Toward New Partnerships In Remote Sensing: Government, the Private Sector, and Earth Science Research (2002)

There is strong support, according to discussions at the March 2001 workshop, for developing more effective public-private partnerships for the production of remote sensing data for scientific research. Such partnerships entail both strengths and weaknesses. A closer examination of the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) and NASA’s Science Data Buy (SDB) can provide a set of “lessons learned” for the development of future public-private partnerships to produce scientific data.¹ Although SeaWiFS collects coastal and ocean data and SDB collects land data, the steering committee’s examination of these cases is focused on the institutional arrangements involved in collecting the data rather than on the particular type of remote sensing data.

Strong economic reasons may exist for entering into a public-private partnership, but the benefits of a successful relationship are not merely financial. On the government side, a public-private partnership can be a means of providing scientists with access to research data that are otherwise unavailable. In the private sector, a partnership through which scientists use private sector data can contribute to the development of new commercial applications of the data. The steering committee learned at the workshop, however, that it is unrealistic for either the government or the scientific community to pursue a public-private data partnership with the expectation that scientists will also be able to obtain new research

funding from the private sector. As discussed at the workshop, it is similarly unrealistic for those in the private sector to expect that a public-private partnership could be a vehicle for obtaining federal subsidies to support an enterprise that is not commercially feasible. The steering committee learned that financial motivations for a public-private partnership cannot be ignored, and that if the partnership does not make financial sense, it will not come about. However, if economic motivations are the only reason for the partnership, the inevitable difficulties encountered in joining public and private sector approaches could create significant disincentives to successful programs. Other issues, such as the long lead time required for government contracts or the small size of the scientific market for data at this time, could also discourage participants from entering into working public-private sector relationships.

The NASA Science Data Buy is an experimental project to evaluate the utility of commercial remote sensing products for Earth science research.² The program was seen by the OMB and Congress as a means of providing scientists with additional sources of data. The SDB is managed by the Commercial Remote Sensing Program (now the Earth Science Applications Directorate) at NASA’s Stennis Space Center, under the direction of the Applications Division within the Office of Earth Science at NASA Headquarters.

The SDB was initiated in 1997 by NASA in response to Senate language in the NASA Authorization Act, FY 1997, that directed the NASA administrator, “where feasible and cost effective, to make acquisitions of space-based and airborne Earth remote sensing data services, distribution, and applications provided by the United States private sector . . . if such acquisitions fully satisfy the scientific requirements of NASA.”³ The Commercial Space Act of 1998 reiterated this sentiment: “The Administrator shall, to the extent possible and while satisfying the scientific or educational requirements of the National Aeronautics and Space Administration . . . acquire, where cost-effective, space-based and airborne Earth remote sensing data, services, and applications from a commercial provider.”⁴ The injunction to obtain data and services from commercial providers was not confined to Earth science data, although neither NASA nor private sector firms have yet provided planetary, space environment, or astronomical data on a commercial basis.

In FY 1997, NASA allocated $50 million for the SDB program to provide commercial remote sensing data to the Earth science community. Authorization to spend the $50 million on SDB continued through FY 2000. In FY 2000, NASA received $20 million in appropriations, under the heading “commercial data buy.”⁵ According to NASA, the SDB program fosters advances in global Earth systems research, strengthening the U.S. economy through the development of remote sensing technologies and new ways of doing business.

In May 1997, NASA’s Stennis Space Center issued a Request for Offers (RFO), soliciting proposals for Phase I of the SDB program. The announcement requested proposals for data sets that would provide important new scientific measurements or more cost-effective ways of supporting NASA’s Office of Earth Science research. These data sets would be available for Earth science and applications research. Information on data price, validation, rights, and applicability to NASA’s Earth science and applications themes (Land-Cover and Land-Use Change Research; Seasonal to Interannual Climate Variability and Prediction; Natural Hazards Research and Applications; and Long-Term Climate: Natural Variability and Change Research) was requested from prospective data providers. NASA received 18 proposals from companies under the Phase I program; 10 companies were selected to provide prototype data sets and products. The objective of Phase I was to evaluate the prototype data sets and to select those to be purchased in Phase II.

The final selection of data sets was conducted by an independent science assessment team that included academic and government scientists organized around NASA’s Earth science themes, looking at data quality, science relevance, data usability, data rights, and provisions for collaborations. NASA management considered the recommendations from the assessment group and, in 1998, selected five companies for the Phase II data purchases, as shown in Table 3.1. Phase II was completed in September 2001.

Scientists who currently have a NASA Earth science research grant can compete for SDB data by electronically submitting tasking requests for data at specific locations and on certain dates. A Science Tasking and Measurement Committee composed of scientists from NASA Headquarters and Stennis Space Center evaluates and prioritizes the data requests to ensure that the data support Earth science and applications research and that they provide valuable and cost-effective products.⁶

The role of Stennis Space Center includes management of the contracts with providers, tasks related to collecting research data, selection of and coordination with the researchers who receive the data, and verification and validation efforts to ensure that the data meet the quality and other specifications required for scientific research. Verification and validation services supplied by Stennis Space Center have not generally been provided by commercial companies.

