PSF Calibrations

In most cases for processing the data you will need an estimate of the point spread function achieved during the observation. For example, stellar photometry in a crowded field will require the PSF for PSF-fitting photometry or, at very minimum, knowledge of what is often a very large aperture correction. For classically-scheduled observations, all necessary calibration data are the responsibility of the observer.

Issues:

By nature the point spread function (PSF) of an adaptive optics system is complex and varying with time and position in the field. The size and time scale of these variations depends on the conditions at the time of the observation. For information on the effect of this PSF variation on data reductions check out the Data Reduction page.

Obtaining a PSF:

A PSF may be extracted from the observed fields if the field contains an isolated point source or if the field contains a number of known point sources which can be combined to give an average PSF. Potential problems are:

• Will the PSF source be saturated? (e.g. the Hokupaa wavefront reference source may be in the field of view but saturated.)
• Will the PSF source be bright enough to obtain an appropriate signal-to-noise ratio?
• Is the field around the PSF source free of other objects? The PSF in an AO image contains a extended halo which is comparable to the uncorrected seeing disk. As such, considerable energy is found at distances >0.5 arcsec. In some cases such as stellar fields it may be possible to extract the PSF from the science observations by combining several point sources within the field. In other cases (e.g. a faint point source near the guide star) this is not possible.
• Are the field variations of the PSF important for the science result?

Separate PSF calibration fields:

If the observed field does not contain suitable PSF sources, then the observer is forced to observe a separate PSF calibration field. In choosing fields for PSF calibrations take the following into consideration (in order of priority):

• By far the most important criteria in choosing a PSF calibrator is that the field must have similar AOWFS/PWFS guide stars (photon flux, colors, and offsets). Note that the magnitudes from stellar catalogs at these typical brightnesses have large uncertainities. You should be prepared with several candidate PSF calibration fields per object. N.B. When you work under less than ideal conditions (e.g. faint guide star, short wavelength, etc.) this is especially important. We have found that it often takes observations of several stars before a good match is found. We have found the GSC-II and USNO catalogs useful in generating lists of potential PSF calibration fields.
• The fields must have appropriate (e.g. isolated, etc.) PSF calibration stars in the field at similar offset relative to guide star. For example, if the science object is located 20" away from its guide star, then the PSF field should be chosen with a similar separation. Ideally the guide star is matched in brightness to the science object's guide star and the off-axis PSF calibration star is bright enough to give a good SNR in a reasonable amount of time.
• The field must be nearby on the sky to the science target (observed at a similar air mass)
• Observations of the science field and the PSF calibration field must be collected over a timescale over which the seeing is relatively constant (~15-30minutes). This can become a substantial overhead for the observing program.

The usefulness of separate PSF calibration fields varies depending on the above considerations and on the stability of the atmosphere during the observations. We have observed large variations in the seeing (and hence performance of Hokupaa) on timescales of minutes but have also observed times of stability which last for hours.

For science targets where the guide star is an extended source, separate PSF calibration field should be treated with care. The correction obtained on a extended source depends strongly on the distribution of the flux. As such the PSF calibration field guide star will rarely mimic performance obtained on the science field. See the Data Reduction page for ideas on how to handle extended object PSFs.

Fields with single PSFs appropriate for PSF calibrations can be found by searching star catalogs for nearby stars with similar magnitude and color to the science field wavefront reference source. The USNO and Guide Star Catalog II may be useful for this purpose. Potential pit falls here are: (1) When you observe this field, don't saturate the PSF calibration source, (2) Make sure you have a list of 2-3 possible PSF calibration fields per science field as these stars often turn out to be binaries/double stars! (3) The short-exposure PSF (total integration time ~ a few sec) can be significantly different than a longer exposure. Make sure that the total integration time on the PSF calibration is at least a 2 minutes. You may also need to integrate to a particular SNR in the wings of the PSF.

We are developing a library of PSF stars from previous observations. Check these out on the Hokupa'a PSF Library.

Last update 11 February 2001; Mark Chun