The CalFUSE Pipeline Reference Guide

Version 2.1, 21 March 2008
Edited by Van Dixon
This document describes CalFUSE v3.2.3.

Note: "CalFUSE Version 3: A Data-Reduction Pipeline for the Far Ultraviolet Spectroscopic Explorer" (Dixon et al. 2007) is the definitive documentation for the FUSE data-reduction pipeline. The CalFUSE Pipeline Reference Guide supplements that paper by providing instructions for running the pipeline and detailed descriptions of its calibration files.

1. Installing the CalFUSE pipeline

1.1 Retrieving the Pipeline
1.2 Installing the Pipeline
1.3 Directory Structure
1.4 Setting Environment Variables

2. Running the CalFUSE pipeline

2.1 Overview of the Pipeline
2.2 Modifying the Default Parameters
2.3 Combining Data from Multiple Exposures
2.4 Detector Distortions in the X Dimension
2.5 FUSE Effective-Area Curves

3. CalFUSE File Formats

3.1 Data Files
3.2 Housekeeping and Jitter Files
3.3 Parameter Files
3.4 Calibration Files

1. Installing the CalFUSE Pipeline

1.1 Retrieving the Pipeline

You may download the files directly from our FTP site using your browser or transfer them via FTP.

   prompt> ftp archive.stsci.edu
          login: anonymous
          password: your_email_address
   ftp> cd pub/fuse/software/calfuse
   ftp> binary
   ftp> get INSTALLING_CalFUSEv3.2.3
   ftp> get INTRO_TO_CalFUSEv3.2.3
   ftp> get cfv3.2.3.sw.tar.gz
   ftp> get cfv3.2.3.cala.tar.gz
   ftp> get cfv3.2.3.calg.tar.gz
   ftp> get cfv3.2.3.calz.tar.gz
   ftp> quit

Uncompress and extract all tarfiles into your calfuse directory.

1.2 Installing the Pipeline

To install and compile CalFUSE, please follow the instructions in the file INSTALLING_CalFUSEv3.2.3.

1.3 Directory Structure

Once you have installed and compiled CalFUSE, you will have the directory structure shown below.

calfuse/v3.2/

   bin/
   calfiles/
       aeff1a100.fit
       aeff1b100.fit
       etc.
   docs/
   idl/
   include/
   lib/
   parmfiles/
       master_calib_file.dat
       parm1a015.fit
       scrn1a015.fit
       etc.
   src/
       Makefile
       analysis/
       cal/
       cfitsio/
       fes/
       fuv/
       libcf/
       slalib/

1.4 Setting Environment Variables

The CalFUSE pipline uses a number of environment variables to determine the locations of various files:

PATH (required)
LD_LIBRARY_PATH (required)
CF_CALDIR (required)
CF_PARMDIR (required)
CF_IDLDIR (optional)

These environment variables can be set in the user's .login or .cshrc files, or the user may edit and source the file calfuse/v3.2/bin/cfsetup.csh. See Section 1.4.6 for more details.

1.4.1 `PATH`

The environment variable PATH tells the pipeline where to find the executable programs. You should include the directory calfuse/v3.2/bin in your PATH.

1.4.2 `LD_LIBRARY_PATH`

The environment variable LD_LIBRARY_PATH tells the pipeline where to find the dynamically-linked libraries. You should include the directory calfuse/v3.2/lib in LD_LIBRARY_PATH. For Solaris installations, the directory /opt/SUNWspro/lib must also appear in LD_LIBRARY_PATH. On Mac OS X machines, this variable is called DYLD_LIBRARY_PATH.

1.4.3 `CF_CALDIR`

The environment variable CF_CALDIR gives the location of the calibration files required by the CalFUSE pipeline. At the conclusion of the installation procedure described in Section 1.1, this directory will be calfuse/v3.2/calfiles.

1.4.4 `CF_PARMDIR`

The environment variable CF_PARMDIR gives the location of the parameter files required by the CalFUSE pipeline. At the conclusion of the installation procedure described in Section 1.1, this directory will be calfuse/v3.2/parmfiles.

1.4.5 `CF_IDLDIR`

The CalFUSE pipeline uses a couple of IDL routines to generate diagnostic plots in GIF or JPEG format. One produces an image of the detector, operplotted with the extraction windows, and the other the count rate for the LiF and SiC target channels. The environment variable CF_IDLDIR gives the location of the IDL plotting routines. By default, these routines live in calfuse/v3.2/idl, but the user may move them to a different location. If the CF_IDLDIR environment variable is not set, the pipeline does not produce any plots. Note that these IDL routines require the IDL Astronomy User's Library.

1.4.6 `cfsetup.csh`

We have provided a pair of shell scripts called cfsetup.csh and cfsetup.sh to set the environment variables discussed above. Depending which shell you are using, copy either v3.2/bin/cfsetup.csh or v3.2/bin/cfsetup.sh to your home directory and set CF_DIR and CF_VERSION to the appropriate values, e.g.:

set CF_DIR="/data1/calfuse"
set CF_VERSION="v3.2"

Add the following to your .cshrc file (or its equivalent):

source cfsetup.csh

2. Running the CalFUSE Pipeline

2.1 Overview of the Pipeline

To run the code, the command is the same for both TTAG and HIST data:

  prompt> calfuse P99901010011attagfraw.fit
  prompt> calfuse P99901010011ahistfraw.fit

The following modules are called by the shell script calfuse.csh:

cf_ttag_init, cf_hist_init: Converts raw-data file into an intermediate data file (IDF) containing a photon-event list, good-time intervals, and a timeline table. Histogram data are converted to a pseudo-time-tag format.

cf_convert_to_farf: Corrects for detector deadtime. Transforms photon coordinates into the Flight Alignment Reference Frame (FARF).

cf_screen_photons: Checks data quality. Assigns status flag to each photon.

cf_remove_motions: Corrects for mirror, grating, FPA, and spacecraft motions.

cf_assign_wavelength: Applies astigmatism and Doppler corrections and assigns a wavelength to each photon.

cf_flux_calibrate: Converts WEIGHT to ERG/CM2 for each photon.

cf_bad_pixels: Applies image-motion corrections to bad-pixel map.

cf_extract_spectra: Extracts LiF and SiC spectra only for target aperture. Wavelength array is user defined; default spacing is 0.013 Å.

