Edited Version of January 10, 2001 Transcript
EIIP Virtual Library Presentation
" Complex Systems in Crisis Management"
Louise K. Comfort, Ph.D.
Professor of Public and International Affairs
University of Pittsburgh
EIIP Technical Projects Coordinator
The original unedited transcript of the January 10, 2001 online Virtual Library presentation is available in the EIIP Virtual Library Archives (http://www.emforum.org/vlibrary/livechat.htm). The following version of the transcript has been edited for easier reading and comprehension. Typos were corrected, date/time/names attributed by the software to each input were deleted but the content of questions and responses are as stated by each participant. Answers to participants questions are grouped beneath the appropriate question to facilitate meaning.
Amy Sebring: Welcome to the EIIP Virtual Library! Our topic today is "Complex Systems in Crisis Management." We are pleased to introduce our special guest, Dr. Louise Comfort. Louise is a Professor of Public and International Affairs at the University of Pittsburgh and author of a new book, "Shared Risk: Complex Systems in Seismic Response." Louise has also participated with the EIIP several times in the past and we are very happy to welcome her back. Thank you for taking the time to join us today, Louise.
Louise Comfort: Thank you, Amy. I am happy to have this opportunity to share some of the concepts from my book.
Crisis management represents one of the most difficult, challenging, urgent tasks that public managers face. Crisis environments are complex, dynamic and place lives, property, and continuity of operations in our communities at risk. Concepts from the emerging science of complex adaptive systems (CAS) help us to understand these environments better and offer fresh insight into the challenging problems of emergency management.
Today, I would like to share with you a brief introduction to some key ideas from this field of complex adaptive systems that are relevant to emergency management. I am responding to a set of questions posed by Amy Sebring, who also shares an active interest in this field.
What is complexity, chaos, self-organizing systems?
The concept of "complex, adaptive systems" (CAS) captures the dynamic processes of change in complex environments. A complex, adaptive system is a set of interdependent units which is capable of reallocating its resources and actions to achieve a stated goal under changing conditions. At the same time as the units within the system change their relationships to one another, the system as a whole changes its relationship to the environment in which it operates.
This definition describes the emergence of a response system in a community following disaster, when the community activates its emergency response plan and first response agencies shift their actions from daily operations to respond to the needs of a community struck by disaster. The challenge is for the emergency response system to shift its performance in coherent ways, so that the separate units --- Police, Fire, Emergency Medical Services, Public Works and others --- move smoothly into coordinated action, given their separate emergency responsibilities.
What ideas from this emerging science have relevance for emergency management?
The concept of CAS notes the iterative effect of small differences in initial conditions upon dynamic processes repeated over time, the startling effect of random events upon the operations of a system and the critical role of interactions among the components of the system in adapting to new demands. It provides a method for monitoring the rate of change and assessing the degree of complexity introduced into a social system by random events.
These ideas are especially relevant to emergency management that seeks not only to incorporate risk reduction into management of daily operations at the local level, but also to facilitate response at the regional or state or national levels, should an adverse event occur. Disaster requires at least two levels of operation and action within the community, often more, and generates a multi-level set of demands that form a complex system of continuing anticipation and response to the problem.
Emergency management addresses the problem of risk that is shared by all members and organizations of a community, whether or not they have contributed to the condition of risk. Shared risk is nonlinear in that small differences in initial conditions, repeated in actions over time lead to unpredictable outcomes. It is also dynamic, in that a change in the performance of one actor in the affected system may directly influence that of other actors in its immediate vicinity, creating a ripple effect of learning or failure throughout the system.
Problems of shared risk are not easily amenable to control strategies, especially if control is externally imposed. Resolution requires processes of communication, coordination and collective learning among multiple groups at different levels of understanding, commitment, and skill, leading to the reciprocal exercise of different types of knowledge, authority, and action among the participating groups.
What is this idea of initial conditions?
Initial conditions represent the existing state of the community at risk prior to a specific hazardous event. It includes the basic resources available for learning and action, as well as the current operating context of the community. These conditions shape the possible courses of development for coordinated response to an actual event.
For example, findings from my study of rapidly evolving response systems following earthquakes show that the information functions of search, acquisition, representation, processing and exchange are essential to interorganizational decision making, and, under certain conditions, lead to self organization and adaptive behavior. But the capacity to carry out these functions depends upon a technical information structure being in place before the disaster occurs, supported by an organizational infrastructure of trained personnel and a current knowledge base.
This is a 'socio-technical system' that includes organizations, machines, computers, and people who use them in the ever-changing context of a modern community. Socio-technical systems, carefully designed, can address multilevel problems of communication and coordination to support organizational learning and action in complex, dynamic systems. Without such infrastructure, communication fails; without communication, coordination fails. Without coordination, the situation slips into chaos.
How does the self-organizing occur in relationship to the timeline of a disaster event?
The concept of `self-organization' indicates the ability of the human members of a sociotechnical system to learn, based on timely, accurate information and feedback on performance. Systematic monitoring of both technical and organizational performance over time and feedback on actions taken in response to hazardous events enhance the capacity of the community to learn improved means of mitigation and response to risk.
This capacity to reallocate resources and action in order to meet new or unexpected demands from the environment enables a sociotechnical system to function more productively and effectively under dynamic conditions.
The interaction between humans and computers is critical to this capacity for self-organization in complex, dynamic environments. Without the carefully designed use of computers, the information processing requirements resulting from a major earthquake, for example, repeatedly overwhelm the capacity of human decision makers, resulting in incomplete, untimely, or inadequate response.
What are the strengths and weaknesses of hierarchical structures? What are the strengths and weaknesses of flatter structures?
The basic assumption underlying hierarchical structures is that if the problem is well defined, if the rules of operation are clearly specified, and if systems can be closed to outside interference and disturbance, they can function without error. Once the system is carefully designed and functioning, considerable effort is placed on the control of error, or any aberrant disruption or behavior. The system is assumed to be in command; all other disturbances need to be controlled or eliminated
The strength of such systems is that they function well in stable environments. Their weakness is that they do not function well in dynamic, changing environments. The rules defined for stable conditions, carefully specified, most often do not fit the dynamically changing conditions of a disaster environment. Worse, personnel in hierarchical structures, trained to follow the rules, have difficulty devising strategies of action that will work when the rules no longer apply. Such organizations often are unable to respond effectively to urgent, time-critical demands. The delayed response of emergency organizations following the Hanshin Earthquake in Kobe, Japan is a classic instance of this weakness.
In flatter structures, the underlying assumption is that no single individual understands the whole problem, but that each member of the organization likely has insight and a responsibility to act on the best knowledge available. Responsibility is shared, and decisions are based upon the most current, complete information available.
The strength of such organizations is that they are mobile, and information flows easily through them to critical personnel. With multiple persons reviewing and checking the information, errors or inaccuracies are discovered more readily. Flatter structures are able to reallocate their resources and personnel more quickly and efficiently, and move more effectively toward self organization.
The weakness of such organizations is that they depend upon fully functioning information infrastructures with well-trained personnel who are capable of acting on their own initiative in ways that are consistent with the system's goals. If this information infrastructure is not fully developed, or if there is a weak link in the personnel chain, organizational performance is hampered. Such breaks need not be fatal, as strong components of the system can move to carry the load for weaker units, but such back-up capacity needs to be designed into the system.
What is the role of leadership?
The primary role of leaders in complex adaptive systems is to articulate the goal of the system in ways that personnel at all levels can understand. In emergency management, this means protecting lives first, property second, and maintaining the continuity of operations for the community. It means enabling each member of the response system to understand how her actions may contribute to, or detract from, the effectiveness of the entire operation, and building a consensus on goals that allows personnel with different responsibilities operating at different locations to work together.
Leaders also have the responsibility to "set the example" of being open to new information, willing to correct their mistakes, ready to shift duties as the situation demands. Such flexibility is critical to the organization's ability to adapt to the requirements of a dynamic situation.
What are the implications for skills an emergency manager should have?
Emergency managers need, importantly, skills of inquiry. They need to be able to sense when a problem is about to occur, to ask the direct questions that will expose a weakness in structural or organizational performance, and to integrate information from different sources to build a comprehensive profile of the status of their community. Most importantly, they will need to develop technical skills in managing information technology in order to use this technology fully to inform their organizations, policy makers and community residents.
What elements of current, recommended EM practice are in line with the implications of ideas about complexity; which ones are not?
