Bandmerge will always produce three (mandatory) output files. They will contain the bandmerged data, the spectral combination counts (QA) and the statistical summary. Other output files (e.g. the Dump files) are optionally produced.
4.1 Bandmerged Data Output File
The bandmerged data output file will include the following header keywords:
\ BANDMERGE Vsn 1.39 A50707 run on 07/01/08 at 15:02:28
Nbands: the number of bands that have been merged,
Total_Merged_Number: the number of merged point sources (Spitzer data only),
NSrc: the number of merged sources (for 2MASS),
Infile-#: the (path and) file names of the input source detection tables,
N_Apertures-#: the number of aperture#bad_pix# columns included (Spitzer only),
Sort_&_Search_Direction: which axis the output bandmerged sources are sorted on (Spitzer only),
Outband-#: the output band numbers (see §3.1) and the corresponding input files (Spitzer only).
The Bandmerged output file will include the following columns:
IDNum: a unique identification number for each bandmerged point source
X, Y, Delta_X, Delta_Y: the refined coordinates with the corresponding uncertainties
CnfFlag: the confusion flags; the character field has a variable width of NBands*(NBands 1) where NBands is the number of bands being merged; thus the width varies from 2 42 characters. If only 2 bands are being merged, the column header is abbreviated to CF. The field is composed of NBands sub-fields of width NBands 1 each. The Nth subfield gives the number of sources in each band that passed the fine position test (see §2.1), in increasing order of band and excluding band N itself. The subfields are also given in increasing order of band. If more than 9 sources in a candidate band passed the fine position test, the symbol X is used. For example, assuming we are merging 3 bands (IRAC-3.6, IRAC-4.5, IRAC-5.8):
If CnfFlag = 112132:
11: the lowest band (IRAC-3.6) found one candidate each in bands 2 and 3 that passed the fine position test.
21: the middle band, band 2 (IRAC-4.5), found 2 candidates that matched in band 1, and 1 candidate in band 3.
32: the highest band, band 3 (IRAC-5.8), found three matching candidates in band 1, and two in band 2.
srcid, flux, deltflux, aperture#, bad_pix# and any optional input columns: all these columns are listed verbatim from the input files, one band at a time. They are given the band number as a suffix.
4.2 Spectral Combination File
The output (QA) spectral combination counts file consists of a table giving all possible spectral combinations and the corresponding number of point sources that were found during bandmerging. An example is given here:
Sp.Comb. Count
======== =====
001 33424
010 5982
011 9521
100 165
101 192
110 146
111 1509
===============
Total: 50939
The information in the above table is interpreted as follows:
There were 33424 point sources in the first (lowest wavelength) input source detection table that were not merged with any other band.
There were 5982 point sources in the second input source detection table that were not merged with any other band.
There were 9521 sources from the first two bands that were merged with each other, but not band 3.
There were 165 sources from the third input source detection table that were not merged with any other band.
There were 192 sources that were merged between bands 1 and 3, but not with band 2.
There were 146 sources merged between bands 2 and 3, but not band 1.
There were 1509 sources merged in all three bands.
4.3 Statistical Summary File
This file is produced for Spitzer data only. It has the following format:
There are two mandatory header lines:
\ BANDMERGE Vsn 1.39 A50707 run on 07/01/08 at 15:02:28
\CHAR BMG_Output_file = Output_1/newtbl.tbl
where BMG_Output_File indicates the path and file name of the bandmerged output file that corresponds to this statistics file.
This is followed by an ASCII table with the following columns:
B-P column: the two bands that are given on the current table line (e.g. 1-2, 2-4 etc.)
SNR_bin column: which S/N ration bin is given on the current table line, e.g. 10-20, >1000. Total, etc.
If the SNR column is not specified in the namelist file then no binning of statistics will be done, i.e. only the Total category will be computed and output. Furthermore, if SNR is present but a given detection has a null value, then that case will be omitted from the statistical computation, no matter how many other bands in the merged source have non-null SNR values (i.e. if any band for a given source has a null SNR then the entire bandmerged point source will be ignored during the statistics processing).
4.4 Dump Output Files
A dump output file will be produced after each completion of the cross-band linkage processing step (see §2.1.1) if the parameter Dump is included in the namelist file. There will be one dump file per input source detection table. The file output depends on the value of Dump (= 1 or 2). If Dump = 0 (default) then no dump file is produced. In both cases 1 and 2, the table is preceded by a one-line header:
\INT Band = <b_value>
where <b_value> = the primary band in this file (a.k.a. the seed band; the bands are numbered from 1-7, in the order of their band index).
If <d_value> = 1 then the output table will have the following format:
| srcid | pbnd | ptrid | chisq | pschisq |
| int | int | int | real | real |
where
srcid is the source identification of the seed (sorted in ascending order).
pbnd is the band of the candidate being pointed to (1-7),
ptrid is the source identification of the candidate (in pbnd),
chisq is the chi-squared value for othe seed-candidate combination,
pschisq is the pseudo-chi-square value for the seed-candidate combination (sorted in ascending order for each srcid).
If <dvalue> = 2 then the output file has the following format:
