Freshwater ecoregion · Africa · river & basin

Upper Zambezi Floodplains

15 species in the atlas are recorded from this freshwater ecoregion. It sits within The Zambezi & Okavango.

Ecoregion boundaries follow the Freshwater Ecoregions of the World (FEOW) framework; species are placed here from the documented range in each species profile.

Where these cichlids live

193 georeferenced records of this ecoregion's cichlids (GBIF + type localities; sampled). Click a point for the species.

Above Victoria Falls the Zambezi behaves nothing like the gorge-cutting river below it. It spreads. For 250 kilometers it slides across the flat, sand-floored Barotse floodplain of western Zambia, a plain so gentle that the annual flood rises only about five meters yet drowns 7,500 square kilometers of grassland. This is the heart of the Upper Zambezi Floodplains ecoregion, a tropical wetland shared by Zambia, Angola, Namibia's Caprivi Strip and northern Botswana, and it is the center of diversity for the largemouth and smallmouth river breams, Serranochromis and Sargochromis. The falls themselves seal this fauna off: a 98-meter wall that fish above cannot cross.

Geography & hydrology

The Upper Zambezi Floodplains ecoregion takes in the floodplains of the Zambezi and Cuando (Kwando/Mashi) rivers above Victoria Falls, reaching across southwestern Zambia, southeastern Angola, the Caprivi Strip of Namibia and the northern edge of Botswana. By convention it begins where the Lungwebungu and Kabompo join the Zambezi; from there the river runs south through the Barotse plain, swings east along the Caprivi, and finally drops 98 meters over Victoria Falls into the Batoka gorge. That fall is the defining feature for the fishes. It is an impassable barrier between the Upper and Middle Zambezi, so the fauna above it has evolved largely in isolation, sharing more with the neighbouring Okavango and Kafue than with the river just downstream (FEOW Ecoregion 556; Bell-Cross 1972; Davies 1986).

The whole basin sits above 1,220 meters on a bed of ancient Kalahari sand, and the gradient is so slight that the Zambezi here was long ago classified as a 'reservoir river.' Instead of confining a flood to a narrow channel, it lets the water spill sideways across the plain, which buffers the rise: the average flood height is only about 5.2 meters and the river seldom shows the violent crests of the Middle Zambezi below the falls. Three floodplains do most of the ecological work. The central Barotse floodplain runs roughly 240 kilometers from Lukulu south to Nangweshi, up to 35 kilometers wide, and at peak flood covers some 7,500 square kilometers and holds on the order of 17 billion cubic meters of water. The eastern Caprivi (southern Barotse) plain stretches about 100 kilometers between Sesheke and the falls and merges with the Chobe-Linyanti system, where Lake Liambezi fills to a 100-square-kilometer sheet of open water when the floods are strong and dries out entirely when they fail (FEOW Ecoregion 556; Jackson 1961; van der Waal & Skelton 1984; Marshall 2000).

The hydrology is a study in delay and reversal. Because most of the rain falls in the northern headwaters in Angola and northern Zambia, where annual rainfall reaches 1,15.5 in against barely 27.5 in near the Namibian border, the flood that swamps the southern plains is largely imported from upstream. The Cuando and Chobe drain in from the west, and in very wet years water even arrives from the Okavango. When the Upper Zambezi floods hard, it can shove water back up the Chobe as far as Ngoma, reversing that tributary's flow; as the flood drops the Chobe turns around again and helps drain the eastern plain. Rain begins in November, the river rises through December to March or April, and the slow draw-down through the dry months sets the calendar that every fish in the system lives by (FEOW Ecoregion 556; Bell-Cross 1972; Jackson 1986; van der Waal & Skelton 1984).

Water & habitat

Floodplain water here is soft, warm, clear and chronically poor in nutrients. The rivers run over Kalahari sand and acidic, leached soils, so they carry little dissolved load. Working on the Barotse plain through 1989, Kirk Winemiller measured Zambezi water at 70–82 °F and floodplain canals from 64 °F on a cold August morning to 82 °F at the surface in late November; pH ran 6.3 to 6.9 in the main river, dropped to about 5.5 in stagnant floodplain pools, and reached 7.0 only in the inflowing Kabompo. The water was almost always near full oxygen saturation and clear, though often stained tea-brown with dissolved plant compounds. Phytoplankton production stays low in these transparent, sand-bottomed waters, which is part of why the system depends so heavily on the flood to deliver food (Winemiller 1991; FEOW Ecoregion 556).

