10 NOVEMBER 2005
or
some reason, 2005 has been the year of living dangerously.
Or maybe for no reason at all. Maybe it's just bad luck, with random earthquakes and hurricanes just striking with more power and ferocity than usual. Bad things happen. But it somehow seems that the past 12 months have produced more disastrous catastrophes than random rolls of the dice would foretell.
Just days before the new year arrived, an earthquake and tsunami rattled the planet and devastated coastal areas in the South Asian Sea region, killing about 300,000 people. More recently hurricanes Katrina and Rita have battered the U.S. Gulf Coast, and the Kashmir earthquake has killed thousands and terrified millions. Now a bird flu pandemic waits in the wings, threatening to overshadow earthquakes the way atomic bombs outblast dynamite. With luck like this, the planet Earth needs no alien enemies.
And maybe 2005 isn't really unusually bad, but just a sign that catastrophes are becoming more common. From floods to power blackouts, flu epidemics to stock market crashes, the world constantly confronts calamities. Scientists, though, have confronted disaster only sporadically.
"Despite many reports on disasters, the scientific investigation of their general features and ways to fight them is still needed," write physicists Dirk Helbing, Hendrik Ammoser, and Christian Kühnert of Dresden University of Technology in Germany.
They note that science has explored several ideas for explaining the increasing occurrence of disasters, including chaos theory, theories of complex self-organizing systems, and more recently the mathematics describing networks.
Nowadays many experts think the network approach is especially promising. Most natural systems are not simple link-to-link chains, but elaborate networks, with multiple pathways connecting the different parts. Disasters happen when the one part of a network goes bad, adversely affecting other parts via those multiple pathways. Chain reactions that result lead to explosive consequences.
"It is often these cascade-like chain-reactions, by which a localized event in time and space causes a large-scale disaster, which may affect . . . people at far remote places of the world," Helbing and collaborators write. And as modern society's networks grow more elaborate, the opportunity for catastrophe grows.
"The tendency of globalization of economic and other systems is likely to increase the frequency of large-scale disasters, as it reduces the diversity required to stop certain chain reactions and to adapt to changing economic and environmental conditions," the physicists note.
Short of returning to the 19th century (or even the 20th, and things weren't so great then, either), coping with disaster demands a better understanding of mathematics of the networks underlying chain-reaction catastrophes. In a paper available online, Helbing and colleagues sketch out several examples of networks underlying disaster scenarios and show how mathematical reasoning can inform the proper public response.
In battling an infectious disease epidemic, for instance, critical decisions arise about proper strategies for vaccinating people or isolating them. Should transportation systems be shut down to keep infected people at home? Maybe, but that might make it hard for health care providers to get to work. Not to mention the garbage collectors. "Waste may contribute to the spreading of the disease, if it is not properly removed," the physicists point out. Shutting down the transportation system will also damage economic activity substantially, which in turn can exacerbate other aspects of the epidemic. Of course, as more people get sick, economic activity also diminishes, and the transportation system and waste disposal will be curtailed by a lack of healthy workers.
You see, disasters really are bad.
But the point is that minimizing the impact of a disastrous epidemic depends on knowing how all the parts of the disaster network fit together. You can't just guess how one strategy will spread its effects throughout the network -- you have to do sophisticated calculations.
Some sample simplified calculations begin to show how to go about dealing with such issues. Say, for example, vaccine is in short supply. How should it be distributed? It turns out that vaccinating bus drivers does not help the situation as much as vaccinating the health workers and waste disposal staff.
"We see that it is more effective to immunize the medical staff than the disposal workers, and the best would be to immunize both groups," Helbing and collaborators show.
Understanding cause-effect networks can help predict what strategies will be
most effective, and should make it possible to identify key steps in the chain
reaction where preventive action might abort a worst-case scenario. 
Beyond
that, the network approach can be applied more broadly to improve overall disaster
response management practices. In responding to a disaster, various networks
must be called into action, including transportation, supply and information
networks. Disaster response involves coordinating a network of networks. The
math can get complicated, but it will surely beat the haphazard approach that
has failed so miserably in recent disaster responses.
In fact, disasters demand a better attitude toward math; the notion that knowing math really isn't very important is one of the reasons the nation isn't very well prepared to cope with catastrophe. So perhaps FEMA should change its name to FEMATH. Federal Excellence in Math would be a valuable agency for ensuring a safer future.
E-mail: tsiegfried@nasw.org
