Tuesday, March 4, 2008

Sel-Organization -Aftershock Article

Self Organization During Disasters


Self organization depends upon ready access to timely, accurate information through an information infrastructure that supports systematic monitoring of critical conditions, feedback to responsible participants, and revision of actions taken in the light of new information. When this is in place, no other emergency management is as successful.” -- Comfort (1995) based on his research of the Great Hanshin Earthquake Japan 1995


Since at least World War II researchers have known that emergencies of all types could be best mitigated by “civilian resources” when available. Self Organization is the key to this remarkable success in civilian response to a wide range of emergencies and have proved effective in many socio-geographic environments around the world. Despite this decades old knowledge, and proven success record, governments, NGOs and corporate elites have continued to foist bureaucratic, centralized and dangerous management schemes on communities around the globe. Aftreshock Action has studied and looked at what makes self-organization such an effective tool in relief and recovery work and what elements support or hamper successful community responses to crises. This article looks at some of the most recent research and points us in ways in which we can optimize our community's ability to self-organize during a disaster.

Though governments, NGOs and corporations are constantly revising and re-branding their emergency efforts, the model they employ remains surprisingly stable. Social science researchers have dubbed their management efforts “command and control systems” as opposed to “community response systems”. These models are often employed together when dealing with an emergency/disaster. Due to the inherent flexibility and quick reactive time community response nearly always occur prior to command/control systems. Sometimes they operate together in geography but almost always remain separate due to the inherent contradictions embedded in each approach. Command/control approaches believe by their own internal logic that they need to eventually dismantle the community response to be effective. So in the end one model usually trumps the others during different periods of a crisis and its aftermath. Both models need to provide some basic functions: evaluation/information; resource management; react/adapt to unpredictable variables; develop and implement decisions; obtain consent of victims, stake-holders and the “public”. Each model based on its organizational principles approaches these functions from very different perspectives.

Control/command systems are characterized by closed information feedback systems. Information is controlled and managed by gate-keepers (e.g.”experts”). This system can often become very fragile and overwhelmed during a disaster by the duality of too much data and ever increasing number of variables. These two problems are exacerbated by the attempt by command/control systems to work from a gestalt (wide angle approach) and the need to manage and feed information to those in a position to analyze it. Obviously these groups know that disasters often compromise information management by damaging infrastructure and rapid changes. To get around these problems they put a lot of stock in predictive models in an attempt to create plans prior to the event. They also seek technology that can provide secure and uninterrupted transmission of data and other information to gatekeepers. Because the loop is closed and in many ways predetermined, there is a heavy emphasis on “scenarios” and “action models”. The closed loop has failed so spectacularly in the past. Control/command models believe that the problem has been a lack of data and information so there has been a greater and greater emphasis on data collection. Soon there was a problem, the amount data was so large it was impossible even for a team of highly educated and specialized experts to not get overwhelmed. To get over this hurdle, they turned to using massive supercomputers for not only crunching the numbers using statistics but to create computer simulations and action modules. In a sense removing any human oversight in the technologically dependent model. The truth is that this model has done poorly. Researchers who have looked at the computer simulations for disasters have found that nearly 80% of predicted outcomes of disasters have been so far off base as to be almost random (2002 Richardson). In fact, over the past 10 years there has not been one successful simulation system that has been able to predict the accurately the damaging effects of any of the past Simley's 1,000 most destructive natural disasters (a common list used by disaster researchers). What is so shocking is that nearly 35% of those disasters on the list are naturally and fairly frequent reoccurring phenomena. Comfort did research showing nearly a decade of computer simulations (from a variety of Government and academic sources) to disaster experts. They were given 100 action plans and damage results devised for 200 actual disasters. The experts then had to assign the simulation to the actual disaster, and they could not do it. It was nearly random and the experts had no agreement among themselves.

