资源环境科技发展态势分析平台

The representation of snow albedo in global models is recognized as one of the key sources of uncertainty in the simulation of mean and variability of climate. The Climate Forecast System version 2 (CFSv2), which is used for subseasonal-to-seasonal forecast of weather and climate, also suffers from large overestimation of surface albedo, particularly over the snow region. In order to improve the snow albedo biases, performances of three snow albedo schemes: the existing maximum snow albedo' (MSA) scheme of land surface model Noah, temperature dependent scheme of Roeckner et al. (The atmospheric general circulation model ECHAM5-part 1. Technical Report 349, Max Planck Institute for Meteorology, 2003) (RK03) and prognostic albedo scheme of Dutra et al. (J Hydrometeor 11:899-916, 2010) (DU10) are evaluated in offline mode. While offline Noah performs better with MSA scheme as compared to RK03 and DU10 in simulating climatological mean albedo, DU10 scheme shows a better skill in simulating monthly albedo and also captures the interannual variability with high fidelity. Therefore, MSA and DU10 schemes are also evaluated in coupled model along with the existing albedo scheme of Briegleb (J Geophys Res 97:7603-7612, 1992) (BR92) in CFSv2. The performance of MSA scheme is found to be better in simulating surface albedo, snow water equivalent and 2m air temperature as compared to those using DU10 and BR92 schemes. Furthermore, the improved albedo have positive impact in the simulation of Indian summer monsoon rainfall, that reduces the dry bias over the Indian subcontinent, and cold bias in sea surface temperature and improves rainfall teleconnection. Thus, among four schemes (MSA, RK03, DU10 and BR92) evaluated here, MSA appears to be the most suitable snow albedo scheme for CFSv2.