Chapter 4 of the ACS Data Handbook describes the use and importance of drizzling for ACS data sets. The pipeline previously constructed drz.fits image products using PyDrizzle. The advantages of MultiDrizzle over PyDrizzle are described in Section 4.5 of the ACS Data Handbook. Additional information can also be found on the MultiDrizzle Software Page:

http://stsdas.stsci.edu/multidrizzle/

The combined images produced by MultiDrizzle in the ACS pipeline have a pixel scale that depends on which camera (WFC, HRC or SBC) was used (0.05 arcsec/pixel for the WFC, 0.025 arcsec/pixel for the HRC and SBC). The orientation for all these images is in the default unrotated detector frame. The drizzled images contain three extensions: [sci] being the drizzled science image, [wht] being the weight image, and [ctx] being the context image, showing which exposures contributed to each output pixel. The background sky value is calculated by MultiDrizzle and stored in the image headers of each individual flatfielded file ("*_flt.fits"). The actual science pixel values in the flatfielded files are not modified by MultiDrizzle. However, pixels that have been identified as cosmic rays are flagged in the data quality array using the bit value 4096. The MultiDrizzle image product drz.fits is generally sky-subtracted, with the exception of narrow-band filters and UV filters that are expected to have a dark sky background.

The MultiDrizzle pipeline products are intended to be scientifically useful in all cases where the standard recommendations on dithering have been followed (for example, obtaining at least four exposures of approximately equal length, and using a pre-defined dither pattern or CR-SPLIT as recommended). While the MultiDrizzle products are also created for datasets with one or two exposures, it should be noted that the high rate of incidence of cosmic rays will generally result in a number of pixels that are unrecoverable because they have cosmic rays in all exposures. It should also be noted that the final PSF is slightly broader, since an additional convolution is introduced in drizzling the input images onto the output plane. As described in Chapter 4 of the ACS Data Handbook, observers can run the MultiDrizzle software outside of the HST Archive to recalibrate the data if necessary.

Last modified February 4, 2005 by Anton Koekemoer