previous sub-section
Chemistry of Sea Water
next sub-section

Inorganic Agencies Affecting the Composition of Sea Water

The factors that may modify the absolute and the relative concen, trations of the substances in sea water are exchange with the atmosphere-inflow of river water, freezing and melting of sea ice, and biological activity. Biological processes and their effects upon the distribution of various elements are considered in chapter VII.

Exchange with the Atmosphere. The distribution of salinity in the oceans, and hence the concentrations of the major elements, is maintained by agencies that are described elsewhere, but one point must be considered at this time. Over the sea and along its shores, spray is continually being swept up into the air, and as the spray represents actual particles of sea water with its dissolved salts, this process affords a mechanism for the removal of salts from the sea. A large portion of the spray undoubtedly falls back into the water or is carried down by rain (Köhler, 1921). However, winds blowing toward the land will carry with them their content of salt, which may be deposited on the land directly or carried down by the rain. Observations by Jacobs (1937) on the chloride content of the air near the sea showed concentrations ranging between 0.07 and 0.5 mg of chloride per cubic meter of air. The amount increased with the wind velocity and was greatest with onshore winds.


A considerable proportion of the dissolved material carried to the sea by rivers is “cyclic salt”—that is, salt that has been carried inland by the atmosphere and then deposited or carried down by rain and snow (Clarke, 1924, Knopf, 1931).

Besides the exchange of salts that takes place between the atmosphere and the ocean as described above, there is an exchange of dissolved gases and nitrogen compounds which may modify the quantity of these substances present in sea water that is in contact with the atmosphere. The factors that affect the exchange of gases are described elsewhere. The exchange of water between the atmosphere and oceans was taken up in chapter IV.

Rain water contains relatively high concentrations of nitrogen compounds, which are believed to be formed from the constituents of the atmosphere by electrical discharges; hence the atmosphere supplies to the ocean, either directly through rainfall or indirectly through run-off from the land, a certain amount of fixed nitrogen. Whether this increment in the amount of fixed nitrogen is balanced by deposition of organic nitrogen in sediments or by the liberation of gaseous nitrogen through the decomposition of nitrogen compounds in the sea is not yet known.

Effects of Rivers on the Composition of Sea Water. The run-off from land is but a part of the cycle of leaching. The precipitation on the land contains only the cyclic salts, dissolved atmospheric gases, and nitrogen compounds. This water acts upon the rocks, contributing to the mechanical break-down of the solid material and extracting from them their more soluble constituents. The nature and quantity of the various elements dissolved depends upon the character of the rocks or soils with which the water comes in contact on its way to the sea. Because the leaching is carried out by water of low salt concentration yet relatively high in carbon dioxide compounds, it is capable of dissolving materials that would not pass into solution if they were in contact with sea water. In addition to dissolved material, rivers carry to the sea colloidal and particulate material in tremendous quantities. A considerable part of this debris is dropped to the sea bottom near shore, and much of the finer material coagulates and settles when mixed with sea water. Sea water reacts in various ways with the colloidal and finely dispersed material, and some of these reactions may affect the relative composition of the dissolved constituents. Interaction between the dissolved constituents of sea water and the sedimentary debris may be subdivided as follows: (1) solution of the constituents of the sediment, (2) adsorption on the sediment, (3) ionic exchange, and (4) reactions to form new substances. Little is known concerning the importance of these processes.

From the magnitude of the land area drained by rivers emptying into the sea and from the composition of the salts dissolved in river waters,

Clarke (1924) has estimated that rivers contribute to the sea each year 2.73 × l09 metric tons of dissolved solids. By weighting the composition of the water of different river systems in proportion to the total supply of dissolved material, he obtained the average composition of river water shown in table 43. Comparison with the composition of the dissolved solids in sea water shows, when referred to the halides, that there is an excess of all of the substances reported. Hence, the effect of river waters will be to modify the relative composition of the dissolved solids in sea water. Possibly a relatively large proportion of the carbon dioxide and nitrate should be regarded as cyclic. In addition, it should be recalled that the cyclic sea salts, carried into the atmosphere as spray and then deposited or carried down by precipitation onto the land, will enter the compilation. If we assume that all of the chloride in the river water is cyclic, the amounts of the other elements must be modified in the proportions in which they occur in sea water. It is hardly possible that this assumption is entirely correct, but it may yield values that are closer to the truth. The average composition of river water adjusted in this way for cyclic salts is included in the table.

Ion River water (weighted average) Sea water River water (less “cyclic” salts)
CO3 35.15 0.41 (HCO3) 35.13
SO4 12.14 7.68 11.35
Cl 5.68 55.04 0.00
NO3 0.90 ............. 0.90
Ca++ 20.39 1.15 20.27
Mg++ 3.41 3.69 3.03
Na+ 5.79 30.62 2.63
K+ 2.12 1.10 2.02
(Fe,Al)2O3 2.75 ............. 2.75
SiO2 11.67 ............. 11.67
Sr++, H3BO3, Br ...... 0.31 .....
100.00 100.00 89.75

It is not known whether the addition of dissolved solids brings about progressive changes in the relative composition of the sea salts or whether there is any progressive alteration of the total salt content or salinity. In any event, both processes must be exceedingly slow. The total amount of dissolved solids contributed by the rivers each year is only an infinitesimal fraction, 5.4 × 10−8, of the total dissolved solids in the ocean.

