Quantifying light absorption and its source attribution of insoluble light-absorbing particles in Tibetan Plateau glaciers from 2013-2015

Amounts of insoluble light-absorbing particles (ILAPs) deposited on the surface of snow and ice can significantly reduce the snow albedo and accelerate the snow melting process. In this study, 67 ice samples were collected in 7 high mountain glaciers over the Tibetan Plateau (TP) regions from May 2013 to October 2015. The mixing ratio of black carbon (BC), organic carbon (OC), and mineral dust (MD) was measured using an integrating sphere/integrating sandwich spectrophotometer (ISSW) system associated with the chemical analysis by assuming the light absorption of mineral dust due to iron oxide. The results indicate that mass mixing ratios of BC, ISOC, and MD show a large variation of 10-3100 ng g-1, 10-17000 ng g-1, 10-3500 ng g-1, with a mean value of 218±397 ng g-1, 1357±2417 ng g-1, 241±452 ng g-1 on TP glaciers during the entire ice field campaign, respectively. The chemical elements and the selected carbonaceous particles were also analyzed of the attributions of the particulate light absorption based on a positive matrix factorization (PMF) receptor model. On average, the industrial pollution (33.1%), biomass/biofuel burning (29.4%), and mineral dust (37.5%) were the major sources of the ILAPs in TP glaciers. Although the mineral dust assumed to be the highest contributor to the mass loading of ILAPs, we noted that the averaged light absorption of BC (50.7%) and ISOC (33.2%) was largely responsible for the measured light absorption in the high mountain glaciers at the wavelengths of 450-600 nm.


Introduction
The Tibetan Plateau (TP), known as the highest plateau in the world, and its surrounding areas contain the largest snow and ice mass outside the polar regions (Qin et al., 2006).
Ample evidence has indicated that the greatest decrease in length and area and the most negative mass balance of high glaciers in the TP regions is associated with the deposition 5 of black carbon (BC) over the past decade (Yao et al., 2012;Xu et al., 2009a;Xu et al., 2006).The unusual increase in temperature over the TP is now considered one of the major contributors to glacial shrinkage (Ding et al., 2006).Climate models indicated that BC heats the troposphere by absorbing solar radiation (Jacobson, 2001;Jacobi et al., 2015), and BC reduces snow and ice albedos when it is deposited on their surface, thus 10 leading to the acceleration of snowmelt (Hadley and Kirchstetter, 2012;Hansen and Nazarenko, 2004;Yasunari et al., 2015;Flanner et al., 2009;Flanner et al., 2007).For example, a mixing ratio of 10 ng g -1 of BC in snow can reduce snow albedo by 1%, which has a similar effect to that of 500 ng g -1 of mineral dust on the albedo of snow and ice at 500 nm wavelength (Warren and Wiscombe, 1980;Warren, 1982;Wang et al., 15 2017).Bond et al. (2014) indicated that the best estimate of climate forcing from BC deposition on snow and sea ice in the industrial era is +0.13 W m -2 with 90% uncertainty bounds of +0.04 to +0.33 W m -2 .In addition to BC, organic carbon (OC) and mineral dust (MD) also substantially contribute to springtime snowmelt and surface warming through snow 20 darkening effects (Yasunari et al., 2015;Painter et al., 2012;Painter et al., 2010;Kaspari et al., 2014;Wang et al., 2014;Wang et al., 2013;Huang et al., 2011).Water-soluble organic carbon (WSOC) and insoluble organic carbon (ISOC) are the major components of organic carbon in the atmosphere, snow and sea ice.In addition to the strong light absorption of ISOC and WSOC may also influence the regional and global climate 25 through the heating and evaporating of clouds and by acting as cloud condensation nuclei (CCN) (Chen and Bond, 2010;Witkowska and Lewandowska, 2016;Alexander et al., 2012).Moreover, WSOC also plays a key role in affecting human health due to its toxic effects (McConnell and Edwards, 2008;Wang et al., 2015).Although the mass mixing ratio of insoluble organic carbon in the snow and ice has been widely investigated in 30 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-86Manuscript under review for journal The Cryosphere Discussion started: 25 May 2018 c Author(s) 2018.CC BY 4.0 License.
previous studies, there are still limited studies that measure the mass mixing ratios of both WSOC and ISOC in snow/ice samples, especially across the TP regions.
Due to the importance of the climate effects by ILAPs, numerous snow surveys have been conducted to investigate the light absorption of ILAPs and their potential source attribution in snow (Clarke and Noone, 1985;Doherty et al., 2010Doherty et al., , 2014;;Hegg et al., 5 2010;Huang et al., 2011).For instance, Hegg et al. (2009) found out that the BC attribution in snow are originated from two distinct biomass burning sources, a pollution source, and a marine source obtained at 36 sites in Alaska, Canada, Greenland, Russia, and the Arctic Ocean in early 2007.Huang et al. (2011) conducted the first snow survey over northern China, and the sources of ILAPs in seasonal snow in the region were 10 explored based on a positive matrix factorization (PMF) with backward trajectory cluster analysis (Zhang et al., 2013a).Wang et al. (2013) indicated that soil dust was found to be the major contributor to snow particulate absorption in Inner Mongolia regions and Qilian mountains over northern China.Recently, vertical profiles of ILAPs in seasonal snow were performed from 67 north American sites, and the sources of particulate light 15 absorption were explored based on the chemical and optical data (Doherty et al., 2014).
However, the assessments of the light absorption and its emission sources of ILAPs on the TP glaciers are sparse due to limited observations.Here, we present a snow survey on collecting the snow/ice samples on 7 high mountain glaciers on the TP regions from 2013-2015.By using an integrating sphere/integrating sandwich spectrophotometer 20 (ISSW) system associated with the chemical analysis, the particulate light absorption of BC, ISOC, and MD in TP glaciers was evaluated.Finally, the relative attribution of emission sources of the ILAPs in these regions was explored based on a positive matrix factorization (PMF) receptor model.and Qiumianleiketage glaciers are also located over the Kunlun Mountains, and these glaciers have an average altitude of 5100 m and 5500 m a.s.l, respectively.The Meikuang glacier is located in the eastern Kunlun Mountains, where is characterized by alluvial deposits and sand dunes (Xiao et al., 2002).The glaciers of Hariqin and Meikuang have similar altitudes but are from different mountains (Li et al., 2016).The Yangbajing glacier 10 is located on the south-eastern margin of the Nyenchen Tanglha Mountains, and seated about 90 km northwest of Lhasa, the capital city of Tibet (Liang et al., 1995).To investigate the enrichment of ILAPs via wet and dry deposition on high glaciers, ~67 vertical profiles of snow/ice samples with seasonal transitions were obtained at 7 high mountain sites from May 2013 to October 2015 (Fig. 1).The collected snow/ice samples were preserved in 0.5-m pure, clean tubes and kept frozen at the State Key Laboratory of Cryospheric Sciences, Cold and Arid Regions Environmental and Engineering Research Institute in Lanzhou.Then, each snow/ice sample was cut vertically into small pieces from the surface to the bottom.Therefore, approximately 189 pieces of the snow/ice samples were analyzed in this study.20

