Predicting and identifying all future SDO eclipses


Senior Member
I've set up a bot to automatically annotate images from the Solar Dynamics Observatory (SDO) satellite once a day and upload the result to a dedicated YouTube channel:
The bot will overlay circles on the images using the expected positions (and sizes) of the sun, moon, and earth calculated using SDO's orbital elements and PyEphem. It also looks ahead to generate predictions of future eclipses caused by either the moon or earth (originally it looked ahead 1 week but I've now extended this to a 2 week search). If it finds an eclipse it hasn't identified before it will generate a video illustrating the eclipse and the time and date and upload it to a separate playlist for future predictions. It's currently set to conduct a search 2 weeks into the future each day and if the start of a previously predicted eclipse changes in timing by more than 30 seconds it will regard it as a new eclipse and publish an updated prediction.

This was made in response to pseudoscientific claims regarding the double eclipse which occurred earlier this month on SDO during the Sunspot Observatory closure. That eclipse was readily shown to be caused by apparent retrograde motion of the moon from SDO, but in the future I think it would be even better if there was a go-to source independent of NASA where it could be shown that the eclipse was predicted ahead of time. There's also a fair bit of conspiracy mongering about the eclipse seasons which occur twice a year on SDO as it passes into earth's shadow, and likewise this bot will predict each of those eclipses before they occur. The source code is available on github: