Data dictionary for MOS ASCII table extension file

Keywords for coordinate systems are derived from the proposed WCS conventions and resolution R3 of the OGIP recommendations.

Those keywords which apply to all fibers and/or all objects in the table are stored in regular FITS header cards. Note the clear distinction between the catalog-specific and the MOS-specific keywords. This is the reason that the table data are separated into different tables.

Primary keys in the tables

Two kinds of primary keys are needed in order for entries in one table to refer to corresponding entries in another table. These keys occur in both the object catalog tables and the fiber configuration table. They also occur in other tables where they serve as foreign keys which reference into the object or fiber tables.
PIVOTID (I3) {TNULL = ' -1'}
Identity of the fiber pivot on the MOS hardware. PIVOTID is the primary key into the fiber table and must be unique. At the MOS focal plane each fiber is threaded through its own pivot. The pivots are arranged in ``banks'' of 10 along arcs of a circle centered on the telescope axis. The pivots are numbered sequentially starting with 0. Pivot number 100 has traditionally indicated the ``big'' guide bundle which is at the north edge of the MOS field.
PIVOTID has a permanent significance so long as the MOS plate is not re-engineered.
OBJ_PKEY (I6) {TNULL = ' 0'}
An unique index assigned to each object entry in the catalog tables. OBJ_PKEY will typically be sequential in the order that the objects were read from the specified catalogs. It has no permanent significance. It serves merely to permit associations between entries in different tables.

General keywords

DATExxxx (A32)
MOS shall use the new form of FITS DATExxxx keywords as proposed by Peter Bunclark on 1996 Nov 19. In short, the new format solves the Year 2000 problem by permitting dates with 4 digit years ('yyyy-mm-dd'). Optionally, it also allows time to be appended ('yyyy-mm-ddThh:mm:ss[.sss...]Z'). Caveat: This format has not yet been adopted by the regional FITS committees nor the IAU.
DATE (A32)
Time of preparation of the FITS header.
DATE_PNT (A32)
Intended time of the observation (UT). It duplicates the information in the MJD_PNT keyword. This is included solely for human readability.

Astrometric keywords

These keywords are related to the astrometry performed on the catalog inputs provided by the observer. They do not relate to the MOS hardware.

Note that the sky fiber positions are originally specified in MOS plate coordinates which are then translated back to refracted GAPPT coordinates; their catalog positions are calculated by the code.

Planetary motions will not be handled. General relativistic effects may be omitted. Precession, nutation, and annual aberration will be applied. This information permits a target selection algorithm to operate.

Header or Table Keywords

These keywords may be found in either the FITS header or as columns in the ASCII table. When they appear in the FITS header their values apply to all entries in the ASCII table. When they appear as columns each of their values applies to its own row only.
RADECSYS (A8) [as specified by FITS WCS] {TNULL = 'NULL '}
Reference frame system which applies to all RAs, DECs. (The values of the keywords EQUINOX and/or MJD_WCS are also required to specify the WCS fully.) For the user input catalogs it will typically be 'FK4' or 'FK5'. The existing fiber assignment programs apply precession, nutation, and annual aberration to all input catalog coordinates into geocentric coordinates of date. These coordinates are then refracted without first applying diurnal parallax and aberration. The WCS draft has no value for this improper scenario; we use 'REFGAPPT' to denote it.
EQUINOX (F13.6) [annum]
Date of the precessional epoch which applies to RADECSYS. According to the FITS WCS draft standard its value depends upon RADECSYS. When RADECSYS = 'GAPPT' the EQUINOX keyword is not required and the keyword MJD-OBS is required. In practice, if RADECSYS = 'REFGAPPT' EQUINOX will generally be the Julian epoch of the date of the observation. If RADECSYS is one of the other possibilities then EQUINOX applies to the catalog reference frame.
MJD-OBS (F17.9) [TAI diem]
The MOS programs do not use this keyword.
According to the FITS WCS draft this is technically the Modified Julian Date of the observation; however, as actually used by the WCS draft this is a required coordinate system keyword when RADECSYS is either 'GAPPT' or 'FK4'. MJD-OBS is not required if RADECSYS is 'FK5'.
The FITS draft WCS standard requires that this keyword serve these multiple meanings and that it be measured in International Atomic Time (TAI). Because such a requirement makes it impossible for FITS to encode data acquired before 1955 there remains some debate over this aspect of the draft standard.
Operation of the MOS target assignment programs does not necessarily result in an actual observation. As catalogs are constructed from many observations over a large time there is further confusion if MJD-OBS is used in conjunction with an 'FK4' catalog. Thus, the MOS programs do not use this keyword.
MJD_WCS (F17.9) [diem] {TNULL = '-999999.000000000'}
This keyword is used in place of MJD-OBS when it is necessary to define the epoch at which a coordinate system is valid. The time evolution of WCS is slow enough that it is not necessary to require a precise time scale such as TAI. In the case of catalogs with proper motions this is also the epoch from which the motions are to be applied.

