How do you know when to evacuate and when to stay put during emergencies?

Evacuations shield populations from the impacts of storms, chemical releases or terrorist acts. Advising residents to remain in place, close windows and turn off air conditioning or heating units until the danger has passed is another form of protective action — one used recently in the Graniteville train accident.

Evacuations pose several challenges that are important in ensuring a timely and effective emergency response.

Hurricanes Katrina and Rita illustrated some of the problems with the execution of mandatory evacuation orders, providing lessons for the nation and for South Carolina:

• It is important to identify your population at risk (including tourists and the homeless), as well as special-needs populations (the infirm, those without cars, the elderly). Both may require extra time and resources to adequately evacuate. We saw that in New Orleans these special needs populations largely were ignored, while in the evacuation of coastal Texas (and Houston) great care was taken to get these people out of harm’s way early in the process.

• Cities and high-density coastal areas take time to evacuate. While it may seem strange to many residents to begin preparations to evacuate coastal areas days in advance of an approaching hurricane, these heavily populated regions require significant lead times to get all evacuees off the roads prior to the arrival of tropical force winds.

The clearance times for coastal areas, among others, are a function of the road network and density, traffic volume and population densities. For many heavily populated coastal areas, clearance times could approach 36 to 48 hours — or two full days in advance of the tropical force winds, let alone the storm itself.

Uncertainties in hurricane prediction do not allow us to pinpoint exactly where the storm will make landfall. At the time the local or state official (in South Carolina, it is the governor) must make the call for a mandatory evacuation, he relies not only on the best science and staff judgment available, but also coordination across state agencies and with county and local government. In this way, all parties have buy-in once the decision is made; more importantly, all are aware of what needs to be accomplished and when. The lack of coordination between local, state and federal officials that we saw in New Orleans would not happen in South Carolina.

Emergency managers always err on the side of precaution in evacuations that, in hindsight, may appear to be exaggerated, but at the time of the initial decision it was prudent. It is better to order an evacuation and cope with the cry wolf syndrome (such as what happened somewhat in Houston) than to call for a mandatory evacuation too late (such as New Orleans) and leave residents to fend for themselves, especially when so many did not have the resources to do so.

• Not everyone needs to evacuate. Some locales, such as barrier islands or low-lying flood-prone areas, are more hazard-prone than others. Some people are more vulnerable than others (the elderly, the infirm, small children or people living in mobile homes) as well. Evacuation zones should be targeted to specific areas and groups and evacuated sequentially, starting with those most at risk.

In the case of Texas, evacuation zones were based on distance from the coast, with Galveston being evacuated first, followed by successive inland areas. Nursing home residents were evacuated first in Texas before the general public.

In the case of New Orleans, there was no geography to the evacuation, and most residents (who could) left at about the same time, clogging roads and creating miles of gridlock. More specific targeting of populations and geographic areas for mandatory evacuation results in better compliance with those orders.

• Residents who do not live in a high-risk area (or the mandated evacuation zone), yet feel the need to evacuate, produce a “shadow evacuation.” Shadow evacuations are common and strain the emergency management system as more people (and cars) get on the road, add to traffic jams, reduce gasoline supplies and fill shelter options (such as motels) that could be used for people who live in riskier areas.

In South Carolina, evacuation shadows were found for Hurricane Floyd (around 28 percent of residents in affected non-coastal counties evacuated), and for the Graniteville chlorine spill an evacuation shadow also was found (59 percent of residents outside the mandated one-mile evacuation zone also left). It is unclear how large the evacuation shadows were for Katrina and Rita.

South Carolina has a good state evacuation plan, one that is improved every year. As a state, we do learn from our evacuation experiences — what went right, what went wrong — and we use that knowledge to improve plans for the next event. As we saw during the past month, it is much better to be safe (Galveston and Houston for Hurricane Rita) than sorry (New Orleans in Hurricane Katrina).

Dr. Cutter is a Carolina Distinguished Professor of Geography and director of the Hazards Research Lab at the University of South Carolina.