SECT. 7] THE PHYSICS OF SEA-ICE 831 



precipitated sodium sulphate increase the strength by about one-third. It will 

 be noted that in the region of "normal strength" (temperatures between —8° 

 and — 23°C and salinities less than 10%o), the temperature dependence is small, 

 salinity being a bigger factor. Sea-ice in this region has a tensile strength 

 varying from about that of fresh ice (about 1.7 x 10'^ dynes cm"^) down to 

 about 60 or 65% of this value. Observations made by SIPRE (Snow Ice and 

 Permafrost Research Establishment) at various locations and temperatures 

 give a range of tensile strength from below 1 x 10^ to over 3 x 10'^ dynes cm~2. 

 All figures refer to small scale tests. 



Less data are available on shear and crushing strengths. Pure shear stress is 

 difficult to apply to ice and none of the methods used is completely satisfactory. 

 Butkovich (1956) used a double shear device in which three snugly fitting, 4-in. 

 long metal cylinders were slid over a 3-in. ice core. With the outer cylinders 

 fixed, a force transverse to the axis was applied to the middle cylinder and 

 increased until fracture occurred. This method gave shear strengths of the 

 order of 2.3x10'^ dynes cm~2 at — 11°C, which is higher than the tensile 

 strength, but the method has been criticized as giving too much support to the 

 ice cyHnder. Most observers report shear strengths lower than tensile strengths. 

 Small-scale, unconfined compression tests are simple to carry out but results 

 invariably show wide scatter. The order of magnitude for compressive strength 

 in the "normal strength" range is 1.2 x 10^ dynes cm~2. 



In situ beam tests usually give flexural strengths of the order of 3 x 10^ c.g.s. 

 The ice beam in these tests has a temperature gradient across it from about 

 — 2°C (in contact with the water) to the air temperature, about — 10°C for the 

 result quoted. 



Representative papers on strength measurements are Assur (1958), Arnol'd- 

 Albiab'ev (1939), Butkovich (1956), Langleben (1959) and Petrov (1955). 



B. Elastic Parameters 



Measurements of the elastic parameters of sea-ice can also be divided into 

 small-scale or large-scale in situ tests, with a further division into static and 

 dynamic methods. Anderson (1958a) reviews the various methods. Because of 

 the visco-elastic behaviour of sea-ice, static loading tests usually give great 

 scatter and are of limited value. Dynamic methods include mechanical vibration 

 of beams, resonance vibration of beams under sonic excitation, pulse methods, 

 in which velocities of longitudinal and transverse waves at supersonic fre- 

 quencies are measured, and seismic methods. The most consistent results are 

 found from acoustic and seismic methods. Values reported for Young's modulus 

 range from about 1.5 x lO^o to 9.1 x lO^^ dynes cm~2 (see Brown and Howick, 

 1958; Oliver e^ aL 1954; Peschansky, 1958; Pomeroy, 1956; and Pounder and 

 Stalinski, 1960a). This parameter is low in the autumn on newly frozen ice and 

 even lower in the spring on deteriorating ice. The maximum values are for cold, 

 hard winter ice. No more specific data on temperature dependence are available. 

 Anderson has plotted the values obtained by several workers against calculated 



