A tropospheric chemistry model has been developed within the Goddard Institute for Space Studies general circulation model (GCM) to study interactions between chemistry and climate change. The model uses simplified chemistry based on CO-NOx-HOx-Ox-CH4 and also includes a parameterization for isoprene emissions, the most important non-methane hydrocarbon. The model reproduces observed annual cycles and mean distributions of key trace gases fairly well. It simulates preindustrial to present-day changes similar to those seen in other simulations. For example, the global tropospheric ozone burden increases 45%, within the 25%--57% range of other studies. Annual average zonal mean surface ozone increases more than 125% at northern midlatitudes. Comparison between runs that allow calculated ozone to interact with the GCM and those that do not shows only minor ozone differences. The common usage of non-interactive ozone seems adequate to simulate ozone distributions. However, use of coupled chemistry does alter the tropospheric oxidation capacity, enlarging the preindustrial to present-day OH decrease by about 10% (-5.3% global annual average uncoupled, -5.9% coupled). Thus simulation of changes in oxidation capacity may be systematically biased (though a 10% difference is within the uncertainty). Global annual average radiative forcing from preindustrial to present-day ozone change is 0.32 W m-2. The forcing seems to increase by ~10% with coupled chemistry. Forcing greater than 0.8 W m-2 is seen over the United States, the Mediterranean area, central Asia, and the Arctic, with values greater than 1.5 W m-2 over parts of these areas during summer. Though there are local differences, the radiative forcing is overall in good agreement with other modeling studies in both magnitude and spatial distribution, demonstrating that the simplified chemistry is adequate for climate studies. ¿ 2001 American Geophysical Union