Improving Experiments Of Snow Albedo Parameterization Scheme In Land Surface Model BCC_AVIM

An18-yr long(1993-2001) snow and meteorological dataset from Col de Porte,France was used to analyze how the snow albedos of different periods withinsnow-covered period vary with elapsed time after snowfall(snow aging) and toinvestigate the effects of air temperature, snow surface temperature and snow depthon snow surface albedo. Combined with observed snow types and liquid water content,an air temperature index was proposed to divide the whole snow-covered into theaccumulating (dry) season and the melting (wet) season. When the mean airtemperature during the day is continuously over2.5℃and the accumulatedtemperature is over18℃, the snow surface temperature is close to or above zerodegree Celsius, then the snow season evolves into the wet period. The results showedthat the broadband shortwave albedo decreased exponentially with respect to elapsedtime after snowfall and snow albedo in wet season decreased more rapidly than that ofdry season. This is because the effective snow grain size and concentration of snowimpurities that essentially affect the snow albedo are significantly larger in wet seasonthan those in dry season. Snow surface temperature, which is often used in empiricalmodel of land surface process, had a good relation with snow albedo reduction rate.However,the albedo reduction rates were very scattered for all temperature ranges andthe albedo of new snow was low for the surface temperatures below-10℃. This isbecause the snow albedo is essentially determined by the grain size, whichisdetermined not only by snow surface temperature. When the snow is very thin, the underlying soil whose albedo is often much lower than that of snow can reduce thesnow surface albedo by absorbing more solar radiation and increase the albedoreduction rate.According to the above observational analyses and the deficienciesinunderestimation of the albedo and overestimation of the snow water equivalentand the snow depth in some years by BCC_AVIM, several improvements wereemployed into the BCC_AVIMmodel:(1) The snow cover fraction (SCF) scheme isreplaced with a new one to decrease the negative discrepancy of SCF;(2)The snowage is reset to zero when the total snowfall exceeds3mm to mimic the extreme ofsnow albedo;(3)The wet season starts when both the accumulated temperature indexis over18℃and the date falls beyond February1stto mimic the slower decrease ofsnow albedo in dry season and faster decrease in wet season;(4) The decay rate ofalbedo accelerates when the snow depth is below40cm to take account of theinfluence of underlying surface. The improved snow albedo parameterization schemecan better represent the snow albedo and its decay with time, but the positive bias insnow depth and snow water equivalent still exists in the simulation of BCC_AVIM.The positive bias is slightly decreased after the energy of rainfall is added to thesnowpack. Due to the small weight of rainfall energy in the overall energy budget ofsnow pack, considering rainfall energy in BCC_AVIM cannot radically improve thesimulation of snow depth and snow water equivalent.