Community response systems do not normally employed super specialists or chugging super-computers to access and use information from disasters. They employ an open feedback loop that is in real-time. This allows groups to process a smaller amount and scene specific information that they as humans can actually understand and use to formulate action responses. The results of the responses can be evaluated and modified in real time, allowing for a reinforced open feedback loop. These systems [inquiry] allow communities to create innovative and effective solutions to problems of resource allocation that are not available to command and control systems. They also increase the amount of “investment” by communities to self-regulate successful strategies during time of crisis by allowing them access to the information the information they need to formulate and execute plans(Tierney and Trainor 2004). Each community can seek out the information that is most needed at the moment, because the occupy the area of action and must confront the situations there. This limits the amount of variables not just the data and allows for effective information control without the need of sophisticated predictive technology or a cadre of highly trained specialists needed to “understand” the entirety of the situation. Because the open feedback loop is natural occurring, there will be information gathering and analysis happening where ever there is a crisis and a community. So the entirety of a crisis can be analyzed but not by anyone group or model so in a sense a real gestalt is created organically.

These differing approaches towards information management have a direct impact of resource allocation.

Command/control systems have a notorious record of resource mismanagement during an emergency that can actually increase the negative effects of a crisis. Resources in this system are employed based on a statistical models derived from their information systems. This creates an almost universal delay effect in moving resources to an afflicted area since information needs to be vetted and analyzed in a closed loop system. The delay can be quite substantial and when combined with the reality that the necessity of resources in the first few hours are exponentially more useful. The delay in resources are bad enough but the actual deployment of resources (even when they are manipulated by disaster-capitalist interests) has too often been a dismal failure. For example, during 9/11 it was decided to send 40% of the FDNY paramedic units and 75% of the trauma doctors to/near WTC. While Red Cross moved nearly 5 tons of plasma from regional areas as far away as Minnesota to NYC. Of course all of these resources sat unused for up to 30 hours. More people died from lack of emergency services (from non-9/11 related medical emergencies) in a single day than anytime in NYC's history. All that blood that was sent, was rushed here and the paperwork got messed up so NYC had nearly 1000% more blood than normal and with no way to send it back to where it was needed. Most of it had to be thrown away because their was no place to store it, while earnest volunteers still lined Times Square to give even more unneeded blood. FEMA sent nearly a dozen trucks of medical supplies (some arriving on 9/14) but no respirators for the responders or residents in and around the World Trade Center. The list goes on and on. Even during “slow” disasters like famines in Africa we see the same mismanagement of resources that could be used elsewhere. Angry citizens tend to blame incompetence of emergency managers but the history is too long and constant to suggest simply idiots are at the helm of emergency management agencies and organizations. It seems to be more systematic, a result of a faulty model.

Fortunately community response model has a better history of resource allocation. It is shocking the amount of aid that is provided by these unfunded groups. 2002 Walther King looked at local church groups' allocation of resources during Hurricane Lili that hit Louisiana and parts of Texas and did an estimated 860 million dollars damage and left 15 people dead. Local Churches had a budget of .005 of the Office of Emergency Management and fewer staff. They provided nearly three times the injury mitigation and protected more homes and businesses from aftermath destruction than the OEM. They used their limited resources (including human labor) very effectively and wasted almost nothing, where a much better funded agency with hundreds of trained specialists were less effective. Over and over again from earthquakes in Japan, to the Tsunami in South East Asia to tornadoes in Nebraska we see the same pattern. Local groups obviously also have the advantage in response time to disasters of all sorts. There has been extensive research on this topic and the results are conclusive, even the OEM and FEMA admit it in their publications, that community groups respond significantly faster to disaster events (often days before). Even resource poor neighborhoods “can maximize their limited resources if they are employed early, providing a initial intervention that has exponential results in disaster mitigation” (National Research Council- Committee on Disaster Research ).

