THE EFFECT OF SEDIMENTARY ROCK OUTCROP AREA ON ESTIMATES OF MARINE PHANEROZOIC DIVERSITY

Since the 1970s, scientists have realized that biases may obscure the true history of biodiversity (richness) in the fossil record. One of the most widely recognized geological biases is change in the amount of preserved sedimentary rock through time. Larger areas of outcrop allow for greater sampling of taxa, leading to inflated diversity estimates. Several researchers, notably Raup (1976) and Peters and Foote (2001), have attempted to assess the effects of outcrop area on recovered diversity, but reliance on proxy data (named formations, sediment volumes) for global outcrop area serves to impart some ambiguity in the interpretation of results. Here, we use a new dataset of global outcrop areas for Phanerozoic sedimentary rock, parsed by age and lithology and derived from the UNESCO Geological Atlas of the World. Global sedimentary rock outcrop areas are partitioned into marine and non-marine classes using rock abundance data from Ronov, and assigned to epoch-intervals using the Gradstein et al. 2004 time scale. This data set allows us to use actual areas of outcrop for marine sedimentary rocks in order to explore the relationship with recorded diversity.

We compare outcrop data to two independent records of the Phanerozoic diversity of skeletonized marine invertebrates – the Sepkoski genus-level compendium and the Paleobiology Database record of sampled-in-bin (not standardized) diversity. Using both ordinary least squares and nonlinear regression analyses, we demonstrate a highly significant power-law relationship between perceived diversity and outcrop area, suggesting that area of available outcrop accounts for a large amount of the variability in the empirical record of diversity through time. The residuals exhibit no distinct temporal trends, and do not support the hypothesis that there was a substantial increase in diversity during the Phanerozoic. The relationship between our diversity residuals and occurrence-standardized diversity estimates generated by the Paleobiology Database is quite strong (r2 = .53 for the globe, .64 for North America), indicating that both approaches to correcting sampling biases are converging on comparable diversity histories.