TIME-AVERAGING AND TAPHONOMY OF DRILL HOLES: BAD NEWS FROM THE COLORADO RIVER DELTA

Time-averaging and taphonomic processes have been long recognized as important biasing factors that can alter ecological patterns or create spurious trends. To explore this issue for one of the commonly quantified paleoecological parameters (drilling predation frequency), 58 bulk samples of the bivalve mollusk Mulinia coloradoensis were analyzed for drill holes. These samples were collected from supratidal beach ridges that have formed recently by accretionary reworking of extensive tidal flats and represent spatially homogenized and extensively time-averaged shell accumulations, which mix mollusks that lived in the area during the last 1000 years. Given this extensive homogenization, it can be expected that drilling frequencies should not vary significantly across samples collected from a single beach ridge: spatial and temporal mixing that occurred when beach ridges formed should mask any spatial patchiness or short-term ecological variation.

The results confirm this prediction: regardless of lateral and vertical position within a beach ridge, all reasonably large samples (n>30) yield similar drilling frequency estimates. This pattern persists when data from all sampled beach ridges are pooled together. Also, a Monte Carlo simulation indicates that the variation in frequency observed across samples is comparable to what would be expected when randomly sampling a homogenous system. In addition, when corrected for shell disarticulation, all samples display absurdly high drilling frequencies (most samples vary between 90 and 115%) and the overall drilling frequency (104.6%) significantly exceeds the maximum biologically possible value of 100%.

These results are all bad news. Not only did beach ridges completely homogenize the short term variations and spatial patchiness that likely existed in local ecosystems, but also that homogenized average is an overestimate, most likely, due to preferential onshore transport of drilled shells. Whereas supratidal beach ridges are ephemeral deposits, which do not constitute an important component of the marine fossil record, the results provide a real life example of the worst-case scenario when time-averaging and taphonomy conspire both to average short term signals and to bias the resulting long-term average.