The diurnal cycle of surface air temperature in simulated present and doubled CO 2 climates

Abstract

The diurnal range of surface air temperature (rT a ) simulated for present and doubled CO2 climates by the CSIRO9 GCM is analysed. Based on mean diurnal cycles of temperature and surface heat fluxes, a theory for understanding the results is developed. The cycles are described as the response to a diurnal forcing which is represented well by the diurnal mean flux of net shortwave radiation at the surface (SW) minus the evaporative (E) and sensible (H) fluxes. The response is modified by heat absorbed by the ground, and by the cycle in downward longwave (LW) radiation, but these effects are nearly proportional to the range in surface temperature. Thus in seasonal means, rT a is approximately given by SW–E–H divided by 6 W m-2/°C. A multiple regression model for (rT a ) is developed, based on quantities known to influence SW, E and H, and applied to both spatial variation in seasonal means, and day-to-day variation at a range of locations. In both cases, rT a is shown to be influenced by cloud cover, snow extent and wind speed. It is influenced by soil moisture, although this effect is closely tied to that of cloud. In seasonal means rT a is also well correlated with precipitable water, apparently because of the latter’s influence on E+H. The regression model describes well the spatial variation in the doubled CO2 change in rT a . The annual mean change in rT a over land on doubling CO2 was −0.36 °C, partly because of a decrease in the mean diurnal forcing (as defined in the theory), but also apparently because of the effect of nonlinearity in T s of the upward longwave emission. A diagnostic radiation calculation indicates that the CO2 and water vapour provide a small increase in rT a through the downward LW response, which partially counters a decrease due to a reduction of SW by the gases.