In October 1976 a cooperative experiment was made to survey the sea-ice topography in the European sector of the Arctic Ocean. H. M. submarine
Sovereign
acquired 4000 km of ice draft data by using an upward-looking sonar, while a Canadian Forces aircraft flew along the submarine’s track and acquired 2200 km of ice elevation data by using a laser profilometer. The two types of profile were processed in corresponding 100 km section lengths, and the following statistical analyses and comparisons were made: (i)
Probability density functions o f ice draft and elevation.
Each distribution shows a peak for young ice and for undeformed multi-year ice. At large ice thicknesses the distributions take the analytical form of a negative exponential. The mean drafts enable two distinct geographical ice regimes to be identified. There is an ‘offshore zone’ of very heavy pressure ridging extending up to 400 km from the coasts of Ellesmere Island and of north Greenland, with mean ice draft in the range 5.0 to 7.5 m , while out in the central Arctic Ocean the mean ice draft is lower (3.9—5.1 m) and the characteristics of the ice cover remain homogeneous over a length scale of 1000 km. The transition between the two regimes is abrupt, taking place in less than 25 km. Data from the same part of the central Arctic taken in March 1971 showed a mean ice draft 0.3 m lower, while data from the central Beaufort Sea showed a mean draft more than 0.8 m lower. (ii)
Level ice distributions.
Ice with a local gradient of less than 1 in 40 was defined as level ice, and used as an indicator of the quantity and thickness distribution of undeformed (i.e. thermodynamically grown) ice in the Arctic Ocean. The distribution has a mode at 3.0—3.1 m draft, and level-ice percentages are in the range 30—40 (bottom side) and 70—80 (top side) in the offshore zone, and 45—55 (bottom) and 85-95 (top) in the central Arctic. Thus about half of the Arctic ice cover consists of deformed ice. (iii)
Pressure ridge spacings.
The spacings of ridge keels fit a negative exponential distribution, characteristic of randomness, except at close spacings where there is a deficit of keels (explained as a geometrical effect) and at very large spacings where there is an excess (due to the contribution of polynyas). The distribution of sail spacings exhibits these two effects, but also differs from a random distribution at moderate sail separations. (iv)
Ridge elevations and drafts.
Keel drafts fit a law of form P(
h
) d oc exp ( — Ah
2
) d
h
, except for an excess of keels at drafts beyond 20 m. There is a positive correlation between mean keel draft and keel frequency. There are 3.5—4.5 keels per kilometre with draft exceeding 9 m in the offshore zone and 2—3 in the central Arctic. Sail elevations fit a law of form P(
h
) d
h
x exp ( — A
h
) d
h
, with a positive correlation between mean sail elevation and sail frequency. The sail elevation and ice elevation distributions can be related by assuming that all thick ice is contained in pressure ridges of triangular cross section. (v)
Keel—sail comparison.
For the 21 corresponding 100 km sections there are positive correlations between mean sail height and mean keel draft, and between keel frequency and sail frequency. From these it is possible to convert a sail distribution (computed from a laser profile) into a keel distribution, enabling sea ice bulk characteristics to be derived from airborne surveys alone. (vi)
Leads and polynyas
. A lead was defined as a continuous sequence of depth points greater than 5 m long and not exceeding 1 m draft. The number density
n
(
d
) of leads per kilometre of width
d
m fits the power law
n
(
d
) — 15 d
-2
. Exceptionally wide leads were concentrated in the offshore zone and in the marginal ice zone close to the open water of the Greenland Sea.
A NOTE ON SUFFICIENT CONDITIONS FOR NEGATIVE EXPONENTIAL POPULATION DENSITIES