CP-2156 Life In The Universe


Origin and Evolution of Continents and Oceans




[101] Continental growth began soon after the formation of Earth and most of its hydrosphere. Perhaps it was only when decay processes became less effective that the continents could maintain integrity and a selectively depleted mantle could begin to develop.


The geometry and dynamics of the ocean basins and continents over the past 200 million years are comprehensible within the framework of the plate tectonic theory. The fundamental features of the theory are that Earth's outer layer, called the lithosphere, is divisible into a number of plates that are moving relative to each other and are being created and destroyed at the plate boundaries. The plates diverge by the creation of new ocean floor, and the resulting convergences with other plates cause destruction of the lithosphere along the edges. The divergence invariably leads to the formation of an ocean basin or an increase in the area of an existing basin. Most convergent plate boundaries have at least one oceanic side-this means that ocean basins are destroyed even where the other boundaries are continental. There are, however, a number of convergences of plates that are both continental, as in the case of the Indian plate and the Asiatic plate. In such cases, continental crust must be destroyed, but by mountain-building and erosion. Thus, in principle, if no other processes were acting, Earth's surface would after a time consist of ocean basins covered with a layer of sediment eroded from the elevated mountains at the continental convergences.

The geological record tells us that this process did not occur efficiently for most of the 3.8 billion years recorded in the rocks. Obviously other processes have been acting to protect the continents, at least since 3.8 billion years ago.

[102] The two processes acting on behalf of continental perpetuity are: (1) the transformation of sediment piles deposited adjacent to the continent or brought to the continental margin by the convergence of ocean- and continent-bounded plates into metamorphic rocks welded to the continental block and (2) the addition of new continental material from the mantle. The geological record tells us that both of these processes have been active throughout the history of Earth.

To the extent that new continental material is formed from the mantle, there has been net continental accretion. The question then becomes: What was the pattern of increase of continental crust with time? A companion question is: In what way was the growth of the volume of oceans related to the growth history of the continental crust?




The best model for the formation of Earth, for a long time, was based on the accretion of material of chondritic, and more specifically carbonaceous chondritic, composition. The subsequent redistribution of the elements was supposed to give rise to the present configuration of hydrosphere atmosphere, crust, mantle, and core.

Two important papers in the 1960s comparing certain properties of terrestrial materials with the supposed chondritic precursor showed that this assumption was too simple. The late Paul Gast, in a classic paper on "the limitations on the composition of the mantle" (Gast, 1960), clearly showed that the 87Sr/86Sr ratio of mantle-derived materials required the mantle source region to have a considerably lower Rb/Sr ratio than any chondrite analyzed. Ringwood (1966) showed that an Earth partitioning from a carbonaceous chondrite bulk composition could not yield the highly oxidized state of the upper mantle and the high concentrations of siderophile elements, such as nickel, found there if the mantle had been formed in equilibium with an iron-nickel core produced by reduction and segregated by gravity.

These ideas clearly required new attitudes with regard to the composition and formation history of Earth's four major inorganic spheres. Since then results on rare gases and their isotopes, the rare earth elements and their isotopes, and the isotopes of oxygen have added additional constraints. In particular, the use of radioactive 147Sm and its daughter 143Nd in a manner similar to the use of the 87Rb-86Sr couple has added much to our knowledge of the development of Earth's crust. The decay constants for 87Rb and 147Sm are 1.42 x 10-11 yr-1 and 6.54 x 10-12 yr-1, respectively.

Several groups have been involved in these studies: the La Jolla group beginning with Lugmair et al., 1975), the Paris group (beginning with [103] Richard et al., 1976), the Caltech group (beginning with De Paolo and Wasserburg, 1976), and the Lamont-Doherty group (beginning with O'Nions et al.,1977). A review of the results to date appears in O'Nions et al. (1979).

The relationship between the 87Sr/86Sr ratio of mantle-derived materials and the 143Nd/144Nd ratio indicates oceanic basalts, specifically the midocean ridge basalts (MORB), are derived from a mantle source that has been depleted for a long time in Rb relative to Sr and in Nd relative to Sm. Since both Rb and Nd are known to be highly concentrated in continental magmatic rocks relative to Sr and Sm, respectively, the assumption is that the depletion is due to the formation of continental crust early in Earth's history.

If one assumes that the bulk Earth has the same relative abundances of the rare earth elements as chondrites (and most other meteorites) and that the initial isotopic composition at least of Sm and Nd was the same throughout the Solar System, then it is possible to extrapolate the correlation line of 87Sr/86Sr vs 143Nd/144Nd to intercept the 143Nd/144Nd expected in the bulk Earth today. This should be the same value found in chondrites. This intercept yields a value of the 87Sr/86Sr ratio for the bulk Earth which can be translated into a Rb/Sr ratio for the bulk Earth, assuming an initial 87Sr/86Sr ratio and the age of Earth based on meteorite studies. The 87Rb/86Sr ratio of the bulk Earth by this model is about 0.09, compared to chondritic values ranging from 0.73 to 0.78, reaffirming Gast's initial observation.

