The snow is melted and filtered; the filters are analyzed in a specially designed spectrophotometer system to infer the concentration of black carbon (BC), the fraction of absorption due to non-BC light-absorbing constituents and the absorption Angstrom exponent of all particles. This is done using BC calibration standards having a mass absorption efficiency of 6.0 m2/g at 550nm and by making an assumption that the absorption Angstrom exponent for BC is 1.0 and for non-BC light-absorbing aerosol is 5.0. The reduction of snow albedo is primarily due to BC, but other impurities, principally brown (organic) carbon, are typically responsible for ~40% of the visible and ultraviolet absorption. The meltwater from selected snow samples was saved for chemical analysis to identify sources of the impurities. Median BC amounts in surface snow are as follows (nanograms of carbon per gram of snow): Greenland 3, Arctic Ocean snow 7, melting sea ice 8, Arctic Canada 8, subarctic Canada 14, Svalbard 13, northern Norway 21, western Arctic Russia 26, northeastern Siberia 17. Concentrations are more variable in the European Arctic than in Arctic Canada or the Arctic Ocean, probably because of the proximity to BC sources. Individual samples of falling snow were collected on Svalbard, documenting the springtime decline of BC from March through May.
Absorption Angstrom exponents are 1.5-1.7 in Norway, Svalbard, and western Russia, 2.1-2.3 elsewhere in the Arctic, and 2.5 in Greenland. Correspondingly, the estimated contribution to absorption by non-BC constituents in these regions is ~25%, 40%, and 50% respectively.
When the snow surface layer melts some of the BC is left at the top of the snowpack rather than being carried away in meltwater. This process was observed in a few locations; it can cause a positive feedback on snowmelt.
The BC content of the Arctic atmosphere has declined markedly since 1989, according to the continuous measurements of near-surface air at Alert (Canada), Barrow (Alaska), and Ny-Alesund (Svalbard). Correspondingly, the recent BC concentrations for Arctic snow are somewhat lower than those reported by Clarke and Noone for 1983-4, but because of methodological differences it is not clear that the differences are significant. Nevertheless, the BC content of Arctic snow appears to be no higher now than in 1984, so it is doubtful that BC in Arctic snow has contributed to the rapid decline of Arctic sea ice in recent years.