NAME

niriinfo -- Reduction scripts for NIRI data

USAGE

niriinfo

DESCRIPTION

The NIRI package contains tasks for processing NIRI imaging and long-slit spectroscopy data, although spectroscopy reduction will also use tasks in the GNIRS package.

This document describes the tasks specific to the NIRI package and gives a description of the NIRI data format. For information on the GNIRS package, see gnirsinfo (since the GNIRS tasks were derived from the NIRI codebase, they should be familiar to experienced NIRI users).

The tasks produce logfiles of the performed processing steps. The name of the logfile may be set in each individual task, or at the package level by setting niri.logfile.

The tasks add header keywords to the output images. These header keywords contain information about the performed processing steps and the values of the critical parameters that were used.

All NIRI images are written as multi-extension FITS (MEF) files. Raw data have a single unnamed extension [1]. Most of the header information is written to the primary header unit [0] (PHU). After being run through nprepare, NIRI data extensions will be named as defined by the parameters. Several additional header keywords will be added. If desired, variance and data quality planes can be added by nprepare.

It is recommended to use imtype="fits". This is set automatically when loading the GEMINI package.

SHORT DESCRIPTIONS OF THE TASKS

NPREPARE - Prepare raw data and create variance and data quality planes.

All data must be run through NPREPARE before further processing. NPREPARE adds certain essential header keywords to the PHU. It also computes the variance and data quality planes, if desired. The data file nprepare.dat contains information about the array characteristics and the regions of the array illuminated by the spectroscopic slits. Although changes are infrequent, it is important that nprepare.dat be up to date. NPREPARE fixes the WCS information in the header for subarray data taken before Feb. 2005.

This task must be used before processing the data with tasks in the GNIRS package, too.

NRESIDUAL - Flag saturated pixels in the DQ planes of subsequent
exposures

When there are saturated or nearly-saturated objects in a frame, image persistence in the array leaves a faint residual that can be confused with a real object in subsequent images. NRESIDUAL takes the DQ planes from the images taken immediately prior to an image and combines them, allowing the effects of saturation to be followed in subsequent exposures. This task is not needed if there are no saturated or non-linear pixels in the exposures.

NIFASTSKY - Derive sky image, median or min/max filtering

NIFASTSKY is used for constructing sky images by median or min/max filtering. There is made no attempt to specifically flag objects in the sky images. The task is mostly used for rapid reductions at the telescope.

NISKY - Derive sky image, includes masking of objects

NISKY is used for constructing sky images. Objects in the input images are identified and masked using OBJMASKS. It is recommended to use NISKY rather than NIFASTSKY for science quality reductions. Because NISKY does a preliminary reduction of each image and identifies objects to mask, it is somewhat slower than NIFASTSKY.

NIREDUCE - Reduce images from NIRI (sky or dark subtract, divide flat)

NIREDUCE is used to sky or dark subtract and flat field divide the science data. The sky image can be scaled before subtraction if desired.

NISUPERSKY - Make improved (deeper) object masks for NISKY

NISUPERSKY takes a stacked image produced using IMCOADD and creates new object masks for use with NISKY. This allows one to mask faint objects that are only detected in the final stacked image. After running NISUPERSKY, one must return to the NISKY step and repeat all steps in the reduction sequence again.

NIFLAT - Derive flat field and bad pixel mask for NIRI images

NIFLAT is used to construct flat field images and bad pixel masks using either GCAL or sky data. For GCAL data, a set of exposures with the shutter closed is combined and subtracted from a set with the shutter open. For sky flats, NIFLAT calls NISKY to mask and remove objects before normalizing. NIFLAT uses the flats and short darks to identify bad pixels.

NIROTATE - de-rotate NIRI images based on WCS information

NIROTATE can be used to correct for image rotation using only the information in header WCS. This is useful mainly for reducing Altair/NIRI AO data taken with the cass rotator fixed in position.

GNIRS.NVNOISE - remove vertical stripes

NVNOISE can be used to independently fit a constant to each of the 32 amplifiers in NIRI or GNIRS. This task helps remove some of the vertical stripe and quadrant bias offset pattern sometimes seen in low-background observations. It is primarily for GNIRS spectra, but can be used with NIRI imaging data in some circumstances.