O'Nions et al. (1979) and McCulloch and Wasserburg (1978) show that continental crust, back through the Archean, has been forming from material of bulk Earth composition as far as the rare earth complexion is concerned and not from depleted mantle as for the ocean basalts.

We can then divide Earth's mantle into two zones: a "depleted" mantle produced from bulk Earth composition by the melting and extraction of continental crust and a mantle retaining its initial bulk Earth composition.

Based on these results, one can argue that continental crust always formed from bulk Earth material and, based on the MORB results, one can argue further that the dominant source of material supplying basaltic rock to the diverging plate boundaries is the "depleted" mantle.

If we extend our definition of "depleted" beyond relative Nd and Rb depletion to the class of elements to which they belong, sometimes called the "large ion lithophiles," then the depleted mantle should be lower than the bulk Earth mantle in U, Th, Pb, K, and other elements that can be grouped with these. Indeed, these elements are low in MORB and are therefore inferred to be low in the MORB mantle source, which is depleted in Rb and Nd.

The oceanic islands have values for 143Nd/144Nd and 87Sr/86Sr intermediate between the MORB value and the bulk Earth value and are also [104] higher in their U and K concentrations. They can therefore be inferred to be a mixture of the two different types of mantle reservoirs.

The close association of Ca, Sr, the rare earth elements, Th, and U during the condensation of solid phases from a solar composition nebula (see Grossman and Larimer, 1974, for a summary) has led various people to estimate the bulk planetary concentration of these elements based on heat flow (O'Nions et al., 1979; Turekian and Clark, 1975). Table 1 shows the concentrations of Nd and Sm for the bulk Earth based on such a calculation together with the continental crust values taken from O'Nions et al. (1979). One can calculate the amount of bulk Earth that had to have been processed to yield the continental crust with its characteristic Sm/Nd ratio if one knows the Sm/Nd ratio of the depleted mantle. If we assume that the rare earth elements are delivered to the midocean ridge basalts from the depleted mantle source with no further fractionation of the rare earths, then the ratio observed in these rocks will provide us with a suitable estimate. If any fractionation occurs during the formation of the MORB magma, then the MORB source (depleted mantle) would have to have a smaller Sm/Nd ratio.

A plot of the Sm/Nd ratio of depleted mantle against the mass of undepleted mantle processed to form depleted mantle relative to the mass of the continental crust is shown in figure 1, based on the appropriate data in table 1. The MORB Sm/Nd ratio is taken to be 0.38 (Sun et al., 1979). This corresponds to a mass of processed mantle relative to continental crust of 30. Since the continental crust is about 2 x 1025 gm, about 60 x 1025 gm of depleted mantle must exist. The model predicts a depleted mantle of 60 x 1025 gm resulting from the formation of the continental crust. Since the mass of the total mantle is about 400 x 1025 gm, about 15% of the mantle is depleted. This calculation is based on one set of data. The amounts of the different reservoirs could change with a different choice of data, but the fact remains that a significant but not overwhelming amount of depleted mantle is consistent with observations.




Mass, 1025 gm

Sm, ppm

Nd, ppm

Sm/Nd, weight ratio


Undepleted mantle

smaller or equal to343




Continental crust










Depleted mantle

greater or equal to60

greater or equal to0.33

greater or equal to0.86

smaller or equal to0.38

Bulk Earth






Figure 1. Plot of Sm/Nd of depleted mantle vs the ratio of mass of the processed deplete mantle (MB) to the mass of continental crust (Mc = 2 x 1025 gm).

Figure 1. Plot of Sm/Nd of depleted mantle vs the ratio of mass of the processed deplete mantle (MB) to the mass of continental crust (Mc = 2 x 1025 gm).


Because MORB comes from shallow parts of the mantle and continental crust from deeper parts, the assumption seems inevitable that Earth is zoned. This is also compatible with models of terrestrial heat production (Clark and Turekian, 1979).

There is a problem with this zone scheme. If the large ion lithophile and generally the low-melting fraction are extracted from bulk mantle materials, then one has the feeling that the residual mantle should be denser than the original mantle and thus not be in the shallower parts of the mantle. But our observational data seem to require just such a conclusion. One way out of this dilemma is to add "lighteners" to depleted mantle material. The most obvious lightener is water. It could be added directly to the processed mantle material from an external source (the ambient ocean?). It could act as lightener either as a molecular additive or as an oxidant with the loss released hydrogen by diffusion. The latter mechanism is preferable. Whatever the mechanism, it implies the existence of an extensive aqueous phase capable of entering the depleted mantle. Earth in that sense may not have been degassing water over time but rather using it up as an oxidant!

