-term observations are needed to obtain more accurateand effective information

4 Conclusion

With the help of various atmospheric saccharides this studypresents the characteristics of BB and various biogenic emis-sions to ambient aerosol in the rural sites of southwest ChinaLevoglucosan was the most dominant species among all thesaccharides with a concentration of 2877 ng mminus3 The ratiosof levoglucosan OC were 19 ndash89 (average 37 )BB contributed to 191 ndash739 of OC (average 413 )The results indicated that domestic biomass fuel burningopen BB events possibly open agricultural waste burningforest fire or sacrificial activity are significant during thespring in this area The total concentrations of five mono-or disaccharides and five sugar alcohols respectively con-tributed 249plusmn 83 and 266plusmn 99 to the total measuredsaccharides Based on the regression analysis these mono-or disaccharides and sugar alcohols were mostly emitted by

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12236 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 6 (a) Factor profile obtained by NMF analysis (b) Source contribution of the five factors to the total saccharides in PM25 samples

direct biogenic volatilization from plant material or surfacesoils rather than BB processes The sampling sites sufferedfrom both local emissions and BB via long-range transportfrom Southeast Asia (Myanmar Bangladesh) and the north-ern Indian subcontinent Five sources of saccharides wereresolved by NMF analysis including BB (34 ) soil mi-crobiota (160 ) plant senescence (210 ) airborne pollen(237 ) and plant detritus (53 ) at rural Lincang in spring

The data indicated that biofuel and open BB activities inrural southwest China and neighboring regions could havea significant impact on ambient aerosol levels In additionto the residential biofuel usage field burning of agriculturalresidues fallen leaves and forest fire were non-negligibleSome new technical measures of biomass resource utilizationare urgently needed to improve the open-burning emissionscenario in rural areas along with strict prohibition policy ofBB Meanwhile the characteristic analysis of saccharides inthe region can serve as a valuable reference for future stud-ies to evaluate temporal variations of biomass combustionand biogenic emission during modeling predictions and pol-icy making

Code availability At present the URLs of the database in our labare still being established The code used in the non-negative ma-trix factorization (NMF) analysis is available upon request from thecorresponding authors

Data availability The meteorological parameters were obtainedfrom the Physical Sciences Laboratory of NOAA (httpspslnoaagovdata last access 8 September 2020 NOAA 2020) 72 h back-trajectories of the aerosol were calculated via the NOAA ARLREADY website (httpreadyarlnoaagovHYSPLITphp last ac-cess 6 September 2020 ARL 2020) Fire pixel counts wereobtained from Moderate Resolution Imaging Spectroradiometer(MODIS) observations on NASA satellites (httpsearthdatanasagov last access 3 September 2020 NASA 2020) The three web-sites have open-access data At present the URLs of the databasein our lab are still being established All data described in this studyare available upon request from the corresponding authors

Supplement The locations of the sampling sites are shown inFig S1 Temporal variations of RH temperature solar irradiationand rainfall are shown in Fig S2 Average concentrations of saccha-ride compounds and the contribution of them for the Datian Dashuand Yakoutian samples are shown in Fig S3 Daily variation on av-erage concentrations of levoglucosan and K+ (panel a) arabitol andmannitol (panel b) PM25 Ca2+ and trehalose (panel c) at the threesites throughout the sampling period are shown in Fig S4 Figure S5shows daily variations on the proportion of the five factors to the to-tal saccharides in PM25 sampled at the three sites during the sam-pling periods Table S1 lists the concentrations of the carbonaceouscomponents and soluble inorganic ions in PM25 during the sam-pling periods of spring 2019 Correlation matrix for the dataset ofthe determined saccharide compounds in PM25 samples is shownin Table S2 The supplement related to this article is available onlineat httpsdoiorg105194acp-21-12227-2021-supplement

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

12234 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Yu 2007 Jia et al 2010) Different levels of glycerol ara-bitol mannitol erythritol and inositol in fine particles havebeen found during the burning of crop residues and fallenleaves as well as indoor biofuel usage for heating and cook-ing (Graham et al 2002 Burshtein et al 2011 Wang et al2011 Yang et al 2012 Kang et al 2018) In this study onlyinositol correlated with levoglucosan (R = 042) suggestingthat inositol might be linked to biomass combustion sourcesHence the primary source of sugar alcohols associated withfine particles was biogenic aerosols at the observation sites

32 Sources and transport

Since the distinct concentration of the studied compoundswas due to different emission sources arising from differ-ent wind directions the 72 h backward trajectories for thesamples at the Dashu site (2412 N 10011 E) and the spa-tial distribution of the fire spots (8 March to 9 April 2019)were calculated to understand the source of saccharides inaerosol (Fig 4) The analysis of air-mass backward trajec-tories suggested that the air mass over Lincang was almostfrom the westerlies during the sampling periods and could beseparated into two episodes of remote western source above2000 m and local western source below 2000 m as shownin wine red and green lines Specifically 516 of air-massbackward trajectories were generally above 2000 m whereas484 of them were below 2000 m

The average concentrations of saccharide compounds aswell as the contributions for the episodes above and below2000 m are shown in Fig 5 The average concentrations oflevoglucosan and mannosan for the above 2000 m samples(3274 and 356 ng mminus3) were higher than those for the below2000 m samples (2503 and 273 ng mminus3) The anhydrosugarsaccounted for 492 and 369 of total saccharides respec-tively for the above and below 2000 m samples This impliedthat the levoglucosan at the observation site was attributedto both the local BB activities and BB smoke transportedfrom the neighboring regions of Southeast Asia (Myanmar)and the northern Indian subcontinent The southwest windfrom the Indian Ocean prevailed at Lincang all year roundIn spring the southwest wind was often affected by the low-temperature downhill wind blowing from the snow-coveredHengduan Mountains The weather frequently alternated be-tween hot and cold with unstable air pressure and strongwind Therefore the lower air could be diluted by the rel-atively clean cold air over the plateau The upper air mainlycame from the westerlies These results were in agreementwith the fact that residents across Southeast Asia use woodas an energy source to cook and generate heat

For glucose fructose and sucrose concentrations were alittle higher in the below 2000 m samples (averages 335264 and 1062 ng mminus3) than those in the above 2000 m sam-ples (averages 292 229 and 678 ng mminus3) This impliedthat biogenic aerosols (such as ruptured pollen) carrying sug-ars could cover long distance which was supported by pre-

Figure 4 Spatial distribution of the fire spots observed by MODISas well as the corresponding 72 h backward air-mass trajectory clus-ters arriving at 1500 m above ground level (agl) during the sam-pling periods for the collected samples The backward trajectorieswere separated into two episodes of remote western source above2000 m and local western source below 2000 m as shown in winered and green lines

vious studies which have observed long-range atmospherictransport of fine pollen from the Asian continent to the re-mote island Chichijima under the influence of the westerlies(Rousseau et al 2008) Although the pollen grains are usu-ally coarse with various shapes and hard shells this results ina relatively short retention time in the atmosphere Thereforeit could be concluded that in addition to the local pollen theconcentration of sucrose in Lincang was also influenced bythe transport of airborne pollen derived from South Asia

