Starlink Manoeuvre Update - July 2023

SpaceX provided an update on its constellation status to the Federal Communications Commission (FCC) on 30 June 2023 (here: https://licensing.fcc.gov/myibfs/download.do?attachment_key=23204338). These semi-annual reports offer an opportunity to analyse the impacts of large numbers of satellites on the safety of the space environment, and I am grateful to SpaceX for its consistent and comprehensive reporting. The recent update, covering the period 1 December 2022 through 31 May 2023, included the number of risk mitigation manoeuvres performed by Starlink satellites for this period: 25,299.

Whilst this is a high number of manoeuvres, it is due (perhaps obviously) to the large number of satellites in the Starlink constellation and in part to the relatively low manoeuvre threshold used by SpaceX. Starlink satellites manoeuvre when the probability of collision is greater than 1-in-100,000 (1E-5), which is a threshold 10 times lower (better) than recommended by NASA (see here: https://nodis3.gsfc.nasa.gov/OCE_docs/OCE_50.pdf).

The evolution of the number of risk mitigation manoeuvres reported by SpaceX since the first semi-annual report in 2021 is shown in Figure 1. The number of manoeuvres reported has approximately doubled from one report to the next, which is typical of exponential growth. Additionally, the number of manoeuvres has increased by a factor of 10 in a period of just two years. The blue line in Figure 1 shows how such exponential growth might lead to very high manoeuvre rates being reported in 2024 and 2025 (e.g., a prediction of nearly 100,000 manoeuvres may be in the report to be filed by SpaceX one year from now, and potentially up to 200,000 in the report to be filed at the end of next year; i.e., 300,000 risk mitigation manoeuvres for the year). As Figure 1 shows, other models provide good fits to the reported data, with both quadratic and cubic trends offering R^2 values greater than 0.998, and indicate that the doubling rate might diminish through time.

Figure 2 shows the manoeuvre evolution as a function of the number of Starlink satellites in orbit at the end of each SpaceX reporting period (this is based on data in the "Up" column on the Starlink page maintained by Jonathan McDowell, here: https://planet4589.org/space/con/star/log.dat). The models used to generate Figure 2 are more robust, in some senses, than those for Figure 1 as they are based on a (likely) stronger causal relationship between the number of Starlink satellites and the number of manoeuvres required. That said, there is still an insufficient number of data points (corresponding to SpaceX semi-annual reports) to form a robust hypothesis about which trend might be the best. I suspect something non-linear based on an expectation of how the number of objects in orbit affects the number of conjunctions, but the trends do change considerably depending on how the number of satellites is calculated. I hope to look into this more deeply and will report back.

The predictions included in Figures 1 and 2 are based on the premise that Starlink launches continue more-or-less unchanged from the way they have in the past (See Figure 3). As we see an increase in the number of Starlink 2nd generation launches, and as other space activity increases, it is entirely possible that the predictions in Figures 1 and 2 become obsolete. Nonetheless, I continue to expect to see an increasing manoeuvre burden, not a decreasing burden.

25,299 is a very large number of manoeuvres for a single operator to have performed within a six-month period, which must also correspond to an even larger number of conjunctions occurring over the timeframe, each with a non-zero collision probability. The non-standard manoeuvre threshold and use of an autonomous collision avoidance system have arguably been successful so far (although it is impossible to credit these exclusively for the lack of collisions), but I wonder how long this might last as we enter the territory of the law of truly large numbers. I will continue to follow the updates from SpaceX (and others, hopefully), develop my analyses and report back.

A quick disclaimer. Whilst the analysis is based exclusively on Starlink, it is because SpaceX is the only entity providing data about the health of its system, in the public domain. SpaceX - rightly in my opinion - calls for other/all operators to provide similar reporting. If they did, we would collectively have a much better understanding of the safety of the space environment.