The habitat is best read as a gradient from moving water to still. The main channel is the realm of swift current, deep sandy holes and high cut banks. Off it lie the permanent lagoons, oxbows, canals and slow tributary mouths, fringed with reed and submerged branches. Beyond those, every wet season, the floodwater climbs out onto a plain of grasses such as Echinochloa, Vossia and wild rice, with stands of Phragmites reed and papyrus in the deeper channels and lagoons. The marginal vegetation matters as much as the open water: it shelters small fish and juveniles, and in the dry season some lagoons vanish under floating mats of Salvinia. Between Nangweshi and Katima Mulilo, and again from Mombava down to the falls, the low-gradient river is broken by long stretches of rapids, and 300 kilometers above Victoria Falls the 21-meter Ngonye (Gonye) Falls forms the only other dry-season barrier to fish movement on the main river (FEOW Ecoregion 556; Roberts 1975; Winemiller 1991).

What ties it together is the flood pulse, and the chemistry of any given spot depends on where it sits in that cycle. At high water the river and plain become one continuous, well-oxygenated sheet; food is abundant and predators are spread thin. As the water falls, fish are squeezed back toward permanent channels and lagoons, isolated pools turn warm and acidic, and the whole fauna concentrates. Lozi fishers read this draw-down precisely, building 'maalelo' — grass-mound dams and reed fences with funnel traps — across the openings where water tries to drain off the plain back into the lagoons, intercepting the fish as the flood retreats. The plain is inundated for long stretches and exposed only briefly, so for most of the year the productive surface is under water (FEOW Ecoregion 556; Winemiller 1991).

The cichlid fauna

This ecoregion is the center of diversity for Serranochromis, the largemouth river breams, and its sister group Sargochromis, the smallmouth breams, and the roster reads as a near-complete cross-section of how those fishes carve up a floodplain. Winemiller's 1989 survey of the Barotse plain found the genus partitioned so cleanly that during low water there was almost no ecological overlap at all. The large, deep-bodied piscivores hold the main river channel; the rest work the lagoons and backwaters; and each has a job. All of them are mouthbrooders, and almost all spawn just before the flood (Winemiller 1991; FEOW Ecoregion 556).

The channel species are the giants. Serranochromis robustus, the nembwe or yellow bream of Lozi fishers, is a diurnal, bottom-hugging piscivore of swift current that grows past 16.5 in and was the only fish observed to specialize on armored Synodontis catfish, crushing their locking spines and bony heads with massive jaws and pharyngeal plates. Serranochromis altus, the hump-back bream described by Winemiller and Kelso-Winemiller from this very system, is a crepuscular hunter of nocturnal mormyrids taken from mid-water and the surface. The roster's other largemouths sort by habitat and prey: S. macrocephalus, the abundant purple-headed njenja of the lagoons, eats mostly nocturnal mormyrids and smaller cichlids; S. angusticeps, the thin-faced mushuna, ambushes passing fish from within weed beds; S. jallae and S. longimanus round out the piscivore guild, the latter uncommon and tied to floodplain lagoons and backwaters. The smallmouth Sargochromis trade fish for harder, slower food. S. giardi, the prized syeo, is a river-dwelling molluscivore that sifts bivalves and caddisfly larvae off the sand; S. carlottae, the rainbow bream, takes aquatic insects and scales from cover; S. greenwoodi and S. mortimeri extend the invertebrate-feeding line. The roster's two haplochromine river breams, Pharyngochromis acuticeps and P. darlingi, are smaller cichlids of the same Zambezian stock.

Beyond the breams the roster carries the floodplain's workhorse cichlids. The threespot tilapia Oreochromis andersonii and the greenhead Oreochromis macrochir are large, herbivorous-detritivorous maternal mouthbrooders that anchor the commercial catch; radio-tracked andersonii in the Caprivi reach moved between the main channel and floodplain with the seasons, exactly the migratory strategy the ecoregion selects for. Coptodon-group substrate spawners are present in the redbreast tilapia's relatives Tilapia sparrmanii, the hardy, widespread banded tilapia, and Tilapia ruweti, the small Okavango tilapia restricted to the upper Zambezi, Okavango and the southern Congo fringe. The dwarf southern mouthbrooder Pseudocrenilabrus philander fills the small-fish niche in vegetated margins. The shape of the whole fauna follows the ecoregion's geography: it shares most of its species with the Okavango and Kafue floodplains, the legacy of a single 'palaeo-Zambezi' drainage later captured and split, and the breams' local diversity is the signature of that isolated arena above the falls (Winemiller 1991; FEOW Ecoregion 556; Skelton 1994; Marshall 2000).