The pipeline expects to find all data files (*raw.fit, *jitrf.fit, *hskpf.fit) in the current directory. It operates on one detector segment at a time. For each input file, it produces the following output files: an intermediate data file (IDF); a bad-pixel map (BPM) with format similar to the IDF; two extracted spectral files, one for each of the LiF and SiC channels; and (if IDL is installed on your machine) a pair of GIF or JPEG files, one an image of the detector and the other a count-rate plot.

For more information, see "CalFUSE Version 3: A Data-Reduction Pipeline for the Far Ultraviolet Spectroscopic Explorer" (Dixon et al. 2007).

2.2 Modifying the Default Parameters

2.2.1 Rejecting Daytime Photons

Say that you are interested only in data obtained during orbital night. Broadly speaking, there are two ways to exclude day-time photons from your data: either modify the screening files before running the pipeline or modify the IDF files afterward. To do the former, simply change the header keyword DAYNIGHT from BOTH to NIGHT in the scrn*.fit files (in the parmfiles directory) and re-run the pipeline. Photons obtained during orbital day will be flagged as bad and excluded from the extracted spectrum.

Because CalFUSE does not discard undesirable photon events, but merely flags them, it is possible to change the screening applied to the data without re-running the pipeline. An interactive tool for doing this is the IDL routine cf_edit, but the C program idf_screen provides the same capabilities and runs considerably faster. It is described in FUSE Tools in C.

2.2.2 Extracting Subsets of an Exposure

Suppose that you have observed a binary star over several orbits and wish to extract only data obtained during a particular orbital phase. Again, you can either modify the screening parameter files before running the pipeline or modify the IDF files afterward. To do the former, change the good-time interval (GTI) keywords in the screening parameter file (described in Section 3.3.3). Up to 99 user-defined GTIs are allowed. To operate on an IDF file, use the routine idf_cut, described in FUSE Tools in C. Note that you can combine multiple IDFs into a single file, then use idf_cut to pull out all of the data obtained during a particular orbital phase.

Now suppose that you want to identify times when the spacecraft was actively tracking a moving target. The housekeeping file contains an array called I_FPD_TRACKING_ON that tells when the Instrument Data System (IDS) is requesting a moving-target motion from the attitude control system. Set your good-time intervals to the times when this flag is 1. It updates once per second. Here's an example:

header = headfits('E93818020011attagfraw.fit')
expstart = sxpar(header, 'EXPSTART')
a = mrdfits('E9381802001hskpf.fit', 1)
i= where(a.I_FPD_TRACKING_ON eq 1, n) ; times when tracking is good
if (n gt 0) then print, a[i].mjd - expstart

2.2.3 Burst Rejection

Burst rejection is performed by the subroutine cf_screen_burst. The program works by rejecting time periods in which the background count rate differs significantly from its median value. These count rates are determined from background regions defined in the BCHR (background characterization) parameter file and stored in the header keywords of the IDF. The use of background regions prevents mistaking variability of the target for bursts and also makes the procedure more sensitive, since the count rate in the background is quite small. We implicitly assume that any burst that affects the observation will also appear in our background regions.

The count rate within the integrated background regions is typically 3-15 counts/second, depending on the phase of the orbit and the position of the source (e.g., objects in the CVZ tend to have higher background count rates than those near the equator). To improve counting statistics, the count-rate array is binned by 10-20 seconds. This limits the time resolution for burst removal. For example, if the time sequence is binned by 10 seconds, then the data affected by bursts will be removed in 10-second chunks.

Since the background varies with time, it is necessary to determine a reference level against which the time sequence can be compared. To do this, we first determine the median count rate for the entire exposure and mark all points more than three times the median as possible bursts. We then pass the time sequence through a median filter (with a typical averaging time of 10 minutes), disregarding time segments that have been marked as possible bursts. We compare the entire time sequence with this filtered reference and mark all time bins that have a count rate more than 7 standard deviations above the running average. Here we assume Gaussian statistics, so that 1 standard deviation is the square root of the number of counts in that bin. The median filter is then re-calculated, again avoiding bins that have been marked as possible bursts. The procedure is repeated until no new bins are marked. Note: the time interval over which the median filter is determined should be longer than the duration of the bursts.

Having established a reference level, we identify as bursts those time segments for which the count rate is 5 standard deviations above the reference level. We consider this to be the minimum detectable event. All photon events that occur during bursts are flagged as bad.

While small bursts may dominate weak sources, they will have virtually no effect on bright sources. For this reason, it is possible to specify the minimum count-rate increase required for an event to be removed. By default, this level is set at 5 counts per second, about the average background level; however, for bright sources it could be increased to several hundred counts per second without affecting the spectrum. It is also possible to specify the rejection criterion as a fraction of the integrated source count rate.

The characteristics of the burst removal process are controlled by keywords in the screening parameter files (described in Section 3.3.3). In most cases, the default values will provide acceptable results, but there are times when the user may want to make adjustments. The most common are

For strong sources, the user may want to increase MNCNT or SRCFRAC, especially if the routine is cutting off data at the end of the exposure. This is often caused by a rapid enhancement of the background resulting from the initial entry of the satellite into the SAA.
If long-duration bursts are found, the user should increase NSMED so that it is longer than the bursts. Long, faint bursts are often seen in data obtained since 2005; an NSMED value of 3600 works well in these cases.
If the background is reasonably bright, then the user may want to decrease the value of NBIN to improve the time resolution of burst removal.

2.2.4 South Atlantic Anomaly

The South Atlantic Anomaly marks a depression in the Earth's magnetic field that allows particles trapped in the Van Allen belts to reach low altitudes. The high particle flux in this region raises the background count rate of the FUSE detectors to unacceptable levels. The subroutine cf_screen_saa compares the spacecraft's ground track, recorded in the LONGITUDE and LATITUDE arrays of the time-line table, with the limits of the SAA, stored in the calibration file SAAC_CAL as a binary table of latitude-longitude pairs, and flags as bad periods when data were obtained while the spacecraft lay within the SAA. Our SAA model was derived from orbital information and onboard counter data from the first 3 years of the FUSE mission. The size of the SAA contours varies on timescales of weeks, and you may find that a larger coutour is necessary for your data. If so, modify the master_calib_file.dat to read saac005.fit, rather than saac004.fit. Both contours are plotted in this figure.

