|
|
|
| Title | Visualization of Flow Behavior in Earth Mantle Convection
(Article) |
|---|
| in | IEEE Transactions on Visualization and Computer Graphics |
|---|
|
| Author(s) |
Simon Schroeder, John A. Peterson, Harald Obermaier, Louise Kellogg, Kenneth I. Joy, Hans Hagen |
|---|
|
| Keyword(s) | Mantle convection, geophysical visualization, particle density, volume rendering, curvilinear grid |
|---|
|
| Year |
October 2012
|
|---|
| Volume | 18 |
|---|
| Number | 12 |
|---|
| Publisher | IEEE |
|---|
|
| URL | http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=6327224 |
|---|
| Pages | 2198--2207 |
|---|
|
| BibTeX |  |
|---|
|
| Abstract |
A fundamental characteristic of fluid flow is that it causes mixing:
introduce a dye into a flow, and it will disperse. Mixing can be used as a
method to visualize and characterize flow. Because mixing is a process that
occurs over time, it is a 4D problem that presents a challenge for computation,
visualization, and analysis. Motivated by a mixing problem in geophysics, we
introduce a combination of methods to analyze, transform, and finally visualize
mixing in simulations of convection in a self-gravitating 3D spherical shell
representing convection in the Earth's mantle. Geophysicists use tools such as
the finite element model CitcomS to simulate convection, and introduce massless,
passive tracers to model mixing. The output of geophysical flow simulation is
hard to analyze for domain experts because of overall data size and complexity.
In addition, information overload and occlusion are problems when visualizing a
whole-earth model. To address the large size of the data, we rearrange the
simulation data using intelligent indexing for fast file access and efficient
caching. To address information overload and interpret mixing, we compute
tracer concentration statistics, which are used to characterize mixing in mantle
convection models. Our visualization uses a specially tailored version of
Direct Volume Rendering. The most important adjustment is the use of
constant opacity. Because of this special area of application, i.e. the
rendering of a spherical shell, many computations for volume rendering can be
optimized. These optimizations are essential to a smooth animation of the
time-dependent simulation data. Our results show how our system can be used to
quickly assess the simulation output and test hypotheses regarding Earth's
mantle convection. The integrated processing pipeline helps geoscientists to
focus on their main task of analyzing mantle homogenization.
|
|---|
|
|
