V33F-01 INVITED
Models for mechanics and morphology of arcuate subduction
We provide an explanation for the polarity, localization, shape, size, and stability of subduction zones on Earth by considering the deformation of a relatively soft, thin doubly-curved lithospheric cap, partially supported by a thick, nearly incompressible mantle. Double curvature leads to an effectively stiffer system and different characteristic subduction geometry than planar plates. Using a combination of scaling concepts, numerical simulation, and analog experiments, we show that freely subducting plate boundaries form strongly localized and cuspate segments shaped by the compromise between bending and stretching in edge buckling modes of thin shells. In contrast, for upper-plate-forced subduction, we find relatively straight subduction zones, consistent with the localized deformations of loaded indented shells. Taken together, these two instabilities match the observed scaling and morphology of subduction arcs, oceanic plates and other large-scale features of the Earth.
V33F-02
Long-lived tendrils in viscous entrainment
Experiments on thermal convection with two layers of miscible liquids have revealed a rich set of dynamics. In particular, there is a stratified regime where thin tendrils of one liquid are entrained by the convective flow in the second liquid. The tendrils remain stable over many convection cycles, reminiscent of the way hot-spots on earth are believed to persist over many mantle convection cycles. To shed light on the how such fragile structures can persist, we examine the steady-state structure formed by viscous entrainment at the boundary between two miscible liquid layers. In general, when the effect of the entrainment penetrates deeply into the lower layer, a broad tendril forms with the interior liquid moving uniformly upwards. When the effect of the entrainment penetrates weakly, the velocity field inside the tendril develops an interior stagnation point. We show that a model for the tendril that assumes the tendril is entirely decoupled from the background flow, i.e. includes only local information, is degenerate, with multiple tendril states for one entrainment condition. Including information about the global geometry of the withdrawal flow removes the degeneracy while introducing only a logarithmic dependence on the global flow parameters. This weak variation with global parameters may be one reason for the tendril's persistence. It also allows us to propose a simple scaling law for the volume flux entrained.
V33F-03 INVITED
Effects of granular charge on flow and mixing
Sandstorms in the desert have long been reported to produce sparks and other electrical disturbances – indeed as long ago as 1850, Faraday commented on the peculiarities of granular charging during desert sandstorms. Similarly, lightning strikes within volcanic dust plumes have been repeatedly reported for over half a century, but remain unexplained. The problem of granular charging has applied, as well as natural, implications, for charged particle clouds frequently generate spectacularly devastating dust explosions in granular processing plants, and sand becomes strongly electrified by helicopters traveling in desert environments. The issue even has implications for missions to the Moon and to Mars, where charged dust degrades solar cells viability and clings to spacesuits, limiting the lifetime of their joints. Despite the wide-ranging importance of granular charging, even the simplest aspects of its causes remain elusive. To take one example, sand grains in the desert manage to charge one another despite having only similar materials to rub against over expanses of many miles – thus existing theories of charging due to material differences fail entirely to account for the observed charging of desert sands. In this talk, we describe recent progress made in identifying underlying causes of granular charging, both in desert-like environments and in industrial applications, and we examine effects of granular charging on flow, mixing and separation of common granular materials. We find that charging of identical grains can occur under simple laboratory conditions, and we make new predictions for the effects of this charging on granular behaviours.
V33F-04
Particle sorting in dense granular flows
Mixtures of particles tend to unmix by particle property. One of the most dramatically destructive examples of this occurs in debris flow: boulders, rocks, and mud tumble down a hillside, and the largest rocks migrate toward the top and then the front of the flow where they do the most damage. Rotating drums and chute flows are two of the most common apparatuses used to systematically study segregation in dense, gravity driven granular flows. In these cases, smaller or, alternatively, denser particles segregate away from the free surface, phenomena that have been modeled using mechanisms such as kinetic sieving and buoyancy, respectively. Other segregation mechanisms have been identified in suspensions and in more energetic systems such as a gradient in granular temperature -- the kinetic energy of velocity fluctuations -- and curvature effects. However, with most experimental systems the dominant segregation mechanism is difficult to ascertain. In typical experimental systems designed to study segregation in dense granular flow (such as chutes and rotated drums), gravity, velocity gradients and porosity gradients coexist in the direction of segregation. We study the segregation of mixtures of particles numerically and experimentally in a split-bottom cell and in a rotating drum to isolate three possible driving mechanisms for segregation of densely-sheared granular mixtures: gravity, porosity, and velocity gradients and their associated dynamics. We find gravity alone does not drive segregation associated with particle size without a sufficiently large porosity or porosity gradient. A velocity gradient, however, appears capable of driving segregation associated both with particle size and material density in dense flows. We present our results and discuss the implications for some particle segregation behaviors observed in natural systems such as debris flows and sediment transport.
V33F-05 INVITED
The 'Unreasonable Effectiveness' of Stratigraphic and Geomorphic Experiments
Experiments involving processes at the scales of channels and channel networks appear to be far more effective at capturing natural dynamics than they should be based on formal classical scaling. After illustrating this with examples, we attempt to understand what it implies for experimental design and scale independence in morphodynamic systems. We use the following set of ideas: (1) Barenblatt's concept of complete and incomplete similarity; (2) Reynolds number independence as an example of a broader class of asymptotic independence of dynamics with respect to dimensionless parameters at their limits; and (3) internal versus external similarity, i.e. the relation between the similarity associated with fractal geometry and similarity among whole systems of different scale. Complete dynamical scaling using classical dimensionless variables guarantees external similarity but is nearly impossible to apply in practice. Two alternatives are partial dynamical scaling, for which the effects of the unmatched dimensionless numbers are often not known; and reliance on internal similarity as an indicator of scale independence and hence of external similarity across a range of the governing dimensionless parameters. Developing a quantitative understanding of the origins and limits of natural similarity, internal and external, seems to us to be a more productive way of fully exploiting the potential of geomorphic and stratigraphic experiments than the classical-scaling approach.
