全球智库信息集成服务系统

PREFACE

This paper was originally written for the conference on "No More Secrets? Policy Implications of Commercial Remote Sensing Satellites," held at the Carnegie Endowment on May 26, 1999. Sponsored by the Endowment’s Project on Transparency, the conference brought together representatives of a score of countries, many U.S. government departments, congressional staff, several nongovernmental organizations, the private sector, and the media for a day-long immersion in the technical and policy issues associated with the emergence of commercial high-resolution satellites. The conference report is available from Ann Florini, director of the Project on Transparency. Highlights of the conference can be seen on the Endowment’s web site at www.ceip.org. The Project on Transparency is deeply grateful to the Rockefeller Brothers Fund for its financial support for the conference.

INTRODUCTION

Ever since the earliest satellites and astronauts started taking pictures of the Earth from space nearly four decades ago, those images have inspired excitement, introspection, and, often, fear. Like all information, satellite imagery is in itself neutral. But satellite imagery is a particularly powerful sort of information, showing both comprehensive vistas and surprising detail. Its benefits can be immense—but so can its costs. The same images that remind us that we all share a fragile planet also enable those who have the images to more accurately aim their weapons at adversaries near and far.

And now, the number of people able to use that imagery is exploding. By the turn of the century, new satellites will likely be in orbit with capabilities approaching those of military spy satellites, but with one key difference: the satellite operators plan to do their best to sell that imagery to anyone able and willing to pay. An America-Russian joint venture is already selling imagery that shows objects as small as two meters across (available through the Web at www.terraserver.com). An American company, a Lockheed offshoot called Space Imaging, launched a substantially more sophisticated satellite with one-meter resolution in April 1999. That satellite never reached orbit due to a problem with the launch vehicle, but Space Imaging has a backup satellite nearly ready to go, and a dozen other U.S. companies have licenses to launch high-resolution imaging satellites over the next few years.

The new commercial satellites will make it possible for the buyers of that imagery to, among other things, tell the difference between trucks and tanks, expose movements of large groups like troops or refugees, and pinpoint the probable location of natural resources. Whether all this amounts to a positive or negative development depends on who chooses to use the imagery and how. On the plus side, governments and nongovernmental organizations (NGOs) may find it easier to respond quickly to sudden refugee movements, to document and publicize large-scale humanitarian atrocities, to monitor environmental degradation, or to manage international disputes before they escalate. The United Nations is even looking into the possibility that satellite imagery could help control drug trafficking and narcotics production. But there is no way to guarantee benevolent uses. Governments, corporations, or small groups could use the imagery to conduct industrial espionage, collect intelligence, plan terrorist attacks, or mount offensive military operations. And even when intentions are good, it can be remarkably difficult to know what you are looking at. The media have already made major mistakes, misinterpreting images of everything from Chernobyl in 1986 to the Indian nuclear test site just last year.

Such bloopers notwithstanding, the new satellite imagery will reveal to many people information to which they never before had access. That makes the advent of these satellites both important in itself and a microcosm of a larger trend now sweeping the world—growing transparency. It is becoming more and more difficult to hide information, not only due to technological improvements, but also because of spreading norms about who is entitled to have access to what information. The idea is gaining ground across issue areas and in many parts of the world that governments, corporations, and other concentrations of political and economic power are obliged to provide information about themselves to others. In politics, several countries are enacting or strengthening freedom-of-information laws that provide citizens the right to information on governmental records. In environmental issues, the current hot topic is regulation by revelation, in which polluters are required not to stop polluting but to reveal publicly just how much they are polluting. Such requirements have dramatic effects. Partly to avoid being labeled as big polluters and partly to stop wasting materials they had not previously realized they were wasting, many companies covered by such laws have drastically reduced emissions. In arms control, mutual inspections of sensitive military facilities have become so commonplace that it is easy to forget how revolutionary the idea was a decade or two ago. As democratic norms spread, as civil society grows stronger and more effective in its demands for information around the world, as globalization gives people an ever greater stake in knowing more about what is going on in other parts of the world, and as technology makes such knowledge easier to attain, transparency would appear to be the ineluctable wave of the future.

The legitimacy of remote sensing satellites is part of this global trend toward transparency. Imagery from high-resolution satellites is becoming available now not only because technology has advanced to the point of making the imagery a potential source of substantial profits, but because governmental policies permit, and indeed encourage, such satellites to be operated. Yet as is always the case with increases in transparency, not everyone benefits and not all uses of the resulting information are benign. Governments remain concerned about just how far this new source of transparency should be allowed to go. The provision of information shifts power from the former holders of secrets to the newly informed, which has implications for national sovereignty, for the ability of corporations to keep proprietary information secret, and for the balance of power between state and nonstate actors.