Emergency management practice has changed significantly within the last ten years, largely due to experience gained in the extraordinary increase in the number and size of disasters within the US and abroad.
Traditional emergency management practice was based on command-and-control concepts from military training. In practice, however, civilian emergency response agencies have discovered that these practices are not effective in the rapidly evolving, dynamic environments of disaster. The US Forest Service, for example, initiated the Incident Command System largely in response to the dynamic requirements of the wildlands fires that sweep through Southern California nearly every fall.
While some structure is essential in order to hold and exchange information, flexibility to adapt to a changing situation is vital in emergency management. While the language of Emergency Management still retains some of the terms of command-and-control organization experienced emergency managers understand the need for flexibility in adapting to the demands of the situation. They also see the importance of building a shared knowledge base for disaster operations, maintaining an environment of continuous learning for their personnel, and enabling personnel to take informed action based on the most current information.
Why is communication so important?
Communication is critical to the process of self-organization and collective learning. Communication of information regarding the current status of the community, as well as the actions in process by participating organizations, allows participants in the response system to make informed decisions about their own actions, and to adjust their actions in accordance with those taken by others to achieve the overall goal of protecting the community and restoring its functions. An informed community is able to adapt its behavior more appropriately to its exposure to risk.
Beyond response, what are the implications for mitigation and recovery?
Once seemingly intractable, problems of shared risk can now be reconsidered in light of advances in information technology that facilitates a community's capacity to adapt to known risks. The technical capacity to order, store, retrieve, and disseminate information to multiple users simultaneously, to represent knowledge visually, and to monitor different types of functions at different levels of performance has created potential new approaches that involve collective learning and self-organization. The same is true for recovery, as with more complete information from multiple sources, communities will have the capacity to rebuild without repeating the mistakes of the past.
I will close with a brief summary of response to the Northridge Earthquake, which shows an emergency response system that adapted successfully, but that moved to the very edge of chaos. This first slide shows the interjurisdictional response system involved.
This next slide shows the actions taken.
The following chaos maps show the performance of the system in three emergency functions for public organizations. This next map shows that the response system performed well in emergency response functions, with the model explaining 65% of the variance and showing strong stability, with a value of k at 1.00.
The next map of the function of communication/coordination, combined for this analysis, is stable, with a value of k at 1.00, but explaining 33% of the variance, which is respectable, but not outstanding.
The last map of the function of financial assistance slips into chaos, with a k value of 3.92. The equation explains only 12% of the variance in this function.
These findings show that the heavy investment in training and equipment for public emergency response organizations in the Los Angeles Region resulted in a strong capacity for emergency response. The weakness in the system appeared in insufficient communication and coordination with the nonprofit and private organizations that make up the total community response system. The functions of financial assistance and response and recovery slipped into chaos. The function of disaster assistance was very near chaos, with a k value of 3.31.
In conclusion, the concepts of complex adaptive systems are useful to emergency managers, as they illustrate the dynamics involved in the emergency management process and offer a means of measurement and monitoring that can assist us in keeping response systems flexible and adaptive in response to sudden, urgent demands from the environment.
That concludes my prepared presentation. I will be happy to answer questions now.
[Audience Questions & Answers]
Amy Sebring: Thank you for that introduction Louise. We now invite your questions or comments.
Claire Rubin: In your book do you use both US and International examples?
Louise Comfort: Yes, I have 3 US case studies and 8 International studies.
Claire Rubin: Does the system work in both settings?
Louise Comfort: Claire, yes, it is a characteristic of complex adaptive systems that the systems evolve differently in each setting, but each earthquake produced a response system.
Amy Sebring: Louise, I understand that even the military is looking at CAS for some ideas. Do you know this to be true?
Louise Comfort: Amy, yes, I have been invited to give presentations to military research groups, the concepts are relevant to military planning as well.
Ray Pena: Dr. Comfort, are you basically saying, like Dr. Quarantelli, that the ICS isn't a good model for a community to use for disaster response?
Louise Comfort: Ray, no, I think that ICS is an 'early version" of a complex adaptive system. It still uses the military language but the concept of adapting to the needs of the environment is consistent with CAS.
Frank Fiedrich: Louise, how important is training for CAS and how can such training look like?
Louise Comfort: Training is very important for CAS, but we need a different kind of training. We need training to search for differences, similarities, to invent more appropriate strategies, to be able to follow the "goal" of the system, rather than precise rules.