2.2.5 Background Subtraction

The algorithm by which CalFUSE constructs a model background is exhaustively detailed in Dixon et al. (2007). Here are a couple of points not discussed in that article:

Sometimes, nebular emission in non-target apertures contaminates the background-sample region but is not bright enough to trigger the bright-background switch in cf_make_ttag_bkgd. To force the program to use this mode, set the keyword BKGDTYPE = -1 in the parameter file scrn*.fit. Both the detector dark count and the scattered-light models will be scaled by the exposure time; no fit will be performed.

To change the background sample regions used by cf_scale_bkgd to scale the empirical background spectrum, simply modify the BKG_MIN# and BKG_MAX# keywords in the IDF header. The IDL tool cf_edit, available from the FUSE IDL Tools Reference Page, provides an easy way to do this.

For faint emission-line spectra, users may prefer not to subtract the background, but instead fit it as an independent spectral component. To turn off background subtraction, set the keyword RUN_BKGD to NO in the parameter file parm*.fit.

2.3 Combining Data from Multiple Exposures

The FUSE project has produced a number of tools to combine data from multiple exposures, both IDFs and extracted spectral files. They are discussed in the document FUSE Tools in C.

2.4 Detector Distortions in the X Dimension

As described in Dixon et al. (2007), the FUSE wavelength calibration was derived from standard optical expressions and fit to the data with only the constant term (the zero-point of the wavelength scale) as a free parameter. The shifted dispersion solution was used to generate the final wavelength-calibration files, while residuals to the fit were assumed to represent distortions in the detector X scale and incorporated into the pipeline's geometric-correction calibration files.

A figure showing these residuals for the X coordinate of the LiF1A LWRS channel appears in the PASP paper. Here are similar plots for all FUSE channels:

**X Distortion Plots for Each FUSE Channel and Aperture**
LWRS	MDRS	HIRS	LWRS	MDRS	HIRS
LiF1A	LiF1A	LiF1A	SiC1A	SiC1A	SiC1A
LiF1B	LiF1B	LiF1B	SiC1B	SiC1B	SiC1B
LiF2A	LiF2A	LiF2A	SiC2A	SiC2A	SiC2A
LiF2B	LiF2B	LiF2B	SiC2B	SiC2B	SiC2B

The small black dots in these figures are the data points for all apertures and both channels for the given detector segment. They are included in the distortion fit, but weighted 100 times less than the data points for the channel/aperture being fitted. This helps control the fits in wavelength regions where the data are sparce or missing.

2.5 FUSE Effective-Area Curves

The FUSE flux calibration is discussed in Dixon et al. (2007), which includes a figure showing the time evolution of the effective area of the LiF and SiC channels on detector segment 1A. Here's a plot comparing the December 1999 and May 2006 effective areas for all eight channels:

Effective area of the eight FUSE channels in 1999 and 2006.

3. CalFUSE File Formats

3.1 Data Files

File formats for FUSE raw data, intermediate data, bad-pixel, and extracted spectral files are described in "CalFUSE v3: A Data-Reduction Pipeline for the Far Ultraviolet Spectroscopic Explorer" (Dixon et al. 2007). Header keywords for all of these files are described in the FUSE Data Handbook.

3.2 Housekeeping and Jitter Files

One pair of housekeeping and jitter files is produced for each FUSE exposure and may be retrieved from the MAST archive. They are used by the pipeline to track a variety of spacecraft and instrument parameters, including spacecraft pointing, detector high-voltage level, and various photon count rates. File formats for the FUSE housekeeping and jitter files are described in the FUSE Data Handbook.

3.3 Parameter Files

The files in the calfuse/v3.2/parmfiles directory allow the user to customize the pipeline processing. These files specify the screening to be performed and the calibration files to be used.

3.3.1 master_calib_file.dat

This file lists each calibration and parameter file that will be used by the pipeline. In general, a separate file is provided for each detector segment. Calibration and parameter files are contained in the directories calfuse/v3.2/calfiles and calfuse/v3.2/parmfiles, respectively. Note the date in the 4th column. For some steps of the pipeline, the optimum calibration file depends on the observation date. FUSE was launched on MJD = 51354.

Data format:

ASCII file; each record is terminated by a newline. Comment lines are indicated by a # in the first column. Data lines contain 5 columns: calibration keyword, detector segment, filename, modified Julian date, and interpolation method.

3.3.2 PARM_CAL (parm1a015.fit)

Keywords in this file give the user (some) control over how the pipeline operates.

SPEX_SIC=                   -1 / SiC extraction window Y center (0-1023)
SPEX_LIF=                   -1 / LiF extraction window Y center (0-1023)
EMAX_SIC=                   40 / SiC extraction window maximum Y movement
EMAX_LIF=                   40 / LiF extraction window maximum Y movement

Setting SPEX_SIC and SPEX_LIF to -1 directs CalFUSE to determine the best locations for the extraction windows. In this case, the EMAX_SIC and EMAX_LIF keywords specify the maximum amount that the windows may move from their expected positions.

To specify the Y position of an extraction window, set the corresponding keyword (SPEX_SIC or SPEX_LIF) to some value in the range 0-1023. You might do this if the pipeline has trouble finding your spectrum, or if you wish to offset the window from your spectrum to obtain a background spectrum (for ttag data only).

The following keywords allow the user to skip some of the pipeline-processing steps.

RUN_PHAX= 'YES'                / PHA X (walk) correction
RUN_BRST= 'YES'                / Burst rejection
RUN_JITR= 'YES'                / Jitter correction
RUN_ASTG= 'YES'                / Astigmatism correction
RUN_BKGD= 'YES'                / Background subtraction
RUN_OPTI= 'YES'                / Optimal extraction

When their keywords are set to 'NO', the walk-correction, burst-rejection, jitter-correction, astigmatism-correction, and background-subtraction steps are not performed. When RUN_OPTI = 'NO', cf_extract_spectra returns a simple, unweighted sum of all good pixels in the extraction window. Bad pixels are not included in the sum.

