4. Creating OTs

OTs can be created in two different ways:

  1. Using a GMOS/F2 pre-image that shows all science targets

  2. Using a RA/DEC catalog and an external image (pseudo-image)

One method’s strength is the other’s weakness, and vice versa.

4.1. Comparing pre-images and pseudo-images

Mask designs based on pre-images and pseudo-images are equally accurate, provided the external catalog used with the pseudo-image has sufficiently high relative astrometric accuracy (0.1” or better). In this case, slit positions between the two methods are well within a single pixel:


Fig. 4.1 Accuracy of the pseudo-image mode. In 2017, the transformations from the RA/DEC system to distorted image coordinates have been re-measured. We observed stellar fields in good seeing, and calibrated the data against GAIA DR1. The new transformations dramatically improve the performance compared to the previous tranformations that were obtained in 2006.

4.1.1. Pre-images: Pros and Cons


  • Always highly accurate relative slit positions

  • Image provided by Gemini, ready to be used in GMMPS

  • Largely immune against proper motions, because science data will be taken soon thereafter.


  • Additional telescope time required, perhaps prohibitively much

  • Lag between the semester start and the start of the MOS observations. The pre-images must be taken before masks can be designed.

4.1.2. Pseudo-images: Pros and Cons


  • Masks can be designed early, no need to wait for pre-images


  • Accuracy of the relative slit positions determined by the accuracy of the targets’ sky coordinates

  • Proper motions may render the mask design invalid if the data is older than about a year.

4.2. Impact of proper motions

Each mask has at least two reference slits (squared, 2” wide) to align the mask on sky, using acquisition stars (Sect. 14). Proper motions may void a mask design in two ways:

  1. One or more of the acquisition stars has high proper motion. In this case it can happen that the mask cannot be aligned on sky anymore. To avoid this problem, GMMPS displays the proper motions (from PPMXL, and in the future from GAIA) of all acquisition stars. An additional warning is shown if any component of the proper motion vector exceeds 100 mas/yr, and an error if 250 mas/yr are exceeded.

  2. Even if the mask was acquired successfully, individual proper motions may drive stellar science targets off their respective slits. In its current version, GMMPS does not check the proper motions of science targets, nor does it correct the slit positions for proper motions.

4.3. Creating OTs based on pre-images

4.3.1. Object detection with SExtractor

Use SExtractor to detect objects in the pre-image. You need the following configuration files, available in the examples/ sub-directory in the GMMPS distribution tree:


Copy all files into the directory with your pre-image. Change to that directory, and then run

sex -c gmmps.conf <image.fits>

This will create a FITS table called preimage.sex.fits, requiring that the SExtractor executable can be found in your PATH variable. Detection parameters can be changed in gmmps.conf.

The preimage.sex.fits table must be converted to OT format before it can be loaded in GMMPS. The conversion is done with stsdas2objt, available in gemini.gmos.mostools in Gemini IRAF. Open an IRAF session and do:

cd /path/to/image
epar stsdas2objt

Configure the task stsdas2objt like this, leaving the other parameters empty or at their default values:

intable = preimage.sex.fits
image   = <image.fits>
fl_wcs  = yes
instrum = gmos (gmos|flamingos)
id_col  = NUMBER
mag_col = MAG_AUTO
x_col   = X_IMAGE
y_col   = Y_IMAGE

You must let stsdas2objt recalculate (RA,DEC) using the WCS in the image headers (fl_wcs = yes). This is because SExtractor calculates RA in degrees, but stsdas2objt wants it in hours instead. The converted FITS table will be called preimage.sex_OT.fits. You may rename it arbitrarily. However, we recommend you keep the _OT.fits suffix.

4.3.2. Object detection with IRAF

IRAF may also be used to create the object tables using daophot.

First, find the objects

daofind mrgN20011021S104_add[1] output=mrgN20011021S104.coo \
     fwhm=12 threshold=100 verify- ccdread="RDNOISE" gain="GAIN" sigma=14.

Aperture photometry using apphot.phot

phot mrgN20011021S104_add[1] coords=mrgN20011021S104.coo \
     output=mrgN20011021S104.mag  ccdread="RDNOISE" gain="GAIN" \
     sigma=14. verify- verbose- inter-

Like before with SExtractor, the resulting daophot table must be converted to OT format. To this end use the task gemini.gmos.mostools.app2objt. Note that app2objt will remove any objects with mag=INDEF since GMMPS cannot handle these values.

app2objt mrgN20011021S104.mag verbose+ image=mrgN20011021S104_add.fits priority="2"

The result, mrgN20011021S104_OT.fits, can be loaded in GMMPS.

