Functional area criteria varied among the 12 types of stewardship data and often focused on content issues. For example, criteria for the existing hazards data type included the following:

• The record should contain scientifically accurate descriptions of particular information.
• The record should have been developed by a site contractor or a credible scientific organization (e.g., EPA, NRC, a national laboratory).
• If the record is a map, it should contain an accurate representation of site facilities and boundaries over time, should be referenced to a commonly applied coordinate system (e.g., USGS), and should reference regional or state landmarks.

Additional criteria apply across all of the individual data types:

• Information should be available in lay terms, with summaries suitable for use by a nonscientific audience or someone not familiar with site and DOE terminology.
• Underlying data should be available to allow independent verification of conclusions and additional analysis as needed.
• Information should be in final form (i.e., approved by information release processes that were active onsite at the time the information was generated, including peer review, line management review, and classification review as appropriate).

To help sites identify stewardship data, DOE may need to provide policy-level guidance that defines stewardship data, describes its importance to DOE, and outlines how to recognize data with stewardship value. The guidance could further describe:

• The types of information to be kept, including record materials, samples, documents, original data, electronic media such as video and audio tapes, and Internet sites.

CHAPTER 5

• How to provide for ownership of records, including responsibilities for generation, retention, and maintenance of retrievable data.
• A comprehensive set of requirements for information format (e.g., electronic, hard copy), consistency among sites, and information controls (e.g., access, availability).
• How to evaluate existing records retention schedules with respect to stewardship data needs.
• How to identify necessary information that may already have been sent elsewhere (e.g., Office of Scientific and Technical Information, National Archives). The Office of Scientific and Technical Information (OSTI) currently is holding many DOE records but is in the process of divesting itself of this responsibility. Some sites have destroyed records assuming the information is being preserved by OSTI. If OSTI destroys the only copy of materials, these records would be lost.
• How to incorporate classified information and otherwise sensitive information of stewardship value (e.g., export controlled, unclassified nuclear, applied technology, litigation sensitive) when special protection for this information is no longer needed.
• How to ensure appropriate stakeholder involvement in identifying stewardship data.

Criteria for identifying stewardship data should be developed as soon as possible to prevent their future loss.

5.2 Modify Existing Records Retention Schedules

As noted in Chapter 4, many of the current requirements and practices for data retention are shorter than the long time periods needed for stewardship purposes. For example, DOE records retention schedules require some records to be archived permanently; others to be destroyed after periods of time that range up to 80 years; and others to be destroyed even before cleanup is complete. Data preservation practices are governed by a variety of laws and regulations, each of which applies to some, but not all, types of stewardship data. While it would be advantageous to attempt to modify data preservation requirements under federal laws such as RCRA, CERCLA, and AEA, DOE has more flexibility in modifying existing records retention schedules.

Existing DOE records retention schedules do not appear to cover all of the 12 stewardship data types (see Appendix B). Even within a particular data category, it does not appear the records schedules require all relevant stewardship data to be preserved permanently. While some of this information will be preserved permanently pursuant to other requirements (e.g., property transfer), stewardship data preservation would be greatly enhanced by:

• Modifying the DOE records retention schedules to cover all stewardship data types.
• Ensuring that all stewardship data are required to be permanently preserved.

Records retention schedules should be revised as soon as possible to prevent the future loss of stewardship data. It will be important to couple any attempts to modify the records schedules with the development of clear criteria for identifying stewardship data. Simply increasing the number of data types to be permanently archived without corresponding criteria to identify the subset of data required for long-term stewardship would lead to an unworkable situation in which nearly every piece of information was permanently archived.

