The mass concentrations of light-absorbing snow impurities at Sapporo, Japan, were measured during six winters from 2007 to 2013. Elemental carbon (EC) and organic carbon (OC) concentrations were measured with the thermal optical method, and dust concentration was determined by filter gravimetric measurement. The measurement results using the different filters were compared to assess the filtration efficiency. Adding NH₄H₂PO₄ coagulant to melted snow samples improved the collection efficiency for EC particles by a factor of 1.45. The mass concentrations of EC, OC, and dust in the top 2 cm layer ranged in 0.007–2.8, 0.01–13, and 0.14–260 ppmw, respectively, during the six winters. The mass concentrations and their short-term variations were larger in the surface than in the subsurface. The snow impurity concentrations varied seasonally; that is, they remained relatively low during the accumulation season and gradually increased during the melting season. Although the surface snow impurities showed no discernible trend over the six winters, they varied from year to year, with a negative correlation between the snow impurity concentrations and the amount of snowfall. The surface snow impurities generally increased with the number of days elapsed since snowfall and showed a different rate for EC (1.44), OC (9.96), and dust (6.81). The possible processes causing an increase in surface snow impurities were dry deposition of atmospheric aerosols, melting of surface snow, and sublimation/evaporation of surface snow.

Key Points

Light-absorbing snow impurities are measured with the thermal optical method
Filtration efficiency of a quartz fiber filter is evaluated
Variations in snow impurities at Sapporo, Japan, are measured

1 Introduction

Light-absorbing snow impurities (LASIs) are important parameters controlling snow albedo. An increase in LASI concentration reduces the visible albedo, and the effect is enhanced for old melting snow with large snow grains [Warren and Wiscombe, 1980]. Additionally, an increase in snow grain size itself reduces the near-infrared albedo [Wiscombe and Warren, 1980]. Therefore, LASI can increase the absorption of solar radiation and accelerate snow grain growth, causing further albedo reduction and a positive feedback. Black carbon (BC), organic carbon (OC), and mineral dust are the main constituents of LASI. BC, which is the most absorptive snow impurity, is emitted during the combustion of fossil and solid fuels and during the open burning of biomass [Bond et al., 2013]. Climate models estimate a positive radiative forcing and equilibrium climate warming by BC in snow on a global scale [Hansen and Nazarenko, 2004; Jacobson, 2004; Flanner et al., 2007, 2009]. There are many sources for OC in the atmosphere: biomass and fossil fuel burning, disintegration and dispersion of bulk plant materials, viable biological microbes (bacteria, viruses, and fungal spores), soil organic matter, and marine aerosol [Jacobson et al., 2000; Cerqueira et al., 2010]. Although the optical properties of OC depend on the optical properties of its components, some organic compounds, such as light-absorbing OC in biomass smoke aerosols, appreciably absorb light in the ultraviolet and blue spectral regions [Kirchstetter et al., 2004; Andreae and Gelencsér, 2006]. Soil dust and desert dust are more weakly absorbing than BC per unit mass [Warren and Wiscombe, 1980; Aoki et al., 2011]; however, because of the relatively high concentration of dust in snow, the effect of dust on snow albedo is also important in regions with seasonal or patchy snow cover. Using a snow energy balance model, Painter et al. [2007] demonstrated that the duration of snow cover in a seasonally snow covered mountain range (San Juan Mountains, USA) was shortened by 18 to 35 days through radiative forcing by deposition of disturbed desert dust. Niwano et al. [2012] also found a shortening of snowpack duration at Sapporo, Japan, by 16 to 19 days owing to radiative forcing by BC and dust in snow.

Measurements of LASI concentrations have been widely performed in the Arctic. For example, Doherty et al. [2010] indicated that the median of BC concentration in Arctic snow during 1998 and 2005–2009 were 3 ppbw in Greenland; 8–14 ppbw in Canada, Alaska, and Svalbard; ~20 ppbw in Russia; and 7 ppbw in the Arctic Ocean. These concentrations were approximately the same as those during 1983–1984: 2 ppbw in Greenland; 15–22 ppbw in Canada, Alaska, and Svalbard [Clarke and Noone, 1985]; and 4–60 ppbw in the Swedish Arctic [Noone and Clarke, 1988]. In contrast, at midlatitudes, measurements of LASI concentrations had been limited to a few sites in North America (1.7–19 ppbw) [Clarke and Noone, 1985; Chýlek et al., 1999] and Europe (115–482 ppbw) [Sergent et al., 1993, 1998; Fily et al., 1997] until the 1990s. In recent years, field campaigns for measurement of LASI concentrations have been widely conducted, especially in Asia. On glaciers in the Tibetan Plateau/Himalayan region, the mean BC concentrations in snow were 41 ppbw in 2001 and 2004 [Xu et al., 2006] and 22 ppbw in 2010 [Wang et al., 2012]. In seasonal snow regions, Wang et al. [2013] measured BC concentrations in snow across Northern China in 2010: median BC concentrations were 117 ppbw in the remote northeast on the border of Siberia, 1220 ppbw in the industrial northeast, and 340 ppbw in Inner Mongolia. The BC concentrations measured at these midlatitudes were generally higher than those in the Arctic, particularly in seasonal snow in Northern China, suggesting that the climate impact of BC in snow at midlatitudes could be larger than that in Arctic snow because of higher BC concentrations and higher insolation. Both BC and OC concentrations were measured on the Tibetan Plateau, with mean OC concentrations of 117 ppbw [Xu et al., 2006] and 223 ppbw [Wang et al., 2012]. Additionally, Wang et al. [2013] indicated that light absorption by BC dominated in northeast China, whereas absorption by OC and dust dominated in Inner Mongolia and the Qilian Mountains. In urban areas in Japan, the mass concentration of total snow impurities during snow-covered seasons was continuously measured at Kitami [Aoki et al., 2003], Shinjo [Motoyoshi et al., 2005] and Sapporo [Aoki et al., 2006, 2007]. From the total snow impurity concentration, the main constituent contributing to the mass concentration would be dust, with normally 1000–10,000 ppbw in the accumulation season and an increase to 100,000 ppbw in the melting season. Reductions in snow albedo were observed with increasing snow impurity concentrations at these sites. These results imply that in midlatitude regions, LASI can strongly affect climate through changes in snow albedo.

Three analytical methods are commonly used to measure LASI concentrations: the optical method, measurement with the single-particle soot photometer (SP2), and the thermal optical method. Several optical methods, such as an integrating plate [Lin et al., 1973; Clarke and Noone, 1985], an integrating sandwich [Clarke, 1982], integrating sphere [Heintzenberg, 1982], and an integrating sandwich/integrating sphere spectrometer [Doherty et al., 2010; Grenfell et al., 2011; Wang et al., 2013], have been developed to analyze particles on a filter. Using the wavelength dependence of the light absorption on the filter, one can derive an estimate of the BC concentration and the fraction of light absorbed by non-BC particles such as light-absorbing OC and dust. An advantage of the optical method is that it measures absorption, which is closely related to the absorption of solar radiation by snowpack. However, the optical method does not definitively quantify the BC concentration separately from the OC and dust concentrations.

