WFPC2 Astrometry FAQs

Q: How can I measure accurate relative positions from WFPC2 images?

A: The IRAF task xy2rd can be used to transform pixel coordinates into RA, Dec. If high precision (better than 0.5 arcsec) is desired, we recommend using the task METRIC which accounts for known geometric distortions in the pixel coordinate system. Note that this will still offer an accuracy of about 0.5 arcsec, due to the uncertainties in the actual coordinates of the guide stars, but the relative precision of the position of different points in different chips should be better than 0.1''.

One possibility for your image alignment is to use the task wmosaic, which will rebin the HST image to a common frame for all four chips, removing known geometric distortions and relative rotations of the chips. This image should not be used for accurate photometry or astrometry, but can be used profitably for cross-identification and other, non-critical tasks. Please note that wmosaic will degrade the PC images severely, since they will be resampled at WFC resolution.

Note that xy2rd, metric, and wmosaic (like all IRAF/STSDAS tasks for WFPC2) were designed for use with WFPC2 images in GEIS format. Consequently, WFPC2 images retrieved from MAST in FITS format must first be converted to GEIS format using the STSDAS task strfits before applying xy2rd, metric, or wmosaic. Using these tasks on FITS images will cause them to fail or yield incorrect results. Note also that the multigroup FITS images can be displayed in FITS viewers like SAOImage DS9, but the celestial coordinates of the image pixels may not be computed correctly from the WCS information in the FITS headers.

Q: I need to determine orientations (within measurement errors). Is the ORIENTAT keyword in the header a global value or specific to each chip?

A: The ORIENTAT keyword in the header accounts for all the inter-chip rotations. It is distinct for each individual chip.

Q: How can I correct the geometric distorsions due to the F160BW and F130LP UV filters in the WFPC2 images? Could you please provide the up-to-date cubic distortion coefficients, as described in "The performance and calibration of WFPC2 on the HST" (Holtzman et al., PASP 1995) for those two specific filters?

A: We do not have specific solutions for the geometric distortion of WFPC2 for individual filters. As you know, the cubic solution published by Holtzman et al was derived for F555W, but it should be reasonably good for all filters. The shifts (coefficients C1, D1) and rotations (C2-C3 and D2-D3, expect possibly for an overall plate scale factor) are expected to apply exactly, independent of wavelength. The plate scale and the geometric distortion within each chip could have a small wavelength dependence due to the field flattener, which is a refractive element. We plan to derive geometric solutions for three filters (F300W, F555W, and F814W) from observations to be taken after the refurbishment mission.

If you would like to estimate the possible wavelength dependence of the geometric distortion, you might want to use the solution derived by Trauger, et al., and published in the Proceedings of the workshop "Calibrating HST: Post Servicing Mission" held at STScI in May 1995. This solution is based on ray tracing, rather than on measurements of stars, and has an explicit wavelength term.

Q: I have an error in my WPFC2 header which is causing a position error.

A: Prior to April 19, 1994, there were errors in the WFPC2 headers which resulted in incorrect positions by about 3". A task in stsdas, uchcoord, can be used to fix the headers.

Note that if you retrieve data from the HST archive at the Space Telescope Science Institute, the headers will be correct (you will not need to run uchcoord on these images). OTFR (On The Fly Reprocessing) will ensure that any updates to the plate scale, cd matrix, etc., even those that occurred after the date given above, will be correctly applied.