In combining the WFPC2 chips into a single mosaic, the relative translational and rotational offsets between the chips were assumed to be those given by the Holtzman et al. IDT report. No further check of these relative offsets has been made at the present time. Because the relative positions of the WFPC2 CCDs is known to drift very slightly with time, it is likely that these mosaics are not exactly correct, astrometrically, but it is believed that they should be quite close (approximately 0.1 arcsecond translational error or better).
Although the HDF v2 images were approximately sky subtracted during the reduction process, small residual offsets between the sky levels remained. Before the mosaics were generated, these offsets were removed by measuring and subtracting constant "global" modal sky values from the images from each chip in each bandpass, to ensure that there would be no substantial discontinuities at the interchip borders.
The photometric zeropoints of the four WFPC2 detectors differ slightly from one another (approximately 3% minimum to maximum variation from chip to chip). In constructing the mosaics, these differences have been removed by multiplicatively scaling the WF2, WF3 and WF4 images to match the zeropoint of the PC1. The resulting zeropoints of the mosaics, then, are:
In addition to the mosaic images themselves, we provide the inverse variance map images which are output by the drizzling process. These provide an approximate measure of 1 / the variance (from sky, dark current and detector readout noise only) at each pixel position.
The World Coordinate System (WCS) of these mosaics has been updated to match the system defined in Williams et al. 1996 (AJ 112, 1335) based on radio interferometry positions for several HDF sources.
For convenience, we also provide rebinned 1024x1024 versions of the mosaics. These have been rebinned with the "wba" (weighted block averaging) routine which is provided with the HDF v2 release. These 1024x1024 images are also matched to the scale of the infrared KPNO HDF/IRIM images .
Mark Dickinson: email@example.com 10/24/96 Richard Hook: firstname.lastname@example.org 10/24/96