2.1 Types of WFC3 Files

Science data taken in orbit by WFC3 are received from the Space Telescope Data Capture Facility at Goddard Space Flight Center and sent to the STScI Operational Pipeline Unified System (OPUS) pipeline, where the data are unpacked, keyword values are extracted from the telemetry stream, and the science data are reformatted and repackaged into raw (uncalibrated) FITS files by the Generic Conversion process (see Section 1.1.1 of the Introduction to the HST Data Handbooks). All WFC3 science data products are two-dimensional images stored in Multi-Extension FITS format files. For each exposure taken with WFC3, there is one FITS file with a unique 9-character rootname followed by a 3-character suffix: rootname_xxx.fits. The rootname identifies the observation and the suffix denotes what type of file it is (see Chapter 5 of the Introduction to the HST Data Handbooks for more details on HST file names).

WFC3 data files are given the following definitions:

• An exposure is a single multi-extension FITS file, the atomic unit of HST data.

• A dataset is a collection of files having a common rootname.
• An association is a set of exposures that can be meaningfully combined into a single product (e.g. a set of dithered exposures, or a series of CR-SPLIT exposures)
• A sub-product is a dataset created by combining a subset of the exposures in an association.
• A product is a dataset created by combining sub-products of an association.

2.1.1 Data Files and Suffixes

The suffixes used for WFC3 raw and calibrated data products are described in Table 2.1 and closely mimic those used by ACS and NICMOS.

Table 2.1: WFC3 data file suffixes.

The initial input files to the calibration pipeline calwf3 are the raw files (RAW) from Generic Conversion and the association (ASN) table, if applicable, for the complete observation set.

For UVIS images, temporary files are created once bias levels are subtracted and the overscan regions are trimmed. These temporary files have suffixes which also relate whether they have been corrected for Charge Transfer Efficiency (CTE) trailing. Files which have been corrected for CTE effects have a "c" as the third letter in their suffix. For example, "_blv_tmp" files have completed the BIASCORR stage, but were not corrected for CTE like the "_blc_tmp" files. If the save option is specified when calwf3 is called on a "_raw" image, then an intermediate "_rac" file will also be saved. This contains the original data with just the CTE correction applied.

The final output files will be renamed with the "_flt" or "_flc" suffixes after the standard calibrations (flat fielding, dark subtraction, etc.) are complete. The "_blv_tmp" files serve as input for cosmic ray rejection, if required. For UVIS CR-SPLIT and REPEAT-OBS exposures, a temporary CR-combined image is created (cr*_tmp) and then renamed with either the "_crj" or "_crc" suffix once basic calibrations of that image are complete. With the addition of the CTE correction, which is the very first step, the pipeline (as of version 3.3) follows two distinct lines of processing, one with the CTE correction applied and one without. Users running raw data through calwf3, which has PCTECORR set to PERFORM, should expect to see both sets of calibrated files when the tasks are complete.

For the IR detector, an intermediate MultiAccum (ima) file is the result after all calibrations are applied (dark subtraction, linearity correction, flat fielding, etc.) to an individual readout of an IR exposure. A final step in calwf3 processing of IR exposures produces a combined, CR-cleaned, image from the individual readouts, which is stored in an "_flt" product file.

The final products of the calwf3 pipeline are not corrected for geometric distortion. AstroDrizzle is used to correct all WFC3 images for geometric distortion, whether they are taken as single exposures or as part of an association. AstroDrizzle is distributed as part of the Drizzlepac package, used for aligning and combining all HST images. For CR-SPLIT and REPEAT-OBS, AstroDrizzle supersedes the calwf3 cosmic-ray rejection processing and uses the individual "_flt" or "_flc" files directly as input, performing cosmic-ray rejection in the process of producing the final drizzled image from multiple exposures (see Table 2.2). This has significant advantages in cases where small numbers of CR-SPLIT images were obtained at a small number of different dither positions, because AstroDrizzle will use all the information from all the input files to produce the best cosmic-ray rejection. The resulting drizzled images should generally be useful for science, although subsequent reprocessing off-line may be desirable in some cases to optimize the data for specific scientific applications.

Table 2.2: The calwf3 and AstroDrizzle input and output products.

