The Autonomous Ocean Sampling Network
The oceans are truly a three-dimensional environment. Yet even the most modern satellites give essentially a two-dimensional view, showing only what's happening within a few meters of the ocean surface. To fully understand this constantly changing environment, oceanographers must use large numbers of instruments or find ways to move their instruments through the water.
The Autonomous Ocean Sampling Network (AOSN) experiments were designed to test a variety of cutting-edge methods for monitoring large areas of the coastal ocean. These experiments have been funded by the U.S. Office of Naval Research since the early 1990s. In 2000, 2003, and 2006, MBARI hosted three complex AOSN experiments in Monterey Bay.
During the first AOSN experiments, many of the sampling tools were in very early stages of development. MBARI Chief Technologist James Bellingham remembers, "In some of the early field experiments, the big achievement was that we took the vehicle to the field and brought it back in one piece."
These early field projects culminated in the MUSE experiment—the first AOSN experiment hosted by MBARI. In the fall of 2000, this ambitious project brought a variety of ships, drifters, and underwater robots to Monterey Bay. Bellingham comments, "MUSE showed us the usefulness of AUVs and gliders, which gave us the ability to be in multiple places at the same time." AUVs (autonomous underwater vehicles) are underwater robots that are programmed at the surface then travel through the water, collecting data as they go.
The 2003 AOSN II experiment
As successful as the MUSE experiment was, much of the project involved getting the various robots and sensors to work simultaneously. It was not until the AOSN II experiment in August 2003 that coordinating the robots and applying the data for scientific purposes came to the forefront.
During the AOSN II field experiment, physical oceanographers, marine biologists, computer-modeling experts, and AUV designers converged on Monterey Bay. These researchers came from more than a dozen prestigious research institutions to study the bay using cutting-edge computer models and high-tech observational tools, including satellites, airplanes, ships, drifters, buoys, autonomous underwater vehicles, and a small fleet of undersea gliders.
The researchers used these diverse research "platforms" to observe and predict the upwelling of cold, nutrient-rich water that occurs near Monterey Bay during the summer months. Such sporadic upwelling events create blooms of marine plankton that support the abundant fisheries and other wildlife in and around the bay.
The 2003 AOSN experiment demonstrated that AUVs and gliders were very useful for tracking plumes of cold water that flow along the coast during upwelling events. The AUVs allowed scientists to gather continuous measurements over large areas and long time periods, even during rough weather. They could also be programmed to “patrol” a section of ocean, waiting for interesting events (such as upwelling) to occur. Once they detected an event, the AUVs were reprogrammed to change their paths and travel to the best location from which to documenting that event. As Bellingham put it, "AUVs allow us to be in the right place at the right time." This process, called "adaptive sampling", has been an important element of many AOSN experiments.
The 2003 AOSN experiment also allowed researchers to test computer models for predicting ocean "weather." Like swirling clouds or eddies in a stream, upwelling currents form complex, rapidly changing patterns. These three-dimensional patterns are difficult to measure, let alone predict. Using powerful supercomputers, researchers ran two separate computer models in real time during the AOSN 2003 field experiment. These models attempted to predict the course of ocean currents as well as concentrations of nutrients and populations of marine algae.
After each day of the month-long AOSN II experiment, the researchers gathered in a "control room" at MBARI to compare data from the AUVs and other sources with the models’ predictions. They also compared the results of the two different models to see which was more accurate. Predictions from the models were then used to decide where to focus sampling efforts on the following day. "AOSN II was where everything came together—the AUVs, the computer models, and the use of real-time data," Bellingham said.The 2006 ASAP experiment
In 2006, even more AUVs returned to Monterey Bay for yet another AOSN field experiment. The 2006 AOSN experiment focused on adaptive sampling and prediction. This "ASAP" project was part of an even larger collaborative effort known as the "Monterey Bay 2006 Field Experiment."
During the ASAP experiment, the AUVs followed much more complex paths than in 2003. Many AUVs were controlled and coordinated by a central computer at Princeton University. The data from the AUVs and other sensor platforms were used to check and increase the accuracy of three different computer models of ocean circulation, running on supercomputers at different institutions across the country.
"One of the greatest achievements of the 2006 experiments turned out to be the discovery that we could do collaborative science without having all the researchers in one room," Bellingham said. Many of the researchers for the 2006 experiment didn't even travel to Monterey Bay, but stayed at their home institutions. Web-based software developed at MBARI and elsewhere allowed researchers to control their AUVS and monitor and discuss the experiment using the Internet.
All of the observational data and model outputs from the 2006 experiment were made available in near-real time on a central data server at MBARI. Plots of observational data and model forecasts were updated every morning on a public web site. This web site also included a "virtual control room" where researchers could discuss the previous day's findings and plan the coming day's research strategy.
The 2006 experiment also showed the importance of developing robust and accessible computer networks for sharing and storing data. As Bellingham points out, "When you get all those observation platforms out there, the cyber-infrastucture becomes the 'glue' that holds the whole thing together." Such collaborative data networks will also be necessary for controlling future large-scale ocean observatories and even larger fleets of mobile undersea vehicles.
Thanks to dedicated funding from the Office of Naval Research and hosting by MBARI, the AOSN experiments have allowed oceanographers to test new ways of studying the ocean. They have not only helped researchers program their undersea robots to work together more efficiently, but have also demonstrated new ways for the researchers themselves to work together.
MBARI contributors to the Autonomous Ocean Sampling Network: Jim Bellingham, Francisco Chavez, Steve Haddock, Thomas Hoover, Michael Godin, Hans Thomas, Duane Thompson
- AOSN web site
- AOSN Data Access web site
- MB06 experiment web site
- August 23, 2006 news release on the MB 2006 experiment
- ASAP experiment web site at Princeton University
- August 12, 2003 news release on 2003 AOSN experiment
- Article on AOSN 2003 Experiment in MBARI's 2003 Annual Report (PDF file)