Paper No. 4
Presentation Time: 8:00 AM-6:00 PM
RECENT ADVANCES IN GEOCHEMICAL FINGERPRINTING OF FOSSIL VERTEBRATES FROM THE EOCENE-OLIGOCENE WHITE RIVER SEQUENCE
Vertebrate fossils on Federal lands are under continuous threat of illegal collection. Short of catching poachers red-handed in the field, methods of determining the provenance of suspected stolen fossils are currently limited to sedimentological and mineralogical comparisons, styles of fossilization, or matching a broken fossil to its complimentary piece. The trace element signature and isotopic composition that is imparted to bone during the fossilization process provides a powerful tool to determine the provenance of stolen fossils. The geochemical fingerprint of a fossil is defined by ratios of rare earth and trace elements (REE/TE) that reflect redox and pH conditions and the degree of sediment-water interaction at the site of fossilization. These ratios are variable on the local and regional scale, and produce distinctive signatures which can be analyzed using discriminant analysis. In an effort to improve the fidelity of this method, we recently began incorporating C and O stable isotopes from the carbonate component of fossil bones into our geochemical database. As a test of this method, we compared vertebrate fossils from the Late Eocene White River Sequence at Flagstaff Rim, administered by the Bureau of Land Management, near Casper, WY to those from temporally equivalent deposits from Toadstool Geologic Park, administered by the U.S. Forest Service, near Crawford, NE. Using only REE/TE signatures, fossils from these two sites could be correctly classified only 93% of the time, but with addition of stable isotopes, the provenance of all of the fossils was correctly determined. We are currently measuring stable isotopes in additional fossils from the White River Sequence of NE and SD for which we already have REE/TE data to determine if the addition of stable isotopes will improve the ability to determine provenance of fossils from sites that previously yielded correct classifications of only 92-93%, and to what degree discrimination varies between stratigraphic units and over local to regional scales.