Optical analysis
An integrating sphere/integrating sandwich spectrophotometer (ISSW) instrument that developed by Grenfell et al. (2011) was used to measure the mass mixing ratio of BC in snow by Doherty et al. (2010Doherty et al. ( , 2014) ) and Wang et al. (2013a).By assuming the major light absorption of mineral dust is due to iron oxides (e.g., hematite and goethite, 25 hereinafter, simply "Fe"), it is possible to evaluate the absorption properties of OC by combining the chemical analysis and optical method (Zhou et al., 2017;Doherty et al., 2014).exponent of all ILAPs (Åtot), as described by Doherty et al. (2010Doherty et al. ( , 2014)).These parameters are defined as follows: 1. C BC max (ng g -1 ): maximum BC is the maximum possible BC mixing ratio in snow by assuming all light absorption is due to BC at the wavelengths of 650-700 nm.
2. C BC est (ng g -1 ): estimated BC is the estimated snow BC mixing ratio derived by 5 separating the spectrally resolved total light absorption.

C BC
equiv (ng g -1 ): equivalent BC is the amount of BC that would be needed to produce absorption of solar energy by all insoluble particles in snow for the wavelength-integrated from 300-750 nm.
4. Åtot: absorption Ångström exponent is calculated for all insoluble particles deposited on 10 the filter between 450 and 600 nm.
5. Ånon-BC: non-BC absorption Ångström exponent is defined as the light absorption by non-BC components of the insoluble particles in snow between 450-600 nm.
6. f − est (%): fraction of light absorption by non-BC light absorbing particles is the integrated absorption due to non-BC light absorbing particles, which is weighted by the 15 down-welling solar flux from snow at the wavelengths of 300-750 nm.

