Some Experiments with a Zonally Averaged Climate Model

Abstract

A simple, zonally averaged numerical model is developed for simulating certain features of the annual mean climate of the Northern Hemisphere. The model is based on the two-level quasi-geostrophic potential vorticity system of equations and a surface beat balance equation. Ale main output consists of the latitudinal variations of temperature at the surface and 500 mb and of zonal wind at 250 and 750 mb. Meridional transport of quasi-geostrophic potential vorticity is simulated by an eddy ~ process using exchange coefficients based on observational data. Solar radiative processes included are absorption by water vapor, ozone and cloud particles, scattering by air molecules and clouds, and reflection by the surface. Longwave radiative processes include absorption and emission by water vapor, carbon dioxide and clouds. Other beat transfer processes?convection, evaporation, condensation and ocean currents?are pammeterized. Using present boundary conditions, the model is used to compute the present climate. Comparison of the computed climate with the observed climate shows good agreement. A special attempt is made to compare some of the radiation quantities computed by the model with satellite observations and radiation budget calculations. The sensitivity of the computed climate to changes in some of the boundary conditions is investigated. These sensitivity experiments are performed with and without an ice feedback mechanism. The ice feedback mechanism is based on empirical relations between the fractions of a latitude belt covered by snow and ice in winter and slimmer and the mean annual surface temperature. When the atmospheric carbon dioxide content is doubled, the hemispheric mean surface temperature increases by 0.5°C in the absence of ice feedback, the largest increases taking place at high latitudes. Ice albedo feedback amplifies the hemispheric average temperature change by about 50%; amplifications as large as several hundred percent are obtained in polar regions. A change in mean surface temperature of ±1°C for a ±1% change in solar constant is obtained in the absence of ice feedback, but this is amplified to ?1.5°C (decreased solar constant) and 1.4°C (increased solar constant) when ice feedback is included. As In the 2?CO2 case, polar amplification factors due to ice albedo feedback are several hundred percent. When hemispheric cloud amount is increased, the surface temperature decreases but in the absence of ice feedback the magnitude of the change approaches zero near the poles. A hemispheric increase in the altitude of the cloud layer causes an increase in surface temperatures. These results are compared with those obtained with other climate models.