Designed to flood
A city street is part of the drainage system. Engineers size the pipe under the road for a storm of a certain size, and everything past that size has to go somewhere. It goes on the pavement. The curb and gutter form a shallow channel, the crown of the road sheds water into it, and the whole cross-section is meant to carry the overflow downhill until the pipe catches up. Streets flood on purpose. The design assumes a driver will see the water and slow down.
The assumption that has aged worst is the one about frequency. Drainage standards are written against precipitation frequency tables, the tables are built from historical rain gauge records, and in much of the country the version still written into local code was assembled from data that stops decades short of the present. The same tables decide how a storm sewer gets sized, which is why a street that has never taken water starts to. Warmer air holds more moisture, and the sharpest signal in the record is in short intense bursts. The practical result is that the storm a street was built to shrug off now arrives more often, and so does the storm that overwhelms it.
The first rain is the dangerous one
There is a second effect that has nothing to do with volume. Traffic engineers have known for a long time that the first rain after a dry spell is more dangerous than the heavier rain that follows it. Oil, rubber, brake dust and exhaust residue collect on the pavement while it is dry. A light rain lifts that film and floats it without washing it away, so the surface is briefly slicker than it will be an hour later once it has been rinsed. Longer dry spells broken by harder bursts is the direction the precipitation record is moving, which means more of that particular transition, and more of it in places where drivers get little practice at it.
The water does not have to be deep to be decisive. The National Weather Service publishes the thresholds, and they surprise almost everyone. Six inches of fast-moving water can knock an adult off their feet. Twelve inches of rushing water will carry away most cars. Two feet will carry away SUVs and trucks. Over half of all flood-related drownings happen when a vehicle is driven into the water. A car that has started to float has no steering and no braking, and it moves with the current until it meets something solid.
Where the road and the creek share the only flat ground
Terrain decides where this lands hardest. The Balcones Escarpment, which runs through central Texas between San Antonio and Austin, is known to hydrologists as Flash Flood Alley: thin soil over limestone sheds rain almost immediately into steep narrow channels, and those channels cross roads at grade. The desert Southwest has the same problem with different geology, where a dry wash is a scenic feature for most of the year and a river for twenty minutes. In much of Appalachia the valley floor is the only buildable ground, so the road, the houses and the creek all sit on it together. None of these places is unaware of the risk. They are built around a version of it that was smaller.
Cities do keep lists. Most public works departments know which underpasses fill first and which crossings close, and the better-funded ones have gauges and automatic barricades at the worst of them. The gap is usually between the list and the budget. A gauge and a swing arm at a single crossing runs into the thousands of dollars, there are counties carrying several dozen crossings with no line item for any of them, and the ones that get funded tend to be the ones where somebody has already died.
What happens afterward is the part worth sitting with. A vehicle that stalls in a flooded underpass at nine at night enters the public record as a traffic event. The crash report has a field for weather and a field for driver behaviour. It has no field for the year the drainage standard was last revised, or for the subdivision approved upstream on the assumption that its detention pond would hold, or for the culvert that was sized in 1974 and never looked at again. The cost then gets settled between two drivers, their insurers and sometimes their lawyers, because the event was filed as a car accident and not as a stormwater failure. Every one of those files is accurate on its own terms. Together they describe a design problem that nobody in the chain has standing to raise.
The fixes are the usual unglamorous ones:
- Gauges and barricades at the crossings that actually close, prioritised by which ones fill first instead of by which ones are most visible from a council chamber.
- A re-survey of which underpasses take water now, because the answer has changed since the last survey and nobody has been asked to check.
- Updating the precipitation tables the local code points at, which is a paperwork job with no ribbon at the end of it.
- Putting the low-water crossings on the same map as the school bus routes and the ambulance routes, which costs nothing and is done in fewer places than you would expect.
None of that is expensive next to a single recovery, and none of it waits on a new standard being written. The street will keep doing the job it was designed to do. The question is whether anyone is told that is the job.