This is a summary of the work I did over the summer of 2000 on the PoPSIcL project. I was primarily involved in data reduction; processing and cleaning up the images from the first PoPSIcL project observing run in April. I have also done a small amount of research into the history of Intracluster Light observations and surface photometry.
(IRAF parameter settings in parantheses)
I began the summer by learning to use UNIX and IRAF, indispensible tools for observational astronomers. Later I learned HTML and produced the PoPSIcL Project web page.
The images I was working with were taken by John Feldmeier and Heather Morrison during the week of April 4th, 2000 on the 2.1m Telescope at Kitt Peak National Observatory in Arizona. Four nights of observing yielded zeroes, dome flats, focus frames, twilight flats, standards, skies, and images for four galaxy clusters: Abell 1234, 1413 , 1914 and MKW 7. The first real step in processing the images was to subtract the bias level from all the images and to trim off the overscan region. IRAF's ccdproc made short work of this, and all the frames were left as 2046x2048 images afterwards.
(ccdproc overscan + trim + biassec=image trimsec=image)
My second task was to combine the zero frames into four Zeroes, one for each night. This was done using IRAF's zerocombine, which averaged each pixel set after rejecting the min and max pixel values. Using the four zero frames, all of the other images were zero subtracted with another pass through IRAF's ccdproc.
(zerocombine combine=average reject=minmax)
(ccdproc zerocor=yes zero=zeron1.fits (for example) )
After zero combination and subtraction, the only frames that were of further interest were the dark skies and the images of the galaxy custers themselves. Further data reduction steps ignore the other accessory frames.
With all of the images now leveled, trimmed, and zero subtracted, we were ready for our first attempt at flatfielding. The initial dark sky flats were created using IRAF's flatcombine tool. We made two dark sky flats, one for nights 1 and 2 and another for nights 3 and 4. Of the 23 dark skies from nights one and two, 3 were rejected (the twenty that were used are listed in /bb_data2/sar9/icl/raw/gdsky12.list ). All fifteen skies from nights three and four were used, although at least one was borderline. The now familiar ccdproc was invoked a third time to flatfield the images, night by night with the appropriate dark sky flat. After this preliminary flatfielding, we took a pause from the data reduction to visually inspect the images. I looked over each of the images from all four nights and noted any unusual brightness gradients, out-of-focus frames, or other abnormalities.
(flatcombine)
(ccdproc flatcor +)
Our subsequent flat-fielding attempts utilized an IRAF script provided by Denise Hurley-Keller (/bb_data2/sar9/icl/script.cl). This script used our two initial dark sky flats to produce a better dark sky flat for each night. The script's procedure is explained in more detail on the PoPSIcL project web page. Several iterations of the script yielded the net result of a pair of much improved dark sky flats (one for nights 1 and 2, another for nights 3 and 4). The "raw" images (only bias-leveled, trimmed, and zero-subtracted; not yet flatfielded) were flattened with the new dark sky flats.
(ccdproc flatcor +)
The next step in our image reduction process was to subtract off the sky level from the cluster images. We entertained several possible methods for this important step, and finally settled on the most complex one. I examined each cluster and found two sections in each cluster image that were relatively devoid of stars and galaxies. One section was above the central cD and one was below. Using IRAf's imcopy, I cut these subsections out of each image and ran them through our spikemode program, which intelligently calculates the mode of each image section. I averaged the two subsection modes for each image, called that the sky count, and subtracted it off of the parent image.
(imcopy)
(imarith)
The final stage in my data reduction program was to combine the images cluster by cluster to get four composite images. Before combining, the images had to be shifted so that they shared the same coordinate system. I visually examined each of the cluster images and found 10-12 reference stars in each frame. The coordinates of these reference stars were recorded in a series of files and used with IRAF's geomap and geotran to shift the images to a common central axis.
(geomap)
(geotran)
With the images spatially aligned, IRAF's imcombine tool allowed me to combine them into a composite image. These four combined images are the final product of my summer's worth of image reduction. The images of clusters Abell 1413 and MKW 7 are of primary interest, mostly because they were produced with cleaner data and had more images per cluster for combination. Our images of Abell 1413 reveal what may be an arc or two from gravitational lensing. Both Abell 1413 and MKW 7 have extremely oblate and flattened envelopes and may have a significant amount of diffuse intracluster light.