Sovereignty is directly challenged by the new satellite systems. If the satellite operators are permitted to image territory anywhere in the world and sell that imagery to anyone, governments lose a significant degree of control over information about their territory. Although both spy satellites and civilian satellites have been doing this for years, the operators of the spy satellites have been remarkably reticent about the information they have collected, making it relatively easy for countries being imaged to ignore them. Pakistan and India may not like being observed in such detail by the United States and Russia, but as long as the high-resolution satellite operators were not showing information about Pakistan to India and vice versa, objections remained muted. Although the civilian satellites that operated before the 1990s did provide imagery to the public, they had much lower resolution, generally not showing objects smaller than 10 meters across. That low-resolution imagery could provide only very limited militarily significant information, nothing like the information that will be available from the new one-meter systems.

Under international law, countries have no grounds for objecting to being imaged from outer space. The existing standards, the result largely of long-standing U.S. efforts to render legitimate both military reconnaissance and civilian imaging from space, are codified in two UN documents. The Outer Space Treaty of 1967 establishes the notion that outer space cannot be claimed as national territory, legitimating satellite travel in space over any country’s territory. And despite years of efforts by then-Soviet bloc and developing countries to claim a right of prior consent in order to review and possibly withhold any data about their territory, in 1986 the UN General Assembly adopting a set of legal principles on civilian remote sensing that make no mention of any right of prior consent. Instead, the principles merely require that "As soon as the primary data and the processed data concerning the territory under its jurisdiction are produced, the sensed State shall have access to them on a nondiscriminatory basis and on reasonable cost terms."

In other words, if a country being imaged knows that it is being imaged and asks for a copy, it is entitled to buy one at the going rate. Even then, countries won’t know who is asking for specific images, or for what purposes. Does a request for imagery of military bases come from an NGO trying to monitor a country’s compliance with some international accord, or from an adversary preparing to strike? There is a major economic concern as well. Corporations may know more about a country’s natural resources than the government does, putting governments at a growing disadvantage when negotiating drilling rights or other agreements. And as we have all seen recently, highly visible refugee flows and humanitarian atrocities can attract intense attention from the international community, even while less visible disasters are ignored. The new imagery poses the possibility that the ability of NGOs and the media to track refugee flows or environmental catastrophes will encourage more in the way of international interventions, even over the resistance of the host government. Whether the rather lackadaisical protection of sovereignty contained in the 1986 legal principles will still satisfy governments whose territory is being observed by large numbers of increasingly observant commercial satellites seems questionable.

Governments are not the only ones who may feel an acute new sense of vulnerability. Corporations may find themselves being observed by competitors trying to keep tabs on their construction of new production facilities around the world and estimate the size of their production runs by looking at their emissions. This is not corporate espionage as usually defined, because satellite imaging is thoroughly legal. But it could in time make it quite difficult for corporations to keep their plans and practices secret.

And it will not be only competitors who want to keep an eye on corporations. Environmentalists, for example, may find the new satellites a useful means of monitoring what corporations are doing to the environment.1 Karen Litfin has pointed out that environmental NGOs already make extensive use of the existing relatively low-resolution imagery that is publicly available to monitor enforcement of the U.S. Endangered Species Act, to document destruction of coral reefs around the world, and to generate plans for ecosystem management.2

Environmental groups are just part of a host of nongovernmental organizations likely to take advantage of this new source of information. Some groups that work on security and arms control have already used, and publicized, satellite imagery, such as the Verification Technology and Information Centre in London and the Federation of American Scientists in Washington. As publicly available imagery improves from five-meter to one-meter resolution, humanitarian groups may find it increasingly useful in dealing with complex humanitarian emergencies and tracking refugee flows.3 In all these cases, nongovernmental groups will find themselves increasingly able to gather and analyze information independently of governments—an important new source of power for civil society.

In short, the emergence of widespread civilian and commercial remote sensing satellites promises both benefits and costs, and thus raises a host of pressing policy questions. Who is regulating them, who should be, and how? Does the new transparency portend an age of peace and stability, or does it create new vulnerabilities that will make the world more rather than less unstable and violent? When should satellite imagery be treated as a public good, to be provided (or controlled) by governments, and when is it a private good to be created by profit-seekers and sold to the highest bidder? Who gets to decide? Is it possible to reconcile the public value of the free flow of information for such pressing purposes as humanitarian relief, environmental protection, and crisis management with the needs of the industry to make a profit by selling that information?

