The current “nonsystem” for providing information for disaster management is not effectively utilizing a wealth of information that resides with various organizations. Existing technologies could deliver to disaster managers important new information products that could save lives, reduce damage to property, and lessen the environmental impacts of natural disasters. Continued improvements in technology should help make information more widely, quickly, and reliably available—and at less cost. The current situation is characterized by numerous shortcomings that inhibit optimal decision-making for disaster management. The inability to access information and the lack of standardization, coordination, and communication are all obstacles that a disaster information network (DIN) could overcome. It is recommended that the Global Disaster Information Network (GDIN) Transition Team move ahead in planning for a disaster information network, taking into account the following conclusions from the present study:

• The need for an improved information network and its potential benefits are clear. Chapter 3 establishes the need for an improved disaster information system. There can be no justification for continuing in the current mode of nonstandard disparate resources when available modern technologies would make their linkage into one system a relatively straightforward matter, with obvious potential payoffs in saving lives and reducing losses if the system is utilized effectively.
• The foundations for an information network are already in place. While a significant undertaking, establishing a DIN would build on a substantial foundation that already exists. The most costly element of building the basic databases is well under way, and the community of users already exists. A network could be established initially by coordinating existing information resources and developing standards and protocols to ensure their reliability and usability and effectively linking with the user

• community. The cost of establishing the information system would be trivial compared with the cost already spent in developing the resources.
  
  

  The existing federal data-gathering and information programs (see Table 2-2) reflect an enormous investment of funds, mostly public, and the dedicated and sustained efforts of many investigators. These databases were derived from a variety of endeavors, including instrumental monitoring, field surveys, data compilations, and laboratory studies. Many of the efforts are of a continuing nature, as data are updated and phenomena are continuously monitored. Altogether, there has been, and continues to be, a very substantial investment of resources in developing and maintaining the databases used for disaster management.

  Despite the importance of these databases, their utility is impaired by a host of problems deriving from incompatible formats, inconsistent geographic reference systems, conflicting standards, and other human-caused factors. Many of these problems could be resolved and the value and utility of the databases for disaster decision-making greatly enhanced through improved organizational and technological coordination with only an incremental increase in cost. It is clearly in the public interest to do this.

• Recent advances in technology provide the mechanism for establishing a network. The Internet and high-speed telecommunications provide the necessary technologies for establishing an information network. Through the Internet, a DIN could be assembled by tapping data and information resources wherever they happen to reside worldwide. Thus, problems associated with assembling resources into a central repository are avoided, and the various organizations that developed the resources can each maintain their identity and roles in the larger system. Among the issues that require attention are determining which organizations need to be included; what datasets should be provided; who is responsible for establishing and updating links, maintaining the accuracy and currency of data, and preserving the data for long-term access; and what technical infrastructure is required at each site to accommodate increased traffic on the different systems.
• Successful implementation of the DIN concept will require a commitment of resources from a broad spectrum of stakeholders. Although the costs of establishing a DIN would not be large compared with its likely benefits or the investments already made in developing data and information resources, maintaining a DIN would require additional expenditures and commitments from the organizations involved. Converting data

• into a standardized format and providing data to the DIN, although a much easier task today because of the Internet than in the past, still require the commitment of resources, especially human resources. Many of the data are created through research or monitoring programs that do not have disaster management as their principal purpose. These organizations should be offered federal incentives to participate in the system because without the raw data the DIN would have nothing to communicate.
  
  

  Successful implementation of a DIN will require a commitment of resources from a broad spectrum of stakeholders and sustained organizational and individual commitment of material, financial, and human resources by DIN users and providers. While resource capability varies and not all DIN participants may be able to commit material or financial resources, each has a professional obligation to actively commit human resources.

• The products of a DIN should be based from the outset on users' needs. The success or failure of an information system is determined by the level of effort and support given to it by its constituency. Thus, it is paramount that broad representation of elements making up the community of users and providers be intimately involved in all aspects of the startup phase of a DIN. This includes such tasks as further defining the initial concept and evolving program plans for the overall system and an operating strategy. The importance of early ownership of a DIN by a wide variety of users and providers cannot be overstated. In this regard, the GDIN Transition Team is encouraged to consider other programs such as the National Aeronautics and Space Administration's Earth Science Enterprise Federation, an activity designed to disseminate information about the earth at a variety of scales (NRC, 1998b).
  
  

  In order for the DIN to provide value-added services and products, it must be responsive to the immediate and future information needs of users. The critical prerequisite is systematic and continuous involvement of an information users' representative from the disaster management community in the design, development, operation, and maintenance of the DIN throughout its evolution. Involvement means establishment of a user/provider forum wherein information users and data providers openly discuss their capabilities and needs and together address each of the major natural hazards in the context of each of the four emergency management system phases (mitigation, preparedness, response, and recovery). Through the forum the needs of existing and new users could be defined through a taxonomy of user profiles that define the type, format, specificity, timeliness, and accuracy of disaster information desired for specific natural hazards.

  Thus, through user/provider activities, opportunities for avoiding redundancy and reducing disaster management costs could be realized by

• improved and better use of existing data collection and data management technologies in activities such as outreach programs and data-sharing cooperatives. An inclusive and ongoing user-provider interface is a prerequisite to the development of an effective DIN.
• A major focus of a DIN should be on integration of various data types. One challenging aspect of a DIN would be to determine how data collected from a large and diverse group of providers can be linked so as to be presented in a meaningful and timely way to an even larger user community. Development of common interfaces can facilitate this process by integrating distributed information sources. Prospective information providers must also have incentives that attract them to this process. Interpretive products such as images, graphs, tables, and maps must be developed with the end users in mind. In designing these “first look” products, emphasis should be on simplicity while maintaining user capability to investigate and question the data more deeply. To accomplish this, standards and protocols should be agreed on by a representative group of potential users and information providers. Existing information integration models should be assessed to facilitate discussion and identify the pros and cons of various approaches. The selected protocols must address communications, database structure, data formatting (by both data providers and DIN servers), hardware/software requirements, networking, and quality control. Consideration should be given to future system growth and the early and continuing evolution of interpretive products. User creativity should be fostered and encouraged by the DIN.

