The basic idea is to calculate a photometric solution for a night, given a set of catalog stars with known magnitudes and a set of detected stars with instrumental magnitudes.
photom catalog_file racol deccol cat_nfilt V=4 B=6
detected_file racol deccol jdcol aircol flagcol
det_nfilt a,b,c d,e,f
mode= [fixk= ... ]
[det_files ... ]
[radius=] [orphans=] [outfile=]
where
catalog_file is name of the ASCII data file containing catalog
of standard star data
racol is the column of the catalog file in which are the
RA values (decimal degrees, J2000)
deccol is the column of the catalog file in which are the
Dec values (decimal degrees, J2000)
cat_nfilt is the number of "filter=column" pairs for the
reference catalog, immediately following.
Must be >= 1
V=4 is a "filter=column" pair; here, "V" is the name
of a filter, and "4" is the index of the
column containing magnitudes in that filter
in the catalog
There must be exactly "cat_nfilt" such arguments
denoting passbands in the catalog file.
detected_file is the name of the ASCII data file containing
stars detected in some image, with instr. mags
racol is the column of the detected file in which are the
RA values (decimal degrees, J2000)
deccol is the column of the detected file in which are the
Dec values (decimal degrees, J2000)
jdcol is the column of the detected file in which are the
Julian Date values
aircol is the column of the detected file in which are the
airmass values
flagcol is the column of the detected file in which are the
"quality flag" values
det_nfilt is the number of "filter=magcol,magerrcol" pairs
for the detected files, immediately following.
Must be >= 2
a,b,c denotes some passband in the detected file:
The values a,b,c are integers, where
a = column with passband name
b = column with magnitude val
c = column with magerr val;
negative value means a
saturated star
A comma "," separates the integers.
There must be "det_nfilt" such arguments denoting
the passbands in the detected file.
mode describes the form of the photometric solution.
Valid options are:
extinct include extinction term,
so equation looks like
V = v + a + b*(color) + kX
noextinct no extinction, but use
one zeropoint per frame j:
V = v + a + b*(color)
j
fixk Valid only in "noextinct" mode, so that we are
NOT trying to determine the extinction coefficients.
Use a fixed value of first-order extinction in
each passband, so that
V = v + a + b*(color) + kX
j
The user must specify the passband name and
first-order extinction value, separated by commas:
fixk=V,0.20 sets k=0.20 in V band
fixk=I,0.08 sets k=0.08 in V band
This option can be used to correct for differential
extinction across a frame, while using standards
within the frame to set the zero point.
det_files ... are optional additional names of files with data
on stars detected in other images.
They are assumed to have the same arrangement
of data columns and filter names as the
first file
radius (optional) radius (in arcseconds) to use when
matching stars in catalog against the
detected stars
orphans (optional) what should we do with stars which are
not detected in all passbands? If set to 0,
we discard them from all further processing.
If set to 1, they are not used in making the
solution, but they will have solution applied
to them (without colorterm) so that they
do appear in the output.
outfile (optional) base name of file into which to
write diagnostic output.
Note that all column numbers should be supplied by the user in 0-indexed form -- this, the first column of data in a file is "0", the second "1", and so forth.
The photom program can produce three different forms of solution for a set of measurements.
V = v + a + b*(v - i)
j
where V is the output, calibrated magnitude,
v is the input, raw magnitude in one passband,
a_j is the zero-point offset for frame j,
and b is the color coefficient.
The term (v - i) is the color of a star
in raw magnitudes.
If there are 20 pairs of V,I frames, then the program will solve for 21 unknowns pertaining to the V-band: 20 zeropoint offsets a_j and a single color term b. It will also solve for 21 different unknowns pertaining to the I-band.
V = v + a + b*(v - i) - k*X
j
where V is the output, calibrated magnitude,
v is the input, raw magnitude in one passband,
a_j is the zero-point offset for frame j,
b is the color coefficient,
k is the first-order extinction coefficient provided
by the user, and X the airmass.
The term (v - i) is the color of a star
in raw magnitudes.
One might use this mode with very wide-field images. There are plenty of stars of known magnitude, but differential extinction causes systematic errors for a fixed zeropoint.
V = v + a + b*(v - i) - k*X
where V is the output, calibrated magnitude,
v is the input, raw magnitude in one passband,
a is the single zero-point offset,
b is the color coefficient,
k is the extinction coefficient,
and X is the airmass of each measurement.
The term (v - i) is the color of a star
in raw magnitudes.
If there are 20 pairs of V,I frames, then the program will solve for 3 unknowns pertaining to the V-band: 1 zeropoint offsets a, 1 color term b, and 1 extinction coefficient k. It will also solve for 3 different unknowns pertaining to the I-band.
Here's an example of the input file formats. Lines which start with a pound character "#" will be ignored. One can use input files which have a different number of columns of data, or columns in different order, as long as the command-line arguments to photom correctly describe the required data columns.