The following keywords control the jitter-screening routine. Times when the tabulated pointing errors exceed the allowed limits are flagged as bad, but only if the tracking flag is greater than or equal to the minimum value. Limits are applied to both the LiF and SiC channels.

TRKFLG  =                    1 / Minimum trustworthy value of TRKFLG
DX_MAX  =                   30 / [arcsec] Allowed pointing error
DY_MAX  =                   30 / [arcsec] Allowed pointing error

The following keywords allow the user to specify the wavelength range and bin size of the extracted spectrum. If not set, default values, stored in the wavelength calibration files, are used for each channel.

LIF_W0  =                      / [A] Requested initial wavelength               
LIF_WMAX=                      / [A] Requested final wavelength                 
LIF_WPC =                      / [A] Requested wavelength per channel           
SIC_W0  =                      / [A] Requested initial wavelength               
SIC_WMAX=                      / [A] Requested final wavelength                 
SIC_WPC =                      / [A] Requested wavelength per channel

3.3.3 SCRN_CAL (scrn1a015.fit)

The keywords in this file control the screening that is performed on a data set.

Keyword Default Possible values

    SAA_SCR      ON         ON/OFF 
    LIMB_SCR     ON         ON/OFF 
    DAYNIGHT    BOTH        BOTH/DAY/NIGHT 
    PHALOW        2         0-31  (integer) 
    PHAHIGH      25         0-31  (integer) 
    BRITLIMB     15.        ---   (float) 
    DARKLIMB     10.        ---   (float) 
    NUSERGTI      0         0-99  (integer) 
    GTIBEG01      0.        >= 0. (float) 
    GTIEND01      0.        >= 0. (float) 
    GTIBEG02      0.        >= 0. (float) 
    GTIEND02      0.        >= 0. (float) 
         ... 
    MNCNT         5         >= 1  (integer) 
    STDREJ        5         >= 1  (integer) 
    NBIN         15         >= 1  (integer) 
    NSMED       600         >= 1  (integer) 
    SRCFRAC       0.001     >= 0. (float) 
    BKGDTYPE=     1         -1/1  (integer)

These keywords are discussed briefly below.

SAA_SCR: SAA screening. When ON, cf_screen_saa flags as bad any time intervals during which the satellite passed through the South Atlantic Anomaly. Exposures do not ordinarily extend into the SAA, so turning this step OFF is unlikely to have much effect.

LIMB_SCR: Limb screening. When ON, cf_screen_limb_angle flags as bad any time intervals during which the telescope line of sight is closer than BRITLIMB degrees to the sunlit Earth limb or closer than DARKLIMB degrees to the dark Earth limb. Exposures do not ordinarily extend below the specified limb constraints, so the user is unlikely to see any effect from setting this parameter to OFF. The main exception is the case of exposures deliberately obtained of the bright earth for purposes of studying airglow emission.

DAYNIGHT: Day/night screening. When set to DAY, only time intervals when the satellite was on the sunlit side of the earth are flagged as good; when set to NIGHT, only time intervals when the satellite was on the dark side of the Earth are set to good; and when set to BOTH, no screening by orbital phase is performed.

PHALOW: Minimum acceptable pulse height for ttag data. (No pulse height information is present in HIST mode.) We no longer recommend tightening the pulse-height limits beyond their default values, as a significant number of "real" photon events can be rejected.

PHAHIGH: Maximum acceptable pulse height for ttag data.

BRITLIMB: The limb angle constraint applied if the Earth limb closest to the telescope boresight is in sunlight. See the discussion for LIMB_SCR above.

DARKLIMB: The limb angle constraint applied if the Earth limb closest to the telescope boresight is in shadow. See the discussion for LIMB_SCR above.

NUSERGTI: The number of user-defined good-time intervals. These can be used, for example, to extract a particular phase in a binary orbit. All other screening (by SAA, limb angle, day/night, pulse height) is performed as usual. There may be up to 99 user-defined good-time intervals (limited by the 8-character length of FITS keywords). The limits of each good time interval are specified by the time in seconds from the start of the exposure.

GTIBEG01: Start of first user-defined good-time interval, specified as time in seconds after the start of the exposure.

GTIEND01: End of first user-defined good-time interval, specified as time in seconds after the start of the exposure.

GTIBEG02: Start of second user-defined good-time interval, specified as time in seconds after the start of the exposure.

GTIEND02: End of second user-defined good-time interval, specified as time in seconds after the start of the exposure.

The next five keywords are used by the module cf_screen_burst. They can be adjusted to optimize performance; for details, see Section 2.2.3.

MNCNT: Minimum enancement of the background (in counts) needed to flag a burst. Currently = 5.

STDREJ: Minimum number of standard deviations from the mean required for a burst to be detected. Currently = 5.

NBIN: Size of time bins (in seconds). Ensures that there are a reasonable number of counts in each time step. Currently = 15.

NSMED: Time interval (in seconds) used in the median-filter algorithm. Currently = 600. This number is used to determine the background level against which the bursts are measured. It should be longer than the typical burst duration.

SRCFRAC: Fraction of the source intensity required for a burst to be removed. For example, SCRFRAC = 0.01 requires that a burst exceed 1% of the integrated source intensity before being removed. Currently = 0.001.

BKGDTYPE: For time-tagged data, the background is generally scaled from measurements of unilluminated portions of the detector. If this keyword is set to -1, the background is instead scaled directly from the exposure time and mean intensities of the detector dark count and scattered light. For details, see Dixon et al. (2007).

3.3.4 Background Characterization File (bchr1a003.fit)

The background-characterization files contain parameters used to construct the background model. Though considered a parameter file, its data are stored as binary extensions and are not meant to be modified by the user.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains a binary table listing the limits of the detector regions used to estimate the background flux.

Data format:

FITS binary table with 3 columns in a single row. Each entry is an 8-element integer array. The first 6 integers are ymin/ymax pairs for three sample regions. The 7th number is the number of rows in the active aperture region, and the 8th is the total number of active rows on the detector.

TTYPE	TFORM	TUNIT
LWRS	8I	pixels
MDRS	8I	pixels
HIRS	8I	pixels

HDU 3

Description:

Lists the detector dark-count rate (in units of 1E-7 counts/s/pixel) as a function of the pulse-height threshold PHALOW, which is set in the SCRN file.

