One brisk day in April 2013, as he drove with colleagues along the southern coast of Patagonia, Mike Kaplan spotted a geologist’s treasure trove — an active gravel pit with freshly exposed walls. He pulled over, grabbed the backpack full of digging tools stowed in the car trunk, and walked into the large hole.

To Kaplan’s south lay the Southern Ocean, stretching toward Antarctica. Strewn around him was evidence of Earth’s most recent ice age: heaps of crushed rock and gravel released by one of the many glaciers that had once covered North and South America. Standing in the pit, Kaplan spotted what he was looking for: a layer of fine gray silt deposited by ice sheets roughly 20,000 years ago.

Photo shows the size of the gravel pit rich in sediments created by ancient glaciers and freshly dug up as part of an unrelated project.

When Mike Kaplan saw a freshly excavated gravel pit just off the road in Patagonia’s Tierra del Fuego region, he wondered if it would yield good samples for his studies of iron dust. It did.

CREDIT: COURTESY OF MIKE KAPLAN

A geologist at Columbia University in New York, Kaplan has spent over a decade collecting the sediments that make dust, and studying how that dust, launched from earth to air to sea, influences Earth’s climate, past and present. Dozens of intriguing samples have made their way home with him, stowed in his suitcase or shipped in a duct-taped cardboard box. As he scraped the dark gray sediment into a plastic bag, he felt a rush of anticipation. Given the sample’s location, he thought that it might be just what he needed to test an aspect of a controversial idea known as the iron hypothesis.

Proposed in 1990 by the late oceanographer John Martin, the hypothesis suggests that flurries of dust — swept from cold, dry landscapes like the glacial outwash where Kaplan now stood, trowel in hand — played a crucial role in the last major ice age. When this dust landed in the iron-starved Southern Ocean, Martin argued, the iron within it would have fertilized massive blooms of diatoms and other phytoplankton. Single-celled algae with intricate silica shells, diatoms photosynthesize, pulling carbon from the atmosphere and transforming it to sugar to fuel their growth. Going a step further, Martin proposed that using iron to trigger diatom blooms might help combat global warming. “Give me half a tanker of iron and I’ll give you an ice age,” he once said half-jokingly at a seminar, reportedly in his best Dr. Strangelove accent.

Thirty years after Martin’s bold idea, scientists are still debating just how much iron dust contributed to the ice age, and whether geoengineering of the oceans — a prospect still lobbied for by some — might actually work. Although it’s now well-established that an uptick in iron fertilization occurred in the Southern Ocean during the last major ice age, for example, scientists still argue about how much it reduced carbon dioxide levels in the atmosphere. And while Martin’s hypothesis inspired 13 large iron fertilization experiments that boosted algae growth, only two demonstrated removal of carbon to the deep sea; the others were ambiguous or failed to show an impact, says Ken Buesseler, a marine radiochemist at the Woods Hole Oceanographic Institution in Massachusetts.

Graph shows fluctuations in dust, carbon dioxide and temperature over past 140,000 years in Antartica.

Scientists have used ice cores and other methods to piece together evidence of fluctuations in carbon dioxide, dust and air temperatures over Antarctica during the past 140,000 years, supporting the idea that iron dust influences climate. As glaciers grew in the last ice age, reaching a maximum about 20,000 years ago, dust levels peaked while temperature and carbon dioxide levels dropped. The end of the glacial period coincides with warming temperatures, a reduction in dust and a rise in carbon dioxide.

In 2008, concerns about possible environmental impacts of iron fertilization, such as toxic algal blooms and damaged marine ecosystems, prompted the United Nations Convention on Biological Diversity to place a moratorium on all large-scale ocean fertilization experiments. The ban “put the kibosh” on such activity, says Buesseler. The problem with that, many scientists now contend, is that the most fundamental questions about iron fertilization — if it can sequester enough carbon to alter climate, and what its environmental consequences would be — remain unanswered.

As atmospheric carbon levels soar past 400 parts per million, some researchers believe that the freeze on iron fertilization experiments should be reconsidered, Buesseler among them. “I'm not a supporter of geoengineering, but I think it is our responsibility to look” at ways of actively removing carbon from the atmosphere, including iron fertilization, he says.

Whether people ever decide to pursue iron fertilization to combat climate change or not, scientists still need to understand the environmental impacts of iron-rich dust and ash from natural sources like volcanoes, and from manmade pollutants, says Vicki Grassian, a physical chemist at the University of California, San Diego. To meet that challenge, labs around the world are studying how iron affects climate and ocean health. Their work spans the scales, from the tiny crystalline structure of iron-peppered nanoparticles to large-scale simulations of global climate. Ultimately, scientists hope to understand the role of iron dust in marine systems, says Kristen Buck, a chemical oceanographer at the University of South Florida. “When you add iron to a system, how does that trigger the system to change?”

