Percent Chance Of Rain And Intensity

A close colleague and friend asked, “How is this flooding possible when you said there was only a 30% chance of rain?” He was confusing probability with intensity. The percent chance of rain is not necessarily an indicator of rainfall intensity. Probability of rain considers the multiplication of area covered times confidence in the forecast.

A National Weather Service document provides a great example. ”If the forecaster was 80% certain that rain would develop but only expected to cover 50% of the forecast area then the forecast would read 40% chance of rain for any given location," noted the NWS website. If widespread rain is expected over 100% of the area but with only 40% certainty, that would also be a 40% chance. The minimum expected amount is one hundredth of an inch, but there is no information about intensity.

‘Rain Or No Rain’ Mental Models

The official report for Wednesday in Atlanta was 0.00 inches of rain. That’s because nothing fell at the official recording site, the Atlanta airport. Up to 2 to 3 inches of rain fell in a scattered shower over downtown Atlanta within a short period of time. This is exactly why we use probability forecasts for rainfall.

Scattered or widespread showers can produce rainfall at isolated locations. Further, as I have written previously, people form mental models about percent chance. My own observations tell me that if the forecast is 30% or less, people often assume it probably won’t rain. For percentages greater than 70%, they assume that it will rain.

Many people assess scenarios as black or white. Will it rain or not? The answer is often both, it just depends on your location. Because people draw conclusions based on their local or personal decisions, I suspect there will be an array of interpretations of whether the 30% chance of rain forecast was accurate in Atlanta that day. Someone driving on Atlanta’s downtown connector experienced a downpour. People only a few miles away experienced nothing.

Isolated Rainstorms Cause Flooding

I am a part of a University of Georgia research team led by my colleague Grace Ahn that is thinking about risk communication. The NOAA-funded project uses advanced immersive reality technologies to convey risks of driving through flooded waters. "Turn Around, Don’t Drown" is a slogan, yet I saw many people in Atlanta driving into flooded roadways. They just wanted to get home, underestimated the danger or assumed because the previous car made it, they also could.

In our research discussions, I have repeatedly stated that this exact type of high-intensity rain event from an isolated storm can be just as dangerous as a widespread coverage rain event. To be fair, some of the Atlanta flooding happened on busy interstate highways. Former student and current NWS meteorologist Jared Rackley noted on my social media post that in that situation, “turning around” may not be a viable option. He has a point, so we need to think through the guidance in this situation, too.

Flash Flooding Is Not Just About The Rainfall

Flooding in New York City and Atlanta this week revealed the flash flood equation that I often discuss:

Flash Flooding = Intense Rainfall + Impervious Surfaces + Stormwater Removal.

When a significant amount of rain falls on an area quickly, it can lead to flooding. Impervious surfaces like roadways and parking lots increase runoff into streams and reduce infiltration into the ground.

When we learned about the water cycle in elementary school, it was unrealistic because no human presence was included. Intense rainfall and impervious surfaces are a recipe for urban flash floods. Now consider a stormwater removal system that may be impeded by blockages or engineering assumptions for last century’s rainstorms, and you have ideal flash flood conditions.

Officials in Georgia issued a statement about the Atlanta flooding and drainage conditions.

A Climate Change Angle?

Is there a climate change angle? Yes, the rainstorm happened due to natural weather processes, but it is very much consistent with prevailing studies that rainstorm intensity is trending upward. In other words, it rains harder, on average, in intense events. Up to three inches of rain falling in less than an hour is an extreme event.

Basic physics explains the connection. “The Clausius-Clapeyron equation shows that for every 1°C temperature increase, Earth’s atmosphere can hold 7% more water…A warmer atmosphere leads to greater presence of evaporated water/moisture to fuel the rainstorms,” wrote David Shultz.

Though flood-climate relationships are more complicated than this simple calculation, virtually all cities have engineered their stormwater design under the assumption of “stationarity.” In other words, they are designed under the assumption that rainstorms like the one this week are just like rainstorms in 1970. They are not. They are “juiced.” Infrastructure design and resiliency planning must consider weather of the present and future, not the past.

Both Drought And Flooding—How Is It Possible?

The state of Georgia, like many places in the South, is experiencing drought. Meteorological drought conditions are defined by prolonged lack of rainfall. Over the next several days, Georgia and much of the southeastern U.S. will experience a weather pattern conducive to increasing rain chances.

That’s great news, because my lawn and vegetable garden need the rain. However, it will take 1 to 2 feet of rain over a month to get parts of Georgia out of drought. We do not want it in the form of these types of downpours. In a 2019 Congressional testimony, I explained that extreme events on both sides of the rainfall ledger could actually become more frequent and/or intense. That’s what disrupts society, more so than averages.

Remember, however, that people often want black or white answers. Is it getting wetter or drier? As we saw this week in Atlanta, “both” is a viable answer.

By the way, dry conditions associated with drought can also amplify flooding because extreme rain falling on hard, dry surfaces increases runoff rate like a paved surface.