Data format:

Nine-element array of floating-point numbers in a FITS image extension. intcr=intcrarray[PHALOW]

3.4 Calibration Files

All calibration files used by CalFUSE are standard FITS format files, except for the FUSE.TLE file, which is in ASCII format. The FITS files generally contain a minimal header, with the actual data being present in extensions. Calibration files that are applied to an entire detector segment (e.g., background, geometric distortions) generally consist of an image the same size as the detector (16384 × 1024 pixels), though some of these files are compressed. Calibration data that are specific to particular spectrograph entrance apertures (e.g., extraction windows, effective areas) are stored in separate extensions of a single file.

The filenames shown are the versions distributed with CalFUSE v3.2.3. In all cases for which separate files are used for data from each detector segment, the filename corresponding to detector 1A is used as an example; similar files exist for segments 1B, 2A, and 2B.

AEFF_CAL - Effective-area curves
AIRG_CAL - Airglow lines
ASTG_CAL - Astigmatism correction
BKGD_CAL - Background image
CHID_CAL - Extraction windows for each aperture
DIGI_CAL - Detector keyword reference file
ELEC_CAL - Electronics keyword reference file
FUSE.TLE - Two-line elements from NORAD
GEOM_CAL - Geometric corrections in X and Y
GRAT_CAL - Grating-motion correction
MIRR_CAL - Mirror-motion correction
PHAH_CAL - Mean pulse height for HIST data
PHAX_CAL - Walk correction
QUAL_CAL - Bad-pixel maps
RATE_CAL - Count-rate-dependent Y stretch
SAAC_CAL - SAA contours
SPEC_CAL - Spectrograph parameters
STIM_CAL - Mean stim-pulse coordinates
TMXY_CAL - Time-dependent drifts in X and Y
VOLT_CAL - Detector high-voltage level
WAVE_CAL - Wavelength calibration
WGTS_CAL - Extraction weights (flux distribution in Y)

AEFF_CAL (aeff1a100.fit)

This file contains effective-area curves (in cm²) for each aperture. Because the FUSE sensitivity changes with time, new AEFF_CAL files are calculated every three to six months.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains the effective-area curve the LiF HIRS spectrum.

Data format:

FITS binary table with 2 columns and 1024 rows. (Data are actually stored as a single row; each table element is an array.)

TTYPE	TFORM	TUNIT
WAVE	1024E	Å
AREA	1024D	cm²

HDU 3, 4, 5

Description:

Contains effective-area curves for the LiF MDRS, LWRS, and PINH spectra.

Data format:

FITS binary table with 2 columns and 1024 rows.

HDU 6, 7, 8, 9

Description:

Contains effective-area curves for the SiC HIRS, MDRS, LWRS, and PINH spectra.

Data format:

FITS binary table with 2 columns and 1024 rows.

AIRG_CAL (airg1a004.fit)

This file lists regions of the detector likely to be contaminated by airglow emission.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains the positions in detector FARF coordinates of the regions affected by airglow and scattered solar emission lines. These regions are rectangular areas defined by minimum and maximum X and Y pixel values. The table also lists the spectral channel and aperture affected by each region and identifies the atomic species and wavelength of the main contributors included in the areas.

Data format:

FITS binary table with 8 columns and a variable number of rows. (Data are actually stored as a single row; each table element is an array.)

TTYPE	FORMAT	TUNIT
XMIN	INT	pixels
XMAX	INT	pixels
YMIN	INT	pixels
YMAX	INT	pixels
CHANNEL	STRING	LiF or SiC
APERTURE	STRING	HIRS, MDRS, or LWRS
ION	STRING	N I, O I, He I, N II, etc.
WAVE	STRING	Å

ASTG_CAL (astg1a011.fit)

This file contains corrections for spectral line curvature caused by astigmatism in the Rowland circle grating mount.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains the X shift for each pixel in the LiF HIRS spectral image needed to correct line curvature for a point source.

Data format:

Image 16384 × 150 pixels. Data consist of real values, stored as two-byte integers and scaled using the BSCALE and BZERO keywords.

Note: the Y dimension of this image ranges from 133 to 191 pixels, depending on the detector and aperture. To determine the image size, read the keyword NAXIS2 from the header of the desired extension. The zero-point row for the astigmatism correction is stored in the header keyword SLIT_CEN.

HDU 3, 4, 5

Description:

Contains X shifts for each pixel in the LiF MDRS, LWRS, and PINH spectral images needed to correct line curvature for a point source.

Data format:

Image 16384 × 165 pixels. Data consist of real values, stored as two-byte integers and scaled using the BSCALE and BZERO keywords.

HDU 6, 7, 8, 9

Description:

Contains X shifts for each pixel in the SiC HIRS, MDRS, LWRS, and PINH spectral images needed to correct line curvature for a point source.

Data format:

Image 16384 × 165 pixels. Data consist of real values, stored as two-byte integers and scaled using the BSCALE and BZERO keywords.

HDU 10-17

Description:

Same as HDU 2-9, except for an extended source. Because no astigmatism correction is presently defined for extended sources, these arrays are filled with zeros.

This file contains the upper and lower limits of the extraction windows.

HDU 1

Description:

Minimal FITS header.

Data format:

Empty

HDU 2

Description:

Contains the extraction limits for each pixel in the LiF HIRS spectral image, assuming a point source. XPIX is the detector X pixel number; YLOW and YHIGH are the lower and upper boundaries of the extraction window. The header keyword CENTROID contains the mean of the YLOW and YHIGH arrays.

Data format:

FITS binary table with 4 columns and 16384 rows. (Data are actually stored as a single row; each table element is an array.)

TTYPE	TFORM	TUNIT
XPIX	16384I	pixels
YLOW	16384I	pixels
YHIGH	16384I	pixels

HDU 3, 4, 5

Description:

Contains extraction limits for each pixel in the LiF MDRS, LWRS, and PINH spectral images, assuming a point source.

Data format:

FITS binary table with 4 columns and 16384 rows.

HDU 6, 7, 8, 9

Description:

Contains extraction limits for each pixel in the SiC HIRS, MDRS, LWRS, and PINH spectral images for a point source.

Data format:

FITS binary table with 4 columns and 16384 rows.

