Dr. Seth Guikema

Hopkins assistant professor Dr. Seth Guikema has been awarded a National Science Foundation grant to create a computer model that will determine the effect of repeated hurricanes and heat waves on the mid-Atlantic region. (Christopher T. Assaf, Baltimore Sun / January 15, 2014)

For openers, a hurricane will likely be unleashed on one neighborhood in Anne Arundel County, and then the disasters will spread from there.

The storm will lash the area with heavy rain, whipping wind, floodwater and power outages, causing road closings and evacuations. The event will be a big mess as it unfolds on a computer map for the benefit of public officials wanting to know how local decisions — on such matters as zoning, building codes, transportation, perhaps even the color of rooftops — affect their community's ability to withstand repeated lashings of extreme weather.

Computerized storm simulations are not new, but this work in the era of climate change aims to take things further with contributions from experts in a wider range of fields. An assistant professor of geography and engineering at the Johns Hopkins University leads a team that includes researchers in public health, civil engineering, earth sciences, human behavior, even a landscape architect from the Maryland Institute College of Art.

"What's new is the integration of all these areas," said Seth Guikema, the Hopkins professor who is principal investigator on this four-year project undertaken with a $3 million grant from the National Science Foundation. It is one of 12 grants totaling $32 million announced by the federal agency in the fall, all emphasizing interdisciplinary research to improve prediction and response and soften the impact of natural disasters.

The studies include work on earthquakes, tsunamis, wildfires, tornadoes, landslides and volcanic eruptions. Guikema's team is focusing on repeated hurricanes and heat waves and their effects in the Mid-Atlantic region, roughly from Virginia to New York.

The work likely will begin with a neighborhood in Anne Arundel County, with its extensive coastline, for the most detailed computer model, Guikema said. The precise area has not been defined. Others areas in the Mid-Atlantic model won't be as detailed, he added.

Some of the $3 million is going to pay for more computing power, but there's a limit to how precise the models can be across a region of several states and tens of millions of people.

For the Arundel neighborhood they can be quite detailed, accounting for every street and driveway, every house, any storm- or heat-resistant features they might have, as well as every store, strip mall, parking lot and public park.

While the model will include homes, commercial and government property, it won't account for heavy industry, because the group decided to limit its scope.

Figuring out what to include in a computer model is a balancing act, Guikema said. Put in too much information and results can be difficult or impossible to interpret. Put in too little and the model poorly reflects reality and doesn't produce useful information.

"All models are approximations of reality," Guikema said. The question is, what "can we safely leave out?"

The work has just begun, but Guikema imagines that the team will create something like a more sophisticated version of a computer game in which players build their own city and see the consequences of their choices.

"It's kind of like Sim City on steroids," and based on real research, he said, for an audience not of gamers but public officials.

The idea is to help authorities see how policy decisions affect the community's ability to cope with extreme weather — now and for decades to come.

Decisions on land use — say, lot size and proximity to coastline — could be entered into the program, and the consequences in terms of storm damage could be shown on a computer map during a simulated storm. Specific house designs and storm-protection features could be tested, along with evacuation plans, transportation networks, the impact of development on flood damage.

Johns Hopkins professors will contribute work on public health impacts, coastal flood surges, climate and the technicalities of building the model itself.

The team includes a landscape architect who teaches environmental design at MICA and an associate professor at Georgetown University who studies how people perceive and respond to risks, including those posed by storms.

Katie O'Meara of MICA said she'll be drawing on the research she's done in this country and overseas on how urban social and physical patterns change over time. For this project, she'll start with current population patterns from South Carolina to New York and go back in time.

At the moment, it's simply a matter of taking stock of "who's where, who's vulnerable" to potential storm effects, including sea-level rise associated with climate change. Eventually, she'll want to consider whether historical patterns can be helpful in looking into the future, as the computer model will be designed to do.

Robin L. Dillon-Merrill of Georgetown University's McDonough School of Business will contribute work on how people behave in emergencies. For years she's studied how people respond to risky situations, warnings of impending danger and the experience of evading a hazard, or what she calls the "near-miss event."

She's looked at how people reacted in the aftermath of Hurricane Katrina, and after disaster was averted when the so-called "underwear bomber" failed to detonate explosives aboard a commercial airliner in 2009.

Her work tells her that people who ignored storm evacuation warnings in the past and emerged unscathed are less likely to heed evacuation orders in the future. Emergency management officials might want to take that into account as they plan disaster response and evacuation warnings.

"We hope to test some of the assumptions they would make as far as the people part of it," Dillon-Merrill said.

The work will be completed in stages, Guikema said, with the first results — probably that neighborhood model of Anne Arundel County — released in about a year.

arthur.hirsch@baltsun.com