GP23B-1314
Paleosecular variation and time-averaged field analysis over the last 10 Ma from a new global dataset (PSV10)
Tuesday, 15 December 2015
Poster Hall (Moscone South)
Geoffrey Cromwell, Occidental College, Los Angeles, CA, United States, Catherine L Johnson, University of British Columbia, Department of Earth, Ocean and Atmospheric Sciences, Vancouver, BC, Canada, Lisa Tauxe, Scripps Institution of Oceanography, La Jolla, CA, United States, Catherine Constable, University of California San Diego, La Jolla, CA, United States and Nicholas Jarboe, Organization Not Listed, Washington, DC, United States
Abstract:
Previous paleosecular variation (PSV) and time-averaged field (TAF) models draw on compilations of paleodirectional data that lack equatorial and high latitude sites and use latitudinal virtual geomagnetic pole (VGP) cutoffs designed to remove transitional field directions. We present a new selected global dataset (PSV10) of paleodirectional data spanning the last 10 Ma. We include all results calculated with modern laboratory methods, regardless of site VGP colatitude, that meet statistically derived selection criteria. We exclude studies that target transitional field states or identify significant tectonic effects, and correct for any bias from serial correlation by averaging directions from sequential lava flows. PSV10 has an improved global distribution compared with previous compilations, comprising 1519 sites from 71 studies. VGP dispersion in PSV10 varies with latitude, exhibiting substantially higher values in the southern hemisphere than at corresponding northern latitudes. Inclination anomaly estimates at many latitudes are within error of an expected GAD field, but significant negative anomalies are found at equatorial and mid-northern latitudes. Current PSV models Model G and TK03 do not fit observed PSV or TAF latitudinal behavior in PSV10, or subsets of normal and reverse polarity data, particularly for southern hemisphere sites. Attempts to fit these observations with simple modifications to TK03 showed slight statistical improvements, but still exceed acceptable errors. The root-mean-square misfit of TK03 (and subsequent iterations) is substantially lower for the normal polarity subset of PSV10, compared to reverse polarity data. Two-thirds of data in PSV10 are normal polarity, most which are from the last 5 Ma, so we develop a new TAF model using this subset of data. We use the resulting TAF model to explore whether new statistical PSV models can better describe our new global compilation.
