Thank you. And thanks for the link. Interestingly one of the datasources was the NASA one.

One reason for the minor difference between AstronomicalData and SunPosition is that the older AstronomicalData code base uses a simplistic conversion to Julian date for its internal computations. SunPosition/Alpha use a more modern/less naive conversion. The result being that you get a slightly different answer with the newer result less affected by errors in this particular component.

Two things: Firstly, I do not really consider NASA to be the gold standard; I used these words only in my reply to Nasser as he seemed to regard them as such.

Secondly, Mathematica delivers the actual sun position, not its apparent position. The latter would imply taking atmospheric refraction into account. At sunrise that amounts to about 0.5 degree, and is at a minimum when the sun is at zenith position. So, 0.3 deg off isn't really much of a problem, or so it seems.

Output seems to be OK now, great! Now, if the speed could be improved somewhat...

I'm investigating. I will let you know if I have found the problem.

I have now asked two users that are in my time zone, they get the same result as me. Another user is on GMT and his results differ from both yours and mine.

It really seems the calculation take your own locale into account, which it shouldn't.

Amsterdam is in TimeZone 1. The way Mathematica is set-up for daylight savings time entails it adds an hour if DST is in effect. On march 1, however, DST is not in effect and TZ 1 is correct for both me and Amsterdam. Anyway, setting TimeZone to two still has the sun rise in the middle of the night.

But I think that's beside the matter. With SunPosition you're asking for the position of the sun at a given day and time above a specific location. All these are the same in my call and in yours and should therefore return the same result. They don't.

In the meantime I've asked another user for his results and he comes up with exactly the same results I got.

No, I'm on a standard V10.0.0

I just executed the code again and still get the same result:

In[6]:= SunPosition[GeoPosition[{52.37`, 4.89`}], 
   DateObject[{2013, 3, 1, #, 0, 0}, TimeZone -> 1]] & /@ Range[0, 23]

Out[6]= {{Quantity[95.7, "AngularDegrees"], 
  Quantity[-5.1, "AngularDegrees"]}, {Quantity[107.6, 
   "AngularDegrees"], 
  Quantity[3.9, "AngularDegrees"]}, {Quantity[120.0, 
   "AngularDegrees"], 
  Quantity[12.3, "AngularDegrees"]}, {Quantity[133.5, 
   "AngularDegrees"], 
  Quantity[19.6, "AngularDegrees"]}, {Quantity[148.3, 
   "AngularDegrees"], 
  Quantity[25.4, "AngularDegrees"]}, {Quantity[164.5, 
   "AngularDegrees"], 
  Quantity[29.1, "AngularDegrees"]}, {Quantity[181.6, 
   "AngularDegrees"], 
  Quantity[30.2, "AngularDegrees"]}, {Quantity[198.6, 
   "AngularDegrees"], 
  Quantity[28.6, "AngularDegrees"]}, {Quantity[214.6, 
   "AngularDegrees"], 
  Quantity[24.5, "AngularDegrees"]}, {Quantity[229.2, 
   "AngularDegrees"], 
  Quantity[18.4, "AngularDegrees"]}, {Quantity[242.5, 
   "AngularDegrees"], 
  Quantity[10.9, "AngularDegrees"]}, {Quantity[254.8, 
   "AngularDegrees"], 
  Quantity[2.4, "AngularDegrees"]}, {Quantity[266.6, 
   "AngularDegrees"], 
  Quantity[-6.7, "AngularDegrees"]}, {Quantity[278.6, 
   "AngularDegrees"], 
  Quantity[-15.8, "AngularDegrees"]}, {Quantity[291.4, 
   "AngularDegrees"], 
  Quantity[-24.6, "AngularDegrees"]}, {Quantity[305.7, 
   "AngularDegrees"], 
  Quantity[-32.6, "AngularDegrees"]}, {Quantity[322.2, 
   "AngularDegrees"], 
  Quantity[-39.2, "AngularDegrees"]}, {Quantity[341.3, 
   "AngularDegrees"], 
  Quantity[-43.5, "AngularDegrees"]}, {Quantity[2.0, 
   "AngularDegrees"], 
  Quantity[-44.8, "AngularDegrees"]}, {Quantity[22.5, 
   "AngularDegrees"], 
  Quantity[-42.9, "AngularDegrees"]}, {Quantity[41.1, 
   "AngularDegrees"], 
  Quantity[-38.0, "AngularDegrees"]}, {Quantity[57.1, 
   "AngularDegrees"], 
  Quantity[-31.1, "AngularDegrees"]}, {Quantity[71.0, 
   "AngularDegrees"], 
  Quantity[-22.9, "AngularDegrees"]}, {Quantity[83.6, 
   "AngularDegrees"], Quantity[-13.9, "AngularDegrees"]}}

Did you see my analysis above? It showed that the actual W|A call also includes the caller's location. If that is somehow used in the calculation that would explain why we see different results. However, the results should not depend on that and it seems likely this is a bug.

Good to hear Jeff! Any suggestions as to how to get SunPosition return the same results as AstronomicalData?

Thanks, Cormullion. If you have a use for my converted sun position code just let me know.

For the sun position and given the required accuracy of a minute or so you don't really need that as most algorithms are good for decades, if not centuries to come. See for instance this site.