33 Source apportionment of saccharides

Based on the compositional data of saccharides and key rep-resentative markers for difference sources five factors asso-ciated with the emission sources of saccharides were finallyresolved by NMF As shown in Fig 6a factor 1 was char-acterized by high levels of levoglucosan (718 ) and man-nosan (787 ) suggesting the source of BB (Simoneit et al1999 Nolte et al 2001) Factor 2 was characterized by tre-halose (999 ) and mannitol (1000 ) and was enrichedin the other saccharide components ie arabitol (441 )glucose (296 ) erythritol (182 ) glycerol (178 ) lev-oglucosan (147 ) and sucrose (86 ) These saccharidecompounds had all been detected in the suspended soil par-

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12235

Figure 5 Average concentrations and contributions of saccharide compounds for the aerosol samples separated above and below 2000 m

ticles and associated microbiota (eg fungi bacteria and al-gae) (Simoneit et al 2004 Rogge et al 2007) A recentstudy found that leaves were a major source of saccharide-associated microbial taxa in a rural area of France (Samakeacuteet al 2020) Hence this factor was attributed to soil and leafmicrobiota Factor 3 had high levels of glycerol (714 ) anderythritol (582 ) and showed loadings of glucose (128 )and fructose (118 ) Kang et al (2018) reported that glyc-erol and erythritol presented large amounts in winter and au-tumn when vegetation is decomposed This factor was at-tributed to plant senescence and decay by microorganismsFactor 4 exhibited a predominance of sucrose (787 ) andshowed loadings of glucose (172 ) and arabitol (118 )This factor was regarded as the source of airborne pollen be-cause pollen was the reproductive unit of plants and containsthese saccharides and saccharide alcohols as nutritional com-ponents (Bieleski 1995 Miguel et al 2006 Fu et al 2012)Factor 5 characterized by the dominance of fructose (882 )was resolved and it was enriched in glucose (382 ) andarabitol (212 ) thus it could be regarded as the source ofplant detritus

The pie chart in Fig 6b shows the contribution of eachsource to total saccharides BB of factor 1 (340 ) wasfound as the dominant contributor to total saccharides Fac-tors 2ndash5 could all be associated with a biogenic source ac-counting for a total contribution of 660 The sources ofsoil microbiota (factor 2) plant senescence (factor 3) air-borne pollen (factor 4) and plant detritus (factor 5) respec-tively contributed 160 210 237 and 53 to totalsaccharides During the sampling periods daily variations inthe proportion of the five factors are shown in Fig S5 Fac-tor 2 soil microbiota emissions could be associated with soilreclamation and cultivation of farming periods whereas fac-tor 3 (plant senescence) and factor 5 (plant detritus) couldbe associated with the harvesting of vegetation or crops

During the observation period of a month along with theweather warming as sunshine enhanced humans left two ob-vious traces of cultivated soil from 9 to 17 March and from27 March to 8 April and a trace of vegetation or crop harvestfrom 17 to 30 March The stronger pollen discharge occurredin March probably due to the flowering of certain plants TheBB emissions peaked on 9 and 16 March and 1 April weremore prone to be open waste burnings

Since there is still some uncertainty in the factor appor-tionment the proportion of sources are only relative anduncertain It is still difficult to distinguish the contributionsof microorganisms and plants to biogenic aerosols Further-more all the above speculations about farming and harvest-ing periods are based on only 1 month of observation andlong-term observations are needed to obtain more accurateand effective information

4 Conclusion

With the help of various atmospheric saccharides this studypresents the characteristics of BB and various biogenic emis-sions to ambient aerosol in the rural sites of southwest ChinaLevoglucosan was the most dominant species among all thesaccharides with a concentration of 2877 ng mminus3 The ratiosof levoglucosan OC were 19 ndash89 (average 37 )BB contributed to 191 ndash739 of OC (average 413 )The results indicated that domestic biomass fuel burningopen BB events possibly open agricultural waste burningforest fire or sacrificial activity are significant during thespring in this area The total concentrations of five mono-or disaccharides and five sugar alcohols respectively con-tributed 249plusmn 83 and 266plusmn 99 to the total measuredsaccharides Based on the regression analysis these mono-or disaccharides and sugar alcohols were mostly emitted by

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12236 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 6 (a) Factor profile obtained by NMF analysis (b) Source contribution of the five factors to the total saccharides in PM25 samples

direct biogenic volatilization from plant material or surfacesoils rather than BB processes The sampling sites sufferedfrom both local emissions and BB via long-range transportfrom Southeast Asia (Myanmar Bangladesh) and the north-ern Indian subcontinent Five sources of saccharides wereresolved by NMF analysis including BB (34 ) soil mi-crobiota (160 ) plant senescence (210 ) airborne pollen(237 ) and plant detritus (53 ) at rural Lincang in spring

The data indicated that biofuel and open BB activities inrural southwest China and neighboring regions could havea significant impact on ambient aerosol levels In additionto the residential biofuel usage field burning of agriculturalresidues fallen leaves and forest fire were non-negligibleSome new technical measures of biomass resource utilizationare urgently needed to improve the open-burning emissionscenario in rural areas along with strict prohibition policy ofBB Meanwhile the characteristic analysis of saccharides inthe region can serve as a valuable reference for future stud-ies to evaluate temporal variations of biomass combustionand biogenic emission during modeling predictions and pol-icy making

Code availability At present the URLs of the database in our labare still being established The code used in the non-negative ma-trix factorization (NMF) analysis is available upon request from thecorresponding authors

Data availability The meteorological parameters were obtainedfrom the Physical Sciences Laboratory of NOAA (httpspslnoaagovdata last access 8 September 2020 NOAA 2020) 72 h back-trajectories of the aerosol were calculated via the NOAA ARLREADY website (httpreadyarlnoaagovHYSPLITphp last ac-cess 6 September 2020 ARL 2020) Fire pixel counts wereobtained from Moderate Resolution Imaging Spectroradiometer(MODIS) observations on NASA satellites (httpsearthdatanasagov last access 3 September 2020 NASA 2020) The three web-sites have open-access data At present the URLs of the databasein our lab are still being established All data described in this studyare available upon request from the corresponding authors