# RA Dec B V I 222.41015 1.50449 9.577 7.751 5.973 221.77236 0.97097 6.685 6.702 6.671 221.85685 2.03581 6.928 6.827 6.684 221.75830 1.39323 9.965 8.751 7.553Column 0 contains the RA, column 1 the Dec, column 2 the B-band magnitude, column 3 the V-band magnitude, column 4 the I-band magnitude.
#ID RA Dec JD air band mag err band mag err flag 0 222.41015 1.50449 6542.2334 1.33 V 7.929 0.009 I 5.973 0.010 0 1 221.77236 0.97097 6542.2334 1.34 V 6.705 0.005 I 6.671 0.004 0 2 221.85685 2.03581 6542.2334 1.40 V 6.841 0.006 I 6.684 0.006 1 3 221.75830 1.39323 6542.2334 1.33 V 8.871 0.010 I 7.553 0.012 257In this example, the Right Ascension is in column 1, the Dec in column 2, the Julian Date of the measurement in column 3, the airmass of the measurement in column 4, columns 5, 6, 7 contain the name, mag, and magerr for the first passband (which happens to be V), columns 8, 9, 10 contain the name, mag, and magerr for the second passband (which happens to be I), and column 11 (the last one) holds the "quality flag" for each star.
The "quality flag" contains information on detections of the star in all passbands. Each passband has a value which is the logical OR of
0 no problems
1 star may be saturated
2 star is close to a bad region of the image
4 star is close to an edge of the image
8 star is not detected in this passband
The values for each passband are combined as follows:
flag for passband 0 is not changed flag for passband 1 is shifted left by 8 bits (<< 8) flag for passband 2 is shifted left by 16 bits (<< 16) flag for passband 3 is shifted left by 24 bits (<< 24)The values are then OR'ed together logically. The resulting 32-bit value is the "quality flag" for the star. In the example above, stars 0 and 1 are okay; star 2 was possibly saturated in V-band; and star 3 was possibly saturated in both V-band and I-band images.
Stars which have non-zero flag values are not included in creating the photometric solution for the night.
We could run the program with a command line like this:
photom refcat 0 1 2 V=3 I=4 rawmags 1 2 3 4 11 2 5,6,7 8,9,10 mode=extinct
We create three types of output file:
photom.coeff by default
Jun04.coeff if outfile="Jun04"
photom_V.comp for passband V by default
photom_I.comp for passband I by default
Jun04_V.comp for passband V if outfile="Jun04"
Jun04_I.comp for passband I if outfile="Jun04"
photom.cal by default
Jun04.cal if outfile="Jun04"
Here are examples of the output file formats. Continuing the example used to illustrate the input file formats (using data files with hundreds of stars), we create the following files:
V=V,(V-I) N 656 a 0 18.235 0.012 b -0.091 0.008 RMS 0.025
V=V,(V-I) N 656 a 1 18.240 0.010 b -0.091 0.008 RMS 0.022
I=I,(I-V) N 656 a 0 19.773 0.021 b 0.183 0.012 RMS 0.055
I=I,(I-V) N 656 a 1 19.709 0.024 b 0.183 0.012 RMS 0.059
The V-band solution is based on the (V-I) color.
There are two lines, one for each pair of V/I frames.
The first V-band frame has a zero-point offset of 18.235 mag,
with uncertainty 0.012 mag; the second V-band frame has a
zero-point offset of 18.240 mag, with uncertainty 0.10 mag.
All the V-band frames share a common color term of -0.091,
with uncertainty 0.008.
The RMS between the magnitudes of stars in the reference catalog
and the calibrated magnitudes of matching measured stars
in the first V-band frame is 0.025 magnitudes, and 0.022 mag
in the second V-band frame.
Note that the I-band solution is based on the (I-V) color, not the (V-I) color!
The numbers in the solution shown above are not those actually produced by running the program on the input files I used to illustrate the file formats. Those were faked-up files in which the input and output values were nearly identical, and so the offset, color term and RMS were all perfectly zero. The values shown above are more likely to resemble those produced with real observations.
0 3 221.7583 1.3932 1.000 8.871 7.553 1.318
8.751 -0.120 8.751 0.010 -0.000
The columns show
34 221.6462 3.0349 100.0 V 9.080 0.010 0 I 7.397 0.010 0
3 221.7583 1.3932 100.0 V 8.751 0.010 0 I 7.553 0.010 0
25 221.7712 4.4628 100.0 V 8.778 0.010 0 I 8.296 0.010 0
126 221.6983 3.1238 100.0 V 99.000 99.000 8 I 12.873 0.185 0
The columns show
The flag column has value zero to indicate no known problems, and non-zero to indicate a possible problem. See the section about input-file format for the meaning of bits in this flag.
Note that the final star in the list above was detected only in the I-band. The photom program must have been run with command-line option orphans=1 for this star to have been kept in the output.
Last modified March 27, 2005, by MWR.