Dams are often built to control floods, but on some types of rivers they can make major floods worse, according to a new study. The finding suggests river managers may need to rethink their flood control strategies on silty and sandy lowland rivers.
“It’s a counterintuitive finding,” says Gordon Grant, a US Forest Service hydrologist and geomorphologist who was not involved in the work. “What this provocative document is showing us is that we don’t fully understand” how dams affect floods, he says.
Dams offer multiple benefits. They can generate relatively clean electricity; they store water and release it in the dry season, which helps farmers and other users, and can hold back flooding. Dams also have drawbacks, such as displacement of people when built, prevention of fish migration and other ecological damage. But no one had suggested that the dams could make the floods worse than before a barrier was built, says Ellen Wohl, a river scientist at Colorado State University, Fort Collins.
In addition to storing water, engineers expect the dams to reduce flood risks by modifying the river downstream. As dams trap sediment, they release relatively clear water that penetrates deeper into the river bed. This incision creates a more spacious channel that can carry more water and prevent floodwaters from spilling onto the river banks.
Hongbo Ma, a geomorphologist at the University of California (UC), Irvine, was interested in how dams alter sediments typical of river beds and this led to the surprising result of his study. The phenomenon involves erosive water released from a dam, which removes more finer particles and leaves larger grains. This swelling of the river bed then creates underwater dunes.
Ma and his colleagues were studying the Yellow River, which flows from the Tibetan Plateau mountains to the East China Sea. As they used sonar to scan the riverbed of the lower course, they were struck by the absence of dunes. This is probably due to the high silt content of the river bed; the Yellow River, the muddiest in the world, takes its name from the heavy loads of silt it carries. Fine particles hinder the formation of dunes. But closer to the Xiaolangdi Dam, the river bed was coarser and had large dunes. “We have a truly amazing example of how bed roughness can change so much,” says Ma.
But he wondered if the rougher, dune-filled river bed would prevent the floodwaters from flowing up, overflowing the riverbanks and spilling out onto the floodplain. To test the idea, he and his colleagues first performed calculations based on the shape of the river channel and other factors. The findings suggested that major floods are now about twice as deep as they were before the dam’s construction in 1999, despite a channel that is up to 3.4 meters deeper, the team reports this month in Nature communications.
By checking the flood logs from 1980 to 2015, the scientists found that the magnitude of moderate and large floods had actually increased. Over the same period, however, the magnitude of minor floods has decreased, probably because the deeper channel of the river is better able to contain them.
Fortunately, major floods have rarely hit the Lower Yellow River since the dam was built. This is because the climate has become drier and the reservoir still has enough capacity to hold back the major flows that come from extreme storms. But climate models suggest that rainfall in the Yellow River basin will increase by up to 30% this century. And because the river continues to dump sediment into the reservoir – it is already 75% full – the dam will have less space to hold floodwaters.
“We need to start thinking about how to reassess the risk of flooding, not only for the Yellow River, but for other rivers as well,” says Ma. The team estimates that the new dams would aggravate major floods on more than 80% of the rivers in plains.
The Yellow River may be an unusual case, notes Cardiff University hydrologist Michael Singer, because most large dams are built on rivers carrying coarser sediments that don’t easily form dunes. However, it would be worth seeing if a similar phenomenon is occurring on other rivers that carry fine sediments, known as alluvial rivers, and have large dams, says Nicholas Pinter, a UC Davis geologist. In the United States, such rivers include Missouri and Tennessee. If so, engineers may need to recalculate local flood risks.
More generally, engineers should also pay more attention to the complex behavior of rivers when designing new dams, says Pinter. “It’s amazing how wrong we have been in thinking these great flood rivers are just pipes,” he says. “We continue to underestimate the importance of bed shapes and roughness.”