The SP2 uses laser-induced incandescence to measure the BC mass in individual particles [McConnell et al., 2007; Kaspari et al., 2011; Schwarz et al., 2012; Moteki et al., 2014]. The size distribution of the BC particles and the BC mass concentration can be quantified with the SP2. However, in SP2 measurements, the uncertainty in the value of the BC concentration in snow samples is substantially larger than that in atmospheric samples. Therefore, a careful assessment of nebulization performance and calibration for large BC particles is required to measure the BC mass concentration in snow [Schwarz et al., 2012].

The thermal optical method is based on a controlled combustion of a filter sample with optical correction for pyrolysis [Chow et al., 1993; Chýlek et al., 1999; Lavanchy et al., 1999; Xu et al., 2006; Kuchiki et al., 2009; Aoki et al., 2011, 2014; Wang et al., 2012]. Using preferential volatilization and oxidation of carbonaceous particles at different temperatures, the concentrations of OC and elemental carbon (EC) can be measured separately. The correction for the pyrolysis of OC is achieved by monitoring the change in the optical signal from the filter sample. The thermal optical method is an accurate and reproducible method for measuring OC and EC concentrations but depends on the efficiency of the filter to collect particles from a liquid sample. Torres et al. [2014] reported that the collection efficiency of a quartz fiber filter, which is used for thermal optical analysis, was less than 38%; however, adding a coagulant to the liquid sample before filtration increased the efficiency to 95%. Therefore, the collection efficiency of the filter used for the thermal optical method must be correctly evaluated. The term EC is operationally defined to refer to the nonvolatile or refractory component of the carbonaceous particles measured with the thermal optical method [Bond et al., 2013]. For atmospheric aerosols in the Asian flow, the EC concentrations measured with the thermal optical method agreed with the BC concentrations measured with an optical method using a particle soot absorption photometer to within 2% [Miyazaki et al., 2007]. Therefore, for snow impurities, if the collection efficiency of a filter is 100%, the EC concentration could be nearly equal to the BC concentration.

In this study, we describe an analytical method for measuring the concentrations of EC, OC, and dust in snow based on the thermal optical method and filter gravimetric measurements. The filtration efficiency of the quartz fiber filter, which directly affects the estimates of snow impurity concentrations, was assessed by comparing the concentration measured when using the different filters. Employing this analytical method, we continuously measured the mass concentrations of EC, OC, and dust in snow at Sapporo, Japan, during six winters from 2007 to 2013. We describe the seasonal and interannual variations of LASI concentrations and discuss the possible physical processes responsible for the variations. The mass concentrations of LASI at Sapporo from 2007 to 2009 have been reported by Kuchiki et al. [2009] and Aoki et al. [2011]. However, because the analytical method was improved by changing the filter and correcting for filtration efficiency, we present measurement results for six years including the winters of 2007–2009.

2 Methods

2.1 Snow Sampling

The snow samples were collected at the meteorological observation field (43°04′56″N, 141°20′30″E, 15 m above sea level) of the Institute of Low Temperature Science, Hokkaido University, located in an urban area of Sapporo, Japan. Meteorological and radiation measurements, including snow pit work and snow sampling during the snow-covered season, have been performed since September 2003 [Aoki et al., 2006, 2007, 2011]. Sapporo is located on the Sea of Japan side of Hokkaido, Japan, where the East Asian winter monsoon brings frequent snowfalls. Snow cover normally persists from early December until early April.

Snow sampling was conducted at approximately 11:00 local time twice a week during the snow-covered period. The vertical profile of snow impurity concentration near the snow surface is important for creating a radiation budget in snowpack, because nearly all absorption of solar radiation by snow occurs at a depth d = 0–2 and 0–10 cm during the accumulation season and melting season, respectively, at Sapporo, Japan [Aoki et al., 2011]. Therefore, we collected snow samples in two snow layers: d = 0–2 and 0–10 cm for the five winter seasons of 2007–2012 and d = 0–2 and 2–10 cm for the 2012–2013 winter season. In the last winter, we completely divided the snow layer into two layers to clearly estimate the vertical profile of the snow impurity concentration. The snow samples were collected with a stainless steel spatula and placed in stainless steel containers for the 2007–2008 winter season and in dust-free plastic bags for the five winters of 2008–2013. The plastic bags were more convenient for in situ snow sampling than the stainless steel containers. Changing the collection vessel did not affect the analytical result.

2.2 Filtering Procedure

The snow samples were transported from Sapporo to the Meteorological Research Institute in Tsukuba, Japan, and stored in a freezer at −18°C. Before filtration, the sample was placed in a glass beaker and melted by immersing the beaker in hot water (2007–2011 winters) or heating the sample in a microwave oven (2011–2013 winters). The samples were completely melted in approximately 10 min. Because the time to completely melt a snow sample differed little among these two melting methods, the change in melting method did not affect the change in OC properties in the meltwater. The melted samples were magnetically stirred and sonicated for 20 min and then prefiltered through a 150 µm mesh filter to remove large particles such as leaves, insects, and clothing fibers.

Nuclepore filters (TMTP02500 and GTTP02500; Millipore Corp., MA, USA), a quartz fiber filter (2500QAT-UP; Pall Corp., MI, USA), and a silver membrane filter (SC45335; Sterlitech Corp., WA, USA) were used for filtration of the melted snow samples. The diameter of the filters was 25 mm. The Nuclepore filters were used to measure the total mass concentration of snow impurities by filter gravimetric analysis (section 2.4). Because of their weak heat resistance, the Nuclepore filters could not be used to measure the EC and OC concentrations by the thermal optical method (section 2.3). The quartz fiber filter and silver membrane filter (pore size: 0.45 µm) were used for gravimetric measurement and thermal optical analysis. According to the manufacturer, the typical aerosol retention of 0.3 µm dioctyl phthalate particles on the quartz fiber filter is 99.9%. However, large discrepancies in collection efficiency for a liquid sample were reported in previous studies. Relatively high collection efficiencies of 71%–99% [Chýlek et al., 1999] and 95% [Lavanchy et al., 1999] were evaluated using reference carbon black particles or commercial soot samples. In contrast, Torres et al. [2014] reported an efficiency of less than 38% for their reference materials generated by fossil fuel and biomass combustion. The collection efficiency might depend on the reference sample used for evaluation because the particle size and mixing state could be different in the reference samples. To assess the collection efficiency of the quartz fiber filter for the melted snow samples, we compared the measurement results using the quartz fiber filters to those using the Nuclepore and silver membrane filters (section 3.1).

The deposition area of the filters was 2.01 cm² (16 mm diameter). The filtration was performed under a reduced vacuum pump pressure (Nuclepore and silver membrane filters) or atmospheric pressure (quartz fiber filter). The atmospheric pressure was employed for the quartz fiber filter owing to its weak resistance to rapid water flow. The quartz fiber filters were preheated in an oven at 900°C for 6 h before filtration to reduce the blank load of carbonaceous components. The amount of liquid for filtration was adjusted according to the degree of contamination so that a sufficiently high mass of impurities were captured on the filter for analysis, which was roughly estimated by the change in color of the filter and the rate of filtration. There is a small potential for loss on the walls of the glass beaker in the filtering procedure [Ogren et al., 1983]. To reduce the loss, the walls of glass beaker were rinsed with pure water at the end of filtration. The amounts of liquid for filtration ranged from 31.4 g for highly polluted samples to 729.0 g for clean samples. After filtration, the impurity-laden wet filters were placed on aluminum foil, dried in a desiccator for at least 3 days, and then stored in petri dishes.

