Creating a model for emergency preparedness

Nov 1, 2000 12:00 PM, JAMES STEVENS

Recent natural disasters and acts of terrorism have focused the nation's attention on emergency preparedness. Severe weather, riots and bombings can challenge the ability of public officials to maintain order and services while preserving lives and property.

In May 1998, the White House issued directives ordering the strengthening of the nation's defense against emerging terrorist threats to the United States. As a result, a number of research projects were launched. The resulting data generated are now being used to simulate and evaluate the nation's emergency response capacity. Important insights can be gained.

The evaluation involves the use of computer models. The Urban Security Project, whose research is partly conducted by Los Alamos National Laboratory, is developing modeling tools that can assess the response of an urban system to changes in the physical environment, sociopolitical setting, and the economy.

The model seeks: - to simulate the interdependent relations of various urban subsystems. (For example, to predict the impact on emergency response routes made by a flood as compared to an earthquake.)

- to accurately predict the compounding effects of multiple urban subsystems (For example, how power system failures impact traffic patterns).

Emergency response teams can make successful predictions based on experience, but only computer models use all available data to provide the most accurate predictions.

Developing a model of urban systems for use in improving urban emergency preparedness involves several phases:

- define the problem;

- implement a modeling method;

- verify the model results;

- analyze the results; and

- implement appropriate changes.

The phases lack clear boundaries, and occasionally the progress in one area results in reevaluation of another area. The implementation of appropriate changes at the emergency responders level, for example, may cause the data on reaction effectiveness to change. Modeling a dynamic system, and especially a system of interdependent subsystems, is an ongoing process as long as model results are being acted upon.

Modeling method Defining the problem to be modeled covers the broad spectrum of urban emergency factors. The model needs to simulate emergency conditions based on inputs and required subsystem reactions. Determining the input situations to be modeled begins with a risk assessment of potential emergencies such as terrorism, civil disturbances, natural and industrial disasters. The subsystem reactions include the impact on resources and infrastructure, and the ability of emergency response personnel to mitigate the negative effects of the emergency. Defining the problem requires a great deal of data collection and drives overall development of the model.

Data collection occurs at many levels. Valuable data comes from historical records, practical experimentation, and the results derived from models of subsystems. Historical records provide information on incident frequency and response effectiveness. The records on earthquakes illustrate the impact on infrastructure and the corresponding response strategy results. Practical experimentation includes testing individual responders and municipal systems. Tools used to train responders also produce data on capacity. For example, JUSTNET, a project sponsored by the National Institute of Justice, produces training simulators for responders that include "Weapons Team Engagement" and "Bomb Threat Training Simulation".

Tests of municipal systems commonly simulate a disaster to evaluate response performance. The Federal Emergency Management Agency (FEMA) simulated a chemical weapons attack at Fort Gordon, Ga. A test of a simulated terrorist attack at the Pentagon involved more than 500 participants to both train personnel and evaluate performance. Models of subsystems that provide data include weather models at the National Weather Service and disaster models like the Consequence Assessment Tool Set (CATS). CATS is an integrated disaster analysis system that estimates damage from technological and natural hazards. Tests and models of subsystems provide building blocks for developing models of larger, more complex systems. The variety of data sources minimizes statistical errors in model inputs. Larger data sets allow more accurate models to be developed.

Models, by design, imitate the behavior of a system. The incremental building of a model includes establishing logical relations, a central clock, distributions of randomness, and input/output nodes. Logical relations must reflect the cause-and-effect patterns of the system being modeled. The central clock or "executive" controls the simulated passage of time and keeps modeled events occurring at appropriately spaced time intervals. Distribution of randomness provides for the input of statistical variation in the data being modeled so that the model can potentially produce the full range of likely outcomes. The input/output nodes provide a means of interacting with the model to simulate various scenarios and retrieve the desired data for analysis. The model will only simulate reality to the extent that these factors of modeling are appropriately applied. The care with which a model is built determines its accuracy.

Models of individual urban subsystems combine to form a model of the total urban system. The Urban Security Initiative, spearheaded by researchers at Los Alamos National Labs, developed a computing architecture that allows different simulation programs to interact using JAVA and COBRA programming languages. The software allows communication among computer models written in different languages and simultaneously running on different machines. Multiple users can log on at the same time to use the simulation package. The model actually incorporates many models of subsystems that are interactive. The master model synchronizes the subsystem models for time and activity using conserved material states such as energy and mass. The rainfall predicted by a weather model, for example, becomes the water flow used in a flood model. This linking of models creates a master urban model. The master model provides a means of simulating a significantly more complex system.

Verification of model results is necessary to validate the model as a tool for determining emergency preparedness. Demonstrating that the model predicts occurrences within statistical confidence intervals for a range of situations proves its ability to predict nearly all situations within that range. A validation of the model might use a high-energy release of "nerve gas simulants" in a subway car to determine if the model accurately predicts the dispersion of a toxic cloud. This test would help verify the interaction of models of explosions, transportation infrastructure, and fluid dynamics within the city for a given day's weather pattern. It is a complex test that demonstrates both the need for validation and the potential value of an effective model. Validating the model provides information about the accuracy of the model, its limitations, and constraints.

Presentation of results An accurate model provides opportunity for experimentation within a virtual urban environment. The ability to quickly and cheaply test scenarios allows civic leaders to evaluate different emergency response plans. The models can provide information on risk exposure for hypothetical changes in response training, vulnerability reduction, facility hardening, and warning methods. Additionally, the simulation results help to improve auditing techniques used in evaluating real-world conditions.

Models make predictions based on input information. Application of developing data trends in simulations provides insight to longer term outcomes. Researchers at Sandia National Laboratories could model the relation between illness rates and the water pollution problem in Bangladesh. Determining a way to curb the arsenic poisoning of millions by even a percentage point would save thousands of lives. Presently the lab is preparing a field test of a model that receives data at hospitals about illnesses in order to predict epidemic outbreaks, infection spread patterns, and victim concentrations. The system is intended to mitigate the impact of biological weapons deployed by terrorists. It would have also been invaluable in the recent and possible reoccurring outbreak of West Nile fever in the New York City area. Modeling developing situations provides response teams the opportunity to be proactive.

The modeling results provide volumes of information on urban emergency preparedness, including information on hazard dispersion, impact to resources and infrastructure damage. The basic life-sustaining resources of food, clean air and water come under threat during emergencies of all kinds, and foreseeing the relation of the developing emergency allows for mitigation efforts. Emergencies also frequently affect energy distribution, transportation, and communications. Not only do these factors concern urban dwellers, but they can impact the effectiveness of emergency response personnel. Having a model of emergency conditions in advance of the emergency provides responders the opportunity to better prepare. Being able to better prepare and mitigate life threatening resource shortages optimizes the response. The models provide a means of designing response strategies engineered according to projected conditions. The models greatly impact emergency preparedness for the better.

Want to use this article? Click here for options!
© 2008 Penton Media Inc.

This month in Access Control

• Opening Up About Door Closers
• An Enterprise Approach
• The Framework For Open Systems
• On A Higher Plane
• More from April's issue