Further examples and explanations can be found in these two IRAF scripts:

4.4. Creating OTs based on target lists and pseudo-images

A mask can also be designed based on a list of RA/DEC coordinates. The relative accuracy of the coordinates should be equal or better than 0.1”. The RA/DEC values are transformed to the x/y coordinates the targets would have in a (distorted) GMOS/F2 pre-image.

You also need an external image with a valid WCS header that covers the area of the target list. This image does not need to be the same image as the one from which the target list was extracted; it merely serves as a visual reference in GMMPS. The image will be transformed into a pseudo-image, mimicking a GMOS / F2 pre-image.

The slit positions are entirely based on the list of sky coordinates.

4.4.1. Configuring gmskcreate

The pseudo-image and the OT for the mask design are built using the Gemini IRAF task gmskcreate. This task is available in gemini.gmos.mostools and must be configured as follows:

indata    = Input ASCII file containing the spectroscopy targets (see below)
inimage   = Input FITS file used to create the pseudo-image (see below)
gprgid    = Your Gemini program ID (e.g. GS-2017A-Q-1)
instrume  = Instrument (gmos-n|gmos-s|flamingos2)
rafield   = RA value of field center (decimal degrees or hours)
decfield  = Dec value of field center (decimal degrees)
pa        = Position angle of field
fl_getxy  = yes
fl_getim  = yes
iraunits  = degrees or hours (units in the input catalog)
fraunits  = degrees or hours (units for "rafield")
outtab    = name for the output OT ("GMI<indata>_OT.fits" if empty)
outcoords = name for an optional file containing x/y positions


RA, DEC and PA must be identical to the ones defined in the phase II observations. You must also verify in the Gemini Observing Tool that a suitable guide star is available for the chosen RA, DEC and position angle.

The file specified by the outcoords parameter contains the instrumental (x,y) coordinates, one object per line. It is not used by GMMPS, but may be useful for overplotting the (x,y) coordinates on the pseudo-image using the IRAF task tvmark.

4.4.2. Input ASCII file for gmskcreate

The input file must contain one line per spectroscopic target, in this order (note that the line below is not included in the file):

ID   RA   DEC   MAG   priority   slitsize_x   slitsize_y   slittilt   slitpos_y/x

ID, RA, DEC and MAG are mandatory columns. The others are optional. Values must be blank separated. Optional values may be set for some objects and omitted for others. If provided, then all five optional values must be set, otherwise they will be defaulted as

priority   = 1
slitsize_x = slitszx
slitsize_y = slitszy
slittilt   = 0
slitpos_y  = 0

where slitsize_x/y are input parameters for gmskcreate. Note that slitpos_y/x refers to the offset of the object along the slit, i.e. this is slitpos_y for GMOS, and slitpos_x for F2.


GMMPS enforces that acquisition stars (priority = 0) have slitsize_x/y = 2.0, slittilt = 0 and slitpos_x/y = 0. Any optional values given for acquisition stars will be ignored.

The RA/DEC coordinates in the target list must match the WCS in the input image, otherwise plotting symbols in GMMPS will not line up properly.

4.4.3. Input FITS image for gmskcreate

The pseudo-image is created from an external image, which must contain the WCS keywords CRPIX1/2, CRVAL1/2, and the four CD matrix entries, CDi_j. The WCS of the external image should correspond reasonably well with the astrometry of the input objects. Otherwise, the slits plotted in GMMPS will not lie on top of the objects. This is not a problem for the mask creation, as the slit positions are based on the target catalog. However, mask checking will be more difficult.

The resulting pseudo-images will always be pre-fixed with the string GMI, so that they can be recognized as such when submitted to the observatory.

4.4.4. Example

The input ASCII file contains the following columns:

ID   RA   DEC   MAG   priority   slitsize_x   slitsize_y   slittilt   slitpos_y

Note that for F2 slitpos_y would be replaced by slitpos_x (offset along the slit).