Supplement The locations of the sampling sites are shown inFig S1 Temporal variations of RH temperature solar irradiationand rainfall are shown in Fig S2 Average concentrations of saccha-ride compounds and the contribution of them for the Datian Dashuand Yakoutian samples are shown in Fig S3 Daily variation on av-erage concentrations of levoglucosan and K+ (panel a) arabitol andmannitol (panel b) PM25 Ca2+ and trehalose (panel c) at the threesites throughout the sampling period are shown in Fig S4 Figure S5shows daily variations on the proportion of the five factors to the to-tal saccharides in PM25 sampled at the three sites during the sam-pling periods Table S1 lists the concentrations of the carbonaceouscomponents and soluble inorganic ions in PM25 during the sam-pling periods of spring 2019 Correlation matrix for the dataset ofthe determined saccharide compounds in PM25 samples is shownin Table S2 The supplement related to this article is available onlineat httpsdoiorg105194acp-21-12227-2021-supplement

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12235

Figure 5 Average concentrations and contributions of saccharide compounds for the aerosol samples separated above and below 2000 m

ticles and associated microbiota (eg fungi bacteria and al-gae) (Simoneit et al 2004 Rogge et al 2007) A recentstudy found that leaves were a major source of saccharide-associated microbial taxa in a rural area of France (Samakeacuteet al 2020) Hence this factor was attributed to soil and leafmicrobiota Factor 3 had high levels of glycerol (714 ) anderythritol (582 ) and showed loadings of glucose (128 )and fructose (118 ) Kang et al (2018) reported that glyc-erol and erythritol presented large amounts in winter and au-tumn when vegetation is decomposed This factor was at-tributed to plant senescence and decay by microorganismsFactor 4 exhibited a predominance of sucrose (787 ) andshowed loadings of glucose (172 ) and arabitol (118 )This factor was regarded as the source of airborne pollen be-cause pollen was the reproductive unit of plants and containsthese saccharides and saccharide alcohols as nutritional com-ponents (Bieleski 1995 Miguel et al 2006 Fu et al 2012)Factor 5 characterized by the dominance of fructose (882 )was resolved and it was enriched in glucose (382 ) andarabitol (212 ) thus it could be regarded as the source ofplant detritus

The pie chart in Fig 6b shows the contribution of eachsource to total saccharides BB of factor 1 (340 ) wasfound as the dominant contributor to total saccharides Fac-tors 2ndash5 could all be associated with a biogenic source ac-counting for a total contribution of 660 The sources ofsoil microbiota (factor 2) plant senescence (factor 3) air-borne pollen (factor 4) and plant detritus (factor 5) respec-tively contributed 160 210 237 and 53 to totalsaccharides During the sampling periods daily variations inthe proportion of the five factors are shown in Fig S5 Fac-tor 2 soil microbiota emissions could be associated with soilreclamation and cultivation of farming periods whereas fac-tor 3 (plant senescence) and factor 5 (plant detritus) couldbe associated with the harvesting of vegetation or crops

During the observation period of a month along with theweather warming as sunshine enhanced humans left two ob-vious traces of cultivated soil from 9 to 17 March and from27 March to 8 April and a trace of vegetation or crop harvestfrom 17 to 30 March The stronger pollen discharge occurredin March probably due to the flowering of certain plants TheBB emissions peaked on 9 and 16 March and 1 April weremore prone to be open waste burnings

Since there is still some uncertainty in the factor appor-tionment the proportion of sources are only relative anduncertain It is still difficult to distinguish the contributionsof microorganisms and plants to biogenic aerosols Further-more all the above speculations about farming and harvest-ing periods are based on only 1 month of observation andlong-term observations are needed to obtain more accurateand effective information

4 Conclusion

With the help of various atmospheric saccharides this studypresents the characteristics of BB and various biogenic emis-sions to ambient aerosol in the rural sites of southwest ChinaLevoglucosan was the most dominant species among all thesaccharides with a concentration of 2877 ng mminus3 The ratiosof levoglucosan OC were 19 ndash89 (average 37 )BB contributed to 191 ndash739 of OC (average 413 )The results indicated that domestic biomass fuel burningopen BB events possibly open agricultural waste burningforest fire or sacrificial activity are significant during thespring in this area The total concentrations of five mono-or disaccharides and five sugar alcohols respectively con-tributed 249plusmn 83 and 266plusmn 99 to the total measuredsaccharides Based on the regression analysis these mono-or disaccharides and sugar alcohols were mostly emitted by

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12236 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 6 (a) Factor profile obtained by NMF analysis (b) Source contribution of the five factors to the total saccharides in PM25 samples

direct biogenic volatilization from plant material or surfacesoils rather than BB processes The sampling sites sufferedfrom both local emissions and BB via long-range transportfrom Southeast Asia (Myanmar Bangladesh) and the north-ern Indian subcontinent Five sources of saccharides wereresolved by NMF analysis including BB (34 ) soil mi-crobiota (160 ) plant senescence (210 ) airborne pollen(237 ) and plant detritus (53 ) at rural Lincang in spring

The data indicated that biofuel and open BB activities inrural southwest China and neighboring regions could havea significant impact on ambient aerosol levels In additionto the residential biofuel usage field burning of agriculturalresidues fallen leaves and forest fire were non-negligibleSome new technical measures of biomass resource utilizationare urgently needed to improve the open-burning emissionscenario in rural areas along with strict prohibition policy ofBB Meanwhile the characteristic analysis of saccharides inthe region can serve as a valuable reference for future stud-ies to evaluate temporal variations of biomass combustionand biogenic emission during modeling predictions and pol-icy making

Code availability At present the URLs of the database in our labare still being established The code used in the non-negative ma-trix factorization (NMF) analysis is available upon request from thecorresponding authors

Data availability The meteorological parameters were obtainedfrom the Physical Sciences Laboratory of NOAA (httpspslnoaagovdata last access 8 September 2020 NOAA 2020) 72 h back-trajectories of the aerosol were calculated via the NOAA ARLREADY website (httpreadyarlnoaagovHYSPLITphp last ac-cess 6 September 2020 ARL 2020) Fire pixel counts wereobtained from Moderate Resolution Imaging Spectroradiometer(MODIS) observations on NASA satellites (httpsearthdatanasagov last access 3 September 2020 NASA 2020) The three web-sites have open-access data At present the URLs of the databasein our lab are still being established All data described in this studyare available upon request from the corresponding authors

Supplement The locations of the sampling sites are shown inFig S1 Temporal variations of RH temperature solar irradiationand rainfall are shown in Fig S2 Average concentrations of saccha-ride compounds and the contribution of them for the Datian Dashuand Yakoutian samples are shown in Fig S3 Daily variation on av-erage concentrations of levoglucosan and K+ (panel a) arabitol andmannitol (panel b) PM25 Ca2+ and trehalose (panel c) at the threesites throughout the sampling period are shown in Fig S4 Figure S5shows daily variations on the proportion of the five factors to the to-tal saccharides in PM25 sampled at the three sites during the sam-pling periods Table S1 lists the concentrations of the carbonaceouscomponents and soluble inorganic ions in PM25 during the sam-pling periods of spring 2019 Correlation matrix for the dataset ofthe determined saccharide compounds in PM25 samples is shownin Table S2 The supplement related to this article is available onlineat httpsdoiorg105194acp-21-12227-2021-supplement

