Sunday, December 23, 2012

Be Very Afraid...

NOAA meteorologist Sarah Jamison plugged March 1913 measurements into today’s computational climate models to reconstruct just what kind of mammoth weather system caused the Great Easter 1913 Flood. The chilling discovery? Such a weather pattern could recur...
“No wonder the Easter 1913 tornadoes and record-breaking floods came as a surprise,” observes Sarah Jamison, Service Hydrologist of the National Oceanic and Atmospheric Administration (NOAA), working from the National Weather Service Forecast Office in Cleveland, Ohio. “It’s almost impossible to forecast weather just from surface measurements of winds, air pressure, and precipitation. You really need to know what’s going on upstairs”—slang for knowledge of the three-dimensional state of the atmosphere from 5,000 to 30,000 feet—“including knowledge of how jet streams steer giant cyclonic systems across the continent.”

Sarah Jamison, NOAA Service Hydrologist, explaining her reconstruction of the Great Easter 1913 storm system using NOAA's Twentieth Century Reanalysis Project (20CR).
[Image credit: Trudy E. Bell] 

Fascinated by floods since her childhood spent on a large river in Maine, Jamison’s day job is keeping track of precipitation and monitoring heights of rivers with stream gauges to produce timely predictions of potential flood conditions around Northern Ohio. Immediately after joining the Cleveland NWS office in April 2010,  she set about making herself smarter about local and regional hydrology, geography, and record floods—and stumbled upon still-existing evidence of the unprecedented magnitude of the 1913 flood around Ohio and Indiana. “What blew me away was how many stream gauges still show a March 1913 measurement as the flood height of record today,” Jamison recounts. “When I plotted those locations on a map [see map just below], I was amazed by the vast area they cover. Just what kind of weather event could produce widespread flood records that have stood for a full century?”

In 1913, the National Weather Service was, by today’s standards, primarily reactive: recording data of what happened at scores of weather stations on the ground. There were no systematic releases of weather balloons twice daily worldwide, as there are today, for radioing back continuous measurements of wind velocities, temperatures, pressures, and other information to construct a miles-high vertical profile of the atmosphere every 12 hours. The jet streams—meandering tubes of strong, fast winds miles high in the stratosphere—had not yet been discovered. No weather satellites were staring down at Earth 24/7 and transmitting photographs of planet-wide weather systems. There was no Doppler radar for detecting precipitation or distinguishing between rain, snow, or hail. There was only limited theoretical grasp of what surface warm fronts and cold fronts portended for storms, and no computers to synthesize and translate observations into real-time forecasts.
Thus, meteorologists in 1913 were astounded by just how much water fell out of the sky in the five days beginning on Easter Sunday, March 23. “There was nothing in the meteorological conditions charted on the daily weather maps previous to the downpour of rain that caused the disastrous floods that gave any indication of the tremendous quantity of rain which fell on subsequent dates,” wrote Alfred J. Henry (professor in charge of the National Weather Service’s Rivers and Flood Division) in Monthly Weather Review for March 1913.

