Characterizing tropical pacific water vapor and radiative biases in CMIP5 GCMs: Observation-based analyses and a snow and radiation interaction sensitivity experiment

Abstract

Significant systematic biases in the moisture fields within the tropical Pacific trade wind regions are found in the Coupled Model Intercomparison Project (CMIP3/CMIP5) against profile and total column water vapor (TotWV) estimates from the Atmospheric Infrared Sounder and TotWV from the Special Sensor Microwave/Imager. Positive moisture biases occur in conjunction with significant biases of eastward low-level moisture convergence north of the South Pacific Convergence Zone and south of the Intertropical Convergence Zone—the V-shaped regions. The excessive moisture there is associated with overestimates of reflected upward shortwave (RSUT), underestimates of outgoing longwave radiation (RLUT) at the top of atmosphere (TOA), and underestimates of downward shortwave flux at the surface (RSDS) compared to Clouds and the Earth’s Energy System, Energy Balance and Filled data. We characterize the impacts of falling snow and its radiation interaction, which are not included in most CMIP5 models, on the moisture fields using the National Center for Atmospheric Research-coupled global climate model (GCM). A number of differences in the model simulation without snow-radiation interactions are consistent with biases in the CMIP5 simulations. These include effective low-level eastward/southeastward wind and surface wind stress anomalies, and an increase in TotWV, vertical profile of moisture, and cloud amounts in the V-shaped region. The anomalous water vapor and cloud amount might be associated with the model increase of RSUT and decrease of RLUT at TOA and decreased RSDS in clear and all sky in these regions. These findings hint at the importance of water vapor-radiation interactions in the CMIPS/CMIP5 model simulations that exclude the radiative effect of snow.