Effects of snow physical parameters on shortwave broadband albedos

Snow pit work of several‐day intervals was performed simultaneously with radiation budget observations during two winters in eastern Hokkaido, Japan. From these data we investigated the effects of elapsed time after snowfall (snow aging), air temperature, snow surface temperature, snow grain size, and snow impurities on the visible and the near infrared albedos for improving the snow albedo scheme in the land surface process from an empirical model to a physically based model. The dependence of albedos on elapsed time after snowfall could be clearly classified by dividing the snow‐covered period into a dry snow season and a wet snow season rather than by snow surface temperature. The albedo reduction by snow aging statistically depends on the snow surface temperature, which is often used to predict the snow albedo in the empirical model of land surface process. However, the albedo reduction rate was very scattered for snow surface temperatures above −10°C. This is because the snow albedo reduction essentially depends on the snow grain size and the concentration of snow impurities. Using the radiative transfer model for the atmosphere‐snow system, the effects of these snow physical parameters on broadband albedos are calculated and compared with the observed ones. The measured broadband albedos fell close to the range of theoretically calculated ones as functions of these snow physical parameters. In particular, the measured near infrared albedo agreed well with the theoretically calculated ones both for the dependence of snow grain size and snow impurities but not as well for the visible albedo in detail. In the near infrared region the light absorption by ice is strong, and thus the snow albedo contains the information of snow physical parameters near the surface where these parameters are measured. In contrast, the visible albedo contains the snow information in the deeper layer because the ice is relatively transparent in the visible region. This suggests the necessity of the multiple‐snow‐layer model for the visible region in the physically based snow albedo model.