Three years ago this September, a storm surge driven by winds from a weakening Hurricane Isabel produced the worst Chesapeake Bay flooding in 70 years.
Waters up to 8 feet above normal tides surged into lower Fells Point and across Pratt and Light streets into downtown Baltimore. Hundreds of basements and businesses flooded.
Hundreds of homes in Bowleys Quarters and elsewhere on the bay shore were badly damaged or destroyed. Property damage reached $410 million in Maryland alone.
But as destructive as Isabel was, recent computer simulations by government scientists - the most extensive ever for the Chesapeake Bay - show that hurricane storm surges here could get much, much worse.
Under some conditions, they discovered, a Category 4 hurricane making landfall in the Carolinas could produce storm surges as high as 18 or 20 feet in Baltimore at high tide. That's at least 10 feet - a full story - above Isabel's high-water mark and enough to carry floodwaters much farther inland.
The storms were simulated on the latest version of the government's computerized SLOSH model. The name is an acronym for "Sea, Lake and Overland Surges from Hurricanes," produced by the National Weather Service and the National Hurricane Center.
The projections show that 18-foot storm surges are also possible all along Baltimore County's shoreline. Parts of Harford and Anne Arundel counties, and the upper tidal reaches of the South, Severn, Patuxent, Potomac and Anacostia rivers are just as vulnerable.
"I guess I'm a little surprised the values are as high as they are," said Wilson A. "Will" Shaffer, chief of the National Weather Service evaluation branch in Silver Spring and a leader of the project.
The precise combinations of tide, storm intensity, track, size and forward speed needed to generate an 18-foot storm surge on the Chesapeake are surely rare, Shaffer cautioned - nothing like it has ever been recorded.
But New Orleans had never before experienced anything like Katrina's flooding, either.
Here, with the new SLOSH model data, the conditions needed to produce an 18-foot storm surge are "within the realm of possibility," Schaffer said. And emergency managers must consider them when they plan for storms.
Those managers agree.
"A 20-some-foot storm surge up the bay is not something we want to take lightly," said Robert A. Ward, all-hazards planner for the Maryland Emergency Management Agency. The new simulations are "a useful planning tool for us that we are going to take seriously."
The agency is working with local officials and the U.S. Army Corps of Engineers to turn the simulation data into updated maps for emergency planning. The new maps will demonstrate how much farther inland storm-surge floods might reach under the right conditions.
"We have to make sure, as emergency managers, that we understand where those areas are, and provide adequate warning to those residents," Ward said.
Managers also will use the data to rethink where and when to order evacuations. They'll reconsider where they can safely establish operations and evacuation centers, store emergency supplies and park emergency vehicles as storms approach.
Shaffer is scheduled to present his findings to the state's emergency managers at a training session Thursday at MEMA headquarters in Reisterstown. He'll also demonstrate the SLOSH model software.
"We are involving a lot more people in this SLOSH training ... so we can be better prepared for these situations," said Lt. Mark Demski of Baltimore County's Fire Department and its office of Homeland Security Emergency Management.
Although he has not seen the most extreme SLOSH simulation, Demski acknowledged that the old model failed to predict the extent of surging with Isabel, which made landfall as a Category 2 storm.
The county has extremely vulnerable neighborhoods in its low-lying, southeastern waterfront communities. In fact, he said, 32 months after Isabel, "we're still in recovery. ... We still have residents that are rebuilding."
The 2006 Atlantic hurricane season opens officially in 4 1/2 weeks. The National Hurricane Center will release its first seasonal forecast May 22. Elsewhere, scientists have predicted that it will be busier than average but not as frenetic as 2005, which saw 27 named storms and 15 hurricanes.
William M. Gray and Philip J. Klotzbach - Colorado State University's famed hurricane prognosticators - predict 17 named tropical storms this year, nearly twice the norm. Nine of those will become hurricanes, they said, and five will become "major" storms (Category 3 or higher, with sustained winds above 110 mph).
Gray's forecast also lists a 64 percent chance that a major storm will strike the East Coast, twice the average risk over the last century.
But it was the memory of Isabel that Shaffer was channeling recently at a computer display of the SLOSH model in his Silver Spring office. With a few clicks, he produced a color map of the Chesapeake region, then conjured up a simulation of the 2003 storm.
