Bowing
Sometimes, rivers suddenly jump their banks, abandoning channels and washing over floodplains in search of new paths. These rare events, called avulsions, can cause catastrophic flooding and threaten lives and livelihoods worldwide.
Avulsions are difficult to predict—the exact factors that cause a river to quickly shift course are not well understood. A recent study published in Nature identifies physical rules that indicate why certain rivers avulse and offers a blueprint to gauge what path an avulsed river might take.
“Scientifically, we’re on the cusp of understanding river avulsions in a way we’ve not been able to for the last century.”
“Scientifically, we’re on the cusp of understanding river avulsions in a way we’ve not been able to for the last century,” said geomorphologist Douglas Edmonds at Indiana University Bloomington, one of the authors of the study.
Using satellite imagery, the researchers mapped avulsions around the world and confirmed an intuition the scientific community had long harbored: Avulsions tend to occur at sites of big topographic changes such as mountain fronts or near shorelines. Out of 174 mapped river avulsions since 1984, 74% occurred in these areas.
Eyes in the Skies
Studying the topography around rivers can be tricky. Rivers are often surrounded by thick vegetation, so traditional aerial or satellite imagery can’t provide visualizations of the elevations of banks or the shape of terrain. But laser beams can penetrate vegetation and bounce off Earth’s surface, creating a 3D model that can be used to measure elevation with a technique known as lidar. So when NASA launched the ICESat-2 satellite in 2018, equipped with advanced lidar technology to survey polar ice sheets and the rest of Earth’s surface, Edmonds and his colleagues realized that the agency had inadvertently created the perfect tool to study the topography around rivers.
Using ICESat-2 and digital elevation models, the researchers were able to get clear topographic data, including the height and slope of channels and banks, for 58 avulsion sites.
Rivers typically avulse for one of two reasons: Sediment builds up on the riverbed and elevates the river above its banks (known as superelevation), or the slope next to the river is steeper than the bed, offering a faster path downhill (known as slope advantage).
“That’s not something that was expected.”
Scientists had thought that both mechanisms could independently cause avulsions. After analyzing the ICESat-2 data, however, the authors concluded that in fact, these two factors operate together but play different roles depending on the river’s location. Near mountains, superelevation drives avulsion, whereas in coastal areas, avulsions are caused predominantly by slope advantage.
Superelevation and slope advantage “seemed to scale opposite. And the only way that works is if they’re inversely related,” said James Gearon, a doctoral candidate also at Indiana University Bloomington and the study’s lead author. Where the value of superelevation was high, slope advantage was low and vice versa.
“That’s not something that was expected,” said geomorphologist Vamsi Ganti of the University of California, Santa Barbara, who was not part of the study. “But the authors make a really good case for it, and it’s very data-driven,” he added.
The researchers define a new avulsion criterion as the mathematical product of the two metrics.
Showing the Path
Flooding triggered by avulsion is different from other kinds of flooding. “It can be quite significant in terms of its volume and its duration because the river has to reestablish its pathway,” Edmonds said. “It can take a long time.” Predicting the river’s new channel before avulsion begins is key to limiting damage and casualties.
To replicate a river’s pathfinding behavior, the researchers used probabilistic modeling and developed an algorithm incorporating two factors: inertia and slope. Their algorithm selected a path that favored the greatest slope and the least change in direction.
“What we showed is that just based on those two simple rules, inertia and slope dependence, you can totally re-create the direction that the river goes,” Gearon said. When tested on 10 real-life avulsions, the paths the algorithm predicted overlapped almost entirely with the observed paths.
The model could be used to predict future avulsion pathways, but “the incredibly important piece of information that we cannot yet predict is, along a given river, where is the next avulsion going to be?” Edmonds said.
Until recently, data on river avulsions were limited to small experiments and a handful of field observations, said Ganti, who has documented avulsions around the world. “Now, we actually have a big dataset that spans the entire globe.” The ICESat-2 data that the new study uses could “potentially be a game changer,” he added.
The authors hope their work will help mitigate avulsion-related hazards, especially in the Global South, where avulsions are more frequent but disaster management resources are fewer. They’d also like to see future flood models account for avulsion. “Flood models are missing this important piece, but I think our understanding of avulsions is catching up,” Edmonds said.
—Sushmita Pathak (@sushmitza), Science Writer
