The effects of temporal resolution on species turnover and on testing metacommunity models.

Patterns of low temporal turnover in species composition found within paleoecological time series contrast with the high turnover predicted by neutral metacommunity models and thus have been used to support nonneutral models. However, these predictions assume temporal resolution on the scale of a season or year, whereas individual fossil assemblages are typically time averaged to decadal or centennial timescales. We simulate the effects of time averaging by building time-averaged assemblages from local dispersal-limited, nonaveraged assemblages and compare the predicted turnover with observed patterns in mollusk and ostracod fossil records. Time averaging substantially reduces temporal turnover such that neutral predictions converge with those of trade-off and density-dependent models, and it tends to decrease species dominance and increase the proportion of rare species. Observed turnover rates are comparable to an appropriately scaled neutral model: patterns of high community stability can be produced or reinforced by time averaging alone. The community attributes of local time-averaged assemblages approach those of the metacommunity. Time-averaged assemblages are thus unlikely to capture attributes arising from processes operating at small spatial scales, but they should do well at capturing the turnover and diversity of metacommunities and thus will be a valuable basis for analyzing the large-scale processes that determine metacommunity evolution.