Table of Contents

Military communication and early-warning satellites in high-altitude orbits play critical roles in enabling nuclear operations—so much so, in fact, that they might be attacked as a prelude to a nuclear strike. However, threats to space-based nuclear C3I capabilities could also arise unintentionally. States periodically reposition their satellites to optimize their performance. If repositioning brought a satellite into proximity with one involved in nuclear operations, it could be misconstrued as preparation for an attack against the latter—especially in a crisis or conflict. To make matters worse, many—perhaps all—satellites involved in nuclear operations are dual-use. As a result, in a conventional conflict, they might be attacked in an attempt to disrupt nonnuclear operations being conducted by their possessor. Such attacks, however, would have the effect of degrading the target state’s nuclear C3I system.

Inadvertent threats to, and attacks on, space-based nuclear C3I capabilities would not be preparations for a nuclear war, but they could risk being interpreted as such—potentially sparking catastrophic escalation.1 In fact, the United States has threatened to resort to nuclear use should its nuclear C3I system come under attack.2 China and Russia are probably less reliant on satellites than the United States for nuclear C3I. Even so, attacks by the United States, or even perceived preparations for them, against any Chinese or Russian satellites involved in nuclear operations would still be very provocative—especially if the target were Russia’s early-warning satellites, given its launch-under-attack posture.3

The American and Russian nuclear C3I systems, and perhaps the Chinese system too, use satellites in two different kinds of high-altitude orbits: geostationary and Molniya. Geostationary satellites remain above a fixed point on the Earth’s equator at an altitude of roughly 36,000 kilometers (22,000 miles). The United States uses this orbit for communication satellites involved in nuclear operations (all of which are dual-use).4 An object in a Molniya orbit (a type of highly elliptical orbit) hangs above the Northern Hemisphere at altitudes approaching 40,000 kilometers (25,000 miles) before it quickly traverses the Southern Hemisphere at much lower altitudes. Russia’s early-warning satellites are located in such orbits.5 Its Unified Satellite Communication System (which is likely used for both nuclear and conventional operations) and the United States’ space-based early-warning system (which is definitely dual-use) comprise satellites in both geostationary and Molniya orbits.6 Less is known about the Chinese nuclear C3I system. Various Chinese military communication satellites and at least one possible early-warning satellite operate in geostationary orbit—though it is not known for sure whether any are involved in nuclear operations.7

Inadvertent threats to, and attacks on, space-based nuclear C3I capabilities would not be preparations for a nuclear war, but they could risk being interpreted as such—potentially sparking catastrophic escalation.

To varying degrees, China, Russia, and the United States have developed, tested, and deployed weapons that are designed, or could be used, to attack satellites.8 The difficulty of attacking satellites increases with altitude. As a result, two of the technologies that pose acute threats to satellites in low-altitude orbits would likely be much less effective against objects in geostationary or Molniya orbits. For the foreseeable future, ground-based directed-energy weapons, which focus energy on a target with, for example, a laser, will simply not be powerful enough to threaten satellites at high altitudes by damaging their sensors or other components.9 Ground-based direct-ascent missiles designed to destroy satellites kinetically could prove somewhat more effective, but because their launches could be detected by a state with space-based early-warning sensors and they require hours to reach high altitudes, potential targets might have time to maneuver and thus evade an incoming interceptor.

In contrast to ground-based weapons, space-based weapons present a significant threat against high-altitude satellites today. So-called co-orbital weapons (sometimes called space mines) would be launched well in advance of any attack and dwell in an orbit that would enable them to reach potential targets relatively quickly. If activated by their possessor, they could then attack other satellites—either by colliding with them or by using a kinetic or nonkinetic standoff weapon. Such attacks could occur more quickly and would be more difficult to detect than operations involving ground-based direct-ascent weapons. Many objects could be used as co-orbital weapons, including some satellites that were not designed for that purpose.10 

No destructive anti-satellite testing has been undertaken against satellites in high-altitude orbits. However, in the last decade, a number of geostationary satellites have been closely approached by others—including by Chinese, Russian, and U.S. satellites.11 These operations may have been explicit demonstrations or tests of an anti-satellite capability, but even if they were not, they demonstrate that such a capability is an inherent consequence of a satellite’s possessing a high degree of orbital maneuverability. The U.S. Defense Intelligence Agency assesses, for example, that “China is developing sophisticated on-orbit capabilities, such as satellite inspection and repair, at least some of which could also function as a weapon.”12

