CFS Press

    This article originally appeared in the January-February issue of Advanced Rescue Technology Magazine

    A New Flood Rescue Response Model

    by Slim Ray

Many of the flood responses in the past few years have fallen into the "too little, too late," category. The scenario is distressingly familiar, especially for flash floods: as the flood hits, residents and local emergency workers struggle to survive and rescue others. Emergency managers struggle under a staggering call load, trying to separate the desperate from the trivial. They call outside resources, but by the time they arrive on scene, the flood is mostly over. This scenario has played itself out many times—in Ft. Collins, Kansas City, the floods of Hurricane Floyd in North Carolina, and elsewhere.

In the aftermath the response is generally justified with "we did all we could" and "there's just no way you can plan for something like this." Yet much of the problems emergency managers have encountered is due to using an antiquated and inadequate response model. In this article we will explore a newer, more responsive model. But first, let's look at some of the characteristics of floods.

Floods generally come in two flavors: river floods and flash floods. Many floods, such as the recent ones in North Carolina in the wake of Hurricane Floyd, have characteristics of both. In general, a river flood is just what the name implies: too much water in a river's watershed. It often affects a wide area, quite often an entire region, and has a slower onset and retreat than a flash flood. This leads to a different time problem than a flash flood, which comes and goes in a hurry. In river floods a major problem is exhaustion of the rescuers over an extended time period.

A flash flood, on the other hand, tends to be an intense but short-term event. The entire incident may last only four to six hours from start to finish, or it may be the start of a long-term river flood. One of the salient characteristics of a flash flood is that there are a very large number of life-threatening incidents in a very short time (more about that later). A flash flood may affect an entire region, or it may be localized to only a few city blocks. Time, always, is the rescuer's enemy, and the major management problem is that there are never enough rescue resources to go around.

All floods, generally speaking, have three phases, which often overlap:

Swiftwater Phase: this initial phase is where the most fatalities (and the most rescues) occur. Large numbers of people, including those in rescue agencies, are caught unaware, often in darkness. Many are swept away. Statistically the two largest groups are people killed are 1) those driving through moving water and 2) children playing near flooded creeks and flood channels. The sheer number of incidents overwhelms dispatch and available units.

Flood Phase: the situation stabilizes somewhat, but there is standing water everywhere and rivers may continue to rise, further restricting access. Those who are able have rescued themselves, at least gotten to positions of less immediate danger like the roofs of houses or the tops of cars. Because of the flooding, rescue units cannot access many locations. Large numbers of people may have to be evacuated from flooded areas and are displaced to shelters. In a river flood, this phase may be prolonged, sometimes for months.

Recovery Phase: the water begins to recede and people return to their homes. The dangers of this stage are more indirect—downed power lines, debris, disease, contaminated water and food, etc. The are major infrastructure problems—washed out roads; loss of power; lack of clean water. Rebuilding begins.

What rescue resources might we expect to use in a flood? The most obvious would be specialized water rescue teams. These, especially any used in the initial phases of the flood, must have swiftwater-specific training. Dive teams or those with only generic water rescue training only put themselves and the victims at risk in fast-moving water. However helpful these specialized teams are, however, there will never be enough of them to go around in flood event. To supplement them, we first need to look close to home.

There are many non-specialized rescuers in any flood zone. These include law enforcement personnel (police, sheriff, etc.), firefighters, rescue squads, and EMS personnel. However, these people need at least minimal training and equipment for simple rescues and to protect themselves. While all these groups will protest that they don't do water rescue, the simple truth is that during a flood, everyone will be doing it. Do not overlook other local resources, like municipal storm water personnel; power crews; and other city/county/state employees whose duties take them into the flood area.

Outside resources include mutual aid from surrounding jurisdictions (who may be having their own problems), and military assistance from the regular military and National Guard, and Coast Guard. Don't overlook qualified volunteers, especially kayakers and raft guides from river companies. These outfits often have a great deal experience in moving water and useful equipment, like rafts and PFDs, also.

