The issue is one of rates not volume. They briefly mention Houston in passing, which is a perfect example. The water is all part of the same cycle and same watershed, but where it goes within that watershed and at what rate become critical. If you make 50% of an area impervious to absorbing water (i.e., you pave it) then the remaining 50% has to absorb much more water, and likely will become saturated. But first, the impervious nature of pavement will cause those low lying areas to flood. Because the water flows faster on the surface than it migrates through heavily packed soil, the result is a water build up that can not only flood the now saturated area, but will in fact flood the lowest areas period. Dirt has an absorption rate that is (somewhat counter-intuitively) actually negatively impacted by becoming excessively dry. This is why Phoenix, notable for rare but massive rain storms, has enormous flooding problems that they deal with mostly by creating buffer areas in the form of intentionally designed flood zones (e.g., parks with berms around them or underpasses).

Houston is designed like this as well (the Addicks and Barker reservoirs). But they are designed based on a given rate of rain over a given area. That assumption wasn't actually violated that badly by Harvey, it was within the margin of error. What was violated in the Houston flood model was the amount of hardscape and where the homes were. The assumptions about the amount of hardscape affects how much of a given rain will reach the reservoir (a lot never does to to evaporation or local use) as well as the rate at which that water reaches the reservoir. That whole system is designed to rate limit the flow of water into the core of the city of Houston. The model and assumptions were built when the core of the city was the main part of the city and the suburbs were much smaller. The water has to go somewhere, and it wants to go into the gulf of mexico...but the rivers and canals in the city core have a limited discharge rate into the gulf. Therefore the goal was based on creating a system that tolerated the hardscape based reduction of the cities 'sponge' and a greater reliance on the sponges of the suburbs. This is all not to mention that they actually built homes INSIDE the reservoirs that are a lot of what we talk about when we talk about the Houston flooding...those homes flooding was the system working as designed, the only people out of the loop were homeowners (slightly more true: and realtors, developers, and local officials who aren't hydrological engineers) [0]. They should have known better but they didn't, because people tend to not notice, understand, or take seriously these details of municipal planning. Then they are shocked when either it works as designed (as a net positive for society at potential costs to individuals) or when it fails because choices made from lack of knowledge have blithly broken a system that is designed for emergencies. Think of it this way, there were people in the Houston gov who knew exactly what was going to happen...do you think talking about a biblical level flood in Houston was going to be taken as a more powerful narrative than 'economic growth, jobs, housing, american dream' when deciding on zoning permits?

TL;DR any sponge can become saturated but the point is that you should not artificially reduce the size of your sponge. Cities know how big their sponge is, its the ground around them. They make predictions and plans on it...and then subdivisions and growth massively seal off parts of the sponge.

[0] https://projects.propublica.org/graphics/harvey-reservoirs

Obviously, but that's not the test.

The test is whether the sponge can absorb rain as fast as it can fall for a long-enough time that other consequences of the rain become serious problems (e.g. deferred outdoor maintenance), while the sponge still isn't saturated.

and imagine what happens when a freeze comes by and what it does to the sponges' life