Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Saturn at 30 to 73 degrees elevation

  • Please log in to reply
12 replies to this topic

#1 Tulloch

Tulloch

    Vanguard

  • *****
  • topic starter
  • Posts: 2,359
  • Joined: 02 Mar 2019
  • Loc: Melbourne, Australia

Posted 08 September 2020 - 03:21 AM

Here in the southern hemisphere we are fortunate at the moment to be able to image the planets at high elevation angles. Jupiter and Saturn are peaking at around 73 degrees for me right now, and I thought I'd take the advantage of imaging one of these planets at a range of angles to measure (if possible) the effect of elevation angle upon colour balance, something I've investigated previously from a theoretical perspective in this thread

 

I had looked at two different methods of determining colour shift, one based upon Solar Irradiance measurements, the other based upon measurements of G2V stars reported in KellySky. But what does this really mean for the planets? So I took images of Saturn at the highest point I could (73.3 degrees) and then at every 5 degrees elevation thereafter as Saturn headed west. I don't normally like imaging to the west, as the planets move over the city of Melbourne (where I live) and I end up pointing right over my hot roof so the seeing deteriorates. However, there was no way to avoid this.

 

The first image below shows an animated gif of the stacked, processed and sharpened videos (5 mins duration @ 100fps, 5000 frames stacked except for 30 and 35 degrees where just 2000 frames were stacked to improve final image quality). Obviously as the elevation reduces the image quality drops, but I was also interested in measuring any colour shift. So I selected the planet's disc and rings in Photoshop and recorded the median red, green and blue values for each of the stacked (but otherwise unprocessed) images. I then plotted the ratios of red/green and blue/green for each image and compared them to the results from the Irradiance and G2V star analysis. I took additional care to make sure I was selecting the same image size (including darker surroundings) for all the images, however there is some error probable with this method, but I tried to reduce it as much as I could.

 

An image showing the process to capture the median red/green/blue values is below, along with the graph showing the measured values compared to the models I discussed earlier. Note that the graph shows the measured ratios compared to the 73.3° elevation angle rather than the 90° value for the models (which I couldn't measure). What the measurements show is that at 30 degrees elevation a shift of around 5% in both red and blue is required to get the colours back to the "true" value, which is slightly less than what both the Irradiance and G2V star analyses suggest.

 

Andrew

 

Celestron Evolution 9.25" SCT @ f21 with ASI 224MC, stacked in AS!3 with 3x drizzle, sharpened in Registax, final touches in Photoshop Elements, final image 50% larger than captured.

Attached Thumbnails

  • 2020-09-05-Elev-30-73-reduced frames.gif
  • Saturn average RGB.JPG
  • Colour shift for Saturn graph.JPG

Edited by Tulloch, 08 September 2020 - 04:25 PM.

  • Magellanico, Kenny V., John Boudreau and 18 others like this

#2 troyt

troyt

    Apollo

  • -----
  • Posts: 1,127
  • Joined: 19 Nov 2008
  • Loc: Australia

Posted 08 September 2020 - 07:09 AM

Interesting Andrew



#3 Tulloch

Tulloch

    Vanguard

  • *****
  • topic starter
  • Posts: 2,359
  • Joined: 02 Mar 2019
  • Loc: Melbourne, Australia

Posted 08 September 2020 - 08:25 AM

And the 30 degree version with the colour correction applied (-5% red, +5% blue) compared to the original.

Attached Thumbnails

  • Sat-30-deg-colour.gif

  • Kokatha man and gfstallin like this

#4 sunnyday

sunnyday

    Voyager 1

  • *****
  • Posts: 11,276
  • Joined: 23 Sep 2019
  • Loc: the Canadian nebula .

Posted 08 September 2020 - 09:12 AM

I too can't wait to see them so high in the sky.
you have made a good account of the difference in height.
thank you .