2.3 Thermal Optical Analysis

Mass concentrations of EC (c_EC) and OC (c_OC) were measured with a Lab OC-EC Aerosol Analyzer (Sunset Laboratory Inc., OR, USA) using the thermal optical reflectance method [Chow et al., 1993]. A 1.0 cm² square sample from a quartz fiber filter was used for analysis. The sample was volatilized at 120, 250, 450, and 550°C in a pure helium atmosphere and then combusted at 550, 700, and 800°C in a 10% oxygen to 90% helium atmosphere according to the Interagency Monitoring of Protected Visual Environments (IMPROVE) thermal evolution protocol [Chow et al., 2001]. It has been found that the EC concentration is different depending on the thermal evolution protocol adopted in the thermal optical method: the IMPROVE protocol typically overestimates the EC concentration compared to the National Institute of Occupational Safety and Health (NIOSH) protocol. The carbon compounds evolved at each temperature-atmosphere combination were converted to methane and quantified with a flame ionization detector. Throughout the analysis, the laser reflectance from the filter deposit was continuously monitored to correct for the OC pyrolysis. The filter reflectance usually decreased as temperature increased in the helium atmosphere owing to pyrolysis of organic material. When oxygen was added, the remaining light-absorbing carbon combusted and the reflectance increased. The split between OC and EC is the point when the reflectance attains its initial value; that is, OC and EC are defined as the carbon evolved before and after the OC/EC split point, respectively. Because we analyzed filter samples on which melted snow samples were filtered, the OC measured in this study was water-insoluble OC.

The instrument was calibrated with a standard sucrose solution every 2 months, and replicate analyses of the standard solution were within 3%. The total carbon (TC = OC + EC) loaded on the sample filters was more than 100 times larger than the lower detection limit of the instrument (0.4 µg/cm²). The blank filters were prepared by passing pure water through quartz fiber filters (the amounts of pure water were comparable to those of snow samples). The averages of the OC and EC concentrations measured with the blank filters were subtracted from those measured with the sample filters, respectively. The ratios of the OC and EC concentrations for blanks to those for all samples were 12.3 ± 12.0% and 0.14 ± 0.52%, respectively.

2.4 Gravimetric Measurement

The total mass concentration of snow impurities (c_TOT) was gravimetrically measured from the difference between the filter weights before and after filtering. Nuclepore filters were used to estimate c_TOT because c_TOT measured using the Nuclepore filter was higher than those using the quartz fiber filter (section 3.1). Thus, c_TOT was calculated from the sum of the net weights of the Nuclepore filters (pore size: 5.0 and 0.2 µm) used for the two-stage filtering system. The filters were weighed on electronic balances with resolutions of 0.01 mg (AG245; Mettler-Toledo, Switzerland; 2007–2011 winters) and 0.001 mg (XS3DU; Mettler-Toledo; 2011–2013 winters). The lowest net weights for the total snow impurities on the Nuclepore filters were 0.03 and 0.041 mg, which were 3 times (2007–2011 winters) and 41 times (2011–2013 winters) larger than the resolutions of the balances, respectively. Finally, we estimated the dust concentration as c_dust = c_TOT − (c_EC + c_OC), where c_EC + c_OC is measured using the thermal optical method [Aoki et al., 2011].

2.5 Evaluation of Filtration Efficiency by Addition of a Coagulant

Adding a high dosage of coagulant to a water sample improves filtration efficiency [Torres et al., 2014]. Torres et al. [2014] demonstrated that the mean collection efficiency of a quartz fiber filter for BC particles increased to 95% when ammonium dihydrogen phosphate (NH₄H₂PO₄) was added to the sample solution, compared with 5% without added NH₄H₂PO₄. To confirm the effect of a coagulant on the collection efficiency of a quartz fiber filter for natural snow samples, we compared the EC and OC concentrations measured with and without adding NH₄H₂PO₄. In this experiment, the melted snow sample was divided into equal parts, NH₄H₂PO₄ (1.5 g per 100 mL of sample, the optimum concentration reported by Torres et al. [2014]) was added to one of the subsamples, and the mixture was magnetically stirred for 10 min. Then, the subsamples with and without NH₄H₂PO₄ were filtered through quartz fiber filters in parallel, and the EC and OC concentrations were measured. The analytical results of a quartz fiber filter, on which both pure water and with the added coagulant, were filtered to ensure that laser reflectance and laser transmittance of the filter were constant throughout the analysis. This result implies that the coagulant did not affect the optical characteristics of a quartz fiber filter in the thermal optical analysis, that is, the OC/EC split point. The OC and EC concentrations in the coagulant sample were the same level as those of the blank filter.

3 Results and Discussion

3.1 Comparison of Snow Impurity Concentrations Measured Using Different Filters

3.1.1 Total Snow Impurities

Figure 1 shows scatterplots of c_TOT measured using the quartz fiber filter (c_TOT^Q) versus c_TOT measured using the two-stage Nuclepore filter system (c_TOT^N) or the silver membrane filter (c_TOT^S). Histograms for the c_TOT^Q/c_TOT^N and c_TOT^Q/c_TOT^S ratios are shown in Figure 2. The snow samples used for intercomparison were collected during the winters of 2008–2009 and 2009–2010. Very large or very small impurity concentrations and anomalous data owing to inhomogeneity of impurities within a sample could cause errors in the intercomparison analysis. To eliminate anomalous data, we selected 89 snow samples (from a total of 106 samples) that were within 2 standard deviations of the mean for the values of log₁₀c_TOT^Q, log₁₀c_TOT^N, log₁₀c_TOT^S, c_TOT^Q/c_TOT^N, and c_TOT^Q/c_TOT^S. Although the data selections using three values (1.0, 1.5, and 2.0) of standard deviations showed nearly consistent results, we adopted the 2.0 standard deviation, in which the number of selected data was the largest. The results showed that the values of c_TOT^Q were generally lower than the values of c_TOT^N and c_TOT^S. The underestimate with the quartz fiber filter was significant for small mass concentrations of impurities. The averages and medians of the c_TOT^Q/c_TOT^N and c_TOT^Q/c_TOT^S ratios were 0.81–0.88, indicating that the quartz fiber filter estimated 12%–19% less total snow impurity concentrations than collected with the Nuclepore or silver membrane filter.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Scatterplots of mass concentrations of total snow impurities using different filters: (a) c_TOT^Q versus c_TOT^N; (b) c_TOT^Q versus c_TOT^S. The slope of the regression line was calculated in linear space.

A possible cause of the underestimate is a loss of net weight from the quartz fiber filter itself during the filtering process. After pure water was passed through the quartz fiber filter, the weight of the filter decreased relative to that before filtration. The net weight losses were approximately 0.05 mg for liquid amounts of 20–500 g and increased to 0.14 mg for a liquid amount of 1000 g; these values correspond to 0.1–2.8 ppmw of equivalent loss in the mass concentration. The loss was attributed to removal of a small amount of quartz fiber by water during the filtering process. We estimated the loss of the quartz fiber filter for each snow sample according to the relationship between net weight loss and liquid amount for filtration. The contribution of the loss of the quartz fiber filter to the measured values of c_TOT^Q was, on average, 9.9%; the contribution was relatively large for low-concentration samples (the maximum was 60% for c_TOT^Q = 0.2 ppmw) and small for high-concentration samples (the minimum was 0.5% for c_TOT^Q = 65 ppmw). When the estimated weight loss of the quartz fiber filter was taken into account by adding it to the measured c_TOT^Q, the averages and medians of the c_TOT^Q/c_TOT^N and c_TOT^Q/c_TOT^S ratios increased to 0.90–0.94, which indicate that the quartz fiber filter still collects 6%–10% less snow impurities compared to the Nuclepore or silver membrane filter. This result implies that another cause for the underestimation of c_TOT^Q was the lower collection efficiency of the quartz fiber filter relative to that of the Nuclepore or silver membrane filter. Thus, the quartz fiber filter was not suitable for measurement of the concentrations of total snow impurity and dust. Although the c_TOT measured using the Nuclepore and silver membrane filters were approximately the same, we used the Nuclepore filter for measurement of c_TOT because of its smaller pore size.

