Quantifying slab evolution and mantle flow using global subduction models with data assimilation

Peng, Diandian

Permalink

https://hdl.handle.net/2142/115526 Copy

Description

Title

Quantifying slab evolution and mantle flow using global subduction models with data assimilation

Author(s)

Peng, Diandian

Issue Date

2022-04-06

Director of Research (if dissertation) or Advisor (if thesis)

Liu, Lijun

Doctoral Committee Chair(s)

Liu, Lijun

Committee Member(s)

• Bass, Jay
• Gregg, Patricia
• Guenthner, William

Department of Study

Geology

Discipline

Geology

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• plate tectonics
• mantle convection
• geodynamic model
• stagnant slab
• flat slab
• sinking rate

Abstract

The solid state convection of Earth’s mantle is a global-scale time-dependent process. While geologic and geophysical observations provide constraints on some snapshots of this temporal evolution, numerical simulation is a natural approach to connecting these snapshots and reproducing the continuous dynamic history. However, numerical models usually suffer simplified model assumptions and uncertain mantle parameters. A model with more realistic Earth-like setting is warranted to reproduce the mantle dynamics. In this dissertation, we build global mantle convection models with sequential data-assimilation, which satisfies the tectonically reconstructed time-dependent plate motion and seafloor ages. To calibrate the model behavior and validate model outputs, we use a multitude of data constraints, including seismic tomography, seismic anisotropy, lithospheric structure, sedimentation history and thermochronological results. We show that the global models successfully reproduced observed slab structures at all major subduction zones, where the fit to data is significantly improved relative to previous studies. With these high resolution global models, we first study the formation of stagnant Pacific and Philippine Sea slabs beneath East Asia whose length can exceed 1500 km based on tomography. We find that previously proposed mechanisms, including trench retreat, mantle resistant forces due to phase transformation and/or viscosity change at 660 km discontinuity, as well as seafloor age variation, are not important for slab stagnation below East Asia. Instead, the dominant mechanism is a westward Cenozoic mantle wind induced by the large lateral gradients of dynamic pressure across the hinge of the Izanagi and Tethyan slabs. This pressure gradient had a long history since the middle Mesozoic: it was progressively enhanced over time as the subduction duration of these two slabs increased and reached a maximum value during the Late Cretaceous, resulting in a continental-scale flat Izanagi slab beneath East Asia. This newly discovered flat Izanagi slab can explain multiple tectonic events and lithospheric structures in East Asia. During the flat subduction, the lowermost lithosphere above the slab was eroded via abrasion and the resulting thinner lithosphere east of the North-South Gravity Lineament (NSGL) than the western side conforms to observation. This explains the latest-Cretaceous formation of the NSGL, a problem debated for decades. At the same time, the lithosphere was thickened locally beneath the NSGL as observed, suggesting that this flat subduction may also the reason for the uplift of Greater Xing’an-Taihang-Xuefeng Mountains along the NSGL. The flat subduction temporally reversed the stress state in the lithosphere above from extension to east-west compression, where the resulting crustal thickening explains the uplift/reversal of sedimentary records in multiple basins in eastern China during 75-65 Ma. Further analysis of the global models generates important insights on multiple fundamental properties of slab evolution and tectonic implications. We show that the descending slabs can migrate horizontally up to 6000 km during the past 200 Ma. This challenges a common assumption that slabs think vertically into the deep mantle, with the implication that the seismically observed geographic location of a slab is not necessarily marking its ancient trench position as usually assumed in tectonic reconstructions based on seismic tomography. In addition, the sinking rate of slabs is not of a constant value, as generally assumed, but varies with subduction duration, depth, and geographic locations. The overall slab sinking rates are notably larger than previously adopted, suggesting that many previous interpretations of ancient tectonic problems based on observed slabs need to be reconsidered. In effect, the quantitative migration history of slabs in our global models could help improving such tectonic reconstructions. We also evaluate the feasibility of regional models in representing the actual mantle dynamics by comparing them with the global models. We find that the two types of models generate similar slab structures and mantle flow below North America and Tonga, both matching seismic tomography. In East Asia and South America, however, incorrect slab structures are produced due to limitations of regional models including the artificial side boundaries, lack of global-scale mantle flow and interaction among different slabs. Overall, these studies show that global mantle convection models with data-assimilation are the most useful for understanding the dynamic evolution of the solid Earth.

Graduation Semester

2022-05

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Diandian Peng

Owning Collections