The central theme of this project is to advance our understanding of the processes controlling magma generation and transport in subduction zones, one of the primary modes for heat/mass transport from the deep Earth to Earth’s surface.   The 55,000 km long system of subduction zones is responsible for processes that define our surface environment: volcanic output, continental growth, mountain building, faulting, major earthquakes, tsunami generation, composition and mass of the atmosphere and oceans.   Our work produces improved models for the interplay between plate tectonics, mantle circulation and magma delivery back to the surface through constructs called volcanic arcs, or island arcs, which has been a challenging problem for past models due to the extreme range in length scales that must be represented.   These range from the 1000 km length scales of tectonic plates moving through subduction zones and related mantle flow, to length scales of 1 km for buoyant upwellings, called diapirs, that are shed from the topside of downgoing plates.  Modeling tools must also represent the scales by which mantle melt (molten rock) moves relative to the material in the rising diapirs, bringing the total length scale range from 0.001-1000 km, or 6 orders of magnitude.      

On this project our team has utilized a multi-model approach, one that combines physical laboratory fluids models with very high resolution numerical/computer models.   A unique aspect of this work is that with our colleagues from Woods Hole (Behn, Zhang), the combination of laboratory geodynamical models and numerical models covers the required, extreme range in scales.  We develop laboratory models of subduction, background mantle flow and diapirs that utilize a real fluid, capable of simulating processes over essentially an infinite range in length scales. However, lab models cannot represent melt migration.    For this we utilize very high resolution numerical models of small scale heat and chemical transport within diapirs, where the evolution of diapirs on their rise within the mantle are prescribed from the lab models.  

The suite of lab geodynamical subduction models are the first ones to represent three-dimensional, time-evolving (or 4-D) circulation of the mantle wedge that include various real world aspects of these systems: a) different modes of plate sinking (constant dip angle, rollback steepening), b) the influence of gaps or tears in the sinking plate and c) the influence of high or low viscosity residuum material, or what is left behind in the wedge after melt/magma extraction has occurred.  Results underscore the importance of including real-world plate motions and moving from 2-D to 4-D representations of subduction zone processes.  For each style of background mantle flow, an extensive suite of models on subduction-diapir interactions in the mantle wedge have been completed.  One particularly interesting outcome is that often used schematic representations of vertical magma ascent in subduction zones is found to be nearly impossible to produce in the 4D models.  Instead, magma expressed through the system of subduction zone volcanoes is far more likely to have followed a complex, circuitous path from deep mantle source to its eruption into the light of day.  A number of factors are shown to influence how deeper magma presents itself at the surface.  The common, and often neglected paddle-like mode of subduction, the so-called rollback motion of sinking lithopshere, produces the most pronounced deflections of diapirs between source and surfacing locations.   An unexpected result is that this simple interaction of rollback flow with diapirs can explain linear, age progressive volcanic tracks observed in regions like the Cascades Convergent Margin and the Western USA (e.g. the High Lava Plains and the Snake River Plain).   The presence of a gap or window in the sinking plate can also lead to severely distorted diapir pathways and timelines, that need ot be considered in building interpretations for surface geologic data sets.   Finally, lab models reveal a new mode for interactions between diapirs that can greatly effect rise rates, rise paths, surfacing locations and geochemical interpretations.   

The complexity of connecting the magma at a specific volcano to the deeper source region of that magma, highlighted so clearly in the lab models, has been incorporated into teaching units.  Students are asked to place passive tracers deep within the 4D subduction/viscous fluid flow models of convergent margins, in locations that allow them to surface beneath a hypothetical volcano.  Through this fun activity students rapidly develop an intuition for 4D subduction flows, and a realization for how unlikely the simple schematic diagrams used in most textbooks are for representing actual transport pathways.  The physical, hands-on subduction lab models developed on this project have been effective at engaging students in geodynamics. Future goals include making the subduction lab models more accessible to a range of university and high school classes using web-based, remote learning technologies. 

Last Modified: 12/12/2017
Modified by: Christopher R Kincaid