3.1.2 Total Carbonaceous Particles

Figure 3 shows a scatterplot of the TC concentration measured with the thermal optical method using the quartz fiber filter (c_TC^Q) versus that measured using the silver membrane filter (c_TC^S). A histogram for the c_TC^Q/c_TC^S ratio is shown in Figure 4. The filter samples used for the intercomparison were the same as those used for the intercomparison of total snow impurities (section 3.1.1). Although it is best to separately compare the EC and OC concentrations measured using the different filters, we only compared the TC concentration, because the OC/EC split is uncertain with the silver membrane filter owing to the change in color and deformation of the filter surface with increasing temperature during thermal optical analysis. As a result, c_TC^Q and c_TC^S correlated with a correlation coefficient of 0.96, while c_TC^Q was slightly overestimated compared to c_TC^S. The slope of the regression line in Figure 3 and the average and median of the c_TC^Q/c_TC^S ratios (Figure 4) were all greater than 1.0. The mass concentrations of total snow impurities were lower, whereas those of TC were higher for the quartz fiber filter than the silver membrane filter, implying that the quartz fiber filter captured more TC but less dust compared to the silver membrane filter. The possible reason for the underestimate of TC for the silver membrane filter is that the silver membrane filter could not capture carbonaceous particles smaller than its pore size of 0.45 µm. As a result, the quartz fiber filter is more efficient for the measurement of TC, even though it is not appropriate for measurements of total snow impurity and dust.

3.2 Effect of Coagulant on Collection Efficiency

Comparisons of the EC and OC concentrations measured without a coagulant (c_EC^Q and c_OC^Q) and with a coagulant (c_EC^Q_COA and c_OC^Q_COA) are shown as scatterplots in Figure 5 and as histograms of their ratios in Figure 6. We used snow samples collected during the winter of 2011–2012 and adopted the data selection method described in section 3.1. Selected samples (52 of 62 samples) were within 2 standard deviations of the mean for the values of log₁₀c_TC^Q, log₁₀c_TC^Q_COA, and c_TC^Q_COA/c_TC^Q, where c_TC^Q and c_TC^Q_COA are the TC concentrations measured without and with a coagulant, respectively. For the EC concentration, the values of c_EC^Q_COA were generally larger than those of c_EC^Q; the slope of the regression line (Figure 5a) and the average and median of the c_EC^Q_COA/c_EC^Q ratio (Figure 6a) were 1.58, 1.58, and 1.45, respectively. This result implies that EC particles are collected on a quartz fiber filter at a higher collection efficiency when a coagulant is added to the sample. In contrast, for the OC concentration, c_OC^Q_COA and c_OC^Q were comparable; the slope of the regression line (Figure 5b) and the average and median of the c_OC^Q_COA/c_OC^Q ratio (Figure 6b) were 0.90, 1.08, and 1.00, respectively. This result implies that there was no apparent effect of the coagulant on the filtration efficiency of OC particles. Additionally, we checked the effect of the coagulant on the optical properties of the filter samples, which could influence the OC/EC split point (pyrolysis correction). The result showed that the difference in the fraction of pyrolysis OC in TC measured between the sample with and without coagulant was 0.01 ± 0.14 on average, implying that the coagulant did not affect the OC/EC split point as a whole. The standard deviation is mainly due to the instrumental precision of the OC/EC split point, which is estimated at 5–10% by Chow et al. [1993]. The uncertainty of the OC/EC split point could possibly cause the dispersions in the scatter plots (Figure 5) and the c_EC^Q_COA/c_EC^Q and c_OC^Q_COA/c_OC^Q ratios (Figure 6).

The effect of adding coagulant in this study—a 1.45-fold (median) increase in the EC concentration measured with a coagulant compared to that without a coagulant—was much smaller than that reported by Torres et al. [2014], who found that adding coagulant increased the collection efficiency for BC particles from 5% to 95% (95/5 times). The possible reasons for this large discrepancy are differences in the particle size and the mixing state of the particles in the two studies. Torres et al. [2014] used reference materials that consisted entirely of BC or BC containing less than 30% OC, which were artificially generated by fossil fuel and biomass combustion. However, we used natural snow samples containing not only BC particles but also OC particles and a large amount of dust particles. It was reported that the BC particle size in snow samples was larger than that of a laboratory BC standard [Schwarz et al., 2012] and relatively large compacted BC particles and a mixture of BC and other aerosol particles (BC particles attached to or embedded within other aerosol particles) were observed in atmospheric aerosol samples [Adachi and Buseck, 2013].

To determine the particle size and mixing state of the BC in our samples, we analyzed a representative sample with a scanning electron microscope (SEM; SU-3500, Hitachi Ltd., Japan). Figures 7a and 7b are SEM images of snow impurities filtered without and with a coagulant, respectively (c_EC^Q_COA/c_EC^Q = 1.98). Without added coagulant, we observed large, compacted BC aggregates attached to dust particles (aggregate size: >2 µm) in addition to relatively small, chain-like BC particles (size ~0.5 µm) (Figure 7a). With added coagulant, in addition to the chain-like BC particles, very large BC aggregates >3 µm were observed (Figure 7b). These SEM images show that compacted BC aggregates and BC dust aggregates often form in snow samples without a coagulant, suggesting that the larger-size snow impurities in the natural snow samples led to the high collection efficiency of the quartz fiber filter even without a coagulant. The higher collection efficiency for our snow samples without a coagulant would result in the apparent smaller effect of the coagulant compared to the reference BC particles.

The mass concentration of total snow impurities was considerably larger with added coagulant (c_TOT^Q_COA) than without added coagulant (c_TOT^Q). The average and median of the c_TOT^Q_COA/c_TOT^Q ratio was 5.13 and 3.32, respectively. The main cause of this difference could be recrystallization of the NH₄H₂PO₄ coagulant on the filter after filtration. After 50, 300, and 1000 g of pure water was added with the coagulant were passed through a quartz fiber filter, the filter weight increased by 0.702, 0.721, and 1.384 mg, corresponding to 14.0, 2.4, and 1.4 ppmw of the equivalent increase in mass concentration, respectively. Although dust particles other than EC and OC possibly coagulate, the effect of the coagulant on collection efficiency for total snow impurities could not be quantitatively evaluated because of this remarkable overestimate by the recrystallization.

