8 Satellite Microwave Radar Observations of Antarctic Sea Ice
8.5.2
Motion Spectra and Temporal Covariance with Wind
Ice drift, with respect to previously measured buoy motion in the Antarctic ice cover,
is known to be highly dependent upon the drifter location with respect to the coast, the
primary characteristics of the ocean circulation, and often the water depth. A large distinction can be made between (a) buoys in deep water under relatively free-drift conditions, and (b) buoys drifting in shallow water coastal regimes. The former implies
largely divergent situations which characterize large portion of the Antarctic ice cover, and where buoys generally suffer less from internal ice resistance to drift. The latter, in contrast, occurs on the continental shelf in situations in which periodic high-frequency tidal currents playa significant role in driving shear motions within the ice cover, and where internal ice stresses can be transferred from the coast to the location of
drift during ice convergence events.
8.5.2.1
Buoys Under Relatively Free Orift
Velocity spectra of several freely drifting deep-water buoys in the Weddell Sea are analyzed in conjunction with geostrophic wind data to investigate the correlation and
spectral relationships between forcing and response. Velocity components of several
mid-winter WWGS '92 buoys (with 6 h resolution data) are computed from I-month
July records in 1992, and concurrent geostrophic wind velocity components estimated
from an optimally interpolated pressure field (ECMWF analysis data fitted to weighted buoy pressure measurements).
Figure 9 shows the results of spectral analysis of the magnitude of the complex ice
and wind velocity vectors. Fig. 9a and Fig. 9b indicate the temporal autocorrelation functions of this scalar ice and wind speed, respectively, and show a temporal e-folding
length scale ofless than one day. The cross-covariance function in Fig. 9C demonstrates
clearly that the wind and ice speeds are highly correlated on times cales of ±1 day and
that there is a slight lag of around 12 h or so between the wind forcing and ice drift
response. This confirms that the drift adjusts in a matter of hours to wind forcing,
explaining the short response time exhibited by the large-scale ice motion in the example shown later in Fig. 11. Similarly, Bartlett-window-filtered power spectra (normalized by the velocity variance) in Fig.9d and Fig.ge show clearly that ice drift speed contains the most energy at periods above 100 hours, and the cross amplitude spectrum in
Fig. 9f shows that the spectra are almost identical in characteristics for this 1 month
period. Accompanying coherency and phase spectra of the complex velocities illustrate
how the correlation between the wind and ice drift increases rapidly to a peak exceeding 0.8 at periods of 45-70 hours (i.e., 2-3 days). This corresponds to the frequency of
passing low pressure systems during this winter period, while in contrast the coherency rises to a plateau exceeding 0.8 at periods exceeding 5 days and is comparable to the
results of Kottmeier and Sellmann (1996).
The phase spectrum reveals that ice consistently drifts to the right of the geostrophic wind at all periods of forcing with respect to the upper level winds, yet with some
scatter. July radiosondes launched from R.V. Polarstern in the vicinity of the buoy array
indicate a relatively constant turning angle of 44
0
between surface (10 m) and upper
8.5.2
Motion Spectra and Temporal Covariance with Wind
Ice drift, with respect to previously measured buoy motion in the Antarctic ice cover,
is known to be highly dependent upon the drifter location with respect to the coast, the
primary characteristics of the ocean circulation, and often the water depth. A large distinction can be made between (a) buoys in deep water under relatively free-drift conditions, and (b) buoys drifting in shallow water coastal regimes. The former implies
largely divergent situations which characterize large portion of the Antarctic ice cover, and where buoys generally suffer less from internal ice resistance to drift. The latter, in contrast, occurs on the continental shelf in situations in which periodic high-frequency tidal currents playa significant role in driving shear motions within the ice cover, and where internal ice stresses can be transferred from the coast to the location of
drift during ice convergence events.
8.5.2.1
Buoys Under Relatively Free Orift
Velocity spectra of several freely drifting deep-water buoys in the Weddell Sea are analyzed in conjunction with geostrophic wind data to investigate the correlation and
spectral relationships between forcing and response. Velocity components of several
mid-winter WWGS '92 buoys (with 6 h resolution data) are computed from I-month
July records in 1992, and concurrent geostrophic wind velocity components estimated
from an optimally interpolated pressure field (ECMWF analysis data fitted to weighted buoy pressure measurements).
Figure 9 shows the results of spectral analysis of the magnitude of the complex ice
and wind velocity vectors. Fig. 9a and Fig. 9b indicate the temporal autocorrelation functions of this scalar ice and wind speed, respectively, and show a temporal e-folding
length scale ofless than one day. The cross-covariance function in Fig. 9C demonstrates
clearly that the wind and ice speeds are highly correlated on times cales of ±1 day and
that there is a slight lag of around 12 h or so between the wind forcing and ice drift
response. This confirms that the drift adjusts in a matter of hours to wind forcing,
explaining the short response time exhibited by the large-scale ice motion in the example shown later in Fig. 11. Similarly, Bartlett-window-filtered power spectra (normalized by the velocity variance) in Fig.9d and Fig.ge show clearly that ice drift speed contains the most energy at periods above 100 hours, and the cross amplitude spectrum in
Fig. 9f shows that the spectra are almost identical in characteristics for this 1 month
period. Accompanying coherency and phase spectra of the complex velocities illustrate
how the correlation between the wind and ice drift increases rapidly to a peak exceeding 0.8 at periods of 45-70 hours (i.e., 2-3 days). This corresponds to the frequency of
passing low pressure systems during this winter period, while in contrast the coherency rises to a plateau exceeding 0.8 at periods exceeding 5 days and is comparable to the
results of Kottmeier and Sellmann (1996).
The phase spectrum reveals that ice consistently drifts to the right of the geostrophic wind at all periods of forcing with respect to the upper level winds, yet with some
scatter. July radiosondes launched from R.V. Polarstern in the vicinity of the buoy array
indicate a relatively constant turning angle of 44
0
between surface (10 m) and upper