Skip to Content
Biotechnology and health

Lake Kivu’s Great Gas Gamble

In a first-of-its-kind endeavor, electricity-starved Rwanda and the Democratic Republic of Congo are trying to get power from a lake—and avert catastrophe.

It’s a Friday afternoon on the Rwandan side of Lake Kivu, and in what was once a quiet cove, a daring venture is taking shape.

Floating just offshore, like a giant mechanical swan, is a nearly completed gas extraction platform: 3,000 tons of concrete and stainless steel that will soon begin capturing a resource not found at this scale in any other lake in the world. Dissolved within Kivu, which straddles the border of Rwanda and the Democratic Republic of Congo (DRC), are approximately 60 billion cubic meters of methane and 300 billion cubic meters of carbon dioxide. The gases, which come from nearby volcanic activity and bacteria decomposing organic material in the lake, represent both danger and economic potential.

If extracted, Kivu’s methane could be used to add up to 960 megawatts of electricity-generating capacity, more than six times what Rwanda has now. For both Rwanda and the eastern DRC, which face crippling power shortages and limited options for expanding their electric grids, that could be an economic game changer, supporting new industries and offering a chance to alleviate searing poverty. If the extraction is done properly and the countries can coöperate, it could even help improve their troubled relations and advance stability in a region long beset by turmoil.

A safety inspector examines the new barge on shore in Kibuye, Rwanda.

Just as critical, removing Kivu’s methane may prevent a possible catastrophe. With methane concentrations rising, scientists warn that Kivu will eventually experience a deadly phenomenon known as an overturn. Also known as a limnic eruption, an overturn can occur if the pressure of the gases in a lake exceeds the pressure of the water at a given depth, causing a chain reaction that releases them with violent results. Only two limnic eruptions are known to have occurred in recorded history—both in small lakes in Cameroon in the 1980s. In the deadlier of the two episodes, at Lake Nyos in 1986, more than 1,700 people were asphyxiated when a cloud of carbon dioxide, which burst from the lake along with a 100-meter fountain of water,spread as far as 25 kilometers from shore. Kivu contains a thousand times more gas than Nyos: if even part of it escaped this way, more than two million people living near its shores would be at risk.

In Kivu it’s the methane, rather than the carbon dioxide, that’s most likely to trigger a gas eruption. That adds urgency to the prospect of harnessing its energy potential, something both Rwanda and the DRC have long sought to do. After decades of little or no progress, gas extraction efforts in both countries have finally gained momentum. On my visit to the lake in February, more than a hundred orange-vested workers were putting the final touches on the first phase of KivuWatt, a $200 million project owned by the U.S. energy firm Contour Global. The lake’s first industrial-scale gas-fueled power project, it is expected to add 25 megawatts of generating capacity by the middle of this year and eventually scale up to 100. Another U.S. company, Symbion Power, is set to begin construction of a 50-megawatt project on the Rwandan side of the lake by the end of the year. In the DRC’s distant capital, Kinshasa, the Ministry of Hydrocarbons is now reviewing bids for that country’s first Kivu gas concession.

Getting the gas out correctly, however, will be tricky. Although the Rwandan government has operated a pilot gas-fired power plant at the lake since 2008, the process of extraction is novel and has been done only on a very small scale. While most experts agree that the lake’s methane should be kept from accumulating further in order to prevent a disaster by the end of the century, a few warn that certain extraction processes could disturb the natural stratification that keeps the bulk of the gases trapped in deep waters. Undertaking them could increase, rather than mitigate, the risk of gas eruption. Until a large-scale extraction operation has commenced, it also remains unclear how efficiently the technology will function and how much electricity Kivu will ultimately yield.

“We are very curious to see how our process works,” says Jarmo Gummerus, a Finnish engineer and KivuWatt’s Rwanda country manager. “Very soon we’ll have a much better idea of the potential of this lake.”

Boosting the grid

Three hours by car over winding roads from KivuWatt, the Rwandan capital, Kigali, does not appear to be a city in the midst of an energy crisis. In the 21 years since the Rwandan genocide, in which an estimated 800,000 people were killed, the city of a million has transformed from a corpse-ridden backwater into a tidy modern metropolis. Today, Kigali is a town of smooth tree-lined streets, sprouting office towers and American-style subdivisions that stretch to the surrounding hills. It’s also the engine of a Rwandan economy that’s grown at an average of 8 percent per year over the last decade—one of the highest rates in the world.