Reducing the threat posed by co-orbital weapons to satellites in geostationary and Molniya orbits would mitigate the danger of inadvertent nuclear escalation. Unfortunately, efforts to establish effective arms control in space—or even to build consensus on what constitutes unacceptable behavior—have largely stalled in multilateral fora.13 A trilateral approach focused on high-altitude orbits may be a fruitful way forward. It would bypass the complexities of multilateral diplomacy and focus on preventing the serious and shared danger of nuclear war.

Solution Concept

Establishing keep-out zones around high-altitude satellites could help reduce the vulnerability of key nuclear C3I capabilities. Specifically, China, Russia, and the United States should commit not to maneuver their satellites within an agreed minimum distance—700 kilometers (430 miles) in any direction—of another participant’s high-altitude satellites (with the exception of repositioning maneuvers conducted one at a time and declared in advance). This agreement would apply only to satellites nationally owned by China, Russia, and the United States and not to privately owned satellites or to satellites owned by other states (so would not contravene the 1967 Outer Space Treaty’s prohibition on “national appropriation”).

Currently, the regulation of high-altitude satellite orbits is minimal. The International Telecommunication Union (ITU), a United Nations agency, allocates slots to geostationary broadcast and communication satellites in order to prevent interference—though these slots can overlap if satellites operate on different frequencies or broadcast to non-contiguous regions on the ground. Participation in the ITU is voluntary and is designed only to minimize broadcast interference.

Establishing keep-out zones would go further than the ITU rules by applying to all Chinese, Russian, and U.S. satellites in both geostationary and Molniya orbits—not just geostationary satellites broadcasting at a particular frequency band—without permitting any overlap. It would begin to establish rules of the road for good behavior in space and help break the deadlock in improving space governance. Even recognizing that keep-out zones could not physically prevent one participant state from attacking another’s satellites in conflict—although the proposed agreement would still apply then—they would still help to reduce escalation risks in three ways.

First, keep-out zones would mitigate the danger that repositioning operations could lead one state to wrongly conclude that one or more of its satellites were under attack—that is, the zones would help to define the difference between innocuous and aggressive actions in space. Even (or perhaps especially) in a conflict, a state that did not intend to attack a nuclear C3I satellite belonging to its adversary would have a clear incentive to abide by rules designed to prevent such threats from arising inadvertently.

Second, even if one participant decided to attack another’s satellites—for whatever reason—keep-out zones could buy time. An attacking satellite would typically have to close in on a target before launching an attack (how close it would need to come would depend on its capabilities).14 This process would not be instantaneous. If the target state detected a violation of its keep-out zones before the attacking satellites were able to execute the attack, it could take preventative action (by, for example, maneuvering its satellites away from the attacking ones). Increasing the warning time of an intentional attack would also reduce the likelihood of escalation resulting from time pressure.

The margin of warning afforded by keep-out zones would depend, in part, on their size. Fuel-efficient maneuvers in geostationary orbit to cross from the edge to the center of a 700-kilometer keep-out zone would require about one day (see appendix B for more details). Faster crossing would be possible by using larger amounts of fuel. For example, the same keep-out zone could be crossed in six hours by expending the same amount of fuel that a communication satellite typically uses each year for station keeping (that is, making minor adjustments so the satellite remains in its correct orbit during day-to-day operations). Larger keep-out zones would buy more warning time and further complicate attacks—but they would be more disruptive to satellite operations. The keep-out distance of 700 kilometers proposed here aims to strike a balance between increasing warning and reducing disruption.

Third, each state could use negotiations to underscore to the others the dangers of attacking its high-altitude satellites. Such messaging could reduce the likelihood of one participant’s deliberately attacking another’s dual-use satellites in an effort to win (or at least not lose) a conventional war because it had underestimated the consequent risk of nuclear escalation.

