HOUSTON - South Texas is currently battling dangerous flooding left over from tropical depression Imelda. It's a scary scene: Parts of the state have been evacuated, and Texas governor Greg Abbott recently declared a state of disaster for 13 Texas counties.
Then, there's this statistic that puts the whole scene into alarming context: Areas like Jefferson County have seen more than 33 inches of rainfall since the storms began earlier this week, and any rainfall totals of more than 30 inches in 24 hours represent a 1 in 500-year rainfall event.
A 1 in 500-year rainfall event seems historic. Catastrophic, even.
However, while there's certainly nothing underwhelming about these dangerous floods, the reality of that "1 in 500" statistic is a little different than you may assume.
It looks like it's time to delve into the fascinating and oft-misunderstood world of weather
Weather stats rely on two things
First of all, let's cover a few common sense issues: Despite sophisticated modeling and projections, we really have no idea what the weather was in any specific point for most of global history, unless someone decided to scribble down the details on a piece of papyrus somewhere. (Weather, of course, being separate from climate, which is a longer-term pattern. For instance, we may know it was generally rainy in the United Kingdom in the 1800s, but we don't know for sure what the exact weather was on September 19,1855, unless someone recorded it.)
"Climate is a long term average of weather," says CNN meterologist Brandom Miller. "So you can make more precise projections over the longer term with higher confidence."
Given that, the whole business of weather statistics relies on two things: Our modern recordings of weather history and the informed assumptions one can make from them.
"A 1 in 500 year event is really based on how frequently an event is likely to reoccur based on its percentage likelihood based on the historical record in that location," Miller says. "Given that we only have 100-150 years of records in the US, some of the extreme low frequency events may have never even occurred before, but are based on extrapolations of what we do have on record."
A different way to look at it
Rather than imagining one particular rain event happening every 500 years, think of it this way: There is a 1 in 500 chance of this rain event happening in a given year, or a 0.2% chance. Put that way, the probability of it happening year over year stays the same.
"Something that has a 20% chance of occurring in a given year, or 1 in 5 chance, would be a 'one in five year event,'" Miller says. "But of course, these can happen much more frequently than that. There is still a 20% chance that it happens the next year, or even at another point that same year. So that can be very confusing to people."
The rains in Texas showcase this very issue: The rainfall brought on by Imelda is statistically high, but it's still not as bad as the rains the area saw during Hurricane Harvey last year. So that "1 in 500 year" rainfall? It's actually happened more than once recently. Very recently. Miller says the Houston region alone has had two or three 1-in-hundreds type weather events in the last five years.
"A 1 in 500 year event has a 0.2% chance of happening in a given year. Having 2 in 3 years (as has happened in Houston area now) would be astronomical," he says.
The big thing throwing these numbers off
Why do these statistics break down so easily? A lot of it, Miller says, has to do with climate change.
"We are seeing, with the climate crisis, many of these extreme events with very long recurrence intervals happening much more frequently."
So something that really may have had a 1 in 500 year statistical probability, over time, is becoming more frequent.
"This tells us that the historical statistics might not be too relevant anymore in a warming climate," Miller says. "The deck is stacked toward more extreme rainfall events in particular."
In other words, these rare, dangerous events are getting less rare. And that's probably scarier than any statistic.
Another thing to know about rain
While we're talking weather stats, here's another one to keep in mind the next time you check your local forecast: That 40% chance of rain probably doesn't mean exactly what you think it means, either.
"In general, PoP (probability of precipitation) is defined as a combination of the percent chance of rain over a given amount of time and the percent coverage over a given area," Miller says. "So if there is a 50% chance that rain will occur and that rain would cover about 40% of the region of interest, there would be a 20% chance of rain for that region the next day.
In other forecast models, the probability of precipitation could be in relation to other weather conditions, as in, when all of those conditions have been met, historically, there has been a 40% chance of rain in your particular area.
Of course, there are different types of forecasts and some of the more specific ones could be better tailored to your exact location. Hey, predicting the weather is complex. And clearly, keeping track of it is too.