The destruction of paleoclimatic isotopic signals in Pleistocene carbonate soil nodules of Western Australia

Stable carbon and oxygen isotopic analyses were conducted on small (< 10 mm), spherical Pleistocene carbonate soil nodules from 53 different levels in a 9.15-m soil auger collected in northern Western Australia. The delta(13)C and delta(18)O values range from -4 to 0parts per thousand and -9 to -6parts per thousand (Pee Dee Belemnite standard), respectively. Comparison of these values to the delta(13)C values of coexisting soil organic matter (SOM) reveals marked differences between the two records. The SOM values record an enrichment of similar to16parts per thousand through the section, whereas the carbonate values record only a 4parts per thousand variability. Application of a diffusion-production model to the data indicates isotopic disequilibrium between the two carbon records. One possible interpretation of the apparent isotopic disequilibrium is that the nodules formed at depths of similar to1 cm, but this interpretation is not supported by field observations. An alternative is that the soil organic matter currently in the profile was emplaced after nodule formation. Such a scenario would have to have occurred numerous times during the Late Pleistocene as the nodules grew below different soil zones through time. A third explanation is that diagenetic alteration of the carbonates in the presence of dissolved inorganic groundwater carbon has occurred. This interpretation is consistent with the widely fluctuating regional groundwater table. It is also supported by petrographic fabrics and heterogeneous cathodoluminescence patterns indicative of initial pedogenic development of the nodules and then at least one dissolution event, multiple precipitation events, prolonged nodule growth by displacive spars and microspars, cementation, and recrystallization of some micrite. The results illustrate potential problems for (1) the use of some pedogenic carbonate isotopic records as proxies for environmental change, and (2) the use of 'altered' delta(18)O values and 'unaltered' delta(13)C values as screening tools for the presence or absence of diagenesis. The results highlight the necessity of understanding regional environmental conditions and diagenetic histories prior to using pedogenic carbonates in paleoreconstructions. (C) 2002 Elsevier Science B.V. All rights reserved.