Keep-Out Zones for Satellites in Geostationary and Molniya Orbits

China, Russia, and the United States should make a joint political commitment that each will maintain a minimum separation between its satellites and the satellites in geostationary or Molniya orbits that belong to, and have been declared by, other participants of the agreement.

Specifically, they should agree not to maneuver any satellite into the keep-out zone of another participant’s satellite:

  • The keep-out zone of a satellite in geostationary or Molniya orbit should be a sphere with a radius of 700 kilometers.
  • If two satellites belonging to different participants have established different Molniya orbits but are expected to pass within 700 kilometers of one another, neither participant should be required to alter the orbit of its satellite, but each should notify the other of the conjunction at least twenty-four hours before the distance between them is due to become smaller than 700 kilometers.

Repositioning maneuvers that bring one satellite into the keep-out zone of a satellite belonging to another participant should be permitted only if the participant conducting the maneuver takes the following steps:

  • Notifies the other participant at least twenty-four hours before the satellite being repositioned reaches the edge of the other satellite’s keep-out zone.
  • Maintains a minimum distance of 250 kilometers (160 miles) between the two satellites at all times.15
  • Minimizes, to the extent possible, the time the satellite being repositioned spends in the keep-out zone of the other satellite.
  • Brings no more than one satellite at a time into any of the keep-out zones of the satellites belonging to each of the other participants.

The participants should annually exchange confidential lists of state-owned satellites in geostationary and Molniya orbits. Only satellites included on this list should be entitled to keep-out zones.

  • For each satellite registered with the United Nations Register of Objects Launched into Outer Space, its owner should provide its international designator.
  • For any other satellite, its owner should provide a designator and its basic orbital parameters (in accordance with the provisions of the United Nations Register of Objects Launched into Outer Space).

By notifying the other participants, each participant may register on or remove from the list a satellite at any time.

The participants should hold an annual meeting to discuss any compliance or implementation issues. They should also commit to discussing urgent compliance concerns through regular diplomatic channels.

Implementing this proposal in geostationary orbit—where one satellite would essentially have to tail another to approach it—would be relatively straightforward. Two satellites in quite different Molniya orbits, however, may occasionally happen to pass close to one another, creating a complication. Because each participant is equally responsible for such a conjunction, there would be no obvious way of deciding which participant should be required to take action to avoid it. For this reason, there should be no requirement to do so, though each participant should be required to inform the other of the impending approach. In any case, conducting an attack during one of these rare conjunctions would be technically difficult since the relative velocities of the satellites involved can be very large.

To ease implementation, only declared satellites should be afforded the protection of a keep-out zone. But because all the participants’ nationally owned satellites, declared or not, must respect keep-out zones, controversary could arise about the ownership of a given satellite. States should commit to discussing such concerns at annual implementation meetings or, if urgent, through diplomatic channels.

Today, a small number of Chinese, Russian, and U.S. geostationary satellites are typically located within 700 kilometers of each other. To establish keep-out zones, some of these satellites would need repositioning—but this should not be onerous and would not meaningfully impact the participants’ capabilities. Table 2 shows the number of geostationary satellites that would need to be moved as of January 1, 2021, depending on the size of the keep-out ones. Zones with a radius of 700 kilometers (just under 1 degree) would only require seven satellites to be moved. Regardless, repositioning is a routine operation, and only minor orbital adjustments would be needed (though the participants would have to negotiate, on the basis of reciprocity, who would move which satellites).

Table 2: Number of Geostationary Satellites That Would Need Repositioning to Implement Keep-Out Zones
United States–Russia Russia-China China–United States Total
0.5 2 1 2 5
1.0 3 1 3 7
1.5 7 1 5 13
2.0 7 2 6 15

Source: Author calculations based on satellite data as of January 1, 2021, courtesy of Union of Concerned Scientists, “UCS Satellite Database,” January 1, 2021,


Participants would verify compliance with this proposal by using their space situational awareness capabilities to periodically measure the positions of satellites in geostationary and Molniya orbits. This data—longitude, latitude, and altitude—enables not only the distance between satellites to be computed but also whether one satellite is drifting in the direction of another. A state would likely want to monitor its own satellites and those belonging to other participants. Failing to detect another participant’s satellite at its expected location would be evidence that it was maneuvering and could cue a search.

