python - PyEphem: dates of solstices and equinoxes and their validity on longer time scales -

My question is how much time has passed, on which pipes for solutes and synonyms and solar geometry dates Provides accurate results.

So far, I got boundary BC from 99 8 9 03-20 AD. In this very informative post on Geethhub 99 99-12-31, and the overall signal is that at present, either on the site, going forward +/- 20,000 years the results become irregular.

I would like to make it clear that I am trying to get the sun's position for a long time - to calculate the solar radiation coming from reality and going to be BP for 10,000 years. Sun's Azimuth at a given place For example, the pyapem provides a good alternative for the works provided by Burger, 1978 (J. Atmosposs. SC, 35: 2362-2367). For me, these algorithms will be an essential advantage of PyEphem at that time the track holds, while the Earth's orbit usually preceding algorithms (for example, is set at a particular moment in Warnal Equinox) March 21 is.

Usually, the algorithms of burgers and others are classified as variables in the orbit of the Earth in the scale of the last glaciers (up to 126,000 years BP). When evaluating PyEphem on that border, I came during some strange behavior for the dates of Solthesis, when the date lies well before the present day:

  import ephem date = ephem.date ((- 59000, 1,1)) orbitPoints = [ 'vernal_equinox_start', 'summer_solstice', \ 'autumnal_equinox', 'winter_solstice', 'vernal_equinox_end'] = {} at [ 'vernal_equinox_start'] = ephem.next_vernal_equinox (Date) Date [Date '[Circle_Esolistis'] = Effam. Date '[winter_synx_start'] date of [[Vernal_vivinx_start ']) [date] [' vernal_equinox_end '] = ephem.next_vernal_equinox (date orbitPoints for orbitPoint [' winter_solstice ']): Date = dates [orbitPoint] distance = body_distance (Sun, date) hlon, hlat = body_hpos (Sun, date) print '% -20s% 30s% 8.4f% 12s% 12s'% \ (orbitPoint, date, distance, Acelelel, Acelatti) days between the print Ikvinoks :% .1f, year length (spring equinox):% .1f '% \ (dates [' autumnal_equinox '] - dates [' vernal_equinox_s tart '], \ dates [' vernal_equinox_end '] - dates [' vernal_equinox_start ']) gives  

:

  vernal_equinox_start -59001/12/20 01:33:21 1.6255 180: 01: 53.6 0: 00: 01.7 Grishm_slstis -59000 / 2/19 04:16:18 1.3380 180: 03: 56.7 0: 00: 09.6 Fall Hritu_vivinks -59000/7/l 09:20 58 0.3704 0: 00: 42.2 0: 00: 09.6 Srdion_solstis -59000/l0/9 07:35: 16.2 days between the equinoxes: 39 1.1335 179: 56: 59.4 -0: 00: 12.4 vernal_equinox_end -58999/10 / 10 07:40:47 1.1348 180: 00: 52.7 0: 00 193.3, year length (spring equinox): 65 9.3  

whenever I date to date = ephem. Date ((- 25000,1,1)), even then that year Area to get a reasonable (?) Values ​​.

If PyEphem actually returns the right result with my period of interest (10,000 years BP) then it will serve my purpose. However, I want to make it clear and open to suggestions for expanding this category, even if it is for verification only. I was seeing Skyfield as an option but it does not appear to provide extended range.

A clear explanation of the limitations of the validity of PyEphem and any suggestions would be greatly appreciated.