Ice in the Sea 281 



adds only about 1-9 km^. This is the amount of land ice that exists in Baffin Bay on a 

 yearly average and drifts southward to melt in Davis Strait, along the Labrador coast 

 and in the Newfoundland area. The amount of sea ice melting during one year can be 

 calculated from the average area covered by pack ice and drift ice. Smith has made an 

 estimate of this kind based on reliable data collected by the Ice Patrol cruises. The bases 

 of this are contained in Fig. 131 which also shows the areas which stand in question; 

 the most important are the shelf areas where the ice-covered area is about 1-6 million 

 km^. Taking the mean thickness of drift and pack ice as about 1-8 m, the total amount 

 of sea ice will be about 3000 km^. In contrast to this, the land ice amounts to only 

 44-65 km^, so that of the average annual amount of ice melting in the north-west 

 Atlantic only between a hundredth and a two-hundredth part comes from icebergs. 

 This is vanishingly small (see Fig. 131). This comparison shows that the amount of 

 pack ice and drift ice is the decisive factor. If for any oceanographic or meteorological 

 problem a consideration of the effects of ice destruction in the north-west Atlantic — 

 which vary considerably from year to year — is needed, it is thus not justifiable to 

 compare it with variations in the ice frequency, as has often erroneously been done. 



In dealing previously with convection processes (see p. 97) two possibilities were 

 discussed for the initiation of such a process, which are of the greatest importance to 

 the thermal structure of the middle and bottom layers of the oceans. It was assumed 

 by Pettersson that the necessary heat loss of the upper water layers was mainly due 

 to the melting of ice in polar and subpolar oceanic regions. However, laboratory 

 experiments by Nansen showed that this hypothesis was untenable. For the special 

 case of the conditions in the north-west Atlantic it is possible, using the values given 

 by Smith to determine directly the amount of heat which is required for the observed 

 yearly melting of pack ice and drift ice and therefore is not available for heating the 

 ocean and the atmosphere. This can be compared, as has been done by Smith, with 

 the amount of heat suppHed during the summer by solar and sky radiation which is 

 required for the increase in temperature of the upper 150 m layer of water (the 

 average depth to which the increase reaches downwards into the sea). From the num- 

 bers given in Fig. 131 it can be seen that the mean summer increase in the tempera- 

 ture of the water masses in this area (down to 150 m) is about 1-2°C. It can also be 

 calculated that the annual melting of pack ice and drift ice in the same area is sufficient 

 to decrease the temperature of the layer down to 150 m depth by 0-6 °C. Thus in the 

 north-western part of the North Atlantic the water is cooled by the melting of the 

 ice by only about half of the amount of the summer increase in temperature due to 

 the absorption of solar and sky radiation. Dynamic treatment of the oceanographic 

 data of the "Marion" and "Godthaab" Expeditions permits the calculation of the 

 amount of the heat deficit at the Newfoundland Banks due to the continuous supply 

 of cold polar water by the Labrador Current. Comparison of this heat reduction with 

 that due to ice melting shows that the latter accounts for only 10% of the cooling 

 effect of the Labrador Current. The dominant factor in the cooling of the water masses 

 of the northern part of the North Atlantic is thus neither the mehing of icebergs nor 

 of the pack ice and drift ice, but much more the continuous advective supply of 

 polar water which the Labrador Current carries southwards towards the warm water 

 of the Gulf Stream. 



The "Meteor" cruise in Icelandic and Greenland waters have given the same 