Even this small fraction does not represent a net gain, as there are many processes which remove material from solution. Certain elements accumulate in the marine sediments and are precipitated either through physical reactions or by biological activity. This is particularly true of the calcium and magnesium carbonate which form calcareous deposits and of the silicon which is found in organic siliceous deposits (radiolarian and diatom oozes). The sodium and potassium may be removed from solution by adsorption on, or ionic exchange with, the clay particles brought to the sea by rivers. However, in the latter case some other element will be released to the water in equivalent amounts. Other aspects of this problem are considered in connection with marine sediments and geochemistry, but it must be admitted that most questions related to changes in composition of sea water are still unsolved. In dealing with these it must also be considered that the amount of water may be changing. Essentially, all the river water is cyclic, but it is known that juvenile water of subterranean origin is continually being added to the surface water. In addition, the amount of water represented by the ice caps may be variable. Goldschmidt (1933) has estimated that for each square centimeter of the earth's surface there are 273 1 of water subdivided as follows:

Sea water … … … … … … … … … … … … … … … 268.45 1
Fresh water.… … … … … … … … … … … … … … 0.1
Continental ice.… … … … … … … … … … … … … 4.5
Water vapor.… … … … … … … … … … … … … … 0.003

The average composition of the river water is of interest in considering the effect on the oceans as a whole and over long periods, but particular investigations must be concerned with the effects brought about by individual rivers whose dissolved solids may differ markedly in composition and concentration from the average. Data can be obtained from Clarke (1924) or similar sources; as an illustration, values for several of the large American rivers are given in table 44.

From these examples it can be seen that the composition of individual rivers may differ considerably from the average. Thus, the Columbia River is low in chloride and the Colorado River is high, and the latter river is high in sodium and sulphate and below average in calcium and carbonate. The effect that dilution will have upon the chlorosity factors will therefore depend upon the character of the river water.

Thus far we have considered only the more abundant elements in the river water. Undoubtedly, all elements are carried to the sea either in solution or as finely divided particulate material. The high production of plant and animal life which frequently occurs near the mouths of rivers has sometimes been ascribed to the plant nutrients introduced by the rivers. Riley (1937) has found that the Mississippi

River water contains higher concentrations of phosphate and nitrogen compounds than the surface sea water in the Gulf of Mexico, and that this has a direct effect upon the amount of life in the waters off the mouth of the river. From studies in the English Channel, Atkins (1923) concluded, on the other hand, that the river water had little effect upon the immediate production, as most of the nutrients were used up by organisms living in the rivers.

Effects of Formation and Melting of Sea Ice on the Composition of Sea Water. A laboratory study of the freezing of sea water was made by Ringer, whose results have been reported by Krümmel (1907) and Johnstone (1928). In these experiments sea water was cooled in the laboratory, and at various temperatures below the initial freezing point the ice and precipitated salts were separated from the mother liquor. Sea water of salinity 35.0‰ begins to freeze at − 1.91°C (p. 66). At first, pure ice crystals separate, and, as the concentration of the brine is increased, the temperature must be further reduced to bring about the formation of additional ice. As the temperature is lowered and the concentration of the brine is increased, the solubility of certain of the dissolved salts is exceeded. At −8.2° the Na2SO4 begins to separate and continues to do so with further cooling. At −23° the NaCl begins to crystallize. In addition, a certain amount of CaCO3 precipitates. Ringer's analyses of the “ice” (including the ice crystals and the precipitated salts) and the brine when the temperature had been reduced to −30° are as follows:

Ion Average Mississippi River Columbia River Colorado River
CO3 35.15 34.98 36.15 13.02
SO4 12.14 15.37 13.52 28.61
Cl 5.68 6.21 2.82 19.92
NO3 0.90 1.60 0.49 ..........
Ca++ 20.39 20.50 17.87 10.35
Mg++ 3.41 5.38 4.38 3.14
Na+ 5.79 }8.33* 8.12 19.75
K+ 2.12 1.95 2.17
(Fe,Al)2O3 2.75 0.58 0.08 ..........
SiO2 11.67 7.05 14.62 3.04
Annual contribution of dissolved solids (metric tons) 100,000,000 19,000,000 13,416,000
Salt content (g/l) 0.166 0.0924 0.702
* Sum


One kilogram of sea water, initial salinity 35.05 ‰, yielded:

Ice crystals ……………………… 931.9 g
NeCl crystals … … … … … … … 20.23
Na2SO4 crystals… …………… 3.95
CaCO3 crystals.… … … … …… Trace
Brine… … … … ………………… 43.95

The brine contained 23.31 g of H2O and

Na+……………… 1.42 g Cl……………… 7.03 g
Mg++……………… 1.31 “ Br……………… 0.08 “
K+……………… 0.38 “ SO4……………… 0.03 “
Ca++……………… 0.39 “

From these data it is readily seen that, when the temperature of the ice and brine is lowered to −30°, there are marked differences in the relative composition of the salts in the “ice” and in the brine. If the cooling is continued to −50°, there is further separation of ice and salt crystals, but some very concentrated brine is still present.