V33F-06
The fluid dynamics of bedform superimposition
Many sedimentary environments possess a range of bedforms, of different scale, that both modify the mean flow and create their own turbulent flow field. Although much research over the past 20 years has documented the mean and turbulent flow field over a range of bedforms, in both air and water, there has been very little study of the dynamics of how such bedforms interact and how the flow field of one bedform may be modulated by the superimposition of another, smaller, bedform. This paper presents results of a laboratory study, using particle imaging velocimetry (PIV), that sought to document the interactions between the flow fields of two bedforms as a smaller ripple becomes superimposed on a larger form. Experiments were conducted in a flume, 0.3m wide, 0.20m deep and 17m long, in which a simplified triangular, two-dimensional bedform (ripple) was placed in the test section with a water depth of 0.06m over the ripple crest. The flow depth: bedform height (Y:h) ratio was 6.7, and the flow field in the leeside of this bedform was quantified using PIV at a sampling rate of c. 11 Hz. A second, smaller, bedform was then superimposed upon the first, at different positions along the stoss side of the principal bedform, until the two merged, yielding a combined bedform with a Y:h ratio of 4. The changing flow fields in the leeside of the larger bedform were again investigated with PIV. This change in bedform height approximates that across the ripple-dune transition in aqueous flows. This paper will detail results on the influence of this bedform superimposition on the characteristics of the flow separation zone generated in the leeside of the larger bedform, and specifically the interactions between the shear layers formed bounding the two separation zones as superimposition progresses. These results show the significant influence of bedform sheltering as superimposition proceeds but that, at a certain critical spacing, interaction of the two separation zone shear layers creates an enhanced turbulent flow field as compared to flow at any other stages of superimposition, including full amalgamation. The paper will detail these flow fields and the implications of this for the behavior of interacting bedforms within a range of sedimentary environments.
V33F-07 INVITED
Integrating multiphase, macroscopic models of particle-laden volcanic flows with experimentally-derived subgrid models
Explosive volcanic eruptions produce turbulent, multiphase flows that encompass a vast range of scales, from micron-scale ash to plumes and density currents that extend for hundreds of kilometers. One of key challenges in understanding these flows is reconciling the role of microphysical processes that occur at the scale of individual particles with the macroscopic and collective dynamics in plumes and pyroclastic flows. Of particular interest is understanding large-scale emergent dynamics that arise from the mass, momentum and energy exchanges at small scales. To account for processes that occur at scales smaller in space and time than those resolved in simulations of large scale dynamics (typically meters, and fractions of a second) we have developed subgrid scales models based on laboratory experiments. we perform the experiments with natural volcanic particles at conditions similar to those that exist in real flows. We identify multiphase dimensionless groups that are important for mass, momentum and energy transfer and show how subgrid scale processes can be included into a multiphase, continuum equations describing large scale dynamics. We illustrate this approach and the importance of subgrid models with three examples: the role of steam generation at the particle-scale when pyroclastic flows enter the ocean, the generation of ash-size particles through particle interactions within pyroclastic flows, the role of boundary conditions (over land versus over water) on the structure and transport distance of pyroclastic flows. All three examples illustrate the two-way coupling that occurs between large scale dynamics and micro-scale phenomena.
V33F-08
3D Numerical Simulations of Coupled Solid and Fluid Mechanics in Volcanic Conduit Erosion and Crater Formation
An essential element of explosive volcanic eruptions is the effect of the evolving conduit and vent on the erupting multiphase flow and the effect of the flow upon the conduit and vent rocks, a 3D geological nozzle problem. This coupling of the host rock solid mechanics with the fluid dynamics of an erupting multiphase fluid has never been directly simulated and is poorly understood. We apply a library of computer codes called CFDLib, which has been developed by the Theoretical Division at Los Alamos National Laboratory. This code provides the unique capability of being able to solve the interaction of an Eulerian fluid with a Lagrangian solid in 3D while treating multiphase turbulence that this interaction generates. Our previous work with CFDLib has been directed at validating results with laboratory experiments, verification against analytical models, and free-jet decompression. This work demonstrated the importance of vent overpressure in determining the characteristics of an erupted column of gas and tephra. However, eruption of an overpressured jet is strongly coupled to the dynamics of the vent shape that in turn is dependent upon conduit dynamics. For this reason most previous computer simulations of volcanic eruptions have assumed pressure-balanced conditions of flow from the vent. Here we demonstrate our progress in simulating vent evolution during eruption of an overpressured multiphase (steam and magma/rock) fluid. With increasing overpressure the evolved vent radius increases with the formation of a crater. The Mach Stem structure of the erupted jet resembles those of our previous simulations from a fixed vent, but the evolving vent nozzle and contributions of eroded material to the jet make its structure more complicated and variable with time. Future work will focus on study of the effects of host rock properties and 3D conduit shape.