1086 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER 1957 



tom-temperature changes of 1°C have occurred in this area. These tem- 

 perature changes are apparently caused by cold water cascading down 

 the continental slope from the shelf off Greenland. It can be presunuMl 

 that this water will flow south along the deep ocean basin. Dependiiii: 

 on its velocity it will be more or less displaced towards the western 

 margin of the basin. As the water masses flow south, lateral mixing; 

 should reduce the temperature contrast with the surrounding water. I 

 In one area south of Newfoundland a study of scattered temperature ' 

 observations made since the year 1900 indicates that water above 3,000 

 meters has shown a temperature increase and that the water below 

 3,000 meters has gradually decreased in temperature. The decrease at 

 the bottom at a depth of 5,000 meters appears to have amounted to 

 0.2°C in fifty years and the maximum increase at depths less than 3,000 

 meters to about 0.5°C. In the high latitudes of the North Atlantic, tem- 

 perature changes probably rarely exceed 1°C at depths greater than 



2.000 meters and changes of a few tenths of a degree are more common. 



V. CATASTROPHIC CHANGES IN THE OCEAN BOTTOM 



5.1 Earthquakes 



Earthquakes may cause damage to submarine cables by triggering 

 the movements of rock and sediments (turbidity currents) and possibly 

 through the effect of the actual earth vibration itself. The most seriousi 

 threat to cables arises not from the direct effect of the earthquake's' 

 energy on the cable but from its ability to generate slumps, slides, and 

 turbidity currents. It can be shown that, in an area of high seismic 

 activity, the slopes will be nearly bare of loose sediment so that earth- 

 quakes in such an area will not result in gravitational displacements of 

 sufficient size to cause serious damage to cable. 



However, in areas such as the continental slopes of the Atlantic where 

 few shocks have been recorded and where loose sediment has therefore 

 accumulated, it can be expected that any quake of moderate or even 

 small size will be sufficient to generate a turbidity current. Thus, quakes j 

 on continental slopes adjacent to cable routes are likely to be extremely 

 destructive. Thei'e is no known method of predicting where earthquakes 

 will occur outside the major seismic belts. 



Fig. 24 shows the distribution of earthquakes in the North Atlantic. 

 The North Atlantic is the best-monitored ocean because of the extensive 

 network of seismograph stations closely adjacent in North America and 

 Europe. All earthquakes reported for the Atlantic Ocean between 1910 

 and 1956 have been compiled and plotted together with the best bathy 



