Northwest China is far inland, surrounded by vast deserts and Gobi sandy beaches. Because of the obstruction of the plateau and mountains to the humid air flow, the climate is characterized as being arid with low precipitation. Snow samples were collected from both northern Xinjiang and the northeastern Tibetan Plateau (Figure 1). The sampling site in northern Xinjiang (89°37′E, 46°57′N, 941 m a.s.l.) was located in the eastern end of the Altay region, near the Altay Mountain Glacial Snow and Environment Observation and Research Station at the southern foot of the Altay Mountains and the northern margin of the Junggar Basin. It has a temperate continental climate characterized by hot and dry summers, snowy and cold winters, and large annual temperature variations. Altay has large elevation differences and is adjacent to Russia, Kazakhstan, and Mongolia. The annual average precipitation in Altay is 150–200 mm, which is mainly concentrated in mountainous areas, accounting for more than 80% of the total annual precipitation (Li et al., 2015b). The average snowfall in winter in Northern Xinjiang is 94.6 mm, and that in the Altay region is more than 200 mm (Yang et al., 2017). The sampling site in the northeastern Tibetan Plateau (100°14′E, 38°0′N, 4,103 m a.s.l.) was located in the middle of the Qilian Mountains and the upper reaches of the Heihe River Basin, near the Heihe Remote Sensing Experimental Research Station. Its average altitude is over 3,000 m, the annual average temperature is about 1°C, and the annual precipitation is about 420 mm, belonging to a typical plateau continental climate. The site has the characteristics of wet in summer, dry and cold in winter and a large annual temperature variation. The unique geographical location, complex topography, and diverse climate types in Northwest China make the study of snow and ice chemistry in this region of more important environmental significance and ecological value.

Samples were collected at fixed locations weekly for two snow seasons, from 7 November 2018 to 20 March 2019 and from 9 November 2019 to 29 April 2020. In Heihe and Altay, 53 and 57 surface snow samples and 142 and 186 snow pit samples were collected, respectively (Table 1). The sampling points were selected in an open and flat place far away from cities to avoid human activity (Wang et al., 2015), and the longitude, latitude, and elevation of the sampling points were measured via GPS. Masks and PVC gloves were worn throughout the sampling to prevent contamination. The snow samples were stratified, and one sample was collected every 5 cm from bottom to top. The collected snow samples were placed in a Whirl-Pak Ziplock bag, and the air was discharged cleanly. The volume of each sample was about 800 mL. The samples were at −18°C before processing and analysis.

All snow samples were tested at the Shaanxi Key Laboratory of Earth Surface System and Environmental Carrying Capacity. The snow samples were allowed to melt in an ultra-clean laboratory at room temperature before testing. To avoid contamination from the air, the experiment was carried out immediately after melting. The samples were filtered by a Polyether sulfone (PES) filter with a pore size of 0.22 μm. The pH and electrical conductivity (EC) of the samples were determined by a portable pH meter (Five Easy Plus™, Mettler Toledo, Switzerland) and a conductivity meter (Seven2Go™, Mettler Toledo, Switzerland), respectively. The anions and cations were analyzed by an ion chromatograph (IC, Dionex-AQ, US). The separation of cations and anions was achieved via Dionex IonPac CS12A (4 mm × 250 mm) and Dionex IonPac AS11-HC (4 mm × 250 mm) analytical columns, respectively. In addition, Dionex IonPac CG12A (4 mm × 50 mm) and Dionex IonPac AG11-HC (4 mm × 50 mm) columns were used as guard columns for cations and anions, respectively. The chromatography columns were kept at 30°C during measurements. The mobile phases of the cation and anion IC system were 20 mmol/L methane sulfonic acid and 30 mmol/L potassium hydroxide solution, respectively, all under a flow rate of 1 min/s. Standard single-cation (Ca²⁺, Na⁺, K⁺, Mg²⁺, and NH₄ ⁺, 1,000 mg/L) and single-anion (SO₄ ^2-, NO₃ ⁻, and Cl⁻, 1,000 mg/L) solutions were obtained from National Non-ferrous Metals and Electronic Materials Analysis and Testing Center, Beijing, China. The analytical standard deviation is less than 5%, and the measurement accuracy is 1 ng/g. All samples were analyzed randomly 2–3 times. The test values for the individual samples were then averaged, yielding an estimated error in ion concentration of ≤10%.

Correlation analysis and enrichment factor were considered to be important methods to determine the source of ions in snow and ice (Haeberli et al., 1988; Sundriyal et al., 2018). In order to more accurately determine the source of the elements to be measured, the enrichment factor (EF) method (Okay et al., 2002; Cao et al., 2009) was adopted. The key of this method was the selection of sea-salt tracer ions (Church et al., 1982; Keene et al., 1986; Zhang et al., 2012) with following rules.