Header Keywords

These keywords are found in the FITS header preceding the ASCII table.
RA_PNT (F11.7) [degree]
Right Ascension of Shane telescope axis as specified by the fiber assignment code. (The RA of the Shane as indicated by its position encoders will be found in the FITS file containing the CCD image.)
DEC_PNT (F11.7) [degree]
Declination of Shane telescope axis as specified by the fiber assignment code. (The Dec of the Shane as indicated by its position encoders will be found in the FITS file containing the CCD image.)
HA_PNT (F11.7) [degree]
The hour angle of the field center to which the the fiber assignment programs precess, nutate, and annual aberrate catalog positions to the night of the observations. Differential refraction is then applied for this hour angle. (The HA of the Shane as indicated by its position encoders will be found in the FITS file containing the CCD image.)
RADECPNT (A8) [as RADECSYS]
In the case where a table has no single RADECSYS this keyword specifies the reference frame that applies to RA_PNT and DEC_PNT. When the fiber assignment code assembles the input catalogs into tables it leaves each coordinate in the system originally provided by the user. The fiber assignment code chooses a coordinate system in which to report its specification for the field center and documents that choice using RADECPNT. For the MOS this will uniformly have the value 'FK5'.
EQUINPNT (F13.6) [annum]
Date of precessional epoch appropriate for RADECPNT. In the case where a table has no single EQUINOX this keyword specifies the reference frame that applies to RA_PNT and DEC_PNT.
MJD_PNT (F17.9) [diem]
MJD_PNT is included as an alternative to the WCS MJD-OBS keyword. The MOS programs use this keyword to specify the time at which the observation is planned to take place. By strict interpretation of the WCS draft the MOS programs will not use this keyword to specify the WCS because RADECPNT will always be the inertial reference frame 'FK5' and there is no proper motion of the planned pointing axis. (If there were a case where RADECPNT defaulted to 'FK4' or 'GAPPT' then MJD_PNT would also specify the epoch of the PNT WCS.)
RA_3M (A13) [sexagesimal hour]
The Right Ascension which should be given to the Shane telescope technician for best pointing. This is an attempt to sidestep the current (1996) deficiencies in the Shane pointing model.
DEC_3M (A13) [sexagesimal degree]
The declination which should be given to the Shane telescope technician for best pointing. This is an attempt to sidestep the current (1996) deficiencies in the Shane pointing model.
EPOCH_3M (A11) [system and annum]
The current (1996) Shane pointing model was developed in 1976 by Rank et al. It predates modern IAU conventions for timekeeping and precession. The _3M positions are given in an attempt to sidestep the Shane this existing precession/nutation code. The value of this keyword will be the Besselian epoch of the (planned) observation.
These next batches of keywords describe various inputs needed for the atmospheric refraction model.
REFWAVEL (F12.10) [meter]
The atmospheric refraction model is applied for a specific observation wavelength. This should reflect the center of the desired spectral bandpass.
REFEPS (EN14) [dimensionless]
The convergence criterion (epsilon) used in the atmospheric refraction code from Pat Wallace's SLALIB.
Keywords describing the telescope.
TELESCOP (A8)
The name of the telescope.
TELGLAT (F11.7) [degree]
The geodetic latitude of the telescope.
TELGLON (F12.7) [degree]
The geodetic longitude of the telescope.
TELGSYS (A8)
Name of the geodetic datum used for telescope position.
TELHIGH (F7.1) [meter]
The elevation of the observatory above sea level.
TELGHGT (F5.1) [meter]
The geoid height above the datum at the telescope.
TELALAT (F11.7) [degree]
The astronomical latitude of the telescope.
TELALON (F12.7) [degree]
The astronomical longitude of the telescope.
Keywords describing the atmosphere.
ATMPRES (F5.1) [mbar]
The atmospheric pressure at the observatory. This should be the actual pressure, not the pressure reduced to sea level.
ATMTEMP (F5.1) [Kelvin]
The ambient temperature at the observatory.
ATMHUMID (F5.3)
The relative humidity at the observatory.
ATMTTLAP (F7.5) [Kelvin/m]
The adiabatic temperature lapse rate in the troposphere above the observatory.