To maximize participation and cooperation by the many potential information providers, the data collection methods must be flexible and relatively easy to implement. Prescribing fixed formats would likely be met with resistance, thereby limiting the number of providers willing to participate in the project. Metadata files (files describing the data) are one means of achieving data acquisition format flexibility. The DIN should provide examples of metadata files and standards and explain how they would be used by the system. This should help encourage participation. Procedures will also be needed to manage and maintain the information that is collected. Routine follow-ups concerning metadata records will be important to ensure system and data integrity. A management structure and staffing requirements should be specified for handling these ongoing tasks.

Integrating and sharing information from a wide variety of sources is an exciting concept that many will embrace. However, using this resource for emergency decision support will require considerable training, confidence on the part of users, and a clear understanding of the network's inherent uncertainties and limitations. It will take time before the DIN can establish

itself as a credible tool for use during critical events. Developers of the DIN should consult with emergency managers, behavioral scientists, and other professionals on how best to address this issue.

Some of the most useful information products for disaster management could be derived by merging real-time with archival information. Some examples are as follows:

• Preparing for an approaching hurricane requires incorporating meteorological data into models that predict storm track and storm surge and then overlaying the results on maps showing population distribution, evacuation routes, deployment of emergency personnel and supplies, and other relevant response information.
• Responding to an earthquake involves rapid determination of magnitude and location followed by modeling to predict ground shaking and damage, which requires soil maps and building inventories, culminating in an assessment of losses and response requirements.
• Assessing the threat of rising floodwaters requires analysis of real-time stream gage data, modeling river basin and channel hydraulics, predicting the ability of levees and other defense structures to contain the flow, and anticipating problems.

Capabilities for integrating information would be especially helpful to disaster managers during the occurrence of compound disasters. For example, an earthquake in Southern California that occurs when Santa Ana winds are blowing off the Mojave Desert could result in widespread and difficult-to-contain wildfires. The capability to integrate earthquake and wildfire modeling could be crucial in responding. Similarly, the combination of a volcanic eruption and ash fall with heavy rain, such as occurred in the Philippines when Mt. Pinatubo erupted and Typhoon Yunya hit in 1991, would require integration of information to predict the weight of ash deposits on roofs, among numerous effects, as well as the ability to move equipment in such conditions. Add to this the occurrence of an earthquake, which did happen later in the Philippines, and the need for integration of information becomes of paramount importance.

Rapid information integration would also be of critical value in predicting or responding to technological or environmental problems caused by a natural hazard. Earthquakes can cause dams to breach or rupture fuel storage tanks, landslides can break pipelines, and fires can destroy wildlife habitats. Timely information delivered to the right decision makers clearly would significantly reduce losses. Special attention should be paid to automatic

data integration, for example, to trigger alarms, so as to assure accuracy of data and avoidance of false alarms.

• Data and information quality and reliability are major issues that need to be addressed. Although the Internet provides the means for accessing information resources, it does nothing to assure their quality and reliability. It would be easy to fall into the trap of thinking that more information is always better; however, bad information likely will lead to bad decisions. Thus, one of the challenges of a DIN would be how to evaluate the reliability of the information in the system.
  
  

  Emergency managers face a particular challenge in using disaster information because of the critical time-constrained nature of their situation. Therefore, a DIN would need to provide mechanisms for emergency managers to evaluate the reliability of data and information they receive. Data provided by government agencies, for example, should come with specific quality assurances, including dating. It is obvious that emergency managers who try to use a DIN for the first time during a crisis would be the most likely to have problems, which points to the importance of training.

• An effective dissemination and access plan is critical to the success of any information system. Time and attention are scarce resources and limit the amount of data that individuals and groups will want to collect. Awareness of the decision-making processes and biases of individuals is needed in order to design a useful disaster information system. Connectivity between users and sources of data and information should be assured in the design of a DIN. This linkage would be provided through the Internet, but in time of emergencies access to the Internet is likely to be disrupted. Because emergency management staffs must have access to communications, robustness and redundancy are critical elements of the system.
  
  

  Although the Internet may provide the common network for dissemination of information, the principles of robustness and redundancy lead to examination of other means and methods to achieve assured connectivity. Potential means and methods include private nets such as Intranet or Extranet, which allow controlled access to special communities, thus avoiding some connectivity problems in time of emergencies. At the source nodes, allowances for emergency managers to have priority access to information in time for decision-making should be implemented. Such procedures are not easily achieved on a national basis and will require additional administrative overhead and support.

• The GDIN Transition Team should focus initially on establishing a national DIN (i.e., with a U.S. focus), but the model should be extended to a global process (GDIN) as soon as it can be demonstrated that a DIN is technically and organizationally feasible. The team members represent agencies with generally limited international missions and experiences. It would be advisable to concentrate on development of a national process where agencies are most familiar and information technologies are well advanced. The process could be extended as soon as the U.S. model works reasonably well both technically and organizationally. Disasters are worldwide issues, and many of the relevant phenomena are global in nature. The International Decade for Natural Disaster Reduction¹ has heightened global awareness of the value of disaster information and opened lines of communication relevant to disaster issues, factors that should help facilitate extension of a DIN from U.S. to global proportions.