We describe the spatial variability of high frequency ice velocity in the W
eddell Sea using satellite-tracked ice-mounted buoys. Ice motion is analyze
d separately for "diurnal" (1/36-1/18 cph) and "semidiurnal" (1/18-1/6 cph)
bands. Ice motion in both bands is caused by a combination of ocean tidal
currents and wind stress. Monthly mean diurnal band ice speeds over the dee
p basin range from 2 to 4 cm s(-1) depending on wind stress variance and ic
e concentration (C-ice). Higher speeds (similar to 10 cm s(-1) are found in
the semidiurnal band in regions of low C-ice, notably the northern Weddell
Sea, where the ice velocity is dominated by the inertial response to wind
stress variations. Monthly mean tidal band ice speeds over the continental
slope and shelves often exceed 10 cm s(-1). We use comparisons between buoy
velocities, moored current meter data, and an ocean tidal model to demonst
rate that ice motion is frequently a good indicator of ocean tidal currents
in strongly tidal regions. The standard deviation of the divergence of oce
an tidal currents estimated from an ocean-only tidal model is small (<0.1x1
0(-6) s(-1)) over most of the Weddell Sea but has values of 1-5x10(-6) s(-1
) along the Ronne Ice Front and the continental shelf break. High frequency
ice divergence is dominated by ice response to wind stress rather than by
tides except along the shelf break and ice fronts. In these tidally dominat
ed regions the periodic divergence maintains a mean lead (open water) area
of similar to 2-5%. This increased lead fraction implies an increase in are
a-averaged winter ocean-to-atmosphere heat exchange rate of similar to 4-10
W m(-2) and an increase in salt flux into the upper ocean as new ice forms
in the leads.