• Ultraviolet Photography
  •  

[SOL] Solar Radiation Spectrum

15 replies to this topic

#1 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 5,094 posts
  • Location: USA

Posted 08 July 2017 - 15:41

Giving this chart its own topic for easy reference.

Note the relative lack of UV in the solar spectrum. Look to the left of the violet band beginning at (unmarked) 400 nm.

Attached Image: post-4-0-97702500-1433605108.jpg
Andrea G. Blum
Often found hanging out with flowers & bees.

#2 Mark

    Member

  • Members(+)
  • 288 posts
  • Location: Massachusetts, USA

Posted 09 July 2017 - 01:01

No wonder we have to push our (UV) cameras so hard outdoors!

#3 Bill De Jager

    Member

  • Members
  • 160 posts
  • Location: Northern California, USA

Posted 09 July 2017 - 18:44

And that's with the sun high in the sky and atmospheric attenuation minimized.

I'm wondering about the solar corona at the eclipse next month. The coronal spectrum must be quite different than normal sunlight given the far higher temperature of the corona. UV ought to be dominant, albeit less intense than in normal sunlight. However, I hear that atmospheric turbulence more strongly affects UV than visible light and that could affect coronal photos taken in UV.

Edited by Bill De Jager, 14 July 2017 - 04:49.


#4 Øivind Tøien

    Member

  • Members
  • 16 posts
  • Location: Fairbanks, Alaska

Posted 10 July 2017 - 22:32

I wonder what the spectrum would look like in the 'blue hour". Of course absolute levels would be lower, but what about relative distribution?
UV reception has been suggested to help some animals living at high latitude during low light conditions.
Øivind

#5 aphalo

    Pedro J. Aphalo

  • Members
  • 26 posts
  • Location: Helsinki, Finland

Posted 07 August 2017 - 20:22

Answering a bit late as usual. ;)
There is a quick on-line calculator for the short end of the solar spectrum (earlier UV only, but current version UV and visible), developed by Sasha Madronich.
http://cprm.acom.uca...nteractive_TUV/
The quick interactive version is handy when wanting to estimate the solar spectrum at a given solar elevation, or at a given location and instant.
For spectral output select "OUTPUT OPTION 2" and "IRRADIANCE, SPECTRAL (W m-2 nm-1)".
The default wavelength range is 280 to 420 nm. The upper limit can be increased as needed up to around 1000 nm. The "increments" input is the number of wavelength steps, rather than step width.
The option related to using 4 vs. 2 streams for calculation should affect results only at very low solar elevation angles, when 4 streams may give a better estimate.

The underlying simulation model (TUV) is also available as source and is written in FORTRAN. This model also simulates atmospheric chemistry. The trickiest part is estimating the "clouds" input data.

When the solar spectrum at ground level is the main interest meteorologists tend to use the model libRadtran that in its latest incarnation (version 2) can simulate the whole solar spectrum from UV to IR. http://www.libradtran.org/ As far as I know there is no interactive version and it does require quite a few inputs.

Edited by aphalo, 07 August 2017 - 20:24.


#6 JMC

    Member

  • Members
  • 128 posts
  • Location: UK

Posted 08 August 2017 - 07:01

Adding to the list of experiments I want to do, I have access to a little Ocean Optics USB spectrometer, which has been calibrated for absolute irradiance, and it has a cosine corrector on the end of the fibre. I'd like to do some measurements of the spectra during the day. However I need some advice on the actual logistics of how to do it, to get consistent, useful data. Do I need to point it at the sun, do I need to avoid the sun, can I just stick the fibre 'up in the air', etc, etc? Also it is slightly limited in that it is mains powered, so I can't just disappear off into the middle of nowhere and take the measurements away from buildings, however I could potentially run it on the roof of my house to get early morning through to early afternoon data without any obstructions.

#7 aphalo

    Pedro J. Aphalo

  • Members
  • 26 posts
  • Location: Helsinki, Finland

Posted 15 August 2017 - 20:59

The problem with array spectrometers is that stray light is not better than 1:000 in the UV region, usually worse. Array spectrometers are always single monochromator instruments and the UV pixels receive some visible light due to internal reflections and other imperfections. Not even the best Ocean Optics spectrometers can be used to measure the solar spectrum down to the shortest wavelengths. One can to some extent get around this problem by measuring the stray light and then subtracting it. The choice of filter for this is crucial, and one may need to compensate for the absorption by the filter of the radiation creating the stray light. We have been using polycarbonate, as much of the stray light in our spectrometer is from the IR, that is in part attenuated by the more common UV long-pass glass filters. Believe it or not, under clear sky conditions a simulation with one of the models I mentioned above, will provide a far better estimate of the UVB region of the solar spectrum than measuring it with a single monochromator spectrometer. So, if you do the experiment with the little Ocean Optics, take note of the exact time and geographical coordinates, so that as a second experiment, the measured and simulated spectra can be compared.

