Paper No. 7
Presentation Time: 3:25 PM

NAILS, JACKS AND CLAMPS: MECHANICAL ANALOGS AS A WAY TO THINK ABOUT ORGANIZING VAST AND VARIED STRATIGRAPHIC INFORMATION SETS FOR COMPUTER ALGORITHMS THAT OPTIMIZE TIME LINES AND CALIBRATE TIME SCALES (Invited Presentation)


SADLER, P.M., Department of Earth Sciences, University of California, Riverside, Riverside, CA 92521, peter.sadler@ucr.edu

The volume of time-stratigraphically relevant information from even a single Paleozoic System threatens to overwhelm manual approaches to correlation and seriation. Hundreds of local stratigraphic sections present different and partial sequences of thousands of biologic, chemical and physical events. It is advantageous to use numerical algorithms that can apply human stratigraphic logic to build a single composite time line of events that respects the local superpositional relationships while recognizing how the freedom to adjust the local observations varies between different classes of stratigraphic information. Graphic correlation and related numerical methods routinely incorporate bentonites and taxon ranges, for example. A bentonite represents a brief moment in time; it has a fixed position at each locality and must be matched exactly. Bentonites exemplify the class of information analogous to nails. A local taxon range, by contrast, matches or under-represents the time interval of the true global range; it may be stretched to fit the global time line. Taxon ranges belong in the class of information analogous to jacks; extending them is work to be minimized. Stable isotope excursions are better handled as conservative uncertainty intervals, a data class with properties and freedoms that are the opposite of taxon ranges. Uncertainty intervals are chosen to match or over-represent truly correlative segments and inflexion intervals in time series of isotopic ratios. They are analogous to clamps and can be squeezed to make all sections fit the same time line of events. The analytical uncertainty of a radioisotopic age acts as a press on the global time scale but is nailed to the local sample analyzed. Sequencing algorithms seek an optimal time line to which all sections can be fit without moving the nails and by the least expenditure of energy to extend the jacks and squeeze the clamps, a property that is straightforward to program as a trial and error search. The resulting time line is parsimonious in the sense that it minimizes implied diachronism and ad hoc adjustments of the local field observations. By these means, fundamentally different kinds of stratigraphic information may be combined efficiently in a mutually supportive fashion without over-stating the precision of any of them.