The authors simulate transient twentieth-century climate in the Goddard Institute for Space Studies (GISS) GCM, with aerosol and ozone chemistry fully coupled to one another and to climate including a full dynamic ocean. Aerosols include sulfate, black carbon (BC), organic carbon, nitrate, sea salt, and dust. Direct and BCsnow-albedo radiative effects are included. Model BC and sulfur trends agree fairly well with records from Greenland and European ice cores and with sulfur deposition in North America; however, the model underestimates the sulfur decline at the end of the century in Greenland. Global BC effects peak early in the century (1940s); afterward the BC effects decrease at high latitudes of the Northern Hemisphere but continue to increase at lower latitudes. The largest increase in aerosol optical depth occurs in the middle of the century (1940s–80s) when sulfate forcing peaks and causes global dimming. After this, aerosols decrease in eastern North America and northern Eurasia leading to regional positive forcing changes and brightening. These surface forcing changes have the correct trend but are too weak. Over the century, the net aerosol direct effect is 20.41 W m22, the BC-albedo effect is 20.02 W m22, and the net ozone forcing is 10.24 W m22. The model polar stratospheric ozone depletion develops, beginning in the 1970s. Concurrently, the sea salt load and negative radiative flux increase over the oceans around Antarctica. Net warming over the century is modeled fairly well; however, the model fails to capture the dynamics of the observed midcentury cooling followed by the late century warming. Over the century, 20% of Arctic warming and snow–ice cover loss is attributed to the BCalbedo effect. However, the decrease in this effect at the end of the century contributes to Arctic cooling.
To test the climate responses to sulfate and BC pollution, two experiments were branched from 1970 that removed all pollution sulfate or BC. Averaged over 1970–2000, the respective radiative forcings relative to the full experiment were 10.3 and 20.3 W m22; the average surface air temperature changes were 10.28 and 20.038C. The small impact of BC reduction on surface temperature resulted from reduced stability and loss of low-level clouds.
