An improved geomagnetic data selection algorithm for global geomagnetic field modelling

A spherical harmonic degrees 60, global internal field model is described (called
BGS/G/L/0706). This model includes a degree 15 core and piecewise-linear secular variation
model and is derived from quiet-time Ørsted and CHAMP satellite data sampled between
2001.0 and 2005.0. For the satellite data selection, a wide range of geomagnetic index and
other data selection filters have been used to best isolate suitably quiet magnetospheric and
ionospheric conditions. Only a relatively simple, degree one spherical harmonic, external field
model is then required. It is found that a new ‘Vector Magnetic Disturbance’ index (VMD), the
existing longitude sector A indices, the auroral zone index IE, and the polar cap index PC are
better than Kp and Dst at rejecting rapidly varying external field signals at low, middle, auroral
and polar latitudes. The model quality is further enhanced by filling spatial and temporal gaps
in the quiet data selection with a second selection containing slightly more disturbed data.
It is shown that VMD provides a better parametrization than Dst of the large-scale, rapidly
changing, external field. The lithospheric field model between degrees 16 and 50 is robust and
displays good coherence with other recently published models for this epoch. BGS/G/L/0706
also shows crustal anomalies consistent with other studies, although agreement is poorer in the
southern polar cap. Intermodel coherency reduces above about degree 40, most likely due to
incompletely filtered signals from polar ionospheric currents and auroral field aligned currents.
The absence of the PC index for the southern hemisphere for 2003 onwards is a particular
concern.