The growth of the continental crust: Constraints from zircon Hf-isotope data

E. A. Belousova*, Y. A. Kostitsyn, W. L. Griffin, G. C. Begg, S. Y. O'Reilly, N. J. Pearson

*Corresponding author for this work

    Research output: Contribution to journalArticlepeer-review

    712 Citations (Scopus)


    A worldwide database of over 13,800 integrated U-Pb and Hf-isotope analyses of zircon, derived largely from detrital sources, has been used to examine processes of crustal evolution on a global scale, and to test existing models for the growth of continental crust through time. In this study we introduce a new approach to quantitatively estimating the proportion of juvenile material added to the crust at any given time during its evolution. This estimate is then used to model the crustal growth rate over the 4.56. Ga of Earth's history. The modelling suggests that there was little episodicity in the production of new crust, as opposed to peaks in magmatic ages. The distribution of age-Hf isotope data from zircons worldwide implies that at least 60% of the existing continental crust separated from the mantle before 2.5. Ga. However, taking into consideration new evidence coming from geophysical data, the formation of most continental crust early in Earth's history (at least 70% before 2.5. Ga) is even more probable. Thus, crustal reworking has dominated over net juvenile additions to the continental crust, at least since the end of the Archean. Moreover, the juvenile proportion of newly formed crust decreases stepwise through time: it is about 70% in the 4.0-2.2. Ga time interval, about 50% in the 1.8-0.6. Ga time interval, and possibly less than 50% after 0.6. Ga. These changes may be related to the formation of supercontinents.

    Original languageEnglish
    Pages (from-to)457-466
    Number of pages10
    Issue number3-4
    Publication statusPublished - Oct 2010


    Dive into the research topics of 'The growth of the continental crust: Constraints from zircon Hf-isotope data'. Together they form a unique fingerprint.

    Cite this