Oxygen isotope geochemistry of cherts from the Onverwacht Group (3.4 billion years), Transvaal, South Africa, with implications for secular variations in the isotopic composition of cherts

δ¹⁸O values for 87 chert samples from the 3.4-b.y.-old Onverwacht Group, South Africa, range from +9.4 to +22.1‰. δ-values for cherts representing early silicified carbonates and evaporites, and possible primary precipitates range from +16 to +22‰ and are distinctly richer in ¹⁸O than silicified volcaniclastic debris and cherts of problematical origin. The lower δ-values for the latter two chert types are caused by isotopic impurities such as sericite and feldspar, and/or late silicification at elevated temperature during burial. Cherts with δ-values below +16‰ are thus not likely to yield geochemical data relevant to earth surface conditions. Fine-grained chert is less than 0.7‰ depleted in ¹⁸O relative to coexisting coarse drusy quartz. Because coarse quartz preserves its isotopic composition with time, the maximum amount of post-depositional lowering of the δ-values of cherts by long-term isotopic exchange with meteoric groundwaters does not exceed 0.7‰ in 3.4 b.y. In response to metamorphism the δ-values of Onverwacht cherts appear to remain unchanged or to have increased by as much as 4‰. Neither metamorphism nor long-term isotopic exchange with groundwaters can explain why Onverwacht cherts are depleted in ¹⁸O relative to their Phanerozoic counterparts. Meteoric waters with a δ¹⁸O range of at least 3‰ appear to have been involved in Onverwacht chert diagenesis. δ-values for possible primary cherts or cherts representing silicified carbonates and evaporites are compatible with the depositional and diagenetic environments deduced from field and petrographic evidence. Onverwacht cherts appear to have formed with δ-values at least 8‰ lower than Phanerozoic cherts. The new Onverwacht data combined with all published oxygen isotope data for cherts suggest a secular trend similar to that initially suggested by Perry (1967) in which younger cherts are progressively enriched in ¹⁸O. However, Precambrian cherts appear to be richer in ¹⁸O than Perry's original samples and can be reasonably interpreted in terms of declining climatic temperatures from ∼70°C at 3.4 b.y. to present-day values, as initially suggested by Knauth and Epstein (1976). This surface temperature history is compatible with existing geological, geochemical, and paleontological evidence.