Iridium 33/Cosmos 2251 Collision

Coverage started 2009 March 5

Updated 2012 June 22

In the report issued on 2009 February 10 at 1502 UTC, SOCRATES predicted a close approach of 584 m between Iridium 33 and Cosmos 2251. This was not the top predicted close approach for that report or even the top predicted close approach for any of the Iridium satellites for the coming week. But, at the time of predicted close approach (1656 UTC), Iridium 33 suddenly went silent. The US Space Surveillance Network (SSN) subsequently reported that they were tracking debris clouds in both the Iridium 33 and Cosmos 2251 orbits, confirming a collision.

This is the first time two satellites are known to have collided on orbit. While Cosmos 2251, a Russian communications satellite, is thought to have ceased operations about two years after it was launched in 1993, Iridium 33 was part of the operational Iridium constellation of 66 satellites at the time of the collision.

As of 2012 June 22, the SSN has cataloged 598 pieces of debris (93 pieces of which have already decayed from orbit) associated with Iridium 33 and 1,603 pieces of debris (222 pieces of which have decayed) associated with Cosmos 2251. The materials below are provided to give a sense of the current relationship between the Iridium constellation and the resulting debris clouds.

Figure 1 shows the current Iridium constellation with the orbits for the operational satellites shown in green, the spares shown in blue, and the inactive satellites shown in red. The Iridium 33 debris is shown in light blue and the Cosmos 2251 debris is shown in orange. An STK Viewer 9 file (see bottom of this page for more information on STK Viewer 9) of the event is also available.

Click for larger image

Figure 1. Screen shot from STK Viewer 9 file of current Iridium constellation and collision debris clouds

Download Current Iridium Constellation Scenario

Gabbard plots of the two debris clouds show the relative distribution of the debris in altitude and suggest that more of the Iridium 33 debris went to slightly higher orbits while much of the Cosmos 2251 debris went to lower orbits, with some of it already reentering the Earth's atmosphere.

Click for PDF file

Figure 2. Gabbard Plot of Iridium 33 Debris Cloud [as of 2010 Apr 28]

Click for PDF file

Figure 3. Gabbard Plot of Cosmos 2251 Debris Cloud [as of 2010 Apr 28]

Together with a 3D interactive plot of the relative velocities at the time of the collision (below), these Gabbard plots should help develop a better understanding of the actual collision geometry.

Tracking a Collision

There has been much discussion about why this collision wasn't reported in SOCRATES or why Iridium didn't act on the information provided in SOCRATES. In reality, SOCRATES did predict a close approach between Iridium 33 and Cosmos 2251 at the time of the actual collision in each of the 14 reports in the week leading up to the event. None of these, however, made the Top Ten list. It is instructive to see just what SOCRATES reported to help understand the limitations of the data used by that system and what needs to be done to improve these types of screenings.

Starting with the report on February 4 at 0401 UTC and continuing right up to the report issued on February 10 at 1502 UTC, SOCRATES did predict a close approach between Iridium 33 and Cosmos 2251 at between 16:55:59.670 and 16:55:59.990 UTC. As can be seen in Figure 4, the predicted close approaches ranged from 117 m to 1.812 km over that interval.

Figure 4 shows the predicted minimum range for the closest conjunction in each report, the closest Iridium conjunction, the closest Iridium 33 conjunction, and the conjunction with Cosmos 2251. The variability of the predictions over the week before the collision speaks to the inherent uncertainty in the TLE data used to make the predictions. While Report Number 5 (February 6 at 0402 UTC) did predict a close approach of 117 m, the predicted close approach grew to 1.243 km the very next day.

Click for PDF file

Figure 4. SOCRATES Min Range

In fact, as can be seen in Figure 4, the Cosmos 2251 conjunction is not even the closest predicted conjunction for Iridium 33 in most of the reports, much less for the entire Iridium constellation. Looking at the relative rankings of these conjunctions can help understand the limitations in using conjunction reports based on data as inaccurate as the TLEs.

Figures 5 and 6 show the rankings in each SOCRATES report for the Iridium 33-Cosmos 2251 conjunction in the total report, against all Iridium conjunctions, and for all Iridium 33 conjunctions. Over the 14 reports, the overall ranking ranges from a low of 1,611 in Report Number 3 to a high of 11—again just the very next day. At the time of the collision, it was ranked 152 overall.