But by keeping meticulous and methodical hourly surface measurements, Henry and his colleagues bequeathed to the future a means of reconstructing history.
“To analyze the great 1913 storm system, I used a tool recently developed by NOAA researchers,” Jamison explains. “It uses historic pressure and surface weather data going back to 1871 to model or simulate the weather patterns that produced them.” This computational tool is the ambitious Twentieth-Century Reanalysis Project—or 20CR for short. Developed in the last half-decade by NOAA and various U.S. and international partner agencies to elucidate long-term relationships between weather and climate, its heart is U.S. Department of Energy supercomputers. Among other functions, 20CR enables today’s meteorologists to use historic surface measurements of atmospheric pressure to reconstruct probable conditions in the atmosphere aloft, thereby gaining insight as to the possible physical causes of historic extreme weather events.
Running the numbers through the 20CR computational simulation, Jamison has retraced what happened March 23-27, 1913. The weather story that emerges is detailed step-by-step below in pictures [see sidebar below “The Meteorology of ‘Our National Calamity’”].
In a nutshell, here’s a quick summary. The mammoth 1913 weather system started out innocently enough, following an ordinary winter-storm pattern over the Midwest. Low pressure areas developed over the Rocky Mountains and Texas, drawing in humid air from the Gulf of Mexico; the lows were steered northeast over the Ohio Valley by the jet streams high in the stratosphere. What was extraordinary about the 1913 winter storm system, however, was its sheer persistence and scale. A series of four low-pressure regions that developed one right after the other over the Rockies and Texas, which drew unusually large volumes of tropical moisture north and east over the Great Plains and Midwest. Such unstable air is ripe for forming families of deadly tornadoes. Moreover, the humid air reached saturation: the point at which absolutely no more moisture can be absorbed. Thus, as the sultry tropical air was pulled north over the cooler Midwest, the air cooled and released vast volumes of water in a great deluge. Meantime, a persistent high-pressure system stalled over Bermuda off the Atlantic coast acted as both a barrier and a focusing mechanism to keep the trough of deep low pressure in one place diagonally across the Ohio Valley and state of Ohio. This setup effectively acted like a powerful meteorological water pump for five days, Jamison says: “The conveyor belt of lows coming out of the Rockies just kept the weather pumping.”
Here’s the scary part: “This 1913 pattern definitely could happen again,” declares Jamison. “In fact, whenever we see a slow-moving winter storm pattern of deep lows and blocking highs, that’s an absolute signal there will be significant flooding somewhere around the Midwest, depending on the details of placement.”
What might be different now, a century later, if a winter storm system of the magnitude of the Great 1913 Easter Storm recurred? “First, there would likely be much lower loss of life,” Jamison points out. The death toll from the Good Friday windstorm [see “The First Punch”], the Easter tornadoes [see “’My Conception of Hell’”], and the Great Easter Flood was at least 1,000 (a future installment to this research weblog will tally the natural disaster’s death and destruction across some 15 states.) Because of today’s knowledge and monitoring of the meteorology of the upper atmosphere, she explains, “today we would have several days of advance warning.
“Also, the nation no longer relies just on wireline communications,” Jamison continues. “Even if thousands of poles and wires were blown down by an intense windstorm two days earlier, as happened in 1913, we now have TV, emergency radio, cell phones, satellite broadcasts, and other alternative ways of warning people to prepare or evacuate. On the other hand,” she cautions, “we would likely have far greater property lost to flooding now than in 1913. Runoff would be far higher and faster because the land today is so much more built up with impermeable surfaces such as paved roads, parking lots, and other infrastructure.”
As a result, Jamison is working with dozens of State and Federal agencies that have formed a consortium called the Silver Jackets ( ). The Silver Jackets are using the year-long 2013 centennial of the 1913 Great Easter Flood as an occasion to raise public awareness about flood risk, preparation, and safety.
Great floods happen. The nearly forgotten 1913 Great Easter Flood was one of them. Despite such technical terms of art as “the thousand-year flood” as a means of expressing a mathematical probability of 0.1 percent in any given year, to the general public such phraseology can be misleading. “Expressions of mathematical probability say nothing about when,” Jamison points out. “The chilling truth is: another thousand-year flood like 1913 could happen any time—even this year.”
©2012–2013 Trudy E. Bell. For permission to reprint or use, contact Trudy E. Bell at

The Great Easter 1913 Flood extended far beyond Ohio... Next time: The Prisoners’ Feast

Sidebar: The Meteorology of ‘Our National Calamity’
The Great Easter 1913 Storm System Reconstructed: The Weather Story in Sarah Jamison’s Step-by-Step Pictures for March 23-27
Figure 1 - Good Friday, March 21, 1913: Two days before the downpour began, a severe windstorm swept the eastern half of the nation from the Great Lakes to the Gulf of Mexico. In this chart, Sarah Jamison of the Cleveland office of the National Weather Service plotted wind speeds for half a dozen Midwestern cities.