Educate your dispatchers on the basics of flood and swiftwater rescue. They can’t make decisions on what to send where unless they know how the system works. Any dispatch system needs a ramp-up capability; that is, the ability to expand its capacity from normal call volume to several hundred calls in a short time. In addition, it is critical for any dispatch system to be able to triage incidents; that is, to separate the trivial from the life-threatening. The newly-devised Natural Disaster Information Cards (see sidebar) are a great help in doing this.

Developed by meteorologist John Weaver at Cooperative Institute for Research in the Atmosphere (CIRA) at Ft. Collins as the result of the flood there in 1997, the Natural Disaster Information Cards are an invaluable tool for helping dispatchers and managers triage incidents and for helping citizens cope with emergencies. They can be downloaded free at and modified to suit local needs.

Incident commanders and emergency managers also need a working knowledge of flood and swiftwater rescue, so they will know what works and what doesn't. Agencies need to have common maps and enough command interoperability to at least be able to talk to each other. A good way to do this is to schedule joint exercises. Most emergency managers, used to the routine of daily operations, have had little experience in managing a large incident.

Okay, most of these things so far are common sense, although unfortunately few agencies do them. But let's look at the response model, especially for a large incident. How do we manage a large flood incident to get rescuers where we need them? One thing we’ve already noticed is the lack of time, especially in flash floods, and the fact that most casualties happen in the first few critical hours.

The traditional response method is often called the "pull" system. In this model, response units remain at their home station until called. Once a disaster occurs, it is sized up and resources are identified to handle the problem. These are then requested through proper channels, alerted and dispatched. With volunteer units, the volunteers must first come to station, don their gear and draw their equipment, then proceed to the incident location. In a large-scale disaster, outside resources are often not notified until local units realize that they can’t cope with the situation, causing additional delays.

The advantage of the pull system is that it is a very efficient way to allocate resources. Rescuers are sent only to actual incidents, and then only as many as are needed. For routine situations this works just fine.

The big disadvantage of the pull system, however, is that it is slow. The cycle of size-up, request, approval, selection, alert, and dispatch is called a decision loop, and it takes time. The more people and the more command levels in the loop, the longer it takes to get people on the road. In a flood this often means people get there too late, and when they do get there, they may not be able to reach flooded areas.

This problem also surfaced in the US Army in the early 1980s, especially for battlefield logistics. Military logisticians found that on the fast-moving air-land battlefield, the pull system was just too slow. By the time supplies had been requested up through the chain of command, authorized, loaded, and sent down to the units, the situation had changed radically and the original requests were no longer appropriate.

To deal with this the army came up with the "push" system. Under this system, unit logisticians would monitor the situation and make an estimate of what the units needed, then dispatch it without being asked. It was not as efficient as the pull system, because the requirements were not exactly known, and sometimes the logisticians guessed wrong. But it was much, much, faster and more responsive.

So what has this got to do with floods? Experience has proved two things—things happen in a hurry and large areas are quickly cut off by rising water. This has meant in practice that the people who are on the spot make the rescues, ready or not. There simply is not time to mobilize and deploy after the flood starts, and often impossible to get there.

By adopting a "push" system, managers have a much better chance of getting people where they need to be. As flooding threatens, units are predeployed to known or suspected trouble spots as determined by weather predictions, historical data, pre-incident surveys, and agency preplans before flooding begins. Only in this way will they be where and when they are needed. It also works in reverse, allowing managers to get citizens out of areas sure to flood.

A good example of this is the system used in Los Angeles. When weather conditions are right for flooding, strike teams are dispatched to walk the flood channels, and swiftwater teams are positioned in likely trouble spots. This cuts out the time needed to identify the incident, dispatch, and travel to the incident. This way they can begin rescue immediately.