Looking at the rankings against all Iridium conjunctions gives a sense of the magnitude of the problem for Iridium—or any satellite operator—in trying to screen with TLEs. The Iridium 33-Cosmos 2251 conjunction ranges from a low ranking of 149 in Report Number 3 to a high ranking of 2 in Report Number 5, with an average ranking of 64. Over this interval, Iridium 33 saw between 10 to 15 times each week when something was predicted to come within 5 km of it—out of the 1,007 to 1,095 such predicted conjunctions for the entire Iridium constellation. Determining that the predicted conjunction for Iridium 33 and Cosmos 2251 was more significant than the many dozens of other Iridium conjunctions for that week is simply not possible using the TLE data.

Click for PDF file

Figure 5. SOCRATES Rank (All)

Click for PDF file

Figure 6. SOCRATES Rank (Top 200)

So, does that mean it isn't possible to provide meaningful conjunction screening for satellites operators? No. SOCRATES was actually established to demonstrate several key points:

  • That it is possible, using off-the-shelf software, standard orbital data, and basic computer hardware to automatically screen large numbers of satellite (over this interval, we screened 2,946 payloads against between 11,761 and 11,848 total objects) and provide access to the reports via a standard web interface.
  • To help people understand that close approaches with satellites in Earth orbit are not uncommon events. Over the period examined here, between 11,169 and 11,493 conjunctions were reported each week—over 1,600 times a day. Even though each conjunction has a low probability of occurrence, over time, a collision with an operational satellite was bound to happen.
  • To illustrate the need for better data to reduce the uncertainty in conjunction predictions, thereby enabling satellite operators to work to protect their assets. Better data is available—directly from satellite operators and from the US and other governments. Our SOCRATES-GEO effort has shown the advantages of operators working together and sharing their best orbital data to reduce risks of collisions, but a more open data sharing policy and wider participation is needed to make this process even better.

Anatomy of a Collision

The following figures illustrate the geometry of the collision of Iridium 33 and Cosmos 2251.

Figure 7 shows the orbital geometry just at the time of the collision (based on the predicted time of closest approach in the last SOCRATES report) at 16:55:59.806 UTC. In this figure, Iridium 33 is moving from the bottom to the top of the image, while Cosmos 2251 is moving from the upper left to the lower right. It can be seen that the objects collided at nearly right angles over northern Russia.

Click for larger image

Figure 7. View of Iridium 33 and Cosmos 2251 orbits at time of collision

Figure 8 shows the initial evolution of the debris cloud just 10 minutes later. This figure uses the first TLEs released by Air Force Space Command as each debris piece was added to the public catalog. Since not all of the debris lines up well at the time of the collision due to limitations of the data (e.g., some of the pieces were just cataloged and are being propagated backward several months), the data was filtered to only show those objects within 100 km of the original parent objects at the time of the collision. That left 209 of the 419 pieces of Iridium 33 debris and 555 of the 998 pieces of Cosmos 2251 debris (764 of the total 1,417 pieces of debris).

Click for larger image

Figure 8. View of Iridium 33 and Cosmos 2251 orbits and debris 10 minutes post-collision

Figure 9 shows the evolution of the debris clouds 180 minutes post-collision, almost two revolutions later. The spread of each debris cloud around its respective orbit is already becoming apparent.

Click for larger image

Figure 9. View of Iridium 33 and Cosmos 2251 orbits and debris 180 minutes post-collision

Of course, the best way to see the evolution of these debris clouds is to view the interactive STK Viewer file of the event, which is provided below.

Download Iridium 33/Cosmos 2251 Collision Scenario

In an attempt to develop a better understanding of the actual collision geometry for this event, a 3D plot was generated of the relative velocities of each piece of debris with respect to its parent object using System Tool Kit (STK). Figure 10 provides a static view of the result.