(1) If the equivalent ratios of Cl⁻/Na⁺ and Mg²⁺/Na⁺ concentrations are greater than those of standard seawater (Cl⁻/Na⁺ = 1.17, Mg²⁺/Na⁺ = 0.23), then Na⁺ can be used as the sea-salt tracer ion.

Finally, Cl⁻ was selected as the sea-salt tracer ion, and the ratios of different ions in the ocean obtained by Keene et al. (1986) were used to determine. Ca²⁺ was selected as the reference element of crustal source (Taylor, 1964).

To further understand the composition of different sources of various ions, the relative contributions of sea-salt fraction, crustal fraction, and anthropogenic sources fraction were calculated using the following formula: E F s e a w a t e r = X / C l − s a m p l e / X / C l − s e a w a t e r (1) E F c r u s t a l = X / C a 2 + s a m p l e / X / C a 2 + c r u s t a l (2) S S F % = X / C l − s e a w a t e r X / C l − s a m p l e × 100 % (3) C F % = X / C a 2 + s e a w a t e r X / C a 2 + s a m p l e × 100 % (4) A F % = 1 − S S F % − C F % (5) Where X is the ion concentration to be measured. In general, elements with EF significantly higher than 1.00 are considered to be enriched relative to the reference source, otherwise are diluted. SSF, CF and ASF refer to Sea-salt Fraction, Crustal Fraction and Anthropogenic Fraction, respectively.

To further trace the potential source areas and transport paths of chemical ions and dust particles in the snow in the study area, the HYSPLIT (Hybrid Single Practical Lagrangian Integrated Trajectory) model developed by NOAA Air Resources Laboratory (http://www.noaa.gov) was also used in this study. The initial height of the trajectory was chosen to be 500 m above the sampling point, and the backward air mass trajectory was pushed back for 3 days. The meteorological data needed for the model calculation came from the National Center for Environmental Prediction (NCEP) and the National Center for Atmospheric Research (NCAR), and the spatial resolution of the data was 1° × 1°.

To some extent, the chemical information recorded in snow can reflect the level of air pollution and climate conditions (Wang P. et al., 2008). Table 2 shows the concentration of water-soluble inorganic ions, pH, and EC values of snow during two snow seasons in Northwest China. It was found that the pH values of snow in Heihe and Altay in 2018/2019 were 6.44–9.42 and 6.40–7.83, with averages of 7.45 and 6.99, respectively. The pH values of snow in Heihe and Altay in 2019/2020 were 6.06–9.03 and 5.43–8.82, with averages of 7.13 and 6.93, respectively. The pH was higher than previous reported values from different locations, such as 6.74 in Tianjin (Wu et al., 2016), China, 5.50 in the Tianshan Mountains (Li et al., 2010), and 5.90 in snow from Himalayan glaciers (Panicker et al., 2019). Previous studies have shown that Ca²⁺ is a key ion affecting the pH of snow (Li et al., 2007; Rao et al., 2019). Acidic ions in snow are neutralized by alkaline ions such as Ca²⁺ and NH₄ ⁺ (Rao et al., 2019; Liu et al., 2020). It is highly possible that the high pH value of snow in this study was related to the vast saline-alkali soil in Northwest China and frequent high-flux dust weather.

Due to the influence of topography and atmospheric circulation, the concentration of water-soluble inorganic ions in snow varies greatly in different regions. In this study, the concentration sum of cations (Σ⁺) was higher than that of anions (Σ⁻). Spatially, the total ion concentration of snow in Altay was higher than that in Heihe. The lower the elevation of the sampling site, the more strongly it can be affected by dust from nearby sources; and the closer it is to the dust source, the more strongly it can be affected by westerly winds. Temporally, the total ion concentration of snow in 2018/2019 was higher than in 2019/2020. The imbalance between the concentrations of anions and cations indicated that in addition to the inorganic anions analyzed in this study, other anions (e.g., CO₃ ²⁻ and HCO₃ ⁻) also contributed significantly (Kim and Cho, 2003). Previous studies have shown that concentration differences between anions and cations in snow can reflect the concentration values of CO₃ ²⁻ and HCO₃ ⁻, where excess cations may come from the surrounding soil dust (Wake et al., 1993; Li et al., 2019).