In ancient seas, iron aplenty

To learn how iron fertilization might work in the future, some researchers are looking at the past, in paleoclimate records such as ice cores and deep-sea sediments. From that perspective, many of the natural iron fertilization experiments have already been run, says Gisela Winckler, a climate scientist at the Lamont-Doherty Earth Observatory at Columbia, and Kaplan’s colleague.

Three billion years ago the ocean was chock-full of iron, ancient mineral deposits show. Iron was plentiful when life first evolved, and the metal was incorporated into a long list of essential cellular functions. Animals need iron to transport oxygen in their blood and to break down sugar and other nutrients for energy. Plants need iron to transfer electrons during photosynthesis and to make chlorophyll. Phytoplankton need it to “fix” nitrogen into a usable form.

Photo shows a reddish sedimentary rock with alternating layers of different hues.

Ancient layers of iron oxides, typically magnetite or hematite, separated by chert (a type of quartz), form sedimentary rocks called banded iron formations. (One shown here from Fortescue Falls in Western Australia.) It’s thought that iron, once abundant in the oceans, began to form such deposits on the ocean floor between 2.5 billion and 1.9 billion years ago, as oxygen levels rose.

CREDIT: GRAEME CHURCHARD / WIKIMEDIA COMMONS

Despite being the fourth most abundant element in the Earth’s crust, iron is vanishingly scarce in the modern ocean. It started disappearing from the seas more than 2.4 billion years ago, when cyanobacteria evolved and started to breathe in carbon dioxide and exhale oxygen. When this happened, dissolved iron rapidly linked up with the newly plentiful oxygen atoms, forming iron oxides such as hematite, a common mineral that contains a form of the element known as iron(III). Most phytoplankton and other living organisms can’t use iron in this state. They require a different form, iron(II), which more readily dissolves and is absorbed by cells.

Hematite has another downside: It sinks. Over billions of years, layer upon layer fell to the sea floor, forming iron ore deposits hundreds to thousands of feet deep. Meanwhile, iron in the waters above diminished to barely detectable levels — an average liter of seawater contains roughly 35 grams of salt, but only on the order of a billionth of a gram of iron. In roughly a third of the ocean, iron is so rare that its absence can hinder the growth of diatoms and other phytoplankton. The Southern Ocean, where Martin developed his hypothesis, is one of the most “iron-limited” oceans in the world. Even with an abundance of other crucial nutrients such as nitrogen and phosphorus, it’s the availability of iron that matters for diatoms and other organisms.

Unless, of course, a gust of wind delivers a plume of iron particles. Standing in the freshly excavated gravel pit in Patagonia, Kaplan was directly upwind of the Southern Ocean — close to where Martin proposed that ice age dust had helped to fertilize the ocean some 20,000 years ago. It was the perfect place to test whether those iron-rich glacial sediments would have made a good fertilizer for diatoms. Researchers already knew that there was more dust-borne iron during the last ice age, much of it freed by melting glaciers. But no one had yet rigorously tested whether the iron was in the form that diatoms can absorb, Kaplan says.

Photo shows a close-up of dirt layers, including rocks on top and fine sediments forming a narrow band below, with another layer of larger sediments below that. A sample collection bag is taped to the rocks.

Sediment exposed in an open gravel pit in Patagonia was deposited by glaciers around 20,000 years ago during the last ice age. After collecting samples from the pit in 2013, Mike Kaplan and his colleagues used it to test whether the iron-rich dust would have been a good fertilizer for diatoms — an important question for understanding iron’s role in the ancient climate.

CREDIT: COURTESY OF MIKE KAPLAN

Kaplan scraped up the dark gray silt and brought it back to Columbia, where he handed it off to then-graduate student Elizabeth Shoenfelt Troein, who is now a postdoctoral fellow at the Massachusetts Institute of Technology. Shoenfelt Troein flew out to the Stanford Synchrotron Radiation Lightsource in Menlo Park, California. There, along with her adviser Benjamin Bostick and fellow graduate student Jing Sun, she spent many long nights zapping the sediment with high-powered X-rays to reveal its mineral composition.

