Lake Kivu

Lake Kivu sits in the Western Rift between Rwanda and the Democratic Republic of Congo, a deep mountain lake with a secret in its lower 400 metres: roughly 60 cubic kilometres of dissolved methane and nearly 300 of carbon dioxide, held down by a density barrier that almost never overturns. That same barrier makes Kivu one of the most completely stratified lakes on Earth and one of the most lopsided ecologically — fish live only in a thin, oxygenated surface skin above water that is anoxic, gas-charged and lethal. The lake is also young and geologically violent, blocked into being by the Virunga volcanoes well under a million years ago, and its cichlids reflect that youth: a small flock of fifteen endemic Haplochromis rather than the hundreds of older rift lakes nearby.

Geography & hydrology

Lake Kivu lies at about 1,1519 ft above sea level in the Western (Albertine) Rift, hemmed by slopes that climb past 2,0 ft on both the Rwandan and Congolese shores. It is roughly 62 mi long with a maximum width near 31 mi, and it is deep — about 1591 ft at its lowest point — which makes it the most completely stratified lake in Africa. Of the western rift lakes it is one of the smaller bodies of water, sharing the loop with Albert, Edward-George and the much larger Tanganyika to the south (Snoeks et al., 1997; FEOW Ecoregion 521).

The drainage history matters more here than the surface dimensions, because it explains why the fish fauna is what it is. Before the East African uplift, the country now under the lake drained westward toward the central Congo basin. A proto-Lake Kivu existed by the mid-Pleistocene and was connected to the ancient Lake Edward basin to the north, so its earliest fishes came from that East Coast lineage rather than from the Congo. Then the Virunga volcanoes erupted across the northern valley — somewhere between roughly 25,000 and 11,000 years ago, depending on whose dating you follow — and dammed the outlet. The closed basin filled, and around 9,500 to 9,200 years ago a new southern spillway, the Ruzizi River, cut through by headward erosion and reversed the lake's drainage entirely. Kivu now empties south into Lake Tanganyika, and through Tanganyika it belongs hydrographically to the Congo system, even though its fishes do not (Snoeks et al., 1997; Haberyan & Hecky, 1987).

That volcanic plumbing is still active, and it feeds the lake from below. Warm, salty, gas-laden groundwater enters through subaquatic springs, much of it along the northern shore, and sinks because of its density. The result is a lake that does not mix from top to bottom: a near-permanent chemical boundary sits around 837–853 ft, and below it the water is older, saltier, warmer at depth than a normal lake, and charged with dissolved gas. Kivu is meromictic, meaning the deep monimolimnion is effectively cut off from the seasonally mixing surface layer above it, and that isolation is what lets the gas accumulate instead of escaping (Bärenbold et al., 2020; Schmid & coworkers, PLOS ONE 2020).

Water & habitat

Most of Lake Kivu is uninhabitable for fish, and that is the single fact that organizes its biology. Only the top 197–213 ft — the biozone, or mixolimnion — carries oxygen and supports life. This surface layer mixes seasonally, warms and cools with the weather, and holds negligible dissolved gas. Beneath it the water column is layered like a stack of glass plates that refuse to slide past one another: each layer is denser than the one above, stabilized by salinity and dissolved carbon dioxide, so the lake resists the annual overturn that ventilates ordinary deep lakes (Bärenbold et al., 2020; Saboorian-Jooybari & Hassanzadeh, 2025).

Go below the main chemocline near 837 ft and the chemistry turns hostile fast. The deep water is anoxic and saturated with gas delivered by the subaquatic springs and, in the case of methane, generated in place by microbes breaking down sinking organic matter. The totals are large: on the order of 60 cubic kilometres of methane and close to 300 of carbon dioxide at standard pressure dissolved across the whole lake, concentrated in the so-called resource zone below about 853 ft. The gas stays in solution only because of the pressure of the water above and the stubborn density gradient holding everything in place — the same condition that gives the lake its limnic-eruption reputation (Bärenbold et al., 2020).

Kivu's surface water is soft but more mineral-rich than the rain-fed lakes to the north; Snoeks and colleagues noted its relatively high salinity, a legacy of those geothermal inputs, as one reason the lake has been hard for new fishes to colonize. So the habitable lake is, in practical terms, a warm tropical surface lens: a rocky and sandy littoral fringe, open pelagic water down to perhaps 197 ft, and then a floor that drops away into dark, dead, gas-filled deep water no fish ever sees. Compared with the structural variety of an old rift lake — the reefs, sand flats, shell beds and deepwater ledges of Tanganyika or Malawi — Kivu offers a narrow stage. There is shoreline, there is the open top of the water column, and below that there is nothing for a fish to use. The animals here have had little room and little time, and both show in the roster (Snoeks et al., 1997; Verbeke 1957, in Snoeks et al.).

The cichlid fauna

Lake Kivu's entire described fish fauna runs to about 28 species across the lake and its affluent streams, of which 19 are cichlids — and 15 of those cichlids are endemic Haplochromis, the only fishes found nowhere else on the planet (Snoeks et al., 1997). The atlas roster is exactly that flock: Haplochromis adolphifrederici, H. astatodon, H. crebridens, H. gracilior, H. graueri, H. insidiae, H. kamiranzovu, H. microchrysomelas, H. nigroides, H. occultidens, H. olivaceus, H. paucidens, H. rubescens, H. scheffersi and H. vittatus. That is a flock, not a radiation. Lake Victoria produced hundreds, and even Kivu's smaller northern neighbours are richer. The shortfall traces back to the lake's youth, its isolation behind volcanic barriers, its salinity, and the recurring tectonic insults — gas upwellings and lava entering the water — that may have wiped out part of whatever fauna the proto-lake once held (Snoeks et al., 1997).