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

12236 Z Wang et al Saccharide composition in atmospheric fine particulate matter

Figure 6 (a) Factor profile obtained by NMF analysis (b) Source contribution of the five factors to the total saccharides in PM25 samples

direct biogenic volatilization from plant material or surfacesoils rather than BB processes The sampling sites sufferedfrom both local emissions and BB via long-range transportfrom Southeast Asia (Myanmar Bangladesh) and the north-ern Indian subcontinent Five sources of saccharides wereresolved by NMF analysis including BB (34 ) soil mi-crobiota (160 ) plant senescence (210 ) airborne pollen(237 ) and plant detritus (53 ) at rural Lincang in spring

The data indicated that biofuel and open BB activities inrural southwest China and neighboring regions could havea significant impact on ambient aerosol levels In additionto the residential biofuel usage field burning of agriculturalresidues fallen leaves and forest fire were non-negligibleSome new technical measures of biomass resource utilizationare urgently needed to improve the open-burning emissionscenario in rural areas along with strict prohibition policy ofBB Meanwhile the characteristic analysis of saccharides inthe region can serve as a valuable reference for future stud-ies to evaluate temporal variations of biomass combustionand biogenic emission during modeling predictions and pol-icy making

Code availability At present the URLs of the database in our labare still being established The code used in the non-negative ma-trix factorization (NMF) analysis is available upon request from thecorresponding authors

Data availability The meteorological parameters were obtainedfrom the Physical Sciences Laboratory of NOAA (httpspslnoaagovdata last access 8 September 2020 NOAA 2020) 72 h back-trajectories of the aerosol were calculated via the NOAA ARLREADY website (httpreadyarlnoaagovHYSPLITphp last ac-cess 6 September 2020 ARL 2020) Fire pixel counts wereobtained from Moderate Resolution Imaging Spectroradiometer(MODIS) observations on NASA satellites (httpsearthdatanasagov last access 3 September 2020 NASA 2020) The three web-sites have open-access data At present the URLs of the databasein our lab are still being established All data described in this studyare available upon request from the corresponding authors

Supplement The locations of the sampling sites are shown inFig S1 Temporal variations of RH temperature solar irradiationand rainfall are shown in Fig S2 Average concentrations of saccha-ride compounds and the contribution of them for the Datian Dashuand Yakoutian samples are shown in Fig S3 Daily variation on av-erage concentrations of levoglucosan and K+ (panel a) arabitol andmannitol (panel b) PM25 Ca2+ and trehalose (panel c) at the threesites throughout the sampling period are shown in Fig S4 Figure S5shows daily variations on the proportion of the five factors to the to-tal saccharides in PM25 sampled at the three sites during the sam-pling periods Table S1 lists the concentrations of the carbonaceouscomponents and soluble inorganic ions in PM25 during the sam-pling periods of spring 2019 Correlation matrix for the dataset ofthe determined saccharide compounds in PM25 samples is shownin Table S2 The supplement related to this article is available onlineat httpsdoiorg105194acp-21-12227-2021-supplement

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12237

Author contributions ZZW JMC and QL designed the researchDW ZZW and RJC conducted the sampling activities ZZW andZYL did the measurements ZZW and XNS analyzed the resultsJMC HDK XL and HLO made suggestions for this paper andZZW wrote the article with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Disclaimer Publisherrsquos note Copernicus Publications remainsneutral with regard to jurisdictional claims in published maps andinstitutional affiliations

Special issue statement This article is part of the specialissue ldquoThe role of fire in the Earth system understand-ing interactions with the land atmosphere and society(ESDACPBGGMDNHESS inter-journal SI)rdquo It is a resultof the EGU General Assembly 2020 3ndash8 May 2020

Financial support This research has been supported by the Na-tional Natural Science Foundation of China (grant nos 9184330191743202 91843302)

Review statement This paper was edited by Ivan Kourtchev andreviewed by three anonymous referees

References

ARL (Air Resources Laboratory) of NOAA HYSPLIT Model[data set] available at httpreadyarlnoaagovHYSPLITphplast access 6 September 2020

Bauer H Claeys M Vermeylen R Schueller E Weinke GBerger A and Puxbaum H Arabitol and mannitol as tracers forthe quantification of airborne fungal spores Atmos Environ 42588ndash593 httpsdoiorg101016jatmosenv200710013 2008

Bieleski R L Onset of phloem export from senes-cent petals of daylily Plant Physiol 109 557ndash565httpsdoiorg101104pp1092557 1995

Burshtein N Lang-Yona N and Rudich Y Ergosterol ara-bitol and mannitol as tracers for biogenic aerosols in theeastern Mediterranean Atmos Chem Phys 11 829ndash839httpsdoiorg105194acp-11-829-2011 2011

Caseiro A Marr I L Claeys M Kasper-Giebl A PuxbaumH and Pio C A Determination of saccharides in atmosphericaerosol using anion-exchange high-performance liquid chro-matography and pulsed-amperometric detection J ChromatogrA 1171 37ndash45 httpsdoiorg101016jchroma2007090382007

Chen J Kawamura K Liu C Q and Fu P Q Long-term observations of saccharides in remote marine aerosolsfrom the western North Pacific a comparison between 1990-

1993 and 2006-2009 periods Atmos Environ 67 448ndash458httpsdoiorg101016jatmosenv201211014 2013

Chen J M Li C L Ristovski Z Milic A Gu Y T IslamM S Wang S X Hao J M Zhang H F He C R GuoH Fu H B Miljevic B Morawska L Thai P Lam Y FPereira G Ding A J Huang X and Dumka U C A re-view of biomass burning emissions and impacts on air qualityhealth and climate in China Sci Total Environ 579 1000ndash1034httpsdoiorg101016jscitotenv201611025 2017

Chen Y and Xie S Characteristics and formation mechanismof a heavy air pollution episode caused by biomass burning inChengdu Southwest China Sci Total Environ 473ndash474 507ndash517 httpsdoiorg101016jscitotenv201312069 2014

Cheng Y Engling G He K-B Duan F-K Ma Y-L Du Z-Y Liu J-M Zheng M and Weber R J Biomass burningcontribution to Beijing aerosol Atmos Chem Phys 13 7765ndash7781 httpsdoiorg105194acp-13-7765-2013 2013

Chow J C Watson J G Chen L W A Chang M C O Robin-son N F Trimble D and Kohl S The IMPROVE_A temper-ature protocol for thermal optical carbon analysis maintainingconsistency with a long-term database J Air Waste Manag 571014ndash1023 httpsdoiorg1031551047-32895791014 2007