Only certain types of minerals yield dust that is rich in soluble forms of iron, including iron(II), the kind that diatoms can easily digest, as Grassian and colleagues described in 2008 in the Annual Review of Physical Chemistry. Clay minerals containing iron, for example, yield iron(II) more easily than hematite, as they’ve found in experiments on dust from around the world, including Africa’s Sahara Desert, Chinese loess and Saudi Arabian beach sand. Winds blowing off the Sahara are one of the most important sources of iron dust in the ocean, supplying more than 70 percent of dissolved iron to the Atlantic, another group has found. But there are several other paths by which iron(II) makes its way to the oceans, including rivers, hydrothermal vents, volcanoes and glacial outwash plains like the one where Kaplan found his sample in Patagonia.

Table lists iron and three forms of iron oxide, with their oxidation states and common names.

Iron is among the most common elements on Earth, but is rarely found in its pure metallic form (Fe). Instead, it readily reacts with oxygen to form various iron oxides, and with other elements to form a wide array of minerals. Hematite is the main source of iron used to make steel but living organisms can more readily use iron in the +2 oxidation state. Adding water to an iron oxide can create rust in what’s termed a hydrated iron oxide.

The glacial sediment contained far more iron(II) than samples deposited during non-glacial periods from the same region, Shoenfelt Troein found. When glaciers grind down bedrock, the resulting freshly ground sediments tend to contain more iron(II) than sediments produced from weathering by wind and water, which are richer in iron(III), Winckler says. Back at Columbia, Shoenfelt Troein fed the iron(II)–rich, glacial sediment to a common species of diatom, Phaeodactylum tricornutum, and the diatoms reproduced 2.5 times as fast as they did on weathered sediment, the team reported in Science Advances in 2017. This would translate into a roughly fivefold increase in carbon uptake compared with the non-glacial sediment, the team calculated.

When the team looked at marine sediment cores from several glacial and interglacial periods spanning 140,000 years, Winckler, Shoenfelt Troein and colleagues found that dust from the glacial periods contained 15 to 20 times more iron(II) than did dust from the current interglacial period. That suggests that the potency of glacial sediment led to a self-reinforcing cycle, in which higher rates of iron fertilization in the oceans reduced carbon in the air, leading to colder temperatures, which in turn, grew glaciers, the team reported in the Proceedings of the National Academy of Sciences in 2018. It also suggests that not all iron is equal when it comes to fertilization, and that freshly mined, fine-ground iron might be more effective than other forms, Winckler says.

In most of the geoengineering experiments in the 1990s and early 2000s, scientists mixed a powdered form of iron called ferrous sulfate with acidic water and fed the liquid off the back of a ship, says David Emerson, a geomicrobiologist at the Bigelow Laboratory for Ocean Sciences in Maine. The fate of ferrous sulfate once it enters the oceans is not fully known, he says, but it’s reasonable to assume that some of it oxidizes to the diatom-disdained iron(III) and sinks, even if some persists in the upper water column for days. Emerson recently proposed using aircraft to distribute a fine iron dust produced by iron-eating bacteria, called biogenic oxide. This form is composed of iron nanoparticles bound to organic compounds, and would likely stay suspended longer than ferrous sulfate in the sunlit surface waters where diatoms grow, he says.

Microscope image shows individuals of the species on a blue background. Green chlorophyll is visible within the transparent cellular structures.

Not all iron is equal when it comes to fertilizing diatoms. When scientists fed this common diatom species, Phaeodactylum tricornutum, a glacier-made sediment rich in soluble iron(II), the phytoplankton reproduced 2.5 times as fast as they did when fertilized with sediment that contained a less soluble form of iron. The higher growth rate would translate into a roughly fivefold increase in carbon uptake, the team calculated.

CREDIT: ALESSANDRA DE MARTINO, ECOLE NORMALE SUPERIEURE, PARIS / NSF

Getting iron to linger in surface waters won’t necessarily ensure that the carbon absorbed by diatoms actually reaches the deep sea, however. Roughly 90 percent of the organic carbon that diatoms create during photosynthesis is released back into the ocean in dissolved form as the algae dies, rots and is consumed by bacteria, zooplankton and fish, Buesseler says. Just 10 percent of the carbon produced by the ocean’s creatures migrates to the depths where it may remain for hundreds to thousands of years — the length of time relevant for climate mitigation. A mere 1 percent gets permanently buried on the seafloor. Critically, no iron fertilization experiment has yet lasted long enough to track how much of the carbon that diatoms do capture actually gets sequestered to the deep ocean, he says.

Location also plays a vital role in whether iron fertilization is effective, Winckler says. Based on marine sediment cores, Winckler and her colleagues have reconstructed a 500,000-year record of iron dust levels throughout the Pacific to see if — and where — notable spikes of iron fertilization occurred in the past. The team knows how much the dust level has changed, and in parallel measures the biological responses to the dust to determine if the phytoplankton “actually cares about the change,” she says. She concludes that the iron hypothesis appears to apply only to some parts of the Southern Ocean — and not other low-iron regions the such as equatorial Pacific, where past iron fertilization experiments have boosted phytoplankton growth but failed to show the degree of carbon capture scientists had expected.

