Few data, sparse data - data mining in the paleomagnetic database

Enlarge

Today, the Earth's magnetic field is permanently observed and measured by observatories and satellites. There is no directly measured data on particular features of the Earth's magnetic field, such as magnetic field reversals. Rocks with certain iron-containing minerals can record the Earth's magnetic field and are therefore our only window to the geomagnetic field in the past. These records are very important for our understanding of the geodynamo over geologic time. However, such global reconstructions of the time-averaged geomagnetic field on a long-term geological scale are only meaningful if sufficient data are available. Unfortunately, the amount of data from paleomagnetic studies is comparatively small because they are quite complex. To reconstruct the geomagnetic field, data from longer periods of time must therefore be bunched together in order to gather enough data to make a valid statement about the entire Earth's magnetic field.

Sanja wants to make progress in our understanding of the geodynamo millions of years ago possible and is therefore evaluating existing paleomagnetic studies for their suitability for such reconstructions. In particular, Sanja would like to compile a good data basis for very long stable phases without field reversals called superchrons and also hyperreversing phases, and model these periods.

Collecting data for special times in the history of the geodynamo

The data from previous palaeomagnetic analyses were compiled in a database. The palaeomagnetic data for certain time periods and/or the spatial distribution across the globe are sometimes quite sparse and also of varying quality. It would of course be desirable to use only high-quality data, but at the same time a sufficiently covering distribution should be maintained in order to be able to make statements about the global geomagnetic field. This data distribution and its quality are the key criteria for determining which time periods are suitable for further analysis.

When analyzing data over geological time periods, however, plate tectonics must also be taken into account: The place where the samples for a paleomagnetic analysis were taken today was not necessarily the same several million years ago, as the earth's plates have moved over time. Sanja applies various plate tectonics models to her data in order to determine their influence on field models.

Abstract

Towards global time-average models of the geomagnetic field over deep geological time

Sanja Panovska¹, Richard Bono², Yael Engbers³, Monika Korte¹

1: GFZ Potsdam; 2: Florida State University Tallahassee, FL, USA; 3: Netherlands Organization for Applied Science TNO, The Hagues, The Netherlands

Studying the Earth’s magnetic field variations throughout geological history is essential for understanding the dynamical processes in the Earth’s outer core and coupling with other layers. Global reconstructions of the long-term variations are only possible when sufficient data are available for a given time period. Recent progress in compiling paleomagnetic data resulted in a database that contains paleomagnetic intensity and directions since 4 billion years ago (PINT). Overall, the paleomagnetic records are sparsely distributed in time and space and are of varying quality. It is essential to select high-quality data but, at the same time, to keep the spatial distribution sufficient for building global models. We consider the data distribution and quality the key criteria in determining which time periods to analyze. When data allow, given periods are modelled with the whole dataset and with data separated into normal and reversed datasets. One further issue that needs to be addressed is the paleospatial data distribution. Different plate motion models are used to reconstruct the site original locations in order to obtain robust field estimates. After we select the time intervals over the Earth’s history, we will build global, time-averaged field models. Available statistical parameters, such as from Giant Gaussian Process models derived for different periods, will be compared to the global models. Our aim is to reconstruct the time-averaged field morphology for intervals when the field was stable for a very long time, or very active, as well as characteristic geological epochs. The results are important for providing information on distinct states of the paleomagnetic field and, thus, on the driving mechanism of the geodynamo.

Towards global time-average models of the geomagnetic field over deep geological time. Sanja Panovska, Richard Bono, Yael Engbers, Monika Korte, Castle Meeting 2024, Utrecht, NL, June 30 – July 6, 2024, Session B: Archeomagnetism; Geomagnetism; Environmental magnetism, 2024