Geoscience Reference
In-Depth Information
Chapter 2
), the Fram Strait outflow merits special attention. A number of efforts
have been made to estimate the ice area or volume flux through the Fram Strait based
on mass balance requirements, measurements of ice drift, and thickness through the
strait, models or a combination of models and observations. The total ice flux is
comprised of a wind-driven and a current-driven component. Determining the vol-
ume flux requires accounting for the ice velocity through the strait, the ice concen-
tration and ice thickness.
Some early estimates based on direct measurements include Wadhams (
1983
)
and Vinje and O. Finnekasa (
1986
). Vinje and Finnekasa (
1986
) also used an indi-
rect method to estimate the flux. A regression equation was developed between
available time series of the observed monthly averaged drift speed profile across
the strait and the monthly difference in sea level pressure between the Fram Strait
(81°N, 15°W) and the central area of the Norwegian Sea (73°N, 5°E). The slope
of the regression equation represents the fraction of the ice velocity attributed to
time-varying winds, and the y-intercept represents the constant effect of the ocean
current. The regression equation was used to compile volume flux time series by
incorporating climatological monthly values of the ice thickness and concentration.
The advantage of this approach is that, because long time series of sea level pressure
are available, long time series of the estimated flux can be obtained.
Time series of estimated ice concentration and velocity are available from sat-
ellite data since the late 1970s. Moored upward-looking sonar (ULS) sites in and
around the strait have provided estimates of ice thickness since about 1990. R.
Kwok and D. Rothrock (
1999
) took advantage of satellite and ULS data to obtain
the cold-season (October-May) area flux over the 1978-1996 period. Ice veloci-
ties across the strait were obtained from an ice tracking procedure using sequential
images from SMMR and SSM/I. The derived velocities were used in conjunction
with SMMR and SSM/I ice concentration data to obtain area flux time series. They
obtained monthly volume fluxes by adjusting the velocities based on profiles of
mean monthly ice thickness from ULS. However, their technique was not applied to
summer months because of biasing of the passive microwave data by melt effects.
Vinje (
2001
) used a regression-based approach along with the best available esti-
mates of ice velocity and ice thickness from SAR, buoy drifts and ULS to provide
an ice volume record extending from 1950 to 2000. The mean annual flux is about
2900 km
3
. Using nine years of ULS data, Kwok, G. Cunningham, and S. Pang
(
2004
) report a somewhat lower value of 2,200 km
3
per year. Vinje estimated that
60 percent of the total annual ice volume flux is wind driven, and about 40 percent
is attributed to the background ocean current. The mean cross-strait ice thickness is
2.56 m. A striking feature of the outflow based on this study is its pronounced annual
cycle (
Figure 7.21
). The largest volume fluxes are found in autumn and winter,
when there are strong northerly winds in the strait. The flux is smallest in summer
when the northerly winds slacken and are replaced by weak southerlies in July and
August. A comparison between the two lines in
Figure 7.16
illustrates the effects of
adjusting for the mean annual cycle of ice thickness. The adjustment accounts for
the thicker ice observed from March through July and the thinner ice for the rest of
the year. A conspicuous feature of Vinje's (
2001
) reconstructed monthly time series
