An important consideration in understanding sea ice mechanics is the i
ntegration of observed sea ice behavior on a flee neighborhood scale (
1-10 km) into ice dynamics on a regional scale O(50 km). We investigat
e sea ice kinematics from October 1993 through April 1994 using relati
ve motions from 13 drifting buoys with Global Positioning System navig
ation in a 20-km array centered on the Sea Ice Mechanics Initiative ic
e camp, and we compare these motions to synthetic aperture radar (SAR)
derived ice velocities over a 100- by 500-km region in the Beaufort Se
a. There is excellent correspondence between the deformation of the bu
oy array and that from the SAR. Inferred ice dynamics from analysis of
the two major northerly wind convergence events of the winter are con
sistent with a granular hardening plastic conceptual model for Beaufor
t sea ice. Under continued northerly winds the ice from the Alaskan sh
ore to the camp failed in shear and convergence, in a progressive mann
er away from the coast. The continuum scale O(10 km) is an order of ma
gnitude larger than the grain, i.e., flee, size O(1 km). The ice motio
n often forms aggregates of 20-200 km separated by narrow (<10 km) she
ar zones, similar to granular materials. At moderate forcing, i.e., wi
nd stress multiplied by fetch, the ice appears to fail along slip line
s that occur at an acute angle to each other and to the direction of t
he wind forcing, characteristic of a plastic material at critical stat
e. With longer fetch the ice appears to fail in compression, perpendic
ular to the wind direction. Sea ice appeared to harden on a regional s
cale after the first event. During the second northerly wind event the
re was a sea ice breakout toward the west, apparently due to a lack of
lateral confining stress. Our observations suggest that the ice flees
advect through relatively stationary stress fields, created by the wi
nd forcing and coastal boundaries. For example, while the SAR and adva
nced very high resolution radiometer images indicated the presence of
the shear feature at the same geographic location for nearly a week, b
uoys would show shearing only for several days as they transited acros
s the region of shear. There is a high correspondence between the majo
r internal ice deformation events and persistent weather patterns on a
3- to 5-day temporal scale. This implies that SAR data collection and
analysis for regional sea ice dynamics should be consistent with the
wind forcing and have a sampling of less than 3 days.