Project 3

Timeframe: 3/14/11 - 4/18/11

VTK 1.6.1
Qt 4.8.1 for Windows - Visual Studio 2008
Visual Studio 2008
CMake 2.8.3
C++ and C#

Pacific Ocean Tsunami Visualization

Project Proposal

Pacific Ocean Tsunami Visualization

  • For the earthquake data, I suspect to have a similar representation as seen for project 1, except this will be focused on Japan, the Alaskan coast, and the American west coast - basically the region pictured below.
  • For the NOAA data I intend to make use of the DART data to collect water column height data as well as use other observatories to get standard meteorological data (Wind, waves, air/sea temperature).
  • I intend to create a visualization that will be able to analyze the effect of the Japan earthquake on the Pacific Ocean - changes in height, water velocity, temperature. I also hope to see a rippling effect from Japan across the Alaskan monitoring stations to Hawaii and the American west coast.
    • The most obvious effects would be water velocity and height, but what about air/sea temperatures?
  • I do have a few concerns about the resolution of the data as the monitoring stations are along the coastlines and there are only a few data stations over the ocean.
    • I might make use of the Tropical Atmosphere Ocean Array (TAO) south of Hawaii since it is a grid of monitoring stations spanning the Pacific.
Pacific Monitoring Stations

Department of Energy Dataset?

  • Examining energy usage? Alternative energy?
  • Annual Electric Power Industry Database

Other Ideas

  • Earth orbiting satellites?
    • Sorting by usage, active, duration?
    • Compare with recent space shuttle orbit? International Space Station?

Environment Setup

No changes in environment setup since project 2.


The following instructions assume was extracted to the Windows desktop.

Run CMake
  • Set the source directory to the TsunamiVis source directory (C:\....\Desktop\TsunamiVis )
  • Set the binaries directory to "C:\....\Desktop\TsunamiVis\build)"
  • Click configure
  • Click "Yes" to create the build directory
  • Use "Visual Studio 9 2008" as the generator.
  • Select "Use Native Compilers", click Finish
  • Set CMAKE_INSTALL_PREFIX to "C:\TsunamiVis" - Make sure there are no spaces or other special characters in this field (like brackets).
  • Your CMake window should look like:

  • Click Configure, click Generate
  • Close cmake
  • Navigate to the build folder (C:\....\Desktop\TsunamiVis\build)
  • Double click on TsunamiVis.sln

In Visual Studio's Solution Explorer:
  • 1) Right click TsunamiVis
  • 2) Select "Set as Startup Project".

3) Press F5 to compile and run TsunamiVis.

If you want to run the binaries directly (\TsunamiVis\build\Debug\bin\TsunamiVis.exe) you will need to copy the Data folder from TsunamiVis into \TsunamiVis\build\Debug\.

The Data

Data Sources:
This project required a large amount of data processing. Three separate data parsers were created: one for each of the data sources. Each of the parsers had multiple steps such as described below.

Example NOAA data:
    2011 03 31 23 45 00 1 5263.477
Example IOC data:
    2011-04-17 20:45:00    1.303
Example USGS earthquake data:
    2010,11,12,021446.22, -6.553, 130.035,5.5,168,PDE-Q
Processed NOAA/IOC data:
Processed USGS data:
  • NOAA Parser
    • Remove 1-second interval data to save space
    • Convert coordinates like 120 W to -120 format IOC and USGS uses.
    • Convert data to comma delimited and append a timestep based on the date/time. This scale will be common for all data sources.
      • 3/1/2011 00:00:00 -> Timestep 0
      • 3/1/2011 00:15:00 -> Timestep 900
    • Merge with IOC data
    • Sort the data by the timesteps.
    • Condense the timesteps so they are sequential
      • 3/1/2011 00:00:00 -> Timestep 0
      • 3/1/2011 00:15:00 -> Timestep 1
    • Interpolate between gaps in the datapoints for smoother time transitions.
    • Generate a lookup table for application and USGS parser to convert from MM/DD/YYYY HH:MM to timestep.
    • Determine the average water height for each sensor
    • Convert to .vtk, load into application

  • IOC Parser
    • Append latitude and longitude data to each file
    • Convert data to comma delimited and append a timestep based on the date/time.
    • Merge with NOAA data

  • USGS Parser
    • Convert time from '' to 'hh,mm,ss' to match NOAA and IOC data
    • Append a timestep based on date/time
    • Use lookup table to convert date/time to timestep
    • Convert to .vtk, load into application
Once the data is loaded into the application the sea level heights are normalized so typical wave behaviors are near 0.0. This allows the sea level of various areas to be comparable. For example NOAA data comes from ocean sensors are in the range of 1k - 5k meters, while IOC data is near land based sensors at a scale of -3 to 3 meters. 

