Data from drifting sea-ice buoys, available during various seasons since 1985 for the Southern Ocean off East Antarctica, have previously been used to characterize the mean sea-ice motion in the region. This mean sea-ice velocity field exhibits intra-annual (winter versus spring/summer/autumn) and also interannual variability. Here we use those data to focus on the composition of the variable component in the ice-velocity field and its statistical representation. Based on the Kolmogorov theorem we interpret the variable component as ‘turbulent diffusion’. In previous studies we estimated that about 58–85% of the ice-motion variability was associated with atmospheric forcing, with lesser contributions due to oceanic forcing. However, variability in differential sea-ice motion cannot be well explained by external forcing, suggesting that nonlinear internal processes, such as crack formation and lead propagation, play in integral role in sea-ice dynamics.