96 BROECKEIt [CHAP. 4 



similar ratio for average surface ocean CO2 and atmospheric CO2 (Cs/Ca), the 

 average rate of exchange can be estimated. From existing data Co/Ca has 

 been estimated as 0.85 ±0.05 and [1 - (Cs/Ca)] as °- 050 ±0-015. The error in 

 the latter stems from two sources: (1) the experimental error in the measure- 

 ments, and (2) the uncertainty in the correction for the change brought about 

 by the addition of industrial CO 2 to the system. The above values yield an 

 invasion rate of 20 + 7 moles m -2 yr -1 . As a 1-m 2 column of air contains about 

 100 moles of CO2 and since about 30% of the earth is covered by land, this 

 corresponds to a mean atmospheric residence time for CO 2 with respect to 

 transfer into the ocean of seven years. 



Consideration must be given, however, to the possibility that the exchange 

 is not geographically uniform. Kanwisher has demonstrated the importance of 

 wind velocity in the rate of exchange. He points out that wind velocities are, 

 on the average, considerably higher in the Antarctic region than over the rest 

 of the ocean. This leads to the possibility that the exchange rate could be, on 

 the average, as much as a factor of five higher in the Antarctic region. If this 

 were the case, an exchange rate of 50 moles m -2 vr -1 would be required for the 

 Antarctic and only 10 moles m~ 2 yr -1 for the remainder of the ocean (as- 

 suming the Antarctic comprises 12% of the ocean surface). The atmospheric 

 residence time for CO2 would rise to about 10 years. This rise occurs because 

 the Antarctic surface water has a 14 C/ 12 C ratio averaging about 8% lower 

 than that for the remainder of the surface ocean. As laboratory measurement 

 of the CO 2 exchange rate in unstirred sea- water (unpublished work by the 

 author) yields a value of 6 moles m -2 yr -1 , the rate at any point in the ocean 

 must be at least this great. In order to maintain the mixed layer exchange rate 

 above this minimum, the contrast between the Antarctic and mixed layer of 

 the ocean cannot exceed a factor of about eight. From the discussion which 

 follows it will become clear that the magnitude of this geographic effect (es- 

 pecially as it affects the ratio of Antarctic to non-Antarctic exchange rate) is 

 critical to the interpretation of the 14 C data. 



5. Oceanic Mixing Models 



The models considered to date are all of the box type. The ocean-atmosphere 

 system is broken up into a series of independent, well-mixed reservoirs. Carbon 

 dioxide with a 14 C/ 12 C ratio typical of the entire reservoir is exchanged by 

 adjacent reservoirs. If the average 14 C/ 12 C ratio for each of the various reservoirs 

 is fixed, the exchange rates required to maintain this distribution of 14 C can 

 be computed. 



The models considered to date can be placed in four groups : (1) the two-layer 

 model of Arnold and Anderson (1957) and of Craig (1957), (2) the outcrop 

 model of Craig (1958), (3) the three-box model and (4) the world-ocean model of 

 Broecker et al. (1960). These models (shown diagrammatically in Figs. 2 

 through 8) will be referred to as models I, II, III and IV for simplicity. In 

 each model the main body of the ocean is divided into a well-mixed surface 