Isabel reached fearsome Category 5 status over the Atlantic Ocean, with top sustained winds of 149 to 161 mph. A brief gust of 233 mph, measured 4,500 feet above the ocean, was the strongest wind ever measured in an Atlantic hurricane.
To the relief of everyone on the coast, Isabel weakened to a Category 2 storm (96 to 110 mph) before it made landfall near Drum Inlet, N.C., around noon Sept. 18.
By the time its center reached southern Virginia, it had degraded to a tropical storm. The highest sustained winds at Thomas Point, Md., were about 48 mph, with gusts to 66 mph.
But as Isabel's weakening remnants spun northward, its center tracked to the west of the Chesapeake. That put its strongest winds over the bay, driving the water north from the Virginia capes - a recipe for big surges.
On Shaffer's computer screen, the blue water in the lower bay begins to turn green, then yellow. These "warmer" colors represent higher water levels. Yellow signals an 8-foot storm surge.
As time advances and the simulated Isabel moves north on the computer, the surge follows, moving up the bay and into the tidal rivers like a ripple on a pond. It takes 14 hours for a storm surge to move from the bay's mouth to the Susquehanna River.
The hottest colors on Shaffer's screen eventually appear at the heads of the tidal rivers and in the upper bay - 8 feet or more at Washington, Bowleys Quarters, Middle River, Fells Point, at the mouth of the Gunpowder River and the north end of the bay.
Shaffer said water tends to "pile up" where the bottom gets shallower and the long tidal rivers - and the bay itself - grow more constricted.
"Isabel surprised us," said MEMA spokesman Edward J. McDonough. "When Hurricane Isabel hit, the storm surge was significantly higher than all the forecast models forecasted. That was the genesis of updating the SLOSH model."
After the storm, a National Weather Service assessment of its forecasters' performance during Isabel also found significant shortcomings.
Although they generally did well with the storm, the review concluded that forecasters in the Baltimore-Washington office in Sterling, Va., failed to provide residents with enough timely and specific information about what the rising storm surge would mean, and where.
"People had no experience with major storm surges on the bay, and as a result they didn't know what to expect with an 8-foot surge," Shaffer said.
SLOSH was invented in the late 1970s to help provide that kind of information. Its creator was Chester P. Jelesnianski, a weather service surge modeler with a penchant for acronyms, Shaffer said. (SLOSH was the successor to Jelesnianski's SPLASH model, for "Special Program to List the Amplitudes for Surges in Hurricanes.")
It takes into account the bay's geography and bathymetry (the shape of its bottom), the storm's intensity, track, size and timing relative to the predicted tides.
It has been updated and refined repeatedly over the years, taking into account discoveries in storm physics, new data on bay terrain and bottom soundings, increasing computer power and storm experience.
It's all captured in complex mathematical formulas that run on powerful computers at the National Hurricane Center in Miami. Scientists typically run the model with storm simulations for long-range planning, and again for forecasters using real storm data, within 36 hours of landfall.
The last SLOSH simulation study was in 2000, with 4,000 to 5,000 storm scenarios tested. After noting the model's shortcomings during Isabel, scientists tweaked the formulas, Shaffer said, then ran 50,000 storm simulations over 18 months.
"Part of the reason for running so many hurricanes is to be sure to have the truth for what we think is the potential flooding for the entire bay," he said.
The new data increased - to extremes no one ever envisioned - the "maximum envelopes of water" expected with storms of increasing strength.
The most sobering outcomes emerged from a composite of storm scenarios, all making landfall in the Carolinas at Category 4, and each producing a storm surge arriving with high tide in the Chesapeake.
The resulting map shows not what any one storm would do, but the worst flooding that could be expected, and where, under such conditions.
Beyond the 18-foot surges in Baltimore, a storm making landfall in the Carolinas at Category 4 could generate potentially devastating floods of 14 to 16 feet on the shores of Kent, Queen Anne's and southern Anne Arundel counties.
Under the right conditions, even Category 3 storms could produce 13-foot storm surges in Baltimore, 15 feet in the South River and in Cecil County, and 18 feet at the mouth of the Gunpowder, the simulations revealed.
There's a message in the findings, said David R. Manning, warning coordination meteorologist in the weather service's Sterling forecast office.
"Isabel was not the strongest hurricane that we could see," he said. "The new study, with all its simulations, is new knowledge. And knowledge is power to those who need to make plans about the eventuality of another tropical cyclone affecting the area."