I am trying whether it is possible to view hot steam sources like power plant cooling towers and resulting cloud using satellite images. I was advised to try Landsat (needed free data for initial research)

I downloaded two Landsat 7 images from various regions with two big power plants (about 4 cooling towers each) and tried to use common band combinations (3-2-1, 4-3-2, 4-5-3, 7-4-2, 5-3-1) in SAGA GIS to create composite images. I though it could show the vapor cloud at least near the power plant but I am not able to identify it. It is possible to somehow filter the images and remove most of the background to highlight what I need?


PS: the goal is to get an approximate result whether it has sense to start more detailed research in this area or not. I have QGIS and SAGA on all of my computers.

  • 2
    You're looking for thermal brightness, so I would suggest using only the infrared bands for any composite image (e.g. 4, 5, 6). It might be easier to examine only one band at a time initially, to verify there is any brightness difference near the plant, and then plan how you'll put those IR bands together into a meaningful composite. – Erica May 6 '14 at 13:43
  • I started examining the bands 4-6 as you suggested: postimg.org/gallery/tjk3o3sq but do not see any clouds yet... – Juhele May 8 '14 at 21:22

If you are looking at temperature differences, the only useful band is the thermal infrared (band 61/62). Unfortunately, the spatial resolution of this band is lower than the other band, 120 m. Therefore you might not have enough details.

On the other hand, the vapor cloud should be white in most wavelenght. It is surprising that you don't see those white patches on you color composites.

  • hi, I know it is strange but I do not see the cloud even on 1-2-3 composite: postimg.org/image/lwqtn6oe1 - will try 2nd image of another power plant. – Juhele May 8 '14 at 21:04
  • the background buildings are already very bright, and the bright areas seem to be saturated. I don't see the clouds either (just a small one near the top left. Please try to change the contrast for the bright areas. – radouxju May 9 '14 at 5:45
  • thanks for hint. Tried different composite processing in SAGA: postimg.org/image/mpggj3bi1 and there are few clouds. Will check also the IR bands. – Juhele May 10 '14 at 23:34

In Landsat 8, when you download the geographic image it comes with a quality band (QB) that is primarily meant to show clouds. Maybe that will help?

  • 2
    Could you tell us how you propose using the QB to answer this question? – whuber May 20 '15 at 14:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.