In situ observations of black carbon in snow and the corresponding spectral surface albedo reduction

Abstract

AbstractBlack carbon (BC) particles emitted from incomplete combustion of fossil fuel and biomass and deposited on snow and ice darken the surface and reduce the surface albedo. Even small initial surface albedo reductions may have larger adjusted effects due to snow morphology changes and changes in the sublimation and snow melt rate. Most of the literature on the effect of BC on snow surface albedo is based on numerical models, and few in situ field measurements exist to confirm this reduction. Here we present an extensive set of concurrent in situ measurements of spectral surface albedo, BC concentrations in the upper 5 cm of the snowpack, snow physical parameters (grain size and depth), and incident solar flux characteristics from the Arctic. From this data set (with median BC concentrations ranging from 5 to 137 ng BC per gram of snow) we are able to separate the BC signature on the snow albedo from the natural snow variability. Our measurements show a significant correlation between BC in snow and spectral surface albedo. Based on these measurements, parameterizations are provided, relating the snow albedo, as a function of wavelength, to the equivalent BC content in the snowpack. The term equivalent BC used here is the elemental carbon concentration inferred from the thermo‐optical method adjusted for the fraction of non‐BC constituents absorbing sunlight in the snow. The first parameterization is a simple equation which efficiently describes the snow albedo reduction due to the equivalent BC without including details on the snow or BC microphysics. This can be used in models when a simplified description is needed. A second parameterization, including snow grain size information, shows enhanced correspondence with the measurements. The extracted parameterizations are valid for wavelength bands 400–900 nm, constrained for BC concentrations between 1 and 400 ng g−1, and for an optically thick snowpack. The parameterizations are purely empirical, and particular focus was on the uncertainties associated with the measurements, and how these uncertainties propagate in the parameterizations. Integrated, the first parameterization (based only on the equivalent BC) gives a broadband (400–900 nm) snow albedo reduction of 0.004 due to 10 ng equivalent BC per gram of snow, while the effect is almost 5 times larger for BC concentrations 1 order of magnitude higher. The study shows that the reconstructed albedo from the second parameterization (including information on the snow grain size) corresponds better to the radiative transfer model Snow, Ice, and Aerosol Radiation albedo than the reconstructed albedo from the first parameterization (excluding grain size information).