Figure 2a - Easter Sunday, March 23, 1913: A low pressure region deepened over Colorado while a high-pressure system over New England departed to the northeast. Between those two features, winds would have increased out of the south, pulling warm humid air from the Gulf of Mexico northward into the Great Plains. (Credit: Sarah Jamison, using the Twentieth Century Reanalysis Project [20CR])

Figure 2b - Easter Sunday, March 23, 1913: The Twentieth Century Reanalysis Project (20CR) calculations show that low-level winds (pressure 850 millibars, corresponding to an altitude of about 5,000 feet) would have produced very strong flow out of the Gulf of Mexico. Such strong winds would have increased wind shear (sudden changes in wind speed and direction) and supported severe and tornadic storms, and brought a high volume of tropical moisture over the Ohio Valley. (Credit: Sarah Jamison, using 20CR)

Figure 3a - Monday, March 24, 1913: As the first low weakened and moved northeast, a second low-pressure region developed and strengthened over the Rocky Mountains. Meantime, a third low formed over Texas. “All,” emphasizes Jamison, “had the Ohio Valley in their tracks.” (Credit: Sarah Jamison, using 20CR)
Figure 3b - Monday, March 24, 1913: Because of the pattern of repeating low-pressure systems drawn up and across the Midwest, low-level winds (altitude about 5,000 feet) would have increased, continuing to feed moisture northward out of the Gulf of Mexico. (Credit: Sarah Jamison, using 20CR)
Figure 3c – Monday, March 24, 1913: Just how much moisture was there? This plot by Jamison shows the precipitable water, or water content calculated to have been in the atmosphere above the Midwestern and Eastern United States at the time of the downpours—between 1.2 and 1.4 inches over Ohio. (Credit: Sarah Jamison, using 20CR)
Figure 3d – Monday, March 24, 1913: Cooler air can hold less water than warmer air. This plot by Jamison shows the climatology of precipitable water over Dayton. “A value of 1.2 to 1.4 inches in late March exceeds the 99th percentile,” Jamison points out. “Essentially, the air mass was saturated as much water as it could hold for that temperature and time of year” leading to the great deluge. (Credit: Sarah Jamison)
Figure 4 – Tuesday, March 25, 1913: First day of horrific flooding around Ohio and other states. “At the surface, two areas of high pressure were locking a front over the Ohio Valley,” Jamison explains. “This area of convergence became the focus for heavy record rains.” (Credit: Sarah Jamison, using 20CR)
Figure 5 – Wednesday, March 26, 1913: The trough was extending over the Northeastern states, and creating record flooding over rivers in New York, including record high water in the Hudson River, which crested Friday, March 28 (to be the subject of a future installment). (Credit: Sarah Jamison, using 20CR)
Figure 6 – Thursday, March 27, 1913: “The surface low was over new England as cold high pressure built in from the west,” notes Jamison. “Cold air and snow accompanied the front, bringing more misery to survivors with no power or heat.” (Credit: Sarah Jamison, using 20CR)

Selected References
First and foremost, thanks go to Sarah Jamison for her time and generosity in supplying information and graphics on her work in reconstructing the Great Easter 1913 Storm System.
Monthly Weather Review, vol. 41, No. 3, March 1913. Alfred J. Henry’s article “Rivers and Floods, March 1913” starts on page 485. Throughout all the reports, especially striking is how accurately the 1913 meteorologists reconstructed what happened, even absent data or knowledge about the upper atmosphere, satellites, and computers.
“ReanalysisProject Targets Once and Future Weather,” Sept 29, 2009, is a lay introduction to the Twentieth Century Reanalysis Project (20CR) at the DOE Office of Science.

More in-depth articles by the 20CR principal investigators include Compo, Gilbert P. Jeffrey S. Whitaker, and Prashant D. Sardeshmukh, “Feasibility of a 100-Year Reanalysis Using OnlySurface Pressure Data,” Bulletin of the American Meteorological Society 87 (2): 175–190 (February 2006) and Compo, G. P., et al., TheTwentieth Century Reanalysis Project,” Review Article, Quarterly Journal of the Royal Meteorological Society 137: 1–28 (January 2011, Part A). Sites directly to the Twentieth Century Reanalysis Project (20CR) tool include and .

Bell, Trudy E., The Great Dayton Flood of 1913, Arcadia Publishing, 2008. Picture book of nearly 200 images of the flood in Dayton, rescue efforts, recovery, and the construction of the Miami Conservancy District dry dams for flood control, including several pictures of Cox. (Author’s shameless marketing plug: Copies are available directly from me for the cover price of $21.99 plus shipping, complete with inscription of your choice; for details, e-mail me at )


  1. You've bought up some really interesting points here. Really hard to find the right post to know about 1913 tornadoes.
    aqua read