Periodic Reporting for period 1 - TEMPO (Combining Tectonics and Machine Learning to Constrain Plate Reconstruction Models Through Time)
Période du rapport: 2019-08-25 au 2021-08-24
The Problem
Plate tectonics processes continuously destroy oceanic crust, which contain the most reliable record of plate motion. There is therefore little data to constrain net rotation of the lithosphere with respect to the deep mantle, constraints on which are required to produce accurate reference frames for plate motion. Accurate plate motion models are required for various geoscience applications including mantle, geodynamo and geochemical modelling of the Earth's interior, and ocean, atmospheric and ecosystem modelling at the surface. Uncertainties in the record of net rotation propagate through plate motion models and eventually limit the certainty on estimates of things such as models of climate change in the geological past.
Society
Plate motion studies have an important role to play in the modelling of climate change and how the Earth will respond to them. Without accurate plate reconstructions, we cannot model the contributions of changing surface configurations to the changes in the temperature of the Earth's surface, sea-level changes and greenhouse gas concentrations. Without accurate knowledge of how the planet responded to changing greenhouse gas concentrations in the geological past, we cannot accurately model how it is likely to respond to anthropogenic global warming.
Objectives
I want to use mantle dynamic simulations to study what drives net rotation, to find how to constrain it and therefore how to reduce the uncertainties in plate tectonic reconstructions.
Plate tectonics processes continuously destroy oceanic crust, which contain the most reliable record of plate motion. There is therefore little data to constrain net rotation of the lithosphere with respect to the deep mantle, constraints on which are required to produce accurate reference frames for plate motion. Accurate plate motion models are required for various geoscience applications including mantle, geodynamo and geochemical modelling of the Earth's interior, and ocean, atmospheric and ecosystem modelling at the surface. Uncertainties in the record of net rotation propagate through plate motion models and eventually limit the certainty on estimates of things such as models of climate change in the geological past.
Society
Plate motion studies have an important role to play in the modelling of climate change and how the Earth will respond to them. Without accurate plate reconstructions, we cannot model the contributions of changing surface configurations to the changes in the temperature of the Earth's surface, sea-level changes and greenhouse gas concentrations. Without accurate knowledge of how the planet responded to changing greenhouse gas concentrations in the geological past, we cannot accurately model how it is likely to respond to anthropogenic global warming.
Objectives
I want to use mantle dynamic simulations to study what drives net rotation, to find how to constrain it and therefore how to reduce the uncertainties in plate tectonic reconstructions.
I used mantle convection simulations to investigate the controlling factors for the magnitude of lithospheric net rotation (LNR) and to find the statistical predictability of LNR in a fully self-consistent convective system. We find that high lateral viscosity variations are required to produce Earth-like values of LNR. When the temperature dependence of viscosity is lower, and therefore slabs are softer, other factors such as the presence of continents and a viscosity gradient at the transition zone are also important for determining the magnitude of net rotation. We find that, as an emergent property of the chaotic mantle convection system, the evolution of LNR is too complicated to predict in our models. However, we find that the range of LNR within the simulations follows a Gaussian distribution, with a correlation time of 5 Myr. The LNR from the models needs to be sampled for around 50 Myr to produce a fully Gaussian distribution. This implies, that within the time frames considered for absolute plate motion reconstructions, LNR can be treated as Gaussian variable. This provides a new geodynamic constraint for absolute plate motion reconstructions.
The Conclusions
I used mantle convection simulations to investigate the controlling factors for the magnitude of lithospheric net rotation (LNR) and to find the statistical predictability of LNR in a fully self-consistent convective system. I find that high lateral viscosity variations are required to produce Earth-like values of LNR. When the temperature dependence of viscosity is lower, and therefore slabs are softer, other factors such as the presence of continents and a viscosity gradient at the transition zone are also important for determining the magnitude of net rotation. I find that, as an emergent property of the chaotic mantle convection system, the evolution of LNR is too complicated to predict in our models. However, I find that the range of LNR within the simulations follows a Gaussian distribution, with a correlation time of 5 Myr. The LNR from the models needs to be sampled for around 50 Myr to produce a fully Gaussian distribution. This implies, that within the time frames considered for absolute plate motion reconstructions, LNR can be treated as Gaussian variable. This provides a new geodynamic constraint for absolute plate motion reconstructions.