Obviously this is a much more difficult challenge than reacting to an existing incident. It requires considerable expertise on the part of managers and a close working relationship between weather forecasters and emergency services—much closer than is now generally the case.

Plans and Preparations

The Chinese military philosopher Sun Tzu said many years ago that most battles are decided before either army sets foot on the field. So it is with floods, and perhaps the most important arrow in the flood manager's quiver is his preplan. Unlike natural disasters like earthquakes and tornadoes, floods can often be predicted. Furthermore, floods are constrained by geography. It's possible, especially in these days of GIS software packages, to map low areas and watersheds and to compare this with real-time meteorological data.

An agency's preplan needs to be a realistic, worse-case scenario of flooding in an entire region. What areas are at risk? In many cases your best data is historic, since some low areas flood every time there is a heavy rain. Where have rescues and drownings occurred in the past?

Where will the water go? You will need maps of the entire watershed of any rivers and creeks, even if they go out of your jurisdiction. When planning deployments, try to anticipate which roads will be underwater, both so they can be blocked off and so that you won't try to deploy through them. Try to choose pre-deployment spots that won't go under water also.

Good maps are a necessity. Federal agencies like U.S. Geologic Survey can be a great help here in showing you where the flood plains are. Don't neglect the smaller watersheds in your jurisdiction, since creeks will, with enough rain, become rivers. As mentioned before, common maps for all agencies are a necessity.

Survey your jurisdiction for obvious flood hazards:

  • Low bridges: These need to be blocked off. Seven people died in one incident in Kansas City when three cars were swept off a low bridge. Low bridges can also become blocked with flood debris and turn into dams, then send a surge of water downstream if they let go.
  • Dams: Low-head dams can become "drowning machines" during floods or high water. How would you rescue someone in one?
  • Flood channels: These present a difficult and dangerous rescue problem, with fast-moving water (up to 30 mph) and smooth banks.
  • Low water crossings: Very common in many western states, these need to be blocked off or at least monitored during high water.
  • Creeks and watersheds (watershed maps): Do you know where the water goes? That insignificant creek running through town may become a major river after a storm.
  • Storm water systems: People and even trucks have been sucked down storm drains.

Remember, though, that your flood problem is never "solved." Like any plan, it has to be periodically re-evaluated. Development, in the form of paving, roads, housing, malls, and other commercial development, can radically change flooding and drainage patterns. So can flood control measures. Often flood controls in one area just move problem somewhere else. Some flood protection measures, like low head dams and flood channels can greatly increase danger to rescuers.

There is one more vital piece of the puzzle—weather prediction. Just as a general must know as much as possible about his opponent, you must know about the weather. Only then will you as a manager be able to accurately allocate resources and make decisions.

New advances in weather forecasting, like Doppler radar and telemetric rain gauges, now make it possible to get real time information about exactly what is happening in the heavens and on the ground. You can now literally watch as storms pass over, filling watersheds and flooding creeks and rivers. For the first time, an incident commander can anticipate a storm's next move. But to do this, you need to develop a "weather intelligence" program. This means establishing a close relationship with weather forecasters, even to having them sitting at your elbow during the incident. In several incidents, large and small, the operation has come to grief because weather information did not get to where it was needed in a timely fashion.

CIRA meteorologist John Weaver calls this "tactical meteorology." By knowing what the storm is doing and when, incident commanders can greatly reduce the duplication of effort inherent in the push system.

If you've done your homework on where flooding is likely to occur, and are able to make a tactical intelligence estimate on where the rain is coming down and where the storm is moving, you can predeploy resources before the roads flood out. You can also warn people in the path of the flood and evacuate them if necessary. One recent innovations is the development of the "reverse emergency calling" systems. These software packages allows EMS to automatically call households in a selected area (which they have determined is going to flood) and play a pre-recorded message warning residents of the danger and perhaps ordering an evacuation. It will never replace knocking on doors, but it does give residents critical advance warning time.

