An Assessment of Robust Holocene Geomagnetic Field Structures

Friday, 19 December 2014
Catherine Constable, University of California San Diego, La Jolla, CA, United States, Monika C Korte, Helmholtz Centre Potsdam GFZ German Research Centre for Geosciences, Potsdam, Germany and Sanja Panovska, IGPP/SIO/UCSD, San Diego, CA, United States
Globally distributed paleomagnetic data from archeological artifacts and young volcanics have been combined with those from lacustrine and marine sediments to produce an increasing number of time-varying regularized geomagnetic field models that span the past 10~000 years. The spatial representation is in spherical harmonics while time variations are parameterised as cubic B-splines, and the model regularization is through quadratic norms, often the Ohmic dissipation norm and the 2nd derivative of the time variations. Results are influenced by global distribution and quality of the data and age constraints and by details of the modeling procedure. The latter include relative weighting according to assigned uncertainties, misfit measure (L1 or L2 norm), and outlier rejection. Calibration of relative paleointensity (RPI) observations and relative declinations is also an important issue, and the results are sensitive to the starting model used for initial data calibrations. It is therefore important to ensure that only absolute observations are used in the initial calibration. The most recent Holocene field models have better uncertainty estimates and improved calibration of both relative declination and RPI data, providing better field representations than earlier generations of models.

Our assessment of Holocene field structure is based on an evaluation of selected models derived from essentially the same dataset as that used to produce the CALS10k.1b model. At Earth's surface robust common features among the various models are a north/south hemispheric asymmetry, with stronger average fields in the northern hemisphere and greater overall variability in the southern hemisphere. Longitudinal structure is also present, with greatest variability in the Atlantic hemisphere, but the signal is not entirely consistent across the various models. Nevertheless at the core-mantle boundary a systematic picture is beginning to emerge of the effects of heterogeneous boundary forcing in the 0-10 ka magnetic field. Correlations with seismic structure in the lowermost mantle are discussed.