THE POLICY HISTORY

To grapple with these questions, it is important to start with a good understanding of what current policy is, how it came about, and who the players are. Because civilian remote sensing has been so heavily dominated by the United States until quite recently, both in technological and policy terms, much of the story is a U.S. story.

That story begins with the National Aeronautics and Space Administration (NASA), which developed the first civilian remote sensing satellite¾ Landsat. After the launch of Landsat 1 in 1972, the nongovernmental sector got its first glimpse of satellite images of earth. However, Landsat 1’s relatively low resolution¾ 80 meters multispectral¾ was too coarse for most commercial purposes. During these early years, research scientists, educators, and government agencies comprised the principal patrons of Landsat imagery.

In an effort to expand the user base for Landsat data and set the stage for the eventual commercialization of the civilian remote sensing industry, the Carter administration issued Presidential Decision Directive (PDD) 54 in July 1979. The PDD transferred the operation of the Landsat systems to the National Oceanic and Atmospheric Administration (NOAA) in the Department of Commerce and directed NOAA to "seek ways to further private sector opportunities in civil land remote sensing activities … with the goal of eventual operations of these activities by the private sector."4 The administration hoped that under NOAA the management cost of the Landsat program would decline significantly, thus improving the prospects for its eventual commercialization. Moreover, it was believed that future availability of 30-meter resolution imagery from Landsats 4 and 5¾ planned to be launched within the next five years¾ would dramatically increase the revenue earned from the sale of Landsat imagery.5 These two factors combined, it was reasoned, would ultimately lead to the creation of an affordable and robust market for remote sensing data, which would encourage the growth of a commercial satellite industry that could develop and operate remote sensing systems for government and private markets.6

Following the election of President Ronald Reagan, the Carter administration's gradual approach to commercialization of the Landsat program was abandoned in favor of a far more accelerated plan. In an attempt to cut federal spending by privatizing government programs, the Reagan administration ignored evidence that suggested that the remote sensing market was not sufficiently mature to sustain an independent commercial remote sensing industry. That evidence included four feasibility studies commissioned by the U.S. government between 1982 and 1983. The first study was undertaken by the Civil Operational Remote Sensing Satellite Advisory Committee (CORSSAC) of the Department of Commerce. The study found that the commercial market for Landsat data was seriously underdeveloped and, therefore, recommended that "commercialization of the Landsat program should be done gradually."7

The other three studies supported the CORSSAC conclusions. Following an in-depth analysis of the satellite imagery market, ECON Incorporated concluded that "full transfer of the civil land remote sensing system to the private sector, with the expectation of a viable self-sustaining enterprise, is premature."8 Similarly, Earth Satellite Corporation declared that "No option was found that would permit the [Landsat] program to be commercialized, today or in the near future, without substantial subsidies or government-guaranteed data purchases."9 Using a much harsher tone, the National Academy of Public Administration argued that the Reagan administration’s decision to transfer operations of the Landsat program to the private sector "fails to meet sensible criteria of preservation of the national security," and represents a "forced premature privatization of these responsibilities."10

Despite these cautionary words, the Reagan administration pressed ahead. Faced with the likelihood that funding for the Landsat program would soon be discontinued, Congress quickly approved the Land Remote Sensing Commercialization Act (P.L. 98-365), which was signed into law on July 17, 1984. The Act directed the Secretary of Commerce to select a contractor to operate the Landsat system; instructed system operators to market the resulting data on a nondiscriminatory basis; required the Department of Commerce to maintain an archive of land remote sensing data for historical, scientific, and technical purposes; and established a licensing and oversight process for the anticipated private remote sensing industry.

Despite the 1984 Land Remote Sensing Commercialization Act, commercial interest in the remote sensing industry remained limited. In an attempt to jumpstart private initiatives in the remote sensing sector, the U.S. government tried to turn the Landsat program into a moneymaking private venture. Thus, in September 1985, NOAA selected Earth Observation Satellite Company (EOSAT), a joint venture of RCA Corporation and Hughes Aircraft Company, to operate the Landsat satellites and market the resulting data for a period of ten years. According to the terms of the contract, the U.S. government would continue to cover the operational costs of the Landsat program through the three-year expected lifetime of Landsats 4 and 5. In addition, because the market for remote sensing data was considered underdeveloped, the U.S. government agreed to subsidize EOSAT in the amount of $295 million over a five year period to develop and launch two new Landsat spacecraft¾ Landsats 6 and 7. Upon the launch of Landsat 6, EOSAT would assume full responsibility for all operational costs of the Landsat program. Policy makers and EOSAT executives believed—or hoped—that during the lifetime of Landsats 6 and 7, EOSAT's revenues would grow sufficiently to allow the company to finance the development, launch, and operation of future land remote sensing systems.11