POTENTIAL SOLUTIONS

5.3 Develop Metadata for Stewardship Data

As noted in Chapter 4, current requirements and practices do not seem to require the preservation of sufficient contextual information to allow future generations to be able to understand the nature and context of stewardship data that are preserved. For hard copy records, much of this contextual information is provided by indexes and other finding tools, but there do not appear to be any requirements for the use of standardized indexes or pointers. Several types of federal metadata standards apply to some types of electronic records (e.g., geospatially referenced information), but not all types of electronic records appear to be covered by these requirements. As DOE sites move toward electronic-based information management systems (see Figure 5-1), improved metadata standards will become more important for stewardship data. Metadata standards also are an appropriate focus because geospatial referencing may be an effective tool for users to locate stewardship data.

Much of the information required to support long-term stewardship activities involves spatially referenced data. For example, future users may wish to know:

• Where were the buildings located in Operable Unit 3?
• Where were the drain pipes located from Building 79 in 1965?
• What is buried in this spot?
• What activities occurred in room 23 of this building?
• What uses are proposed or planned for each portion of the site?

The metadata standards developed for spatial data are a logical starting point for preserving the essential contextual information about stewardship data. The existing Federal Geographic Data Committee (FGDC) Content Standards for Digital Geospatial Metadata primarily describe published data sets, describing their structure, content, and data quality issues without making specific refer-

CHAPTER 5

ence to the particular needs of stewardship for DOE facilities. For stewardship purposes, DOE needs to develop metadata standards that are well-suited to describing both published and non-published data. DOE also will require additional stewardship-specific information not explicitly referenced in the generic FGDC standards.

DOE needs at least two different types of metadata standards. First, in the near term, there is a need to screen a broad range of data sets for their potential stewardship value. By necessity, this screening must be a rapid process and may often need to depend on comparatively sketchy information in existing indexing systems or on documentation prepared by individuals who are not subject matter experts. Second, once a data set is identified as having stewardship value, the content, quality, condition, and other characteristics of the data set need to be documented sufficiently to ensure the long-term usability of the data.

The FGDC core metadata elements (i. e., the mandatory elements) are an appropriate starting point for rapid screening of data sets for stewardship value. These elements are relatively easy to complete for a data set, and if formal records of the data sets already evaluated have been maintained, the amount of detail provided by the core metadata elements should be sufficient to avoid repeatedly re-evaluating the stewardship potential of the same data sets. Moreover, providing the amount of detail required by the full FGDC structure can take substantially more time, particularly when creating after-the-fact documentation for existing data sets. Therefore, the full FGDC metadata elements do not appear appropriate during the screening process. Table 5-2 lists the data fields included as sets of metadata elements whose output meets the minimum data collection requirements of the FGDC core elements.

Data Fields
Identity of this entry (for future tracking and updating) Originator Publication date Title of data set Edition Presentation form (e.g., map, atlas) Publication place Publisher Online linkage (URL) Abstract Purpose Supplemental information Beginning and ending dates Currentness reference Progress Intended data set maintenance and update frequency Bounding coordinates (West, East, North, South)	Theme keywords Theme keywords reference Place keywords Place keywords reference Limits on data accessibility Limits on use of data Browse graphic URL Browse graphic caption Browse graphic file type Spatial data type Distribution organization Distribution contact position/ person Address type Street address, city, state or province, postal code, country, phone, fax, e-mail Data set name as known by distributor Liability held by distributor Date of last metadata entry or update (year, month, date)
On-screen forms are available on the Internet at: http:/ www/ fgdc/ gov/ Clearinghouse/ MetadataESystem/ metaform. html

POTENTIAL SOLUTIONS

Although data sets can be documented comparatively quickly using only the mandatory data fields, the full FGDC structure is designed to describe fully the characteristics of a data set and would be required for future use of the data. Data sets with likely stewardship value merit published descriptions sufficient to let potential users identify the suitability of a data set to their purpose, obtain the information, and contact the creators of the data for further information if necessary. This is the essence of the FGDC and Government Information Locator Service (GILS) approaches outlined above. However, the original creators of the data will not be available over the long life-cycle of stewardship data, so it will be important to complete metadata documentation sufficiently to ensure utility of the data over decades or centuries. Therefore, even the full set of FGDC metadata standards may not be sufficient for stewardship data.

