Andreas Nilsson

Paleomagnetic data, although undeniably limited, have been used to constrain complex time-varying geomagnetic field models. Here we present four different versions of a time varying geocentric dipole model for the past approximately 9000 years based on a limited selection of sedimentary paleomagnetic data with good spatial coverage. The robustness of the modeling approach is evaluated using modern magnetic observation station data provided by the World Data Centre for Geomagnetism in Edinburgh and we discuss the magnitude of the error introduced by insufficient data coverage. The prediction in terms of directional variations is tested against paleomagnetic data from 63 different sedimentary sites and compared to results from recently published spherical harmonic geomagnetic field models. We find that the dipole estimate can explain the paleomagnetic data almost as well as the more complicated spherical harmonic models. Specific areas and records are identified where there are significant differences between the performance of the dipole estimate and the spherical harmonic models. Our results suggest that many of the differences between the dipole model predictions and the paleomagnetic data can be related to poor paleomagnetic quality and/or the lack of accurate and precise time control inherent to some of the records. These results agree with Valet et al. ( 2008) who suggested that the CALS7K. 2 model probably underestimates the dipole contribution in order to reduce the misfit between the inversion and the data. We conclude that better paleomagnetic data and associated time control are needed to improve global geomagnetic field models.