ESTIMATES OF TIME-AVERAGING IN TERRESTRIAL GASTROPOD ASSEMBLAGES FROM THE QUATERNARY OF CANARY ISLANDS

Shells collected from a single stratigraphic horizon tend to vary in age from hundred to thousand years. Such temporal mixing (time-averaging) reduces the resolution of the fossil record and may generate false patterns. Thus, evaluating the magnitude and structure of time-averaging is essential for reconstructing past environments and ecological conditions.

Quantitative estimates of time-averaging in gastropod shell accumulations from Quaternary eolian deposits of Canary Archipelago were obtained by direct dating of individual shells collected from exceptional well-preserved dune and paleosol land snail assemblages. A total of 203 shells of the helicids Theba geminata and T. arinagae, representing 44 samples (=stratigraphic horizons) from 14 sections, were dated using amino acid (isoleucine) epimerization ratios calibrated with 12 radiocarbon dates. Most samples reveal a substantial variation in shell age that exceeds the error that could be generated by dating imprecision, with the mean within-sample shell age range of 6,670 years and the mean standard deviation of 2,920 years. Even the most conservative approach (Monte Carlo simulations with a non-sequential Bonferroni correction) indicates that at least 25% of samples must have undergone substantial time-averaging (e.g., age variations within those samples cannot be explained by dating imprecision alone). Samples vary in shell age structure, including both left-skewed (17 out of 44) and right-skewed distributions (26 out of 44). Dispersion and shape of age distributions of samples do not show any notable correlation with the stratigraphic age of samples, suggesting that the structure and scale of temporal mixing is time invariant. The statistically significant multi-millennial time-averaging observed here is consistent with previous studies of shell accumulations from various depositional settings, and reinforce the importance of dating numerous specimens per horizon in geochronological studies. Unlike in the case of marine samples, typified by right-skewed age distributions (attributed to an exponential-like shell loss from older age classes), many of the samples analyzed here displayed left-skewed distributions, suggestive of different dynamics of age mixing in marine versus terrestrial shell accumulations.