The sequencing results (Figures 2, 3) showed that Ca²⁺ was the dominant cation, which may be related to the arid and semi-arid dust area in Northwest China close to central Asia; the order of cation concentrations was Ca²⁺ > Na⁺ > Mg²⁺ > K⁺, which was basically consistent with the order of crustal elements (Ca²⁺ > Na⁺ > K⁺ > Mg²⁺), indicating mainly terrestrial sources. In 2018/2019, SO₄ ²⁻ was the dominant anion in snow in both areas, accounting for 25.25% of the total ion concentration. In 2019/2020, Cl⁻ and NO₃ ⁻ were the main anions in the snow in the Heihe and Altay regions, accounting for 14.7% and 29.02% of the total ion concentration, respectively, and NH₄ ⁺ and NO₃ ⁻ concentrations increased significantly compared with those in 2018/2019. The major ions in snow around the world are different. For example, the main cation and anion in snow from Northwest China (Li et al., 2010; Wang et al., 2016), central Japan (Dong et al., 2011; Kawamura et al., 2012), and the Satopunath Glacier of the Himalayas (Panicker et al., 2019) were Ca²⁺ and SO₄ ²⁻, respectively. Na⁺ and Cl⁻ were the major cations and anions in snow from Wasatch Mountain in Utah, western United States, respectively (Hall et al., 2014).

Upon comparing the ion concentration of the two areas, the concentrations of SO₄ ²⁻ and NO₃ ⁻ in the Altay region were much higher than that in the Heihe region, whereas the concentrations of Cl⁻ and Na⁺ showed the opposite trend. This is because the Altay region is a relatively well-developed region in Xinjiang and even in Northwest China; bordering Russia and other countries, is the main route for the circulation of goods. Compared with the Heihe region, the population density of Altay is higher, and industry is more developed. Coal, coke oven gas, and traffic fuel are important pollution sources in Xinjiang, China (Turap et al., 2019; Yu et al., 2019). When the westerly circulation passes through many petroleum industrial areas and urban agglomeration in Xinjiang, SO₄ ²⁻ and NO₃ ⁻ become enriched in the snow. In contrast, the Heihe region is located in the northeastern Tibetan Plateau, close to the Qaidam Basin and surrounded by many large and small salt lakes such as Qinghai Lake and Hala Lake. The content of Cl⁻ and Na⁺ was high due to the input of salt mineral dust and salt lake materials (Li et al., 2015c; He et al., 2019).

The average NH₄ ⁺ concentration in the Altay region was higher than that in the Heihe region during the two consecutive snow accumulation periods (0.85 mg L⁻¹ > 0.55 mg L^−1; 0.72 mg L⁻¹ > 0.27 mg L⁻¹). Research has shown that NH₄ ⁺ is not only derived from agricultural activities but also from vehicle exhaust emissions (Bo et al., 2019). This is because automobile exhaust catalytic converters produce NH₃ when converting CO and NO_x, and NH₃ is mainly generated in the process of vehicle acceleration (Gandhi et al., 1991; Huai et al., 2003; Burgard et al., 2006). The Altay region is more prosperous than the Heihe region in terms of tourism, industry, and urbanization, so the chemicals in the snow are a direct result of human activities.

In addition to the atmospheric circulation pathway, other factors such as the different sources of water vapor and the post-deposition process of snowpack can also affect the ion concentration in snow pits and surface snow. In this work, the average value of each ion concentration in surface snow was generally lower than that in snow pit (Figure 4). The vertical distribution of ion concentration may be due to the fact that surface snow represents recent precipitation events, whereas snow in snow pits represents precipitation over a longer period of time after atmospheric chemical deposition (Liu et al., 2020).

Determining the sources of ions in Northwest China is complicated due to the obvious differences in topography and population density. In this study, the sources of ions were analyzed by the enrichment factor method. Enrichment Factor (EF) values are given in Table 4, which lists the contributions of sea-salt, crustal, and anthropogenic fractions to each ion. The EF_(seawater) values of all ions were greater than 1, especially Ca²⁺, indicating that ions were enriched relative to seawater and mainly came from soil. The EF_(seawater) and EF_(crustal) values of SO₄ ²⁻ and NO₃ ⁻ were both greater than 1, indicating that they were relatively enriched compared to soil and seawater, and NO₃ ⁻ was enriched to a higher degree. The EF_(crustal) values of K⁺ were all much less than 1, indicating that it was mainly from terrestrial sources, with a small proportion of marine sources. Previous studies have shown that K⁺ is mainly derived from crustal sources in addition to biomass combustion (Jain et al., 2000; Zhang et al., 2007). Upon further calculating the contribution proportion of ion sources (Table 4), the results showed that Cl⁻ and Na⁺ in snow in the study area were mainly imported from sea sources, but also from land sources, which may be related to the abundant salt dust in the surrounding area (Zheng et al., 2008). However, the AF of Mg²⁺ in snow in the Altay region was 29.87%, which may be related to the burning of coal by residents in winter given the high magnesium content in coal (Zhang et al., 2018).

