Whether you are observing for yourself or another astronomer, you want the images to be as good as possible.
1) Remember Wallerstein's Rule: "Four astronomers, five opinions." If you ask many observers what you "should" do, you will get hopelessly confused. But the advice given below is pretty standard stuff.
What are some useful things to keep in mind to image a given field?
2) Do I have the right coordinates? Do I have the right equinox for those coordinates? Have I entered them into the telescope control system correctly?
If you take an image of, say, a galaxy that has (or had) a supernova in it, and there is no galaxy in the image, you should suspect that you had the wrong coordinates, wrong equinox, or entered the coordinates incorrectly.
This goes too, of course, for a field of standard stars. The most commonly used list of optical standards:
Landolt, 1992, AJ, 104, 340
This contains charts of the fields.
Note that there "Landolt standards" and "stars observed by Landolt". Look at columns 10 and 11 of Table 2 of the Landolt paper. Those are the numbers of observations and the numbers of nights a given star was observed. A field like Rubin 149 is great because it was observed on 30 nights. Other great fields are T Phe, PG0918, PG1323, SA 110 (near star 503), and SA 112-805/822. Why are these so good? Because these fields were observed many times, and you get multiple stars over a range of color. But SA 97-42, for example, all by its lonesome, is not so good because it's a very red star (could be variable), and was observed by Landolt on one night only.
I should mention that the Landolt star PG1323A is variable. It is a W Uma star with a range of at least 0.38 mag in V. But the other stars of the PG1323 field are good as standards.
I have observed the field PG0231+051 many times even though Landolt did not observe it too many times. The blue star (PG0231+051 itself) has a V-I color of -0.348 +/- 0.013, not the value that Landolt gives (-0.534 +/- 0.1221).The last 6 columns of Table 2 of Landolt's paper give the mean errors of the mean of the photometry. This blue star was too faint for him to observe with single channel photometry in a red filter. Hence the large uncertainty.
In a subsequent paper (Landolt, 1992, AJ, 104, 372) there is UBVRI photometry of spectrophotometric standards. These are of comparable brightness to the other Landolt standards, and should be observed if you are taking Z-band observations. A list of Z-band magnitudes of these stars is to be found in Appendix B of Mario Hamuy's Ph. D. thesis.
The charts for the spectroscopic standards are to be found in Stone & Baldwin, 1983, MNRAS, 204, 347.
Note that some of these objects have rather sizeable proper motions. Some of the charts have vectors with tick marks on them. They indicate the position of the star in roughly 1950, 1985 and 2000.
3) It is useful to set up your image display screen by flipping it up/down and/or left/right, so that you can easily compare your charts (usally done north = up, east = right) with the screen.
4) Do I have the right exposures? In my experience, the exposure time calculators usually give you exposure times that are too short to obtain a given signal to noise ratio for faint objects.
For standard stars like V = 12 mag, typical exposure times on the 60-inch telescope are something like U = 40, B = 20, V = 12, R = 8, I = 12, Z = 30 sec. But don't take my word for it.
Especially for the first image in a multi-filter sequence, after verifying that you are looking at the right place on the sky (I usually start with the V filter rather than U), use imexam with the "r" option to make a radial display of the key stars in the field. If you are getting 500 counts maximum, that's not enough, so try longer exposures. If you're getting more than 55,000, that's too much. If the profile has a flat top, that means you've saturated and the measurement can't be used at all. So try a shorter exposure. Of course, a field of standards has stars of different brightness, so there is some optimum exposure to get all of the stars of interest. Adjust accordingly. That's why I start with V and scale from there.
5) At some telescopes, the latest image obtained is not automatically displayed. I try to display all the images I take on a given night. And inspect them. Why? there could be an airplane trail right across my object. Or - there could be a whopping cosmic ray right next to the supernova I am interested in. In those cases you will want to take subsequent images in the same filters.
6) At what airmass should I observe my standards?
If you are trying to calibrate a field that could not be observed very high in the sky, then you should observe standards at that airmass too. For example, if you are observing at CTIO and your field of interest is at +16 Declination, then even when transitting the meridian, it will be at 1.7 air masses. (That's the secant of the zenith angle.) If all of your standards were observed high in they sky (like at 1.1 airmasses), then you will not be able to derive the atmospheric extinction. Using assumed extinction values is only advisable if all of your fields (program and standards) are observed at approximately the same elevation above the horizon. If your standards were observed high in the sky and you assumed extinction values to calibrate fields observed low in the sky, you could end up with serious systematic errors in the photometry, on the order of 0.10 mag. That's no good if you are quoting photometric accuracy of 0.02 mag.
7) If it is definitely not photometric, it generally is a waste of time to observe standards. Standard fields can be used on non-photometric nights to derive color terms for given filters. But without zeropoints, you still need observations on photometric nights.
8) Ideally, your stellar images are 2 pixels wide. This has to do with something called Nyquist's theorem. If your images are sharter than two pixels, you are not extracting all the information available. If you have 5 pixel wide stars, you are spreading the light out over many pixels, and you will be much more limited as to the faintest point source you can observe.
If you display an image with ximtool and use imexam with the "r" or "a" options, you can get the size of the stars in pixels.
If this number is more than 4 pixels, either you didn't focus well, or the focus has changed (and you need to refocus), or it's just a night with crummy seeing. With some telescopes you need different focus offsets for the various filters. Perhaps your B-band images always have stars of FWHM = 4 pixels but in the other filters it is 2 pixels. Then you need to look at your focus offset for the B filter.
9) When taking sky flats, immediately after sunset you can get the U filter. If you're giong to do Z-band photometry, get Z-band sky flats next. That's because a CCD chip is not sensitive at the short and long wavelength limits. By the time you get some Z-band flats, the sky is then dark enough to do B or whichever filter you want to do next.
I myself prefer sky flats over dome flats. But if you are observing in more than 4 filters, there just isn't time to get 5 sky flats in all filters. The sky gets dark too fast.
Sky scattering in the atmosphere is minimum 120 degrees from source of light (like the Sun). So I take my sky flats at a Declination equal to my latitude, and 2 hours east of the meridian. You should offset 10 arcsec or more between sky flat exposures in case the stars are appearing in the frames. Then medians of the multiple images can get rid of the stars.
10) Often you start observing in the west and work your way east. It's good to determine ahead of time when your objects will be beyond 2 airmasses (setting in the west) or when they rise above 2 airmasses in the east. Ideally, everything is observed high in the sky, where the images are sharpest and the stars are brightest, but that's not what happens in real life. Still, you have some control over what order you observe things in.
11) If it is photometric, it is good to observe a decent field of standards at the very start of the night (immediately after focussing), every 90 minutes thereafter, and as the very last field of the night. That way you have the best chance to prove it was photometric all night or to know when it stopped being photometric.
Go back to Kevin Krisciunas home page by clicking here .