Participants would presumably want to be able to observe satellites in geostationary and Molniya orbits frequently enough that, in the time between observations, a co-orbital weapon could move only, say, one-half or one-third of the distance from the edge of a keep-out zone to the satellite at its center. There can be no definitive estimate of this time period since it would depend on how much fuel is available to the attacking satellite and, of that, how much the attacker would be willing to expend. Nonetheless, ideally, each participant would probably want to image any satellites near its own every two or three hours.

There are various means to detect the positions of satellites in high-altitude orbits.16 In each case, multiple sensors are required to monitor all relevant satellites. For ground-based systems, these sensors must be spread around the Earth.

  • Ground-based optical sensors—telescopes—can detect satellites by observing reflected light from the sun or from a laser used to illuminate the target. They are the simplest, most inexpensive, and most readily available means to detect satellites, though they suffer from various limitations. Most notably, they cannot operate in daylight or when skies are overcast. However, commercially available infrared telescopes have proven capable of tracking geostationary satellites during the day, thus addressing a key weakness of existing optical telescopes.17
  • Space-based optical sensors are capable of high-cadence imaging of other satellites, though they are considerably more expensive than ground-based telescopes. Their performance depends on the number of available sensors and their orbits.
  • Ground-based radio telescopes can be used to locate satellites by intercepting their communications. Such observations are not affected by weather or time of day but are only possible when a satellite is transmitting.
  • Some ground-based radars are capable of detecting and imaging high-altitude satellites. They are not affected by weather or time of day, but they are expensive and have a particularly limited field of view. They are therefore poorly suited to wide-area searches, but they can accurately measure the position of satellites whose approximate locations are known.

China, Russia, and the United States have long had an interest in developing effective space situational awareness capabilities. These capabilities are shrouded in opacity, particularly in the cases of China and Russia—though some significant inequalities are apparent. (Appendix C summarizes what is publicly known about each state’s capabilities.) For its part, the United States is likely already capable of verifying the proposed agreement—at least for most of geostationary orbit (it may lack radar coverage over parts of the Eastern Hemisphere). The opacity of Russian and Chinese capabilities makes it difficult to determine whether they could also verify the agreement. Russia has a fairly extensive space situational awareness system, though it likely has less capability than the United States. Chinese capabilities, meanwhile, are more opaque still and likely less sophisticated than Russia’s.


Technical feasibility. The primary technical challenge to verification is uncertainty about the adequacy of existing space situational awareness capabilities—China’s and Russia’s, in particular. (Making minor adjustments to the positions of a small number of satellites to establish keep-out zones should be straightforward, though deciding who would move which satellite could become somewhat contentious.)

Even if not all the participants have adequate verification capabilities today, however, they are likely on a trajectory to acquire them. Accurately tracking the location of highly valuable satellites—and potential threats to them—is clearly in the national interests of China, Russia, and the United States, regardless of whether they try to mitigate those threats cooperatively. Moreover, acquiring effective space situational awareness capabilities is critical to the development of anti-satellite weapons, which all three states appear to want. At the same time, cheaper ways of monitoring high-altitude satellites, such as infrared telescopes, are emerging. Any participants currently lacking the required verification capabilities might still be able to detect noncompliance, albeit without high confidence.

Political feasibility. One key political challenge is that China and Russia appear to want the ability to hold U.S. satellites in high-altitude orbits at risk, while the United States may desire a similar capability against China and Russia. These incentives are probably asymmetric right now as the U.S. military relies more heavily on satellites, which could present tempting targets for China and Russia during a conflict.

But the incentives may be evening out. China and Russia are investing heavily in military satellites, including in high-altitude orbits. Notably, Russia is rebuilding its space-based early-warning system and has offered assistance to China, which is developing one for the first time.18 Russia relies on a launch-under-attack policy to help ensure the survivability of its nuclear forces, while China appears to be moving in the same direction. While both states also have a network of ground-based early-warning radars, they should seek (as the United States does) to have confirmation of an incoming attack from two physically independent detection systems before deciding to launch their nuclear forces.