HDU 10-17

Description:

Same as HDU 2-9, except for an extended source.

Data format:

FITS binary table with 4 columns and 16384 rows.

DIGI_CAL (digi001.fit)

This file lists the expected values of 16 parameters for each detector. The subroutine cf_check_digitizer compares keywords in the data file header with these values and issues a warning message if they are discrepant.

HDU 1

Description:

All data are stored in header keywords:

               Detector 1 parameters                                  
DET1ASCL=       17.0 / DET1 segment A  Time (X) image scale factor    
DET1BSCL=       91.0 / DET1 segment B  Time (X) image scale factor    
DET1AXOF=      183.0 / DET1 segment A  Time (X) image position offset 
DET1BXOF=      107.0 / DET1 segment B  Time (X) image position offset 
DET1AUCT=      227.0 / DET1 TDC-A Upper Charge Threshold setting      
DET1BUCT=      228.0 / DET1 TDC-B Upper Charge Threshold setting      
DET1ABWK=      165.0 / DET1 TDC-A Begin CFD Walk setting              
DET1BBWK=      139.0 / DET1 TDC-B Begin CFD Walk setting              
DET1AEWK=      165.0 / DET1 TDC-A End CFD Walk setting                
DET1BEWK=      125.0 / DET1 TDC-B End CFD Walk setting                
DET1ABSL=      227.0 / DET1 TDC-A Charge Baseline Threshold setting   
DET1BBSL=      228.0 / DET1 TDC-B Charge Baseline Threshold setting   
DET1ALCT=        7.0 / DET1 TDC-A Lower Charge Threshold setting      
DET1BLCT=        7.0 / DET1 TDC-B Lower Charge Threshold setting      
DET1ALTT=      169.0 / DET1 TDC-A Lower Time Threshold setting        
DET1BLTT=      143.0 / DET1 TDC-B Lower Time Threshold setting        
                 Detector 2 parameters                                
DET2ASCL=       65.0 / DET2 segment A  Time (X) image scale factor    
DET2BSCL=       78.0 / DET2 segment B  Time (X) image scale factor    
DET2AXOF=      107.0 / DET2 segment A  Time (X) image position offset 
DET2BXOF=       80.0 / DET2 segment B  Time (X) image position offset 
DET2AUCT=      227.0 / DET2 TDC-A Upper Charge Threshold setting      
DET2BUCT=      228.0 / DET2 TDC-B Upper Charge Threshold setting      
DET2ABWK=      113.0 / DET2 TDC-A Begin CFD Walk setting              
DET2BBWK=      111.0 / DET2 TDC-B Begin CFD Walk setting              
DET2AEWK=      114.0 / DET2 TDC-A End CFD Walk setting                
DET2BEWK=      112.0 / DET2 TDC-B End CFD Walk setting                
DET2ABSL=      228.0 / DET2 TDC-A Charge Baseline Threshold setting   
DET2BBSL=      228.0 / DET2 TDC-B Charge Baseline Threshold setting   
DET2ALCT=        7.0 / DET2 TDC-A Lower Charge Threshold setting      
DET2BLCT=        7.0 / DET2 TDC-B Lower Charge Threshold setting      
DET2ALTT=      159.0 / DET2 TDC-A Lower Time Threshold setting        
DET2BLTT=      149.0 / DET2 TDC-B Lower Time Threshold setting        
END

Data format:

Empty

ELEC_CAL (elec005.fit)

This file contains header keywords describing the detector electronics, including stim-pulse positions and the limits of the active region of each detector. It is read by a number of CalFUSE subroutines.

HDU 1

Description:

All data are stored in header keywords:

ACTIVE_L=          512 / Left edge of active area (inclusive)           
ACTIVE_R=        15871 / Right edge of active area (inclusive)          
TTAG_BUS=       8000.0 / Maximum IDS count rate in TTAG mode            
HIST_BUS=      32000.0 / Maximum IDS count rate in HIST mode            
ABORT_1A=     2.76E-06 / Electronics abort time for 1A                  
CLOCK_1A=     1.50E-06 / Electronics clock time for 1A                  
STATE_1A=    11.68E-06 / Electronics state time for 1A                  
ABORT_1B=     4.66E-06 / Electronics abort time for 1B                  
CLOCK_1B=     1.66E-06 / Electronics clock time for 1B                  
STATE_1B=    10.29E-06 / Electronics state time for 1B                  
ABORT_2A=     5.23E-06 / Electronics abort time for 2A                  
CLOCK_2A=     1.27E-06 / Electronics clock time for 2A                  
STATE_2A=     8.37E-06 / Electronics state time for 2A                  
ABORT_2B=     4.23E-06 / Electronics abort time for 2B                  
CLOCK_2B=      0.0E-06 / Electronics clock time for 2B                  
STATE_2B=     6.31E-06 / Electronics state time for 2B                  
STIMWDTH=          128 / Half width of stim box                         
STIMTOLR=         0.01 / Stim position tolerance                        
STIMLX1A=          236 / Stim 1A left X position                        
STIMLY1A=          805 / Stim 1A left Y position                        
STIMRX1A=        16125 / Stim 1A right X position                       
STIMRY1A=          781 / Stim 1A right Y position                       
STIMLX1B=          262 / Stim 1B left X position                        
STIMLY1B=          757 / Stim 1B left Y position                        
STIMRX1B=        16148 / Stim 1B right X position                       
STIMRY1B=          780 / Stim 1B right Y position                       
STIMLX2A=          242 / Stim 2A left X position                        
STIMLY2A=          618 / Stim 2A left Y position                        
STIMRX2A=        16119 / Stim 2A right X position                       
STIMRY2A=          627 / Stim 2A right Y position                       
STIMLX2B=          251 / Stim 2B left X position                        
STIMLY2B=          639 / Stim 2B left Y position                        
STIMRX2B=        16121 / Stim 2B right X position                       
STIMRY2B=          632 / Stim 2B right Y position                       
ACTVL_1A=          800 / Active left edge of 1A (inclusive)             
ACTVR_1A=        15583 / Active right edge of 1A (inclusive)            
ACTVB_1A=            0 / Active bottom edge of 1A (inclusive)           
ACTVT_1A=         1023 / Active top edge of 1A (inclusive)              
ACTVL_1B=          800 / Active left edge of 1B (inclusive)             
ACTVR_1B=        15583 / Active right edge of 1B (inclusive)            
ACTVB_1B=            0 / Active bottom edge of 1B (inclusive)           
ACTVT_1B=         1023 / Active top edge of 1B (inclusive)              
ACTVL_2A=          800 / Active left edge of 2A (inclusive)             
ACTVR_2A=        15583 / Active right edge of 2A (inclusive)            
ACTVB_2A=            0 / Active bottom edge of 2A (inclusive)           
ACTVT_2A=         1023 / Active top edge of 2A (inclusive)              
ACTVL_2B=          800 / Active left edge of 2B (inclusive)             
ACTVR_2B=        15583 / Active right edge of 2B (inclusive)            
ACTVB_2B=            0 / Active bottom edge of 2B (inclusive)           
ACTVT_2B=         1023 / Active top edge of 2B (inclusive)              
FIFO_DRN=         3470 / IDS FIFO drain rate

