To understand the role of the diurnal cycle in maintaining the mean state of the climate system, two 10-year simulations were carried out with the UCLA atmospheric general circulation model (resolution 4 degrees latitude by 5 degrees longitude), one with a diurnal cycle, and one without. In the simulation with a diurnal cycle (control), sunshine varies according to the hour of day, as the model was originally designed. In the other simulation, the sunshine a given location normally receives is distributed uniformly throughout the 24-hour day. Both simulated climates therefore receive the same amount of insolation on average. Since the focus of this investigation is the diurnal cycle, we first compared the simulated diurnal cycle of temperature to observations. We find that the model’s diurnal cycle of temperature is very similar in amplitude and phase to observations, with large amplitude in dry regions, and maximum temperatures occurring in the early to mid afternoon nearly everywhere.

The precipitation and circulation fields in the experiment without the diurnal cycle are markedly different from those of the control experiment, particularly in the deep tropics over land. The region where this difference is most striking is in sub-Saharan Africa. Here precipitation and convergence are drastically reduced in the model without the diurnal cycle during the wet season. In the control experiment as in the observations, a pronounced diurnal cycle of precipitation occurs here, with most of the precipitation falling in mid to late afternoon. At this time of day, the air in the moist boundary layer warms and reaches a critical level of buoyancy relative to the air above, generating deep convection and precipitation. When sunshine is uniformly distributed over 24 hours, the surface layer rarely becomes warm enough to initiate this convective overturning, dramatically reducing overall precipitation. Similar threshold effects influence the precipitation and circulation fields in other continental and nearby ocean regions when the diurnal cycle is removed.

The diurnal variation in sunshine also affects the mean depth of the planetary boundary layer (PBL). PBL depth is dramatically lower in the control model in deserts and mid-latitude continents during summertime. In these regions in both observations and the control simulation, the PBL grows slowly during daytime as it eats into the overlying stratified atmosphere, then collapses when the sun sets and a cool dense layer of air forms at the surface. Because PBL depth comes into equilibrium with its environment more slowly when conditions favor growth of the PBL than when they favor decay, the PBL does not reach nearly as high on average in the control run where a strong daily cycle of surface buoyancy forcing occurs.