Furthermore, while attacks on high-altitude satellites could provide military benefits, they could also create serious risks. In fact, because they could undermine nuclear C3I capabilities, they would be even more escalatory than attacks on satellites orbiting at lower altitudes (which would hardly be risk-free). As a result, an agreement that focuses narrowly on enhancing the survivability of satellites in geostationary and Molniya orbits—and thus reduces the shared risk of nuclear war—may be of interest to Beijing, Moscow, and Washington.


1 Acton, “Escalation Through Entanglement.”

2 U.S. Department of Defense, “Nuclear Posture Review,” 21.

3 Arbatov, Dvorkin, and Topychkanov, “Entanglement as a New Security Threat,” 38–39.

4 “Advanced Extremely High Frequency System,” Air Force Space Command, March 22, 2017,

5 Future satellites may be placed in geostationary orbit. Pavel Podvig, “Fourth Tundra Early-Warning Satellite Is in Orbit,” Russian Forces (blog), May 22, 2020,

6 V.A. Grigoryev and I.A. Khvorov, “Military Satellite Communications Systems: Current State and Development Prospects,” Military Thought 16, nos. 3–4 (July 1, 2007): 149–150; Jana Honkova, “The Russian Federation’s Approach to Military Space and Its Military Space Capabilities,” George C. Marshall Institute, November 2013, 26–28; and Office of the Secretary of Defense, “Report to the Defense and Intelligence Committees of the Congress of the United States on the Status of the Space Based Infrared System Program,” U.S. Department of Defense, March 2005, 3,

7 Stephen Clark, “China Launches Military Satellite Toward Geostationary Orbit,” Spaceflight Now, February 7, 2021,

8 Brian Weeden and Victoria Samson, eds., “Global Counterspace Capabilities: An Open Source Assessment,” Secure World Foundation, April 2021, xiv–xxiii,

9 Satellites in Molniya orbits spend a portion of their orbit at low altitudes near the South Pole where they may be vulnerable to direct ascent and directed-energy weapons—though their high speeds during this part of their trajectory helps mitigate this vulnerability.

10 David Wright, Laura Grego, and Lisbeth Gronlund, “The Physics of Space Security: A Reference Manual,” American Academy of Arts and Sciences, 2005, 136,

11 Weeden and Samson, eds., “Global Counterspace Capabilities,” 1-7–1-11, 2-10–2-13, 3-6 –3-10.

12 Defense Intelligence Agency, “Challenges to Security in Space,” [February 2019], 21,

13 Michael J. Listner, “The International Code of Conduct: Comments on Changes in the Latest Draft and Post-Mortem Thoughts,” Space Review, October 26, 2015,

14 Some co-orbital weapons may need to physically rendezvous with a target to damage it. Others may be armed with standoff munitions weapons that could attack a target from a distance. However, maneuvering to allow the use of a standoff munition with a relatively short range, say 50 kilometers (31 miles), would not require significantly less time than manuevering to reach the satellite itself. Long-range standoff weapons would be more technically difficult to develop and deploy.

15 Maintaining such a minimum distance is possible because repositioning maneuvers involve a change in altitude. A minimum distance of 250 kilometers corresponds to a slow drift rate of about 1 degree per day. Faster repositioning maneuvers would increase the distance of closest approach.

16 George Veis, “Optical Tracking of Artificial Satellites,” Space Science Reviews 2, no. 2 (August 1, 1963): 250–296; Grant H. Stokes et al., “The Space-Based Visible Program,” Lincoln Laboratory Journal 11, no. 2 (1998): 207–236; Chuck Livingstone, “Radar Systems for Monitoring Objects in Geosynchronous Orbit,” Technical Report TR 2013-009, Defence Research and Development Canada, June 2013,

17 Jeffrey Shaddix et al. “Daytime GEO Tracking With ‘Aquila’: Approach and Results From a New Ground-Based SWIR Small Telescope System,” 2019 Advanced Maui Optical and Space Surveillance Technologies Conference, Maui, Hawaii, September 17–20, 2019.

18 Christopher Weidacher Hsiung, “Missile Defense and Early Warning Missile Attack System Cooperation: Enhancing the Sino-Russian Defense Partnership,” IFS Insights, Norwegian Institute for Defense Studies, January 1, 2020,