As Rwanda and its capital have developed, however, the country’s electricity grid has struggled to keep pace. Although installed capacity has doubled in the last five years, it remains a scant 156 megawatts. Today, nearly 80 percent of Rwanda’s 12 million people, including the vast majority of rural residents, still lack a connection to the grid. Families and business that do have power, meanwhile, face some of the highest electricity prices in the region—in part because nearly a third of the country’s power is generated from imported diesel and heavy fuel oil, which arrive by truck from Kenya and Tanzania. According to the World Bank, Rwandan companies pay an average of 24 cents per kilowatt-hour, compared with 15 cents in Kenya and 17 cents in Uganda. The average industrial user in the United States pays less than seven cents.

Kivu, 1,460 meters above sea level, is part of a system of lakes along the Great Rift Valley.

Hoping to reduce its widespread poverty and boost its small industrial base, Rwanda has set ambitious electrification targets. The country’s second Economic Development and Poverty Reduction Strategy, launched in 2013, assumed a nearly fourfold expansion of the power grid, to 563 megawatts, by the end of 2017. Given financial constraints and limited domestic energy resources, however, this will be difficult to pull off. Aside from KivuWatt, the only significant power project nearing completion is a 15-megawatt plant that will burn peat. Although work has begun on another 80-megawatt peat facility, and financing is being arranged for two large-scale regional hydroelectric projects, it’s not clear if any will be on the grid by the 2017 target. Rwanda might also have significant geothermal resources, if preliminary surveys are correct, but two exploratory wells drilled in 2013 came up empty. And although Rwanda recently inaugurated East Africa’s first utility–scale solar field and authorities are working to bring off-grid solar installations to rural homes, schools, and hospitals, it’s unlikely that solar will be able to meet a significant portion of industry’s demands. Out of desperation, Rwanda could soon become a significant electricity importer. According to the Ministry of Infrastructure, arrangements are in the works to purchase 30 megawatts from Kenya this year and, eventually, up to 400 megawatts from Ethiopia.

Across the border in the eastern part of the Democratic Republic of Congo (formerly known as Zaire), the power crisis is even more acute. The DRC, a country of 77 million people in a territory roughly the size of Western Europe, contains extensive hydroelectric resources. If fully tapped, the Congo River’s Inga Falls could yield an estimated 40,000 megawatts, nearly twice the capacity of the world’s largest power station, the Three Gorges Dam in China. Today, however, the DRC’s aging grid has an installed capacity of just 2,400 megawatts, roughly half of which is routinely unavailable because the transmission infrastructure is in such poor shape. In the war-torn east, power is particularly limited. Goma, the largest city on Lake Kivu, has an available capacity of less than five megawatts—a meager amount for a town of a million residents and a situation, some argue, that helps promote conflict. If boosting eastern Congo’s grid can spur the development of industries, says Bantu Lukambo, an environmental activist based in Goma, that would reduce the appeal of the region’s dozens of armed groups, which are magnets for youth with no other employment prospects. In addition, he says, more development could weaken the market for illicit charcoal, a trade that generates millions of dollars per year for local militias and leads to extensive deforestation.

Gases from the lake will enter the gray stainless-steel tube to be separated.

The volcanoes responsible for much of Kivu’s gas loom over Goma and its environs. In 2002, an eruption of Nyiragongo, a volcano located 20 kilometers north of town, destroyed a fifth of the city, leaving tens of thousands of people homeless and depositing lava that’s still being used as a building material. On a drive west from town, Mathieu Yalire, chief geochemist at the government-run Goma Volcano Observatory, shows me several depressions known to contain lethal seepages of carbon dioxide that are concentrated near the ground at the edges of past lava flows and occasionally asphyxiate children. At Kasinga Primary School in the town of Sake, 25 kilometers west of Goma, principal Batchoka Lubungo shows us a photo, displayed on the wall of his office, of a young victim.

“One morning we found the boy dead over there,” he says, pointing to a known danger zone just outside his window. “We keep this picture here as a warning to the students.”

The presence of mazuku is a reminder of Lake Kivu’s potential to sow disaster. But the carbon dioxide is not the only danger. The lake’s geochemistry is unusual, largely as a consequence of local subaquatic springs that absorb carbon dioxide from the region’s volcanic soil and feed the gas into Kivu’s deepest waters. Much of the methane comes from decomposing organic matter; the rest comes either from the volcanic soil or from bacteria converting the carbon dioxide to methane. Critically, these springs are saline, while the water sources feeding the lake’s upper layers are fresh. Since saline water is much denser than fresh water, this creates density gradients that prevent the gases from diffusing upward and into the atmosphere. Although this stratification is stable at present, the gas accumulation it makes possible has apparently led to limnic eruptions in the distant past. If nothing is done, it is likely to do so in the future.

Top: Women dry tiny sambaza, or sardines.
Bottom: A sign in Kibuye explains KivuWatt.

