Measurement report: Molecular composition, optical properties, and radiative effects of water-soluble organic carbon in snowpack samples from northern Xinjiang, China

Abstract. Water-soluble organic carbon (WSOC) in the cryosphere has an important impact on the biogeochemistry cycling and snow–ice surface energy balance through changes in the surface albedo. This work reports on the chemical characterization of WSOC in 28 representative snowpack samples collected across a regional area of northern Xinjiang, northwestern China. We employed multimodal analytical chemistry techniques to investigate both bulk and molecular-level composition of WSOC and its optical properties, informing the follow-up radiative forcing (RF) modeling estimates. Based on the geographic differences and proximity of emission sources, the snowpack collection sites were grouped as urban/industrial (U), rural/remote (R), and soil-influenced (S) sites, for which average WSOC total mass loadings were measured as 1968 ± 953 ng g−1 (U),
885 ± 328 ng g−1 (R), and 2082 ± 1438 ng g−1 (S), respectively. The S sites showed the higher mass absorption coefficients at 365 nm (MAC365) of 0.94 ± 0.31 m2 g−1 compared to those of U and R sites (0.39 ± 0.11 m2 g−1 and 0.38 ± 0.12 m2 g−1, respectively). Bulk composition of WSOC in the snowpack samples and its basic source apportionment was inferred from the excitation–emission matrices and the parallel factor analysis featuring relative contributions of one protein-like (PRLIS) and two humic-like (HULIS-1 and HULIS-2) components with ratios specific to each of the S, U, and R sites. Additionally, a sample from site 120 showed unique pollutant concentrations and spectroscopic features remarkably different from all other U, R, and S samples. Molecular-level characterization of WSOC using high-resolution mass spectrometry (HRMS) provided further insights into chemical differences among four types of samples (U, R, S, and 120). Specifically, many reduced-sulfur-containing species with high degrees of unsaturation and aromaticity were uniquely identified in U samples, suggesting an anthropogenic source. Aliphatic/protein-like species showed the highest contribution in R samples, indicating their biogenic origin. The WSOC components from S samples showed high oxygenation and saturation levels. A few unique CHON and CHONS compounds with high unsaturation degree and molecular weight were detected in the 120 sample, which might be anthraquinone derivatives from plant debris. Modeling of the WSOC-induced RF values showed warming effects of 0.04 to 0.59 W m−2 among different groups of sites, which contribute up to 16 % of that caused by black carbon (BC), demonstrating the important influences of WSOC on the snow energy budget.


covers up to 40% of Earth's land seasonally (Hall et al., 1995). Snowfall is a crucial fresh water, the chemical compositions and optical properties of WSOC from this area snowpack is likely 139 different from those reported for the remote regions with more persistent snow coverage. 140 In this study, seasonal snow samples were collected across the northern Xinjiang region  (Pu et al., 2017). The rest of the 158 sites were assigned to R group, most of them were from desert area or barren grasslands located 159 at least ~50-100 km from major cities; hence, they were mostly influenced by natural sources. 160 The S sites are a subgroup of the R group, they correspond to specific locations where the 161 snowpack was visibly patchy and shallow, so local soil could be blown into snow by strong 162 winds. For the S samples, the coarse mineral particles of yellow/brown color were clearly seen 163 on the filters following snow water filtration (Fig. S1), consistent with the expected high 164 loadings of soils at these sites. Out of R group, a sample from site 120 was considered water. Snow depths, snow density, and snow temperature were also measured for each snow 186 pit. All collected samples were then stored in a freezer (< -20 ℃) until further processing.

214
The absorbance at 600 nm was subtracted from the whole spectrum to correct the 215 scattering effects and baseline shifts of the instrument (Chen et al., 2019). The BrC mass 216 absorption coefficients (MAC; m 2 g -1 ) related to WSOC contributions were calculated by: Where λ is the wavelength, A is the base-10 absorbance measured by the spectrophotometer,

236
Moreover, the sum of residual error decreased significantly when the number of components 237 increased from 2 to 3 (Fig. S4). The spectra of derived fluorescent components appeared 238 consistent with those commonly found in other studies (Table S2). for ESI(+/-) modes.

273
The raw experimental data files were acquired by Xcalibur software (Thermo Scientific  Table S3.

315
To evaluate influence of BrC attributed to WSOC on the snow albedo, optical properties   their broad range from 478 to 7069 ng g -1 with an average of 1775±1424 ng g -1 (arithmetic 352 mean ± 1 standard deviation, and same below). The U and the S sites showed higher 353 concentrations with averages of 1968±953 ng g -1 and 2082±1438 ng g -1 , respectively, while 354 the value of R sites (885±328 ng g -1 ) was approximately a factor of two lower (Table S1). Of source for the covered ecosystems.

364
As shown in Fig. 1c, the U sites were also associated with the highest BC concentrations 365 among all four groups (mean: 707±651 ng g -1 ). Furthermore, the mass contributions of sulfate 366 ions at U sites (Table S1, mean: 33%±7%), which is a commonly-used marker for fossil fuel  Table S1). Although WSOC concentrations in R samples were relatively 374 low, they were still higher than most of the values from high-altitude or high-latitude regions higher than those of U (0.39±0.11 m 2 g -1 ) and R (0.38±0.12 m 2 g -1 ) samples, respectively (Table   387   S1). The information on MACBrC related to WSOC in snow and ice is yet very scarce in were 1.32±0.24 m 2 g -1 and 1.02±0.21 m 2 g -1 for U and R samples, respectively (Table S1). The       (Table S1). In addition, the relative intensities of  Table S4). These results suggest terrestrial origin (soil dust) of HULIS-1, which is consistent       In ESI+ mode, CHO+ and CHON+ were the main components in all samples, 539 accounting for 35% to 61% and 20% to 28% of total formulas, respectively (Table 1). 540 The U samples showed the lowest CHO+ abundance (mean: 42%±5%) while the

554
The abundance of CHO-was highest in ESI-with a range of 41% to 61%. The U 555 samples and the site 120 sample showed the lowest (mean: 47%±4%) and highest (61%) 556 fractions of CHO-, respectively. The CHON-and CHOS-compounds account for 557 roughly equal contributions with ranges of 16% to 27% and 14% to 22%, respectively.

558
The detected CHOS compounds were more abundant in ESI-than those in ESI+.

559
Furthermore, CHOS-compounds shows much higher oxidation level and lower  BrC have profound impact on the reduction of the snow albedo in Northern Xinjiang.

769
As shown in Fig. 9, due to the stronger wavelength dependence of BrC absorption, the   anthraquinone structures. Therefore, the special spectroscopic features of BrC from the 842 site 120 sample was attributed to biogenic organics from plants. 843 The RF due to BrC in snow was reported, for the first time, based on the field data