Chemical analysis
Previous studies on these parameters have concluded that ILAPs are primarily derived from BC, OC, and Fe.We assume that the mass absorption coefficients (MACs) for BC, 20 OC, and Fe are 6.3, 0.3, and 0.9 m 2 g -1 , respectively, at 550 nm and that the absorption Ångström exponents (Å or AAE) for BC, OC, and Fe are 1.1, 6, and 3, respectively (Grenfell et al., 2011;Doherty et al., 2010Doherty et al., , 2014;;Wang et al., 2013).Meanwhile, to quantify WSOC, about 10 ml of the filter liquor was injected into a total carbon analyzer (TOC-V, Shimadzu).The method detection limit (MDL) used was 4 μg/l with a 25 precision of ±5% (Cong et al., 2015).The definition of TOC (Total Organic Carbon) in this study is calculated as the total WSOC measured by carbon analyzer and the ISOC calculated by ISSW instrument.
Meanwhile, for the filtrated snow/ice samples, we measured the major anions (Cl -, NO 2 -, NO 3 -, SO 4 2-) and cations (Na + , NH 4 + , K + , Mg 2+ , Ca 2+ ) with an ion chromatograph using a CS12 column for cations and an AS11 column for anions at the Institute of Tibetan Plateau Research in Beijing.All the detection limit of the ions was 1 μg/l.In 10 addition, except for the anions and cations and trace elements, CL salt , MD and biosmoke K (K Biosmoke ) were determined to assess the mass contributions of the major components in the snow/ice samples.CL salt was estimated as follows in accordance with Pio et al. (2007) Where (Na/Al) Crust =0.33 , and represents the Na/Al ratio in the dust materials (Wedepohl, 1995).With 0.12, 0.038, 0.038, and 0.25 being the mass rations in seawater 20 of magnesium to sodium, calcium to sodium, as well as potassium to sodium and sulfate to sodium, respectively.
The MD content was calculated by a straightforward method, and the Al concentration in dust was estimated at 7% (Zhang et al., 2013b): MD=Al/0.07 (3) 25 We determined K Biosmoke as follows (Pu et al., 2017): Where (K/Al) Crust is 0.37 and represents the K/Al ratio in the dust materials (Wedepohl, 1995) and Na Ss is estimated by Eq. ( 2).

Enrichment factor (EF)
To evaluate the relative contributions of trace elements from natural (e.g., mineral and 5 soil dust) versus anthropogenic sources, an inter-annual comparison of EF c values, which represent the enrichment of a given element relative to its concentration in the crust of the earth.The primary uncertainty in these calculations is attributed to the differences between chemical compositions in the snow and the reference crustal composition.The EF c is defined as the concentration ratio of a given metal to that of Al, 10 which is a reliable measure of crustal dust, normalized to the same concentration ratio characteristic of the upper continental crust (Wedepohl, 1995), calculated with the following equation:

Source apportionment
The Positive Matrix Factorization (PMF 5.0) is considered as a generally accepted receptor model to determine source apportionment of the ILAPs when source emission profiles are unknown (Paatero and Tapper, 1994).Details of the PMF procedure used in this study are also similar to the previous work as discussed in Hegg et al. (2009Hegg et al. ( , 2010)).20 Generally, the mass concentration of the chemical species and the uncertainty were used as the input.The final data set used for the PMF analysis contained 189 samples with 18 elements whereby only elements that have high recovery were used.The uncertainty value of each variable in each sample estimated from an empirical equation.The PMF model was run for 3 to 6 factors with 6 random seeds, but only a three-factor solution of 25 the ILAPs in TP glaciers could provide the most meaningful results.Q values (modified values) for the 3-factor solution (both robust and true) were closest to the theoretical Q value of any of the factor numbers for which the model was run, suggesting that the 3-factor solution was optimal.