Responsibility for the operation of the Landsat program was transferred to EOSAT in October 1985. But the government failed to keep its end of the bargain. The Reagan administration deleted additional subsidies to EOSAT in its FY 1987 budget proposal, arguing that the remaining $125 million called for in the EOSAT-NOAA contract would have to come from EOSAT or other private sources. As a result, EOSAT began laying off employees in December 1986 and ceased all marketing and spacecraft development activities. The Congress again intervened in order to ensure the survival of the Landsat program beyond Landsat 5. In FY 1987, Congress appropriated $62.5 million to continue the development of Landsat 6; however, the funds could not be released until NOAA drafted and Congress approved a new Landsat commercialization plan.

In June 1987, NOAA submitted its new commercialization proposal to Congress. It called for only one additional Landsat satellite to be developed by EOSAT as well as a feasibility study for a second satellite that might include more advanced or commercially oriented sensors. Congress initially rejected this proposal, but acquiesced in October 1987. Between November 1987 and April 1988, NOAA renegotiated its contract with EOSAT. The revised contract directed EOSAT to develop Landsat 6 and all associated ground systems. The government agreed to subsidize the project up to $220 million; any additional costs would have to be absorbed by EOSAT. Under an innovative "payback" arrangement, EOSAT also agreed to refund $10.8 million to the government over a period of approximately four years. In addition, EOSAT waived all rights to data from follow-on civil remote sensing spacecraft beyond Landsat 6. With this agreement in place, Congress finally released the $62.5 million appropriated for the development of Landsat 6 in the FY 1987 budget.

As soon as the Landsat 6 development dilemma was resolved, the Landsat program faced yet another crisis. The 1984 Land Remote Sensing Commercialization Act had instructed NOAA to finance the operations of Landsats 4 and 5 through the expected lifetime of the two satellites, which was due to expire at the end of 1987. Consequently, NOAA had no plans to fund either satellite after the expiration date, regardless of whether the satellites were still operational. Thus, NOAA did not request any funds for the Landsat program in its FY1989 budget. Congress quickly appropriated $9.4 million to fund Landsats 4 and 5 for the first half of the fiscal year and asked NOAA to secure funding for the second half. Unable to obtain sufficient funds, NOAA directed EOSAT to turn the satellites off in April 1989. This proposal drew strong protests from the Congress, foreign governments, and data users in the United States and around the world. In response to the outcry, the National Space Council, chaired by Vice President Dan Quayle, drafted an interim funding plan which asked government agencies that used Landsat imagery to provide money to NOAA. Moreover, the National Space Council recommended that the federal government ensure continuity of Landsats 4 and 5 operations so long as these satellites were the only source of civilian remote sensing data.12 As a result, NOAA rescinded the shutdown order in March 1989.

The same routine played out during the fiscal years 1990 and 1991. In both years, Congress provided $9.5 million for the first six months of the each year and asked other government agencies which utilized Landsat imagery to provide the remaining funds. Finally, on June 1, 1989, President George Bush "approved funding for continued operations of Landsat satellites 4 and 5 and for the completion and launch of Landsat 6."13 In addition, President Bush "directed the National Space Council and the Office of Management and Budget to review options with the intention of continuing Landsat-type data collections after Landsat 6."14

The Undoing of Landsat Commercialization

By the early 1990s, several factors compelled the U.S. government to review the Land Remote Sensing Commercialization Act of 1984. First, the forced commercialization of the Landsat program had faltered badly. Instead of spending $295 million to acquire and deploy two Landsat satellites, the U.S. government had committed itself to spending $245.7 million to develop and launch only one. Moreover, Landsat commercialization had dramatically increased the price of Landsat imagery. As a result, the number of requests for MSS data dropped from 35,272 in 1984 to only 8,000 in 1990.15 The $4,400 price tag for a single thematic mapper scene deterred many scientists and academics from purchasing Landsat data.