Data format:

Empty

FUSE.TLE

Description:

Contains NORAD two-line elements characterizing the orbit of the FUSE satellite. There is typically one entry for each day of the mission since launch (1999 June 24).

Thermal instabilities cause the FUSE gratings to wobble on orbital, diurnal, and precessional (60-day) timescales. An additional long-term, non-periodic drift is also apparent. For each of these timescales, this file contains a set of four file extensions, providing grating-motion corrections for each of spectrograph's four gratings. At present, only corrections for the long-term and orbital drifts are available.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains correction to position (in both X and Y) of the LiF1 spectrum (segments A and B) due to rotation of the LiF1 grating as a function of time throughout the mission. This motion is non-periodic.

Data format:

FITS binary table with 3 columns and any number of rows. At present, the Y correction is always zero.

TTYPE	TFORM	TUNIT
MJD	1D	days
XSHIFT	1E	pixels
YSHIFT	1E	pixels

HDU 3, 4, 5

Same as HDU 2, but for the SiC1, LiF2, and SiC2 gratings, respectively.

HDU 6

Description:

Contains correction to position (in both X and Y) of the LiF1 spectrum (segments A and B) due to rotation of the LiF1 grating as a function of orbital phase.

The grating-motion solution depends on three parameters: beta angle (the angle between the target and the anti-sun vector), pole angle (the angle between the target and the orbit pole), and spacecraft roll angle (east of north, determined from the file-header keyword APER_PA). The pipeline routine cf_grating_motion compares the beta, pole, and roll angles of the spacecraft with the grid of values contained in this extension. Finding a match, it reads the corresponding XCOEFF and YCOEFF values (up to five of each) and uses them to generate pair of series (one each for X and Y) from which the motion corrections are computed.

Data format:

FITS binary table with 10 columns and any number of rows. Columns 7 and 8 contain arrays of Fourier coefficients. Columns 9 and 10 record Χ² values for the fits to the measured X and Y coordinates of the Lyman β line.

TTYPE	TFORM	TUNIT
BETALO	1E	degrees
BETAHI	1E	degrees
POLELO	1E	degrees
POLEHI	1E	degrees
APERPALO	1E	degrees
APERPAHI	1E	degrees
XCOEFF	5E	pixels
YCOEFF	5E	pixels
XCHISQ	1E
YCHISQ	1E

HDU 7, 8, 9

Same as HDU 6, but for the SiC1, LiF2, and SiC2 gratings, respectively.

MIRR_CAL (mirr002.fit)

Data in this file are used to correct for spectral drifts in the X dimension due to mirror motions during an exposure. The file was constructed assuming that the guide channel is LiF1, so its mirror motions are corrected by the spacecraft itself. Corrections are thus given only for the SiC1, LiF2, and SiC2 channels. Since 2005 June, the guide channel has been LiF2. In practice, this is not a problem, because the correction for LiF2 is zero. Also note that the corrections for SiC1 and SiC2 are identical (though opposite in sign).

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains corrections to the X positions of SiC1 photons.

Data format:

FITS image extension containing a single array of 100 numbers, representing the correction to be added to the observed X position of each SiC1 photon, computed once per minute for the 100-minute orbit of the FUSE satellite.

HDU 3

Same as HDU 2, but for LiF2 spectra.

HDU 4

Same as HDU 2, but for SiC2 spectra.

PHAH_CAL (phah1a002.fit)

For each aperture, this file lists the mean pulse height of photons obtained in TTAG mode as a function of time through the mission. This mean pulse height is assigned to HIST data so that it can receive a rudimentary walk correction.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains mean pulse heights for photons observed through the LiF HIRS aperture.

Data format:

FITS binary table with 2 columns and an arbitrary number of rows.

TTYPE	TFORM	TUNIT
MJD	D	Modified Julian Date
PHA	B	Arbitrary units (0-31)

HDU 3, 4, 5

Description:

Contains mean pulse heights for photons observed through the LiF MDRS, LWRS, and PINH apertures.

Data format:

Description:

Contains locations and dimensions of the dead regions within the detector active area. The dead regions are specified as ellipses. The table contains the coordinates of the center of each ellipse and its semi-major and semi-minor axes.

Data format:

FITS binary table with 4 columns. (Data are actually stored as a single row; each table element is an array.)

TTYPE	TFORM	TUNIT
X	INT	pixels
Y	INT	pixels
RX	INT	pixels
RY	INT	pixels

RATE_CAL (rate1a002.fit)

This file gives the stretch in the Y scale as a function of observed Y value and count rate.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Shifts are stored in a 102 × 31 image. Rows (i) are the observed Y value / 10. Column numbers (n) are related to the count rate as n=10 × [log10(count rate) - 1]. Thus, the shift for a Y value of 800 and a global count rate of 500 will be located in the array element [80,17].

Shift is applied according to the recipe new Y = old Y + image[i, n]

Data format:

Image 102 × 31 pixels. Data consist of real values.

SAAC_CAL (saac004.fit or saac005.fit)

These files contain the outer contour of the SAA used by the subroutine cf_screen_saa. By default, the pipeline uses saac004.fit. The file saac005.fit contains larger contours.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains a binary table with the longitude and latitude, marking the extent of the SAA.

Data format:

FITS binary table with 2 columns and 10 rows.

