Jitter in slopes of the HR model values.

[pacesm]

For the lack of a better place, I put this issue here, as this concerns datasets retrieved through the VirES Python client. However, this issue is likely produced on the server when evaluating the magnetic models.

The Swarm HR model values contain a suspicious jitter in the estimated first derivative (see example below Swarm MAGA_HR_1B times and locations, 2016-01-01 ~12:40, using CHAOS 8 - Core model)

It is not completely clear what causes this variation. It seems like a numerical artifact, probably a buildup of rounding errors (timestamps?).

It needs to be investigated.