How large a flood could occur on the Red River?

Guest post by Phil Gerla, Geology & Geological Engineering

As we ride out the 48-foot crest of the 2019 Red River flood, the communities of Grand Forks and East Grand Forks can relax a little.

As described in an earlier post by Fred Remer (“The perfect melt“), the conditions since mid-March have been ideal for curtailing the spring flood — temperatures generally below normal, freezing at night that slows melting during the day, few storms, dry air masses that have sublimated the snow and sopped up much of the melt water.

In addition, the Corps of Engineers’ extensive flood protection work almost 20 years ago has provided great relief to our cities, which are now protected to 60 feet. Prior to the Corps’ work, a flood of 48 feet would have meant rapid construction of temporary clay dikes and a need for tens of thousands of sandbags . In fact, the 1979 flood, which reached only a few inches higher (48.8 feet) than this year’s crest required an extensive flood fight.

As a geologist, I tend to think in terms of deeper time and wonder about the nature of the Red’s spring runoff events since glacial Lake Agassiz drained from the Red River Valley about 9,000 years ago. We have, at best, a 150-year direct record of floods along the Red River, perhaps going back a little longer if we include the anecdotal observations from Winnipeg and old Fort Garry.

Clearly, there is only a miniscule record of river flows, covering less than 2% of the Red River’s history. Nevertheless, there have been many floods that we would consider as “large” during that short record, and as we experience large floods, their statistical recurrence time keeps getting shorter. It suggests that large floods, like that of 1997, may not be particularly uncommon during the millennial history of the Red.

It raises the question: Combining all the perfectly bad conditions, just how large a flood could occur? Imagine an unusually wet summer and fall, soils fully saturated going into the winter, and an early, hard, and deep freeze sets in, rendering the ground impervious to spring melt. In addition, the region is hit with multiple blizzard “bomb cyclones” all winter and spring arrives with a vengeance — temperatures climb abruptly, melting the snow. To make matters worse, widespread heavy convective precipitation occurs at the same time. A nightmare scenario that’s highly improbable. But possible? Yes.

In 2008, John Costa and Robert Jarrett, two world-class flood experts, published a fascinating report that describes their thorough analysis of 30 extreme floods that occurred in the U.S. during the mid-20th century. They pored through the U.S. Geological Survey and local agency records, many of which were never published and remained hidden in dusty filing cabinets. Their goal was to find and characterize exceptional floods for watersheds of all sizes, from a square mile or two to hundreds of thousands of square miles, and see if there is any consistent upper limit on flood size. After either verifying, correcting, or refuting data, Costa and Jarrett tabulated and graphed the results, which showed a clear trend in the upper bound for these monster floods. Their graph, reproduced here, shows records from all over the nation.

The contributing area or watershed that lies above Grand Forks is roughly 26,000 square miles. Interpolation of the graph suggests that one of these extreme events in the Red River basin would result in a peak discharge of about 1,000,000 cubic feet per second (cfs). That’s roughly an astounding seven times the peak flow estimated for the 1997 flood (137,000 cfs). A humbling observation!

A wide range of landscapes and climates covers the U.S., so each region may have its own characteristic curve. Perhaps the Red’s is not so bad? True, but the smooth pattern on the graph suggests that differences across the country do not greatly influence these extraordinary floods . Furthermore, our place in the mid-continent with its crazy weather might mean that our region’s curve lies a little higher compared to other areas. Regardless, an extreme event would certainly be devastating to the region, as it would be anywhere. The chances of its occurrence are extremely remote, fortunately, but underscore the fact that flood protection is not flood proofing —  the potential for future large floods means we need to stay aware and vigilant.

Costa & Jarrett’s 2008 report can be accessed here: