G40 



DATA OF GEOCHEMISTRY 



leached for a long time. In table 79, these perennial 

 rivers are represented by the water of the Eose River. 

 The intermittent rivers pose something of a problem. 

 The Murray River has a very low discharge rate, as 

 befits a river flowing through a semiarid land, but it is 

 surprisingly dilute. This river shows considerable 

 fluctuation from year to year — during some years it 

 ceases to flow at all — and one cannot help wondering 

 if the discharge figures represent dry years and the 

 chemical analyses wet years. If this is the case the 

 chemical denudation for Australia will be underesti- 

 mated to some extent, but even these figures of doubtful 

 reliability suffice to show that the smallest continent 

 contributes only a very small amount to the world 

 total for chemical denudation. It appears to yield 

 about 6 metric tons per square mile of total area. 



Table 79. — Discharge and chemical denudation of Australia 



SOUTH AMERICA 



There are no new data of consequence on which to 

 base an estimate of the chemical denudation of South 

 America. It seems more reasonable to use the Parana 

 as representative of the less well watered parts of the 

 continent than to use the very dilute Uruguay, as 

 Clarke did; this leads, on the basis shown in table 80, 

 to a figure of 73 metric tons of dissolved substance 

 being removed per square mile per year. 



Table 80. — Discharge and chemical denudation of South America 



WORLD SUMMARY 



By multiplying the average chemical denudation of 

 the continents by their size, one arrives at a figure of 

 3,905,000,000 metric tons for the total amount of 

 mineral material carried in solution each year by the 



rivers flowing into the sea. This is almost 1,200,000,- 

 000 tons greater than the estimate of Clarke in 1924, but 

 it is substantially below the earlier estimate of Sir John 

 Murray (quoted in Clarke, 1924b, p. 63) which was 

 estimated on rather meager data. It is evident from 

 the difference between various independent estimates 

 that an accurate assessment of the chemical substance 

 carried by the rivers has not been reached. Although 

 the current estimate rests on more accurate basic infor- 

 mation than the earlier ones, there is some possibility 

 that it may be appreciably too high. Our knowledge 

 of the chemistry of rivers for much of the world, par- 

 ticularly the humid tropics, rests upon a very small 

 number of samples, rather than upon long-term studies 

 during all seasons. Most of the annual discharge of 

 a river having a variable discharge rate occurs during 

 short periods of flood when the dissolved salt content 

 is at its lowest, so infrequent haphazard sampling is 

 likely to result in an overestimate of the mean salinity 

 of the water which the river carries. This error is not 

 likely to affect the figures for Europe and North Amer- 

 ica, but it might influence the total for South America, 

 Africa, and Asia. 



MEAN CHEMICAL COMPOSITION OF WORLD RIVER WATER 



Weighted means have been calculated for the chemi- 

 cal composition of the river waters of various continents 

 as well as the world on the same basis as the calculations 

 for chemical denudation. In a few cases it was neces- 

 sary to supplement incomplete analyses with estimates 

 for silica based on the silica content of similar waters. 

 The Rose River analysis, used as a basis for computing 

 the composition of the permanent rivers of Australia, 

 has had 3.5 ppm of silica added to it, a quantity which 

 is found in the similar dilute waters of Tasmania. 

 The mean figures of Alekin and Brazhnikova (1957) 

 for the Arctic and Pacific drainage systems of the 

 U.S.S.R. lack silica. These have been supplemented 

 by assuming that their silica contents were similar to 

 those of the Mackenzie and St. Lawrence Rivers, 

 respectively. Some important analyses lump sodium 

 and potassium. In calculating world mean composition 

 the combined Na + K of these analyses has been parti- 

 tioned according to the Na/K ratio of the rest of the 

 waters of the world. 



The results of this computation are summarized in 

 table 81. The principal differences between the conti- 

 nents are in the amounts of calcium and bicarbonate 

 ions. The great variation in the nitrate and iron 

 contents is insignificant for reasons that were dealt 

 with at length above. The world mean for nitrate, 

 however, may be of the correct order of magnitude. 



The mean for the river water of the world, 120 ppm, 

 is somewhat lower than Clarke's 1924b data suggested, 



