Impact of diurnal cycle simulation on mean state and interannual variability in a CMIP5 model: CanAM4

Evaluation of physical processes within CMIP5 models is a critical task that enables (1) a better interpretation and application of the simulation results and (2) future model improvements. The diurnal cycle is a fundamental mode of Earth system variability involving radiative, convective, turbulent, and cloud processes; all are critical for modeling climate change. This study evaluates the representation of cloud, precipitation, and radiation diurnal cycles in an AMIP simulation spanning 2000-2009 by the fourth version of the Canadian Centre for Modelling and Analysis Atmospheric GCM (CanAM4)—a CMIP5 participant—using a suite of satellite observations of top-of-atmosphere (TOA) radiation, clouds and precipitation in the tropics. CanAM4 simulates the general patterns of observed first diurnal cycle harmonic amplitude in clouds, precipitation, and TOA radiation well. However, large differences in TOA outgoing longwave radiation (OLR) and longwave cloud forcing (LWCF) diurnal cycles are found in convective regions (e.g., central South America and central Africa). Significant low cloud amount diurnal cycle amplitude biases are also found over oceans. The impact of these diurnal cycle errors on the simulation of the mean climate state and interannual variability are discussed.