From these experiments it would appear that the formation of sea ice might have a pronounced effect upon the relative composition of the salts in the water. The brine from which it formed would be modified in one direction, and, if melting took place in water other than that from which the ice formed, the effect would be in the opposite direction. However, the formation of sea ice in nature is not reproduced by these laboratory experiments. Let us suppose that, in a region where the depth to the bottom is moderate or great, sea water of normal composition is subjected to cooling at the surface. The resulting increase in density gives rise to convection movements that continue until the water at the surface reaches the freezing point, and then ice will begin to form. The brine, being of greater concentration but having virtually the same temperature, will sink and new water will be brought toward the surface and into contact with the ice. At first, isolated, elongated ice crystals are produced, but as the freezing continues these form a matrix in which a certain amount of the brine is mechanically included. The ice crystals themselves are at this stage probably “pure ice.” If the freezing proceeds rapidly, the brine will accumulate in separate cells within the body of the ice and, as the temperature of the ice near the surface is reduced, more ice crystals are formed, the cells decrease in size, and the concentration of the brine in the cells increases (fig. 16, p. 72). This may continue so far that solid salts crystallize in the cells. From this it can be seen that there is not necessarily any relative change in the composition of the dissolved salts in the sea water and in the sea ice (ice crystals plus the enclosed brine).

The salinity of the ice, using the same definition as applied to sea water, has been shown to depend upon the rate of freezing. Malmgren (1927), from the observations made by the Maud Expedition, gives the

following average values showing the relation between the salinity of new ice and the air temperature, where the latter is used aa a measure of the rate of freezing.

Air temperature (°C) Salinity of ice (‰)
− 16 5.64
− 28 8.01
− 30 8.77
−40 10.16

The salinities are based upon chlorinity determinations made on samples of the melted ice. The salinity of the surface water was about 30 ‰. The effect of the rate of freezing is also shown by the analyses of samples obtained in April from an ice floe that had started to form the preceding November:

Distance from surface of ice (cm) 0 6 13 26 45 82 95
Salinity of ice (‰) 6.74 5.28 5.31 3.84 4.37 3.48 3.17

The lower salt content of the deeper ice is related to the slower rate of formation. When ice is formed with extreme rapidity, its salinity will approach that of the water from which it is produced.

According to Ringer's experiments the cooling of ice containing cells of brine leads to formation of additional ice crystals and, if the temperature is reduced sufficiently, to separation of salt crystals within the ice. With very rapid freezing, brine and salt crystals may accumulate on the surface of the ice, making the surface “wet” at temperatures of −30° to −40°C and greatly increasing the friction against sled runners or skis.

In such rapidly frozen ice the cells containing the brine are large or numerous. If the temperature rises, the ice surrounding thg cells melts and the separated salt crystals are dissolved, but before complete solution has taken place the brine cells may join, permitting the brine to trickle through the ice. Under these conditions some of the solid salts may be left in the ice, and the composition of the water obtained by melting will differ from that of normal sea water. If, on the other hand, the temperature of the ice is raised to 0°C, all salts dissolve, the cells grow so large that the ice becomes porous, all brine trickles down from the portions of the ice above the sea surface, and the exposed old ice becomes fresh and can be used as a source of potable water.

Analyses by Wiese (1930) indicate that the processes that have been described may be effective in changing the relative composition of the salts, He found that the sulphate and alkalinity factors were greater in the ice than in the water and were greater in old ice than in newly frozen ice. This indicates that small amounts of sulphate, probably present as Na22SO4, have remained in the ice during the process of ageing and, probably, that the relative amounts of CaCO3 have changed.


Results of the Maud Expedition, reported by Malmgren (1927) and Sverdrup (1929), are not in agreement with the findings of Wiese. Chlorinities of water obtained by melting ice were systematically higher when determined by titration than when computed by means of Knudsen's Hydrographical Tables from observations of density. This discrepancy was interpreted to mean that the sea ice contains an excess of chlorides, but it may arise from the application of Knudsen's Tables to water that has been diluted by essentially distilled water, as explained on p. 59. The fact that the SO4/Cl ratio was nearly the same in the ice and the sea water (Malmgren, 1927, p. 9) supports the latter explanation and indicates that no changes in relative concentration had resulted from processes of freezing and melting. The problem, however, cannot be considered solved, and it offers opportunities for further laboratory investigations and observations in the field.

previous sub-section
Chemistry of Sea Water
next sub-section