Examination of the interactive 3D scenario (provided below Figure 10) shows large out-of-plane relative velocity components, the apparent result of coupling of the two masses, despite the hypervelocity nature of the collision. There are also a large number of pieces of Cosmos 2251 debris with significant radial (downward) relative velocities, although it is not apparent why this situation would be the case. It is hoped that the availability of this data set will help researchers with expertise in hypervelocity impacts develop a more complete description of the collision geometry for this event.

Click for larger image

Figure 10. 3D Plot of Debris Relative Velocities

Download Iridium 33/Cosmos 2251 Debris Relative Velocities Scenario

Video captured by Kevin Fetter on 2009 March 12 provides additional insight into the collision geometry for this event. His video shows the main piece still associated with Iridium 33 in the SATCAT (NORAD Catalog Number 24946) followed by the inactive Iridium 28 satellite. The double flashes from Iridium 33 suggest that at least two of the MMAs (Main Mission Antennas, seen at the bottom of the satellite in Figure 11) on that object survived the collision relatively intact.

Click for larger image

Figure 11. View of Iridium satellite showing MMAs (bottom)

The video was taken from Brockville, Ontario, Canada (44.6062 N, 75.6910 W) looking at 8h 45m RA, +04° 56' Dec between 00:54:03 and 00:54:26 UTC. The screen shot seen in Figure 12 verifies the basic conditions of this observation, including that the necessary visibility conditions were satisfied.

Click for larger image

Figure 12. Snapshot of Kevin Fetter's observation conditions as seen in video

Impact on the Space Environment

Loss of the Iridium 33 satellite and its near-term impact on users of the Iridium constellation, while significant, only hints at the long-term implications of this event. A search of SOCRATES on 2009 August 5 shows 154 conjunctions (within 5 km) between the 66 operational and 8 spare Iridium satellites and Iridium 33 debris and another 33 conjunctions between the 30 operational and 6 spare Orbcomm satellites and Iridium 33 debris, over the upcoming 7-day period. Looking at Cosmos 2251 debris, there are a further 334 conjunctions with Iridium satellites and 108 with Orbcomm satellites. And there are many other satellites which operate in this orbital regime seeing similar increases in the number of conjunctions, as well.

Unfortunately, the problem will not be getting better any time soon. Figures 13 and 14 show the distribution of the debris clouds for Iridium 33 and Cosmos 2251, respectively. Each dot represents the apogee and perigee height of each object listed in the public catalog on 2009 August 5. Objects just below the 800-km apogee line (horizontal) represent objects which were knocked into lower orbits as a result of the collision, while those just to the left of the 800-km perigee line (vertical) represent objects which were knocked into higher orbits. As we saw with the Gabbard plots, more of the Iridium 33 debris was knocked into higher orbits while more of the Cosmos 2251 debris was knocked into lower orbits.

Click for larger image

Figure 13. Apogee-Perigee Lifetime Plot of Iridium 33 Debris (average size)
[as of 2009 Aug 28]

Click for larger image

Figure 14. Apogee-Perigee Lifetime Plot of Cosmos 2251 Debris (average size)
[as of 2009 Aug 28]

Overlaid onto the plots of orbital distributions are contours of the expected orbital lifetime values (in years) for objects with the same average physical characteristics. Objects to the left of the first contour should decay within 1 year of 2009 August 5 while those to the right of the last contour are predicted to be in orbit 100 years from now. For the Iridium 33 debris, over half of it is beyond the 100-year contour. For the Cosmos 2251 debris, several dozen pieces (beyond those which have already decayed) should reenter in the next couple of years, but significant numbers will likely remain in orbit 25–50 years from now.

These contours are based on the average characteristics of the objects cataloged to date. To determine those characteristics, the overall dry mass and volume of each original object, Iridium 33 and Cosmos 2251 (as shown in Table 1), was used to compute the average mass and volume of the resulting debris.

Table 1. Pre-Collision Satellite Characteristics

SatelliteNumber of
Total Volume
Dry Mass
Iridium 334193.38855686
Cosmos 22519987.84190074

Interestingly enough, the average volume yields an almost identical size of 12.33 cm radius for the Cosmos 2251 debris and 12.45 cm radius for the Iridium 33 debris (likely a result of the limiting sensitivity of the radar sensors in the US Space Surveillance Network). The average masses are also close, with the average being 1.327 kg for Iridium 33 debris and 0.902 kg for Cosmos 2251 debris (consistent with the original average densities of the two spacecraft).