As to the direction in which to point the cosine diffuser, it just depends on what you are interested in. Most meteorological data is collected with the surface of the diffuser perfectly horizontal. However, if you want to know the spectrum of the light being received by a flower or a wall, you would need to position the surface of the cosine diffuser parallel to the surface of your interest. Obviously the spectrum of direct radiation from the sun is different to that of scattered radiation from the sky. Oversimplifying things, by changing the direction the cosine corrected diffuser points to, you change the mix ratio of light form these two sources. In my experience the most frequent source of errors is shading by the operator (or reflections from those wearing white lab coats) followed by shading by other obstacles. And in the near infrared reflection from vegetation can be a significant problem even when the vegetation is not shading the sensor at all.

For reproducible and comparable output: keep the surface of the cosine sensor perfectly horizontal, make sure you do not disturb the reading by keeping yourself below the level of the sensor or several meters away. Avoid other obstacles by either being far from them, or measure from the roof of a building or other elevated place. Around local solar noon, the unexpected disturbances are less, but when the sun is low in the sky measurement becomes very tricky because, 1) sensor levelling errors disturb the measurements a lot, and 2) it is almost impossible to avoid disturbances from the surroundings unless the sensor is situated on a high elevation.

If you point the diffuser straight into the sun, the spectrum you will measure will change less through the day (under clear sky) as direct radiation will predominate always, but it may take some effort to orient the diffuser correctly, and to find/model spectral data to compare to.

By the way, which model of Ocean Optics spectrometer will you use? All those I know of are USB powered, and can be used in the field tethered to a laptop. And the very small OM modules require only low voltage DC as power. Even if sold with a mains adapter, it should be feasible to use them with a small LiPo battery or a power bank depending on the voltage required.

Edited by aphalo, 15 August 2017 - 21:06.


#8 JMC

    Member

  • Members
  • 128 posts
  • Location: UK

Posted 16 August 2017 - 06:25

Thanks for the advice Pedro, very useful. It's an Ocean FX model, bought for something different, and is mains powered. I just thought as it was something I had to hand, I would try it. The need for mains power shouldn't be too much of an issue as I could run it on the roof of my house, using an extension lead.

#9 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 5,094 posts
  • Location: USA

Posted 16 August 2017 - 16:53

Or a portable battery pack?
Andrea G. Blum
Often found hanging out with flowers & bees.

#10 JMC

    Member

  • Members
  • 128 posts
  • Location: UK

Posted 16 August 2017 - 17:20

Yep, I suppose that would work too. Added the work to the list :)

#11 JMC

    Member

  • Members
  • 128 posts
  • Location: UK

Posted 14 October 2017 - 10:50

When will we get consistent blue sky in the UK? Oh, sorry I forgot, we don't.... I had a play about with the OO FX this morning, took it outside and tried to get a spectra from the light in the back garden. 100ms integration time, with each spectra being an average of 50 scans, FX spectrometer with 25um slit, 600um fibre and cosine corrector. Calibrated for spectroradiometric measurements. The end of the cosine corrector was held horizontal about 1.5m off the ground. This was done just before 10AM on 14th Oct 2017, 51.4 degrees north, 0.5 degrees west, weather was overcast, varying from cloud, to varying degrees of hazy sunshine, as the clouds blew through. There was no proper 'blue sky', it is the UK after all. Anyway, thought I would share what I saw.
Attached Image: Sun spectra.jpg
Looks as though at least I can pick up differences as a result of cloud cover :), and as mentioned above in one of the posts, no wonder we struggle with getting light for UV images.

#12 Cadmium

    Member

  • Members(+)
  • 893 posts

Posted 15 October 2017 - 05:25

Jonathan, Very nicely done!

#13 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 5,094 posts
  • Location: USA

Posted 15 October 2017 - 15:45

What is that large dip in the charts around 770 nm?
Andrea G. Blum
Often found hanging out with flowers & bees.

#14 JMC

    Member

  • Members
  • 128 posts
  • Location: UK

Posted 15 October 2017 - 18:08

Thanks Steve. Andrea, I 'believe' they are something to do with clouds/water vapour. However if anyone has a reference to back that up please post it up. Interesting to note that down at 293nm there is nothing - Andrea, that may well explain your difficulty getting good images with the 293 filter. Obviously a sunnier day, would I guess produce more UV down there, but even so, when comparing it to the region covered by the Baader U, it will be much much less intense.

#15 UlfW

    Ulf W

  • Members
  • 41 posts
  • Location: Sweden, Malmö

Posted 15 October 2017 - 19:30

I think the dip at "around 770nm" are the oxygen Fraunhofer lines designated A. (more close to 760nm).

https://en.wikipedia...raunhofer_lines

Edited by UlfW, 15 October 2017 - 19:33.

Ulf Wilhelmson
Curious and trying to see the invisible.

#16 Andy Perrin

    Member

  • Members(+)
  • 818 posts
  • Location: United States

Posted Yesterday, 00:54

JMC, definitely it goes down fast beyond 350, but it's hard to tell if there's an actual spectral absorption line there with the graph scaled as it is. You'd have to plot just the 250-350 nm range to even tell, and probably it should be in direct sunshine without haze. I think water absorbs in that region, so there may be more there on a dry clear day.