 |
| Titel |
Holocene global geomagnetic field reconstruction based on archeomagnetic data: Assessing error sources and uncertainties |
| VerfasserIn |
Roman Leonhardt, Karl Fabian, Elisabeth Schnepp |
| Konferenz |
EGU General Assembly 2010
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 12 (2010) |
| Datensatznummer |
250039915
|
|
|
|
|
|
| Zusammenfassung |
| Using a Bayesian inversion technique, which minimizes the total variational power at the
core-mantle boundary under data constraints, several spherical harmonic geomagnetic field
models up to degree five are established for the Holocene period covering the last 5000 years.
These models are based on different collections of archeomagnetic data and historic
observations. Using a bootstrap type statistical analysis the influence of data quality upon the
reconstruction of the Gauss coefficients is analyzed. In particular, the influences of
uncertainties in ages, magnetic field vectors as well as spatial and temporal distribution are
investigated. Besides Gaussian data scatter, also the influence of systematic measurement bias
is discussed.
The first step in field reconstruction is the selection of data based on various quality
parameters. It is shown that this data selection significantly affects the resulting
characteristics of the model. The analysis confirms that age uncertainties can lead
to a significant masking of short term field variations. The enormous spread in
archeointensity and related ages uncertainties obfuscates underlying magnetic field variations
completely for some regions. Including only the most trustworthy data into the
inversion reduces the scatter in regional data and, most importantly, the possible bias.
However, it also reduces temporal and spatial data distribution significantly. The
analysis suggests regions and time intervals for which a re-evaluation of existing
data, or an extension of the existing data base is of particular importance in order
to enhance the validity and reduce the uncertainty range of a global field model. |
|
|
|
|
|
|