Conservation

The Barotse fishery is one of the most important inland fisheries in Zambia, and it is in trouble. It supports around one million people directly and indirectly, and fish supply more than half the animal protein in Zambian diets — local consumption on the floodplain runs several times the national average. Roughly 80 fish species are recorded from the system. Yet across three decades catch per unit effort has fallen from about 3.1 to 1.2 tonnes per fisher as the number of fishers climbed, and the historic yield of perhaps 2,000 to 3,000 tonnes a year is sustained now mainly by adding more boats to a shrinking resource. The decline is poorly studied but widely pinned on overfishing and weak management rather than any single environmental shock (WorldFish; ICCA Consortium 2016; academic value-chain study; Marshall 2000).

Gear is at the center of it. The shift from multifilament to monofilament gillnets, which are on average three times more effective, multiplied each fisher's reach; drifting gillnets and sefa-sefa dragnets sweep channels indiscriminately. The most damaging change has been mosquito-net fishing: fine-mesh bed nets, distributed for malaria control, are now dragged through floodplain shallows and nursery grounds, taking juveniles and small species wholesale and forcing fishers to target ever-smaller fish as the larger breams thin out. The big, late-maturing channel cichlids are the most exposed. Winemiller documented Sargochromis giardi, the highly sought syeo, being fished down in the central floodplain within a single season because it sits in accessible deep-shoreline water, while the swifter-water nembwe stayed partly out of reach. Serranochromis species mature slowly — most of the roster's large breams at around three years — which leaves them little margin against rising effort.

What has buffered the fishery so far is the flood itself. Because most breams spawn just before the water rises and disperse their mouthbrooded young across a plain where predators are scarce, the fish repopulate overfished lagoons each wet season as the flood drains back. Zambia leans on that rhythm with an annual fishing ban, closing the Zambezi and other rivers from 1 December to the start of March to protect spawning; enforcement is uneven, and patrols still seize illegal catches. Newer pressures stack on top of the old ones. Agriculture is expanding onto the plain as droughts make rain-fed farming less reliable, drawing down water and adding runoff; the broader Zambezi basin faces continued damming and flow regulation; and climate variability is making the floods that drive the system less dependable. Across the border, Fish Protection Areas in Namibian conservancies, with trained community guards, show that locally enforced no-take refuges can work — and most observers now agree that community control of fishing practice, not top-down rules alone, is the realistic path to keeping the Barotse breams in the water (WorldFish; ICCA Consortium 2016; Larsen et al. 2018; Lusaka Times 2026; Marshall 2000).

Sources

  1. Upper Zambezi Floodplains — Freshwater Ecoregions of the World (FEOW Ecoregion 556; boundaries, hydrology, Barotse floodplain, endemism)
  2. Comparative ecology of Serranochromis species (Teleostei: Cichlidae) in the Upper Zambezi River floodplain (Winemiller, Journal of Fish Biology 39, 1991)
  3. Serranochromis altus, a New Species of Piscivorous Cichlid from the Upper Zambezi (Winemiller & Kelso-Winemiller, Copeia 1991)
  4. Challenges in fisheries management in the Zambezi, one of the great rivers of Africa (review; Barotse fishery decline)
  5. Physical losses of fish along the value chain in Zambia: a case study of Barotse Floodplain (Int. J. Fisheries & Aquaculture; CPUE 3.1 to 1.2 t/fisher, 80 species)
  6. Widespread mosquito net fishing in the Barotse floodplain (Larsen, Welsh, Mulenga & Reid, PLoS ONE 2018)
  7. Empowering Zambezi communities to manage their own fisheries (ICCA Consortium, 2016; historic yield, monofilament nets, Fish Protection Areas)
  8. Improved Fisheries Management in the Barotse Floodplain of Zambia (WorldFish Center report)
  9. Zambezi fisheries crackdown nets 200kg fish during the annual fishing ban (Lusaka Times, 2026)
  10. Tilapia ruweti, Okavango tilapia (FishBase; upper Zambezi / Okavango distribution)

Cichlids recorded here

← All cichlid waters