Thickness and Roughness Variations of Arctic Multiyear Sea Ice

• Main Authors: Ackley,S. F., Hibler,W. D. , III, Kugzruk,F. K., Kovacs,A., Weeks,W. F.
• Other Authors: COLD REGIONS RESEARCH AND ENGINEERING LAB HANOVER N H
• Format: Text
• Language: English
• Published: 1976
• Subjects: Snow; Ice and Permafrost; *SEA ICE; MATHEMATICAL MODELS; DENSITY; THICKNESS; LOAD DISTRIBUTION; TIME DEPENDENCE; SURFACE ROUGHNESS; ANALYSIS OF VARIANCE; DEPTH; CRACK PROPAGATION; PROFILES; ELEVATION; SALINITY; BENDING STRESS; ARCTIC OCEAN; ISOSTATIC CHANGE; Ice mechanics; Multiyear sea ice; Ice floes; Arctic; Arctic Ocean; Ice; permafrost; Sea ice
• Online Access: http://www.dtic.mil/docs/citations/ADA028086; http://oai.dtic.mil/oai/oai?&verb=getRecord&metadataPrefix=html&identifier=ADA028086

Three surface elevation and ice thickness profiles obtained during the 1972 Arctic Ice Dynamics Joint Experiment on a multiyear ice floe were analyzed to obtain relationships between surface elevation, thickness and physical properties of the ice. It was found that for ice freeboards from 0.10 m to 1.05 m above sea level a linear relationship between ice density and freeboard could be postulated. The equation for the regression line is: Ice density = -194f' + 974 kg/cu m where f' is the ice freeboard plus snow depth in ice equivalent at the point in question. This statistical relationship is consistent with observed physical properties, which indicate that as the ice freeboard increases, ice salinity decreases and the higher freeboard or thicker ice therefore decreases in density. Using this variable density with freeboard relationship, a model was constructed to predict ice thickness, given ice freeboard and snow depth alone. This prediction is desirable, since snow depth and freeboard are relatively easy to obtain, whereas ice thickness can usually be obtained only by drilling through the ice. The model was compared with two other models. It was found that the variable density prediction model gave the best approximation to observed ice thickness, with a standard error between the measured and predicted value of about 0.4 m, compared with errors from 50 to 100% higher for the other two models.