Bug in Lunar Eclipse Program between 2000BC and 1000AD

[Alexander Wolf]

Original topic: https://sourceforge.net/p/stellarium/discussion/278769/thread/ad20da89/

Message #1: Total Lunar Eclipse Date UTC Difference @ Max Eclipse Stellarium vs. NASA tabulated data (standard)

-1997 6-May -13
-1798 22-Aug -11.5
-1599 14-Jun -10
-1397 3-Feb -9
-1198 23-May -8
-999 14-Mar -7
-796 17-Apr -6
-597 9-Feb -5
-399 1-Dec -4
-199 20-Mar -3
3 4-May -3
199 21-Oct -2
400 17-Dec -2
600 28-Sep -1
799 21-Jul -1
900 13-Aug -1
951 23-May -0.5
-----------------------
998 6-Nov 0
1200 22-Dec 0
1251 1-Oct 0
1298 21-Sep 0
1400 3-Oct 0
1599 6-Aug 0
1798 29-May 0
2000 16-Jul 0
2015 28-Sep 0
2199 2-Nov 0

Message #2: Originally I was looking at lunar eclipses between 25 AD and 35 AD (Jesus's Death) for additional correlations. I found that there was a (3 hour difference) between the UTC time for maximum lunar eclipse and what the latest version of Stellarium predicts.

I then wondered how pervasive the error was. I also remember fainty that in the last 1-2 years there was a similar issue with a time offset that was corrected for the 2014 and 2015 lunar eclipses.

In short, I sampled the nearest total lunar eclipese every 200 years starting in 2000BC through 2015 AD and was happy to find that for the years 1000AD-Future, the lunar eclipses agree perfectely with the NASA tabultated values for maximum solar eclipse time (UTC).

As this thread title notes, there seems to be a slow degredation in the simulated stellarium times versus the NASA tabulated values in the years 2000BC and 1000AD.

I listed the approximate number of hours difference, finding that the error continously increased as more years passed from near 0 hours (UTC) in 1000AD to -13 hours (UTC in 2000 BC).

Message #3: The Total Lunar Eclipses checked can be found at this link to the NASA website's Lunar eclipses from 2000BC to 3000AD.
http://eclipse.gsfc.nasa.gov/LEcat5/LEcatalog.html

[gzotti (georg-zotti)]

Dear anonymous (poster of SF report),

which times do you compare?

NASA/Espenak (The URL you gave) lists TD of greatest eclipse. If you switch off DeltaT correction (Config/Navigation), your "time panel time" minus zone offset is used as TD.

-1997-05-06 is shown at mid-eclipse around 18:09 (TD) then. Offset against Espenak is less than 1 hour, which is not perfect, but not as dramatic. Reason may well be the different models implemented by Mr. Espenak and Stellarium. As always, there is still room for improvement.

[gzotti (georg-zotti)]

Seems solved. Misunderstanding...