Hailed by The New York Times as the greatest revolution in ocean observing, one organisation has sent out 4,000 robots to traverse up and down the ocean’s upper 2 km, measuring its heat content and temperature. This moniker is well earned, as the ocean remains a mystery despite the important role it plays in food production, climate regulation and tourism.

Having sent out their first robots in 2008, equipped with nitrogen and oxygen sensors, the team at Monterey Bay Aquarium Research Institute measures the ocean’s basic metabolism. “If you feel unwell and go to the hospital, the doctor does not immediately throw you in the MRI scanner. They take basic vital signs first, and that’s what we’re doing with the ocean,” explained Ken Johnson, senior scientist at the Monterey Bay Aquarium Research Institute. Using gas sensors, they measure oxygen and carbon dioxide, which allows them to measure the metabolism, growth and death of plants and animals in the ocean.

As in humans, monitoring the ocean’s vital signs alerts scientists to when processes become disordered, which can lead to devastating impacts. For instance, higher temperatures strip the ocean of oxygen, leading to changes in migratory patterns and a decline in species diversity, making the ecosystem less resilient. Oxygen-poor environments are more attractive to greenhouse gas-producing bacteria, compounding the problem of rising temperatures.

Plant growth is at risk as ocean temperatures rise. Plants are reliant on nutrients from the deepest waters of the ocean, which are transported upwards. As the upper ocean gets warmer, it becomes harder for cold water to be lifted up and the flow of nutrients changes. “Our sensors are our version of an MRI to see the problems in the ocean,” said Mr Johnson.

Historically, vast areas of the ocean could never be sampled, while others were sampled every decade. “If you are really interested in ocean health, once every ten years is not enough. Oceanographers tend to sample only in summer, so the metabolism of the ocean is heavily biased to summer; no one wants to go in winter,” Mr Johnson said. The ability to sample regularly throughout the year now means that oceanographers can measure how the rate of ocean acidification varies based on location.

In the 1990s, satellite remote sensing technologies became integral for monitoring oceanographic changes. Now, technologies include AI-based systems, sail drones and wave gliders. Mr Johnson’s robots use profiling floats that operate by changing their buoyancy, allowing them to descend to 1 km, drifting for five to ten days, and then diving as deep as 2 km before returning to the surface.

Extreme marine temperatures, known as marine heatwaves, are now occurring in different areas of the ocean. These cause harm to fish, seabirds, coral and coastal ecosystems. Marine heatwaves doubled in frequency between 1982 and 2016. And in August 2023 the average daily global sea surface temperature surpassed the 2016 record, peaking at 20.96°C.

This picture contrasts with the fact that sea ice in Antarctica reached record highs in 2013 and 2014, followed by a sharp decline from 2016 to 2020. “The ocean is more variable than people realise, the signals of health can go up and down, the annual cycles change,” explained Mr Johnson. “We don’t have the data to say how variable it is. We have almost a 30-year record of flying satellites that measure the colour of the ocean, which is chlorophyll, but that’s just on the surface. Most chlorophyll is deep down where the nutrients are; we don’t see that on satellite, so we don’t really know what is happening.”

Ocean colour sensors, measuring how the sunlight reflects off the water, were first launched in 1978, but there is a gap in the records from 1986 to 1996. After that point, Japan and the US launched sensors that have been operational since. They function best in the open ocean; coastal waters present a more complex picture due to suspended particles from river run-off and terrestrial activities.

The ocean soaks up 30-40% of all gases produced by human activities, making it the world’s main carbon dioxide sink. The ocean’s effectiveness at removing greenhouse gases has prompted hope for slowing climate change by accelerating the natural process of ocean carbon dioxide uptake.

How carbon dioxide (CO₂) affects the ocean

Research teams now are exploring methods for removing carbon dioxide from the ocean and reversing acidification. Currently, this is done by converting dissolved bicarbonate to molecules of carbon dioxide, which are then removed under vacuum. This conversion involves applying a voltage across a stack of membranes, a process that temporarily acidifies the water. A team of MIT researchers has devised a membrane-free method, which reverses this acidification, making the water alkaline before it is released back into the ocean. It is hoped that this process could slowly start to reduce ocean acidification locally. The carbon dioxide removed can be used as feedstock, although the majority is expected to be placed in underground geologic storage areas.