Christner B C Morris C E Foreman C M Cai R and SandsD C Ubiquity of biological ice nucleators in snowfall Science319 1214 httpsdoiorg101126science1149757 2008

Claeys M Graham B Vas G Wang W Vermeylen R Pashyn-ska V Cafmeyer J Guyon P Andreae M O Artaxo Pand Maenhaut W Formation of secondary organic aerosolsthrough photooxidation of isoprene Science 303 1173ndash1176httpsdoiorg101126science1092805 2004

Despreacutes V R Huffman J A Burrows S M Hoose C Safa-tov A S Buryak G FroehlichNowoisky J Elbert W An-dreae M O Poumlschl U and Jaenicke R Primary biologicalaerosol particles in the atmosphere a review Tellus B 64 1ndash58httpsdoiorg103402tellusbv64i015598 2012

Ding X Wang X M Xie Z Q Zhang X and Sun LG Impacts of Siberian biomass burning on organic aerosolsover the North Pacific Ocean and the Arctic Primary and sec-ondary organic tracers Environ Sci Technol 47 3149ndash3157httpsdoiorg101021es3037093 2013

Ding X He Q F Shen R Q Yu Q Q Zhang Y Q Xin JY Wen T X and Wang X M Spatial and seasonal variationsof isoprene secondary organic aerosol in China significant im-pact of biomass burning during winter Sci Rep-UK 6 20411httpsdoiorg101038srep20411 2016

Draxler R R and Hess G D An overview of the HYSPLIT-4 modelling system for trajectories dispersion and depositionAust Meteorol Mag 47 295ndash308 1998

Elbert W Taylor P E Andreae M O and Poumlschl U Contribu-tion of fungi to primary biogenic aerosols in the atmosphere wetand dry discharged spores carbohydrates and inorganic ions At-mos Chem Phys 7 4569ndash4588 httpsdoiorg105194acp-7-4569-2007 2007

Engling G Carrico C M Kreindenweis S M ColletJ L Day D E Malm W C Lincoln E Hao WM Iinuma Y and Herrmann H Determination oflevoglucosan in biomass combustion aerosol by high-performance anion-exchange chromatography with pulsed

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12238 Z Wang et al Saccharide composition in atmospheric fine particulate matter

amperometric detection Atmos Environ 40 S299ndashS311httpsdoiorg101016jatmosenv200512069 2006

Engling G Zhang Y N Chan C Y Sang X F Lin M HoK F Li Y S Lin C Y and Lee J J Characterization andsources of aerosol particles over the southeastern Tibetan Plateauduring the Southeast Asia biomass-burning season Tellus B63 117ndash128 httpsdoiorg101111j1600-0889201000512x2011

Fabbri D Torri C Simoneit B R T Marynowski LRushdi A I and Fabianska M J Levoglucosan andother cellulose and lignin markers in emissions from burn-ing of Miocene lignites Atmos Environ 43 2286ndash2295httpsdoiorg101016jatmosenv200901030 2009

Fu P Q Kawamura K and Barrie L A Photochemical andother sources of organic compounds in the Canadian high Arcticaerosol pollution during winter-spring Environ Sci Technol43 286ndash292 httpsdoiorg101021es803046q 2009

Fu P Q Kawamura K Kobayashi M and Simoneit B R TSeasonal variations of sugars in atmospheric particulate matterfrom Gosan Jeju Island significant contributions of airbornepollen and Asian dust in spring Atmos Environ 55 234ndash239httpsdoiorg101016jatmosenv201202061 2012

Fu P Q Kawamura K Chen J Charriegravere B and SempeacutereacuteR Organic molecular composition of marine aerosols over theArctic Ocean in summer contributions of primary emissionand secondary aerosol formation Biogeosciences 10 653ndash667httpsdoiorg105194bg-10-653-2013 2013

Fu P Q Zhuang G S Sun Y L Wang Q ZChen J Ren LJ Yang F Wang Z F Pan XL Li X D and Kawamura K Molecular markers ofbiomass burning fungal spores and biogenic SOA in theTaklimakan desert aerosols Atmos Environ 130 64ndash73httpsdoiorg101016jatmosenv201510087 2016

Fraser M P and Lakshmanan K Using levoglucosan as amolecular marker for the long-range transport of biomasscombustion aerosols Environ Sci Technol 34 4560ndash4564httpsdoiorg101021es991229l 2000

Graham B Mayol-Bracero O L Guyon P Robert G CDecesari S Facchini M C Artaxo P Maenhaut W PK and Andreae M O Water-soluble organic compoundsin biomass burning aerosols over Amazonia 1 Characteri-zation by NMR and GC-MS J Geophys Res 107 8047httpsdoiorg1010292001JD000336 2002

Graham B Guyon P Taylor P E Artaxo P Maenhaut WGlovsky M M Flagan R C and Andreae M O Organiccompounds present in the natural Amazonian aerosol character-ization by gas chromatography-mass spectrometry J GeophysRes-Atmos 108 4766 httpsdoiorg1010292003JD0039902003

Harrison R M Beddows D C S Hu L and Yin J Compar-ison of methods for evaluation of wood smoke and estimationof UK ambient concentrations Atmos Chem Phys 12 8271ndash8283 httpsdoiorg105194acp-12-8271-2012 2012

Hennigan C J Sullivan A P Collett Jr J L and RobinsonA L Levoglucosan stability in biomass burning particles ex-posed to hydroxyl radicals Geophys Res Lett 37 L09806httpsdoiorg1010292010GL043088 2010

Holden A S Sullivan A P Munchak L A Kreiden-weis S M Schichtel B A Malm W C and Col-

lett Jr J L Determining contributions of biomass burn-ing and other sources to fine particle contemporary carbon inthe western United States Atmos Environ 45 1986ndash1993httpsdoiorg101016jatmosenv201101021 2011

Jacobson M C Hansson H C Noone K J and Charlson R JOrganic atmospheric aerosols review and state of science RevGeophys 38 267ndash294 httpsdoiorg1010291998RG0000452000

Jaenicke R Abundance of cellular material andproteins in the atmosphere Science 308 73httpsdoiorg101126science1106335 2005

Jaenicke R Matthias-Maser S and Gruber S Omnipresence ofbiological material in the atmosphere Environ Chem 4 217ndash220 httpsdoiorg101071EN07021 2007

Jia Y L and Fraser M Characterization of saccharides in size-fractionated ambient particulate matter and aerosol sources thecontribution of primary biological aerosol particles (PBAPs)and soil to ambient Environ Sci Technol 45 930ndash936httpsdoiorg101021es103104e 2011

Jia Y L Bhat S G and Fraser M P Characterizationof saccharides and other organic compounds in fine parti-cles and the use of saccharides to track primary biologi-cally derived carbon sources Atmos Environ 44 724ndash732httpsdoiorg101016jatmosenv200910034 2010

