Leopards in Nyerere National Park (Selous): What the Science Actually Says

Ask most safari operators how many leopards live in Nyerere National Park and you will get a confident number. Ask a scientist, and you will get a longer, more useful answer.

Until recently, nobody knew. The leopard (Panthera pardus) is the most widely distributed big cat in Africa and one of the least well counted. It is solitary, largely nocturnal, and lives at low densities across enormous areas — the precise combination of traits that makes a species hard to survey and easy to guess about. For a park the size of Belgium, guessing had been the norm.

That changed with a study published in early 2026 in Ecological Solutions and Evidence, which produced the first spatially explicit leopard density estimates for the Selous–Nyerere ecosystem. It is worth understanding what it found, and equally worth understanding what it did not.

The ecosystem in question

The old Selous Game Reserve, protected since 1896 and gazetted in 1922, was for a century the largest protected area on the African continent — more than 50,000 km² including buffer zones, and a UNESCO World Heritage Site from 1982.

In 2019 the Tanzanian government split it. The northern and central photographic section, roughly 30,893 km², became Nyerere National Park under TANAPA administration. The remaining ~20,000 km² to the south and east continued as Selous Game Reserve under TAWA, retaining its hunting blocks.

The two together — referred to in the literature as the Selous–Nyerere ecosystem — remain one of the largest intact wilderness areas on Earth. The habitat is a mosaic: miombo woodland across most of the interior, with Vachellia (Acacia) savannah, grassland, riverine forest and seasonal swamp along the Rufiji and its tributaries, including the Beho Beho.

That mosaic matters for what follows.

How you count an animal that does not want to be counted

The method is spatially explicit capture–recapture, or SECR, applied to camera trap data. It exploits a convenient fact of leopard biology: every leopard's rosette pattern is unique and stable through life, and the pattern differs between the left and right flank. Paired cameras facing each other across a trail capture both flanks simultaneously, which allows individual identification without ever touching the animal.

From 2020 to 2022, a team led by Charlotte Searle of Lion Landscapes and the University of Oxford's WildCRU, working with the Tanzania Wildlife Research Institute, Frankfurt Zoological Society, TANAPA and TAWA, ran seven camera trap surveys across the ecosystem. Grids of paired stations were set roughly 3–5 km apart, each survey running for two to three months. Across the whole effort, more than 4,600 km² was surveyed and 373 individual leopards were identified.

SECR then models where each individual's activity centre sits and how detection probability declines with distance from it, producing a density estimate with genuine confidence intervals rather than an extrapolated guess. It is the current methodological standard for big cat monitoring, and it is why these numbers can be taken seriously.

The results

Density is reported as adult and subadult leopards per 100 km².

Sector	Protected area	Density (per 100 km²)	SE
Matambwe	Nyerere NP	8.08	± 1.54
Miguruwe	Selous GR	7.38	± 1.26
Msolwa	Nyerere NP	6.05	± 0.78
Liwale	Selous GR	5.93	± 0.88
Kingupira (west)	Selous GR	5.58	± 0.87
Kingupira (east)	Selous GR	5.22	± 0.71
Kalulu	Nyerere NP	3.80	± 0.64

Leopard density by survey sector, Selous–Nyerere ecosystem (Searle et al. 2026). Values are adult and subadult leopards per 100 km², ± standard error.

The range across the ecosystem is therefore roughly 3.8 to 8.1 leopards per 100 km², with 95% confidence intervals spanning 2.73–5.29 at the low end and 5.57–11.71 at the high end.

Two things stand out.

First, these are high numbers for this habitat type. The authors note that these are the highest leopard densities yet documented in miombo woodland — a biome that accounts for close to one fifth of the leopard's remaining African range. Miombo is nutrient-poor and supports lower ungulate biomass than the mesic savannahs of southern Africa where most leopard density work has been done, so the expectation had been for sparse populations. The data say otherwise, at least here.

Second, the highest density recorded anywhere in the survey was in the Matambwe sector, in the north of Nyerere National Park — the prey-rich country of the northern lakes and the Rufiji tributary system. This is the sector of the park most travellers actually visit.

The more interesting finding

Run the leopard numbers alongside the lion numbers from the same camera trap grids, published by the same team in 2025, and something odd appears.

In Matambwe, lion density was 6.27 individuals over one year of age per 100 km² — the highest in the ecosystem. Leopards there were also the highest, at 8.08.

But in the dry eastern Miguruwe sector of Selous Game Reserve, lion density collapsed to 0.33 per 100 km², the lowest recorded anywhere in the complex. Leopard density in the same sector was 7.38 — the second highest.

Lions and leopards, in other words, are not tracking together. The study found leopard density strongly correlated with the relative abundance of preferred prey, whereas lion density was not correlated in the same way. The authors are careful here — the sample is seven survey sites, which is a small number of data points for inference — but the implication they draw is that the two species are not being uniformly affected by the same anthropogenic pressures. Leopards, with their broader diet, smaller prey preference and greater tolerance of disturbance, appear able to hold density in places where lions cannot.