2.1.2 Auxiliary Data Files

Association Tables (asn)

Association tables are useful for keeping track of the complex set of relationships that can exist between exposures taken with WFC3, especially with REPEAT-OBS, CR-SPLIT, and dithered exposures. Images taken at a given dither position may be additionally CR-SPLIT into multiple exposures (e.g., UVIS observations). In these cases, associations are built to describe how each exposure relates to the desired final product. As a result, WFC3 association tables can be used to create one or more science products from the input exposures. The relationships defined in the association tables determine how far through the calibration pipeline the exposures are processed and when the calibrated exposures get combined into sub-products for further calibration.

An association file has a single extension that is a binary FITS table. That table has three columns:

• MEMNAME, the member name
• MEMTYPE, the role which that member plays in the association
• MEMPRSNT, a boolean value for whether the member is present or not.

The different MEMTYPE values are summarized in Table 2.3.

Table 2.3: Exposure types in WFC3 associations. The suffix "n" is appended to the MEMTYPE to denote multiple sets are present within a single association.

EXP-CRJ

EXP-CRn

PROD-CRJ

PROD-CRn

EXP-RPT

EXP-RPn

PROD-RPT

PROD-RPn

EXP-DTH

PROD-DTH

A sample association table for a two-position dithered observation with CR-SPLIT=2 is presented in Table 2.4 This example shows how both MEMNAME and MEMTYPE are used to associate input and output products. The MEMTYPE for each component of the first CR-SPLIT exposure, IxxxxxECQ and IxxxxxEGQ, are given the type EXP-CR1. The sub-product Ixxxxx011 is designated in the table with a MEMTYPE of PROD-CR1. The last digit of the product filename corresponds to the output product number in the MEMTYPE. A designation of zero for the last digit in the filename is reserved for the dither-combined product.

The column MEMPRSNT indicates whether a given file already exists. For example, if cosmic ray rejection has not yet been performed by calwf3, the PROD-CRn files will have a MEMPRSNT value of "no". The sample association table in Table 2.4 shows the values of MEMPRSNT prior to calwf3 processing.

Table 2.4: Sample association table ixxxxx010_asn.

Table 2.5: Extensions in UVIS _flt or _flc files after standard processing with calwf3 and AstroDrizzle.

Note 1: All UVIS filters except for the quad filters have a calibrated filter-dependent distortion correction and associated NPOLFILE.  Images produced by the calwf3 pipeline for these 52 filters have 15 extensions; images for the quad filters will have only 11 extensions (they have no WCSDVARR). 

Note 2: With the release in Dec 2019 of improved astrometric corrections for WFC3, additional "headerlet" extensions, which encapsulate WCS information, will be appended to the flt/flc files. The astrometric improvements primarily reduce the pointing errors present in WFC3 data (generally a few tenths of an arcsecond), but do NOT affect the distortion solution. Several astrometric solutions may be available for a given dataset, each stored in its own headerlet. Products retrieved from the MAST archive will have the best solution applied by default. For details on absolute astrometry in HST images, see Section 4.5 of the DrizzlePac Handbook.
 
In order to create a geometrically correct, drizzle-combined product, PROD-DTH exposures are combined only with AstroDrizzle, which executes after calwf3 has finished processing all members. PROD-RPT and PROD-CRJ products are combined using wf3rej and all output files have the “cr” extension.

Support Files (spt)

The support files contain information about the observation and engineering data from the instrument and spacecraft that was recorded at the end of the observation. A support file can have multiple FITS image extensions within the same file. Each extension holds an integer (16-bit) image containing the data that populates the *_spt.fits header keyword values.

Trailer Files (trl)

Each task used by calwf3 creates messages during processing that describe the progress of the calibration and are sent to STDOUT. In calibration pipelines written for other HST instruments, trailer files were created by simply redirecting the STDOUT to a file. Because multiple output files can be produced in a single run of calwf3, creating trailer files presents a unique challenge. Each task within calwf3 must decide which trailer file should be appended with comments and automatically open, populate, and close each trailer file.

Note that calwf3 will always overwrite information in trailer files from previous runs of calwf3 while preserving any comments generated by Generic Conversion. This ensures that the trailer files accurately reflect the most recent processing performed. The string "CALWF3BEG" will mark the first comment added to the trailer file. If a trailer file already exists, calwf3 will search for this string to determine where to append processing comments. If it is not found, the string will be written at the end of the file and all comments will follow. Thus any comments from previous processing are overwritten and only the most current calibrations are recorded.

