Toward an Earth Science Enterprise Federation: Results from a Workshop (1998)

NASA is experimenting with new arrangements for managing data and information from its Earth Science Enterprise (ESE) program, formerly known as Mission to Planet Earth. A federation is an association of autonomous partners that agree to abide by certain interface standards, business practices, and expectations of conduct to achieve a common goal. Federation provides a mechanism for representing the interests of a broader community. Although federations have been used as an organizational model for centuries, the concept has rarely been applied to scientific data management. Consequently, William Townsend, then Acting Associate Administrator of NASA's Mission to Planet Earth, requested that the NRC conduct a workshop to educate users and producers of ESE data about federations. In this report, the Steering Committee on a Workshop for an Earth Science Enterprise Federation examines the federation concept, compares different governance models, and offers some lessons for managing scientific data in an ESE federation.

The roots of NASA's Earth Science Enterprise began in the 1980s. For the first time, earth scientists from government, industry, and academia used advances in computer technology to develop numerical models to understand the Earth as an integrated system of land, oceans, air, ice, and ecosystem processes. Simultaneously, advances in spacecraft and sensor technology led to a new generation of satellites that could provide vast quantities of remotely sensed data.

From these revolutionary developments in science and technology, the Earth Observing System (EOS) was born with the goal of providing new information on earth system processes and so foster an interdisciplinary research environment. In order for this research environment to thrive, NASA realized that it also needed a system for data acquisition, initial processing, back-up archiving, and distribution. The EOS Data and Information System (EOSDIS) was established to create this environment. (A brief history of EOSDIS is contained in NRC [1995b], NRC [in press]). NASA, with input from its earth science research community, defined a set of standard products, which will be stored and distributed by the Distributed Active Archive Centers (DAACs). At the time it was conceived, NASA used a traditional top-down, centralized management model for EOSDIS.

In the years since the inception of this model, several factors prompted NASA to consider changes in its approach:

• The growth and widespread diffusion of the Internet and World Wide Web brought vast amounts of information to a wide range of users. These new information technologies enhanced access to data services and transformed the ways scientists, businesses, agencies, and other organizations communicate.
• Advances in the capabilities of workstation computers enabled scientists to be both producers and users of data from their desktops.
• Demand for EOSDIS data products and information extended beyond the earth science community to embrace a broad range of users--policy makers, educators, business people, and the general public. The major constituents currently interested in this data and information are described in Box 1.
• The U.S. government encouraged commercial applications of earth science data.

The National Research Council (NRC) also recommended that NASA consider changes in direction. In 1994 an NRC report pointed out that the centralized architecture of EOSDIS would not allow users to combine data from different sensors, modify standard products to meet new scientific needs, or revise algorithms to process data for different purposes (NRC, 1994). The report concluded that the EOSDIS architecture was too rigid to support the scientific community for which it was built and recommended that the products be designed and controlled in part by the customers of the system. A 1995 NRC report went further, recommending that "responsibility for product generation, publication, and user services should be transferred to a federation of partners selected through a competitive process open to all" (NRC, 1995a). A follow-on report also recommended that NASA consider implementing the federation concept in stages (NRC, 1996). The report urged NASA to implement an "initial limited set of pilot or prototype federated projects, ... in the near term, while continuing to develop the framework of a fully federated system for the long term."

In response to changing conditions and NRC recommendations, NASA has decided to test the federation concept (see Box 2) as a means for restructuring EOSDIS. The federation experiment will involve more than EOSDIS functions; it will embrace all of the ESE program. Although NASA recognizes the potential benefits of a federated ESE, it is also mindful of the potential dangers associated with transferring major scientific functions of EOSDIS outside of the federal government. In order to make sure that EOSDIS continues to fulfill its broad public purpose, NASA intends to transfer functions in phases and to evaluate success along the way. If the experiment proves successful, EOSDIS functions will be awarded through a competitive bidding process. Meanwhile, NASA will continue to develop and launch satellites, ensure that standard data products are produced and distributed, and foster development of the federation.

The first phase of the experimental federation is a Working Prototype Federation of Earth Science Information Partners (ESIPs). NASA recognizes three types of ESIPs, which overlap with the four ESE constituencies described in Box 1.

• Type 1 ESIPs. These ESIPs are responsible for standard data and information products whose production, publishing/distribution, and associated user services require emphasis on reliability and adherence to schedules. Type 1 ESIPs include DAACs and data producers.
• Type 2 ESIPs. These ESIPs are responsible for producing innovative science information products and services, which primarily serve the global change and earth science communities. Type 2 ESIPs include data producers and global change scientists.
• Type 3 ESIPs. These ESIPs are responsible for providing innovative, practical applications of earth science data to a broad range of users beyond the global change research community. Type 3 ESIPs include knowledge brokers and for-profit businesses.

Members of the prototype federation consist of the Type 2 and Type 3 ESIP winners of two 1997 NASA Cooperative Agreement Notices (see Appendix D). According to these notices, the objective of the prototype federation is to experiment with and evolve processes to make earth science data easy to preserve, locate, access, and use for all beneficial applications, including those for research, education, and commerce. It was against this background that the workshop on an ESE federation was held.

As the earth science community begins the transition from the original EOSDIS model to a federation model, it is important to consider options for governance and other federation issues. This report is based on a workshop held in February 1998, background materials prepared by the Committee on Geophysical and Environmental Data (CGED), and a seminal paper on federations by Charles Handy (1992). The workshop was organized by the Steering Committee on an Earth Science Enterprise Federation, which operated under the auspices of the CGED. Representatives of existing federations (libraries, international organizations, industry, government, and academia) described the characteristics of their organizations in plenary sessions, and working

groups discussed broad federation concepts. More than eighty participants, including federation experts, representatives of existing federations, ESIP winners, DAAC managers, and the broader scientific community, attended the two-and-one-half-day workshop. The agenda for the workshop and the list of workshop participants are in Appendixes B and C, respectively.

Chapter 2, Working Group Summaries, raises the following questions:

• What are the objectives of an ESE federation?
• What are the major governance issues to be considered in the development of an effective ESE federation?
• What are the potential benefits and costs of joining an ESE federation? What are the criteria for judging the success of an ESE federation as a management model?

Chapter 3, Models for an ESE Federation, compares six federation models and provides lessons for managing ESE data and information. The models are described in Appendix A.