All data were retrieved through the MAST HST data archive and the full set of uncalibrated data products and best reference files were requested for each observation.
All data were process using the NICMOS data reduction pipeline NICRED. The NICRED pipeline includes nine processing steps from basic calibration to generating final co-added registered mosaics (See table below). Among these nine steps are procedures for successfully handling three of the most common NICMOS anomalies: electronic ghosting, the pedestal effect and cosmic-ray persistence. Electronic ghosts occur when a bright object is observed in one of the four quadrants on a NICMOS detector. This results in an echo of the bright object in the other three quadrants. The step undopuft attempts to remove these artifacts. The pedestal effect is the result of variable biases in each of the four NICMOS quadrants - leaving a significant pedestal signature in the processed data. The step calped removes this effect. NICMOS observations that occur directly after HST passage through the South Atlantic Anomaly (SAA) suffer from significant cosmic ray persistance. The step saaclean removes these persistent cosmic-ray hits.
| Processing Step | Description |
|---|---|
| undopuft | Removes electronic ghosts (a.k.a. the “Mr. Staypuft” effect) |
| calped | Basic calibration, sky subtraction/pedestal effect removal, cosmic-ray rejection |
| saaclean | Removes cosmic-ray persistence signal if target was observed shortly after the SAA |
| medsub | Removes residual instrument signatures by subtracting a “super-median” reference image |
| flatten | Uses bicubic spline to flatten the background |
| nonlincor | Corrects for count-rate non-linearity |
| align | Uses object matching algorithms to improve image alignment and registration |
| weightmap | Creates accurate RMS maps for use with MultiDrizzle |
| mdrizzle | Creates final CR-cleaned, distortion-free drizzled image mosaics using multidrizzle |
The undopuft step was used to remove the electronic ghosts that can appear when observing a bright source. For details see Electronic Ghosts: Mr. Staypuft, Ringing, and Streaking.
The calped step performed basic instrumental calibration (dark current subtraction, flat fielding, conversion to count rates, and cosmic ray identification and rejection) and attempted to remove the NICMOS pedestal effect. These task were performed by the STSCI IRAF package tasks calnica and pedsky.
The NICMOS pedestal effect is the result of variable biases in each of the four NICMOS detector quadrants these varying bias levels can leave a significant pedestal signature in the processed data. For details see the NICMOS anomaly page Residual Bias (Pedestal).
NICRED runs all of the calibration steps provided by calnica in the default sequence with the exception of one additional step. Before the calnica cosmic ray identification and removal step CRIDCALC is run NICRED runs an additional step to improve the cosmic ray rejection. For NICMOS MultiAccum mode observations, CRIDCALC assumes that accumulating background counts over the entire observation is a linear function. This assumption may not be the true for all observations. Depending on circumstances of the observation the background count rate may vary over the duration of the observation. In order to determine if the background count rate is sufficiently non-linear, NICRED computes the median of the first and last three readouts of the MultiAccum observation. If the NIRCED finds the count rate has varied it applies the additional step of running pedsky on each of the individual readouts in the MultiAccum observation. This additional step assures the background count rate is linear before running the CRIDCALC step.
The saaclean step was used to remove cosmic ray persistence due to observations following an HST transit of the South Atlantic Anomaly (SAA). See Removing Post-SAA Persistence in NICMOS Data. NICRED uses the PyRAF task saaclean to perform this processing.
The medsub step was used to remove any residual instrument signature left over after basic calibration by subtracting a “super-median” reference image. These super-median images are created by median stacking a large number of images that have been processed by the NICRED modules undopuft, calped and saaclean. Many super-median reference images (based on various camera, sample-sequence, observation window or HST proposal ID, and filter combinations) were generated for this project and are provided in nicred_reffiles.tgz.
medsub step used the following criteria for determining which super-median image to use:
The flatten step was used to remove any discontinuities between the four quadrants of the NICMOS camera 3 image. Discontinuities between quadrants can occur when an exposure contains a large bright object in one of the quadrants.
The nonlincor step corrects NICMOS images for their count-rate dependent non-linearity. It used the header keywords CAMERA and FILTER to determine the non-linearity parameter. It corrects the first image, and in the case of a multi-extension image, the second image as well, with the appropriate power law. For details see Correcting the NICMOS count-rate dependent non-linearity.
The align step used the NICRED external package superalign to determine the internal shifts and rotations for the final mosaic images. The superalign package was used because it offers several useful advantages relative to other alignment programs:
- It does not require that all images be contiguous with a single reference image. This allows one to construct arbitrarily large mosaics out of individual images.
- Input catalogs can include substantial (>80%) contamination from cosmic rays.
For more details on superalign see the NICRED documentation appendix section superalign.
The weightmap module generates an inverse variance weigh map image of each input image as input to MultiDrizzle.
See the NICRED weightmap module documentation.
The mdrizzle step was used to perform the cosmic ray rejection and combination of dithered observations using the STScI software package MultiDrizzle. For a complete discussion of MultiDrizzle and the Drizzle alorgithm for combining dithered imaging data see the MultiDrizzle Handbook Wiki.