Observations and model simulations of snow albedo reduction in seasonal snow due to insoluble light-absorbing particles during 2014 Chinese survey

Abstract. A snow survey was carried out to collect 13 surface snow samples (10 for fresh snow, and 3 for aged snow) and 79 subsurface snow samples in seasonal snow at 13 sites across northeastern China in January 2014. A spectrophotometer combined with chemical analysis was used to quantify snow particulate absorption by insoluble light-absorbing particles (ILAPs, e.g., black carbon, BC; mineral dust, MD; and organic carbon, OC) in snow. Snow albedo was measured using a field spectroradiometer. A new radiative transfer model (Spectral Albedo Model for Dirty Snow, or SAMDS) was then developed to simulate the spectral albedo of snow based on the asymptotic radiative transfer theory. A comparison between SAMDS and an existing model – the Snow, Ice, and Aerosol Radiation (SNICAR) – indicates good agreements in the model-simulated spectral albedos of pure snow. However, the SNICAR model values tended to be slightly lower than those of SAMDS when BC and MD were considered. Given the measured BC, MD, and OC mixing ratios of 100–5000, 2000–6000, and 1000–30 000 ng g⁻¹, respectively, in surface snow across northeastern China, the SAMDS model produced a snow albedo in the range of 0.95–0.75 for fresh snow at 550 nm, with a snow grain optical effective radius (R_eff) of 100 µm. The snow albedo reduction due to spherical snow grains assumed to be aged snow is larger than fresh snow such as fractal snow grains and hexagonal plate or column snow grains associated with the increased BC in snow. For typical BC mixing ratios of 100 ng g⁻¹ in remote areas and 3000 ng g⁻¹ in heavy industrial areas across northern China, the snow albedo for internal mixing of BC and snow is lower by 0.005 and 0.036 than that of external mixing for hexagonal plate or column snow grains with R_eff of 100 µm. These results also show that the simulated snow albedos by both SAMDS and SNICAR agree well with the observed values at low ILAP mixing ratios but tend to be higher than surface observations at high ILAP mixing ratios.