ASCII Table Column Keywords

These document the TTYPEn, TFORMn, TUNITn, and TNULLn, cards of the catalog ASCII tables.
UNAME (A32) [No embedded spaces]
A unique, relatively short name of an object. Used to tag object in simulators and observation planning software -- names as short as possible reduce clutter.
RA_OBJ (F11.7) [degree]
Right Ascension of an object in the coordinate system specified by RADECSYS, EQUINOX, and MJD_WCS. The coordinate system keywords may be in the FITS header of the table or in other columns of the table.
HA_OBJ (F11.7) [degree]
Hour Angle of an object in the coordinate system specified by RADECSYS, EQUINOX, and MJD_WCS. The coordinate system keywords may be in the FITS header of the table or in other columns of the table.
DEC_OBJ (F11.7) [degree]
Declination of an object in the coordinate system specified by RADECSYS, EQUINOX, and MJD_WCS. The coordinate system keywords may be in the FITS header of the table or in other columns of the table.
PRIO (I4)
Priority of object. This is used in automated target assignment algorithms to determine the desirability of acquiring each object. The value 0 indicates an object potentially useful as a guide star. Values 1 through 98 indicates target objects in decreasing priority. The value 99 indicates an object intended to be used as the field center. Values other than these may take on special meanings to the assignment software.
MAG (F7.3)
Magnitude of an object (in any consistent system). At present this is not used by the automated assignment algorithms.
NAME (A128)
Full name of an object including any commentary information
PM_RA (F9.4) [arcsec/(Julian annum)] {TNULL = '-999.9000'}
Proper motion in right ascension
PM_DEC (F9.4) [arcsec/(Julian annum)] {TNULL = '-999.9000'}
Proper motion in declination
PARALLAX (F7.4) [arcsec] {TNULL = '-9.9999'}
Parallax
RAD_VEL (F13.3) [m/s] {TNULL = '300000000.000'}
Radial velocity
CAT_FILE (A128)
Name of the file from which the data in the row were read

MOS hardware keywords

These keywords are related to the MOS hardware and/or the Shane telescope. They are not related to any astrometric corrections applied to the catalogs of objects.

Header Keywords

These keywords describe the machined components of the MOS. They are used when calculating whether or not a configuration will result in a collision.
BUTNDIAM (F8.6) [meter]
Diameter of the normal fiber buttons.
HYPODIAM (F8.6) [meter]
Diameter of the hypodermics in the pivots containing normal fibers.
GYDBDIAM (F8.6) [meter]
Diameter of the big guide button.
GYDFDIAM (F8.6) [meter]
Diameter of the hypodermic carrying the big guide bundle.
LEEWAY (F8.6) [meter]
Length of hypodermic which must remain outside the pivot to prevent dislodging it.

ASCII Table Column Keywords

These document the TTYPEn, TFORMn, TUNITn, and TNULLn, cards of the fiber ASCII tables.
SLITID (I3)
Position of fiber within ``slit''. At the entrance to the spectrograph the fibers terminate in parallel along a slit. This documents the order in which they appear on the CCD images. (Is this zero-indexed or one-indexed?)
FIBERID (A5)
Name of the fiber. During construction each fiber was individually identified.
OBJ_POSX (F9.6) [meter]
X position of the fiber button as specified to the MOS. This permits diagnosis of coordinate transformation problems which may correlate with low signal.
OBJ_POSY (F9.6) [meter]
Y position of the fiber button as specified to the MOS. This permits diagnosis of coordinate transformation problems which may correlate with low signal.
FIBSTAT (I3)
State of the fiber. A value of 1 indicates a normal fiber in working condition. A value of 0 indicates a broken (unusable) fiber whose pivot and hypodermic are still present and must be avoided. A value of -1 indicates a nonexistent fiber (no button or hypodermic to avoid). A value of 99 indicates the big coherent guide bundle. A value of 19 indicates a little hex-pack guide bundle.
PIV_POSX (F9.6) [meter]
X position of the fiber pivot as manufactured.
PIV_POSY (F9.6) [meter]
Y position of the fiber pivot as manufactured.
PRK_POSX (F9.6) [meter]
X position of the fiber button parking spot.
PRK_POSY (F9.6) [meter]
Y position of the fiber button parking spot.
HYPOLEN (F9.6) [meter]
Length of the hypodermic threaded through the pivot
ANGMAX (F7.3) [degree]
Maximum permissible angular deviation of the pivot from its parked location.