There are many complex factors involved in determining where iron fertilization might work, including upwelling currents that deliver iron from deeper waters and the availability of other vital nutrients. Yet “people often just look at one piece of this puzzle, and then make big conclusions,” Winckler says.

Two maps of the world’s ocean track the distribution of soluble iron, the form of the metal that is most easily absorbed by marine organisms.

Before the GEOTRACES effort to map the global sources of iron, the metal’s path was somewhat unclear. In a 2019 paper, scientists used data from GEOTRACES to “fingerprint” the origins of soluble iron in particles of dust that fall on the ocean. Human activities such as burning coal or gasoline contributed as much as 80 percent of the soluble iron that landed on the sea surface throughout the world's oceans — far more than a previous model (shown on the left) had suggested. Mineral dust swept from dry regions like the Sahara made up a smaller portion of the ocean’s iron.

CREDIT: T.M. CONWAY ET AL / NATURE COMMUNICATIONS 2019

Grassian studies yet another factor that can influence iron fertilization in unexpected ways: the chemical reactions that transform particles containing iron as they fly through the sky, exposed to air, water and sunlight. At her lab in San Diego, she simulates the effects of water vapor and airborne pollutants on iron particles. She and her colleagues have discovered that chemicals like sulfur dioxide and nitric acid make iron more soluble — and thus easier for diatoms to absorb — by coating them in acid.

Iron particles produced by manmade pollution are also potent fertilizers, she and others have found. Iron flecks in coal fly ash, for example, are amorphous globs that dissolve more easily than the crystals found in mineral dust. The result is that even if you have less overall iron in coal fly ash, its impact on algae could be just as important as that of mineral dust, Grassian says.

Iron can rapidly alter its molecular composition or state as it moves from the Earth’s crust to the ocean, and such changes determine whether iron is in a chemical form that diatoms and other photosynthetic algae can use — and thus, how much carbon they capture. Yet, for decades, climate and atmospheric chemistry models have overlooked iron’s complexity, which includes the many forms of iron present in dust as well as how dust is altered by aging and chemical exposures. “As physical chemists, we're trying to understand the details … to get away from thinking about things in a too-simplistic fashion,” Grassian says.

Other researchers are studying what happens when dust-borne iron dissolves into the ocean. When water molecules come up against the abrupt transition to air, many can no longer find partners for all their hydrogen bonds. As a result, one of every four water molecules has something akin to a grasping limb — a single hydroxyl (OH) group — pointing up into the air with nothing to bind to, creating an uneven chemical landscape. That variability can affect how iron transforms into one of its myriad chemical identities, and then how organisms such as diatoms interact with the metal, says Heather Allen, a physical chemist at Ohio State University.

Microscopic image shows multiple silica-shelled diatoms on a black background.

Diatoms are a form of phytoplankton, marine algae that transform atmospheric carbon dioxide into organic carbon. When they die, their ornate silica shells sink, transporting some of that carbon from surface waters to the deep ocean and ocean floor.

CREDIT: SCENICS & SCIENCE / ALAMY STOCK PHOTO

Sometimes iron doesn’t interact only with the water, but also encounters a millimeter-thick gel of carbohydrates, proteins and lipids known as the sea surface microlayer, or the ocean’s “skin.” This layer can concentrate trace metals such as iron, particularly when oily pollutants common along shipping routes, such as hydraulic fluids, are present, says Allen.

Iron is so scarce in the ocean that even a bit of rust flaking off a ship’s hull can throw measurements off by a factor of 10. The instruments used to detect iron are sensitive: “If a whale poops, there goes your whole experiment,” Buck says. Through a project called GEOTRACES, Buck and an international consortium of other scientists have examined more than 20,000 measurements to map where iron comes from in the ocean, where it goes and how it changes. To avoid contamination, scientists process seawater samples in plastic-enclosed bubble labs that appear more suited to studying deadly microbes than one of Earth’s most abundant elements.

They’ve found that most naturally produced iron dust blows off the Sahara and other deserts, but large amounts are also released in plumes of hot dissolved minerals from hydrothermal vents. Volcanoes, which can spew thousands of kilograms of iron into the atmosphere in a single eruption, are another important source. Although the evidence is circumstantial, iron fertilization from volcanic ash may have contributed to the brief hiatus in carbon buildup in the atmosphere after the 1991 eruption of the Philippines’ Mt. Pinatubo, says Emerson. Unfortunately, there was no monitoring at the time to determine if this led to a large-scale iron fertilization event, he says.

