SKOT. 2] GASKS 317 



steady-state mixing in the ocean, with ])eriods of ra])id mixing alternating with 

 periods during which no mixing occurs. Duration of periods is assumed to be 

 of the order of 100 years. Fergusson (1958) beheves that the residence time is 

 less than 5 years, based on observation of local changes in oceanic ^'^C in the 

 surface-water layers. Also Bolin and Eriksson (1959) favor a residence time of 

 five years. 



Over the last five years a large number of ^^C determinations of the dissolved 

 bicarbonate in ocean water have been carried out by the Lament Geological 

 Observatory (Broecker and Olson, 1959), by the New Zealand Radiocarbon 

 Laboratory (Burling and Garner, 1959) and at the Scripps Institution of 

 Oceanography (Bien, Rakestraw and Suess, 1960). The new data, however, 

 tend to demonstrate the uncertainties in our knowledge rather than to allow a 

 more precise evaluation. The most important new observation is that i'*C 

 concentrations in Pacific surface and deep waters north of the Antarctic con- 

 vergence differ by a larger factor than in the surface and deep water in the 

 Atlantic. 



Fig. 2, taken from a paper by Bien, Rakestraw and Suess (1960), illustrates 

 the situation. The radiocarbon concentrations in the Atlantic cover a much 

 more narrow range, indicating faster mixing. The situation in the Pacific is 

 simplified by the fact that cold deep and bottom water is supplied only from 

 the Antarctic. The northward movement of water at 3500-m depth can be 

 recognized by an increase in apparent i^C-age from 1500 years at latitude 30°S 

 to 1700 years at latitude 30°N. The fact that deep Pacific water shows a greater 

 apparent age than Atlantic water suggests a somewhat slower exchange and a 

 longer residence time of CO2 in the atmosphere than was assumed previously. 



A second new observation that demonstrates our ignorance is that the 

 decrease in the specific ^^C activity measured in tree-rings, which was assumed 

 to be caused by burning of fossil fuel, did not begin at about 1850 but roughly 

 a century earlier (deVries, 1958, 1959; Broecker, Olson and Bird, 1959; Suess, 

 1960). DeVries suspected that this decrease in i^Q activity had something to 

 do with the general glacial advances during the 18th century. In his opinion, 

 an inverse correlation exists between I'^C in the atmosphere and glacial advances 

 and retreats. With such a correlation, the effects from industrial coal combus- 

 tion should in reality be larger than the observed decrease of 2% in i-*Q because 

 the glaciers began to retreat again during the middle of the last century, so 

 that a reversal of the general trend should have occurred. This reversal was 

 compensated for by artificial fuel combustion. De Vries estimates that the 

 total effect of fossil fuel combustion should be about 2.7%. However, the 

 present ^''C values in the atmosphere can be obtained by a mere extrapolation 

 of the trend observed during the 18th and 19th centuries. Although it still 

 seems probable that the observed decrease of 2% over the past 100 years is at 

 least in part due to combustion of fossil fuel, it is now impossible to draw any 

 quantitative conclusions from the magnitude of the effect. 



There is good reason to believe that, during the coming years, a third method 

 will become available to determine the residence time of CO 2 in the atmosphere 