Regional averages
Over 67 snow/ice samples were collected at 7 sites from 2013 to 2015 across the Tibetan plateau field campaign.Each vertical snow/ice sample was cut into several pieces.The , Åtot, and Ånon-BC of the snow/ice samples are given in Table 1 for each glacier.The lower median values of C BC est could be found in the Tanggula, Hariqin, and Yangbajing glaciers on the south of the TP regions, while the other glaciers shows a relative higher range (94-172 ng g -1 ) on the north edge of the TP regions.Details of the vertical profiles of all snow/ice samples collected in each 10 site could also be found in Table S1.During the field campaign from 2013-2015, the lowest concentration of BC in the snow/ice samples is found in the Tanggula glacier, with a value of C BC est ~10 ng g -1 .In contrast, the highest values of C BC est , C BC max , and C BC equiv are 3100 ng g -1 , 3600 ng g -1 , and 4700 ng g -1 , respectively, taken in the Yangbajing region.
We note that there are no apparent differences in ILAPs between the cold and warm 15 seasons in high glacier regions.
It is well known that the aerosol composition and the size distribution are key parameters that affect the absorption Ångström exponent.Doherty et al. (2010) reported that the value of the absorption Ångström exponent of OC was close to 5, which is consistent with previous studies with values ranging from 4-6 (Kirchstetter et al., 2004).Several 20 studies indicated that the absorption Ångström exponent of mineral dust ranged from 2 to 5 (Fialho et al., 2005;Lafon et al., 2006).The variation in the absorption Ångström exponents for urban and industrial fossil fuel emissions is typically in the range of 1.0-1.5 (Millikan, 1961;Bergstrom et al., 2007), which is slightly lower than that of biomass burning, which primarily falls in the range of 1.5-2.5 (Kirchstetter et al., 2004;Bergstrom 25 et al., 2007).Although the source attribution of the insoluble light-absorbing particles in the samples is not a dominant determinant of the value of the absorption Ångström exponent, fossil fuel burning may have a lower absorption Ångström exponent (<2) than 2-5 (Fialho et al., 2005;Millikan, 1961).Generally, the median of the absorption Ångström exponent for total particulate constituents (Åtot) exceeds 1.0 at all locations (Fig. 30 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-86Manuscript under review for journal The Cryosphere Discussion started: 25 May 2018 c Author(s) 2018.CC BY 4.0 License.
2).As shown in Fig. 2a, the lowest median value of Åtot (~2.1) is found in the Tanggula glacier, while the other glaciers exhibit much higher values (2.5-2.9).The results indicated that the emission of the ILAPs in the Tanggula glacier likely originated from the combustion sources, which is also consistent with the previous studies (Bond et al., 1999(Bond et al., , 2001;;Bergstrom et al., 2007;Schnaiter et al., 2003Schnaiter et al., , 2005;;Clarke et al., 2007).5 Except Hariqin glacier, the other glaciers show an increased trend of the absorption Ångström exponent for non-BC particulate constituents (Ånon-BC) from the south to north regions in the TP regions (Fig. 2b).Åtot and Ånon-BC for all snow/ice samples were in the range of 1.4-3.7 and 1.9-5.8,respectively (Table S1).The lower absorption Ångström exponent (Åtot <2) found in the Meikuang (site 30), Tanggula (sites 51, 56, 57, 58) and 10 Yangbajing glaciers (site 65) suggested that the sites were primarily influenced by fossil fuel emission, whereas the other sites were heavily influenced by soil dust and biomass burning.Another notable feature is that a lower Ånon-BC suggests a higher percentage of mineral dust in all snow/ice samples, while a higher Ånon-BC reflects that the non-BC ILAPs in snow and ice were mainly dominated by OC (Wang et al., 2013;see Eq. 3).15 Histograms of the absorption Ångström exponent by region are shown in Fig. 3.In the Yuzhufeng and Tanggula glaciers, there is a large variation of the absorption Ångström exponent (~1-4), reflecting that the ILAPs are not only dominated by BC in these regions but also influenced by non-BC absorbers such as OC and mineral dust.In contrast, a common feature in the other regions is that they show a less variable of the absorption 20 Ångström exponent, ranging from 2.5-3.the depth of these samples ranged from 15-45 cm (Table S1).As shown in Fig. S2, most values of C BC est in this region range from ~100-1000 ng g -1 , with a few values lower than 100 ng g -1 .One notable feature is that the highest concentrations of C BC equiv and C BC max for the surface layer are 2600 ng g -1 and 1600 ng g -1 , respectively, at site 41. while the concentration of ISOC and Åtot are ~8600 ng g -1 and 3.41.Based on the non-BC fraction 20 ranging from ~20-670 ng g -1 and with a median value of 130 ng g -1 .The median value of the mass concentration of ISOC is ~600 ng g -1 in the Meikuang glacier.The fraction of total particulate light absorption due to non-BC constituents is typically ~16-62%, and Ånon-BC (5.12) in this region is highly similar to that found in the Yuzhufeng glacier (5.06).
This suggests that the ILAPs have similar emission sources in the Yuzhufeng and 5 Meikuang glaciers.In addition, there appears to be no significant difference in the mixing ratios of ILAPs in the snow/ice samples via dry and wet deposition in these glaciers between the cold and warm seasons (Fig. S2, S3, S4, S5, S6).
In the Qiyi glacier (Fig. S4), the C BC est are much similar than those in the Meikuang glacier, with values ranging from ~20-720 ng g -1 , which does not include the highest 10 value of 1900 ng g -1 at site 13.The fraction of total particulate light absorption due to non-BC constituents f non-BC est is typically ~20-70%, with a median value of 41%.This information along with the lower Åtot (2.6) indicates that BC plays the dominant role in influencing the light absorption in this region.The vertical profiles of ILAPs in the snow/ice samples were collected in the warm season from 2014 to 2015.In the warm 15 season, the mixing ratios of ISOC and Fe ranged from 80-10100 ng g -1 and 20-340 ng g -1 , respectively, and the mixing ratios of ILAPs in most of the snow/ice samples increased remarkably from the top to the bottom.This result is highly consistent with a previous study by Doherty et al. (2013).Fig. S5 shows that the vertical profiles of the mass mixing ratios of BC, ISOC, and Fe for the snow/ice samples in the Tanggula glacier were more 20 complicated than those for the other regions.With the exception of the surface layer at sites 53 and 54, most values of C BC est ranged from 10 to 280 ng g -1 in the Tanggula glacier; therefore, this glacier was the cleanest region of all the studied glaciers.At sites 52-54, a notable feature is that the surface mixing ratios of C BC est are significantly larger than those in the sub-surface layers, possibly because of the accumulation of BC via dry deposition 25 on the snow/ice surface.Doherty et al. (2013) also found that the ILAPs could be scavenged with the snow meltwater, therefore leading to a much higher concentration of BC in the surface snow.At sites 56-58, f non-BC est was lower than 38%, and Åtot ranged from 1-2.5.These results are consistent with the fossil fuel combustion source due to industrial activities.Because single layer samples are not shown, the vertical profiles of 30 S6i, and S6h, the other sites reveal the trapping and scavenging effects of a higher mass concentration of BC in the surface layer due to the melting process.Previous studies have also illustrated that the ILAPs could become trapped and integrated at the surface of the 5 snowpack due to melting and sublimation to enrich the surface concentrations (Conway et al., 2002;Doherty et al., 2013;Painter et al., 2012).