Kalogridis A C Popovicheva O B Engling G Di-apouli E Kawamura K Tachibana E Ono K Ko-zlov V S and Eleftheriadis K Smoke aerosol chem-istry and aging of Siberian biomass burning emissions ina large aerosol chamber Atmos Environ 185 15ndash28httpsdoiorg101016jatmosenv201804033 2018

Kang M J Ren L J Ren H Zhao Y Kawamura K Zhang HL Wei L F Sun Y L Wang Z F and Fu P Q Primary bio-genic and anthropogenic sources of organic aerosols in BeijingChina Insights from saccharides and n-alkanes Environ Pollut243 1579ndash1587 httpsdoiorg101016jenvpol2018091182018

Kleindienst T E Jaoui M Lewandowski M OffenbergJ H Lewis C W Bhave P V and Edney E OEstimates of the contributions of biogenic and anthro-pogenic hydrocarbons to secondary organic aerosol at asoutheastern US location Atmos Environ 41 8288ndash8300httpsdoiorg101016jatmosenv200706045 2007

Lai C Y Liu Y C Ma J Z Ma Q X and He H Degrada-tion kinetics of levoglucosan initiated by hydroxyl radical underdifferent environmental conditions Atmos Environ 91 32ndash39httpsdoiorg101016jatmosenv201403054 2014

Lee D D and Seung H S Learning the parts of objects by non-negative matrix factorization Nature 401 788ndash791 1999

Lee T Sullivan A P Mack L Jimenez J L Kreiden-weis S M Onasch T B Worsnop D R Malm WWold C E and Hao W M Chemical smoke marker emis-sions during flaming and smoldering phases of laboratoryopen burning of wildland fuels Aerosol Sci Tech 44 1ndash5httpsdoiorg101080027868262010499884 2010

Li Q L Wang N L Barbante C Kang S C Calle-garo A Battistel D Argiriadis E Wan X Yao P PuT Wu X B Han Y and Huai Y P Biomass burn-ing source identification through molecular markers in cry-

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12239

oconites over the Tibetan Plateau Environ Pollut 244 209ndash217 httpsdoiorg101016jenvpol201810037 2019

Li X Chen M X Le H P Wang F W Guo Z G IinumaY Chen J M and Herrmann H Atmospheric outflow ofPM25 saccharides from megacity Shanghai to East China SeaImpact of biological and biomass burning sources Atmos Envi-ron 143 1ndash14 httpsdoiorg101016jatmosenv2016080392016a

Li X Jiang L Hoa L P Lyu Y Xu T T Yang XIinuma Y Chen J M and Herrmann H Size distributionof particle-phase sugar and nitrophenol tracers during severe ur-ban haze episodes in Shanghai Atmos Environ 145 115ndash127httpsdoiorg101016jatmosenv201609030 2016b

Liang L L Engling G Duan F K Cheng Y andHe K B Characteristics of 2-methyltetrols in ambientaerosol in Beijing China Atmos Environ 59 376ndash381httpsdoiorg101016jatmosenv201205052 2012

Liang L Engling G Cheng Y Duan F K Du Z Y and HeK B Rapid detection and quantification of fungal spores in theurban atmosphere by flow cytometry J Aerosol Sci 66 179ndash186 httpsdoiorg101016jjaerosci201308013 2013a

Liang L L Engling G He K B Du Z Y Cheng Y andDuan F K Evaluation of fungal spore characteristics in BeijingChina based on molecular tracer measurements Environ ResLett 8 014005 httpsdoiorg1010881748-9326810140052013b

Liang L L Engling G Du Z Y Cheng Y Duan FK Liu X Y and He K B Seasonal variations andsource estimation of saccharides in atmospheric particu-late matter in Beijing China Chemosphere 150 365ndash377httpsdoiorg101016jchemosphere201602002 2016

Medeiros P M Conte M H Weber J C and SimoneitB R T Sugars as source indicators of biogenic or-ganic carbon in aerosols collected above the Howland Ex-perimental Forest Maine Atmos Environ 40 1694ndash1705httpsdoiorg101016jatmosenv200511001 2006

Miguel A G Taylor P E House J Glovsky M M and Fla-gan R C Meteorological influences on respirable fragment re-lease from Chinese elm pollen Aerosol Sci Tech 40 690ndash696httpsdoiorg10108002786820600798869 2006

Miyazaki Y Fu P Q Kawamura K Mizoguchi Y and Ya-manoi K Seasonal variations of stable carbon isotopic com-position and biogenic tracer compounds of water-soluble organicaerosols in a deciduous forest Atmos Chem Phys 12 1367ndash1376 httpsdoiorg105194acp-12-1367-2012 2012

Mochida M Kawamura K Fu P Q and TakemuraT Seasonal variation of levoglucosan in aerosols overthe western North Pacific and its assessment as abiomass-burning tracer Atmos Environ 44 3511ndash3518httpsdoiorg101016jatmosenv201006017 2010

NASA EARTHDATA of NASA (National Aeronautics and SpaceAdministration) Active Fire Data [data set] available at httpsearthdatanasagov last access 3 September 2020

Nishikawa M Hao Q and Morita M Preparation and evaluationof certified reference materials from Asian mineral dust GlobalEnviron Res 4 103ndash113 2000

Nolte C G Schauer J J Cass G R and Simoneit B R THighly polar organic compounds present in wood smoke and in

the ambient atmosphere Environ Sci Technol 35 1912ndash1919httpsdoiorg101021es001420r 2001

Nirmalkar J Deshmukh D K Deb M K Tsai YI and Sopajaree K Mass loading and episodic varia-tion of molecular markers in PM25 aerosols over a ruralarea in eastern central India Atmos Environ 117 41ndash50httpsdoiorg101016jatmosenv201507003 2015

NOAA Physical Sciences Laboratory (PSL) of NOAA (NationalOceanic and Atmospheric Administration) Gridded ClimateDatasets [data set] available at httpspslnoaagovdata lastaccess 8 September 2020

Pashynska V Vermeylen R Vas G Maenhaut W and ClaeysM Development of a gas chromatographicion trap massspectrometric method for the determination of levoglucosanand saccharidic compounds in atmospheric aerosols Appli-cation to urban aerosol J Mass Spectrom 37 1249ndash1257httpsdoiorg101002jms391 2002

Pietrogrande M C Bacco D Visentin M Ferrari S andCasali P Polar organic marker compounds in atmosphericaerosol in the Po Valley during the Supersito campaigns-Part2 Seasonal variations of sugars Atmos Environ 97 215ndash225httpsdoiorg101016jatmosenv201407056 2014

Puxbaum H Caseiro A Saacutenchez-Ochoa A Kasper-Giebl AClaeys M Gelencseacuter A Legrand M Preunkert S andPio C Levoglucosan levels at background sites in Europefor assessing the impact of biomass combustion on the Eu-ropean aerosol background J Geophys Res 112 D23S05httpsdoiorg1010292006JD008114 2007

