29-10-2012, 12:20 PM
Volume Rendering with Animation of Gulf Stream Currents
Volume Rendering.pdf (Size: 389.37 KB / Downloads: 19)
Abstract
This paper describes the novel use of volume rendering for
the visualization of MICOM oceanographic data and the
tracking of mesoscale eddies. We present methods for generating
3D volumetric data and describe how to use VIRVO
to build an animation for MICOM Gulf Stream motion data.
This put the data in a format that easily gives oceanographers
the ability to visualize mesoscale eddies from any
angle, including an undersea aspect, over any duration,
and by combinations of spatial dimensions. One advantage
of this technique is that only minimal processing of
the oceanographic data is required. High or low density
objects, corresponding to heat-laden currents, can be enhanced
or masked by modifying a dynamic transfer function,
also providing the ability to visualize cooler return waters.
While the data format is relatively simple and the software
is easily installed on a personal computer, the same
animations can be viewed in an immersive virtual reality
environment.
Introduction
Understanding the ocean through computer models is important
and numerous models exist. Several, including MICOM,
the Miami Isopycnic Coordination Ocean Model,
and HYCOM, the HYbrid Coordinate Ocean Model, use
an isopycnic (equal density) model that has been successful
in modeling currents. Other isopycnic models include the
Hallberg Isopycnal Model (HIM), the Navy Layered Ocean
Model (NLOM), and the Parallel Oregon State University
Model (POSUM).
The DYNAMO Project compared three ocean models
and only isopycnic models predicted the Azores Current
and a correct trajectory for the North Atlantic Current [1].
Oceanographers study the adiabatic properties of the sea,
looking at variables such as heat, velocity, density and salinity.
Visualizing synthetic variables, such as heat index [2],
a function of velocity and temperature, can be difficult. The
velocity component itself is a function of Eastward and
Poleward current movement, so visualization can be quite
useful.
Previous Work
The earliest visualization tools were developed at the (National
Center for Atmospheric Research (NCAR) in the
1960s. NCAR’s Man-Computer Interactive Data Access
System (McIDAS) and the AFOS (Automation of Field Operations)
system from the National Weather Service date
back to the 1970s. During the 1980s systems like PROFS
concentrated on developing highly interactive applications.
Despite these advances, geoscience requirements for 3D
visualization have always been challenging. Data collected
hourly and over vast areas can easily overwhelm databases
and conventional modeling and visualization tools. Both
the atmosphere and ocean are relatively thin when compared
with the earth’s vast surface, and stretching of the zaxis
is required to visualize geophysical phenomena. Viewing
multiple attributes of geospatial data simultaneously requires
associating each with either a color or transparency
so that each may be viewed separately. Representing flow
must often employ 2D techniques such as arrow plots,
streamlines, advected particles, wire frame overlays, digital
elevation models, color-mapping techniques, and quiver
plots, but they may not easily adapt to 3D.
3D Modeling Techniques
Recent computational projects involving global ocean models
are numerous and span most of the ocean models. Although
most ocean models are 3-dimensional, many visualization
projects model current activity only in two dimensions.
A number of researchers have developed 3D visualization
tools, but we are particularly interested in 3D volumetric
approaches which model the ocean as a solid with
variations in density and temperature.
While the following is not intended to be a comprehensive
survey, some of the efforts in 3D volume rendering
visualization of oceanographic include cloud water density
mapping of LAMPS data [5], template matching and time
tracking applied to NOAA/Levitus ocean data [6], color
mapping of volumetric data [7], POP and MICOM data [8],
VRML for volume visualization [9], OVIRT [10], motion
estimation in volumetric data [11], and ParVox [12]. Volume
rendering applications in other domains, such as visualization
of astrophysics [13] or atmospheric data have
been explored. Most, however, require either servers, mainframes,
supercomputers, special-purpose workstations, immersive
virtual reality environments, or custom application
software. None met our needs for an applications programming
interface with an open architecture that runs on a desktop
compactly, provides real-time rendering, time series, animation,
and has immersive virtual reality and parallel processing
options.
Methodology
The MICOM data product we used is from ECMWF (European
Centre for Medium-RangeWeather Forecasts) data.
Available variables are southward barotropic velocity, eastward
barotropic velocity, barotropic pressure, mixed layer
Montgomery potential, mass weighted average mixed layer
sigma theta, mass weighted average southward velocity,
mass weighted average eastward velocity, layer thickness
in pressure units, mass weighted average layer temperature,
and mass weighted average layer salinity.
MICOM ECMWF products also can provide wind stress
vector, wind velocity, surface radiation, humidity, air temperature,
and precipitation data. Heat flux is calculated from
surface radiation, air temperature, specific humidity, wind
speed and sea surface temperature. Fresh water flux is calculated
from evaporation estimates, specific humidity and
sea surface temperature.
One feature of the MICOM model is a top, mixed layer.
This layer interacts with not only the other layers but the
wind and solar system, hence the name. Oceanographers
who use isopycnic models are interested in properties of
these layers. We are particularly interested in thermal transport
characteristics and the function of eddies and jets.
Building a 3-dimensional representation of an area of interest,
then having a visual framework to check our assumptions
is invaluable.
Conclusion
We have presented a method of visualizing MICOM data
using the VIRVO volume rendering API. Volume rendering
is an intuitive approach to representing 3-dimensional data.
Data is easily converted for visualization and user software
requires minimal hardware and software resources. MICOM
data can be rendered easily and diagnostics and other
surfaces can be added to the volumetric object. Immersive
virtual reality and parallelization are current options with
VIRVO. Using VIRVO, even desktop PCs can host small,
fast, easily-prepared and easy-to-use oceanographic visualizations.