A waterlogged weather pattern, Ellicott City’s flood-prone geography and climate change all conspired to make this latest flash-flood horror show.
A torrent of rain
The image below shows the devastating pocket of extreme rain that befell a small region of central Maryland, bracketing Ellicott City, Catonsville and the campus of the University of Maryland Baltimore County over a nearly three-hour period.
The radar estimates 9.6 inches of rain fell midway between Ellicott City and Catonsville, with somewhat lesser surrounding amounts. It indicates about 6 inches fell in Ellicott City proper.
But weather radars notoriously underestimate rainfall. Automated rain gauges in and around Catonsville, at ground zero, recorded nearly 13 to 15 inches of rain. The National Weather Service received a gauge report of 8.4 inches in Ellicott City.
To put this 2018 storm in perspective, about 5 inches of radar-indicated rain fell in Ellicott City July 30, 2016, while gauges back then captured 8.22 inches. So in terms of rainfall magnitude, Sunday’s storm is a likely rival, if not even more intense.
In the coming days, we will have access to more data collected by gauges across the region, which will allow us to refine this analysis.
Ellicott City’s flood-prone geography made the flooding far worse
Flash floods are inherently multi-factor disasters: Rain falls from the sky, and the land surface must process all that water. If the land surface is already saturated (from a month of heavy rain during May), or is heavily urbanized (in the case of Ellicott City), that water cannot infiltrate — it must run off. The meteorology is just one component. The nature of the land surface and the topography are equally critical.
As we wrote after the 2016 flood, Ellicott City sits at the bottom of a topographical funnel, at the confluence of several streams feeding into the Patapsco River.
Now add to that acres of impervious land surface: blacktop, concrete, rooftops, channelization, developed areas around the rim of the funnel. Add inches of water in a few hours, and you have a guaranteed recipe for a flood disaster.
The following image shows how bad this can get, taken from a stream gauge along the Patapsco, downstream of Ellicott City.
The graph plots stream discharge vs. time. Discharge is the estimated volume of water passing the gauge location, per unit time. For the past week, discharge has been coming down after recent rains. Then late Sunday afternoon, the hydrograph explodes upward, a vertical wall of change. This is as dramatic as stream hydrographs get.
An atmospheric conspiracy
Less than two weeks ago, the atmosphere delivered several inches of water in a few hours over a small region encompassing Frederick, Md. A flash flood and terrible damage ensued throughout the historic town. The atmospheric setup from Sunday’s Ellicott City flash flood was nearly identical.
We have been stuck in a late July weather pattern, one in which the jet stream has built a ridge of high pressure over the Mid-Atlantic. This has allowed high levels of afternoon heat and humidity to build through our region. Afternoon warming becomes realized as the buoyant energy in thunderstorm clouds. The high humidity comes from the notorious Bermuda High, which pumps vast amounts of Atlantic moisture our way.
But the vestiges of winter hang on, paradoxically. The waters of the North Atlantic, off New England, are still chilly. Now and then, pockets of this cool air slip south, in the form of a backdoor cold front. These fronts are so-named, in contrast to ordinary cold fronts, which arrive from the west-northwest.
A surge of cool North Atlantic air moving southward into the Mid-Atlantic can push a cold front into our region — curiously — from the north, during late spring and early summer. It’s a quirk of our region’s geography.
This setup is shown in the image below.
A backdoor front with an accompanying area of low pressure straddled the D.C.-Baltimore region Sunday morning. The low pressure drew in air from all directions, creating a convergence of moisture at ground level. As moist air rushed inward, it flowed upward. Rising moist air leads to rain. Add instability (the buoyant energy of a warm afternoon), and the air ascends more forcefully, as updrafts of thunderstorm clouds.
The image also shows a great pooling of atmospheric moisture, called precipitable water. Take a one-meter-by-one-meter square air column, from the ground to the top of the troposphere, and condense all the water vapor. The liquid equivalent at the bottom is a measure of the total water content. On Sunday, our precipitable value was nearly record-breaking for this region, for May 27.
Now it’s time for some simple geometry. The winds in the layer of the atmosphere that steers thunderstorm cells were blowing from west to east. This air current was parallel to the backdoor front pushing into the area. As moisture on south winds impinged on the front, near Ellicott City, storm cells blossomed in the unstable afternoon air. They were then blown eastward, moving along the stalled front. Ellicott City lay beneath the trajectory. What resulted was a train of rain-bearing storm cells, moving one after the other over Ellicott City.
This exact process happens commonly, in all parts of the United States, so often that we call it “cell training.” But for the second time in the space of less than two weeks, this same process unfolded in central Maryland (the first time in Frederick), with disastrous consequences.
Given this setup, both the National Weather Service and Capital Weather Gang recognized, as early as Saturday, the potential for heavy rainfall and flash flooding in the region on Sunday. But realizing potential is one thing; trying to narrow the threat down to a specific town, hours before, is still nearly impossible, given the state of our understanding, lack of weather observations and limitations of the computerized forecast models. The National Weather Service did successfully highlight the general zone where flooding rain was most likely by mid-Sunday afternoon.
Is this climate change?
Yes and no. This was an extremely localized, small, relatively short-lived storm. Climate change unfolds on time scales of decades and over very large regions, by comparison. Climate change did not “cause” this thunderstorm complex.
However, climate change has probably altered the larger environment in which these small thunderstorms are embedded. Notably, the water vapor content of the atmosphere, as a whole, has increased and scientific studies have shown a statistically meaningful uptick in the frequency of extreme rain events over the eastern United States. Statistically, over the long term, these types of extreme floods are probably becoming more common, in areas that are normally rainy as a result of global warming.
The 2016 Ellicott City cloudburst was deemed a “thousand-year rain event” in terms of the probability of recurrence. That an event of this magnitude unfolded in the same spot, two years later, is what it is, and statistics be damned.
Reality check: A “thousand-year event” can happen any time. It boils down to rotten luck that Ellicott City got singled out two times in two years. Let’s not forget the horrific flooding Sunday that also unfolded on the streets of nearby Catonsville and Arbutus, and the campus of the University of Maryland Baltimore County. These spots received even more water.
No doubt, there will be soul-searching for those who live and work in lower Ellicott City. Personally, I would not have expected to “relive” (vicariously) such a devastating flood, in the same spot, over the course of my lifetime. Yet, it happened. Sometimes lightning does strike twice.