Poumlschl U Martin S T Sinha B Chen Q Gunthe S SHuffman J A Borrmann S Farmer D K Garland R MHelas G Jimenez J L King S M Manzi A MikhailovE Pauliquevis T Petters M D Prenni A J Roldin PRose D Schneider J Su H Zorn S R Artaxo P and An-dreae MO Rainforest aerosols as biogenic nuclei of cloudsand precipitation in the Amazon Science 329 1513ndash1516httpsdoiorg101126science1191056 2010

Reyes-Villegas E Bannan T Le Breton M Mehra A PriestleyM Percival C Coe H and Allan J D Online chemical char-acterization of food-cooking organic aerosols Implications forsource apportionment Environ Sci Technol 52 5308ndash5318httpsdoiorg101021acsest7b06278 2018

Rogge W F Hildemann L M Mazurek M A Cass G R andSimoneit B R T Sources of fine organic aerosol 4 Particulateabrasion products from leaf surfaces of urban plants EnvironSci Technol 27 2700ndash2711 1993

Rogge W F Medeiros P M and Simoneit B R TOrganic marker compounds in surface soils of cropfields from the San Joaquin Valley fugitive dust char-acterization study Atmos Environ 41 8183ndash8204httpsdoiorg101016jatmosenv200706030 2007

Rousseau D D Schevin P Ferrier J Jolly D An-dreasen T Ascanius S E Hendriksen S E and PoulsenU Long-distance pollen transport from North Americato Greenland in spring J Geophys Res 113 G02013httpsdoiorg1010292007JG000456 2008

Rybicki M Marynowski L and Simoneit B R T Compositionof organic compounds from low-temperature burning of ligniteand their application as tracers in ambient air Chemosphere 249

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021

12240 Z Wang et al Saccharide composition in atmospheric fine particulate matter

126087 httpsdoiorg101016jchemosphere20201260872020

Samakeacute A Jaffrezo J-L Favez O Weber S Jacob V Albi-net A Riffault V Perdrix E Waked A Golly B SalamehD Chevrier F Oliveira D M Bonnaire N Besombes J-L Martins J M F Conil S Guillaud G Mesbah B RocqB Robic P-Y Hulin A Le Meur S Descheemaecker MChretien E Marchand N and Uzu G Polyols and glu-cose particulate species as tracers of primary biogenic organicaerosols at 28 French sites Atmos Chem Phys 19 3357ndash3374httpsdoiorg105194acp-19-3357-2019 2019

Samakeacute A Bonin A Jaffrezo J-L Taberlet P Weber S UzuG Jacob V Conil S and Martins J M F High levels ofprimary biogenic organic aerosols are driven by only a few plant-associated microbial taxa Atmos Chem Phys 20 5609ndash5628httpsdoiorg105194acp-20-5609-2020 2020

Sang X F Chan C Y Engling G Chan L Y Wang X MZhang Y N Shi S Zhang Z S Zhang T and Hu MLevoglucosan enhancement in ambient aerosol during spring-time transport events of biomass burning smoke to SoutheastChina Tellus B 63 129ndash139 httpsdoiorg101111j1600-0889201000515x 2011

Sang X F Zhang Z S Chan C Y and Engling G Sourcecategories and contribution of biomass smoke to organic aerosolover the southeastern Tibetan Plateau Atmos Environ 78 113ndash123 httpsdoiorg101016jatmosenv201212012 2013

Schkolnik G Falkovich AH Rudich Y Maenhaut W and Ar-taxo P New analytical method for the determination of levoglu-cosan polyhydroxy compounds and 2-methylerythritol and itsapplication to smoke and rainwater samples Environ Sci Tech-nol 39 2744ndash2752 httpsdoiorg101021es048363c 2005

Shang X Zhang K Meng F Wang S Lee M Suh I KimD Jeon K Park H Wang X and Zhao Y Characteristicsand source apportionment of fine haze aerosol in Beijing dur-ing the winter of 2013 Atmos Chem Phys 18 2573ndash2584httpsdoiorg105194acp-18-2573-2018 2018

Shen R R Liu Z R Liu Y S Wang L L DongL Wang Y S Wang G A Bai Y and Li XR Typical polar organic aerosol tracers in PM25 overthe North China Plain Spatial distribution seasonal varia-tions contribution and sources Chemosphere 209 758ndash766httpsdoiorg101016jchemosphere201806133 2018

Simoneit B R T Schauer J J Nolte C G Oros D R Elias VO Fraser M P Rogge W F and Cass G R Levoglucosan atracer for cellulose in biomass burning and atmospheric particlesAtmos Environ 33 173ndash182 httpsdoiorg101016S1352-2310(98)00145-9 1999

Simoneit B R T Elias V O Kobayashi M KawamuraK Rushdi A I Medeiros P M Rogge W F andDidyk B M Sugars-Dominant water-soluble organic com-pounds in soils and characterization as tracers in atmosphericparticulate matter Environ Sci Technol 38 5939ndash5949httpsdoiorg101021es0403099 2004

Sullivan A P May A A Lee T McMeeking G R Kreiden-weis S M Akagi S K Yokelson R J Urbanski S P andCollett Jr J L Airborne characterization of smoke marker ra-tios from prescribed burning Atmos Chem Phys 14 10535ndash10545 httpsdoiorg105194acp-14-10535-2014 2014

Tang M Gu W Ma Q Li Y J Zhong C Li S YinX Huang R-J He H and Wang X Water adsorption andhygroscopic growth of six anemophilous pollen species theeffect of temperature Atmos Chem Phys 19 2247ndash2258httpsdoiorg105194acp-19-2247-2019 2019

Tanoue E and Handa N Monosaccharide composition of marineparticles and sediments from the Bering Sea and Northern NorthPacific Oceanol Acta 10 91ndash99 1987

Tominaga S Matsumoto K Kaneyasu N Shigihara A KatonoK and Igawa M Measurements of particulate sugars at urbanand forested suburban sites Atmos Environ 45 2335ndash2339httpsdoiorg101016jatmosenv201009056 2011

Tsai Y I Wu P L Hsu Y T and Yang C R Anhydrosugar andsugar alcohol organic markers associated with carboxylic acidsin particulate matter from incense burning Atmos Environ44 3708ndash3718 httpsdoiorg101016jatmosenv2010060302010

Tsai Y I Sopajaree K Chotruksa A Wu H C and KuoS S Source indicators of biomass burning associated with in-organic salts and carboxylates in dry season ambient aerosolin Chiang Mai Basin Thailand Atmos Environ 78 93ndash104httpsdoiorg101016jatmosenv201209040 2013

