The Trajectory Assimilation Model (TRAM) is a Lagrangian model of Arctic sea ice that consists of cells, each of which is defined by the attributes of position, velocity, mean ice thickness, and ice concentration. The continuum dynamic sea-ice model is forced by geostrophic winds and surface ocean currents from Eulerian model output. The ice internal stress is determined from a viscous plastic rheology using strain rates found with the Smooth Particle Hydrodynamics formalism.

Radarsat Geophysical Processor System (RGPS) data consist of thousands of trajectories with positions determined roughly every 3 days. The irregular temporal and spatial sampling of the velocity field make the data awkward to use in model intercomparison studies. The TRAM model is used to regularize the temporal sampling by associating each RGPS trajectory with a model cell and forcing the model cell to follow the observed trajectory. This is accomplished through the use of a corrective force that is proportional to the difference in position between the model trajectory and the observed trajectory. We will show results for interpolating the RGPS data to a regular spatial grid with daily time steps over a one year period, the SHEBA year. In addition, the model stress field associated with the observed strain rates and the model ice thickness field will be described.