Still, much about this risk remains uncertain. Studies of Kivu’s sediment record suggest that the lake has experienced at least five overturns in the last 6,000 years. It’s not clear, however, whether these events involved all the lake’s layers of water, thus releasing all its gas, or just portions of its upper layers. In addition, though recent measurements have found that the concentration of methane is increasing—at a rate that could bring the gas close to saturation by the end of the century—it’s not yet known why this is happening or whether it will continue. Complicating matters, Kivu consists of five different basins of varying depths, each with distinct physiochemical properties.

It is clear, though, that an eruption in Kivu’s main basin could cause a disaster of apocalyptic proportions. If all the methane and carbon dioxide currently dissolved in Kivu were released into the atmosphere, they would cover the entire lake in a cloud of gas more than 100 meters thick. If even a small fraction of the gas were to get out, it could suffocate entire towns along the lake shore. This can happen if water at a given depth becomes fully saturated with gas and is lifted by a big earthquake, a volcanic eruption, or another external disturbance to a depth where the water pressure is not great enough to keep the gas dissolved.

Extraction

Whatever the extent of Kivu’s eruption risks, its methane has long been of commercial interest. From 1963 until 2006, Rwanda’s lakeside Bralirwa Brewery fueled its boilers with methane extracted 800 meters offshore. In the 1980s, researchers from the Netherlands tapped Bralirwa’s excess gas to fuel a fleet of cars, though the project eventually foundered. By the early 1990s, cross-border efforts to utilize the gas for electricity had begun to gain momentum, but progress was cut short by the Rwandan genocide and subsequent wars in eastern Congo. Eventually, with the return of stability in Rwanda, the government in Kigali entered into partnership with a Scottish firm to build a small pilot facility on the lake. The plant, known as KP1, began working intermittently in 2008 and produces a few megawatts of electricity. Another project briefly produced 2.4 megawatts in 2010, but the equipment was removed from the lake after it was damaged in a storm.

KivuWatt is the country’s first attempt at large-scale gas extraction. Even though its technology is novel, the concept is relatively simple. A barge will be anchored to the lake bed 13 kilometers offshore, where four plastic pipes will draw up water from 350 meters below the surface. As the water rises, bubbles of methane and carbon dioxide will begin to form; eventually, roughly 80 percent of the methane and 40 percent of the carbon dioxide will be siphoned off inside a subsurface horizontal chamber known as a separator. From there, the partially degassed water will be reinjected deep into the lake, and the gas—at this point roughly 30 percent methane and 70 percent carbon dioxide,with trace amounts of hydrogen sulfide—will continue upward into one of four towers on the barge. Here, “wash water” taken from a depth of 40 meters will be mixed with the gas to remove as much of the remaining carbon dioxide as possible. This water will be returned to a depth of 60 meters, shallow enough for some of the carbon dioxide to eventually diffuse into the atmosphere. The end product, a gas composed of roughly 85 percent methane, will then be pressurized and sent to a power plant on shore.

For a region in such dire need of electricity, Kivu’s gas is an attractive proposition. According to Gummerus, the engineer overseeing the project, KivuWatt will sell power generated in its first phase to Rwanda’s state-owned utility for less than 15 cents per kilowatt-hour. That’s competitive with the rates expected from the country’s forthcoming peat projects and less than half the cost of power generated from imported fossil fuels. (Later the rate is expected to fall to less than 12 cents per kilowatt-hour.)

As desirable as the project seems for both economic and safety reasons, however, it could pose environmental risks of its own—including the chance that the degassing operations could change the structure and properties of the lake.

Children of fishermen play around fishing boats off the shores of Gisenyi.

According to the Management Prescriptions for the Development of Lake Kivu Gas Resources, a 2009 document—known as the MPs—that both Rwanda and the DRC have adopted as general guidelines for gas extraction, the risks have less to do with the removal of the gas itself than with the reinjection of the degassed water. Since Kivu’s methane-rich deep water is saline, dense, and abundant in nutrients, releasing it near the surface could damage the lake’s ecosystem and weaken its density stratification. To mitigate these risks, the report dictates that all water extracted from the lake’s deep-water “resource zone” be reinjected at least 260 meters below the surface so that it remains under enough pressure. Because KivuWatt’s design was approved before the report was released, however, the project isn’t subject to this requirement, and its degassed water will be returned above that level, at a depth of 240 meters. Although this distinction may seem trivial, Philip Morkel, a South African engineer and a member of the five-person expert committee that wrote the guidelines, believes otherwise. “Once you punch through those gradient layers, you start damaging the protective mechanism that the lake has to preserve itself,” he says. “On a large scale it becomes seriously problematic.”

Some experts see less reason for alarm. Dario Tedesco, an Italian volcanologist with extensive knowledge of the lake, tells me the quantity of water reinjected is unlikely to be large enough to create a serious disturbance. Alfred Johny Wüest, head of the aquatic physics research group at the Swiss Federal Institute of Aquatic Science and Technology and another member of the MPs committee, says it erred on the side of caution, which means that even the water from KivuWatt will be reinjected deep enough for safety.

