Land Surface Albedo Bias in Climate Models and Its Association with Tropical Precipitation

The influence of land surface albedo on tropical precipitation has been widely appreciated for decades, but bias in the representation of surface albedo in weather and climate models has been studied much less than bias in sea surface temperature or soil moisture. This study shows that CMIP5 simulations of the late 20th century exhibit large multi-model mean bias and intermodel variability in surface broadband shortwave albedo averaged over boreal summer. Intermodel variability in surface albedo is coherent on global scales and correlates with intermodel variability of precipitation over large parts of the tropics. Although enhanced precipitation would be expected to reduce soil albedo, evidence is presented in support of the alternate hypothesis that intermodel albedo variability instead causes intermodel precipitation variability. Further study is needed to elucidate the cause of surface albedo bias in individual models, but these results suggest that reducing that bias will improve simulations of low-latitude precipitation.