Using the same set of data, I also studied the effect of changing plate configuration on carbon storage in the crust. This is an element of the carbon cycle, locking away carbon from the surface and atmospheric carbon reservoir for tens of millions of year, before it is either degassed through volcanoes or locked away for much longer periods in the Earth's interior. I found that volatile storage increase when mid-ocean ridges form or change direction. The primary driver of changes in flux of volatile for degassing or subduction is the state of the ridge system 10s of millions of years previously, rather than the rate of subduction.
In the process of investigating measures of net rotation, I developed a machine learning tool that accelerates the calculation of seismic anisotropy as result of mantle flow. This was developed as a side project and is currently being used by research groups in Lyon and Montpellier. An article about the tool is in preparation and the tool will be released as open source software at the same time.
The Conclusions
I used mantle convection simulations to investigate the controlling factors for the magnitude of lithospheric net rotation (LNR) and to find the statistical predictability of LNR in a fully self-consistent convective system. I find that high lateral viscosity variations are required to produce Earth-like values of LNR. When the temperature dependence of viscosity is lower, and therefore slabs are softer, other factors such as the presence of continents and a viscosity gradient at the transition zone are also important for determining the magnitude of net rotation. I find that, as an emergent property of the chaotic mantle convection system, the evolution of LNR is too complicated to predict in our models. However, I find that the range of LNR within the simulations follows a Gaussian distribution, with a correlation time of 5 Myr. The LNR from the models needs to be sampled for around 50 Myr to produce a fully Gaussian distribution. This implies, that within the time frames considered for absolute plate motion reconstructions, LNR can be treated as Gaussian variable. This provides a new geodynamic constraint for absolute plate motion reconstructions.
Using the same set of data, I also studied the effect of changing plate configuration on carbon storage in the crust. This is an element of the carbon cycle, locking away carbon from the surface and atmospheric carbon reservoir for tens of millions of year, before it is either degassed through volcanoes or locked away for much longer periods in the Earth's interior. I found that volatile storage increase when mid-ocean ridges form or change direction. The primary driver of changes in flux of volatile for degassing or subduction is the state of the ridge system 10s of millions of years previously, rather than the rate of subduction.
In the process of investigating measures of net rotation, I developed a machine learning tool that accelerates the calculation of seismic anisotropy as result of mantle flow. This was developed as a side project and is currently being used by research groups in Lyon and Montpellier. An article about the tool is in preparation and the tool will be released as open source software at the same time.
I produced the first data set explicitly designed to study net rotation. It has allowed me to treat net rotation as a statistical variable with a distribution and uncertainty. This therefore allows the net rotation to be treated as a statistically constrained variable in plate reconstructions, rather than an arbitrarily chosen parameter.
I am also in the process of using statistics to fill the gaps in the geological record qua oceanic lithosphere creation and destruction. Statistical methods have never previously been applied to this topic.
The anisotropy machine learning tool will accelerate the calculation of anisotropy to such a degree that geodynamic models that require continuous recalculation of anisotropy will now be computationally feasible. This is of particular importance for seismic hazard studies because it will allow for better models of the likelihood of large scale plate motion and possibility of earthquakes at different locations.
I am also in the process of using statistics to fill the gaps in the geological record qua oceanic lithosphere creation and destruction. Statistical methods have never previously been applied to this topic.
The anisotropy machine learning tool will accelerate the calculation of anisotropy to such a degree that geodynamic models that require continuous recalculation of anisotropy will now be computationally feasible. This is of particular importance for seismic hazard studies because it will allow for better models of the likelihood of large scale plate motion and possibility of earthquakes at different locations.