3.3 Seasonal and Annual Variations of Snow Impurities

Figure 8 (middle and bottom) depicts the mass concentrations of snow impurities in two snow layers measured from 2007 to 2013. On the basis of the coagulant effect on the filtration efficiency, the EC concentrations measured during the 2007–2012 winters, in which no coagulant was used, were corrected by multiplying the measured EC concentrations by 1.45 (the median of the c_EC^Q_COA/c_EC^Q ratio). In contrast, the OC concentrations measured during the 2007–2012 winters were not corrected because there was no apparent effect of the coagulant on the filtration efficiency. For the 2012–2013 winter, we used the EC and OC concentrations measured with the added coagulant. The snow depth and air temperature are shown in Figure 8 (top). Although the maximum snow depth varied from year to year, the snow cover generally started in early December, and the snow depth gradually increased until the end of February. Beginning in March, the snow depth rapidly decreased, and the snow completely melted by the end of March or early April. Thus, we refer to the period from December to February as the accumulation season and the period from March to April as the melting season. The air temperatures were generally around or below 0°C from January to February and then often increased to above 0°C after March.

The mass concentrations of EC, OC, and dust in surface snow (d = 0–2 cm) ranged from 0.007 to 2.8, 0.01 to 13, and 0.14 to 260 ppmw, with medians of 0.19, 0.31, and 3.6 ppmw, respectively, during the six winters. The snow impurity concentrations varied seasonally in each year with fluctuations from day to day or week to week. They remained relatively low during the accumulation season and gradually increased during the melting season. The seasonal variation is clearly depicted in Figure 9, which shows the monthly averages of snow impurity concentrations in two snow layers. The monthly averages and standard deviations for the d = 0–2 cm layer were larger than those for the d = 0–10 or 2–10 cm layer, although the seasonal variations were similar to each other, implying that the mass concentration of snow impurities and its short-term (daily or weekly) variation were larger in the surface than in the subsurface. The short-term variation would be controlled by the physical processes, which caused an increase of the snow impurity concentrations after snow fall. We discuss the possible processes in section 3.5.

Figure 10 shows monthly and overall averages of the mass fractions of EC, OC, and dust in the total impurities. The overall averages of EC, OC, and dust fractions were 5.4%, 8.5%, and 86.1%, respectively. The dust fraction was the dominant mass fraction in all of the cases. The mass fractions of EC, OC, and dust also varied seasonally: the mass fractions of EC and OC decreased, while the mass fraction of dust increased from the accumulation season to the melting season.

3.4 Interannual Variations of Snow Impurity Concentrations During the Winters of 2007–2013

The annual median snow impurity concentrations together with the maximum snow depth, the total amount of snowfall, and the average air temperature for each snow period are depicted in Figure 11. The mass concentrations of snow impurities showed no statistically significant trend from 2007 to 2013; however, they varied from year to year, depending on the meteorological conditions. For example, the median dust concentration was highest in the winter of 2009–2010, when the amount of snowfall was small and the air temperature was relatively high, whereas it was lowest in the winter of 2012–2013, when the amount of snowfall was large and the air temperature was low. The mass concentrations of OC and dust in the d = 0–2 cm layer and the mass concentrations of EC, OC, and dust in the d = 0–10 or 2–10 cm layer were negatively correlated with the total amount of snowfall (correlation coefficients: −0.39, −0.82, −0.66, −0.95, and −0.89, respectively). These results are because surface snow impurities are generally more diluted with an increase in snowfall. In contrast, the EC concentration in the d = 0–2 cm layer was not correlated with the total amount of snowfall (correlation coefficient: 0.10). This result is possibly attributed to the relatively small effect of dilution by snowfall for the EC concentration in surface snow because the EC concentration in snowfall would be relatively high due to the predominance of wet deposition, which is discussed in the next section.

3.5 Variation of Impurity Concentrations With Time After Snowfall

To discuss the physical processes controlling the snow impurity concentrations, we examined the variations of impurity concentrations in the d = 0–2 cm layer and d = 0–10 or 2–10 cm layer as a function of days elapsed since snowfall (Figure 12). This snowfall was defined as an increase of snow depth more than 1 cm from the previous day. All impurity concentrations gradually increased after snowfall, which are statistically significant correlations at the 1% significance level (the correlation coefficients are shown in Table 1). The increase in snow impurity concentration after snowfall affects the seasonal variations and short-term variations. As a result, the snow impurity concentrations increased during the melting season, in which intervals between snowfalls were typically longer than in the accumulation season (Figure 9). The increases in impurity concentrations were fit by a linear regression (Figure 12; the regression coefficients a and b are listed in Table 1). The values of (a + b)/b (the rate of increase in impurity concentration normalized by its initial concentration) are larger for the d = 0–2 cm layer than for the d = 0–10 or 2–10 cm layer, implying that the impurity concentration in the surface snow increases faster than that in the subsurface snow. This is because physical processes causing the variation would first strongly affect the surface and then the subsurface but in a more moderate manner if percolation takes place. Furthermore, the rate of increase in impurity concentration is different for the EC, OC, and dust. Hereafter, we discuss the possible processes that could enhance surface snow impurity concentrations: dry and wet deposition of atmospheric aerosols, melting of surface snow, and sublimation and evaporation of surface snow. The difference in the increase rate between EC and dust concentrations is also discussed. Processes that increase the OC concentration in surface snow are more complicated because OC particles originate from a large variety of sources (biomass and fossil fuel burning, plant materials, viable biological microbes, soil organic matter, and marine aerosol) [Jacobson et al., 2000; Cerqueira et al., 2010], and OC on the snow surface is also potentially affected by the growth of snow algae during the melting season [Takeuchi, 2013]. Therefore, we do not discuss variations of OC concentration in this study.

Table 1. Regression Coefficients and Correlation Coefficients for the EC, OC, and Dust Concentrations in Figure 11a

Snow Impurity	Snow Layer	a (day⁻¹)	b	(a + b)/bb	Correlation Coefficient
EC	0–2 cm	0.072	0.165	1.44	0.662
OC	0–2 cm	0.346	0.039	9.96	0.742
Dust	0–2 cm	5.69	0.979	6.81	0.626
EC	0–10 or 2–10 cm	0.035	0.132	1.27	0.663
OC	0–10 or 2–10 cm	0.083	0.229	1.36	0.585
Dust	0–10 or 2–10 cm	1.16	4.40	1.26	0.427

^a Regression lines are expressed by c = a t + b, where c is the mass concentration of the snow impurity, t is the elapsed time (days) since snowfall, and a and b are the regression coefficients.
^b Rate of increase of the impurity concentration normalized by its initial concentration.

The normalized rate of increase in concentration in the d = 0–2 cm layer for dust (6.81) was 4.7 times the corresponding rate for EC (1.44) (Table 1). This result indicates that the increase in surface snow impurity was faster for dust than EC. The most probable cause is the difference in depositions between EC particles and dust particles. In general, dust particles in the atmosphere are mainly removed by dry deposition, whereas EC particles are largely removed by wet deposition, which is evaluated by models [Tanaka and Chiba, 2005; Huang et al., 2010]. The rate of increase in dust concentration was larger for the d = 0–2 cm layer than for the d = 0–10 or 2–10 cm layer, while that in EC concentration was approximately the same for the two snow layers (Table 1), supporting that dry deposition was dominant for dust particles. The EC mass fraction in total impurities averaged for snowfall days (7.0%) was larger than that for nonsnowfall days (4.7%), which also suggests that wet deposition was dominant for EC particles. Wang et al. [2014] estimated that 87% of EC in the surface snow was attributed to dry deposition at Changbai Mountain, northeastern China. The larger amount of snowfall at Sapporo (mean snow depth was 46.1 ± 25.5 cm) than at Changbai Mountain (mean snow depth 17.2 ± 7.7 cm) would result in the dominance of wet deposition for EC particles at Sapporo. According to these results, the dominance of dry deposition for dust would cause the large rate of increase in dust concentration after snowfall compared to EC. Ohta and Okita [1990] and Kaneyasu et al. [1995] determined that atmospheric EC particles increased in winter (November–February) due to domestic heat use, while atmospheric soil particles increased in spring (March–May) due to suspension from dry land in Sapporo. These variations in the local emissions of atmospheric aerosols would also affect the increase in EC and dust concentration in the surface snow.

