Feature Articles4D Visualization and Analysis Of Seafloor Vents and Plumes
By Maurice Doucet
Chief Systems Architect
Chief Technology Officer
IVS 3D Inc.
Portsmouth, New Hampshire
Multibeam sonar technology has advanced to the stage of providing high-resolution acoustic return information not only from the seabed, but also from the intervening water column. This ability will potentially allow multibeam sonars to address a number of critical ocean problems, including the direct mapping of fish and marine mammals, the location of water column targets and a wide range of physical oceanographic processes.
To exploit water column data, an efficient means of reading, processing and analyzing the data is required. Marine researchers currently have a limited view of water column data in real time and a limited capacity to store them, replay them or run further types of analysis. The water column data need to be integrated with other sensor assets such as bathymetry, bottom backscatter, sub-bottom, seafloor characterizations and other datasets so that a complete picture of the marine environment under analysis can be realized.
IVS 3D has developed a format-independent processing tool, FMMidwater, for extracting features of interest from the water column files of multibeam and single-beam sonars, which is available in version 7.2 of IVS 3D’s Fledermaus suite. FMMidwater’s graphical user interface (GUI) tools allow rapid feature isolation and extraction to Fledermaus time-aware objects that can be used for visualization and analysis.
Exploitation of water column data has inherent challenges, not the least of which is the lack of sonar format standardization in the marine industry. While there are some common data formats for bathymetry and backscatter, notably the Generic Sensor Format (GSF), the authors are not aware of any existing formats for water column data.
Each sonar manufacturer has a specific data-logging file format that contains discrete packet types representing data from a specific component within the system, such as navigation, attitude, backscatter, multibeam and water column data. Navigation and attitude information is typically logged in separate packet types and occasionally logged to a completely separate file.
This results in further fragmentation of an instantaneous snapshot of the water column. In order to compute where each sample of the water column from a single ping exists in time and space, one needs to integrate information from multiple packets and, sometimes, multiple files.
Another challenge is the volume of data produced by some shallow-water high-frequency systems. Even short survey lines can yield gigabytes of information that needs to be processed, requiring researchers to have both an efficient means of storage as well as efficient tools to mine the survey data.
Generic Water Column Format
To address the primary challenges that exist in the exploitation of water column data, IVS 3D designed the Generic Water Column (GWC) format with help from the Center for Coastal & Ocean Mapping Joint Hydrographic Center (CCOM) and Kongsberg Maritime (Kongsberg, Norway). The GWC format can be considered analogous to the GSF format.
The GCW format was designed as a unified way of storing water column data in a compressed, subsampled and integrated manner. During the source conversion process, the water column packets are reintegrated along with proper time-based navigation and attitude such that the visualization environment will have access to all of the relevant data of any particular ping. The time-series data for each beam can also be subsampled in order to reduce the overall footprint of the processed file. The GWC is intended to be an open format and is available for download and comment on the IVS 3D website.
Feature Extraction and Visualization
Although the raw acquisition rates of water column data collected from multibeam sonars can be very high, there is generally much less information in the midwater column. An important aspect of any tool used to examine this type of data is the ability to rapidly review and extract features of interest.
In the IVS 3D FMMidwater tool, this is done by providing multiple ways to view and filter the data prior to displaying them in a 4D environment. The most basic view is a typical “swath view” display of the water column data. Using the timeline control within the tool, one can move through the data serially or jump to any point in time within the file and see the swath display for that ping. Another option for data review is by way of a single-beam along-track plot. Here the data through the water column are viewed as if an observer were perpendicular to the track line. The final option for data review is using the “stacked” view. This view allows the operator to see the same type of single-beam along-track plot with the exception that the sample values are filtered to display the maximum value from all beams at a given ping and range.
Before transferring features to the Fledermaus visualization environment, it is often useful to use some threshold filtering techniques to further reduce the amount of undesired data within the water column. Using FMMidwater, the operator is able to filter out specific beams, sections of the beam range and specific signal levels.
As an example, if one is passing over a deepwater plume, some of the outer beams may contain no relevant information. Also, even though the water column time series may cover thousands of meters, the plume may exist in only the final few hundred meters. In that case, the entire near range of the sonar time series can be eliminated. Finally, water column targets typically exist in a range of energy levels as determined by their acoustic properties. Histogram filtering controls can be used to eliminate low-level noise returns and provide for better color contrast over targets of interest. All of these filtering techniques are directly accessible through the primary GUI interface of FMMidwater.