Relationship between BC and ISOC
The largest contribution to both BC and OC emissions is from biomass burning, but 10 biofuel and fossil fuel burning play key roles in influencing the mass concentration of BC and OC (Bond et al., 2004(Bond et al., , 2006;;Chen et al., 2010;Akagi et al., 2011).The OC/BC ratio could also represent the relative abilities of aerosol light scattering and absorption; therefore, this ratio is an indication that the origins of the carbonaceous particles are due to different emission sources during the transport and deposition processes (Chow et al., 15 1996;Turpin and Huntzicher, 1995).For instance, Bond et al. (2004) revealed that the higher ratio of OC/BC (7-8) was mainly due to open biomass burning and biofuel combustion rather than fossil fuel combustion.Koch (2009) revealed that an OC/BC ratio of ~4.0 was assumed to indicate emission sources due to fossil fuel combustion, while an OC/BC ratio higher than 12 was due to residential coal combustion (Cao et al., 2005).20 Further studies have also been performed to analyze the variations in the abundance of OC and BC among sources in the atmosphere such as coal, diesel, gasoline, and wood combustion (Watson et al., 1994(Watson et al., , 2001)).However, only few studies have revealed the mass mixing ratios of OC and BC in the Loess Plateau, especially in the high glacier regions via dry and wet deposition.As illustrated in Fig. 5, the ISOC/BC ratios range 25 from 1 to 18 during the warm season and from 0.05 to 18.4 during the cold season, with median values of 6 and 5, respectively.High ISOC/BC ratios (>3) in all TP glaciers mainly revealed the primary emissions of carbonaceous aerosols due to fossil fuel and biofuel burning.The highest ISOC/BC ratio (18.4) found at site 52 resulted that the emission sources is mainly due to residential coal combustion.Only slight differences were found in the ISOC/BC ratios (6 and 5) between warm and cold seasons, respectively.
The lowest values for the ISOC/BC ratios (<2) varied substantially at several sampling sites, suggesting that the natural combustion sources were the dominant factor.In order to better illustrate the spatial distribution of the mixing ratio of the ISOC/BC in snow/ice, the median values of the mixing ratio of ISOC/BC in each glacier is given in Fig. S7.The 5 result indicates that the relative lower mixing ratios of ISOC/BC are found in the southern glaciers, while much higher in the northern glaciers.Furthermore, the ISOC concentration exhibited a strong correlation with BC (y=4.350x+343.80,R 2 =0.54) (Fig. 6), which suggested influences from similar emission sources (e.g., biomass burning and residential and commercial coal combustion).10

Water-soluble organic carbon
It is well known that the WSOC component has a large variation ranging from 20% to 99% of total carbonaceous particles (Mayol-Bracero et al., 2002;Saxena and Hildemann, 1996).There is a very large fraction representing the average WSOC (>80%) to TOC, 15 which can absorb solar light and enhance cloud formation through their direct and indirect climate effects.These results are highly consistent with previous studies showing that fossil fuel combustion plays a key role in leading to the higher fraction of WSOC to total organic carbon (TOC) (Andreae and Rosenfeld, 2008).The concentration of WSOC during the field campaign varied from 400 to 43600 ng g -1 , with a median of 1400 ng g -1 20 (Table 1).The median values of the WSOC/TOC ratio were 0.85 for the warm season and 0.9 for the cold season during this field campaign (Fig. 7).No remarkable correlation was observed between WSOC and ISOC in both cold and warm seasons.As WSOC is considered as a stable indicator of the primary biomass burning, with a small contribution of fossil fuel combustion, we indicate that a mixture emission sources from the biomass 25 burning and fossil fuel combustion might be contributing significantly to the TOC concentrations in these TP glacier regions.