Keeping records can be helpful for reshaping coastal ecosystems. “The coastal ocean has recovered in many places, this is generally true for systems that we pay attention to as humans: when we are fully aware and we see things, we fix them, but if we don’t pay attention to it, we not to even know there is a problem or make the effort to fix it,” said Mr Johnson.

An example is seen in Abu Dhabi, where they are safeguarding the world’s second-largest dugong population by restoring several thousand hectares of coastal areas. They realised that dugongs, the dolphin-like creatures that may have been mistaken for mermaids in ancient times, thrive in conditions that also allow corals, turtles and several other species to thrive.

Similarly, the Indian government led an initiative to invest in waste management, water treatment and engaging communities in conserving the ecosystem along the banks of the Ganges River. Over 370 km of riverbank has been restored to date with US$4.25bn invested so far.

Country-wide measures

The data collected by Mr Johnson’s sensors have several applications, including fisheries management and oxygen distribution patterns. But there is a new field of environmental DNA (eDNA), which could have far-reaching positive implications. “Everyone is sloughing off DNA and that’s the field that’s going forward. At the moment it tells the health of a system, but it can’t yet say what the system is,” he explained. eDNA technology is also helpful for detecting new species, as it measures the cellular material shed by organisms.

Other advances are being made with acoustics on sensors, although this can be a national security problem for some nations due to Navy restrictions. Nevertheless, this is an exciting time for ocean monitoring technology, as more sophisticated and expensive solutions are being deployed.

Fish play an important role in global food security and nutrition, and sustainable systems must be implemented to include dealing with overfishing and unequal access to resources. Aquatic food production is forecast to increase by 15% by 2030, but before this increase can happen, the current state of fish stocks needs to be understood. Data on global fish stocks are, however, incomplete, especially in regions like Asia, Africa and Latin America.

In order to determine sustainability, certain benchmarks are assessed, such as unfished biomass and the amount that would be a sustainable fish biomass. In a context of limited data, this creates a situation whereby a big belt of areas across the tropics, where a lot of people rely on food from coral reefs, do not have good estimates on the state of their fisheries.

Fish stock assessments are carried out to manage fisheries, such as setting catch limits that are sustainable for specific fish populations. One way of doing this is the ABC method, which looks at the abundance of fish, biological factors such as age and population growth rate, and catch data. However, this method is expensive and labour intensive. The cost is a barrier for less developed countries, making it mostly restricted to wealthy countries.

Another method, which has the most abundant data, involves measuring the catch on a beach or landing site. Trying to estimate the global status of fisheries is therefore problematic, because the best data come from that gathered empirically by research organisations in wealthier countries.

Now, a new AI method is allowing researchers to estimate coastal fish stocks and will hopefully allow countries that have poor fisheries data to determine the sustainability of their stocks. The approach, spearheaded by Dr Tim McClanahan, director of Marine Science at Wildlife Conservation Society, estimated fish stocks with 85% accuracy in a pilot region in the Western Indian Ocean.

Dr McClanahan carried out the work over a number of years, and began by counting fish in protected areas where there was no fishing, allowing him to estimate recovery rates. He used over 1,000 transects to count fish stocks and collected satellite data such as water temperature, distance from shore, ocean productivity, depth of the water and existing fisheries management.

He trained the model with 70% of the information and tested it against the remaining 30%. “When we did that, we came to the shocking recognition that it had 85% predictive ability,” said Dr McClanahan. “Nobody had probably ever realised that you could predict fish populations from basically reflections off the surface of the sea, knowing the distance and the management system it’s in. This is extremely powerful,” he carried on enthusiastically.

The beauty of the method is that anyone, anywhere in the world, can use the tool to determine the health of fish stocks, the time they need for recovery, and even how much money is lost by fisheries mismanagement. In the pilot study, the amount recouped after using the tool was US$50m-150m annually. Dr McClanahan expects that the method will chart the way for more applications of AI in fisheries research and is already using his model to predict changes in global climate change.

Normally, the level of dissolved oxygen in the ocean ranges from 7-8 mg/L. If the concentration falls below 4 mg/L, fish and other species will migrate out of the area or avoid it. Areas of low oxygen are known as dead zones and areas where oxygen levels are at or below 2 mg/L are referred to as hypoxic.

