The Statistics Behind the 100-Year Storm
By Emma Collins
If you live in a flood-prone region, you probably hear a lot about something called the “100-year storm.” But what is that exactly?
Flooding in Conway, SC after hurricane Florence (Jason Lee/Sun News/AP)
What is the 100-year storm?
The 100-year storm is the number of inches of rainfall one could expect to occur in a certain period of time (usually 24 hours) with a 1% probability any given year. Note that this is not the same as the 100-year flood, which is the water surface elevation that one could expect to experience with a 1% probability in any given year. This is a theoretical number derived from years of observed data, and while a common misconception is that a precipitation event of this magnitude is a once-in-a-hundred-year storm, it’s better to think in terms of probability. A 1% annual probability does in fact mean that the 100-year storm will occur once in a span of 100 years on average, but this is often erroneously presented as an event that will happen at most once every 100 years.
For example, the 100-year rain event for Charleston, SC, is 10.2 inches in 24 hours. The historic weather events in October of 2015 brought 11.50 inches of rainfall in 24 hours to the Charleston area, but the 100-year storm event was also exceeded in September of 1998, when Charleston experienced 10.52 inches in a day. So, what happened? Is it a fluke that in a span of 17 years a city could have two storms that exceed the 100-year rainfall? Statistically speaking, this shouldn’t be a huge surprise. There is a 15.7% chance that a 100-year event will occur at least once in a 17-year period. More specifically, the probability that the 100-year storm will occur at least once is as follows:
How well does this number represent reality?
This is a highly debated question. In most cases, the 100-year event is either extrapolated from a dataset covering less than 100 years, or taken from a dataset of around 100 years. The underlying assumption of this type of analysis is temporal stationarity, which is to say, rainfall patterns will not change throughout time. If extreme weather events start to occur with greater frequency, then using historical weather data to predict future weather data will not be possible. It would be more accurate to represent the 100-year storm as a storm that has historically occurred once in a span of 100 years on average.
What this means for Stormwater Management
Urban areas often struggle during unusually large storm events. Relative to rural areas, cities are usually characterized by high amounts of impervious surface (i.e. area in which rainfall cannot soak in to the ground). This means that during storms most precipitation gets directed into gutters and sewer drains, which for large enough events can overwhelm a city’s stormwater system, backing up ditches, storm drains and culverts. This is especially a concern in areas that are experiencing rapid development. Many stormwater systems are designed to handle a 100-year storm’s worth of runoff based on the amount of impervious cover, but that may become overwhelmed in a 10-year storm if new development converts formerly vegetated area to impervious surface (such as rooftops and pavement). Likewise, if changing climate patterns were to cause the formerly 10% annual probability storm to occur with a 50% annual probability, the stormwater system in place would quickly become insufficient.
While stormwater systems can be engineered to handle any amount of flow, there is a point of functional and economical diminishing return. Municipal systems generally aren’t built to withstand the 1000-year flow because even though the impact can be catastrophic, they are statistically unlikely to occur. However, in the event large volumes of runoff become increasingly common, stormwater systems must be designed to accommodate the moving point of catastrophe.