Automating a 20-year survey of deep-sea animals
November 14, 2016
A new MBARI robot that can traverse the deep sea, collecting data all on its own, is providing a boost to one of MBARI’s longest running research projects – the midwater-time series project. The robot carries a new video system called “i2MAP” (Investigations of Imaging for Midwater Autonomous Platforms). This new tool will add to the scientific observations MBARI researchers have been making with remotely operated vehicles (ROVs) for nearly two decades.
MBARI’s midwater time-series project consists of video surveys of animals in the upper kilometer of the water column in Monterey Bay. Trained biologists review the surveys, identifying and documenting every animal seen. The resulting data can be analyzed to understand what species live in the midwater region, and how the populations change over time. In addition to species counts, MBARI researchers are also collecting data such as oxygen content, temperature, and salinity.
We know very little about the life in the deep sea. MBARI’s midwater surveys provide baseline information on water chemistry and variety of sea life, and show how these things change over time. Many animals live in the midwater and form important links in marine food webs. The trends observed from MBARI’s time series will help scientists at MBARI understand how midwater food webs change in response to climate change.
“We see some stuff that changes on the short term, seasonal changes. We also see how communities build up and how a very common animal suddenly disappears for a year or two,” said Senior Research Specialist Kim Reisenbichler, who has been working with Senior Scientist Bruce Robison on this project.
Until about 1990, midwater research was conducted by pulling nets through the water. The nets gave scientists a way to count fish, squids, and hard-bodied animals, but underestimated jellies. “It was like taking Jell-O through a strainer,” Reisenbichler said.
Since then, Robison and Reisenbichler have used ROVs to conduct midwater surveys. During each survey the large ROV, equipped with bright lights and video cameras, is deployed from research vessel Rachel Carson and dives as deep as 1,000 meters. Aboard the Rachel Carson, in a small dark room, two ROV pilots and two or three scientists watch large monitors. The monitors show video from the ROV’s cameras. The scientists look for signs of life, from small jellies to large fish. Some of the animals they see are very common, while others have never been seen before.
The ROV surveys give a better count of the jellies than do net trawls, but some, agile species such as some fish, squid, and crustaceans, may swim away when they sense the slow, brightly illuminated, and noisy ROV approaching.
The team at MBARI began working on designing a new survey method with the hope of creating an autonomous underwater vehicle (AUV) that would require less manpower, run quieter, and take higher quality footage than the ROV—a task more challenging than it sounds.
An autonomous underwater vehicle operates without human assistance, programmed to independently steer its way through the water column. In creating an AUV for video surveys, one challenge MBARI engineers faced was that the lights aboard the AUV take a lot of power. Although the ROV has ample power provided through its tether to the Rachel Carson, the AUV would be on its own, with all of its power supplied by a large battery.
Most importantly, the new AUV had to collect data that was compatible with ROV data collected over the past 20 years, by producing equal or better quality video. It has taken the team four years of designing, testing, redesigning and retesting to complete the new video system, a tool that meets their standards.
Based on initial field trials, the data from the i2MAP AUV far surpass those from MBARI’s ROV. The ROV is much slower than the AUV and an entire research team must put in nine hours at sea for each survey. On the other hand, the AUV is much faster and, aside from deployment and recovery, operates with less human assistance. This allows the research team to utilize the ROV for tasks such as exploration and experimentation while the AUV conducts the more routine video surveys.
The ROV is noisy and slow, and puts out more light than is required for midwater surveys. However, the AUV is nearly noiseless and its light system is designed specifically for midwater surveys, reducing survey bias. Finally, the AUV is equipped with a better camera and the ability to upgrade to higher quality video formats in the future.
The midwater time series helps researchers understand the community of animals in association with the properties of the water. It can also help scientists understand changes over time. Some changes are part of seasonal cycles or sporadic events such as El Niño. Other changes are indicators of long-term climate change. By observing changes over time, scientists can better predict what might happen to deep-sea ecosystems in the future.
“Now, with the reality of climate change, and seeing that it is going to be changing for the foreseeable future, it is important to have a record of how it is affecting the water column,” said Reisenbichler.
MBARI’s midwater time series project is the only data set, or collection of records, of its kind in the world. Reisenbichler hopes that the i2MAP will change that. He is looking forward to the day when other researchers, universities, or institutions will follow MBARI’s lead and use AUVs to launch their own time-series projects, leading to comparable datasets in other regions of the world. “If I look back 10 years from now and see other people doing the things we are doing, then I’ll feel like the time we put into this project was really worthwhile,” Reisenbichler said.
For now, the new i2MAP will continue to build on MBARI’s longest running research project, adding to our understanding of life in the deep sea, and how it is changing over time.
Article and video by Teresa Carey