A key function in emergency response is the ability to adapt/react to changing situations on the ground with imperfect information. Again we find this ability to be severely impaired with the command/control system of crisis management. This impairment is a direct result of the centralization problem inherent in government agencies and NGO organizations. This is similar to the closed information loop, in that, these control/command systems seek to predict and pre-plan the response of its responders. It seeks to substitute on-the-scene decision-making with simulations and regulations. This is undoubtedly also partially a result of the overspecialization of emergency professionals. A simple google search of the types of emergency professionals brings up scores of types, most involving at least 3 years of specialized schooling. Emergency professionals also continuously drill, up to 15 hours a week yet they are incredibly limited in making real decisions in the field. Since they are dependent on the command part of the model they must wait (like resources) for the experts to allocate and plan their actions. During Katrina, dozen of helicopter pilots waited over three days for orders before they could do limited rescue runs despite having the fuel, helicopters, equipment and skills to do it. This is not an aberration but the normal course of events. The reaction time and the adaptation needed to respond effectively to changing situations during a disaster are removed from specialized emergency workers.

The community model draws upon the diverse skills of its members, that almost always lack emergency training. This lack of basic skills (e.g. First Aid) has had a negative impact on the community model but that has not always been the case. During the heyday of Civil Defense (in the 1950's) most Americans had had a basic first aid course. Public schools taught modules on fire suppression, first aid and other useful emergency skills. We know that when emergency skills are disseminated they can be very effective. CPR and the Heimlich maneuver have saved countless lives by employing a community response model. Researchers suggest that nearly 3 times as many victims have been saved by ordinary folks using CPR than all the professionals (doctors, paramedics, life-guards, police, firemen, etc) combined. The community-response model relies on individuals to make decisions and implement action plans based on the vents surrounding them. They have no need for constant drilling or waiting for commands to provide much needed services on the ground. Nearly 1/3 of civilian helicopter pilots (tourist pilots, news helicopter pilots, etc.)during Katrina jumped into action hours after the disaster while better prepared and trained emergency pilots sat waiting for orders. There is more that can be done, we need to provide more basic emergency skills to various parts of our communities to ensure more opportunities for effective action. Despite this obvious weakness there can be no doubt that during an emergency you are more likely to be saved and your home protected by your neighbors than highly trained professionals with high tech gadgets.

All emergency aid efforts are ultimately dependent on complicity and involvement of a variety of groups:victims, stake-holders and the “public”. Even emergency managers understand that the public is an important factor in determining the overall success of a disaster operation. After the dismal failure of FEMA to act during Katrina they sub-contracted with a New York PR firm to “fix” their image and employed their services during the recent California wildfires to some positive effect. The emergency management agency has added “public trust” to its logo, highlighting the role of public opinion on relief efforts. Needless to say trust is not the first word to pop into most Americans' minds when they think of government emergency agencies (NGOs do much better but still need to use slick public relations firms). The above problems with the command/control model undoubtedly play a role in why emergency agencies can not gain the support (or complicity as the social science researchers say)of various segments of the public but the very model of control/command makes this task very difficult.

Command/control models rely heavily on both specialization and centralization that are often at odds with public support. Specialization is by necessity removes itself from the common experience of most people. The knowledge of the specialists is often obscure to everyday educated people, and thus the motivations for their actions often are not understandable to those outside. Centralization also creates decision-makers that are removed both socio-economically and geographically from most of the population and the victims of disasters. A study 2006 by McWilliams, a social psychologist at Harvard,found that among nearly every class (gender, socio-economic, regional, etc.) people tended to trust those they could relate to even if they were more poorly equipped to solve their problems. When people are not invested or do not trust decision makers they tend to unconsciously thwart action plans, a point well understood by emergency managers. Instead of trying to get more people to have trust in their authority (which may not even be possible) they use force and fear to create a passive population. This has negative consequences in that it interferes with their information gathering and fails to utilize the resources, labor and skills of the populations. They give these up in turn for creating passive complicity.