TYPICAL IMAGING REDUCTIONS

For typical reductions the user will need appropriate flat fields and short dark images, on-target science images and sky images. Flat field images taken with the calibration unit are usually taken both with the IR lamp on and off to allow separation of the instrumental thermal signature from the sensitivity response. Short darks are used to identify bad pixels.

If the target is not extended and the field is uncrowded the on-target images may be used as the sky images. This assumes that sufficiently large dither steps were used during the observations.

1. Use NPREPARE and NIFLAT to derive normalized flat fields and derive a bad pixel mask. NIFLAT will not run if the data have not first been NPREPAREd. Be sure to exclude the first frame in each sequence.

2. Use NPREPARE to update the raw data headers and create the variance and data quality planes of the science data. Use the bad pixel mask created by NIFLAT to generate the DQ planes. The other tasks will not run if the data has not been NPREPAREd.

3. If there are non-linear or saturated pixels in the image, then it is best to run NRESIDUAL. This makes sure that persistence images from saturated objects are tracked in the subsequent DQ planes.

4. Use NISKY (or alternatively NIFASTSKY) to derive sky images. The near-IR sky level and structure varies on time scales of a few minutes, and care should be taken to combine only sky images that are close enough in time to each other and to the relevant science images that such variations are non-significant. This may take some experimentation. Use fl_keepmasks=yes to check the object identification step. Once again, always be sure to check and exclude the first frame from an observing sequence.

5. Use NIREDUCE to subtract the sky images and apply the flat fields. Data from each filter should be processed separately, as NIREDUCE takes only one sky image at the time as its input. With NIRI data, better sky subtraction can often be achieved by scaling the sky image to the same mean level as the science frame using fl_scalesky=yes (assuming the science frame is sky-dominated).

6. If the data were taken with the cass rotator fixed, and there are no objects to register the images, use NIROTATE to de-rotate the images using the WCS information in the headers. (This step will not normally be needed.

7. For low-background imaging or spectroscopy, there are sometimes vertical stripes seen in some quadrants. The GNIRS task NVNOISE is designed to fit a constant offset for each of the 32 amplifiers, and can help remove this pattern. NVNOISE should be run after flat fielding and sky subtraction. This task should not be used unless a problem is apparent.

8. The processed images may be co-added with the task GEMTOOLS.IMCOADD. Photometry may be derived with any suitable photometry package. Once again, be sure to exclude the first bad frame in a sequence.

9. For deep imaging, NISUPERSKY should be run on the output image produced by IMCOADD to create better object masks. The interim steps are then deleted, and the reduction procedure re-started from the NISKY step (step 4 above).

TYPICAL SPECTROSCOPIC REDUCTIONS

For most spectroscopic reductions the user will need appropriate flat fields, arc lamp spectra, and short dark images, on-target science images and possibly sky images. Short darks are used to identify bad pixels. If the target is not extended the on-target images may be used as the sky images. This assumes that sufficiently large dither steps were used during the observations.

1. Use NPREPARE to update the raw data headers and create the variance and data quality planes, if desired. The other tasks will not run if the data has not been NPREPAREd. NPREPARE will write the appropriate SPECSECs to the headers.

2. If there are saturated pixels (such as a setup frame on a bright telluric standard), you should run NRESIDUAL at this point.

3. Process the data with the GNIRS package. For more information see both gnirsinfo (general information) and gnirsinfols (advice specific to longslit spectroscopy).

BUGS AND LIMITATIONS

The tasks in the NIRI package are designed to operate on MEF FITS images that have been processed using NPREPARE. NPREPARE will not run on data from instruments other than NIRI. The NIRI tasks will not run on simple FITS files.

The NIRI spectroscopic tasks have been adapted for use with GNIRS, and are still being tested and debugged. Please use them with caution while we fix the bugs and determine the best way to use the scripts.

SEE ALSO

nprepare, nresidual, niflat, nifastsky, nisky, nireduce, nirotate, nisupersky, gnirs.nvnoise, gnirs.nscut, gnirs.nsflat, gnirs.nsreduce, gnirs.nsstack, gnirs.nscombine, gnirs.nsappwave, gnirs.nswavelength, gnirs.nssdist, gnirs.nstransform, gnirs.nsextract, gnirs.nstelluric, gnirs.gnirsinfo, gnirs.gnirsinfols