To document key data sets for future stewardship use, it will be necessary to go beyond the existing FGDC metadata standards. The FGDC process provides a suitable framework, because supplemental profiles tailored to a specific type of data can be formally proposed and approved. However, additional information to be documented for stewardship may include:

• Factors regarding the overall quality or importance of information in the data set that will preserve information on the relative importance of data sources for future users:
  • Whether the metadata author is a subject matter expert, or whether the author was able to consult the original data set creators.
  • How the data set compares to others with information for the same category or categories of stewardship data.
  • Desrciption of the importance of the information in the data set.
  • Information on the priority of this data set for inclusion in a permanent archive.
• Factors to assist future users in locating pertinent stewardship data include:
  • Which types of stewardship data are covered by the data set.
  • Whether the data set includes onsite and/ or offsite information.
  • Whether there are any restrictions on the public release of information in this data set.
• Factors to assist in preserving stewardship data in the near term until the data sets are transferred to a permanent repository include:
  • Whether the data set is presently housed in an environment where it is likely to be safely preserved.
  • A date, if any, when the current location at which the data set is housed is scheduled to be shut down.
  • Contact information for the person currently responsible for preserving the data set.
  • Whether the physical media or computers holding this data set have been tagged as containing stewardship data.
  • Specific data set tracking information to document the screening process for identifying data sets with stewardship value.

Depending on the eventual design of a system to manage stewardship data, several other types of metadata elements may be required. The FGDC standards, for example, would allow a user to identify and obtain all records that pertain to a given building at a site for a particular time period. However, the standards by themselves would not ensure that the user would be able to merge the data from a variety of sources into a single data set. If a decision is made to merge a number of stewardship data sets together into a unified data set, then additional information will need to be col-

CHAPTER 5

lected and preserved. This would include the complete specifications defining the information content of the data set together with the transformation rules used to incorporate the source data into the unified data set. A stewardship archive could take this approach or preserve the source data sets in their original form.

5.4 Develop a System to Access Stewardship Data

As noted in Chapter 4, perhaps the most difficult challenge involving stewardship data is their accessibility. Under current requirements and practices, people get access to archived information primarily by request. Information is preserved in a number of places, so it is difficult for users to know that relevant data may exist and where to look for these data. When information is located, it may take more than a year to retrieve it. The lack of adequate contextual information may make it difficult or impossible to use any information that is retrieved.

This discussion outlines the elements of a system DOE could adopt to deliver stewardship data to appropriate users. The discussion considers both the requirements of an overall system to manage stewardship data and the types of roles, responsibilities, and other practices that need to be established to manage and operate such a system.

5.4.1 ELEMENTS OF A STEWARDSHIP DATA SYSTEM

Any system for managing stewardship data must be able to perform two key functions: (1) maintain physical control of stewardship data from the time they are identified until they are no longer needed (if such a time can be identified); and (2) enable appropriate users to find and retrieve these data in a timely manner. The first requirement is essentially an inventory or asset control problem; the system must be able to track the location and status of all physical and electronic units of stewardship data and ensure that these data are being adequately preserved. The second requirement is essentially an accessibility problem; the system must allow appropriate users to identify, find, and obtain all units of stewardship data that may be of interest.

Both physical control and accessibility must be maintained throughout the full life-cycle of stewardship information. One of the unique challenges for stewardship data is maintaining accessibility even when physical control is transferred from one entity to another. There are five key elements to include in a stewardship data system:

1. Electronic archive. All records that have stewardship value would be scanned or otherwise preserved in an electronic format and included in the electronic archive. Hard-copy versions of these records would continue to be archived under the existing requirements and practices (e.g., through NARA). Some design principles for electronic archives stress a need to "plan for chaos" (i. e., distribute copies of records to many repositories) to maximize the likelihood long-term survival.