True investment, as demonstrated by the community response model, requires transparent and local decision making processes. It also requires an open information exchange system (not just the choreographed and control management of information so common among bureaucratic agencies and organizations). The fact that the community response approach draws from a shared well of experiences, resources and understandings make their actions and motivations instantly comprehensible at least to the local victims, stakeholders and the regional population. The community response approach often fails to generate large public support due to its regional focus and sphere of influence, the very factors that make it successful on the local level can work against it at larger levels. This problem is somewhat mitigated that most of the community response models resources, actors and interests are local and thus the need to have complicity by an outside generalized public is not as detrimental as it might be for the control/command approach. Activity is probably the easiest and surest way to ensure investment in a given set of actions during an emergency. People that are involved tend to feel more favorable about emergency rescue, relief and recovery efforts than those less involved. The community response model allows a maximum of participation at all stages of disaster relief, thus increasing the likelihood of investment of various populations. Since the communities tend to make decisions along more decentralized and horizontal means, this also promotes a joint sense of ownership of the actions even of groups that one is not directly involved in. For example volunteer fire-fighters and paramedics tend to rate the services of each other much higher than professional co-workers do and in fact there seems to be great distrust among professional departments where competition for budget and resources create toxic work environments (Young, Stanfield, et al 2001).

These two models are so opposite that there is very little reason to try combine these forms. A socio-technological, used by large NGOs and government agencies, approach requires a shift in the conception of response systems as reactive, command and control to ones of inquiry based systems that promote self-organization and self-evaluation. These systems create open feedback systems that do not easily exist in traditional command and control systems. These systems of inquiry allow communities to create innovative and effective solutions to problems of resource allocation that are not available to command and control systems. They also increase the amount of “investment” by communities to self-regulate successful strategies during time of crisis in other words any response to disaster must promote not inhibit emergent structures and activities during disasters.” (Tierney and Trainor 2004). Yet governments and NGOs suggest that disasters be left to professionals with the knowledge, skills and technology to ensure that there is not destructive panic. The control/command model rests on the myth that the public is ill-prepared emotionally and in resources to provide for their own rescue, relief and recovery. They use the boogey-man of panic to suggest that we can not count on our friends, neighbors or co-workers to create viable community responses to emergencies.

Social science research finds little justification that people will respond qualitatively differently to man-made versus natural disasters. The variables that effect response are: Speed of onset; geographical spread; access to information and previous experience or training for disasters. Furthermore all social research from the past 20 years suggest that people do not panic. The population as a whole tends to make very rational choices based on the information and experience they have available. Researchers have looked at variety of cases from the Spanish Influenza, to Chernobyl to 9/11 and find that people respond similarly during a crisis as people say they would react if there was a crisis. So we can predict how people will react to an unpredictable event and they tend to act reasonably. This would suggest that regular citizens can and should be trusted to participate at all levels in their own rescue and recovery operations. Panics do occur and have occurred, but they almost uniformly are located in liminal spaces highlighted by supposedly total environment. There have been recent panics a on cruise ships, airports, stadiums, hospitals and even prisons. One of the reasons for this, is that when authority breaks down AND there is a lack of organic communities people make calculated decisions based on their own awareness and needs. This combined with no reliable sense of information creates panics that can turn deadly.

In the end, it is clear that our survival during an emergency depends on the strength of our local communities. Despite the overwhelming research on this point, NGOs and governments have only allocated token support for local community organizing efforts. This is a great shame because the amount of resources needed to adequately prepare communities is relatively small when compared to the bloated budgets of NGOs and government agencies. Instead of strengthening communities' responses to disasters the government has actually created policies to retard such naturally occurring self-organizing. It would be foolish to look to reform government or NGOs which have committed themselves for decades to a dysfunctional control/command (and profited well on it). We must find ways to support and expand local organizing efforts. The uncanny way a collective community action occurs prior to, during and after a disaster demonstrates the power of organic, and indigenous organizing. These type of communities are not only effective in helping out their neighbors but also revitalizing communities and re-establishing community norms and structures (Drabek 1986). This is the goal of Aftershock Action.