Thread by @ProfHughLewis

A follow-on from yesterday's thread with a note about averages. In a #SOCRATES report from 30 June 2022 the average collision probability for each #Starlink conjunction was 3.7E-6 but the range of values can be broad (chart shows data since 2019) [1/n]

https://twitter.com/ProfHughLewis/status/1543986914004533252

#SOCRATES predicted some events with a collision probability > 1E-2 (1-in-100) & some with a probability < 1E-7 (1-in-10,000,000). The average value might seem to be almost negligible & you might think all conjunctions would be similar, but that's not the case [2/n]

In addition, some #Starlink & #OneWeb satellites experience more conjunctions than others. Most satellites experience relatively few encounters but a few satellites are involved in a relatively large number (charts shows data for 7 days from 30 June 2022) [3/n]

The pattern (i.e. frequency) distributions seem clear & I expect them to be similar (not exactly the same) for predictions coming from @18thSDS based on high accuracy ephemerides. Indeed, some other work I have done recently mostly confirms this. [4/n]

@18thSDS The insight/conclusion that should be drawn is that averages only tell part of the story. A low average collision probability does not mean that all encounters will have a low collision probability (and vice versa) [5/n]

@18thSDS Quoting an average collision probability for a satellite does not mean that all satellites in the same constellation will experience the same probability in all encounters. [6/n]

@18thSDS In the case of #Starlink, a probability of 3.7E-6 (the average I quoted above) is not at the level at which a manoeuvre might be performed, yet Starlink satellites do manoeuvre (more than 12,000 times so far) in response to encounters with other objects. [7/n]

Hope this addition was useful and interesting! Thanks again to @TSKelso for access to & support with the #SOCRATES data, and thanks also to @mikeclindsay for some very helpful discussions. [8/8; fin]