HOW MUCH IS TOO MUCH? EXPLORING THE EFFECTS OF TIME-AVERAGING ON PALEO-COMMUNITY DATA

Time-averaging is the bane and blessing of paleontology. The depth of time is the most important aspect of paleontology, however biologists often cite the lack of temporal resolution as the biggest weakness of the fossil record. To investigate this issue, twenty data-sets were selected to represent environmental disturbances across a gradient of spatial and temporal scales. Environments include lake and open ocean systems, hosting phyto- and zoo-plankton communities, during acidification or rapid warming. Sampling resolution ranges from biweekly to 100,000 years, with 15-300 community samples in each dataset. This gradient allows comparison across multiple time-scales with similar environmental disturbances. In addition to the chosen variation in temporal resolution, this study applies artificial time-averaging, testing for the level of time-averaging that disguises change caused by disturbance. Community data was compared using ordination techniques, evenness, and rank-abundance curves/species-abundance distributions.

At the smallest scale (acid lakes with biweekly sampling), each metric shows extensive variation or noise. After the environmental perturbation, variation increases greatly. This pattern is reflected in each data-set, however the amount of background noise before disturbance is lessened as spatial and temporal scales are increased. The change in community structure (regardless of metric) is always greater during the disturbance than before the disturbance. This pattern is strengthened when the time-averaging within a data-set is artificially increased (mixing each sample with specimens from the samples directly before and after). These results indicate that time-averaging has a positive effect on paleoecological analysis of fossil communities. Although time-averaging does dampen the magnitude of change resulting from an environmental disturbance, it also dampens the normal ecological variation (noise) inherent in biological systems. If the environmental disturbance is across a large spatial scale and lasts more than 1,000,000 years, time-averaging on the scale of 100,000 years will not obscure the biotic response to the event. Therefore, time-averaging is actually useful in community analysis and rarely obscures biotic response to environmental disturbances.