The orbital lifetime contours were generated using the Lifetime Tool in System Tool Kit (STK). The average area-to-mass ratios resulting from the data in Table 1 (assuming the objects to be spherical) were used together with a grid of orbits, each of which were initialized for the selected apogee and perigee (using the appropriate inclination from Table 1), and assuming the argument of perigee, right ascension of the ascending node, and mean anomaly were all zero (sensitivity analysis of these assumptions showed measurable differences but did not significantly affect the overall results). A coefficient of drag of 2.2 was assumed for all calculations.

As was done in our analysis of the Chinese ASAT test debris, the Jacchia-Roberts atmospheric model was used (again after performing sensitivity analysis on the results) and the latest Schatten predictions from March 2009 through May 2039 were used to estimate future space weather conditions. Orbital lifetime predictions were evaluated looking forward 150 years (which means the Schatten predictions were repeated as needed to span this interval).

For objects smaller than the average value, the orbital lifetime contours shift to the right, and for objects larger than the average value, the contours shift to the left. Figures 15 and 16 show the results assuming debris of the same average density as the original satellite but now using a radius of 5 cm (the approximate limit of the US Space Surveillance Network radar sensors). This is the equivalent of increasing the ballistic coefficient (CDA/m) by a factor of 2.56 (the area-to-mass ratio varies as the inverse of the spherical radius for a constant density), so it could also account for variability in the estimates of the coefficient of drag or density. Figures 15 and 16 likely represent a best case for the decay of the cataloged debris from this collision.

Click for larger image

Figure 15. Apogee-Perigee Lifetime Plot of Iridium 33 Debris (5-cm radius)
[as of 2009 Aug 28]

Click for larger image

Figure 16. Apogee-Perigee Lifetime Plot of Cosmos 2251 Debris (5-cm radius)
[as of 2009 Aug 28]


The collision of Iridium 33 and Cosmos 2251 is a prime example of the consequences of not being good stewards of the near-Earth space environment. Not only was an operational communications satellite lost, but thousands of pieces of debris larger than 1 cm were produced, putting many other communications and earth resources satellites in similar orbits at increased risk for many decades to come.

Space-faring nations must be more proactive toward reducing the amount of material launched into or left in Earth orbit. All satellites will eventually fail and satellite operators simply cannot expect natural forces to remove them in a timely fashion. The international community must continue to work toward establishing and adopting standards for mitigating the generation of space debris.

Until such time as someone can come up with a cost-effective way to actively remove space debris from orbit, the international community must also work more closely together to share orbital information to improve overall space situational awareness, with the goal of mitigating the risk of collision to other operational satellites. While the SOCRATES-GEO and new SOCRATES-LEO services—offered by the Center for Space Standards & Innovation (CSSI) through the GEO and LEO Data Centers—are already supporting 17 satellite operators by screening over 250 active satellites, progress in this area has been far too slow.

The ultimate success of these risk mitigation services depends on the fullest participation of all satellite operators—together with existing space surveillance systems—to provide the best space situational awareness picture possible. After all, why wouldn't we want to share the best orbital information available if we thought it would have helped avert this collision and its consequences? Without this type of cooperation, our chances of avoiding a repeat of this event are limited. And, until then, we will have no idea of whether simply sharing data would have made it obvious that this event was exceptional or if some more ambitious cooperative undertaking is necessary.

Note: Larger versions of all the images provided on this page are available by clicking the images. The interactive STK Viewer files of these scenes are also provided to give you a far better sense of the overall environment by allowing you to zoom in and out and move around the Earth while watching all the satellites moving in their orbits.

Note: STK Viewer is a free product which allows anyone with a Windows computer to view an STK (System Tool Kit) scenario. With it, you can animate a scenario forward or backward, pause the animation, and zoom or pan the view for a more complete understanding of the event. Just like with Adobe Acrobat, where the authoring software requires a license but the Adobe Reader is free, STK can produce STK Viewer files—also known as VDFs—which can then be viewed by anyone with the STK Viewer software. You can find the free STK Viewer 9 on CelesTrak here.