We're talking catastrophic floods, twice in less than two years.

“Yet more rain and dangerous flooding are affecting parts of the Midwest and South tonight.”

“Take you across the Meramec River, notorious for flooding very quickly.”

“Towns like Eureka, and others, well they're in for a rough few days.”

But, just 12 miles downstream is a town called Valley Park.

It stayed pretty dry thanks to a levee.

A giant artificial embankment that surrounds the city, designed to keep water out.

Valley Park got their levee  largely because the city was willing and able to pay for it.

It kept their own town dry but here's what flooding looked like in neighboring towns,

including Eureka, that didn't have levees.

“The valley Park levee withstood the water.”

“That levee holding could be part of the problem.”

“They suspected the same levee that protected Valley Park may have been to blame for devastating

flooding in Fenton, Eureka, and Arnold.”

Can a levee protect one town while making flooding worse for others -- especially towns

that can't afford a levee?

A fluid mechanics lab, a 13-foot long model of a river, and some adorable tiny houses

will help us find out.

People love levees.

There's about 100,000 miles of these embankments across the US.

Sometimes concrete, sometimes earthen for centuries, we've built them between humans

and rivers.

We even write songs about them.

But it turns out, even when levees are effective, they can still be devastating.

As far back as 1852, levees came with a very important warning.

Charles Ellet Jr. - a famed US civil engineer - cautioned that levees confine rivers and

cause them to "rise higher and flow faster."

And relying on levees “encourages a false security.”

But he was largely ignored.

Levees became the default for flood control - mostly funded and constructed by local entities.

Some riverside communities that could afford it, built taller levees for more protection,

while less fortunate neighbors dealt with the devastating effects.

To show you what we mean we went to the banks of the Mississippi River.

To a fluid mechanics lab, where a team of engineers from the University of Minnesota

built us this landscape model to test flood scenarios.

Overhead, a scanner collects 3D data to measure exactly what's happening.

It's a generic model of a river with no levees.

That means, when the water level increases, it overflows.

And spreads across the floodplain, often creating important wetland habitat that's home to

a variety of species.

Putting in levees cuts rivers off from this land, destroying floodplains and wetlands.

It allows people to convert these areas into farmland or build houses on them.

And while levees protect these communities from flooding, they constrict the river into

a narrow channel, making the water flow faster and higher.

That creates a bottleneck leading to additional flooding upstream.

If all the levees are the same height, both sides should be about equally protected from

the average flood.

And if the river rises so much that the water overtops the levee, then both sides should

flood pretty equally.

But let's say, people on one side of the river lobby for higher levees.

Now, instead of both sides flooding, only one floods.

The side with the lower levees is at a clear disadvantage.

So what can people in flood zones do?

We can't just pick up and move major cities.

We need levees to protect places like these.

There is an alternative.

We could build levees farther back, so rivers could still expand and create wetlands.

These “setback levees” ease flooding on both sides, rather than protecting one city

at the expense of another.

This approach is common in other parts of the world - like Holland for example - but

not in the U.S.

Here, we tend to build levees right next to rivers.

Some communities that can afford to build higher levees do so at the risk of others

with little oversight.

We do have the Army Corps of Engineers -- a federal agency tasked with regulating a fraction

of all levees.

They have to ensure -- at least on paper -- that federal levees won't dramatically raise local

flood levels.

But those engineering predictions don't always match reality.

For instance, the Army Corps designed Valley Park's levee in the early 90s with data

and software from that era.

Engineers estimated the levee's impact on neighboring areas would be minimal.

But by the time it was completed in 2005, the region had grown significantly.

More people had built alongside the river, increasing the risk of flooding - which wasn't

considered in the original plan.

While we don't know the precise impact of Valley Park's levee on neighboring towns

just yet, since the levee was built, the region outside Valley Park has suffered two of the

worst floods in its history.

The Army Corps says they've done nothing wrong, and that the levee meets all state

and federal laws.

But there's a growing body of research that shows levees push flooding onto surrounding

communities that have lower levees or no levees at all.

Researchers measured water levels around 13 levees in the Midwest and found they all increased

flooding -- some by over five feet.

As the climate changes and cities push for higher levees, flooding is only expected to

get worse.

Especially along the Mississippi River, which is almost entirely lined with levees.

Some of these embankments have been substantially raised since their completion, against federal

rules, making flooding worse across the river and upstream.

So, yes, we need some levees.

But the system for regulating them is broken.

Even though the science overwhelmingly  shows that constructing higher levees makes flooding

worse in the long-term.

We keep building them taller, passing our problems upstream.