Next, we discuss the process by which the melting of surface snow enhances the snow impurity concentrations (snow melt amplification). As snow melts, a fraction of snow impurities are retained at the snow surface, which increases the mass concentrations of impurities in the surface snow. Conway et al. [1996] found that the degree of snow melt amplification was dependent on the particle sizes of the snow impurities. They demonstrated that ash particles larger than approximately 5 µm remained at or near the snow surface, whereas many of the submicron-sized soot particles were flushed through the snow with the meltwater. Our observation shows that the rate of increase in concentration for dust was larger than that for EC, which is consistent with the demonstration of Conway et al. [1996]. The degree of snow melt amplification is determined not only by the particle size but also by the mixing state and hydrophobicity [Doherty et al., 2013]. Furthermore, the melt amplification would depend on the physical properties of snow as well because water percolation, which could move the impurity to the down layer, is affected by snow type, snow layer, snow density, and liquid water content [Colbeck, 1972, 1979; Coléou and Lesaffre, 1998]. Therefore, these other factors would also influence the melt amplification.

Sublimation of dry snow and evaporation of wet snow also increase the concentration of impurities in surface snow by removing snow mass and leaving impurities on the snow surface. Aoki et al. [2014] observed an increase in the concentration of surface snow impurities on the northwest Greenland ice sheet from 28 June to 12 July 2012, which can be explained by the effects of sublimation, evaporation, and snow melt amplification. To assess the effects of sublimation and evaporation on enhancing the snow impurity concentrations at Sapporo, we calculated the snow mass loss in the d = 0–2 cm snow layer by sublimation/evaporation and water percolation during the melting seasons of the six winters of 2007–2013 using a multilayered physical snowpack model, named Snow Metamorphism and Albedo Process (SMAP) [Niwano et al., 2012, 2014], with input parameters of meteorological and radiation data measured at Sapporo. The SMAP model calculates temporal evolution of energy balance, mass balance, and internal physical states of snowpack taking snow settlement, phase changes, water percolation, and snow metamorphism into account. In this study, the model layer thickness was set to range between 0.5 and 3 cm following Niwano et al. [2012]. The snow mass losses in the surface 2 cm by sublimation/evaporation during these melting seasons were 5.2, 11.6, 24.1, 25.4, 27.5, and 32.9 mm liquid equivalent, respectively, whereas water percolation accounted for losses of 147.3, 128.3, 161.8, 256.6, 231.4, and 456.2 mm liquid equivalent, respectively. These data indicate that there was less sublimation/evaporation than percolation in terms of the surface mass loss. However, the efficiency of enhancement in surface snow impurities would be higher for sublimation/evaporation than percolation; the former perfectly leaves impurities at the snow surface but removes only snow mass, whereas the latter leaves some fractions of impurities at the snow surface but moves the residual below with the meltwater. The scavenging fraction of snow impurities due to the percolation of meltwater is estimated to be 10–30%, which was different with site and properties of impurities [Doherty et al., 2013]. Therefore, the sublimation/evaporation and percolation would cause the increase in snow impurity concentrations; however, it is beyond the scope of this study to distinguish the separate effect because of the uncertainty of the scavenging fraction by percolation and the lack of detailed snow physical properties.

4 Conclusions

Mass concentrations of EC, OC, and dust in snow at Sapporo, Japan, during six winters from 2007 to 2013 were measured with the thermal optical method and filter gravimetric measurement. We compared the measurement result using the different filters to assess the collection efficiency of a quartz fiber filter. The results implied that the quartz fiber filter underestimated the concentration of total snow impurities and dust compared to a Nuclepore filter, while the quartz fiber filter captured more carbonaceous particles than a silver membrane filter with a pore size of 0.45 µm. Adding NH₄H₂PO₄ coagulant to melted snow samples enhanced the collection efficiency of the quartz fiber filter for EC particles by a factor of 1.45 but had no apparent effect on the measurement of OC.

The mass concentrations of EC, OC, and dust in surface snow (d = 0–2 cm) ranged from 0.007 to 2.8, 0.01 to 13, and 0.14 to 260 ppmw, respectively, during the six winters. The averages of EC, OC, and dust mass fraction were 5.4%, 8.5%, and 86.1%, respectively. The snow impurity concentrations showed clear seasonal variations: the concentrations were relatively low during the accumulation season and increased during the melting season. The mass concentration of impurities and its short-term variation were larger at the surface than at the subsurface. There was no statistically significant trend in the snow impurity concentration during the six winters. However, the snow impurity concentrations varied from year to year and were negatively correlated with the total amount of snowfall. In general, impurity concentrations increased with the number of days elapsed since snowfall, probably due to dry deposition of atmospheric aerosols, snow melt amplification, and sublimation/evaporation of surface snow. The rate of increase in the surface concentration was different for EC (1.44), OC (9.96), and dust (6.81). The larger rate for dust compared to EC was caused by the fact that dry deposition and melt amplification were more effective for the dust particles than the EC particles.

Acknowledgments

This study was supported in part by the Experimental Research Fund for Global Environment Conservation, Ministry of the Environment, Government of Japan; the Grant for Joint Research Program, Institute of Low Temperature Science, Hokkaido University; and the Japan Society for the Promotion of Science (JSPS), Grant-in-Aid for Scientific Research (S), 23221004. We would like to thank Teppei Yasunari, Masahiro Takahashi, Yuki Sawada, Tomoyasu Kuno, Kou Shimoyama, Taro Nakai, Tetsuo Sueyoshi, Shun Tsutaki, Tsutomu Watanabe, Hayato Oka, Kazuhiro Okuzawa, Junko Mori, Tatsuya Nakayama, Hirokazu Hirano, Chusei Fujiwara, Niyi Sunmonu, Kohei Otomo, Masaki Okuda, Makoto Shimamura, Takehiro Fukuda, Akane Tsushima, Kenta Tone, Hirotaka Sasaki, Momoko Nakayama, Ikumi Oyabu, Yukito Hayashi, Kosuke Tsubokura, Naoya Kanna, Daisuke Nishimura, Takako Saeki, Satoshi Yoshida, Shiho Sato, Daiki Sakakibara, Satoshi Matsuno, Satoshi Omiya, Nobuyoshi Konishi, Takeshi Ishikawa, Masahiro Minowa, and Etsuko Tanaka for the snow sampling and snow pit work at the Institute of Low Temperature Science, Hokkaido University. We would also like to thank Masae Igosaki for support in analyzing the snow impurities. The data used for this study are available upon request from Teruo Aoki, Meteorological Research Institute, Japan ([email protected]).

