Every year floods make their way through populated areas, costing lives and millions of dollars in damages, devastating communities, and grinding local economies to a halt. If you’ve ever experienced one yourself, you know how powerless it feels to be up against mother nature. And if you haven’t, be careful in thinking it can’t happen to you. Nearly every major city across the world is susceptible to extreme rainfall and has areas that are vulnerable to flood risk. Luckily, we’ve developed strategies and structures over the years to reduce our vulnerability and mitigate our risk. We still can’t change how much it rains (at least in the short term), but we’ve found lots of ways to manage that water once it reaches the earth to limit the danger it poses to lives and property. Today, we’re talking about how large scale flood control structures work on rivers.

We all know generally what a flood is: too much water in one place at one time. But, I think there’s still uncertainty in how floods actually occur. Part of the reason for that confusion, I think, is the huge variety of scales we have when talking about flooding. Most river systems are dendritic. The topography of the land and the long-term geologic processes mean that streams join and concentrate the further you move downstream just like the branches of a tree. A watershed is the entire area of land where precipitation collects and drains into a common outlet; it’s a funnel. And as you move downstream, those funnels start to combine. The further you go, the larger the watershed becomes as more and more streams contribute to the drainage. So watersheds can be tiny or gigantic.

Your front yard is a watershed to the gutter on the street. If it happens to be raining hard directly on your house, the gutter will flood, maybe even overtop the road onto the sidewalk. At the complete opposite end of the spectrum, more than a million square miles (or three million square kilometers) make up the drainage area of the Mississippi River in the U.S. A big rainstorm in one city is not going to make a dent in the total flow of this river. But, if everywhere in the basin is having an unseasonably wet year, that can add up into major flooding as all that water concentrates into a single waterway. This seems simple, but it is a real conceptual challenge in understanding flooding, not to mention trying to control it. Smaller watersheds only flood during single intense storm events, called flash floods. Usually, this water is already long gone by the time the next storm comes. In contrast, large watersheds flood in response to widespread and sustained wet weather. They aren’t really affected by single storm events. Of course, in a dendritic system, there’s everything in between which means a flood can be a local event affecting a few houses and streets for a couple of hours during an intense thunderstorm or a months-long ordeal impacting huge swaths of land and multiple communities.

Riverine flooding is also a challenge because it’s not linear. In a cross section through a river, you have the main channel where most normal flows occur. Every unit of rise in the river doesn’t equal that much extra width in inundation. Plus there’s not much development within the banks of a river: maybe some low bridges and a few docks. But, above the channel banks, things change. The slopes aren’t so steep and you end up with wide, flat areas of land. And you know what we humans like to do with wide, flat areas near waterways - we build stuff, like entire cities. That or use it as farm land. The problem is that, once a channel overbanks, every unit of rise in the river equals much wider extents of inundation. You can see now why this is called the floodplain. And looking at a cross sectional view, it’s easy to see one of the most common structural solutions to flooding: levees. If overtopping the banks of the river creates the problem, we can just make the banks of the river higher by building earthen embankments or concrete walls. Levees protect developed areas by confining rivers within artificial banks. That means areas outside the levees flood less frequently. It doesn’t mean they have zero flood risk at all, since it’s always possible to have an extreme event that overwhelms the levees. For earthen structures, overtopping of a levee can cause erosion and even failure (or breach) of the berm. That can make the flooding even worse than it would have been otherwise, especially if people weren’t evacuated from the area ahead of time. So, even though they are a pretty simple solution to the problem of flooding, levees aren’t perfect.

Sometimes getting that water out of the channel is exactly what you want though. Another tried and true flood control technique is diversion canals. These are human-made channels used to divert flood waters to undeveloped areas where it won’t be as damaging. Often it’s not possible to widen an existing river because there’s already too much development or for environmental reasons. So instead, we create a separate channel to divert floodwater around developed areas and back into the natural waterway downstream. In most cases there will be some kind of structure at the head of the diversion channel to help control which route the water takes. For normal conditions, water will flow through the natural river, but when a flood comes, most of that water will be diverted, reducing the flood risk to the developed areas.

But, it would be nice if all that water didn’t make it into the river in the first place. That’s only possible with the other major type of flood control infrastructure: dams. These are structures meant to impound or store large volumes of water, creating reservoirs. Dams meant for flood control are kept partially or completely empty so that, when a major flood event occurs, all that water can be stored and released slowly over time. The theory here isn’t too complicated. We can’t change the volume of water that comes from a flood, but with enough storage, we can change the time period over which it gets released into the river. Big sloshes of water into this bucket come out slowly over time. As long as the sloshes are far enough apart and the bucket is big enough, you almost never see significant flooding out on the other side. But, not all dams are built specifically for flood control. Many reservoirs are intended to stay as full as possible so the water can be used for hydropower, supplying cities, or irrigation of crops. If a water supply reservoir happens to be empty at the time of a big flood, it will work just like a flood control reservoir, storing the water for later use. But, if the reservoir is already full, they have to open the floodgates to let the water through. This can be frustrating for the residents downstream who may have thought they had protection from the dam.

In many cases, a dam can serve multiple purposes at the same time. Different zones, called pools, are established for the different uses. One pool is kept full to be used for hydropower or water supply and one is kept empty to be used for storage in the event of a flood. Finding the right balance point between how much storage to keep full versus empty is a complicated challenge that considers climate, weather, the maximum amount of flow that can be released without damaging property downstream. Some dams vary the size of these pools over the course of a year depending on the seasonality of flooding, and some even use risk indicators like the depth of the snowpack within the watershed to dynamically adjust the volume available to store a potential flood.

I’ve been using the term “flood control”, but the truth is that term is falling out of favor. Now if you ask an engineer or hydrologist, they’re more likely to talk about “flood risk management.” Our ability to quote-unquote “control” mother nature is tenuous at best, and the more we try, the more we realize this: even if expensive infrastructure is helpful in a lot of circumstances, at best it is an incomplete strategy to reduce the impacts of flooding over the long term. For one, flood control structures (especially levees) can protect some areas while exacerbating flooding in other places. For two, overbanking flows are actually beneficial in a lot of ways. Just like wildfires, flooding is a natural phenomenon that has positive effects on the floodplain like improving habitat, ecology, soils, and groundwater recharge. And for three, we are understanding more and more the true value of resiliency - that is instead of reducing the probability of flooding, instead reducing the consequences. This is normally accomplished with strategic development like reserving (or converting) the floodplain for natural wetlands, parks, trails, and other purposes that aren’t as easily damaged by flooding. In fact flood buyouts where high risk property is purchased and converted to green space is often the most cost effective way to reduce flood damages in the long term (even if not the most politically popular strategy).

It’s not likely we’ll ever have the ability to reduce the volume of rainfall during major storms, and in fact, many locations are already experiencing more extreme rainfall events than they ever have due to climate change. But, we will continue to develop strategies, both structural and non-, to reduce the risk to lives and property posed by flooding.