The Canadian Regional Climate Model with oxidant chemistry (CRCM) has been used to study the climatology and transport of ground-level ozone in eastern North America. The climatology is derived from simulations of five consecutive Julys from 1990 to 1994. Biogenic emissions are calculated each time step, while anthropogenic emissions are taken to be identical for each simulated July. Comparisons of daily ozone maximum for each individual month suggest that the model is simulating satisfactorily the month-long periods and usually correctly reproduces the high and low ozone regimes associated with the different parts of the domain and/or the various synoptic conditions, the Ohio region being the exception. The climatological number of exceedences above 80 ppbv, calculated as the monthly average for the 5 years, yields better agreement with the number of exceedences calculated from the observed climatological O3 field than those for the individual months. This behavior suggests that a large fraction of the errors present in each simulation are randomly distributed. Thus it should be possible to obtain fields that are more representative of reality, for applications such as emission reduction scenario studies, by averaging over several long simulations. ¿ 1999 American Geophysical Union