← Latest papers
📄 earth_science

Causes of the Desiccation of Lake Victoria at 17 ka

This study uses a high-resolution triple-nested model to demonstrate that the desiccation of Lake Victoria 17,000 years ago was primarily driven by regional climate responses to reduced greenhouse gases, cooler temperatures, and altered insolation, which collectively disrupted moisture convergence and collapsed the lake's characteristic precipitation cycles rather than by remote forcing.

Original authors: Kerry H Cook, Daniel Peppe, Edward Vizy, Patrick Andrews

Published 2026-07-13
📖 5 min read🧠 Deep dive

Original authors: Kerry H Cook, Daniel Peppe, Edward Vizy, Patrick Andrews

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine Lake Victoria as a giant, shallow bathtub sitting in the heart of East Africa. It's so wide and shallow that it's incredibly sensitive to the weather. About 17,000 years ago, this bathtub didn't just get a little low; it completely dried up, leaving behind a cracked, dusty floor. Scientists have been trying to figure out why the water vanished. Was it a distant storm in the North Atlantic pulling the water away? Or was it something happening right in the neighborhood?

A team of researchers used a super-powerful computer model to build a "time machine" simulation. They recreated the climate of 17,000 years ago with high precision, zooming in so closely they could see the tiny breezes swirling over the lake and the mountains. Their simulation suggests a very specific story: the lake didn't dry up because of a remote, far-away cause. Instead, it was a local neighborhood problem caused by a perfect storm of three specific changes in the air and water.

The Great Drying: A 51% Drop
In their simulation, the total amount of rain falling on the Lake Victoria Basin in that ancient era dropped by a massive 51% compared to today. The annual rainfall fell from 592 mm down to 289 mm. The researchers note that water budget models suggest if rainfall drops by just 25%, a lake like this will dry up within centuries. So, a 51% drop is more than enough to explain why the lake vanished.

The Culprits: Three Local Changes
The simulation identified three main "villains" that worked together to steal the rain:

  1. The Chilly Ocean: The water in the western Indian Ocean was cooler than it is today. Think of this like a cold front sitting just off the coast. This cooling changed the air pressure, creating a strong push of dry, easterly winds blowing right across the lake.
  2. The Thin Air: The atmosphere had much less greenhouse gas. Specifically, carbon dioxide (CO2) was only 203 ppmv (compared to 398 ppmv today), methane (CH4) was 421 ppbv (vs 1825 ppbv), and nitrous oxide (N2O) was 193 ppbv (vs 327 ppbv). This thinner air meant the high mountains around the lake cooled down much faster at night, creating a strong, dry wind that rushed down the slopes.
  3. The Sun's Shift: The Earth's tilt and orbit were slightly different, meaning the sun was about 3.8% stronger in the spring and 3.7% weaker in the fall.

How the Rain Got Stolen: The Day and Night Cycle
The magic (and tragedy) of the drying happened in the daily rhythm of the weather.

  • The Nighttime Disaster: Usually, at night, the lake is warmer than the land, creating a gentle breeze that pulls in moist air and creates a rainstorm over the water in the early morning (around 03:00–05:00 local time). But in the 17,000-year-ago simulation, the cooler ocean and the dry mountain winds broke this cycle. The dry air from the mountains and the strong easterly winds from the ocean pushed the moisture away from the lake instead of letting it gather. The nighttime rainstorm simply never happened.
  • The Afternoon Woes: During the spring afternoons, the sun heated the high mountains more than the lower land. This usually helps create rain, but in the simulation, it made the dry winds from the east blow even harder over the mountains and into the basin, drying out the land to the east of the lake.

What It Was NOT
The researchers were very clear about what didn't cause the drying. They explicitly argue against the idea that a distant event in the North Atlantic (like a massive iceberg melting) was the main culprit. While some other studies have suggested a link to the North Atlantic, this simulation shows that the drying was driven by regional responses to the local climate changes, not a remote signal from far away. They also found that the lake didn't dry up because evaporation increased; in fact, evaporation went down because the air was cooler. The problem was that the rain stopped coming in, not that the water boiled away faster.

The Seasonal Story
The drying wasn't the same all year. The biggest hit came during the spring rains (March, April, May), where rainfall plummeted by 72%, dropping from 305 mm to just 87 mm. The fall rains also suffered, dropping by 45%. Interestingly, the winter and summer seasons changed in opposite directions, but the spring and fall losses were so severe that they doomed the lake.

How Sure Are They?
The authors are quite confident in the spring drying results because their computer model ran multiple times and gave the same answer every time. The physics of the wind and moisture match up perfectly with the changes in temperature and greenhouse gases they programmed. However, they are a bit less sure about the fall season because the different runs of the simulation didn't agree as closely on the monthly details.

In short, this study suggests that 17,000 years ago, Lake Victoria didn't need a global catastrophe to dry up. A local combination of a cooler ocean, thinner air, and a shifted sun created a "dry wind machine" that blew the rain away, turning a massive freshwater lake into a dust bowl.

Drowning in papers in your field?

Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.

Try Digest →