Coordinate system transformation keywords

All of the remaining keywords appear in the FITS header, not in the columns of the tables. These keywords describe the transformations which are applied to get from refracted geocentric apparent coordinates to MOS fiber positioner coordinates. It is evident that portions of the transformation have many different origins over a long period of time.

Correction for Polar Axis Alignment

The polar axis of the Shane is not precisely aimed at either the refracted or the actual pole. This induces small rotations in the field that are described using these keywords. Bob Kibrick notes that this simple model ignores any rotations caused by fork flexure of the Shane. These corrections are handled using the notation of Pat Wallace's TPOINT code.

Hank Donnelly's version of the assign code has used a MA-like term with a sign opposite to that found in the Pat Wallace TPOINT code. Fortunately the MA term for the Shane is zero, so this has had no effect.

TP_ME (F9.3) [arcsec]
The misalignment in elevation (ME) term of the TPOINT code. This is a rotation around the axis connecting E and W points on horizon. In the N hemisphere ME < 0 means telescope pole above unrefracted pole. In general: ME > 0 is a rotation in the direction of zenith->vacuum_NCP .
TP_MA (F9.3) [arcsec]
The misalignment in azimuth (MA) term of the TPOINT code. This is a rotation around the axis thru (HA=0,Dec=0) . In the N hemisphere MA > 0 means telescope pole is east of due north. In general: if azimuth is N->E, MA has sign of (actual az - pole az) .

Tangent Plane Projection

These keywords describe the transformation from refracted apparent coordinates to MOS tangent plane coordinates
ADC3MSCL (F7.4) [arcsec/mm]
The plate scale of the Shane prime focal plane field center with the ADC in place.
MOSANGLE (F6.4) [degree]
There is a small rotation between the north point of the Shane telescope structure and the positive Y axis of the MOS.

Distortion by the ADC

These keywords describe the transformation from MOS tangent plane coordinates to MOS plate coordinates. In 1994 June the Shane prime focus Atmospheric Dispersion Corrector (ADC) was used to expose several plates. The plates were held in a test rig intended to match the future position of the MOS plate. Burt Jones produced a plate solution that included a radial polynomial distortion. See that solution for definitions of these terms.

Note that the values of X0 and Y0 here were determined from the glass plates in a test rig. These probably do not bear much relation to the center of distortion on the MOS plate. Burt Jones has suggested that recalculation of this transformation at the beginning of each observing run might be a better idea than producing the polynomial warp.

ADCALPHA (E10.7) [mm**-2]
radial distortion coefficient
ADCBETA (E10.7) [mm**-4]
radial distortion coefficient
MOSADCX0 (F8.4) [mm]
X position of center of distortion
MOSADCY0 (F8.4) [mm]
Y position of center of distortion

Polynomial Warp

These keywords describe the 2-d polynomial warp which is applied after the tangent plane projection and distortion as calculated by Burt Jones. These terms may be compensating for scale differences or nonlinearities in the MOS positioner and for changes in the focal length of the Shane primary mirror.

The astute reader will note that this transformation alone could replace most of the previous steps after the polar axle correction.

MOSADCAn (E14.11) [integer powers of micrometers]
The index n runs from 0 to 5. The An coefficients describe the MOS-x part of a 2 dimensional polynomial warp which is applied to the MOS plate coordinates after all traditional astrometric corrections. The equation in use:
x_mos = a0 + x * a1 + y * a2 + xy * a3 + x**2 * a4 + y**2 * a5
MOSADCBn (E14.11) [integer powers of micrometers]
The index n runs from 0 to 5. The Bn coefficients describe the MOS-y part of a 2 dimensional polynomial warp which is applied to the MOS plate coordinates after all traditional astrometric corrections. The equation in use:
y_mos = b0 + x * b1 + y * b2 + xy * b3 + x**2 * b4 + y**2 * b5

MOS flop

The positive Y axis on the MOS has always pointed very nearly to the north point on the sky. In the early implementation of the MOS as mounted at Cassegrain on the Nickel 1-m telescope the positive X axis pointed almost due east. In the final version of the MOS at the Shane the positive X axis points almost due west. The code handles this by optionally negating the X component. If West is Positive then the X coordinate is negated.
MOSWESTX (L1)
Does the positive X axis of the MOS point west?
Back to the MOS Data Products page.
Back to the MOS home page.
Steve Allen <sla@ucolick.org>