Workshop participants reported that the SDB is working well in many ways. Scientists said that the application process is well designed and that requests are easy to submit. NASA officials said that they have more applicants than they can serve under the current program. Private sector representatives were enthusiastic about the opportunity to work with scientists. Most importantly, however, scientists at the workshop who had participated in the SDB expressed satisfaction with the quality of the data and with their spatial, spectral, and radiometric resolution. They said that the data are making significant contributions to their research and teaching. The high-resolution data from the private sector are used by some

scientists to calibrate less-expensive, lower-resolution data from sensors produced and managed by the government. In this way, the SDB not only provides new data for research, but it also enhances the usefulness of lower-resolution data produced in the public sector.

All participants saw the need for the increased communication between scientists and data producers in the commercial sector. Representatives of some private sector firms that participate in the SDB said that the opportunity for direct communication with their scientist “customers” is an important strength of the program. Others acknowledged that communication with scientists is important but too infrequent. Some participants thought that when scientists and private sector data providers are able to interact directly, it is more effective than when such communication is mediated by government officials.

Workshop attendees identified several areas that they believe need improvement. Some of these areas relate to differences in the way public sector and private sector remote sensing programs operate, and others relate to the way the SDB program operates. In the first category, scientists stated concerns about the impact of the private sector practice of obtaining data by task orders. In some cases, the concern was that science data requests through the SDB are postponed or given lower priority in the tasking queue when they conflict with the needs of commercial customers. Some participants also expressed concern about timely receipt of their data, which is related to tasking. If every order for science can only be filled by tasking and if obtaining the data is subject to delays caused by weather or cloud cover, the time between the SDB award to a scientist and having the data in hand can be lengthy. Some scientists reported delays of a year or more. Concern was expressed not only about delays in receiving data, but about what this signifies regarding future private sector support for scientific data. If, they reasoned, private sector data producers have not given sufficient priority to scientists’ data requests during the experimental phase of the SDB, the data buy might limit scientific applications of commercial data in the future. In the longer term, delays in obtaining science data could undercut the utility of public sector data purchases for scientific research.

The lack of historical or “heritage” data from the commercial remote sensing sources under the SDB was another issue raised at the workshop. While this problem may limit the usefulness of the data, particularly in global change research, it could diminish over time as the availability of commercial data for scientific research increases. However, an inherent limitation is that commercial data are not routinely collected, as they are usually tasked to meet customer requirements. This characteristic may limit the routine production of data acquired at multiple time periods that can be archived and made available in future years.

Workshop participants also expressed concern about the lack of consistency in intellectual property provisions among the various data providers in the SDB. Each of the five companies that participates in the program negotiated separate intellectual property provisions with NASA. This is not surprising, given that standard practices for the treatment of intellectual property in the remote sensing industry are still evolving. However, since intellectual property issues are a pervasive concern in the scientific community, this area was seen as one that needs improvement. Participants suggested that, in the future, a single, standard agreement for all participants in the data buy program would be more useful to scientists and would promote the development of common policies.

Differences in data purchase policies of firms in the program also posed a problem for some workshop participants. One vendor used Landsat data to provide orthorectified global land data sets using data collected during the mid-1970s and in 1990. Other vendors, who provided data through aerial and spaceborne platforms, needed a minimum-sized ground area to obtain the requested data. This meant that scientists who wanted data over a small study site might have to accept data for an area 10 to 100 times the size of their request. Consequently, they would have to search through a large data set for the requested data and store data that were not required for the study. Scientists reported that paying for data that are irrelevant to their scientific needs is an inefficient use of government resources; they would rather obtain repeat coverage of their study sites than coverage over areas that are not of scientific interest to them.

Workshop participants from federal agencies expressed some concern that participation in the SDB was limited to NASA grantees instead of being open to the broader Earth science research community in both academia and government. Scientific data that NASA collects from its own satellites (e.g., Terra) are, in contrast, open to anyone. These participants noted that scientific and policy research in government agencies is increasingly based on remote sensing data, that the cost of data is a problem for some agencies, and that they could also benefit from participation in the SDB.

Originally envisioned as a data buy arrangement, the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) is the first example of a public-private partnership to procure scientific data.⁷ In 1991, NASA signed a fixed-price contract with Orbital Sciences Corporation (OrbImage, a subsidiary of Orbital Sciences, purchased the spacecraft from Orbital after launch) for $43.5 million that guaranteed the purchase of five years of ocean color data. These data would provide

detail on global ocean bio-optical properties that are important to Earth science research.⁸ The agreement grew out of concern in the ocean biogeochemistry community over the availability of follow-on data to those collected by the Coastal Zone Color Scanner on Nimbus-7 that operated from 1978 to 1986. Ocean color data are important for understanding the process by which oceans store carbon, a cycle believed to influence the global climate; the data provide evidence of phytoplankton concentrations and thus also indicate where fish feed; and, the data are commercially valuable. Consequently, a public-private partnership seemed a good vehicle for obtaining the data.