References

Adachi, K., and P. R. Buseck (2013), Changes of ns-soot mixing states and shapes in an urban area during CalNex, J. Geophys. Res. Atmos., 118, 3723–3730, doi:10.1002/jgrd.50321.
Andreae, M. O., and A. Gelencsér (2006), Black carbon or brown carbon? The nature of light-absorbing carbonaceous aerosols, Atmos. Chem. Phys., 6, 3131–3148, doi:10.5194/acp-6-3131-2006.
Aoki, T., A. Hachikubo, and M. Hori (2003), Effects of snow physical parameters on broadband albedos, J. Geophys. Res., 108(D19), 4616, doi:10.1029/2003JD003506.
Aoki, T., H. Motoyoshi, Y. Kodama, T. J. Yasunari, K. Sugiura, and H. Kobayashi (2006), Atmospheric aerosol deposition on snow surfaces and its effect on albedo, SOLA, 2, 013–016, doi:10.2151/sola.2006-004.
Aoki, T., H. Motoyoshi, Y. Kodama, T. J. Yasunari, and K. Sugiura (2007), Variations of the snow physical parameters and their effects on albedo in Sapporo, Japan, Ann. Glaciol., 46, 375–381, doi:10.3189/172756407782871747.
Aoki, T., K. Kuchiki, M. Niwano, Y. Kodama, M. Hosaka, and T. Tanaka (2011), Physically based snow albedo model for calculating broadband albedos and the solar heating profile in snowpack for general circulation models, J. Geophys. Res., 116, D11114, doi:10.1029/2010JD015507.
Aoki, T., S. Matoba, S. Yamaguchi, T. Tanikawa, M. Niwano, K. Kuchiki, K. Adachi, J. Uetake, H. Motoyama, and M. Hori (2014), Light-absorbing snow impurity concentrations measured on northwest Greenland ice sheet in 2011 and 2012, Bull. Glaciol. Res., 32, 21–31, doi:10.5331/bgr.32.21.
Bond, T. C., et al. (2013), Bounding the role of black carbon in the climate system: A scientific assessment, J. Geophys. Res. Atmos., 118, 5380–5552, doi:10.1002/jgrd.50171.
Cerqueira, M., C. Pioa, M. Legrandb, H. Puxbaumc, A. Gieblc, J. Afonsoa, S. Preunkertb, A. Gelencsérd, and P. Fialhoe (2010), Particulate carbon in precipitation at European background sites, Aerosol Sci., 41, 51–61, doi:10.1016/j.jaerosci.2009.08.002.
Chow, J. C., J. G. Watson, L. C. Pritchett, W. R. Pierson, C. A. Frazier, and R. G. Purcell (1993), The DRI thermal/optical reflectance carbon analysis system: Description, evaluation and applications in U.S. air quality studies, Atmos. Environ., 27(8), 1185–1201, doi:10.1016/0960-1686(93)90245-T.
Chow, J. C., J. G. Watson, D. Crow, D. H. Lowenthal, and T. Merrifield (2001), Comparison of IMPROVE and NIOSH carbon measurements, Aerosol Sci. Technol., 34(1), 23–34.
Chýlek, P., L. Kou, B. Johnson, F. Boudala, and G. Lesins (1999), Black carbon concentrations in precipitation and near surface air in and near alifax, Nova Scotia, Atmos. Environ., 33(14), 2269–2277, doi:10.1016/0004-6981(85)90113-1.
Clarke, A. D. (1982), Integrating sandwich: A new method of measurement of the light absorption coefficient for atmospheric particles, Appl. Opt., 21(16), 3011–3020.
Clarke, A. D., and K. J. Noone (1985), Soot in the Arctic snowpack: A cause for perturbations in radiative transfer, Atmos. Environ., 19(12), 2045–2053, doi:10.1016/0004-6981(85)90113-1.
Colbeck, S. C. (1972), A theory of water percolation in snow, J. Glaciol., 11(63), 369–385.
Colbeck, S. C. (1979), Water flow through heterogeneous snow, Cold Reg. Sci. Technol., 1, 37–45.
Coléou, C., and B. Lesaffre (1998), Irreducible water saturation in snow: Experimental results in a cold laboratory, Ann. Glaciol., 26, 64–68.
Conway, H., A. Gades, and C. F. Raymond (1996), Albedo of dirty snow during conditions of melt, Water Resour. Res., 32(6), 1713–1718, doi:10.1029/96WR00712.
Doherty, S. J., S. G. Warren, T. C. Grenfell, A. D. Clarke, and R. E. Brandt (2010), Light-absorbing impurities in Arctic snow, Atmos. Chem. Phys., 10(23), 11,647–11,680, doi:10.5194/acp-10-11647-2010.
Doherty, S. J., T. C. Grenfell, S. Forsström, D. L. Hegg, S. G. Warren, and R. Brandt (2013), Observed vertical redistribution of black carbon and other insoluble light-absorbing particles in melting snow, J. Geophys. Res. Atmos., 118, 5553–5569, doi:10.1002/jgrd.50235.
Fily, M., B. Bourdelles, J. P. Dedieu, and C. Sergent (1997), Comparison of in situ and Landsat Thematic Mapper derived snow grain characteristics in the Alps, Remote Sens. Environ., 59, 452–460, doi:10.1016/S0034-4257(96)00113-7.
Flanner, M. G., C. S. Zender, J. T. Randerson, and P. J. Rasch (2007), Present-day climate forcing and response from black carbon in snow, J. Geophys. Res., 112, D11202, doi:10.1029/2006JD008003.
Flanner, M. G., C. S. Zender, P. G. Hess, N. M. Mahowald, T. H. Painter, V. Ramanathan, and P. J. Rasch (2009), Springtime warming and reduced snow cover from carbonaceous particles, Atmos. Chem. Phys., 9(7), 2481–2497, doi:10.5194/acp-9-2481-2009.
Grenfell, T. C., S. J. Doherty, A. D. Clarke, and S. G. Warren (2011), Light absorption by particulate impurities in snow and ice determined by spectrophotometric analysis of filters, Appl. Opt., 50(14), 2037–2048.
Hansen, J., and L. Nazarenko (2004), Soot climate forcing via snow and ice albedos, Proc. Natl. Acad. Sci. U.S.A., 101(2), 423–428, doi:10.1073/pnas.2237157100.
Heintzenberg, J. (1982), Size-segregated measurements of particulate elemental carbon and aerosol light absorption at remote arctic locations, Atmos. Environ., 16(11), 2061–2469.
Huang, L., S. L. Gong, C. Q. Jia, and D. Lavoue (2010), Importance of deposition process in simulating the seasonality of the Arctic black carbon aerosol, J. Geophys. Res., 115, D17207, doi:10.1029/2009JD013478.
Jacobson, M. Z. (2004), Climate response of fossil fuel and biofuel soot, accounting for soot's feedback to snow and sea ice albedo and emissivity, J. Geophys. Res., 109, D21201, doi:10.1029/2004JD004945.
Jacobson, M. C., H. C. Hansson, K. J. Noone, and R. J. Charlson (2000), Organic atmospheric aerosols: Review and state of the science, Rev. Geophys., 38(2), 267–294, doi:10.1029/1998RG000045.
Kaneyasu, N., S. Ohta, and N. Murao (1995), Seasonal variation in the chemical composition of atmospheric aerosols and gaseous species in Sapporo, Japan, Atmos. Environ., 29(13), 1559–1568.
