Subject: The new cloud camera

From: Russet McMillan

Submitted: Fri, 7 Feb 2003 04:20:37 -0700 (MST)

Message number: 658 (previous: 657, next: 659 up: Index)

Hi all,

We've had several comments lately about the difficulty of interpreting
cloud camera images.  The truth is, even those of us who work with it
every night are having some difficulty.  The instrument belongs to
Sloan and therefore we (the ARC community) can't make changes to the
hardware or the fundamental software, but Bill Ketzeback is working on
a new image display that should be better for interpreting some sky
conditions.  That should be available very soon (perhaps within the
week).  In the meantime, Bill has added some more explanation to the
main IRSC web page, and I thought a short message to the community
might also help.

I think the main source of confusion arises from the fact that the
image display is actually a difference of the latest image minus a
combination of several previous images.  This helps to get rid of a
number of stable, warm structures which otherwise would swamp out
everything else: the trees, the domes, and the supports for the camera
itself.  This differencing also makes the camera much more sensitive
to very thin clouds, so that it's easy to tell the difference between
photometric and non-photometric skies.  But when thicker clouds are
present, the image gets confusing.

The camera is configured so that white is warm and black is cold, but
the differencing adds a layer of complexity: white on a difference
image indicates a position where a cloud or other warm object is NOW,
while black on a difference image indicates a position where a warm
object used to be and is no longer.  Grey on the difference image
indicates no change: either an unchanging clear sky, or very thick,
featureless overcast.  Experience over the last few months has led us
to believe that images with a standard deviation of less than 10
usually indicate photometric skies; but under some circumstances,
extremely thick clouds can also produce a low standard deviation.

When just a few clouds are around, their trails are easily visible on
the cloud camera, and the direction and speed of their motion is
apparent on a single image.  Multiple clouds tend to make a messy
image, and it's hard to judge the thickness of the clouds from the

The next generation image, which will be available soon, is a raw IR
image with the stable structures carefully masked off.  The masking
isn't perfect, so this approach is less sensitive to thin clouds;
however, it should be much easier to judge the thickness and behavior
of heaver clouds using the second image.  The plan is to display both
images side by side, for use in different circumstances.

I hope this explanation is helpful to some of the folks who've been
confused.  Be on the lookout over the next few weeks or months for
lots of useful stuff to be added to the web page, thanks to Bill.


APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 658 in the apo35-general archive. You can find
APO  the archive on
APO  To join/leave the list, send mail to
APO  To post a message, mail it to