Full text loading...
Abstract
The fossil record preserves a wide range of events that might be used to build timescales and correlate strata from place to place. The events include the originations and extinctions of species, the occurrence of distinctive faunal assemblages, magnetic field reversals, changes in ocean chemistry, and volcanic ash falls. The fundamental task is to determine the global sequence of all these events. Modern computer algorithms can build high-resolution timescales by sequencing and calibrating thousands of such events from hundreds of localities. Each successful sequencing algorithm can be understood in terms of a simple two-dimensional visual device. Graphic correlation, permutable sequences, permutable matrices, and slotting devices are each suited to a different data problem, such as contradictory evidence of sequence, a lack of information about sequence, or uncertain correspondence of events. Algorithms based upon permutable sequences and evolutionary programming heuristics have the flexibility to optimize sequences with a wide variety of event types and data problems; they are slower than the more narrowly tailored methods. All methods will be challenged to keep pace with the amount of biostratigraphic information that is accumulating in the latest generation of shared databases. Currently, it is possible to sequence sufficiently large numbers of events to imagine a potential resolution of 10,000 to 50,000 years over time spans on the order of 50 to 100 million years. The actual resolving power is less because the solutions to these sequencing problems are not unique. Multiple equally good solutions typically emerge and, to extract a consensus sequence with which all solutions agree, some runs of events must be collapsed into unresolvable clusters. Nevertheless, quantitative methods have been shown to improve resolution up to tenfold over traditional methods that discard many potentially useful events.