Reverse-trajectory calculations initialized with ozone observed by the Upper Atmosphere Research Satellite Microwave Limb Sounder (MLS) provide high-resolution ozone profiles for comparison with lidar and ozonesonde observations from the Arctic Stratospheric Observatory facility near Eureka in the Canadian Arctic (~80 ¿N, 86 ¿W). By statistical measures, calculated profiles show a small average improvement over MLS profiles in the agreement of small-scale structure with that in ground-based observations throughout the stratosphere, and a larger (although still modest) and more consistent improvement in the lower stratosphere. Nearly all of the calculated profiles initialized with daily gridded MLS data show some improvement in the lower stratosphere. Even in cases where overall agreement between profiles is mediocre, there are frequently one or more individual features in the calculated profiles that strongly resemble laminae in the ground-based observations. Differential advection of ozone by the large-scale winds leads to lamination in three distinct ways. Filamentation results in lamination throughout the stratosphere, with comparable features arising from initializations with gridded MLS data and with potential vorticity/&thgr;-space reconstructions of MLS data (reconstructed (RC) fields). Laminae also form in the middle and lower stratosphere in conjunction with intrusions into the vortex; while calculations initialized with RC fields produce laminae, the agreement of structure calculated using gridded MLS initialization data with ground-based observations is distinctly better. Inside the lower stratospheric vortex, laminae form by advection of local features in the MLS initialization fields; RC-initialized calculations fail to produce any significant features since these local ozone variations are not strongly correlated with potential vorticity. That local features observed by MLS are needed to produce laminae resembling those in independent ground-based observations at Eureka indicates that both datasets are capturing real atmospheric features. ¿ 1998 American Geophysical Union