2007 GSA Denver Annual Meeting (28–31 October 2007)

Paper No. 15
Presentation Time: 5:15 PM

STRATIGRAPH, A NEW R PACKAGE FOR ANALYSIS AND DISPLAY OF STRATIGRAPHICALLY DISTRIBUTED PALEONTOLOGICAL DATA


GREEN, Walton A., Department of Geology and Geophysics, Yale Univ, P. O. Box 208109, Yale Station, New Haven, CT 06520 and JARAMILLO, Carlos, Center for Tropical Paleoecology and Archeology, Smithsonian Tropical Research Institute, Balboa, Ancon, 0843-03092, Panama, walton.green@yale.edu

The basic data behind stratigraphic paleontology and community paleoecology consist of times (dates), taxa, and locations. All three of these variables are scale-dependent, and in real-life situations such data are always `noisy'; that is, incorrect dates, misidentifications, and approximate locations are common features of such data sets. Procedures for identifying a geological or biological signal through such noise fall into the framework of exploratory data analysis. stratigraph is a software toolbox, now available in pre-release versions, to analyze and display such data. Specifically, functions are currently available to read pollen or other micropalaeontologial data from and write them to standard interchange formats, and to plot pollen diagrams. Current software tools for this kind of analysis include expensive commercial packages like Rockware, StratBugs, and CANOCO, but menu-driven compiled software does not provide the generality or flexibility needed to tune analyses to particular data sets. stratigraph is intended to show the advantages of interactive data analysis via open-source scripting in a high-level programming language. While menu-driven software may eventually be needed by some end-users, the programmers writing such software are seldom scientists or familiar with the detailed questions that their programs will be used to answer. Instead, community-supported, open-source software is an ideal locale in which to develop new methodologies or implement established methods in stratigraphic paleontology.