Anyone who has spent time in the southern circuit will recognise the pattern intuitively. It is useful to see it quantified.

One more thing the cameras found

Among those 373 identified individuals was a single female with a rare colour morph: a "strawberry" or erythristic leopard, in which the normally black rosettes appear red or brown. The condition is thought to arise from a mutation in the TYRP1 gene.

She was recorded in Selous Game Reserve, and she represents the first documented record of this phenotype anywhere on the African continent outside South Africa — one animal in 373, or 0.3% of the population sampled. She was published as a short note in Ecology and Evolution in 2024.

It is a small thing. It is also a good illustration of what large-scale camera trap work turns up incidentally, and of how thin our baseline knowledge of this ecosystem still is.

What this does not tell us

Here is where honesty is required, because this is the point at which most safari websites reach for a headline figure.

These studies do not produce a total leopard population for Nyerere National Park, and the data do not support inventing one.

The reasons are straightforward. The surveys covered roughly 4,600 km² of an ecosystem exceeding 50,000 km². Density varied more than two-fold between the sampled sectors, and it varied in a way that correlates with prey abundance — which is itself patchy and unevenly distributed across the landscape. Multiplying any single density figure by the total park area would produce a number that is arithmetically simple and ecologically meaningless. The authors explicitly frame their estimates as a baseline for future monitoring, not as a census.

What can be said with confidence is this: the Selous–Nyerere ecosystem holds a relatively high-density leopard population, it constitutes an important stronghold for the species in Africa alongside its established importance for lion and African wild dog, and there is now a defensible reference point against which future surveys can measure change.

The threats

The study identifies two principal pressures.

Habitat conversion in boundary areas, which is accelerating. The ecosystem's edges — particularly where agricultural land meets the protected area — are where leopard density would be expected to decline first. Earlier work in the nearby Udzungwa Mountains found leopard density increasing with distance from the protected area boundary, a proxy for both reserve extent and distance from human settlement.

Bushmeat poaching, which affects leopards two ways: indirectly, by suppressing the prey base that the density data show leopards depend on, and directly, through accidental capture in wire snares set for ungulates. Snaring is indiscriminate, and a snared leopard is usually a dead one.

Neither threat is dramatic in the way that ivory poaching is dramatic. Both are more corrosive over time.

What it means if you are going to Nyerere

A word of caution against a natural but incorrect inference: high density does not translate directly into high sighting rates.

Eight leopards per 100 km² is a genuinely good figure — but the animals are nocturnal to crepuscular, solitary, and this is thickly vegetated country for much of the year. Leopards here are not habituated to vehicles the way they are in parts of the Serengeti or the Sabi Sand, where decades of concentrated tourism have produced cats that ignore Land Cruisers entirely. Camera traps see far more leopards than people do, which is precisely why the method was used.

What the science does support is this: when you are in northern Nyerere, you are in country holding one of the densest leopard populations recorded anywhere in the miombo biome. The cats are there in numbers. Whether they choose to show themselves is a separate question, and the answer usually involves dawn, dusk, riverine thicket, and patience.

Where Wanyamapori Porini Camp sits in this picture

Wanyamapori Porini Camp stands on the Beho Beho River, in the Matambwe sector in the north of Nyerere National Park — the same block of country in which the 8.08 figure was recorded. That is a statement about location, not a promise about sightings, and the distinction matters. What it does mean is that guests here are on ground holding the highest leopard density documented anywhere in the miombo biome.

It also shapes how we guide. Density this high announces itself in signs long before it produces an animal: fresh tracks on the road at first light, a drag mark heading into a thicket, baboons or impala calling at something moving ahead of you that never shows itself. On a walking safari — which is how the country around Beho Beho is best covered — reading those signs is much of the point. A leopard you did not see is still evidence, and a guide who knows the ground will tell you what passed through in the night.

References

• Searle, C.E., Strampelli, P., Parsais, S.N., Haule, L., Olesyapa, K., Salum, N.D., Mtoka, S., Hape, G., Mathayo, D., Elisa, M., Lobora, A.L. & Dickman, A.J. (2026). Population status of leopard in one of Africa's largest wilderness areas and the challenge of monitoring at scale. Ecological Solutions and Evidence.

• Searle, C.E. et al. (2025). Spatially explicit camera trap-based lion monitoring in Tanzania's Selous–Nyerere landscape. Journal of Zoology.

• Searle, C.E. et al. (2024). New record of strawberry leopard (Panthera pardus) in Selous Game Reserve, Tanzania. Ecology and Evolution, 14(7).

• Rovero, F. et al. (2019). Reserve size and anthropogenic disturbance affect the density of an African leopard (Panthera pardus) meta-population. PLOS ONE, 14(2): e0209541.