As each image is processed, an accompanying trailer file with the "*_trl.fits" suffix will be created. Further processing with calwf3 will concatenate all trailer files associated with an output product into a single file. Additional messages will then be appended to this concatenated file. This duplicates some information across multiple trailer files but ensures that for any product processed within the pipeline, the trailer file will contain processing comments from all the input files.

Linking trailer files together can result in multiple occurrences of the "CALWF3BEG" string. Only the first, however, determines where calwf3 will begin overwriting comments if an observation is reprocessed.

Jitter Files (jit/jif)

Jitter files are Observation Log Files that record pointing, jitter, and other Pointing Control System (PCS) data taken during an HST observation. The _jif.fits file is a FITS image of the jitter during the exposure; the associated FITS header contains information about file structure, observation details, modeled background light, point control system, jitter summary, orbital geometry, and problem flags and warnings. The _jit.fits file accompanies the _jif.fits file, and contains reconstructed pointing, guide star coordinates, derived jitter at the science instrument aperture, pertinent guiding-related flags, orbital data (e.g., latitude, longitude, limb angle, magnitude field values, etc.) and fine guidance sensor flags. Please refer to Section 5.1 and 5.2 of the Hubble Space Telescope Data Handbooks for more information.

2.1.3 Hubble Advanced Products

In late-2020, the HST data calibration and archive pipelines began producing new Hubble Advanced Products (HAP) to be distributed through MAST. These include two new types (or levels) of products for a given dataset, referred to as 'Single Visit Mosaics' (SVMs) and 'Multi Visit Mosaics' (MVMs), available for download from either the MAST Discovery Portal or the MAST HST Search Interface.  An example of the file naming convention for standard (HST) and advanced (HAP) products is shown in Table 2.6.

SVMs comprise the data from a single HST visit which are aligned to a common astrometric reference frame and then drizzled onto the same north-up pixel grid. Compared to standard HST data products, these include improved relative alignment across different exposures (and/or filters) acquired within the same visit, enabling easy comparison of the drizzled products. When possible, sources in the HST images have been aligned directly to an external reference catalog (e.g. Gaia eDR3, GSC v2.4.2, or 2MASS) and include improved absolute astrometry in the image World Coordinate System (WCS), carried as HDRLET extensions in the FITS files (see Table 2.5).  SVM data products with both relative alignment (by filter) and absolute alignment to Gaia will contain the string 'FIT_SVM_GAIAeDR3' in the 'WCSNAME' keyword in the science extension of the image header.  The SVM drizzled images are used to generate point source and segment catalogs during pipeline processing. These catalogs supersede those produced by the Hubble Legacy Archive and will be the basis of the next version of the Hubble Source Catalog.  Availability of these SVM data products was announced in the December 2020 MAST Newsletter article, and more details on the software used to compute these data products is available in the DrizzlePac documentation for Single Visit Mosaic Processing.

In mid-2022, MAST began distributing a new type of Hubble Advanced Product (HAP) referred to as Multi-Visit Mosaics (MVM). These are cross-visit, cross-proposal mosaics, which combine public observations of fields observed multiple times by ACS/WFC, WFC3/UVIS, or WFC3/IR.  MVM data products combine all exposures falling within a pre-defined 0.2° x 0.2° ‘sky cell’ for each detector+filter drizzled onto a common, pre-defined pixel grid.  Note that the MVMs are large files (up to 1TB, depending on the number of input exposures in each sky cell). Because they combine observations acquired over a range of dates, the MVMs may have photometric errors of several percent or systematic alignment errors when combining visits with different catalog solutions. These products are therefore recommend to be used as ‘discovery images’ for comparing observations in different detectors and passbands and not for precise photometry.  Availability of the MVM products was announced in the May 2022 MAST Newsletter article. More detail on the software used to compute these data products is available in the DrizzlePac documentation for Multi-Visit Mosaic Processing.  

For statistics on the alignment for each HST detector, uncertainties in the WCS solutions when aligning to different reference catalogs, and examples of SVM and MVM data products, see 'Improved Absolute Astrometry for ACS and WFC3 Data Products' (WFC3 ISR 2022-06).

Table 2.6: MAST standard data products (HST) and Hubble Advanced Products (HAP) available for a sample dataset 'idkl0l020' taken in the F140W filter.