Contributions to particulate light absorption by ILAPs
The Cryosphere Discuss.BC and OC were mainly emitted from biomass burning and biofuel and fossil fuel combustion, while mineral dust was emitted from the local soil or desert regions (Bond et al., 2013;Kaspari et al., 2014;Painter et al., 2007;Chen and Bond, 2010;Streets et al., 2001;Pu et al., 2017).The contributions to particulate light absorption by BC, OC, and Fe have been investigated using the ISSW measurements across northern China and 5 North America (Wang et al., 2013;Doherty et al., 2014).The fractional contributions to absorption by BC, ISOC, and Fe at 450 nm in surface high glaciers are shown in Fig. 8, and the concentrations of BC, ISOC, and Fe are given in Table S1.BC plays a dominant role in particulate light absorption with values ranging from 20-87% in all high glacier regions.ISOC is the second highest absorber in high glacier regions, and there are large 10 variations of light absorption of ISOC during the field campaign with values ranging from 0.5-58%.The light absorption due to ILAPs in the TP glacier regions is not only from industrial and biomass burning but also with a small contribution from local soil dust.Note that the median fraction of light absorption due to Fe is ~13%, with the highest light absorption of iron being higher than 40% in the Yangbajing glacier.This result is an 15 indication that mineral dust plays a key role in affecting the spectral absorption properties due to the soil and mineral dust at site 67.Although the total light absorption was dominated by BC and ISOC in all selected glaciers, we note that the relative spatial distribution of the total light absorption due to Fe (>17%) also play key roles to affected the snowmelt in the southern glaciers than that of northern glaciers across the TP regions 20 (Fig. S8).

Enrichment factor (EF)
Briefly, EF values ranging from 0.1 to 10 indicate significant input from crustal sources.
Conversely, EF values that larger than 10 exhibit a major contribution from 25 anthropogenic activities.Referring to the EF analysis (Fig. 9), the mean EF of Fe < 5 in each glacier can be assumed to customarily originate from crustal sources.Recent studies have also indicated that light-absorbing particles in snow are dominated by local soil dust in some typical regions over northern China (Wang et al., 2013), and northern America (Doherty et al., 2014).Comparable with Fe, the other trace metals with the mean EF of 30 ≥5.0 were moderately to highly enriched predominantly from anthropogenic emissions (Hsu et al., 2010).For example, Pacyna (2001) reported that fossil fuel combustion is a major source of Cr.Cu primarily originates from emissions from fossil fuel combustion and industrial processes, while Pb and Zn are known to be drawn from the traffic-related activities and coal burning (Contini et al., 2014;Christian et al., 2010).Together with 5 high EF values observed for Cu, Zn, and Cd in our snow/ice samples clearly suggested that the TP glaciers have already been polluted by human activities, such as biomass burning, fossil fuel burning, and the coal burning.For instance, high level of Pb and Cr have demonstrated a link to coal combustion (Mokhtar et al., 2014;Zhang et al., 2013aZhang et al., , 2013b)), abundant Cu and Cd were associated with traffic-related dusts (Cheng et al., 10 2010).Therefore, we concluded that the natural dust source and anthropogenic emission source are both non-negligible to the ILAPs in the TP glaciers.

