The IRAS / GGSE 4 Close Approach

At LeoLabs, we track nearly 14,000 objects in Low Earth Orbit (LEO) and continuously screen for potential collision events 24/7. While we always have notable events of interest to study, the close approach this week of the IRAS and GGSE 4 satellites turned out to be one of the closest calls we have seen to date, and one with high potential impact due to the significant sizes of the defunct satellites. In this post we will summarize the details of this event and the data we collected for it.

Event Details

It’s immediately worth asking — for close approaches like this, exactly how close is “close”? What’s normal, and what’s considered high risk? While there is no right or wrong answer here, a general metric used by many satellite operators is that a close approach with a miss distance of <1km is worth paying attention to, and events with <100 meters miss distance are of very high concern and usually warrant performing a collision avoidance maneuver, if possible.

For the IRAS / GGSE 4 event, the LeoLabs system first recognized a conjunction event on January 23, 2020 at 05:58:18 UTC, six days and 18 hours before the Time of Closest Approach (TCA). Our first reported metrics told us these two objects would pass within 63 meters of one another on January 29th. Our margin of error on this computed miss distance was +/- 86 meters RMS.

We continued to monitor this event and over the next few days, watched as subsequent updates showed computed miss distances go as low as seven meters. (As there is always a margin of uncertainty in these measurements, some fluctuation from one update to the next within the uncertainty bounds is normal and expected.)

Extremely low miss distances such as this warrant high attention. It is important to note that computed miss distances are point-to-point (center-of-mass to center-of-mass of both objects), and do not take into account the objects’ sizes. Thus, a computed miss distance of seven meters implies that the actual separation of the closest edges of each object to one another at TCA may be lower than this. (Probability of collision calculations do, however, take into account approximate sizes of the objects.)

With multiple report updates showing miss distances of <20m, we also noted the size of IRAS which has a mass of over one ton and dimensions of 3.6m x 3.24m x 2.05m. LeoLabs’ measured Radar Cross-Section (RCS) of 5.3 m² agrees with that inferred size from its physical dimensions. Translation: it’s a really big satellite.

Our RCS for GGSE 4 is 1.3 m²; which is common for medium-sized payloads. However, we subsequently learned that this satellite may have 18 meter-long booms extending from it, likely in the nadir or zenith directions.

This news, combined with the already-low reported miss distance, indicated a serious threat of a collision as a real possibility. On January 27, we first shared news and details of the event on Twitter:

Public interest in the event very quickly garnered, and over the next 24 hours this became a leading story on many news outlets.

One interesting element to this story was where over Earth this event took place. Conjunctions are most common to occur near the North or South poles, where polar orbiting satellites tend to intersect very often. Most other events are scattered around the world, with many occurring over oceans or uninhabited areas.

LeoLabs’ visualization of detected close approaches worldwide over a 24-hour period

In contrast, this prominent event had a TCA occurring not only over the United States, but over a major US city (Pittsburgh), and in the dusk hours where visibility of this event was actually a possibility if skies were clear.

Location of event TCA occurring over Pittsburgh, PA

We were excited to see all the traction and interest this generated, not just from those near Pittsburgh, but from all around the Northeastern US! A few keen-eyed observers (and skilled photographers) actually captured the event on camera, including Twitter user @Greg_NJ:

Post-Event Analysis

Our team of data scientists manually reviewed the data as soon as it came in, checking to ensure our radar signature returns for each satellite showed what we expected — only one detected object, and not many (as would be the case if new debris were present).

This plots shows our radar return signature for IRAS, in one of the first radar passes following the event:

LeoLabs radar signature returns for IRAS after the event

The x-axis shows elapsed time (in seconds), y-axis shows range from the radar to the satellite (in km), and the color map indicates signal-to-noise ratio returns from the radar. The black dot shows the expected approximate location of IRAS as predicted by our system, while the lighter pixels just to the right of it show the actual detection of the satellite. This and additional data products were used to confirm that only one object was detected, precisely where it was expected to be, giving strong indication that no new debris was generated. This same analysis was performed on GGSE 4 radar observations following the event.

So what was our final assessment of this event? Our post-TCA analysis showed a predicted 18m (+/-47m) miss distance. Taking into account the objects’ unique sizes and dimensions, it is reasonable to estimate that these satellites may have come within just a few meters of actually hitting each other. While a full head-on collision was less likely, there was a very real chance that one of the GGSE 4 extended booms could have hit IRAS as it passed below it at a relative velocity of 14.7 km/s (32,900 mph). This would have created many new pieces of debris that would then pose a danger to other operational satellites in similar orbits. Thankfully, there is no evidence that this occurred.

All total, we computed 69 collision risk updates in the form of Conjunction Data Messages (CDMs) over the eight-day period that we tracked this event. We saw highly consistent data (within expected error bounds) as both objects were reliably tracked by our radars several times per day, allowing us to generate updates on the event every 2.8 hours on average.


Though this was a unique and somewhat harrowing event, the message we want to convey to our readers is that this was not an isolated incident.

In fact, our system detects similar high-risk events on a regular basis. Just this week alone, while all eyes were on IRAS and GGSE 4, other close approaches occurring around the same time include:

COSMOS 1365 FUEL CORE (13594) vs. COSMOS 970 DEB (10576)

JAS 1B (FUJI 2) (20480) vs. SL-8 R/B (4784)

THOR ABLESTAR DEB (120) vs. DMSP 5D-2 F13 DEB (23535)

SL-8 R/B (11870) vs. SL-8 R/B (12792)

Each of the above examples involve very near misses between two objects that are non-operational (defunct satellites, spent rocket bodies, or pieces of debris). They cannot maneuver out of the way of one another, and they pose a very real threat to the LEO space environment and the health of vital operational satellites.

Moral of the story: these types of near-miss events happen all the time, and the vast majority of them have historically gone unnoticed up until now, despite how close each one may be to becoming the next big collision.

Events like these highlight the need for responsible de-orbiting of satellites upon their mission completion, as well as the need for active debris removal (ADR) technologies to take large derelict objects like those listed above and bring them safely down to a low altitude where they can re-enter Earth’s atmosphere and no longer pose a threat to operational satellites.

Written by Matthew Shouppe, LeoLabs Director of Commercial Space

Tracking space debris in Low Earth Orbit.