Video Overview

Using the Application

This application is designed to allow the user to simultaneously view earthquake and sea level data during March 2011. The focus of the application is to visualize the events leading up to and following the March 11, 2011 magnitude 9 earthquake  off the east coast of Japan.

Pictured above is the startup screen. The essential components of the interface is the main view, the main (right) toolbar, and the timeline (bottom) toolbar.

  • The Main View
    • The main view the primary vtk window used to view the data. The window uses the standard vtk mouse interactor:
    • Rotate scene - hold left mouse button
    • Translate scene - hold middle mouse wheel or shift-left mouse button
    • Zoom - middle mouse wheel or hold right mouse button

  • The Main Toolbar (right)
    • Sea Level Data
      • Adjust the opacity of the sea level map generated by the sensor points
      • Show/hide the sensor stations represented by colored cubes
      • Show/hide the Water Level legend
      • "Show tsunami detection instead of sea level"
        • By default this feature is unchecked and will color the sensor cubes based on the difference in the sea level at that location.
        • If toggled on, points will show if the sensor's tsunami detection algorithm was triggered. When this occurs, the cube will turn white.
        • The sensor is triggered based on the water pressure at the location, however newer models of the sensor can be remotely triggered - so this may not be the best indication that a tsunami wave is currently at that location.
        • This will take a minute to switch between modes.
      • Adjust the scale of the Water Level
    • Earthquake Data
      • Show earlier earthquakes, or only the ones detected at a particular timestep
      • Show/hide the legend
      • Show earthquake data as either points or spheres
      • Adjust the depth clipping plane
        • This is useful for cutting through the depths of earthquake data and measuring a particular depth.
        • You can also flip the normal of the clipping place
    • Animation
      • Start/stop an animation of water and/or earthquake data over time.
      • The position of the timeline slider is used as the starting point.
      • The animation will play until the end of the timeline or until stopped.
      • The animation speed can be adjusted from 1ms per timestep to 1 second.
    • Satellite  Map
      • Adjust the opacity of the terrain map
    • Interesting Timesteps
      • This will jump the view to a particular point on interest.
      • This will also adjust certain viewing settings to emphasize that event.
      • Caution: This may change settings that may not be reflected in the GUI controls.
      • Specifics of these points of interest are discussed later.

  • The Timeline Toolbar (bottom)
    • Scroll through all available timesteps ~7000 spanning March 1 - March 31, 2011
    • The current date/time is indicated above the toolbar.
    • There are two tabs:
      • "Month of March 2011" covers the entire time range.
      • "March 8, 2011 - March 13, 2011" covers a more limited range, but allows greater precision when selecting times.

Interesting Features

Visualizing the distribution of earthquakes below Japan during March 2011. In the points view, the subduction zones can be seen. In the spheres representation, we get a better look at how the different magnitude quakes wrap around each other - almost helix-like.

In looking at this area over time, you can see minor 0-4 magnitude quakes slowly forming at various depths leading up to March 11. After the 9.0 appears, a series of 6-8 magnitude quakes quickly follow.

Also by using the depth clipping plane, we find that the 9.0 magnitude quake was at a depth of about 32 kilometers.

3/11/11 6:30 UTC
3/11/11 10:55 UTC 
3/12/11 3:15 UTC 

When examining the sea level data, the tsunami waves cause a very subtle disturbance on most sensors against the normal water wave frequencies. Sensors very close to Japan and Midway Island detected  a significant change in water height. This is shown in the image sequence above by the region of dark blue moving east.

If I had more time, I would have included additional sensors on then western American coast as coastal sensors tend to pick up the tsunami wave more clearly.

A particular point of interest when using the tsunami trigger times rather than water height is the 7.3 magnitude earthquake that hit on March 9, 2011  2:00 UTC. While this was considered a typical quake in the region and caused minor changes in the water level, the pressure wave from the quake was strong enough to trigger the tsunami detectors in the area - as illustrated above.
Arthur Nishimoto,
Apr 18, 2011, 9:43 PM