This season Hurricane Harvey slammed Houston and surrounding southeastern Texas with torrential rains that broke records and created what felt to many like biblical-scale flooding. Harvey was an unusually wet and potent storm by today’s standards—and it may provide a glimpse of North America’s future. A new study predicts the continent will experience more storms that dump similarly huge volumes of rain by the end of the century, thanks to climate change.

For the study, published Monday in Nature Climate Change, researchers looked at how global warming will affect storms, called mesoscale convective systems, in North America. These are clusters of smaller storms that grow into one another and eventually act as one. Hurricanes can fall into this category, but the majority of such events in North America are actually giant thunderstorms. They mainly occur east of the Continental Divide, are relatively long-lived and span at least 60 miles (about 100 kilometers)—although they can grow large enough to cover a state as big as Kansas, according to Andreas Prein, the study’s lead author and atmospheric scientist at the National Center for Atmospheric Research.

These systems, however, do provide crucial water supply—about 60 percent of the central U.S.’s summertime rainfall, Prein says. “They are very important from the hydrological viewpoint, and also from their extreme perspective, because they’re related to things like flash flooding and hail,” he explains. “That’s why it’s really important we understand how these storms might change in the future.”

Until now it has been difficult for climate modelers to study how global warming might influence these storms systems in a precise way—standard models did not offer high enough resolution to simulate them. For their new study, Prein says he and his team developed better models and had more computational resources to realistically simulate the storms, with a resolution of four kilometers. “That’s 20 times finer than traditional climate models,” Prein says. They ran the models (in computer time) for 13 years in today’s climate as well as for 13 years in the predicted climate for the end of the century, if global temperature rises 4 to 5 degrees Celsius. That’s the “business as usual” scenario, which assumes greenhouse gas emissions will continue to rise until 2100.

When Prein’s team compared storm models for today versus the future, they found mesoscale convective systems are likely to triple in frequency in North America by century’s end. They also discovered these future storms will likely release even larger volumes of rain than today’s do. There are a number of reasons for this—for one, storm rainfall intensity will increase in a warmer world. The researchers found the U.S. Northeast and mid-Atlantic regions as well as parts of Canada will see a 25 to 40 percent increase in maximum rainfall rates, with a 15 to 20 percent increase in other parts of the continent. The team also found the size of these storms will likely grow, particularly in the southern U.S. Together those two factors could cause storms to release much more water—a 40 to 80 percent rise in rainfall volume for lower latitudes and 20 to 40 percent increase for middle and high latitudes.

The researchers offer a real-world example of what these numbers could mean: “We looked at all of the rain [that would fall] within an hour over New York City,” Prein says. “And you get about 60 percent more rainfall in this area in the future. [That’s the same as] adding six times the discharge of the Hudson River.”

To make matters worse, the models showed parts of Canada and the U.S. (in particular, the Midwest and mid-Atlantic) can expect more slower-moving storms, too. “This is really bad news,” Prein says. “Slow-moving storms mean that you will get another load of rainfall, because [the storm] doesn’t move out.” This is exactly what Houston experienced with Harvey; its plodding pace helped make the city’s floods so catastrophic, because it had plenty of time to drop huge volumes of rain.

Past studies have looked at climate change’s effect on storm rainfall rates but they have not as closely examined these storms as a whole—including their size, motion and overall volume. “Considering storm size in terms of both magnitude and spatial extent is particularly unique,” says Jeremy Pal, a professor of civil engineering and environmental science at Loyola Marymount University who was not involved with the study. “It suggests that the impacts of climate change on extreme precipitation and flood events may be severely underestimated in previous studies.” Prein feels similarly about the conclusions of his work. “We need to look at the storm as a whole,” he says. “What we [predict] for future flooding might not be extreme enough.”