“There are multiple events that lead to ocean hypoxia, the most common is excessive nutrient flow from land,” explained Dr Yingjie Li, human-environment scientist at Michigan State University, who monitors global coastal dead zones. “Another reason is definitely climate change; as the temperature increases, oxygen will escape from the water. There are many other reasons but these two are the most important ones,” he continued. The growing number and size of these hypoxic zones are an increasing threat to the health of the marine ecosystem, which in turn affects human well-being.

Dr Li uses satellite imagery to measure oxygen levels in the ocean, a more convenient method than the common and labour-intensive practice of taking water samples and testing oxygen levels in the lab. Their images of algal bloom on the ocean surface, which changes based on oxygen content, can be taken daily with high temporal and spatial resolution. Images of algal changes can then be used to build a map of hypoxic zones, which can be of use to policymakers, allowing for targeted actions. Satellite image data were obtained from NASA’s Earth Observing System, which is freely available to scientists.

To map hypoxic zones, the machine learning models needed to be trained. “Satellite imagery is great, but ground-level data is also critical. We collaborated with marine scientists from Louisiana State University who collected several water samples. We also used available water samples from NOAA and these samples were all critical for training our machine learning models,” explained Dr Li.

Now, the team plans to release their tool as a platform, allowing other scientists to use a similar approach in other regions. They expect it to be of use to more than just policymakers. “This has great implications globally,” explained Jianguo Liu, also from Michigan State University. “It can inform farmers about the fertilisers they use and food they produce.

Fertilisers may come from places far away, then farmers use them to produce food and extra fertiliser makes its way to the ocean. It’s not just the farmers responsible for the products, we are all responsible. Our tool will play a hugely important role to inform the public and policymakers to translate knowledge into action,” he said optimistically.

In the fight against climate change, experts believe that reducing emissions alone is not enough, and we now need to actively sequester carbon dioxide to effectively reach negative carbon emissions. The result is that carbon dioxide removal technologies have become a burgeoning industry, which needs to reach US$1trn annually to be effective.

One way to sequester carbon dioxide from the atmosphere is through biological processes; with photosynthesis, plants absorb carbon dioxide and convert it to glucose, releasing oxygen back into the air. In this way, they capture carbon dioxide and store it organically.

There has been a reliance on terrestrial ecosystems such as forests for carbon storage, but coastal ecosystems such as mangrove forests, seagrass meadows and kelp forests also store carbon. Kelp or seaweed may be one of the answers, as they are believed to sequester nearly 200 million tonnes of carbon dioxide annually—equivalent to what New York State emits in a year. By their simple existence, their physiological mechanisms take pollutants out of the atmosphere and use them for food.

“The momentum has been growing for the past six to seven years,” explained Schery Umanzor, an assistant professor in the College of Fisheries and Ocean Sciences at the University of Alaska Fairbanks. “The initial narrative was more about food production as a protein source, then the narrative of biofuels came into play, and nowadays it is towards [the] production of bioplastics and the capacity to remove carbon from the air and take it into the deep ocean.”

In her three years of working in Alaska alongside kelp farmers, Dr Umanzor realised how much carbon capture varies across geographical regions and species. “I wasn’t expecting to see this much difference in the dynamic of carbon and nitrogen uptake,” she said, referring to an exploratory experiment on the ribbon and sugar kelp species.

These differences may have implications for whether kelp is a sink or source of carbon, as this is likely to vary with time and across species. For this reason, Dr Umanzor is cautious about making claims that kelp will mitigate ocean acidification and anything related to carbon. “It will vary based on when kelp is harvested and by region,” she explained.

Rather than focus on kelp’s impacts on carbon, Dr Umanzor is focusing on its ability to remove nitrogen from the ocean, which it does effectively. Nitrogen is a nutrient that favours marine productivity, but it is not always beneficial for marine life. It is the cause of floating brown seaweed known as Sargassum, which can smother corals and seagrasses. It also releases hydrogen sulphide, which can cause issues for marine ecology and tourism. Kelp farms can offset the large concentrations of nitrogen that contribute to this, despite the concept of carbon renewal being more developed than nitrogen removal.

Money is pouring into the kelp industry, described as Alaska’s newest gold rush, while Amazon intends to fund the world’s first commercial-scale seaweed farm off the coast of the Netherlands, costing €1.5m. Meanwhile, the European Maritime Fisheries Fund has provided €2.1m in funding to a project run by a Scotland-based company aiming to develop a sustainable EU seaweed industry.