Source apportionment
Given the importance of the climate effect in our understanding of the ILAPs in TP 15 glaciers, we present a PMF receptor model to analyze the source attribution of ILAPs in these glaciers.In this study, two datasets including the mass concentrations of the chemical components and the ILAPs in snow/ice and the associated uncertainty datasets were used to run the PMF 5.0 model.The details of the techniques have already been illustrated by Hegg et al. (2009Hegg et al. ( , 2010) ) and Pu et al. (2017).The factor loadings 20 (apportionment of species mass to individual factors) for the 3-factor solution of the source profiles based on the PMF 5.0 model are given in Fig. 10 (in both measured mass concentration and the % total mass allocated to each factor).It is evident that the first factor (top panel) was obviously characterized by high loadings of Cl − , CL salt , SO 4 2-, and NO 3 -, which are well known markers for the urban or local industrial pollutions 25 (Alexander et al., 2015).Although Cl − to Na + are usually considered as a potential product of emission source of sea salt, but also a high loading of Cl − to CL salt indicated another source in addition to sea salt such as industrial emission and coal combustion (Hailin et al., 2008;Kulkarni, 2009).Additionally, the highest loading of NH 4 + is also suggested as an indicator of coal combustion (pang et al., 2007).Compared with the first factor, the highest loading of Al (90.3%) and Fe (87.3%) are well-known markers for the urban or regional soil dust (Pu et al., 2017).Therefore, the second factor or source profile is easily interpretable as a natural soil dust source.But a notable feature is that the relative high mass loading of C BC max (76.4%) to that of the previously identified soil dust source as reported by Pu et al. (2017).It is well known that K + and K Biosmoke are the 5 major indicators of biomass burning source (Zhang et al., 2013a).Therefore, it is easily interpretable that the highest loadings of K + and K Biosmoke are well representative the biomass burning source (Fig. 10c).However, it is also important to note that the lowest mass loading of C BC max in this factor is a bit unexpected.Indeed, the C BC max is not only attributed to the biomass burning emission, but also associated with the industrial 10 activities associated with the local soil dust (Bond et al., 2006).Therefore, we interpreted the third factor normally considered a predominantly biomass burning product.However, the major emission of BC in TP glaciers originated from the local soil source instead of the biomass burning and industrial pollution than previous studies (Pu et al., 2017;Zhang et al., 2013a).15 Finally, the chemical composition and mean source apportionment of the ILAPs to the three sources in the TP glaciers were given in Fig. 11.Note that the apportionment is of the light absorption by insoluble particles in the surface glaciers.On average, the source appointment of the ILAPs in all TP glaciers by soil dust is close to 37.5%, while the industrial emission and biomass burning contributes 33% and 29.4%.Specifically, the 20 largest biomass burning contribution of the light-absorption of ILAPs was found in the Qiyi glacier, which is close to the human activity regions (Guan et al., 2009).In the Meikuang, Qiumianleiketage, Yangbajing, and Tanggula glaciers, the soil dust contribution of light absorption is much larger (>47.9%)than that of industrial pollution and biomass burning, especially in the Meikuang glacier.In these regions, the percent of 25 the MD light absorption is ranging from 20.4-31.1%,while the light absorption by biomass burning is in the range of 18.5-35.8%.Industrial pollution constitutes a major fraction in the Yuzhufeng glacier.Chemical analysis shows that the percentages of the chemical species in the Yuzhufeng and Meikuang glaciers are much similar.The attribution of the total anions by chloride, nitrate, and sulphate is higher 52% and 48% 30 than the other chemical species in the Yuzhufeng and Meikuang glaicers.In the Hariqin glacier, the largest attribution of the sulphate is up to 45.4%.As shown in Fig. 11, the source apportionment of the light absorption by insoluble particles in the surface glaciers is dominated by soil dust and the industrial pollution in most glaciers, only with a large fraction of the light absorption due to biomass burning in the Yuzhufeng glacier.These 5 results are highly consistent with the previous studies (Andersson et al., 2015;Li et al., 2016).They found that the contributions of coal-combustion-sourced BC are the most significant for the TP glaciers.Based on the model simulations, Zhang et al. (2015) revealed that the largest contribution to annual mean BC burden and surface deposition in the entire TP regions is from biofuel and biomass (BB) emissions in South Asia, followed 10 by fossil fuel (FF) emissions from South Asia.

Conclusions
In this study, the ILAPs observations in 7 high glacier regions across the Tibetan Plateau are presented using the ISSW technique along with chemical analysis.Approximately 67 15 vertical profiles of snow/ice samples are analyzed during the warm and cold seasons from 2013-2015.There are no apparent differences in the mixing ratios of ILAPs in the snow/ice samples during seasonal transitions.The results indicate that the variations of Åtot and Ånon-BC for all snow/ice samples range from 1.4-3.7 and 1.9-5.8,respectively, excluding site 16.The lower absorption Ångström exponent (Åtot <2) suggested that the 20 sites 30, 51, and 56-58, 65 were primarily influenced by fossil fuel emission, whereas the rest of the sites were heavily influenced by mineral dust and biomass burning.Another notable feature is the large variation of the ILAPs in the snow/ice samples.By excluding some of the highest ILAPs values in the snow/ice samples, the values of C BC est , ISOC C , and Fe C range from 100-1000 ng g -1 , 10-2700 ng g -1 , and 10-1000 ng g -1 , respectively.