Don't forget the media. It can be friend or foe, but it is absolutely vital for public education. Insist on a two-way relationship. If they want you to keep them informed, they need to help you keep the public informed. Citizens should be told the basics of flood safety. Probably the most important thing they need to know is not to drive through flooded roads, since the majority of flood fatalities happen in cars. Some states have produced videos to discourage this, and maybe you can get the local TV station to do a spot.

Don't wait until the flood hits and then try to figure out what to do. This has been tried and it doesn't work. Modern management techniques make it possible to get ahead of the flood, and that's where you must be for effective flood incident management. If you don't, you're sunk, literally. Once you get behind, you'll never catch up during those critical opening hours, where you are most likely to lose the most people, both civilians and your own.

To reiterate the differences between old and new models:


  • Reactive
  • "Pull" system — resources sent only when requested
  • Incident command designed to deal with serial incidents, copes with multiple simultaneous incidents only with difficulty.
  • Long response time
  • Responds to incidents already in progress
  • Resource deployment hampered or prevented by flooded roads
  • Efficient use of resources
  • Mangers trained "on the job" for routine operations only


  • Proactive
  • "Push" system — attempts to anticipate needs rather than reacting to them
  • Extensive preplanning and use of weather intelligence data
  • Incident command able to deal with multiple simultaneous incidents, has "ramp-up" capability
  • Short response time
  • Reacts to predictions of flooding
  • Resources predeployed when possible
  • Less efficient use of resources, some duplication unavoidable
  • Emphasis on multi-agency, multi-jurisdictional operations
  • Managers trained for large-scale operations

Let's look at some actual examples.

In the floods in Fort Collins, Colorado (1997) and Kansas City (1998) the dispatch system quickly became clogged by the large number of incidents. Emergency managers ran out of local resources in short order. In both cases the response model at the time did not allow expansion of dispatch capability to cover the increased number of incidents, nor was outside help summoned until it was too late. In Kansas City there was no agency preplan or much interagency cooperation, leading to a failure to close access over a low bridge. Seven people in three cars were washed off the bridge and drowned.

The next year North Carolina found its response model was simply inadequate for dealing with a disaster the size of Hurricane Floyd. State policy is to relegate rescue responsibility to the counties, with state emergency managers acting only as resource coordinators. When rains from the hurricane swamped the eastern part of the state, county emergency management was quickly overwhelmed, but state policy prevented mobilizing resources until the counties actually asked for them, virtually guaranteeing that adequate rescue resources would not be available during the first critical night. By the time teams outside the area were actually dispatched, flooded roads prevented them from getting into many areas. Thirty-seven counties flooded, yet there was no overall state management of the overall crisis, at least in the initial stages. Each county struggled along as best it could, attempting to deal with incidents on a case by case basis. Had it not been for massive intervention by military rescue helicopters, it is likely that the death toll would have been much higher than the fifty-two who actually lost their lives.

Another common failure of the rescue agencies in both Kansas City and North Carolina was the lack of flood rescue training and equipment. National Guardsmen, transportation workers, and others were deployed to the flood zone without any training or PPE. According to a CDC report, ten percent of the fatalities in the Hurricane Floyd flooding were rescue workers.

It is time for us, the rescue community, to recognize that existing response models, both for floods and large-scale disasters, are simply not adequate. No longer should we pretend that they are meeting the needs either of citizens who look to us for rescue, or for keeping our own safe. Our task for the future is to plan, train, and make rescues in the safest and most expedient way possible.

© Slim Ray 2001 All Rights Reserved

© Slim Ray All rights reserved

Back to CFS Press Home Page Order Information Higgins and Langley Memorial Awards About Slim Ray, Author
   Swiftwater RescueSwiftwater Rescue Field GuideAnimal Rescue in Flood and Swiftwater Incidents

CFS Press 8 Pelham Road, Asheville, NC 28803
Phone 828-505-2917
Fax 828-505-1030