2. Electronic index. An electronic index would include a standard outline/ format, a uniform set of metadata (with appropriate standards), a system for geospatially referencing all holdings, and a standard thesaurus of index terms. The index would allow users to know what data are available, sort through these data in a preliminary manner, and find and retrieve all records of interest. The index also would include "pointers" to records that were reviewed for stewardship value but not included in the electronic archive. These might be records with no stewardship value (e.g., records of parking permits) or records that had potential stewardship value but were superseded by more accurate or up-to-date records (e.g., draft reports where a final report was issued, documents that contain secondary references or reproductions of primary sources).

POTENTIAL SOLUTIONS

3. Delivery system. A delivery system would provide users with electronic copies of records contained in the electronic index upon request and would allow users to request hard copies of material, if needed. A key requirement would be rapid response; users would need to obtain electronic materials in a matter of minutes or hours.

4. User interface with a search engine. A user interface would enable users to access the stewardship data system, search for stewardship data, and retrieve electronic materials from the stewardship archive. The interface would need to facilitate searches to be conducted at varying levels of geographic detail (e.g., at the national level, for a state or region, for a particular site, and for a particular facility or portion of a site).

5. Maintenance system. A maintenance system would be necessary to ensure that formats for maintaining electronic information, search engines, user interfaces, and all system components remained current with existing technologies. Without a systematic approach to maintenance, obsolescence would pose a significant risk to the long-term viability of stewardship data.

Two systems under development may be useful to consider when designing a stewardship data system. First, the National Environmental Data Index (NEDI) is being developed to assist in integrating a diverse set of environmental data generated by several federal agencies. The design principles for this index may provide guidance for how to integrate a diverse set of stewardship data, currently in multiple formats, across all DOE sites. Second, DOE has begun to develop a data repository and retrieval system for the proposed geologic repository at Yucca Mountain, Nevada. This system is designed to improve the accessibility, traceability, and transparency of data critical for decisions related to granting permits for the proposed geologic repository. The system includes an electronic archive, indexing system, interface, and search engine, and currently handles requests for about 15,000 pages of information per month. The Openness Advisory Panel of the Secretary of Energy's Advisory Board concluded that existing technologies and expertise are sufficient to extrapolate experience at Yucca Mountain to the entire Department. ¹

Design Principles for the National Environmental Data Index (NEDI):

Support multiple metadata standards Use Internet and other communication links, as needed Develop the system as a distributed data index Support distributed searches (through FIPS 192/ Z39.50) Use existing standards and off-the-shelf software Support multiple interfaces to span the range of user needs Allow for multiple access points to NEDI

5.4.2 PRACTICES AND PROCESSES

A system for managing stewardship data will need to establish and codify practices to ensure physical control and accessibility of all information from the time it is identified as stewardship data until it is no longer needed for stewardship purposes, or indefinitely. These practices must be clear and simple enough to be followed by DOE and current contractors as well as future site stewards, particularly when responsibility for a given activity shifts from one entity to another. These practices also need to ensure that appropriate users can access stewardship data for decades or centuries. Any system for managing stewardship data must establish a process for:

• Approving the release of information into the system for managing stewardship data during cleanup and long-term stewardship.
• Determining who retains responsibility for data identification, preservation, and accessibility during cleanup and long-term stewardship.

¹ The Prospects for Introducing a Comprehensive Electronic Records Management System into the Department of Energy. Secretary of Energy Advisory Board, Openness Advisory Panel, Draft Subgroup Preliminary Assessment Report, November, 1997.

CHAPTER 5

• Ensuring adequate geospatial referencing of stewardship data.
• Determining the extent to which hard or electronic copies of records should be preserved in locations other than NARA archives.
• Preserving a directory that points to the locations of all available data, including those preserved electronically and those preserved in hard copy.
• Determining who retains responsibility for developing and maintaining user interfaces that allow access to stewardship data during cleanup and long-term stewardship, and ensuring that the user interface remains intuitive and easily understood by non-DOE entities.
• Ensuring that existing standards, protocols, and requirements are being met, particularly when non-federal entities are involved.
• Determining formats that are appropriate for the long-term preservation and accessibility of electronic data, particularly with regard to the eventual obsolescence of hardware and software.
• Determining technology solutions that can ensure adequate infrastructure for data transfer/distribution, user interfaces, and data integration/ conversion software.
• Ensure adequate funding for stewardship data needs.