25
Among the samples, the lower concentrations of BC were found in the Tanggula, Hariqin and Yangbajing glaciers, with the median concentrations of 33 ng g -1 , 24 ng g -1 , and 28 ng g -1 , respectively.The large contribution of high ISOC/BC ratios (>3) at all high glacier sites mainly indicated that the primary emissions of carbonaceous aerosols were due to fossil fuel and biofuel burning.30 BC and ISOC play dominant roles in particulate light absorption, with values ranging from 20-87% and 0.5-58%, respectively, in all high glacier regions.However, the relative spatial distribution of the total light absorption due to Mineral dust also plays key roles to affect the snowmelt in the TP glaciers.In this study, we also present a PMF receptor model to analyze the source attributions of ILAPs in these glaciers.We found that the 5 anthropogenic air pollution is much heavy across the northern glaciers due to human activities, but the major emissions of the light absorption by insoluble particles in TP glaciers originated from the local soil and industrial pollution sources, followed by the biomass burning source.We note that the C BC max is not only attributed to the biomass burning emission, but also associated with the industrial activities associated with the 10 local soil dust.Therefore, the natural dust source and anthropogenic emission source are both non-negligible to the ILAPs in the TP glaciers.
As shown in Fig.1, the spatial distribution of aerosol optical depth (AOD) retrieved from Moderate-resolution Imaging Spectrometer (MODIS) sensors are ranging from 0.1 to 0.4 from south to north near the high mountain glacier regions over TP regions.The Qiyi 30 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-86Manuscript under review for journal The Cryosphere Discussion started: 25 May 2018 c Author(s) 2018.CC BY 4.0 License.glacier (39°14' N, 97°45' E) is located in the eastern part of the TP above the equilibrium line altitude (ELA) at 4130 m a.s.l.The Tanggula glacier (33°04' N, 92°04' E) is located at 5743 m a.s.l. in the central Qinghai-Tibetan Plateau, and the average snowline in the Tanggula glacier is 5560 m a.s.l.The Yuzhufeng glacier (35°38' N, 94°13' E) is the highest peak across the Kunlun Mountains, with an elevation of 6178 m.The Meikuang 5
Fig. 4c.The Yuzhufeng glacier is located in the northern part of the Loess Plateau, close to the Meikuang glacier.Twelve snow/ice samples were collected in the Yuzhufeng glacier, and 15 of the light absorption (f non-BC est ) is 0.56, we pointed out that the light absorption in the surface glacier at site 41 wasn't only influenced by BC, but also possibly related to the ISOC, and MD.The large variation of the absorption Ångström exponent distribution is an indication of the complicated emission sources of ILAPs in the snow/ice samples.In the Yuzhufeng glacier, Åtot generally varied between ~2 and 3.7, and the average value of 25 f non-BC est is close to 50%, so these results also reveal that the ILAPs in snow/ice samples are heavily influenced by anthropogenic air pollutants.Large variations of ISOC (measured by ISSW) and WSOC are also observed, with values ranging from ~10-17000 ng g -1 and ~410-43000 ng g -1 , respectively.Except for site 23, the values of C BC est in the Meikuang glacier are much lower than those in the Yuzhufeng glacier, with values 30 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-86Manuscript under review for journal The Cryosphere Discussion started: 25 May 2018 c Author(s) 2018.CC BY 4.0 License.

The
Cryosphere Discuss., https://doi.org/10.5194/tc-2018-86Manuscript under review for journal The Cryosphere Discussion started: 25 May 2018 c Author(s) 2018.CC BY 4.0 License.C BC est are plotted in Fig. S6 for all snow/ice samples, which were collected in the Qiumianleiketage, Hariqin, and Yangbajing glaciers.Except for the sites in Fig. S6d, S6e,

Figure 3 .
Figure 3. Histograms of the frequency of Å tot (450-600 nm) for snow/ice samples in each of the glacier region.Samples from all vertical profiles are included.

Figure 4 .
Figure 4. Box plots of the region variations in (a) BC concentration, (b) ISOC concentration, and (c) Fe concentration of the seven glaciers.QY, QM, MK, YZF, HRQ, TGL, and YBJ represent the following glaciers: Qiyi, Qiumianleiketage, Meikuang, Yuzhufeng, Hariqin, Tanggula, Yangbajing glaciers, respectively.Error bars are 10 th , 25 th , median, 75 th , and 90 th percentiles of the data.The dot symbol represents the average concentrations of the ILAPs in snow/ice sample in each glacier.

Figure 5 .Figure 6 .
Figure 5.The mass mixing ratios of ISOC and BC for each surface snow/ice sample.

Figure 7 .
Figure 7.The mass fraction of WSOC to TOC of each snow/ice sample.

Figure 8 .
Figure 8. Relative contributions to total light absorption by BC, ISOC, and Fe oxide (assumed to be in the form of goethite) for surface snow/ice samples in each sampling site (See Table1).

Figure 9 .Figure 10 .
Figure 9. Average enrichment factors of trace metals in surface snow/ice samples at each region.

Figure 11 .
Figure 11.Chemical composition and source apportionment for the seven glaciers in the TP regions.Note that the apportionment is of the light absorption by insoluble particles in the surface glaciers.