Kaspari, S. D., M. Schwikowski, M. Gysel, M. G. Flanner, S. Kang, S. Hou, and P. A. Mayewski (2011), Recent increase in black carbon concentrations from a Mt. Everest ice core spanning 1860–2000 AD, Geophys. Res. Lett., 38, L04703, doi:10.1029/2010GL046096.
Kirchstetter, T. W., T. Novakov, and P. V. Hobbs (2004), Evidence that the spectral dependence of light absorption by aerosols is affected by organic carbon, J. Geophys. Res., 109, D21208, doi:10.1029/2004JD004999.
Kuchiki, K., T. Aoki, T. Tanikawa, and Y. Kodama (2009), Retrieval of snow physical parameters using a ground-based spectral radiometer, Appl. Opt., 48, 5567–5582, doi:10.1364/AO.48.005567.
Lavanchy, V. M. H., H. W. Gäggeler, U. Schotterer, M. Schwikowski, and U. Baltensperger (1999), Historical record of carbonaceous particle concentrations from a European high-alpine glacier (Colle Gnifetti, Switzerland), J. Geophys. Res., 104(D17), 21,227–21,236, doi:10.1029/1999JD900408.
Lin, C. I., M. Baker, and R. J. Charlson (1973), Absorption coefficient of atmospheric aerosol: A method for measurement, Appl. Opt., 12(6), 1356–1363.
McConnell, J., R. L. Edwards, G. L. Kok, M. G. Flanner, C. S. Zender, E. S. Saltzman, J. R. Banta, D. R. Pasteris, M. M. Carter, and J. D. W. Kahl (2007), 20th-century industrial black carbon emissions altered Arctic climate forcing, Science, 317, 1381–1384, doi:10.1126/science.1144856.
Miyazaki, Y., Y. Kondo, S. Han, M. Koike, D. Kodama, Y. Komazaki, H. Tanimoto, and H. Matsueda (2007), Chemical characteristics of water-soluble organic carbon in the Asian outflow, J. Geophys. Res., 112, D22S30, doi:10.1029/2007JD009116.
Moteki, N., Y. Kondo, and K. Adachi (2014), Identification by single-particle soot photometer of black carbon particles attached to other particles: Laboratory experiments and ground observations in Tokyo, J. Geophys. Res. Atmos., 119, 1031–1043, doi:10.1002/2013JD020655.
Motoyoshi, H., T. Aoki, M. Hori, O. Abe, and S. Mochizuki (2005), Possible effect of anthropogenic aerosol deposition on snow albedo reduction at Shinjo, Japan, J. Meteorol. Soc. Jpn., 83A, 137–148.
Niwano, M., T. Aoki, K. Kuchiki, M. Hosaka, and Y. Kodama (2012), Snow Metamorphism and Albedo Process (SMAP) model for climate studies: Model validation using meteorological and snow impurity data measured at Sapporo Japan, J. Geophys. Res., 117, F03008, doi:10.1029/2011JF002239.
Niwano, M., T. Aoki, K. Kuchiki, M. Hosaka, Y. Kodama, S. Yamaguchi, H. Motoyoshi, and Y. Iwata (2014), Evaluation of updated physical snowpack model SMAP, Bull. Glaciol. Res., 32, 65–78, doi:10.5331/bgr.32.65.
Noone, K. J., and A. D. Clarke (1988), Soot scavenging measurements in Arctic snowfall, Atmos. Environ., 22(12), 2773–2778.
Ogren, J. A., R. J. Charlson, and P. J. Groblicki (1983), Determination of elemental carbon in rainwater, Anal. Chem., 55, 1569–1572.
Ohta, S., and T. Okita (1990), A chemical characterization of atmospheric aerosol in Sapporo, Atmos. Environ., 24A(4), 815–822.
Painter, T. H., A. P. Barrett, C. C. Landry, J. C. Neff, M. P. Cassidy, C. R. Lawrence, K. E. McBride, and G. L. Farmer (2007), Impact of disturbed desert soils on duration of mountain snow cover, Geophys. Res. Lett., 34, L12502, doi:10.1029/2007GL030284.
Schwarz, J. P., S. J. Doherty, F. Li, S. T. Ruggiero, C. E. Tanner, A. E. Perring, R. S. Gao, and D. W. Fahey (2012), Assessing single particle soot photometer and integrating sphere/integrating sandwich spectrophotometer measurement techniques for quantifying black carbon concentration in snow, Atmos. Meas. Tech., 5, 2581–2592.
Sergent, C., E. Pougatch, and M. Sudul (1993), Experimental investigation of optical snow properties, Ann. Glaciol., 17, 281–287.
Sergent, C., C. Leroux, E. Pougatch, and F. Guirado (1998), Hemisphericaldirectional reflectance measurements of natural snow in the 0.9–1.45 mm spectral range: Comparison with adding-doubling modelling, Ann. Glaciol., 26, 59–63.
Takeuchi, N. (2013), Seasonal and altitudinal variations in snow algal communities on an Alaskan glacier (Gulkana glacier in the Alaska range), Environ. Res. Lett., 8, 035002, doi:10.1088/1748-9326/8/3/035002.
Tanaka, T., and M. Chiba (2005), Global simulation of dust aerosol with a chemical transport model. MASINGAR, J. Meteorol. Soc. Jpn., 83A, 255–278.
Torres, A., T. C. Bond, C. M. B. Lehmann, R. Subramanian, and O. L. Hadley (2014), Measuring organic carbon and black carbon in rainwater: Evaluation of methods, Aerosol Sci. Technol., 48, 238–249, doi:10.1080/02786826.2013.868596.
Wang, M., B. Xu, H. Zhao, J. Cao, D. Joswiak, G. Wu, and S. Lin (2012), The influence of dust on quantitative measurements of black carbon in ice and snow when using a thermal optical method, Aerosol Sci. Technol., 46, 60–69, doi:10.1080/02786826.2011.605815.
Wang, X., S. J. Doherty, and J. Huang (2013), Black carbon and other light-absorbing impurities in snow across Northern China, J. Geophys. Res. Atmos., 118, 1471–1492, doi:10.1029/2012JD018291.
Wang, Z. W., J. C. Gallet, C. A. Pedersen, X. S. Zhang, J. Ström, and Z. J. Ci (2014), Elemental carbon in snow a Changbai Mountain, northeastern China: Concentrations, scavenging ratios, and dry deposition velocities, Atmos. Chem. Phys., 14, 629–540, doi:10.5194/acp-14-629-2014.
Warren, S. G., and W. J. Wiscombe (1980), A model for the spectral albedo of snow. II: Snow containing atmospheric aerosols, J. Atmos. Sci., 37(12), 2734–2745.
Wiscombe, W. J., and S. G. Warren (1980), A model for the spectral albedo of snow. I: Pure snow, J. Atmos. Sci., 37, 2712–2733.
Xu, B., T. Yao, X. Liu, and N. Wang (2006), Elemental and organic carbon measurements with a two-step heating gas chromatography system in snow samples from the Tibetan Plateau, Ann. Glaciol., 43, 257–262, doi:10.3189/172756406781812122.

Citing Literature

Volume120, Issue2

27 January 2015

Pages 868-882

Elemental carbon, organic carbon, and dust concentrations in snow measured with thermal optical and gravimetric methods: Variations during the 2007–2013 winters at Sapporo, Japan

Abstract

Key Points

1 Introduction