5.4.3 INSTITUTIONAL FRAMEWORK FOR STEWARDSHIP DATA

Although DOE sites can take many steps now to begin implementing practices and processes for addressing stewardship data needs, a more systematic approach is needed to coordinate and focus efforts across all DOE organizations and sites. To do this effectively, DOE needs to develop an institutional framework to generate, preserve, and provide access to stewardship data. Since the stewardship mission differs significantly from missions of existing organizations within DOE, a specialized stewardship data entity would likely be the most effective means of providing for stewardship data needs. A distinct stewardship data entity also would mean that funding for long-term stewardship can be addressed directly through the annual budget process, rather than dispersed as an indirect cost in a variety of DOE offices.

It is impossible to determine how many entities might be involved in stewardship at the local, state, regional, and/or national levels or whether these would be government agencies, non-governmental organizations, or commercial enterprises. A variety of options exist for developing an institutional framework and for distributing responsibilities associated with managing stewardship data among current and future stewardship entities. The following three options for managing stewardship data describe the range of possibilities for designing such a framework:

1. The dispersed option. In this option, numerous entities would be responsible for identification, preservation, and accessibility of stewardship data. The primary advantage of such an option is its flexibility. Sites would be allowed to establish agreements with a variety of entities so that responsibilities could be matched with site-specific needs and circumstances. The primary disadvantage would be coordination. It would be difficult to ensure that existing and future requirements, standards, and protocols are followed adequately, particularly when governmental and non-governmental entities are involved.

2. The concentrated option. In this option, a single, national entity would be responsible for identification, preservation, and accessibility of stewardship data. The primary advantages of such an option are efficiency and continuity. A single entity would make it easier to maintain standards and practices and could function much like a library does today. Needed modifica-

POTENTIAL SOLUTIONS

tions such as hardware/ software upgrades could be accomplished in a coordinated and systematic manner. The primary disadvantage would be inflexibility. A "one size fits all" approach to managing stewardship data may not be the most appropriate response to the needs of a specific site. For example, stakeholders have expressed a strong desire to maintain permanent archive repositories near the sites, which might be difficult under this option.

3. The hybrid option. In this option, some responsibilities would be concentrated in a single entity, while others would be dispersed among multiple entities. For example, several entities might be responsible for maintaining data archives, while a single entity might be responsible for identifying stewardship data, ensuring they are preserved in data archives, and ensuring data accessibility and re-distribution. This option would be intermediate in terms of flexibility, efficiency, and the need to maintain coordination. Under this option, DOE could establish a stewardship data entity that maintained the overall responsibility for stewardship data but retained the ability to distribute specific responsibilities to other entities (e.g., the Department of Interior could maintain an archive for all records related to endangered/ threatened species).

One of the key functions of a stewardship data entity would be coordination of stewardship information management activities at all the sites for which DOE is responsible as they complete cleanup and other missions and prepare for closure or transfer and for long-term stewardship. Another key function would be to maintain the electronic archive and indexing/ metadata system critical for data accessibility. This entity or function could be located at the field level, preferably at a site with a well-defined long-term mission. Ideally the site would already have the resources, personnel, expertise, and technologies needed for the stewardship data functions.

Establishing a new data function in a central facility would require some investment of effort and funds. Funds are currently being spent on data retention with no assurance that the systems and data needed by long-term stewards will be available. Given the findings cited in Chapter 4, a single, effective data preservation system would reduce costs because it would prevent the loss of records, eliminate the need to regenerate information, and possibly help avoid site closure delays.