Jonathan's Space Pages

For each Starlink satellite:

• I define the active lifetime as the end of life date minus the launch date.
• I define the end of life date to be the date at which the satellite stops controlled manuevers and begins uncontrolled decay. This can be many months after the satellite is lowered out of the active constellation,

Therefore, the active lifetime as defined here is an upper limit to the useful operational lifetime of the satellite. A future enhancement would be to record 'date reached operational orbit shell' and 'date last departed operational orbit shell' and use that for the 'useful lifetime'. But for now, I'll stick to the definition above.

However, a majority of the Starlinks have, at this writing, not reached their end of life dates. I therefore use the Kaplan-Meier estimator (`censored statistics', `survival analysis') to infer the lifetime population percentiles from both completed Starlink lifetimes and (for still-operating satellites) the lower limits to those lifetimes.

The X axis in the plot below is the fraction of the inferred Starlink population, and the Y axis is the lifetime. The current version of the plot shows that we infer 20 percent of the population (X=0.2) has a lifetime of 4.2 years (Y=4.2) or less.

There is not yet enough data to determine a median lifetime; we can constrain it as being greater than 4.5 years based on this data.

Data file