The driving force for all this activity must be the heat generated in Earth by radioactive decay of U, Th, and K (and possibly other shorter-lived [106] radionuclides in its early history). With the decrease, by decay, of the heat-generating radionuclides, the process must slow and the rate of production of new continental crust must decrease.

As undepleted mantle is transformed into depleted mantle by the formation of continental crust, the depth of the boundary between the two increases. This could inhibit the formation of additional continental crust by decreasing the efficiency of mass transport from this depth to effect heat loss. The loss of Earth's heat at the present time may be more efficiently mediated by convection and volcanic activity in the upper depleted zone than by the eruption of new continental material from the deeper undepleted zone. The eruptions called "hot spots" or "plumes" as inferred from the location of nonridge oceanic islands may result from release of substances from the deep undepleted zone. In this they may resemble, in a small way, the mechanism for the production of continents at an earlier time.




Are oceans increasing in volume, or did they appear essentially in their present volume at Earth's surface early in its history? This question has troubled geologists for a long time. The case for continuous degassing was detailed by Rubey (1951). The case for instantaneous degassing has been based mainly on the antiquity of rocks with sedimentary features on cratons (Armstrong, 1968) and on calculations based on 40Ar in the atmosphere (Fanale,1971).

The case for continuous degassing of some gases from the interior, notably the rare gases, has received support from a recent 3He measurement in rocks, hot springs, and ocean water (for reviews see Craig et al., 1978 Tolstikhin,1978).

The fact that oceanic basalts actually add more gas than they lose during hydrothermal alteration (Dymond and Hogan, 1973) eliminates degassing as a new source of gases for all but 3He and Ne, for which the calculation had been made anyway.

There are therefore few constraints on devising a model for the way in which the ocean arrived at Earth's surface. My prejudice, shared with others (e.g., Anders and Owen, 1977), is that the hydrosphere arrived as part of a low-temperature veneer added to the accreting planet and that it was released mainly at the time of infall as the result of gravitational heating either as it passed through the growing atmosphere or during impact.

If this is the case, then the growing (undepleted) mantle could have started out more anhydrous than the bulk material arriving at Earth. [107] Certainly any water that became a part of the mantle would later be involved in the Segregation of continental crustal material from the mantle.

There is no reason to assume a priori that the volatile-bearing material infalling to Earth had to be coupled to the alkali-metal-bearing phases required for continental crust formation. The data for chondrites (from Minster and Allegre, 1979) show that as one proceeds from the least-equilibrated H-group chondrites (H-3) to the most equilibrated (H-8), the Rb/Sr ratio and the K concentration increase. This is in exactly the opposite sense of the distribution of the volatile components such as carbon and the rare gases (Mart), 1967). Clearly, K and Rb do not follow the "low temperature", volatiles in chondrites, and there is no reason to expect them to have done so during the accretion of Earth.

I contend therefore that water and some other volatiles were released in large part to Earth's surface early in its history, perhaps at the time of accretion, and that there need be no correlation with continental crust formation and the volumetric growth of the oceans. Indeed, as implied in the previous section, the availability of a supply of water in Earth's outer spheres may have provided the opportunity to oxidize Earth's iron distributed in the mantle as the result of accretion. This would elevate the oxidation state of the rocks of the upper, depleted mantle while guaranteeing the retention there of the otherwise siderophile elements such as Ni. I confess that the exact mechanism of planetary oxidation is not completely clear to me, but certainly it involves the incorporation and dissociation of H2O with subsequent loss of H2 by diffusion from the upper mantle and ultimately from Earth altogether.




If the continent-making process is coupled to heat production in Earth, then one would expect to see continental material dating back to the beginning of Earth or at least rocks with a radiogenic isotopic imprint that remembered the early days. As we have seen, neither of these have been found, and so another explanation must be sought. I shall conclude by offering one that may be difficult to prove in light of the paucity of data for the early days of Earth's history.

At the present time, the mantle convective cycle involves release of basalts from the depleted upper zone of the mantle. In the past, when continental crustal material was forming, it was forming from bulk Earth composition (or undepleted) mantle. Just as at the present time at the convergence areas the oceanic basalts are injected back into the mantle, so in the [108] early days when the continental crust was forming it may have been destroyed by reinjection into the mantle. It would then essentially return the depleting mantle to its primitive undepleted state. Only when the process of resorption became less efficient did the continents attain their integrity and the process of formation of depleted mantle begin in earnest.