Where did these fifteen come from? They are placed within the great Victoria-Edward-Kivu super-flock, the same haplochromine assemblage that fills the East Coast lakes, and their nearest relatives often turn out to be in other lakes rather than within Kivu itself. That detail has driven a long argument about whether the Kivu fish descend from a single colonizing ancestor that diversified in place, or from several lineages that entered separately and now share the lake. Enzyme work and scale morphology have hinted at a single monophyletic group, which would imply one founding event; the broader super-flock hypothesis, with lineages cutting across modern lake shores, argues against that tidy picture. Snoeks and colleagues left the question open and called for molecular data, and later work has even questioned how many valid species the flock really contains (Snoeks et al., 1997; Walker, Eawag thesis 2013).

The species sort themselves out by feeding and form rather than by the rock-versus-sand niches of an old lake. Tooth-shape studies of three sympatric Kivu endemics — H. kamiranzovu, H. insidiae and H. astatodon — found each carries a distinct dentition, evidence that even in this small flock the fish have partitioned diet finely enough to coexist (Hata et al., 2020). Beyond the haplochromines, the cichlid list is thin and mostly borrowed. One tilapia is native — Oreochromis niloticus, the Kivu-Edward-Tanganyika form sometimes called eduardianus — while three others (Oreochromis macrochir, O. leucostictus and Tilapia rendalli) are introductions that escaped from fish ponds in the Kivu drainage from the late 1940s onward and established themselves. The non-cichlids are a short list of widespread East Coast cyprinids and catfishes (small Barbus, Clarias liocephalus, an Amphilius) plus one Tanganyikan immigrant, the cyprinid Raiamas moorii, thought to be the single fish that made it up the young Ruzizi from the south. None of the native non-cichlids is endemic. The endemism of Kivu lives entirely in the Haplochromis (Snoeks et al., 1997; Hata et al., 2020).

Conservation

Kivu's defining feature is also its defining hazard and, lately, its industry. Because the deep water holds so much dissolved gas behind so fragile a density barrier, the lake belongs to the rare category that produced the 1986 Lake Nyos disaster in Cameroon, where a sudden release of carbon dioxide suffocated more than 1,700 people. Nyos and Monoun are tiny crater lakes; Kivu is vastly larger and has roughly two million people living along its shores, which is why its gas has been studied so hard. The current scientific reading is reassuring on the near term: an intercomparison led by Bärenbold and Schmid found methane and carbon dioxide close to steady state, with no measurable increase over the preceding 45 years, and a 2025 modelling study put the chance of a buoyancy-driven overturn or spontaneous degassing as very low over the next five centuries. Methane is the one component still slowly climbing, because microbes keep generating it in the deep water (Bärenbold et al., 2020; Saboorian-Jooybari & Hassanzadeh, 2025).

That renewability is the basis of an energy business. Rwanda extracts the methane to burn for electricity: the KivuWatt plant, run by ContourGlobal and online since 2016, pulls deep water up from around 984 ft on a floating gas-extraction facility, depressurizes and washes the methane out, then pipes it ashore to generate about 26 MW — by some accounts close to a third of the country's installed capacity, with further plants meant to push methane toward a third of national demand. The appeal is obvious: power for a country whose electricity access was in the single digits a generation ago, and a controlled drawdown of a gas that would otherwise only accumulate. The catch is that extraction means deliberately disturbing the very stratification that keeps the lake stable, and experts remain split on how to reinject the stripped water without nudging the density structure toward instability. Rwanda and the DRC signed an accord in 2020 to coordinate safe extraction across the border, though Congolese projects have moved slowly (Bärenbold et al., 2020; ContourGlobal KivuWatt; Equal Times, 2024).

For the fishes, the pressures are more ordinary but real. The pelagic fishery rests almost entirely on an introduced animal: the Tanganyika sardine Limnothrissa miodon, stocked from Lake Tanganyika in 1959 to fill an empty open-water niche, now the lake's main commercial catch and a food staple known locally as isambaza or sambaza. Its sister species Stolothrissa tanganicae was introduced at the same time and failed to take. The fishery is worked at night from lamp-lit trimaran rigs, but it has strained under illegal fine-mesh gear that scooped up larvae and eggs, prompting Rwanda to mandate minimum mesh sizes by species. The endemic Haplochromis are the conservation concern that is easy to overlook beside the gas and the sardine: a flock of fifteen fishes restricted to a single, geologically twitchy lake, sharing the water with introduced tilapias and an introduced clupeid, and dependent on the thin oxygenated surface layer that any serious disturbance to the lake's stratification would put at risk (Snoeks et al., 1997; Lamboeuf et al., on Limnothrissa sustainability; RAB Rwanda fisheries).

Where these cichlids live

Geography & hydrology

Water & habitat

The cichlid fauna

Conservation

Sources

Cichlids recorded here

Haplochromis · 16