Lec8_VariabilityInCoastalObservatories
Download
Report
Transcript Lec8_VariabilityInCoastalObservatories
Visualizing Spatial and Temporal
Variability in Coastal Observatories
Walter H. Jimenez Wagner T. Correa
Claudio T. Silva Antonio M. Baptista
OGI School of Science & Engineering,
Oregon Health & Science University
Originally Published in the Proceedings of the
14th IEEE Visualization Conference (VIS’03)
Bob Armstrong
MSIM 842
28 Feb 2007
Summary of this Paper
This paper is about
improvements to the
Columbia River's
environmental
observation and
forecasting system,
called CORIE.
Their new tools add
3D and 4D
visualization to
CORIE.
How We'll Proceed
The Problem
The Motivation
Their Approach
Conclusion
Questions
The Big Picture
SEATTLE
This is the area
of
CORIE sensors
PORTLAND
The Big Picture
Take off, you hoser.
This is the area
of
CORIE sensors
The CORIE Sensor Field
Jetty A CORIE Station
The Problem
In 2003, there was a mismatch between CORIE's
simulation capabilities and generated data
System based on high-resolution time-varying 3D
unstructured grids
Also included a visualization component that for the
large part only generates 1D or 2D plots
Some 3D information can be inferred from depth
plots, which slice the 3D data
CORIE is the region's Environmental Observation
and Forecasting Systems (EOFS)
Forecast of Velocity Magnitudes
Location of CORIE Stations
Environmental Observation and Forecasting
Systems
Seek to generate and deliver quantifiably reliable
information about the environment
Rely heavily on modeling and visualization
CORIE
Observation Network
Advanced Modeling System
Real-Time telemetry from over 20 stations
Models the complex circulation in the river and plume
Information Management System
Web Publishing of Processed Data
Why should we care about this problem?
Interesting visualizations
They desire volume rendering versus isometric
surface renderings
Very easy to interpret output
Vector treatment is logical
Colorization is logical and very telling
Moving towards real-time representation of largescale fluid fields
Preprocessing of CORIE Output
CORIE output data must be converted in order to
be visualized
Volume rendering of the salinity scalar
Requires an unstructured grid of tetrahedrons
Each vertex is associated with one salinity scalar
Rendering of the bathymetry requires a grid of
triangles representing the ground surface
Velocity field visualization
Requires an unstructured grid of points
Vector attributes associated with each point
Sample CORIE Data
Station
: Fort Stevens Wharf (USCG day mark
red26)
Identifier
: red26
Latitude
: 46 12.447 N
Longitude
: 123 57.084 W
Instrument depth
: 7.5m
Year
: 2006
Month
: December
Time reference
: Pacific Standard Time
Last revision
: 1/29/2007
Data reviewed by
: cseaton
Expunged temperature measurement : 0 %
Expunged salinity measurement
: 0 %
Expunged depth measurement
: 0 %
Records removed for time
: 0.03 %
Comments:Depth data referenced to NGVD29 MSL.
See http://www.ccalmr.ogi.edu/CORIE/data/publicarch/methods_meanings.html
for meanings, formats and quality control procedures
yyyy/mm/dd hh:mm:ss saliniy
temp
depth
(PST)
(PSU)
(C)
(m)
#########################################
2006/12/01
00:00:21 23.3
9.2
-9999.00
2006/12/01
00:01:03 23.7
9.2
-9999.00
2006/12/01
00:02:06 23.0
9.2
-9999.00
2006/12/01
00:02:48 23.6
9.2
-9999.00
More CORIE Model Output
Unstructured Grid
Triangles are elements
Verticies are nodes
Column of points is variable due to depth
Attribute value at each node can be
Scalar -> salinity
Vector -> velocity
Volume rendering is performed on tetrahedrons
derived from wedges
CORIE Output Data
Bathymetry
Volumetric Rendering
Wedges look like slices of pie
Wedge is divided into 3
tetrahedrons
Polyhedron projection algorithms are used in
order to economically volume-render the
unstructured grid
Wedges into Three Tetrahedra
n2
n1
n3
CORIE and Bathymetry
The unstructured grid “stops” on the
bottom surface of the river/ocean
The grid of triangles that represents
the bathymetry is constructed using
the vertices at greatest depth
Different colors are applied to
indicate differing depth
Columbia River Topology and Bathymetry
Columbia River Insertion into Continental
Shelf
Representation of Salinity Fields
Rely on volume rendering
Allows for study of fine detail between high and low
salinity regions
Blue = high salinity
Yellow = low salinity
Red = interface regions
Salinity Intrusion During Flood Tide
Maximum
Gradients
of Salinity
Freshwater Plume during Ebb Tide
Velocity Field Representation
Flow vectors visualized as a set of oriented lines
Lines have the same length
Colors represent vector magnitude
Orientation represents flow direction
Easier to view a small subset of the vector field
Velocity Field during Ebb Tide
Use of the Model in Simulation
Historical data is used extensively in simulation
Simulation can be “validated” by the use of
drifters
Drifters are floating data and position collection
devices
Helpful to validate simulation behaviors
Observed & Simulated Trajectories of a Drifter
Simulated Drifter
Real Drifter
Performance
2.53GHz Pentium 4
Nvidia GeForce 4200
Render 700k to 800k tetrahedrons per second
Typical grids include around 6 million cells
Sims run in weekly increments
One time step is rendered and dumped for each
15 minutes of simulation
Future Work
Holy Grail: Real-time frame rates
Make the visualization machine independent
Reduce the complexity of visualization output
production
Explore advanced vector visualization techniques
Questions
Could this type of visualization be used for
atmospheric data?
Is 3D volumetric rendering worth the cost?
Why is volumetric rendering the right approach
here?