Infrared Parameterization and Simple Climate Models

Abstract

Empirical studies of total outgoing infrared radiation IR and surface temperature T have shown them to be well correlated for large time and space scales. An analysis of one year of Nimbus-6 data shows that the simple form IR = A + BT (with A = 204 W m⁻², B = 1.93 W m⁻²K⁻¹) explains 90% of the area-weighted variance in the annual mean and annual cycle of the zonally averaged IR field. The geographical distribution of the annual cycle in IR shows a large amplitude over the continental interiors, as is found in the observed temperature field, and the ratio of the large amplitudes (B_local) is approximately 2 W m⁻²K⁻¹. This helps to explain our recent success in modeling the geographical distribution of the annual cycle in T with a two-dimensional, time-dependent energy balance climate model (EBCM) which makes use of the A + BT rule. The parameterization works well in regions where the thermal inertia is small and the annual cycles of T and IR are large and in phase. Those regions where B_local differs markedly from 2 W m⁻²K⁻¹ are where the IR is strongly affected by the cloudiness of seasonal precipitation regimes. This effect is especially evident over the tropical oceans where the parameterization fails; but that is where the thermal inertia is large, the seasonal cycle in T is small, and even large errors in the radiative cooling approximation will have little impact on seasonal cycle simulations by simple climate models.