Shifting is in a river’s nature. But when a river breaks free of its channel and carves a new path across the landscape, devastating floods may descend upon communities with little to no warning.
For decades, researchers have struggled to explain exactly how river channels become primed for such sudden diversions, or avulsions. A study published September 18 in Nature may have finally quelled the debate, showing how two factors work together to stage the rerouting of a river. Building on their findings, the researchers also developed a promising algorithm that can predict the new path of a river that has avulsed.
“These are monumental floods, civilization-changing floods in some cases,” says sedimentologist Douglas Edmonds of Indiana University in Bloomington. In 2010, avulsions on the Indus River in Pakistan contributed to flooding that forced roughly 20 million people from their homes. Nonetheless, flood hazard models remain unable to predict where rivers will reroute, Edmonds say. “It’s really an invisible flood hazard.”
Avulsions require a setup and a trigger — an overburdened camel’s back and a final straw (SN: 6/28/24). “The trigger could be a flood, an earthquake, it could be a logjam in a river,” Edmonds says. The setup refers to how the deposition of sediment has primed a river for diverting — and it’s the fundamental cause of avulsion, Edmonds says. “Rivers experience floods all the time, but they don’t avulse all the time.”
The new study focused on defining the setup, for which there had been two competing hypotheses. One held that avulsions happen when a river becomes superelevated, or the deposition of sediment raises a river’s water level above the surrounding land. The other contended that avulsions occur when there is a slope advantage, or once the slope of a new, potential path becomes steeper than that of the river’s current path.
Edmonds and his colleagues began by using satellite data to investigate roughly 170 avulsions, noting how far downstream rivers tended to divert. Avulsions were roughly three times as common near river mouths and mountain fronts than in between, they found.
Focusing on 58 river channels for which high-resolution topographical data were available, the researchers measured the superelevation and slope advantage prior to avulsions. They found that superelevation best explained avulsions near the mountains, while slope advantage best explained those near river mouths and deltas.
There’s so much sediment flowing out of the mountains that the rivers just pile it up until they’re superelevated and spill over, Edmonds says. Meanwhile in deltas, there’s a lot of cohesive mud that forms very steep natural leaves around deep channels, and avulsions need a steep slope advantage to start cutting through the levee, he adds.
These two factors — slope advantage and superelevation — work together in an inverse fashion, the researchers found. The more superelevated a river becomes, the less of a slope advantage it needs to avulse, and vice versa. “It is the first time that anyone’s been able to show that with data,” says Penn State geologist Elizabeth Hajek, who was not involved in the study.
Avulsions occurred when the mathematical product of the two factors surpassed a threshold value, the researchers found. So long as precise topographical measurements of a river’s channel are available, which is more likely for larger rivers and in places with clear skies, you could probably use that threshold metric to identify where avulsions are likely to occur, says geomorphologist Vamsi Ganti of the University of California, Santa Barbara, who was also not involved in the study.
The researchers developed a computer algorithm that highlighted where on a map an avulsed river might go, factoring in the steepness of the terrain and momentum of the river. When tasked with predicting the pathways of 10 past avulsions, the algorithm correctly captured the path of each one. “It’s a really nice tool,” Hajek says. “It could be really, really helpful for identifying areas of concern.”
The plan is to develop avulsion hazard maps for the globe or vulnerable regions, Edmonds says. “Now that we have this metric, we can go measure it on rivers all over all over the world.”