HR: 1340h
AN: U43A-0826    [Abstracts]
TI: Optimizing the Depth-Age Transformation
AU: * McMillan, D G
EM: dgm@yorku.ca
AF: Department of Earth and Space Science and Engineering, York University 4700 Keele Street, Toronto, ON M3J 1P3 Canada
AU: Stoner, J S
EM: jstoner@coas.oregonstate.edu
AF: College of Oceanic and Atmospheric Sciences, Oregon State University 104 COAS Administration Building, Corvallis, OR 97331-5503 United States
AB: Much of our understanding of Earth's past comes from cores of marine and lacustrine sediments. In order to properly reconstruct the temporal evolution of the paleosystem we must produce a time series of measurements by making a transformation from the depth domain to the age domain. Ideally, this requires comprehensive knowledge of the sediment accumulation rate over the entire history of the core. But in the absence of such precision, we are forced to make simplifying assumptions about the accumulation rate. The most common, and to date probably the most reliable, method of depth-age transformation is to tie known ages, based on complementary data, such as oxygen isotope events, directly observed climate events or radiometric determinations, to a finite number of depths and assume the accumulation rate is constant between these tie points. The resulting piece-wise continuous accumulation rate function has a discontinuity at the depth of each tie point---a feature which is not required by the age data. Such jumps have been shown to cause significant decorrelation when comparing, for example, two relative paleointensity time series, whose only difference is their tie point data. In addition, the ages we obtain from the complementary data have uncertainties associated with them and therefore, there is no reason to desire an accumulation rate that fits those data exactly. We now recognize that we have all the elements of the classical geophysical inverse problem: an unknown function of one spatial variable (accumulation rate as it varies with depth); a geophysical data set (the tie point ages and their uncertainties); and a method of forward modeling (recovering age data given an accumulation rate function). In this work, we develop and implement several methods for optimizing the depth-age transformation, all of which are based on the well-known ideas of geophysical inversion. We compare results from minimum norm, minimum deviation and minimum structure solutions and assess the ability of each method to reproduce an accumulation rate function by using a statistical model of random variations in sediment accumulation and sub-sampled, radiocarbon dated depth-age data from high resolution marine cores. New chronologies based on the optimized transformations yield improvements in common geomagnetic signal, which are illustrated by examining the coherence between several pairs of relative paleointensity time series.
DE: 0520 Data analysis: algorithms and implementation
DE: 1105 Quaternary geochronology
DE: 1165 Sedimentary geochronology
DE: 1507 Core processes (1213, 8115)
DE: 1560 Time variations: secular and longer
SC: Union [U]
MN: Fall Meeting 2005