David Crews: Has CAS been applied to other natural disaster scenarios besides seismic, e.g. hurricanes, floods, etc.?
Louise Comfort: Paul Bourget is interested in applying it to floods and will be doing research on it. Omar Cardona in Colombia is interested in applying CAS to the range of disasters, including civil war that Colombia faces.
David King: As I understand it, you are saying emergency managers need to be trained in flexibility. What and how do we find training in that specialty? Any examples you can suggest?
Louise Comfort: We really need to rethink and recreate training in flexibility. I am doing a trial demonstration with practicing managers in the Pittsburgh Metro Region this year. And they are very eager to incorporate these concepts into their training.
Amy Sebring: Regarding initial conditions, does your research suggest that communities that have strong working relationships on a daily basis will function better in disaster situations?
Louise Comfort: Amy, yes, this is very important. Trust is crucial in uncertain situations. And building that trust among working organizations can best be done outside emergency situations.
Cam King: Is your book available and, if so, were can it be purchased?
Louise Comfort: Shared Risk is available through <http://www.Elsevier.com>, or more easily through <http://www.Amazon.com>. Regrettably, it is very expensive. I've tried to argue with Elsevier about reducing the price.
Amy Sebring: I would like to mention that there is a link on our background page where you can get more information about Louise's book.
Roger Fritzel: Louise, can you tell us more about the "six different types of information technology" that were used?
Louise Comfort: The types included EBS, Emergency Broadcasting System; OASIS, California's satellite system, a direct satellite link between Pasadena's DFO and FEMA, the CUBE-REDI system, California's direct seismic monitoring system and the direct newsline with radio and TV stations.
Amy Sebring: Regarding recommended EM practice, would you not say that the concept of an EOC, where different organizations come together to communicate face to face, is in line with CAS?
Louise Comfort: Yes, it is, Amy, but, in fact, the concept of an EOC is changing with practice given the increased use of information technologies in the separate agencies, the chiefs prefer to stay in their agencies. Consequently, they send their second or third in command to the EOCs. This means decisions still need interactive communication among departments.
Avagene Moore: Louise, has the need for flexibility in training been shared with those who are conducting EM training and developing courses? If so, what has been the reaction? Can we expect some adaptations or revisions to today's training? i.e., FEMA's EMI.
Louise Comfort: I hope so. I spoke with Wayne Blanchard just yesterday. I think he is willing to listen.
Frank Fiedrich: If risk is shared, responsibility should be shared as well. How can this practically be implemented in disaster plans if you do not know how the system will behave in a certain disaster situation?
Louise Comfort: Planning is truly a process, a way of learning how to think. The critical task is to know what the damage is, what resources are available, and what constraints a manager faces. It is not the "system" that needs to work as much as the individuals in the system who communicate with one another quickly and accurately in order to coordinate their actions. Certain kinds of planning are important and they do need to be tested.
Amy Sebring: Louise, do you have any plans yourself for further research or application of your research in this area?
Louise Comfort: We are planning a trial demonstration with practicing agencies this year, it will be both training experience and research. This will give us a better understanding of how quickly practicing managers can absorb new information and adapt to it.
David Crews: Frank brings up an interesting point. Social systems have a lot to do with response. Having worked at NORAD a number of years ago, systems were in place that could meet many of the CAS systems that are needed in civilian operations.
Amy Sebring: Thank you very much for being with us today, Louise. Please stand by a moment while we take care of some announcements.
Louise Comfort: Thank you all for your interest and attention.
Amy Sebring: For any first-timers, we will have a text transcript posted later this afternoon, which you will be able to access from the Transcripts link on our homepage. Then on Friday we will have a reformatted transcript in both html and in Word for download. If you are not currently subscribed to get our program announcements, and would like to, please go to <http://mail.wces.net:81/guest/RemoteListSummary/EIIP>.
Next week we will be trying something a little different. We are going to do something like a group discussion. However, you will not get discussion questions in advance. This will be more of a pop quiz format and is called "Can You Pass the Road Test to Drive on the Information Highway?" It should be fun, but will also be a way to get your input as to what skills are most needed, so please join us then.
Thanks to all our participants today. It is good to be back. We will adjourn the session for now. Please help us thank Louise for her presentation.