There are several lines of evidence that the dams failed, including bits of volcanic glass found embedded in the canyon walls. When the lava hit the river, the cold water was "hydro-explosive," she says, cooling the basaltic lava so quickly that it formed glass that in turn exploded into fragments.
More lava poured on top of the fragments until the dam was formed, sometimes spanning the entire canyon. But glass, of course, isn't the best building material for a dam.
"That isn't a dam that's constructed by people," Fenton says. "It's a natural dam. There's been no engineering behind it."
As she searched through the canyon, she found glass deposits of the type that were formed when the hot lava hit the cold water, and the deposits were scattered everywhere. That suggests, quite strongly, that the dams failed from the bottom up.
"We think there's evidence in our flood deposits that suggest the dams actually failed near the base," she says.
"These were high dams," adds hydrologist Robert H. Webb of the Geological Survey. "We estimate some were more than 1,500 feet tall."
When a dam fails, the water doesn't all come out in a steady stream. The water level is much higher near the dam than farther downstream, resulting in something called an "exponential decay curve." Evidence of that curve was found in the canyon walls, and the record there is almost a perfect match with the record left behind when Idaho's Teton Dam failed in 1976.
"We have that curve preserved from a lava dam that failed in the Grand Canyon 165,000 years ago," Webb says.
Big and Getting Bigger
Webb, by the way, is a leading proponent of the idea that the canyon was formed somewhat intermittently, rather than at a steady pace. The end of a glacial period would send higher amounts of water down from the Rocky Mountains, for example, allowing the river to carve much more rapidly than during drier periods.
But that doesn't mean the river is through with its work. Geologist Joel Pederson of Utah State University, Logan, argues in the August issue of the journal Geology that scientists were wrong in thinking that the Grand Canyon was "largely finished" about 1.2 million years ago.
Pederson says his research shows the erosion is continuing at "healthy" rates, even today.
Various scientists are now concluding that some of the canyon's most famous features, including the colorful Marble Canyon are much younger than the earlier estimates of 3 million to 5 million years old. They may be as young as 600,000 years old, based on newer dating techniques.
That's a mere child, as geological features go.
Now we know it's had more than a few temper tantrums. Some years have been harder than others. And it still has some growing to do.
But it's hard to imagine that it could get any better.
Lee Dye’s column appears weekly on ABCNEWS.com. A former science writer for the Los Angeles Times, he now lives in Juneau, Alaska.