Stochastic optimisation of absolute geomagnetic palaeointensity determinations

I propose to develop a new mathematical tool that will make measurements of the Earth's magnetic field strength in the ancient past much more reliable and efficient. The magnetic field of the Earth extends far into space and is important to humans for many reasons. It is used by us and other species for navigation and it also protects human technology from the 'solar wind' - a stream of high-energy particles emitted by the sun. The interaction of the solar wind with the Earth's magnetic field causes aurorae (the Northern and Southern Lights) and other 'space weather' phenomena. The initiation of a strong, global 'geomagnetic' field more than three billion years ago may have been a crucial factor in allowing the first life on Earth to appear. Before then, the atmosphere might not have been able to form because of continual erosion by the solar wind. The importance of the Earth's magnetic field to human civilisation and life in general means that it is very important that we study it and learn as much about it as we can. Another good reason for this is that it can tell us a great deal about the place where it is generated: the deep interior of the Earth. At any one point on the Earth's surface, the magnetic field has both direction and intensity which both vary rather erratically in time. To study the present-day behaviour of the Earth's magnetic field, we use specialist satellites and a global network of magnetic observatories. To study the field in the distant past, we must turn to the geological record and volcanic rocks in particular. These lock-in the direction and intensity of the field at the time and place that they cool from molten lava and therefore provide a globally-distributed 'palaeomagnetic' record for almost the whole of Earth's history. The ancient direction of the Earth's magnetic field as recorded in rocks is much easier to measure than is its ancient strength (called its 'palaeointensity'). However, absolute palaeointensity records are essential for allowing us understand the geomagnetic field and its history. The problem with measuring the palaeointensity is that the rocks which are used can be affected by many complex physical factors which can bias the result. Furthermore, the precise way the measurement is carried out can also affect its reliability. A lot of recent work has gone into improving our understanding of these problems but a lack of synthesis means that palaeomagnetists still do not agree on which rocks and experimental methods produce reliable palaeointensity measurements. There is also some disagreement over which of the thousands of palaeointensity measurements which have already been published can be trusted and which should be disregarded as unreliable. These disagreements could largely be overcome if we had objective, quantitative information about the likely success of any particular palaeointensity experiment. My proposal is to provide this by developing an entirely new 'stochastic' (i.e. partially random) numerical model of palaeointensity experiments which can optimise: experimental design, analytical and procedure, and objectively determine the reliability of published data. I will rigorously constrain and test this model using new and published experimental data and ultimately, I will employ it to obtain important new information including the strength of the Earth's magnetic field more than 3 billion years ago. The benefits of this work will be considerable. It will enable future palaeointensity studies to be performed with much greater efficiency and will also allow us to get the most out of the thousands of palaeointensity determinations which are already published. Our understanding of the Earth's magnetic field, its formation in the outer core, and its protection of society and life as a whole will all be improved as a result.