Average Recurrence Intervals (ARI) and Floods
The average recurrence interval (ARI) is how often does a storm of a given magnitude return. A 1-in-100 year storm is expected once ever 100 years. This is often interpreted as a 1% chance in any individual year.
Ellicott City, Maryland was founded in the late 1700s on the Patapsco River, south and west of Baltimore. It is a wooded, hilly landscape typical of the Western Shore of the Chesapeake Bay, which is just a few miles downstream. The hills and small canyons, ravines really, were part of the reason the settlement thrived. They provided falling water to power for mills and manufacturing. From the beginning, during intense rainfalls, the town was flood prone.
In 2016, long after Ellicott City’s time as a milling and manufacturing center, a flash flood washed through the Old Town, ruining businesses and homes. People died. The intense precipitation, 6 inches in 2 hours, was funneled through the small canyons. Aside from the severity of the precipitation, development of land upstream from the city for houses and commerce increased the runoff.
The precipitation was objectively intense. The flood has been characterized as both a 1-in-500 year and a 1-in-1000 year event1 this is the average recurrence interval (ARI); it is not an easy characterization to make because we don’t have data records that long. In either case, for Ellicott City, it was unprecedented – a rare event. Even in the best of circumstances, the storm water systems would not have been designed to manage that much water.
After the flood, Ellicott City with aid from governments at all levels, set about to rebuild, recover, and restart. That is what we do.
In 2018, it happened again.
One of the most robust signals of our warming climate is the increase in extreme precipitation. It is straightforward physics that in an environment with an abundance of water, when the atmosphere and the surface warm up, there is increased evaporation. The atmosphere contains more water vapor. When the conditions are right for rain, the likelihood of large amounts of water falling in a short amount of time is increased.
This type of flash flood has captured people’s attention.
It is a type of flood that, with little ambiguity, follows from extreme rainfall. The depth of the flood waters and the areal extend will depend on, for example, the ability of the soil/surface to absorb water; however, the onset of the flood followed from excessive rainfall.
Ellicott City, Maryland, is a small geographic area. Recent floods in Houston, Texas and Central Europe 2rapid attribution study. of are at much larger scale.
Different types of storms are causing the rains; the warming world has abundant amounts of water vapor to supply them all. 3storms in Rules and Tools
The average recurrence interval (ARI) is how often does a storm of a given magnitude return. A 1-in-100 year storm is expected once ever 100 years. This is often interpreted as a 1% chance in any individual year.
There is no ambiguity that there is a type of flood caused by extreme precipitation that is characterized by events that have a probability of 1-in-100 years or less – i.e., rare events.
That these rare events might be returning in, say, 1-in-10 years, or that in any given year they are occurring all over the country, or they are occurring simultaneously around the globe is a strong indicator of a warming climate.
With the categorization that there is a type of flood caused by rare ARI events, if we establish a trend in ARI, then we do establish a trend in this type of flood.
There are many papers establishing the increase in extreme precipitation. DeGaetano and Tran (2022) 4DeGaetano_Changes_Average_Recurrence_Interval_ARI_JAppMeteorolClim_2022 establish trends in ARI in the Mid-Atlantic United States. The abstract is provided here:
ABSTRACT: Increases in the frequency of extreme rainfall occurrence have emerged as one of the more consistent climate trends in recent decades, particularly in the eastern United States. Such changes challenge the veracity of the conventional assumption of stationarity that has been applied in the published extreme rainfall analyses that are the foundation for engineering design assessments and resiliency planning. Using partial-duration series with varying record lengths, temporal changes in daily and hourly rainfall extremes corresponding to average annual recurrence probabilities ranging from 50% (i.e., the 2-yr storm) to 1% (i.e., the 100-yr storm) are evaluated. From 2000 through 2019, extreme rainfall amounts across a range of durations and recurrence probabilities have increased at 75% of the long-term precipitation observation stations in the mid-Atlantic region. At approximately one-quarter of the stations, increases in extreme rainfall have exceeded 5% from 2000 through 2019, with some stations experiencing increases in excess of 10% for both daily and hourly durations. At over 40% of the stations, the rainfall extremes based on the 1950–99 partial-duration series show a significant (p . 0.90) change in the 100-yr ARI relative to the 1950–2019 period. Collectively, the results indicate that, given recent trends in extreme rainfall, routine updates of extreme rainfall analyses are warranted on 20-yr intervals.
Goffin et al. 2024 take a different measure on days of extreme precipitation, the number of days needed to reach 50% of yearly totals (WD50). They find when considering the number of extreme wet events that affected different parts of Europe in 2011-2022 that observations showed that the floods occurred when a large fraction of the annual precipitation was reached in 2–3 of the wettest days. 5Goffin_Intensification_Wettest_Days_Europe_GRL_2024