Re: [TASS] photometric vs. non-photometric

[mgutzwiller@LANVISION.COM]

For the Mark III there is a fairly simple metric called the background
variation.  Since the Mark III image was a drift scan image each row was
taken at a different time.  If the cloudiness changed during the exposure
the background would change also.

This is one of the characteristics that StarPost uses to filter the images.


Mike G.


                -----Original Message-----
                From:   Stupendous Man [mailto:richmond@A188-L009.RIT.EDU]
                Sent:   Thursday, September 02, 1999 2:13 PM
                To:     TASS@LISTSERV.WWA.COM
                Subject:        photometric vs. non-photometric

                  Peter Mount asks:

                > What does everyone thing the criteria are that determine
what
                > type a particular
                > night are, and how would you deal with either a night that
starts as
                > Photometric quality, but deteriorates and visa versa?

                  Terminology: "photometric" is an astronomical term which
means
                "skies are clear and stable".  In practice, it often means
                "Fred went outside once an hour and never saw a cloud."

                  Technically speaking, it's complicated.  I do this sort of
work
                for a living, so I'd be happy to go into MORE gory details
than any
                sane person would want to know -- just ask.  But let me give
                two succinct definitions, and skimp on the details for now.

                  Rock-solid method to see if night is photometric: collect
observations
                of stars with known magnitudes at a range of positions
around the sky.
                Run the observations through a procedure which checks to see
if they
                obey equations of the form

                     instrumental mag  =  known mag +  A  + c*(color)  +
k*airmass

                where A, c, and k are constants, one per passband, and
"color" is the
                color of the star, calculated from its known magnitudes in
two passbands.

                  If the observations obey the above equation, then the
night was
                photometric; if not, it wasn't.

                  Quick n' dirty method, which can be done during a night:

                         Step 1: carry out rock-solid method over a few
weeks

                         Step 2: calculate mean values for constant A, c, k
in each passband

                         Step 3: during a night, use the mean values to
calculate the
                                 expected instrumental magnitude for stars
as the telescope
                                 takes images.  If the expected magnitude
matches the
                                 instrumental magnitude, then the night is
photometric.


                  So, if the plan is for a user to be able to specify,
"Please take
                this action only if the conditions are photometric", one
must

                             a. take pictures
                             b. reduce pictures to "clean" versions
                             c. measure and identify standard stars in the
images
                             d. compare measured magnitudes to a known model

                all during the night!  It's a tough task.

                  A better idea is to have a VERY simple program scan each
new image
                and look for "cloudiness" -- Peter McCullough finds that
Stardial
                images during cloudy weather are easy to distinguish from
those taken
                during clear conditions, by several metrics.  If any image
shows
                hints of "cloudiness", then declare the night "not
photometric".

                  One might consider classifying two halves of the night
(early and late)
                separately, or even each hour .... but that's overkill, to
my mind.

                                                          Michael Richmond