Imagine a devastating, summertime windstorm that tracked across 10 states, from Iowa to the Atlantic Coast, traversing the region in just 18 hours. Along its path, the windstorm generated extreme gusts, from 70-100 mph, felling millions of trees and snapping untold numbers of utility poles. At the end of the day, nearly 4.5 million power outages left people sweltering amidst a vicious, early-summer heat wave. Twenty-two people lay dead from falling trees, windblown debris, and lightning. All told, the NWS received over 1,100 reports of wind damage.
What Is a Derecho?
This was not a deadly tornado outbreak, and not an “inland” hurricane. The severe summer windstorm was called a derecho. As it raced across the landscape at 50-60 mph, it was fueled by the oppressive heat and humidity of a multiple-day heat wave spanning the upper Midwest, Ohio Valley, and mid-Atlantic. Until this storm, the term “derecho” was largely unfamiliar to the general public. The phenomenon was first reported in the scientific literature during the late 1800s. Meteorologists describe a derecho as a long-lived convective wind storm, producing continuous, severe wind damage (gusts > 58 mph) along a corridor at least 250 miles long. The level of damage is comparable to that of a weak-to-moderate tornado, but imagine a damage swath hundreds of miles long by many tens of miles wide. Instead of rotary winds, derechos produce intense straight-line winds, with localized pockets of diverging (fan-like) wind called downbursts. While tornadoes are embedded in thunderstorm updrafts, downbursts and straight-line winds are the products of intense downdrafts.
Swift Impact, Then Days of Misery
Figure 1 illustrates the June 29 derecho, from the vantage of a weather satellite. This massive, even outright menacing, storm crossed the Ohio Valley during the day, surmounted the Appalachians during the evening, then struck the Washington, D.C.-Baltimore, Maryland, metropolis at 10:30 p.m. As the derecho approached, observers described a curtain of nearly continuous lightning, followed by a thunderous “wall of wind” and horizontal sheets of rain. Whole trees bent over, snapping or uprooting. Limbs and debris were flung into the air. Utility lines and power poles crashed down under heavy limbs.
Figure 1. Stunning false-color infrared image of the June 29, 2012, derecho in its mature phase over the Ohio Valley. Black and white colors correspond to the highest cloud tops.
Within 30 minutes, the storm was over. Millions of households were plunged into darkness, then endured several days of sweltering heat as crews worked to rebuild the electrical grid. Many businesses lost significant revenue (including spoiled frozen food in grocery stores and restaurants), transportation networks were paralyzed, and cell towers were overwhelmed with callers. Even vital 911 call centers were knocked out.
Figure 2 shows the large number and widespread nature of power outages, throughout a 12-state area. Maryland, Ohio, Virginia, and West Virginia were the hardest hit, averaging close to one million outages per state. While this was a multi-day outage for many locations, outages became a multi-week affair in the most remote locations, particularly across West Virginia. Access to fallen utility lines was significantly hampered by the sheer volume of tree debris that first needed to be cleared from roadways.
Figure 2. Recovery curve of massive power outages created by the June 29, 2012, derecho. Millions lost power in what turned out to be a multi-week restoration effort for many during a prolonged heat wave.
We return to the vantage of earth's orbit once more. Figure 3 illustrates a before-and-after sequence of nighttime lights across the Washington-Baltimore region. The reduction in intensity and coverage of suburban lighting in the immediate wake of the derecho is striking. In a few places, whole towns appear to have disappeared. Even as the electrical grid was hastily rebuilt, high electrical demand in the ongoing heat wave continued to stress the crippled grid, causing secondary outages in the days following the derecho.
Figure 3. Nighttime images of the Washington-Baltimore urban corridor, home to nearly 10 million inhabitants, before and after the June 29, 2012, derecho. The temporary, partial loss of the region's utility grid is striking.
Life History of the Derecho
The continuous network of United States NEXRAD (NEXt generation RADar) sites allowed scientists to stitch together a complete visualization of the derecho, from initiation to decay. It's a spectacle to view the complete animation, but the storm's rapid growth and steady progression are captured by the radar mosaic shown in Figure 4. Radar echo intensity—a measure of heavy rain (and embedded hail)—grades from light green to yellow to red. Surface observations of peak wind gusts (mph) are shown throughout the storm's progression.
Figure 4. Mosaic of radar “snapshots” illustrating progression of the June 29, 2012, derecho across the Ohio Valley and Mid Atlantic. Prominent is the arc-shaped bow echo. Individual wind gust reports (mph) are shown in the small squares.
While the first thunderstorm cells can be traced to Iowa, the mosaic picks up where these cells became particularly severe in the vicinity of South Chicago, Illinois. This initial cluster moved into northern Indiana around 3 p.m., producing violent wind gusts in excess of 90 mph. A separate cluster of cells rapidly developed over central Indiana, then merged with the northern storm. This merger was a key, defining step in the assembly of the derecho, as it doubled its effective length. Such aggregation of cells is termed “upscale growth,” and because it's a random process, derecho prediction is very difficult.
