Robotic floats — more than 100 of them, in fact— are currently in the ocean and tracking an important measurement for climate models: oxygen and carbon dioxide.

The gas-cylinder-sized devices being developed at Monterey Bay Aquarium Research Institute  monitor a biological activity known as primary productivity, or the rate at which phytoplankton convert carbon dioxide into organic matter through photosynthesis. Primary productivity is the first step in building a marine ecosystem, says MBARI Senior Scientist Ken Johnson.

"The phytoplankton use their produced organic matter to grow and divide into a larger population," Johnson told Tech Briefs. "Zooplankton then eat the phytoplankton, small fish eat the zooplankton, large fish eat the small fish, and on up to whales."

Johnson and his team want to measure primary production by analyzing the daily cycle of oxygen in the surface waters. Specifically, the researchers use the floats to measure the increase in oxygen during daylight (due to photosynthesis and primary productivity), and then the decrease at night (when respiration consumes the oxygen).

"Tracking carbon is essential if you want to understand fisheries, or the role of the ocean is absorbing CO2 from the atmosphere," said Johnson. "About 25% of the fossil fuel CO2 emitted each year to the atmosphere goes into the ocean and much of that flux is prompted by CO2 taken up during primary productivity."

In a short Q&A with Tech Briefs below, Ken Johnson explains more about how the robotic floats work, where in the ocean they'll be sent, and what kinds of clues they provide about ocean health.

Tech Briefs: What do these robotic floats look like?

Ken Johnson: The robot – a profiling float – is pretty simple, which is also a key to their success. About the size of a gas cylinder (see the above image), they are free drifting and control only their depth by pumping a small amount of vegetable oil from an internal reservoir to a rubber bladder outside the pressure cylinder. Like a helium balloon, inflate the bladder and it goes up. Deflate the bladder by moving the oil back into the pressure housing, like a balloon sucking the helium back into a gas cylinder, and it goes down. The difference is, because density is so uniform in the ocean, the float needs to move only about 1 cup of oil (~240 ml) to move 2 kilometers vertically.

Tech Briefs: What is each robotic float collecting, and how frequently?

Ken Johnson: Our floats have enough batteries to do this about 250 times. If the float only comes up every 10 days, it can live for 2500 days, or almost 7 years. They park at 1000 meters for most of the 10 days where currents are weak, and the floats don’t move much. They then descend to 2000 meters, turn on their sensors, and come to the surface, making measurements along the way.

At the surface they get a GPS fix and send the data home by the Iridium satellite communication system. They don’t need any maintenance during this time so operating costs are small – mostly just the cost of processing and distributing data. All the data is freely available within 24 hours of each vertical profile. Floats could cycle faster than 10 days, but then they would run out of power too fast, so 10 days is a compromise between float life and getting more data.

Tech Briefs: How many floats are in the ocean, and where are they being deployed?

Ken Johnson: There are about 4000 of these floats in the ocean right now (see a map at https://argo.ucsd.edu/ ), but most of those only measure temperature and salinity; these are core-Argo floats. Only about 400 of the 4000 are measuring chemical and biological properties, or BGC-Argo floats.

Core and BGC-Argo floats are everywhere, from Antarctica to about 70 degrees north. For a couple of technical reasons, there aren’t many in the Arctic. (They don’t like permanent ice cover – they need to get to the surface to phone home).

The bad news for us is that only about 170 of the 400 BGC-Argo floats have the right timing parameters on their 10-day cycle to be used to measure primary productivity. Our goal is to get 1000 floats in the ocean with the right timing to really monitor primary productivity with seasonal resolution around the globe.

We’ve just been funded to acquire 500 floats through a program called G0-BGC: Global Ocean Biogeochemistry Array. They will go everywhere in the ocean, in time. Our first ones are in the Atlantic. We’ve also been funded over the past 8 years to put 200 of these floats in the Southern Ocean, which formed the basis for much of our study.

Besides oxygen, our floats also measure temperature, salinity, pressure, nitrate (a key fertilizer for phytoplankton), pH, chlorophyll, optical backscatter (plankton abundance) and some also measure downwelling light (water clarity).

Tech Briefs: What are the strengths and weaknesses of a robotic method of measurement vs. alternative ways of measurement?

Ken Johnson: The strength is that the robots run for years, all year round making a direct measurement of primary productivity. With even 170 floats, we can sample much of the ocean.

Most direct, primary-productivity measurements have been made by scientists on ships. They have to collect a water sample, add some radioactive carbon (carbon 14) to the sample, and then incubate the sample for 1 day, filter the sample, and count the carbon 14 that has been incorporated in the cells. That’s great, but most research cruises are only about 30 days long, so one never samples over a whole year. Very few research cruises go in the winter or even spring when large plankton blooms occur. That misses much of the seasonal cycle. And there are really very few of these cruises in the open ocean. So ships just don’t generate much of the required data.

Tech Briefs: What about satellites?

Ken Johnson: Satellites give great high-resolution images of productivity, but the results are model-based, not direct measurements, and they don’t see very deep in the upper ocean, missing much of the signal. The satellite measures the color of the ocean; one computes chlorophyll from the color, and then runs a model that has been calibrated by the shipboard data and observed chlorophyll. If the plankton physiology is changing (perhaps in a warmer world), then the models can be wrong and trends in estimated primary productivity might not be real.

So, bottom line, we get a direct measurement of primary productivity, and we can do this down below the surface. But the best answer (we emphasize in the paper ) is to use the float-based measurements to calibrate the satellite observations in quasi-real time so the satellite productivity models are verified, and then use the high-resolution satellite data (way higher spatial resolution than one would get from 1000 floats) to get the best of both worlds.

A robotic float at the surface. (Image Credit: MBARI)

Tech Briefs: Are the robotic floats’ analysis one piece of the overall data picture? If the floats’ measurements are contributing to environmental climate models, are there other measurements also supplementing that data?

Ken Johnson: For sure floats are one piece of a multi-platform data set. The float sensors aren’t perfect, and we really depend on high-quality ship data to help calibrate our sensors and the ships are essential to deploy floats. The synergism of floats and satellite observations will also result in much better assessments of primary productivity.

The real contribution of floats is that they can stay at sea for years, sampling through all kinds of weather, including under clouds where satellites can’t see. But they only measure a few chemical and biological parameters. We need ship-based studies to amplify our understanding of what floats observe. Those float values set the stage for scientists to investigate ocean change using much more complex observations from ships, but which will span only a short period of time. We need satellites to increase our spatial resolution.

Finally, environmental models really only want to assimilate quasi-continuous data streams. Those come from floats and from satellites and not so much from ships. So to the extent that models will assimilate chemistry and biology data, it’s going to be coming from floats and satellites. And it will be important if you want to forecast weather more than a few days; you have to get the ocean correct since it is changing on a period of weeks to months. And to get the ocean right, adding chemical properties can really improve their accuracy. Same for fishery models.

Tech Briefs: What’s next? Where else do you want to deploy these floats?

Ken Johnson: Next up for us is expanding the US part of this global array to 500 floats through GO-BGC. We have cruises leaving shortly going into the eastern tropical Pacific and equator region, which can be hard to get to. But really, we want an evenly distributed array around the world ocean. So we’re building floats and sensors as fast as we can.

What do you think? Share your questions and comments below.