In the agreement between NASA and Orbital Sciences, the private sector partner incurred the costs of instrument and satellite development, placement in orbit, and data acquisition. NASA scientists developed the specifications and design for the sensor, and the agency was responsible for disseminating data to the scientific community. Through an open competition, NASA appointed a science team associated with SeaWiFS and also staffed a project office in the agency that played pivotal roles in data processing, instrument calibration, and product validation.

Since 1991, the relationship of the partners has evolved in ways that make SeaWiFS an instructive example of a public-private partnership, or “hybrid” approach. NASA’s upfront investment in SeaWiFS provided an “anchor tenant” for Orbital Sciences. Although the original plan called for launch of the satellite in 1993, it was not successfully launched until 1997. The launch was late because of delays in construction of the satellite and problems with the Pegasus XL launch vehicle. The delay cost NASA approximately $2 million to $3 million per year over a four-year period to keep the SeaWiFS project office running. During this time, NASA focused on data preparation and documentation.⁹ As a result of this preparatory work, NASA was capable of providing high-quality scientific data in a matter of hours after initial data collection from the sensor.

From the point of view of a data user, the SeaWiFS partnership of scientists, government, and the private sector has been a success in many ways, providing a constant data stream with global coverage of ocean chlorophyll since late 1997. From the scientists’ perspective, SeaWiFS provides high-quality data unmatched by those from any other ocean color sensor. It is less clear that the government and private sector view SeaWiFS as a success, given delays of launch following

development of the contract in spring 1991 and unrealized revenues from the projected commercial market that have made SeaWiFS unprofitable from a business perspective.

Reports to the workshop from NASA and OrbImage highlighted successful aspects of the partnership, focusing on the essential contribution of each partner in specific areas of calibration, validation, processing, archiving, and dissemination of data. Criticisms voiced by participants at the workshop focused on inadequate estimates and lack of full accounting of the true costs of the partnership, which actually included significant investment beyond the contract price. A fair assessment of SeaWiFS must take into account that it was not a simple data buy, because without contributions from the scientific and government partners, no usable products could have been generated. From a scientific perspective, it provides global data on ocean chlorophyll that are important for research.

In establishing the partnership, stipulations on data distribution attached to SeaWiFS dictated a delay of 14 days to allow the private sector partner to market real-time data to prospective users of ocean chlorophyll products—particularly the commercial fishing industry. Because the market did not mature as forecast, the impact of these stipulations on the commercial venture was unclear. For most scientific purposes, the delay posed little or no impediment, as a great deal of the research relying on satellite ocean color data addresses plankton dynamics and requires analysis over periods of weeks to years, rather than immediate access to data. Exceptions to the embargo policy are made for users that have real-time needs on a request basis (for example, oceanographic projects requiring imagery may request data in real time), and a number of such requests have been honored by the project office and private sector provider.

Scientists would like to extend the SeaWiFS mission beyond its intended 5-year duration, and serious attention is now being given to a contract extension to enable continued operation. Government and industry partners reported that they would like improvements in the terms of a new contract—for example, that a longer timeframe be specified so that the private sector might realize sufficient revenues to cover costs and make a profit. It is clear that this approach to acquiring Earth science data from space by partnering with the private sector has pluses and minuses. It is also clear that, with the extensive participation on the part of the government in data processing and distribution, the output of SeaWiFS has been a resounding scientific success.

Both partners in this public-private partnership also find room for improvement. From the perspective of the private sector partner, the 5-year data purchase contract was too short, given the extension of the satellite beyond its 5-year design. In addition, during the development of SeaWiFS, a number of contract changes were imposed by the government. Private sector representatives at the

workshop argued that their costs were increased by contract changes, changes in the regulations that affect the contract in future missions, or delays in the contracting process. Other issues that arose included a need to price the product to reflect the investment, based on market analyses that reflect the true costs of providing data both to government-sponsored users and to commercial users. That is, market analyses should estimate both government and nongovernment revenues. It would also be useful to have the government guarantee in writing that it will meet its part of the contract over the life of the program.

From the government partner perspective, additional investments—which included carrying the science team over an extended period—were required because of delayed data delivery. There is a residual, yet unanswered question of whether a realistic cost accounting of the system would have shown that the partnership concept reduced the total public-private investment required for the mission below what would have been required by a government-only system.

The SDB and SeaWiFS represent two examples of the federal government’s efforts to create public-private partnerships that lead to new types of remote sensing data for scientific research. In SeaWiFS, the government is involved throughout the process of producing the data by virtue of its role as a principal customer, or “anchor tenant” of the program. In the SDB, remote sensing data are produced by the private sector, and the government agrees to purchase specific amounts of data to be used for scientific research. Although the two partnerships represent different approaches, some issues related to public-private partnerships are common to both examples and arise in other types of public-private partnerships. These cross-cutting issues are discussed in Chapter 4.