Once the water column data have been properly filtered and the operator has focused on a particular area of interest, this information can be transferred into the Fledermaus 4D environment. Four primary visualization metaphors were developed to provide the most effective means of reviewing water column features. These include the beam fan, beam line, point cloud and volume objects.
The beam fan object is an across-track view or swath of the water column. This object is temporally aware and is rendered in the 4D environment dynamically from the GWC file. The beam fan represents color-coded time-series data for each beam of the sonar. It is color-coded according to the signal level at a given temporal and spatial point and can be changed to display amplitude, power, volume scattering or target strength.
The beam line object is an along-track curtain similar in nature to a seismic curtain in the Fledermaus environment. The primary difference is that the curtain is aligned to the selected beam angle. The beam line is another time-aware object that allows visualization of the beam from the start time to another given time. One can also set up a time window that will clip the displayed data to a desired time interval between two times.
The point cloud object contains 3D points that represent the geolocated bore-sight positions of a particular time-series sample. The points are time-aware and can be viewed using a time window similar to the beam line.
The volume object is an isosurface representation of the filtered water column target information. During export, the selected water column data is ray-traced into a voxel “brick” of data that can be used to generate isosurfaces at user-specified levels. This visual metaphor can be used in a nested mode to show the internal structure of a complex object such as a seafloor geothermal vent.
All of these capabilities are integrated into FMMidwater as an easy-to-use graphical tool that enables an operator to extract the maximum amount of useful water column data in the minimum amount of time. The operator can easily see the water column display along with a map display that shows the current location of the view according to the time position on the timeline control. One can easily switch between swath, beam and stacked views; add or remove survey lines to the project; and adjust the visualization based on the described threshold filtering techniques. The processed data can then be transferred into Fledermaus for 4D analysis and combined with the bathymetry and backscatter data using normal system tools.
A prototype version of FMMidwater was used during a cruise of the NOAA Ship Okeanos Explorer. While on the cruise to test the new Kongsberg Maritime (Kongsberg, Norway) EM302 multibeam sonar in May 2009, researchers discovered a 1,400-meter-high plume rising from the seafloor. The feature was noticed in the online display of the water column data of the sonar, and further analyzed by the prototype FMMidwater tool.
The ship returned to the area in July, verified that the plume was still active and detected a number of other plumes ranging in height from 700 to 1,400 meters in a 15-kilometer area around the original discovery.
Dr. James V. Gardner of CCOM and Mashkoor Malik of NOAA participated on the cruise and provided details of the discovery in the journal Eos, Transactions, American Geophysical Union. The discovery of this plume is just one example of the emerging use of these types of multibeam water column data.
FMMidwater was also used operationally this summer during the Deepwater Horizon incident on several NOAA and commercial vessels in the Gulf of Mexico to aid in the detection and mapping of petroleum in the water column. It also helped discover a number of previously unknown seeps in and around the target area of the investigation. Integration of the output of FMMidwater with existing assets such as water sample data, bathymetry and backscatter into the 4D temporal visualization capability of Fledermaus helped provide the involved scientists with a true big picture of a dynamic marine environment.
The capabilities of multibeam sonars have been extended in recent years to enable them to map the water column along with the seafloor. The development of suitable analysis tools has lagged behind sonar functionality and limited the use of sonars by scientists investigating a number of critical ocean issues.
FMMidwater seeks to fill this gap by providing scientists with a means of not only managing the large volume of data produced by these sonars, but also an efficient means of exploiting the data for scientific knowledge. In addition, the development of the new visualization metaphors provides a unique analysis environment for the integration of water column data with existing multibeam data (bathymetry and backscatter) and other oceanographic and mapping data.
The authors gratefully acknowledge the support of the development by a grant from the New Hampshire Innovation Research Center. Thanks also to Jim Gardner at CCOM, Mashkoor Malik of NOAA and Dr. Jens Greinert of the Royal Netherlands Institute for Sea Research.
For a full list of references, please contact Maurice Doucet at firstname.lastname@example.org
Maurice Doucet, chief systems architect for IVS 3D, worked on a joint project with the Center for Coastal & Ocean Mapping Joint Hydrographic Center to address the data extraction, analysis and visualization of water column data, the result of which was FMMidwater. He develops new software technologies and architectures for IVS 3D.
Mark Paton is responsible for the overall Fledermaus product design and development, as well as leading ongoing research and development activities. Prior to establishing IVS 3D, he performed his graduate studies at the University of New Brunswick, where he developed the core functionality of the Fledermaus software suite.