TTYPE	TFORM	TUNIT
LATITUDE	1E	degrees
LONGITUDE	1E	degrees

SPEC_CAL (spec002.fit)

This file lists spectragraph parameters used by the subroutine cf_fpa_position, which shifts spectra in X to account for offsets of the focal-plane assembly.

HDU 1

Description:

All data are stored in header keywords:

ALPHALIF=   25.0000 / [deg] Spectrograph entrance angle (LiF)        
ALPHASIC=   24.0000 / [deg] Spectrograph entrance angle (SiC)        
SIGMALIF=   1869.20 / [A] Central groove spacing (LiF)               
SIGMASIC=   1734.00 / [A] Central groove spacing (SiC)               
DIAM    =   1652.00 / [mm] Rowland circle diameter

Data format:

Empty

STIM_CAL (stim1a002.fit)

This file lists the X and Y coordinates of the stim pulses as a function of time. If the stim pulses for a particular data set are corrupted, cf_thermal_distort uses these values to estimate the thermal-distortion correction for that exposure.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains a binary table listing the X and Y coordinates of the stim pulses for various dates throughout the mission.

Data format:

FITS binary table with 5 columns and a variable number of rows.

TTYPE	TFORM	TUNIT
MJD	1D	Modified Julian Date
X0	1D	pixels
X1	1D	pixels
Y0	1D	pixels
Y1	1D	pixels

Data format:

Data are stored as a real-valued image 103 pixels wide and an arbitrary number of pixels high. Each row represents the correction for a particular MJD. The first column contains the MJD value for that row, the second column contains the shift for pixels with 0 <= Y < BIN_FACT, the third column for pixels with BIN_FACT <= Y < 2*BIN_FACT, etc. The value of BIN_FACT is stored as a keyword in the header of the HDU. To prevent pile-ups at boundaries, we interpolate between columns; however, we do not interpolate between rows.

VOLT_CAL (volt1a009.fit)

This file lists the detector bias voltage values used for observations and SAA passes. The nominal values change through the mission, so are given as a time-ordered list. Both cf_fuv_init and cf_screen_high_voltage use this file to confirm that the full detector voltage was used throughout an observation.

HDU 1

Description:

All data are stored in header keywords:

HVGOODLM=        0.850000 / HV/FULL < HVGOODLM is questionable             
HVBADLIM=        0.850000 / HV/FULL < HVBADLIM is bad                      
MJD0    =     51403.84931 / [MJD] New voltage uplinked this date           
SAA0    =              70 / SAA voltage level                              
FULL0   =             129 / Full voltage level                             
MJD1    =     51416.68750 / [MJD] New voltage uplinked this date           
SAA1    =              70 / SAA voltage level                              
FULL1   =             129 / Full voltage level

Data format:

Empty

WAVE_CAL (wave1a023.fit)

This file contains the wavelength of each detector pixel.

HDU 1

Description:

Contains keywords giving history of file.

Data format:

Empty

HDU 2

Description:

Contains the wavelength for each pixel in the LiF HIRS spectral image. Useful file header keywords include

FPACXPOS=              117.000 / [microns] LiF FPA X position                    
FPACZPOS=             -34.7000 / [microns] LiF FPA Z position                    
DISPAPIX=           0.00673558 / Average dispersion (Angstroms per pixel)        
W0      =              984.994 / [A] Recommended initial wavelength             
WMAX    =                 1085 / [A] Recommended final wavelength                
WPC     =            0.0130000 / [A] Recommended wavelength per channel

Data format:

FITS binary table with 2 columns and 16384 rows:

TTYPE	TFORM	TUNIT
PIXEL	1I	pixels
WAVELENGTH	1E	Å

HDU 3, 4, 5

Description:

Contains wavelengths for each pixel in the LiF MDRS, LWRS, and PINH spectral images.

Data format:

FITS binary table with 2 columns and 16384 rows.

HDU 6, 7, 8, 9

Description:

Contains wavelengths for each pixel in the SiC HIRS, MDRS, LWRS, and PINH spectral images.

Table of Contents

1. Installing the CalFUSE Pipeline

1.1 Retrieving the Pipeline

1.2 Installing the Pipeline

1.3 Directory Structure

1.4 Setting Environment Variables

1.4.1 PATH

1.4.2 LD_LIBRARY_PATH

1.4.3 CF_CALDIR

1.4.4 CF_PARMDIR

1.4.5 CF_IDLDIR

1.4.6 cfsetup.csh

2. Running the CalFUSE Pipeline

2.1 Overview of the Pipeline

2.2 Modifying the Default Parameters

2.2.1 Rejecting Daytime Photons

2.2.2 Extracting Subsets of an Exposure

2.2.4 South Atlantic Anomaly

2.2.5 Background Subtraction

3. CalFUSE File Formats

3.1 Data Files

3.2 Housekeeping and Jitter Files

3.3 Parameter Files

3.3.1 master_calib_file.dat

3.3.2 PARM_CAL (parm1a015.fit)

HDU 1

HDU 2

HDU 3

3.4 Calibration Files

HDU 1

HDU 2

HDU 3, 4, 5

HDU 6, 7, 8, 9

HDU 1

HDU 2

HDU 1

HDU 2

HDU 3, 4, 5

HDU 6, 7, 8, 9

HDU 10-17

HDU 1

HDU 2

HDU 3

HDU 1

HDU 2

HDU 3, 4, 5

HDU 6, 7, 8, 9

HDU 10-17

HDU 1

HDU 1

HDU 1

HDU 2

HDU 3

HDU 1

HDU 2

HDU 3, 4, 5

HDU 6

HDU 7, 8, 9

HDU 1

HDU 2

HDU 3

HDU 4

HDU 1

HDU 2

HDU 3, 4, 5

HDU 6, 7, 8, 9

HDU 1

HDU 1

HDU 2

HDU 1

HDU 2

HDU 1

HDU 2

HDU 1

HDU 1

HDU 2

HDU 1

HDU 2

HDU 3

1.4.1 `PATH`

1.4.2 `LD_LIBRARY_PATH`

1.4.3 `CF_CALDIR`

1.4.4 `CF_PARMDIR`

1.4.5 `CF_IDLDIR`

1.4.6 `cfsetup.csh`