Wan E C H and Yu J Z Analysis of sugars andsugar polyols in atmospheric aerosols by chloride attach-ment in liquid chromatographynegative ion electrospraymass spectrometry Environ Sci Technol 41 2459ndash2466httpsdoiorg101021es062390g 2007

Wang G Chen C Li J Zhou B Xie M Hu S Kawa-mura K and Chen Y Molecular composition and size dis-tribution of sugars sugar-alcohols and carboxylic acids inairborne particles during a severe urban haze event causedby wheat straw burning Atmos Environ 45 2473ndash2479httpsdoiorg101016jatmosenv201102045 2011

Wang Q Q Shao M Liu Y William K Paul G LiX H Liu Y A and Lu S H Impact of biomassburning on urban air quality estimated by organic tracersGuangzhou and Beijing as cases Atmos Environ 41 8380ndash8390 httpsdoiorg101016jatmosenv200706048 2007

Wang X Shen Z X Liu F B Lu D Tao J Lei Y LZhang Q Zeng Y L Xu H M Wu Y F Zhang R Jand Cao J J Saccharides in summer and winter PM25 overXirsquoan Northwestern China Sources and yearly variations ofbiomass burning contribution to PM25 Environ Res 214 410ndash417 httpsdoiorg101016jatmosres201808024 2018

Womiloju T O Miller J D Mayer P M and Brook J R Meth-ods to determine the biological composition of particulate mat-ter collected from outdoor air Atmos Environ 37 4335ndash4344httpsdoiorg101016S1352-2310(03)00577-6 2003

Wu J Kong S F Zeng X Cheng Y Yan Q Zheng HYan Y Y Zheng S R Liu D T Zhang X Y Fu P QWang S X and Qi S H First High-resolution emission in-ventory of levoglucosan for biomass burning and non-biomassburning sources in China Environ Sci Technol 55 1497ndash1507httpsdoiorg101021acsest0c06675 2021

Xiao M X Wang Q Z Qin X F Yu G Y and DengC R Composition sources and distribution of PM25 sac-charides in a coastal urban site of china Atmosphere 9 274httpsdoiorg103390atmos9070274 2018

Atmos Chem Phys 21 12227ndash12241 2021 httpsdoiorg105194acp-21-12227-2021

Z Wang et al Saccharide composition in atmospheric fine particulate matter 12241

Xie Y L Hopke P K Paatero P Barrie L A and LiS M Identification of source nature and seasonal vari-ations of Arctic aerosol by positive matrix factorizationJ Atmos Sci 56 249ndash260 httpsdoiorg1011751520-0469(1999)056lt0249IOSNASgt20CO2 1999

Yan C Q Zheng M Sullivan A P Shen G F Chen Y JWang S X Zhao B Cai S Y Desyaterik Y Li X Y ZhouT Gustafsson Ouml and Collett Jr J L Residential coal combus-tion as a source of levoglucosan in China Environ Sci Technol52 1665ndash1674 httpsdoiorg101021acsest7b05858 2018

Yan C Q Sullivan A P Cheng Y Zheng M Zhang YH Zhu T and Collett Jr J L Characterization of sac-charides and associated usage in determining biogenic andbiomass burning aerosols in atmospheric fine particulate matterin the North China Plain Sci Total Environ 650 2939ndash2950httpsdoiorg101016jscitotenv201809325 2019

Yang Y Chan C Y Tao J Lin M Engling G ZhangZ Zhang T and Su L Observation of elevated fun-gal tracers due to biomass burning in the Sichuan Basinat Chengdu City China Sci Total Environ 431 68ndash77httpsdoiorg101016jscitotenv201205033 2012

Yttri K E Dye C and Kiss G Ambient aerosol concentrationsof sugars and sugar-alcohols at four different sites in NorwayAtmos Chem Phys 7 4267ndash4279 httpsdoiorg105194acp-7-4267-2007 2007

Yttri K E Simpson D Noslashjgaard J K Kristensen K Gen-berg J Stenstroumlm K Swietlicki E Hillamo R Aurela MBauer H Offenberg J H Jaoui M Dye C Eckhardt SBurkhart J F Stohl A and Glasius M Source apportion-ment of the summer time carbonaceous aerosol at Nordic ru-ral background sites Atmos Chem Phys 11 13339ndash13357httpsdoiorg105194acp-11-13339-2011 2011

Zhang K Shang X N Herrmann H Meng F Mo Z YChen J H and Lv M L Approaches for identifying PM25source types and source areas at a remote background site ofSouth China in spring Sci Total Environ 691 1320ndash1327httpsdoiorg101016jscitotenv201907178 2019

Zhang T Claeys M Cachier H Dong S P Wang W Maen-haut W and Liu X D Identification and estimation of thebiomass burning contribution to Beijing aerosol using levoglu-cosan as a molecular marker Atmos Environ 42 7013ndash7021httpsdoiorg101016jatmosenv200804050 2008

Zhang T Engling G Chan C Y Zhang Y N Zhang Z S LinM Sang X F Li Y D and Li Y S Contribution of fungalspores to particulate matter in a tropical rainforest Environ ResLett 5 024010 httpsdoiorg1010881748-9326520240102010

Zhang Y Obrist D Zielinska B and Gertler A Particulateemissions from different types of biomass burning Atmos Envi-ron 72 27ndash35 httpsdoiorg101016jatmosenv2013020262013

Zhang Y L Wei Q Zhou C P Ding M J and Liu L SSpatial and temporal variability in the net primary production ofalpine grassland on the Tibetan Plateau since 1982 J Geogr Sci24 269ndash287 httpsdoiorg101007s11442-014-1087-1 2014

Zhang Y X Shao M Zhang Y H Zeng L M He L YZhu B Wei Y J and Zhu X L Source profiles of par-ticulate organic matters emitted from cereal straw burningsJ Environ Sci 19 167ndash175 httpsdoiorg101016S1001-0742(07)60027-8 2007

Zhu C Kawamura K and Kunwar B Effect of biomass burn-ing over the western North Pacific Rim wintertime maximaof anhydrosugars in ambient aerosols from Okinawa AtmosChem Phys 15 1959ndash1973 httpsdoiorg105194acp-15-1959-2015 2015

Zhu C S Cao J J Tsai C J Zhang Z S and TaoJ Biomass burning tracers in rural and urban ultrafine par-ticles in Xirsquoan China Atmos Pollution Res 8 614ndash618httpsdoiorg101016japr201612011 2017

Zhu Y Yang L Chen J Wang X Xue L Sui X WenL Xu C Yao L Zhang J Shao M Lu S and WangW Characteristics of ambient volatile organic compounds andthe influence of biomass burning at a rural site in NorthernChina during summer 2013 Atmos Environ 124 156ndash165httpsdoiorg101016jatmosenv201508097 2016

httpsdoiorg105194acp-21-12227-2021 Atmos Chem Phys 21 12227ndash12241 2021