• News
  • Learning in Machines & Brains

The underlying forces behind river formation

by CIFAR Dec 19 / 12

Until recently, the precise mechanisms that govern the branching of rivers were a mystery. But now, a team of geologists, including CIFAR Scholar Taylor Perron (MIT) and newly appointed Junior Fellow Ken Ferrier (Harvard), have developed a model to understand two opposing forces controlling river formation.

Branching river networks in the Allegheny Plateau in Pennsylvania and West Virginia.
IMAGE: Taylor Perron

One force is the flow of water that erodes rock or soil, which causes river valleys to deepen and widen. The second is the movement of soil down valley slopes, which fills in the eroded landscape and prevents the valley from expanding.

For the first time ever, the team discovered that there are “tipping points” among the two forces that determine whether a single river channel in a valley branches into a network of channels or remains as one. The team tested their model on rivers in two regions in the U.S. and found that the model’s predictions were accurate. Their findings were published in the prestigious journal Nature.

“Our study teaches us how to read Earth’s history as it is written in the landscape,” explains Perron. “River networks are also one of the most widespread features of erosion in the Solar system, and we are intrigued by the idea of using our model to learn something about landscapes on places like Saturn’s moon Titan.”

The team modeled the tug of war between the two opposing forces of erosion. When they tested the model on a simulated landscape, they found that a fine balance existed between the two forces. For example, if they increased erosion from water flow past a very specific point, the river valley grew at the expense of neighboring valleys. The resulting widening of the river valley triggered a second tipping point, at which the valley’s single river channel branched into a network of rivers.

The team predicted that a river would form branching networks more easily in a place with more runoff or soft rock. They tested their prediction with the model in two locations with different environments: the Allegheny Plateau, in southwest Pennsylvania, and Gabilan Mesa, in California’s Salinas Valley. The Gabilan Mesa has a drier climate, softer rock and less permeable soil than the Allegheny Plateau. This means that when it rains hard in the Gabilan Mesa, water accumulates faster and erodes the landscape more easily, leading to more extensive river branching at a finer scale.

“We found that our model accurately predicted the way the rivers branched in the two locations,” says Perron. “We also showed how river networks are shaped by different rock types, climates, and biological communities. Understanding how river networks originate and change is key to understanding how landscapes have evolved in the past, and will help us make more accurate predictions of what may happen in the future.”