After 4 p.m., the arc of severe thunderstorms continued to intensify and expand southward. A steady-state arc is a derecho's defining radar signature, which is termed a bow echo. Clusters of downbursts create a powerful, fan-shaped surge of wind at the line's mid-section, surging ahead into an arc shape. The bow complex in this case raced steadily east-southeast, at times exceeding 60 mph. Summertime derechos are often composed of a single, massive bow complex that regenerates continuously as it lifts warm, humid air ahead of it—literally “scooping up” highly unstable air. There are other dynamics at play, including large, counter-rotating vortices on either end of the arc. These pull in cold, dense air at high levels behind the storm. This chilled, descending current helps maintain the forward-directed surge of winds at the surface.
By 5-6 p.m., meteorologists in communities ahead of the derecho realized that quite a formidable weather system was fast approaching. With the system nearing the western slopes of the Appalachians, forecasters in the Washington-Baltimore region struggled with whether to brace for a significant impact. They had good reason to question the integrity of the system, which would enter the region well after sunset. With the loss of the sun's heating, perhaps the atmosphere would stabilize, and the vigor of the derecho's convective cells would begin to wane. Or perhaps the irregular rampart of the Appalachians would disrupt the storm's circulation, hastening a weakening trend.
Lines of severe thunderstorms frequently meet their demise when crossing the Appalachians. On June 29, meteorologists had several tools to assess the likelihood of weakening. One was the latest update cycle of numerical prediction models, those that focus on the smaller motion scales including terrain interactions. These are called mesoscale models and are discussed later in this article. The other critical forecast aid was an ongoing assessment of the atmosphere's instability that evening. This is done by launching weather balloons (called radiosondes) at standard times. The 8 p.m. launch made at Washington, D.C.'s Dulles airport sent a chill down the spine of duty forecasters. The sounding (vertical profile of temperature and humidity) revealed a deep, extremely unstable air mass. Like a keg of black powder, the atmosphere over Washington-Baltimore was primed for explosive convective activity. If the derecho survived the Appalachians intact, it would very likely regenerate, perhaps even thrive, raising the specter of widespread, highly destructive winds well after dark—just as people were turning in for the night.
Why was the atmosphere east of the Appalachians so unstable? The air mass was accumulating nearly unprecedented levels of heat energy and moisture in its low levels, with part of the heat dome parked over the eastern two-thirds of the United States. The high temperature had hit 104°F in Washington, D.C., that afternoon. Additionally, this energy-rich low layer was being held at bay, capped by a strong stable layer termed an inversion. The inversion layer could be traced as far west as Wyoming, and it “bottled” the unstable air, preventing its premature ascent and release during the afternoon. Now, with the derecho rapidly approaching, it seemed likely that the massive “scoop” of chilly air travelling ahead of the storm would ingest some of the most energetic summertime air ever observed east of the Appalachians.
A Wall of Wind
Countless severe thunderstorm warnings were issued for the Washington-Baltimore region and Virginia. The densely populated area braced for a severe wind impact, and the derecho delivered. The approach of the derecho—which did indeed rejuvenate once on the other side of the mountains—was monitored by NEXRAD. Figure 5 shows the radar operator's view around 10:30 p.m. as the storm arrived on Washington, D.C.'s doorstep. The left panel is the conventional radar view, with oranges and reds indicating heavy rain cells. Taken by itself, it suggests a strong squall line of thunderstorms, but notably absent is the telltale arc or bow shape. However, the Doppler wind image (right panel) foretells a serious wind impact. In this panel, the black dot to the west of Washington, D.C., marks the location of the radar. Green colors show winds approaching the radar; red and magenta colors indicate receding winds. Put the two together, and you can visualize a massive sheet of wind sweeping over the radar from the northwest.
Figure 5. Approach of the June 29, 2012, derecho, as documented by the National Weather Service Sterling, Virginia, Doppler radar. Left panel shows ordinary reflectivity (rain intensity) while right panel portrays the dramatic “wall of wind” sweeping from west to east across the region.
The razor-sharp leading edge of high velocity air is the “wall of wind” that strikes like a hammer blow. Immediately behind this edge, Doppler indicated winds in the 60-70 mph range just a few thousand feet above the surface, with embedded pockets in excess of 80 mph. Such was the suddenness and intensity of wind that electrical substations were being knocked out upwind of communities, minutes before the arrival of the actual wind front!
Forecasting and Warning for the Derecho
Derechos pose a significant forecasting challenge. Meteorologists are confronted with a windstorm that creates damage on par with that of a landfalling hurricane, but which congeals in a matter of hours, as opposed to days of advance warning provided by an approaching Atlantic storm.
