We present a Lagrangian model-based technique to reconstruct ozone profile and column in support of the interpretation of ozone lidar observations made at the Arctic Lidar Observatory for Middle Atmosphere Research (ALOMAR). High-resolution ozone profiles as well as column ozone are reconstructed locally at ALOMAR, several times a day, regularly throughout the winter and spring of 1997/1998. The approach consists of calculating a large number of back trajectories to determine the origin of air parcels above ALOMAR and of using satellite observations to determine their ozone content. The blend of satellite observations with limited spatial and temporal coverage, and global gridded meteorological data is akin to a simplified form of data assimilation. The usefulness of the approach is demonstrated by a systematic comparison between reconstructed and observed ozone profiles and column and, in particular, their day-to-day variability. Abrupt changes in reconstructed and observed profile shape are caused by polar vortex displacements, deep intrusions of midlatitude air, or vortex edge filamentation. Prominent laminae are seen in the spring, as the vortex breaks down. Short-lived, large-amplitude total ozone peaks are seen in the model and observations, most prominently in February and March, when they are in excess of 100 Dobson units. They are shown to result from a combination of a lowering of isentropes in the lowermost stratosphere and of advection of ozone-rich air from the base of the polar vortex. ¿ 2001 American Geophysical Union