- Anders, E.; and Owen, T.: Mars and Earth: Origin and Abundance of Volatiles. Science, vol. 198, Nov. 4, 1977, pp. 453-465.

- Armstrong, R. L.: A Model for the Evolution of Strontium and Lead Isotopes in a Dynamic Earth. Rev. Geophys., vol. 6, May 1968, pp. 175

- Clark, S. P., Jr.; and Turekian, K. K.: Thermal Constraints on the Distribution of Long-Lived Radioactive Elements in the Earth. Phil. Trans. R. Soc. Lond., Ser. A, vol. 291,1979, pp. 269-275.

- Craig, H.; Lupton, J. E.; and Horibe, Y.: A Mantle Helium Component; Circum-Pacific Volcanic Gases: Hakone, the Marianas, and Mt. Lasser In: Terrestrial Rare Gases, E. C. Alexander, Jr., and M. Ozima, eds., Center for Academic Publications, Tokyo, Japan, 1978, pp. 3-16.

- De Paolo, D. J.; and Wasserburg, G. J.: Nd-Isotope Variations and Petrogenetic Models. Geophys. Res. Lett., vol. 3, May 1976, pp. 249-252.

- Dymond, J.; and Hogan, L.: Noble Gas Abundance Patterns in Deep-Sea Basalts- Primordial Gases from the Mantle. Earth Planet. Sci. Lett. vol. 20, Sept.1973, pp. 131 -139.

- Fanale, F. P.: A Case for Catastrophic Early Degassing of the Earth. Chem Geol., vol. 8, Sept. 8,1971, pp. 79-105.

- Gast, P. W.: Limitations on the Composition of the Upper Mantle. J. Geophys. Res., vol. 65, Apr. 1960, pp. 1287-1297.

- Grossman, L.; and Larimer, J. W.: Early Chemical History of the Solar System. Rev. Geophys. Space Phys., vol. 12, Feb. 1974, pp. 71-101.

[109] - Lugmair, G. W., Scheinin, N. B.; and Marti, K.: Search for Extinct 146Sm,1: The Isotopic Abundance of 142Nd in the Juvinas Meteorite. Earth Planet Sci. Lett., vol. 27, Aug.1975, pp. 79-84.

- Marti, K.: Trapped Xenon and the Classification of Chrondites. Earth Planet. Sci. Lett., vol. 2, 1967, pp. 193-196.

- McCulloch, M. T.; and Wasserburg, G. J.: Sm-Nd and Rb-Sr Chronology of Continental Crust Formation. Science, vol. 200, June 1978, pp. 10031011

- Minster, J-F.; and Allegre, C..: 87Rb-87SR Chronology of H Chrondites Constraint and Speculations on the Early Evolution of Their Paren Body. Earth Planet. Sci. Lett., vol. 42,1979, pp. 333-347.

- O'Nions, R. K.; Carter, S. R.; Evensen, N. M.; and Hamilton, P. J.: Geochemical and Cosmochemical Applications of Nd Isotope Analysis. Ann. Rev. Earth Planet. Sci., vol. 7,1979, pp. 11 -38.

- O'Nions, R. K.; Hamilton, P. J.; and Evensen, N. M.: Variations in 143Nd/144Nd and 87Sr/86Sr Ratios in Oceanic Basalts. Earth Planet. Sci. Lett., vol. 34, Feb.1977, pp.13-22.

- Richard, P.; Shimizu, N.; and Allegre, C. J.: 143Nd/146Nd, a Natural Treasure: An Application to Oceanic Basalts. Earth Planet. Sci. Lett., vol. 31,1976, pp. 269-278.

- Ringwood, A. E.: Chemical Evolution of the Terrestrial Planets. Geochim. Cosmochim. Acta, vol. 30, Jan.1966, pp. 41-104.

- Rubey, W. W.: Geologic History of Sea Water. Bull. Geol. Soc. Am., vol. 62, 1951, pp.1111-1147.

- Sun, S.-S.; Nesbitt, R. W.; and Sharaskin, A. Y.: Geochemical Characteristics of Mid-Ocean Ridge Basalts. Earth Planet. Sci. Lett.,1979.

- Tolstikhin, I. N.: A Review: Some Recent Advances in Isotope Geochemistry of Light Rare Gases. In: Terrestrial Rare Gases, E. C. Alexander Jr., and M. Ozima, eds., Center for Academic Publications, Tokyo Japan,1978, pp. 33-62.

[110] - Turekian, K. K.; and Clark, S. P., Jr.: The Non-homogeneous Accumulation Model for Terrestrial Planet Formation and the Consequences for the Atmosphere of Venus. J. Atmos. Sci., vol. 32, June 1975, pp. 1257-1261.