The actual input file would look like this (note that the column names listed above are absent):

10    201.67725788    -47.65156978    17      2   1.0   15.0  0.0   3.0
11    201.68830528    -47.64194528    17
12    201.66749228    -47.65391578    17      2   1.0   15.0 -6.0   0.0
13    201.68427878    -47.66114059    17
14    201.71123928    -47.64720188    17      3
15    201.69364588    -47.62953180    17
16    201.69952048    -47.63118018    17
17    201.69640768    -47.66783558    14.6    0   1.0   5.0   0.0   0.0
18    201.71233788    -47.63684878    14.3    3   1.0   20.0  3.0
20    201.64417083    -47.68341666    14.7    0   2.0   2.0   0.0   0.0
21    201.73536249    -47.61178055    14.6    0   2.0   2.0   0.0   0.0

Here, the optional values for objects 11, 13, 15, and 16 have been omitted, and default values will be used. Likewise, for objects 14 and 18, incomplete optional values are provided, resulting in all of them being defaulted. Objects 10 and 12 have 15” long slits, number 12 is tilted by -6 degrees, and object 10 offset by 3.0”x along the slit.

Objects 17, 20 and 21 are acquisition stars (priority = 0), meaning all other optional values provided will be ignored and slitsize_x/y = 2.0, slittilt = 0 and slitpos_x/y = 0 enforced. The aquisition objects can be changed interactively in GMMPS. It is important that the OT contains sufficiently many potential acquisition sources to allow for flexibility in the mask design.

Configure gmskcreate as follows:

indata     = test
gprgid     = GS-2017A-Q-1
instrument = gmos-s
rafield    = 201.68298
decfield   = -47.64762
pa         = 35
fl_getxy   = yes
fl_getim   = yes
inimage    = sdss.fits
iraunits   = degrees
fraunits   = degrees

The output files are:

  • GMItest - A file containing the GMOS-S x,y coordinates for the pseudo-image

  • GMItest_OT.fits - The Object Table

  • GMIsdss.fits - The pseudo-image

  • gmos.log - a log file

GMItest_OT.fits contains

#     ID      RA        DEC           x_ccd     y_ccd       MAG       priority   slitsize_x   slitsize_y  slittilt  slitpos_y
#             H         deg           pixels    pixels      mag                  arcsec       arcsec      degrees   arcsec
1     10      13.44515  -47.65157     3291.15   2561.97     17.       2          1.0          15.0        0.0       3.0
2     11      13.44589  -47.64194     2924.58   2088.10     17.       1          1.0           5.0        0.0       0.0
3     12      13.44450  -47.65392     3615.19   2677.60     17.       2          1.0          15.0       -6.0       0.0
4     13      13.44562  -47.66114     3058.21   3033.21     17.       1          1.0           5.0        0.0       0.0
5     14      13.44742  -47.64720     2163.62   2347.12     17.       3          1.0           5.0        0.0       0.0
6     15      13.44624  -47.62953     2747.09   1476.67     17.       1          1.0           5.0        0.0       0.0
7     16      13.44663  -47.63118     2552.11   1557.89     17.       1          1.0           5.0        0.0       0.0
8     17      13.44643  -47.66784     2655.82   3363.20     14.6      0          2.0           2.0        0.0       0.0
9     18      13.44749  -47.63685     2126.77   1837.18     14.3      1          1.0           5.0        0.0       0.0
10    20      13.44294  -47.68342     4390.63   4133.77     14.7      0          2.0           2.0        0.0       0.0
11    21      13.44902  -47.61178     1358.33    599.68     14.6      0          2.0           2.0        0.0       0.0

4.4.5. Data sources

Often, input catalogs and images for gmskcreate are related. E.g., both catalog and image are based on a particular survey. This is not a requirement, though. One could also use e.g. the RA/DEC positions from the UCAC4 catalog, and an SDSS FITS image of the same area.

4.5. Recommendations and requirements

  • Targets and acquisition stars must be selected from the same object catalog with relative astrometry better than 0.1”. Otherwise it is most likely that the mask design is critically flawed (partial or full slit losses).

  • At least 2 acquisition stars are required for masks created from pre-imaging.

  • At least 3 acquisition stars are required for masks created from pseudo-imaging.

  • More than 4 acquisition stars increase overheads unnecessarily.

  • It is a good idea to include a brighter star in one of the science slits, easily detected in a single exposure. Put a note in the phase II observing tool asking the observer to inspect the first science exposure, and to abandon the observations if no obvious spectrum from the bright star is seen. This is particularly useful if your science targets are very faint. It also allows you to judge the accuracy of the reduction and stacking of the spectra. The star might also serve for telluric calibration. For GMOS, select two stars near the right and left edges of the field to cover the whole wavelength range (not necessary for F2).


If the mask acquisition went fine and if the science targets are too faint to be seen in a single exposure, then we will continue observing. If the PI determines after the fact that the mask design was bad, then the time will be charged to the program and the partner country. A new mask design may be submitted to use the remainder of the time allocation.