Wüest has other concerns, however, including that the project could harm the lake’s ecology. That worries some fishermen, too. Fishing will be forbidden in an exclusion zone around KivuWatt, but even outside that area, the reinjection of wash water and the noise and vibrations from the extraction project could be noticeable. Fishermen near the KP1 pilot plant told me that their catches, mainly of sardines known as sambaza, have fallen significantly since the facility began to operate. It’s difficult to isolate the cause there, however, since a drop in sambaza numbers has been documented across the lake—a decline linked to a rise in unregulated boats and the introduction of a predator species.

Workers attend to the final construction phase of the KivuWatt barge.
A portrait of Rwanda’s president looms over controls in the KP1 pilot plant.

The most pressing concern from an energy standpoint is how well the gas extraction technology will perform. The MPs’ authors estimate that the lake is capable of providing between 160 and 960 megawatts of generating capacity over a period of 50 years. After that, the gas that continues to accumulate could be harnessed for up to 100 megawatts of capacity in perpetuity.

All of this, though, hinges on the efficiency of the extraction process and the onshore power conversion technology. KivuWatt should extract about two-thirds of the methane from the water it draws up, with the rest lost during the separation process and in the washing towers. Although some of that gas will be returned to the lake in the reinjected water and could theoretically be reëxtracted at a later date, it is likely to settle in smaller concentrations, which could make it uneconomical to capture. But these projections are based only on simulations, and the real efficiency won’t be known until the barge begins to operate. Essentially, the lake’s potential is still highly uncertain.

Disappointing results could spur disputes between Rwanda and the Democratic Republic of Congo. Although the two countries signed an agreement in 1975 to share the methane equally, Rwanda’s head start in exploitation means it could potentially infringe upon its neighbor’s resource, particularly if the lake yields less power than expected. (Because gas levels in the lake’s main basin are uniform across a given depth, it’s impossible to extract from Rwandan waters without affecting the quantity available in the DRC.) The two countries do not exactly have an amicable history. Since the current Rwandan government seized power at the end of the genocide, Rwanda has twice launched rebellions to topple governments in Kinshasa—once successfully. It has also been implicated in the support of numerous proxy militias in the DRC’s east as well as the widespread smuggling of Congolese minerals, including gold, tin, and coltan, an ore mined for use in electronics manufacturing. In this context, some worry, Kivu’s gas may simply be another means by which Rwanda, a small yet highly organized state, manages to profit at the expense of its larger, dysfunctional neighbor.

After the barge is completed, it will be towed 13 kilometers onto the lake to begin extracting gas.

If the project is successful, though, it could help mend cross-border relations. The two countries, which already share power from a hydroelectric plant on their border south of Kivu, have long sought to implement a joint gas-to-power project, according to Augusta Umutoni, head of Rwanda’s Lake Kivu Monitoring Program, a government body that oversees the extraction process. Despite the vagaries of politics, she says, energy officials from the two countries collaborate well at a technical level. Both, however, are waiting until the concept has been shown to work on a commercial scale.

With so many questions lingering as KivuWatt prepares for launch, it’s easy to forget that KP1, the pilot facility, is now in its eighth year of operation. Although the project, originally conceived to generate five megawatts, struggled to produce more than a single megawatt for years, it is now consistently generating between two and three—giving authorities a slight boost of confidence.

On a February evening, I join Olivier Ntirushwa, KP1’s plant manager, on a tour of the barge, which floats a kilometer from shore, within sight of the Rwanda-DRC border. Looking north, I see the city of Goma and its erratic patchwork of lights. Further afield is the steaming cone of Nyiragongo. Over the noise of the barge’s machinery, Ntirushwa gives me a brief history of the project: how it operated well below capacity for years; how recent improvements to its separator finally boosted its power output. When I ask about his predictions for KivuWatt, Ntirushwa says he doesn’t know much about the project, or how much electricity the lake might eventually yield. Instead, he stresses how it feels to be involved in such a high-stakes experiment.

“It’s exciting because we are the pioneers of this technology,” he says as we stare out over the water. “Nobody else has ever done this.”

Jonathan W. Rosen is a journalist based in Kigali.

Deep Dive

Biotechnology and health

Scientists are finding signals of long covid in blood. They could lead to new treatments.

Faults in a certain part of the immune system might be at the root of some long covid cases, new research suggests.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

The next generation of mRNA vaccines is on its way

Adding a photocopier gene to mRNA vaccines could make them last longer and curb side effects.

Ready, set, grow: These are the biotech plants you can buy now

For $73, I bought genetically modified tomato seeds and a glowing petunia.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.