Until 10 years ago, forecasters had to rely on the larger-scale, regional prediction models to pinpoint general regions that were primed for severe thunderstorm initiation. This was supplemented with hourly observations culled from radiosondes, surface stations, and satellites. Still, forecasters had to interpolate the point observations in order to “fill in the gaps,” and thunderstorm cells are small enough to develop inside those gaps.
Enter faster computer processors and real-time mesoscale prediction models: these mathematical representations of the atmosphere focus on smaller geographical domains. They use a finer grid spacing, and can be run every one to three hours. They begin to approach the scale of larger convective disturbances, such as the bow echoes that lead to derechos.
The Storm Prediction Center (SPC) in Oklahoma City, Oklahoma, realized the growing potential for a widespread severe wind threat downwind of the derecho, and issued severe thunderstorm watches and a “moderate risk” outlook for communities east of the Appalachians. Forecasters were aware of the tremendously unstable air mass and hedged that the derecho would hold together crossing the Appalachians. An example of a mesoscale model prediction run at 1 p.m. on June 29 and valid nine hours later for 10 p.m. is shown in Figure 6. Indeed, the model continued to simulate a very intense, bow-shaped convective system after sunset.
Figure 6. The NOAA High Resolution Rapid Refresh (HRRR) prediction of the derecho for the afternoon of June 29, 2012.
Severe thunderstorm warnings were punched out ahead of the derecho for large swaths of counties just ahead of it. These warnings—totaling nearly 300 for this single convective event—are shown in Figure 7. More so than in the radar mosaic of the storm (Figure 4), one can appreciate the true scale of the derecho. Its origins are traced back to that spot of convective cells in Iowa, early in the morning. The merging of two separate “severe trails” is apparent over eastern Indiana, where the two main clusters of storms coalesced into the single, large bow echo. Amazingly, the derecho continued to expand in length, nearly continuously, through the afternoon and evening. Never before have so many individual warning polygons been welded into a continuous, “mega-warned” region, outlining a single, massive convective event. The coordination and rapidity of warning issuance are a true testament to the efficacy of the modern NWS forecast network.
Figure 7. Nearly 300 separate severe thunderstorms warnings were stitched together into an enormous storm-warning envelope.
Despite mesoscale models and high-resolution observations, it is still difficult to accurately predict the formation of a derecho more than 12-18 hours in advance. Convective initiation is tricky to nail down. Upscale growth by coalescing cells may be predicated more on stochastic (random) interactions than deterministic predictions. Our understanding of whether a bow echo will survive intact across the Appalachians is only now receiving serious study. But a derecho can produce societal damage and disruption comparable to a tornado outbreak or Category 1 hurricane at landfall. Not every derecho is as devastating as the June 29 system, but the United States remains quite vulnerable to these summertime events, which occur by the dozens every year. The onus is on research scientists and operational meteorologists to increase their predictive skill.
Derechos and Global Warming
The June 29 derecho swept over the Atlantic after midnight, leaving millions of people to pick up the pieces along a damage swath that extended nearly 950 miles. On this day, much of the heat dome's thermal energy was swept up by a massive storm system and converted into the kinetic energy of deadly, damaging winds. Nature has ways of dealing with excess heat accumulation, and these violent events provide a “safety valve” of sorts.
What about future derechos—their incidence, frequency, and tracks—given decades of climate change, and its upward trajectory? It's increasingly clear that summertime derechos develop along the northern periphery of large, stagnant heat domes. Sometimes they erupt in families, such that locations along “derecho corridors” experience repeat hits. It's logical to assume that as climate warms, summertime heat waves will become more frequent, and more intense. What does this say about derecho activity?
Recently, there has been a push to better understand whether tornado outbreaks in the Midwest and mid-South will become more frequent or virulent as a result of global warming. Strong and violent tornadoes originate from rotating thunderstorms (supercells), which require large amounts of thermal instability and wind shear to develop. A case can be made that the atmosphere will become more unstable, but wind shear may actually decline (since shear is created by strong thermal contrasts between the poles and tropics, and global warming is slowly weakening this hemispheric difference). So it's conceivable that there may not be much change in tornado frequency.
Derechos also require very unstable air, but their dependence on wind shear may be less. More moderate values have been associated with quite violent derechos. The wind does not need to veer, or change direction, but only to increase in speed with altitude. While there appears to be no study that examines the future incidence of derecho activity in the United States, it is possible to expect greater derecho activity in the coming decades. Combined with misery of a mid-summer heat dome, that would be very troubling news indeed.
JEFFREY B. HALVERSON is a professor in the Department of Geography and Environmental Sciences, University of Maryland Baltimore County. He is also serves as the Severe Storms Expert for the Washington Post's Capital Weather Gang.