High Accuracy Sun Position Calculation

[Kiwi Geoff]

(09-23-2015 10:58 PM)Marcel Wrote:  Hi Geoff,
With Astro Lab 4 (from Meeus book..), I found
R: 05:46:14
T: 11:54:08
S: 18:03:01
in local time.
Marcel.

Thank you Marcel, this is getting very interesting.

I thought the "Sunrise/Sunset" issue I found only applied to SPA, however if Astro Lab 4 is derived from the J. Meeus book (and not SPA) then it implies that the algorithm issue goes back to the J Meeus book.

From your Astro Lab4 output, we can see that it is "sunset" time that does not agree with my JPL DE430 derived time.

So who is correct here ?

Perhaps the most accurate resource is:

JPL Solar System Dynamics.

which gives the "sunset" (using parameters given previously in the thread) as

Date Time (-7) Az(r-appr) Elev
2015-Sep-20 18:01:21 Cs 271.8562 -0.2141

I also found that we can use the:
SPA online calculator
to generate the Solar position every 1 second, and then "manually" extract the R.T.S.

SPA generated SOLAR position:
R 9/20/2015 05:46:16 87.884450 (degrees Azimuth)
T 9/20/2015 11:54:09 180.002206 (degrees Azimuth)
S 9/20/2015 18:01:22 271.858286 (degrees Azimuth)

This now agrees within ONE HALF SECOND (of time) with the JPL DE430 ephemeris.

JPL DE430 ephemeris gives the following:
R 2015-9-20 05:46:15.930 Az = 87.884
T 2015-9-20 11:54:08.515 El = 51.252
S 2015-9-20 18:01:21.472 Az = 271.858

Thus we can see that the SPA "sun position code", does not agree with the SPA (and Lab4) "Sunrise/Sunset" times.

I feel this "proves" my above assertion - that something is indeed wrong with the Sunrise/Sunset algorithm given by J Meeus. I wonder if there was an erratum at some stage - or am I the first to find this issue with the "Sunrise/Sunset" algorithm?

Regards, Geoff Hitchcox, Christchurch, New Zealand.