  • gfstallin likes this

#5 RedLionNJ

RedLionNJ

    Skylab

  • *****
  • Posts: 4,189
  • Joined: 29 Dec 2009
  • Loc: Red Lion, NJ, USA

Posted 08 September 2020 - 10:49 AM

Nice quantitative experiment, putting actual numbers on a well-known phenomena.

 

It's always been obvious significantly lower objects are redder, but it never occurred to me they were also "less blue".

 

This may well help future processing (if the seeing at low elevations ever cooperates).

 

Thanks!

 

Grant


  • Tulloch likes this

#6 BQ Octantis

BQ Octantis

    Skylab

  • *****
  • Posts: 4,022
  • Joined: 29 Apr 2017
  • Loc: Red Centre, Oz

Posted 08 September 2020 - 06:28 PM

Pretty fun science!

 

I would chalk up the shift to basic Rayleigh scattering: blue is scattered the most by the atmosphere and red the least, and the lower the elevation angle the greater the airmass moment arm through which this is happening. This is why the moon—grayscale at zenith (without processing heroics)—is yellow at the horizon. (Indeed, I'd suggest the moon as a more reliable, repeatable test target for your atmosphere than Saturn!) With your atmosphere being above a major city roughly at sea level, there are also all the other factors to consider (water vapor, smog/smoke/pollution, light pollution, to name a few).

 

I was surprised to not easily find a plot of color temperature by elevation angle/altitude/zenith distance on the Google. Photographers have raved about the the "golden hour" since color photography began. And remote sensing for Earth science is old news—the first Landsat was launched in 1972, and they would have had to consider the atmospheric absorption by graze angle (and by polarization). The best I could find was Roger Clark's plot on solar irradiance by wavelength at 20˚, 60˚, and 80˚ (fairly close to your data points). It's not a nice plot of color shift by solar zenith distance split out by color (like yours), but you could very easily derive it from the chart—it even includes the ground truth spectrum for our very own G2V star:

 

plot.atm-sun-modtran-3zenith.a.tgif-s.gi

[Source]

 

The MODTRAN stamp likely explains why I can't find anything on the Google: in professional engineering, knowledge is money—and standardized knowledge is a lot of money. smile.gif

 

BQ


  • happylimpet and Tulloch like this

#7 Tulloch

Tulloch

    Vanguard

  • *****
  • topic starter
  • Posts: 2,359
  • Joined: 02 Mar 2019
  • Loc: Melbourne, Australia

Posted 08 September 2020 - 06:58 PM

Nice quantitative experiment, putting actual numbers on a well-known phenomena.

 

It's always been obvious significantly lower objects are redder, but it never occurred to me they were also "less blue".

 

This may well help future processing (if the seeing at low elevations ever cooperates).

 

Thanks!

 

Grant

Thanks Grant, it is fun to get back and do a bit of science now and then smile.gif

 

Pretty fun science!

 

I would chalk up the shift to basic Rayleigh scattering: blue is scattered the most by the atmosphere and red the least, and the lower the elevation angle the greater the airmass moment arm through which this is happening. This is why the moon—grayscale at zenith (without processing heroics)—is yellow at the horizon. (Indeed, I'd suggest the moon as a more reliable, repeatable test target for your atmosphere than Saturn!) With your atmosphere being above a major city roughly at sea level, there are also all the other factors to consider (water vapor, smog/smoke/pollution, light pollution, to name a few).

 

I was surprised to not easily find a plot of color temperature by elevation angle/altitude/zenith distance on the Google. Photographers have raved about the the "golden hour" since color photography began. And remote sensing for Earth science is old news—the first Landsat was launched in 1972, and they would have had to consider the atmospheric absorption by graze angle (and by polarization). The best I could find was Roger Clark's plot on solar irradiance by wavelength at 20˚, 60˚, and 80˚ (fairly close to your data points). It's not a nice plot of color shift by solar zenith distance split out by color (like yours), but you could very easily derive it from the chart—it even includes the ground truth spectrum for our very own G2V star:

 

 

 

The MODTRAN stamp likely explains why I can't find anything on the Google: in professional engineering, knowledge is money—and standardized knowledge is a lot of money. smile.gif

 

BQ

Thanks BQ, that Irradiance analysis is exactly what I did previously (in this post I referenced earlier), by interpolating between the points on the graph to give me intensity levels at 30 - 90 degrees at 10 degree increments). It is possible to do this yourself (you just need to download and understand how to use the SMARTS program), but it all became too complicated and I just went with a "close enough is good enough" approach.

