We document a new middle atmosphere general circulation model that includes ozone photochemistry. The dynamical model component is based on the NCAR middle atmosphere version of the Community Climate Model. The chemistry model component simulates the evolution of 24 chemically reactive gases. The horizontal resolution is approximately 3¿ in latitude and 6¿ in longitude. It includes 44 levels, with a maximum vertical grid spacing of about 2.5 km and a top level at around 75 km. The chemical model distinguishes between species where we judge transport to be critical and those for which it may be neglected. Nine longer-lived species (N2O, CH4, H2O, HNO3, N2O5, CO, CIONO2, HCl, and HOCl) and four chemical families (NOy, NOx, Ox and Clx) are advected. Concentrations of 15 species which are typically shorter-lived or are members of the chemical families are diagnosed using quasi-equilibrium assumptions (O(1D), OH, Cl, O(3P), O3, HO2, NO2, ClO, NO, HNO4, NO3, N, OClO, Cl2O2, H2O2). Distributions for a number of other species are prescribed. Results are presented from a 2-year simulation, which include only gas phase photochemical reactions and in which the ozone distribution forecast from the chemistry module does not affect the radiative forcing of the dynamical fields. The calculated distributions of trace species and their seasonal evolution are often quite realistic, particularly in the northern hemisphere extratropics.

Distributions of long-lived species such as N2O and CH4 correspond well to satellite observations. Some features, such as the double peak structure occurring during equinoxes, are not reproduced. The latitudinal variation and seasonal evolution of the ozone column abundance is quite realistic. The calculated vertical distribution of the ozone mixing ratio exhibits significant differences from measured values. The model underestimates significantly the ozone in the upper stratosphere (40 km) and in the extratropics, where the maximum values occur at too low an altitude. The model reproduces the key features expected in the distribution of fast reacting nitrogen and chlorine compounds. Problems in the southern polar simulation (which we attribute to the lack of a comprehensive parameterization of gravity waves) are discussed. The model currently parameterizes only stationary gravity waves generated by flow over orography.Other sources of gravity waves may be required to improve the model simulation. ¿ American Geophysical Union 1995