Up to 122 million people around Lake Victoria could face dangerous heat by 2100

The Lake Victoria basin holds one of the densest rural populations in Africa, spread across the shorelines of Kenya, Uganda and Tanzania. A study published this month in Environmental Research Letters models what dangerous heat stress there looks like by the end of the century — and finds that the answer depends as much on how many people are living there as on how hot it gets [s1].

The method

The researchers used a convection-permitting climate model — a class of model that resolves storm-scale processes explicitly rather than approximating them, which matters over a large lake with strong local circulation — combined with two heat stress indices, humidex and the heat index, and high-resolution population projections [s1]. The scenario is SSP5-8.5, which the authors interpret explicitly as the high end of the climate change signal rather than a central expectation [s1].

That framing should travel with every number below. This is a deliberate upper-bound exercise, not a best estimate.

The projection

By the end of the century, up to 122 million people — around 44% of the region's projected population — may experience dangerous heat stress for more than 5% of the time annually, which the paper translates to roughly 18 days a year [s1].

The comparison point is what makes that figure land. For the 2005–2016 baseline period, about 1% of the population, roughly 1 million people, met that same threshold [s1].

At a higher threshold, up to 28% of the population — about 78 million people — would experience dangerous heat for 15% of the time, roughly 55 days a year [s1].

The decomposition, which is the more useful finding

The study attributes 66% of the increased population exposure to the combined effect of rising temperatures and rising total population in the region [s1]. Heat and demography are not separable contributors here; they multiply. More hot days over more people produces exposure growth faster than either trend alone.

That is the paper's central methodological argument, and the authors state it directly: assessments of heat stress need to consider both climate and population dynamics [s1]. A projection that holds population fixed at today's level would substantially understate exposure in a region where population is growing quickly.

Geographically, the model identifies the northern and southern shores of Lake Victoria and urban areas as the highest heat-risk zones [s1] — which is also where much of the population and most of the fishing and market economy sit.

How it fits a wider pattern

A Nature paper published a few days earlier projected global wildfire smoke mortality and reached a structurally similar conclusion about Africa. Under a moderate emissions pathway (SSP2-4.5), it projects 1.40 million premature deaths annually from wildfire smoke during 2095–2099 (95% CI 0.66 to 2.25 million), roughly six times current levels — with Africa seeing the greatest rise in fire-related deaths, an elevenfold increase [s2].

The mechanism the Nature authors give for Africa's outsized increase is the same two-part one: changes in emissions, and an ageing population [s2]. Two independent modelling exercises, on two different exposures, both find that the African health burden grows disproportionately because environmental change and demographic change compound.

What this is not

Neither result is an observation. Both are projections to the end of the century, and both inherit uncertainty from their emissions scenarios, their climate models and their population forecasts. The Lake Victoria study's "up to" phrasing is doing real work — 122 million is an upper figure under a high-end scenario, not a central expectation under current policy.

There is also a gap between exposure and harm. The Lake Victoria paper projects how many people experience dangerous heat stress by a threshold definition; it does not project deaths, illness or lost work. How much harm follows a given number of dangerous-heat days depends on housing, occupation, water access, health services and adaptation that the model does not attempt to forecast.

What to watch

The most consequential near-term question is whether heat-health surveillance in the basin improves enough to test these projections against observation. Modelled exposure is only actionable if there is mortality and morbidity data to check it against, and East Africa's heat-health records are thin compared with the European and North American datasets on which most temperature-mortality methods were developed.

Sources

• Projected population exposure to dangerous heat stress around Lake Victoria under a high-end climate change scenario — Environmental Research Letters, 2025-09-23
• Global warming amplifies wildfire health burden and reshapes inequality — Nature, 2025-09-18