According to [1] the Space Force released binaries for an updated version of their SGP4 propagator. I haven’t had a chance to compare yet, but it would be worthwhile to compare with Orekit. Just want to make sure the Orekit community is aware of another SGP4 variant.
Regards,
Evan
[1] https://twitter.com/SpaceTrackOrg/status/1325950520733863937
Thank you Evan for this remark.
Indeed, it could be important to see if Orekit follows this new version and if modifications are needed in our implementation.
Bryan
As stated in the comments of the tweet, it’s a pity that only binaries are provided 
It won’t be so easy to upgrade Orekit if there are any discrepancies …
My guess is that like the past Vallado and others will modify their code as to be as close as possible to the updated SGP4 binaries. Hopefully this won’t take too long to happen.
My understanding is that the calculations will be identical.
I suspect that this newer version just supports the alpha-5 NORAD ID format designed to allow NORAD IDs > 99,999. link to space-track documentation
There was an announcement about a higher precision SGP propagator (SGP4-XP), but that requires a special type of TLE that isn’t publically available.
I’ve had a go at comparing the Space Force, orekit and Vallado SGP4 propagators. They all look consistent (apart from one outlier that is the same in orekit and Vallado) for the data that I’ve tried.
Code is here sgp4-comparison.
Rather than using TLEs, I’ve used individual GP elements, which is the preferred method. Plus the orekit TLE parser doesn’t like non-numerical NORAD IDs (like “T0005”, for example).
Great!
Could you open a feature request on Gitlab so we do not forget to support non-numerical NORAD IDs?
I guess it is a simple change.
Feel free to contribute the change by yourself if you want!
Wow! That is a very thorough comparison! Good catch with the new non-numeric IDs.