How I set These Up

[theskyhound]

Hi Razvan,

I'll defer to BMD on the question of ephemeris calculations in HORIZONS since I only use ST4v to calculate the ephemeris. The trick with ST4v is to input the appropriate epoch for the osculating elements. If the epoch is too far off the ephemeris positions will be incorrect, as you're aware. Back in July 2021 (earlier in this thread) BMD & I tried to input elements from HORIZONS at fractional days, like Jul 9.75 with limited success. This would be the ideal way to deal with close approaches using ST4, enter osculating elements at for example, October 9.0, 9.25, 9.50 & 9.75, then have ST4 choose the closest epoch for the calculations. Currently ST4 doesn't seem to keep epochs that are too close together in time as this would require. I imagine having the program try to keep this all straight would really make things complicated, as if orbital mechanics aren't already complicated.


Phil S.

You are correct that SkyTools will look at the epoch of osculation and use the orbit with the closest time. The only limitation that I am aware of, for epochs being close together recognized as the same epoch, comes from the a test for round off error in the times so that epochs that are indeed the same are recognized as such. I just checked the code and it is set to consider epochs that are within one minute of each other to be the same epoch. So unless my conversion of one minute to Julian Centuries is off, then epochs that are more than a minute apart will be stored separately.

So I suspect that the orbits are being combined for other reasons. The orbital elements are also checked against each other, and if there is little difference, then they are determined to be the same orbit (epoch). So for example, if you took two orbits for a main belt asteroid that were calculated for times a day apart, they would probably have almost the same elements, and presumed to be the same orbit. So it may simply be that your orbital elements aren't changing as quickly as you think they are.

SkyTools 3 had a less accurate orbital model, so when we had close passes in years past, I would use Horizons to generate orbital elements like you guys are doing in order to improve the accuracy. I found that producing elements once per hour was sufficient.