 

Oh, and the moon doesn't get up higher than about 30 degrees for me at the moment smile.gif

 

Andrew


Edited by Tulloch, 08 September 2020 - 07:03 PM.


#8 BQ Octantis

BQ Octantis

    Skylab

  • *****
  • Posts: 4,022
  • Joined: 29 Apr 2017
  • Loc: Red Centre, Oz

Posted 08 September 2020 - 07:31 PM

Your data appears to indicate the KellySky correction vectors are overstated. Your data points track the irradiance curves much more closely…

 

BQ



#9 Tulloch

Tulloch

    Vanguard

  • *****
  • topic starter
  • Posts: 2,359
  • Joined: 02 Mar 2019
  • Loc: Melbourne, Australia

Posted 08 September 2020 - 07:44 PM

Your data appears to indicate the KellySky correction vectors are overstated. Your data points track the irradiance curves much more closely…

Maybe - the KellySky numbers are better for blue, but the irradiance values are better for red. It's interesting that my measured data points are quite symmetric across the y=1 axis, something that is not reflected in the other analyses.



#10 BQ Octantis

BQ Octantis

    Skylab

  • *****
  • Posts: 4,022
  • Joined: 29 Apr 2017
  • Loc: Red Centre, Oz

Posted 08 September 2020 - 10:40 PM

Could it be that your reference for 1.0 is 73˚ and not 90˚? Since extinction is a nonlinear exponential curve ( X = (cos z + 0.025 e-11cos z )-1 , where X is the number of air masses and z is the zenith angle), that offset might actually put it right on the irradiance curves…

 

BQ



#11 BQ Octantis

BQ Octantis

    Skylab

  • *****
  • Posts: 4,022
  • Joined: 29 Apr 2017
  • Loc: Red Centre, Oz

Posted 08 September 2020 - 10:43 PM

εύρηκα!

 

https://www.fortlewi...tail.asp?ID=210

 

BQ



#12 Tulloch

Tulloch

    Vanguard

  • *****
  • topic starter
  • Posts: 2,359
  • Joined: 02 Mar 2019
  • Loc: Melbourne, Australia

Posted 09 September 2020 - 02:28 AM

Could it be that your reference for 1.0 is 73˚ and not 90˚? Since extinction is a nonlinear exponential curve ( X = (cos z + 0.025 e-11cos z )-1 , where X is the number of air masses and z is the zenith angle), that offset might actually put it right on the irradiance curves…

Maybe, but the offset between 90 and 70 degrees for both the KellySky and irradiance values is only around +/-0.005 for each colour, so while it may make some difference, I don't think it will be that much. Also, the KellySky difference is greater in the red, while the irradiance difference is greater in the blue. 

 

It appears that this analysis is only based Rayleigh scattering and doesn't take any other atmospheric factors into consideration. Also, the variations are based upon BVR (photometric) filters, rather than RGB so there is another difference. Finally, the ratios for red to green don't really change until you get up to 45 degrees and the blue ratios are much larger, something I don't see in any other analysis (or my measurements).



#13 BQ Octantis

BQ Octantis

    Skylab

  • *****
  • Posts: 4,022
  • Joined: 29 Apr 2017
  • Loc: Red Centre, Oz

Posted 09 September 2020 - 02:45 AM

I meant eureka! on finding the bloody plot! lol.gif

 

On the other hand, the plot does beg, why do the other two sources use green as the reference green when it is red that is the least affected? The asymmetry in both plots and the symmetry of yours seem like an artifact of the mind, given the ugly equation just for extinction…

 

BQ




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics