READING THE FOSSIL RECORD: USING THE LITERATURE TO ANALYZE ECOLOGICAL PATTERNS IN CENOZOIC TAXONOMIC DIVERSITY, NEW ZEALAND

Literature-sourced data are now used commonly in a broad range of paleontological investigations, particularly in the construction of diversity curves at a variety of temporal and spatial scales. These data are also applicable to the investigation of ecological trends on a macroevolutionary time-scale.

A collection-based occurrence and taxonomic database has recently been compiled from the literature on marine macroinvertebrate faunas in the late Cenozoic of New Zealand to investigate temporal paleoecological patterns and their relationship to changing biodiversity. These data include macrofaunal assemblage lists compiled from the literature spanning the Oligocene to Pleistocene, and comprise 27,000 taxon occurrences, representing 600 genera (including bivalves, gastropods, scaphopods and articulate brachiopods), from 2,400 stratigraphic localities. Overall, raw diversity patterns derived from these data are consistent with those of other published diversity estimates, suggesting that this published subset of data is representative of broad-scale temporal patterns in this region. As a test of the quality of these data, a number of additional diversity indices were analyzed with respect to geography and a range of paleoenvironmental controls, including water depth, sedimentary facies, and paleotemperature. Two important features are apparent: (1) data are not evenly distributed with regard to geography and paleoenvironmental settings through time; and (2) alpha diversity within these settings is highly variable. Understanding of such variation in the structure of biodiversity datasets is essential to meaningful interpretation of temporal diversity and paleoecologic trends.

The construction of this dataset has highlighted a range of problems inherent to assemblies of literature-sourced data in biodiversity and paleoecologic investigations. These include the broad range in quality of taxonomic data (both accuracy and nomenclatural issues), and availability and quality of supplementary data (i.e. age, sedimentologic, geographic, and paleoecologic information). A summary of these issues, their potential influence on diversity and paleoecologic analyses and the procedures developed to avoid them will be discussed here.