Second, the emergence of foreign competitors to Landsat imagery eliminated the United States’ comfortable position as the sole provider of remote sensing data. France successfully launched SPOT-1 in February 1986 and SPOT-2 in January 1990. Both satellites could provide higher resolution images of earth and had a shorter revisit time than the U.S. Landsat system. Although SPOT lacked Landsats’ broader swath width and more numerous spectral bands, many in the United States still worried that SPOT would dominate the remote sensing market. This concern gained new urgency as SPOT’s sale of remote sensing imagery surpassed that of EOSAT by 1989.16 Many within and outside the U.S. government began to question the wisdom of a policy that allowed foreign countries to surpass the United States in an industry that was pioneered by the United States

Third, Landsat imagery proved remarkably useful during the planning and execution of the 1990-1991 Persian Gulf War. As D. Brian Gordon of the Defense Intelligence Agency testified, "There were significant contributions by Landsat … to the success of Operation Desert Storm."17 According to some estimates, the U.S. Department of Defense spent between $5 to $6 million on Landsat imagery during the Gulf War.18 Throughout the conflict, Landsat imagery was used by allied forces for terrain analysis, operational planning, and concealment detection.19

All these pressures led the U.S. Congress to pass the Land Remote Sensing Policy Act (P.L. 102-555), which was signed into law by President Bill Clinton on October 28, 1992. The Act recognized that "the continuous collection and utilization of land remote sensing data from space are of major benefit in studying and understanding human impacts on the global environment, in managing the Earth’s natural resources, in carrying out national security functions, and in planing and conducting many other activities of scientific, economic, and social importance."20 The Act further acknowledged that "despite the success and importance of the Landsat system, funding and organizational uncertainties over the past several years have placed its future in doubt and have jeopardized United States leadership in land remote sensing."21 Therefore, the Act rejected full commercialization of the Landsat program "within the foreseeable future"22 and transferred control of the Landsat system to the National Aeronautic and Space Administration (NASA) and the Department of Defense (DOD). Authority for licensing private remote sensing satellites was left with the Secretary of Commerce, as it had been under the 1984 Act, with advisory roles given to the Secretaries of State and Defense.

Shortly after the passage of the Land Remote Sensing Policy Act, disagreements arose between NASA and the DOD over what type of sensors should be placed on board the Landsat 7 satellite. NASA favored the cheaper Enhanced Thematic Mapper (ETM) Plus capable of acquiring 15-meter images of earth. DOD, on the other hand, pushed for a new sensor, the High Resolution Multispectral Stereo Imager (HRMSI), which could collect 5-meter-resolution data of particular interest to the Department of Defense. The disagreement had not yet been resolved when disaster struck.

At approximately 11:08 a.m. (PST) on October 5, 1993, nearly 13 minutes after the launch of Landsat 6, a ruptured hydrazine manifold prevented fuel from reaching the satellite engine and the $256.5 million spacecraft plunged into the Pacific Ocean.23 Following the failure of Landsat 6 to reach orbit, NASA concluded that the high cost of developing the HRMSI sensor could undermine the timely development and deployment of Landsat 7. Thus, in September 1993, NASA officially rejected placement of the HRMSI sensor on board Landsat 7. Consequently, the Department of Defense pulled out of the Landsat program.

With the Department of Defense out, the question of which agency or agencies should develop and operate the Landsat system and market the resulting data once again came to the forefront. This issue was finally resolved by the White House in a Presidential Decision Directive (PDD-3) that was announced on May 10, 1994. According to the PDD, NASA would be responsible for developing and launching Landsat 7, NOAA would operate the spacecraft and all relevant ground systems, and the Department of Interior would archive and distribute the data at the marginal cost of reproduction.

As this paper is being written, however, yet another rearrangement of responsibilities for the Landsat program is underway. A new PDD has been drafted that will transfer all operational responsibility of the Landsat 7 program from NOAA to the Department of Interior’s U.S. Geological Survey (USGS). Given that USGS has a more obvious interest in pursuing land remote sensing activities than NOAA, the transfer is expected to be approved in the very near future.

With the successful launch of Landsat 7 on April 15, 1999, the United States came full circle, restoring a significant government subsidy to civilian remote sensing, although now with a substantially improved satellite. Landsat 7 will provide 15-meter panchromatic and 30-meter multispectral images of earth for the next five years. It will have the ability to re-image areas of interest every 16 days with its more accurate ETM Plus sensor. More importantly, Landsat 7 images will be made available to all consumers of satellite imagery "at the cost of fulfilling user requests."24 At present, these prices are set at $475 per scene for minimally processed data and at $600 per scene for radiometrically and geometrically corrected data¾ over fifty percent cheaper than any other comparable commercially available civilian or private satellite imagery.25 It is hoped that Landsat 7, launched nearly 30 years after the deployment of the first Landsat spacecraft, will once again place the United States in the forefront of the international civilian remote sensing industry.

The Emergence of a U.S. Commercial Remote Sensing Industry

Although the lega