To further understand the source and transformation of water-soluble inorganic ions in snow, a Pearson correlation analysis was conducted for the main inorganic ions in this study (Figure 12). In general, the correlation coefficients of ions in the Heihe and Altay regions were significantly different during the two snow seasons, indicating different sources. Figures 12A, C show the correlation between ions in snowpack in the Heihe and Altay regions in 2018/2019, respectively. The results of the EF method show that SO₄ ²⁻ and NO₃ ⁻ were driven by anthropogenic sources, and considerable portions of Na⁺ and Cl⁻ were from marine sources. The significant correlation among Cl⁻, Na⁺, SO₄ ²⁻, and NO₃ ⁻ indicated that they also had other sources besides anthropogenic and marine ones. Previous studies have shown that an important source of SO₄ ²⁻ in snow is soil dust, followed by human pollution (Kulshrestha et al., 2003; Kumar et al., 2016). NO₃ ⁻ can also be combined with dust to form nitrate substances such as Ca(NO₃)₂ and Mg(NO₃)₂ (Dentener et al., 1996), and Mg²⁺ is also closely related to dust (Kang et al., 2004). Therefore, the input of mineral dust from numerous salt lakes (e.g., Ulungur Lake, Qinghai Lake, and Hala Lake) and saline-alkali land around Heihe and Altay had certain contributions to Na⁺, Mg²⁺, Cl⁻, SO₄ ²⁻, and NO₃ ⁻. However, SO₄ ²⁻ and NO₃ ⁻ showed a more significant correlation than other ions, and their AF values were relatively high. SO₄ ²⁻ may be derived from the combustion of fossil fuels and biomass (Kulshrestha et al., 2005; Kumar et al., 2016), and NO₃ ⁻ may come from the photochemical reaction of NO_x emitted from automobile exhaust (Singh et al., 2014).

Figures 12B, D show the correlation between ions in snow cover in the Heihe and Altay regions in 2019/2020, respectively. As mentioned above, Na⁺, NO₃ ⁻, Cl⁻, Ca²⁺, Mg²⁺, and K⁺ may come from dust input. Compared with other ions, a weak correlation was seen for NH₄ ⁺. Previous studies have shown that NH₄ ⁺ is mainly derived from terrestrial biological processes and alkaline soil release (e.g., agricultural cultivation and animal feces) (Xu et al., 2008). According to the Statistical Bulletin of National Economic and Social Development of Altay Region in 2019, the total sown area of crops in the Altay Region in 2019 was 262,406 ha. Therefore, the significant correlation between NH₄ ⁺ and NO₃ ⁻ in this study indicated that NH₄ ⁺ may be related to agricultural sources, existing in the form of NH₄NO₃ aerosol in snow (Kulshrestha et al., 2005). The weak correlation between SO₄ ²⁻ and other ions in snow in Altay indicated that their sources were different. Given the developed industries in Altay, SO₄ ²⁻ may be related to acidic gas input generated by factory emissions (Kulshrestha et al., 2003).

To understand the potential source of water-soluble inorganic ions in snow, the HYSPLIT model was used to simulate the backward trajectory of the air mass 500 m above the sampling point, and the simulation time was divided into three stages according to the time variation characteristics of ion concentration, namely November to December; January to February; March and April (Figures 13, 14). The results showed that the air mass mainly came from the direction of west to north. In the vertical direction, the air mass mainly showed a gradual sinking movement and was obviously affected by local air mass.

The Heihe region was mainly influenced by air masses from the Taklimakan Desert and the Qaidam Basin in the west. Part of the air mass in the northwest originated from the southern foot of the Ural Mountains. The air mass passed through Kazakhstan and other places, crossed the Tianshan Mountains, and reached the Heihe region. The northeast direction was influenced by the detour flow from the Mongolian Plateau, and the southwest direction was also influenced by the air mass originating from Nepal over the Himalayas. The Altay region was mainly affected by nearby air masses, which was more significant in 2018/2019 compared to 2019/2020. This explained the higher ion concentrations in snow in 2018/2019 than that in 2019/2020. The long-range air masses in the west and northwest originated mainly from central and eastern Kazakhstan and the Russian border, and partly from the southern West Siberian Plain of Russia. In general, air masses passed through arid and semi-arid areas such as deserts and Gobi, carrying large amounts of sand and dust as well as Ca²⁺, Na⁺, Cl⁻, SO₄ ²⁻, and other chemicals. In addition, the wind speed was faster in winter and spring, leading to an increase of dust and ions in the snow. Westerly winds may also carry water vapor from Central Asian salt lakes and deserts, increasing the concentration of salts and alkaline metal ions in the snow. The industries in the densely populated areas of Kazakhstan and southern West Siberia are dominated by coal and oil. When the westerly circulation passed through these places, NO₃ ⁻ and SO₄ ²⁻ continued to be input along the way and were finally stored in the snow.