The ocean’s iron miners

Given how quickly iron rusts and sinks, there should be very little dissolved iron in seawater, including the highly soluble iron(II). Yet GEOTRACES has detected more of it than scientists predicted. Buck and others believe that some of these scant traces of dissolved iron can be explained by an active effort made by living things to scavenge it. In addition, they point to the presence of organic molecules called ligands, which lock up iron in a soluble, diatom-friendly form. One common example of a ligand is found in siderophores, chemical compounds that bacteria secrete to break down iron particles.

Some organisms actively mine iron from dust. On the northernmost end of the Red Sea, for example, marine biogeochemist Yeala Shaked of the Hebrew University of Jerusalem is studying how a stringlike, reddish kind of phytoplankton called Trichodesmium takes advantage of iron-rich dust that blows in from the Sahara. This Trichodesmium species assembles into puffball-shaped colonies, each composed of tens to thousands of individual filaments. When this dust lands, the colonies shuttle the iron-rich mineral particles into the center of the colony and start extracting iron(II). A colony can transform a pool of iron(III) into iron(II) in 30 minutes, Shaked and her colleagues have found in lab experiments.

These puffball-like colonies of Trichodesmium, a kind of phytoplankton, actively mine iron(II) from mineral dust. Studies by climate scientist Yaela Shaked and colleagues suggest the organisms can sense and select dust particles that are rich in the bioavailable iron(II), and also shuttle particles of mineral dust that contain iron inside the colony. There, the colony can rapidly convert iron(III) into the more easily dissolved iron(II).

CREDIT: COURTESY OF YEALA SHAKED

Even small changes in the abundance and productivity of phytoplankton could have a significant impact on marine life and the rate of global warming, so organisms such as Trichodesmium are key to global climate models. An ambitious effort at MIT, for example, is attempting to incorporate many different phytoplankton species into their simulations.

Despite iron’s assumed influence over climate, climate models still don’t currently include much detailed information about the element, says Andreas Schmittner, a climate scientist at Oregon State University. Although it’s now well-established that iron fertilization occurred in the ancient Southern Ocean, for example, there’s still lively debate over how much it affected past carbon dioxide levels. Some scientists have argued that iron fertilization wasn’t particularly important, and that most of the roughly 100 ppm drop in carbon dioxide during the last ice age can be explained by changes in ocean currents and sea ice.

But in June 2019, Schmittner and colleagues published a different take in Science Advances, calculating that cooler temperatures and iron fertilization were responsible for most of the decrease, and ocean circulation and sea ice had “close to zero” impact, he says. Iron fertilization alone accounted for a 25 to 35 ppm decrease in atmospheric carbon during that period, “a larger effect than we expected,” he says.

Once scientists have pieced together more about iron’s complex chemistry, they will still have to learn when to turn certain factors off and on in these climate models to accurately simulate reality, Grassian says. Better models will also depend on fine-tuning countless other factors that could affect how much carbon dioxide sequestration occurs in response to a phytoplankton bloom, including how layers of ocean water mix, and the presence of zooplankton, tiny marine organisms that graze on algae.

Several iron fertilization experiments favored certain phytoplankton species over others, a consequence that could inadvertently reorganize marine food webs. Large algal blooms both natural and manmade have also been known to deplete oxygen in the water, creating dead zones. One risk is that iron fertilization could damage ecosystems downstream, by depriving them of nutrients that normally would have reached them, Buesseler says. “What happens when that water upwells somewhere else and [a] fishery collapses because … you’ve kind of stripped away all the juicy nutrients in one part of the ocean?”

Meanwhile, the controversy over iron fertilization as a geoengineering approach rages on. As the vision of a climate-tweaking tool has waned, some companies have attempted to apply the idea to revitalize fisheries. In a highly controversial 2012 example, American businessman Russ George persuaded members of the Haida Nation to fund the dumping of roughly 100 tons of iron sulfate off the coast of Canada, fertilizing a 10,000-square-kilometer algae bloom. George sold the controversial project as a way to boost salmon populations and sequester carbon, but follow-up studies failed to find conclusive evidence that it worked.

In 2013, the London Protocol, an international treaty that prevents ocean dumping, adopted amendments allowing researchers to apply for exceptions to the moratorium on iron fertilization experiments. Winckler does not advocate using iron fertilization as a geoengineering tool, but she is among those who think that more rigorous experiments are necessary to establish the approach’s efficacy and potential risks and benefits, even if people decide never to use it. “We are in a climate crisis, and we’ve got to think about these questions,” she says.