The Second U.S. Naval Observatory CCD Astrograph Catalog (UCAC2)

Table of Contents

1) Introduction

2) Summary of the Project
  a) Observations and Instrumentation
  b) Reductions of UCAC Observations
  c) Computation of Proper Motions

3) Properties of the Catalog and Important Notes For the User
  a) Sky Coverage
  b) Completeness
  c) Positions, Proper Motions and Error Calculations
  d) UCAC2 Magnitudes
  e) 2MASS Photometry
  f) Non-stellar Data
  g) Provided Utility Software
  h) Arrangement of the Data Files
  i) Star Identification Numbers

4) Files Contained on the CDs
  a) File readme.txt
  b) Directory "info"
  c) Directory "sw"
  d) Directory "u2"

5) Data Formats
  a) Files z001 to z288 (directory "u2")
  b) File u2index.da (directory "u2")
  c) File u2index.txt (directory "info")

6) Description of Utility Software

7) Acknowledgment

8) References


1) Introduction
This is the "readme" file of the Second U.S. Naval Observatory CCD
Astrograph Catalog, UCAC2.  Each of the 3 CDs contains this readme 
file as well as an identical copy of the "info" directory (index files
and plots) and the "sw" directory (utility software).

The UCAC2 is the second release of the ongoing UCAC project, designed 
to observe the entire sky for R magnitudes of about 7.5 to 16.  The 
observed positional errors are about 20 mas for the stars in the 10 
to 14 magnitude range, and about 70 mas at the limiting magnitude 
of R ~16.  For up-to-date information on the project, see our web page
at .  This web page will also serve as
the location that the UCAC team will post addenda to the UCAC2 catalog.

The UCAC2 is a high density, highly accurate, astrometric catalog of 
48,330,571 stars covering the sky from -90 to +40 degrees in declination
and going up to +52 degrees in some areas.  The northern limit is a
function of right ascension.  Proper motions and photometry are
provided for all stars.  Positions and proper motions are on the
ICRS (International Celestial Reference System) and given at the epoch

The UCAC2 has a number of major differences with respect to UCAC1.  
These differences include:  
 - much larger sky coverage
 - reduced systematic errors of CCD observations
 - positions given at a standard epoch (J2000.0)
 - the addition of several new catalogs for improved proper motions
 - photometry in the J, H, and K_s bands from the 2MASS project
 - data in binary format for direct access 
 - inclusion of software to aid users in quick access of the data  
Additional details of the data are found in Sections 3, 4, and 5 of
this document.

UCAC2 is the last intermediate data release before the final, all sky 
catalog will be constructed.  Observations will end around mid 2004; 
the final catalog is expected out in 2005.

The UCAC project, observations, and first data release are described in
detail in the paper Zacharias et al. AJ 120, 2131 (2000).  Versions 
of that and other related papers are placed on our Web page
(  The paper describing UCAC2 is in
preparation (AJ, 2003).  The data are freely available; however, any
publication utilizing these data should make proper reference to the
project and papers.

2) Summary of the Project

2a) Observations and Instrumentation
The UCAC is an observational program, using the U.S. Naval Observatory
Twin Astrograph and a 4k by 4k CCD camera, covering just over one square
degree per frame with a scale of 0.9"/pixel.  The red-corrected, 20 cm 
aperture, 5-element lens of the astrograph provides a 9 degree (diameter)
field of view (designed for photographic plates), thus only a fraction is 
utilized with our CCD camera, centered on the optical axis.

The CCD is a thick, Kodak device with 9 micrometer pixel size.  The 
camera, made by Spectral Instruments, is peltier cooled to -18 C.  
The raw data are severely affected by a low charge transfer efficiency
(CTE) of our otherwise cosmetically excellent CCD chip.  To mitigate this
problem, a relatively warm operating temperature is used, causing a
significant dark current.

Observations started in January 1998 at Cerro Tololo Interamerican 
Observatory (CTIO) in Chile, where the entire southern sky and ~ half
of the northern sky were observed.  In October 2001 the instrument was
moved to the Naval Observatory Flagstaff Station (NOFS) in Arizona to
continue the northern sky.  Complete coverage from pole to pole is
expected by mid 2004.  This second data release contains data observed
through 08 December 2002 at which time about 86% of the entire sky was

A 2-fold, center-in-corner overlap pattern was adopted on a 0.5 degree
grid, starting at the South Celestial Pole.  Each field was observed 
with a long (100 to 150 sec) and a short (20 to 30 sec) exposure.
Extensive quality control routines led to the rejection of over 15% of
the frames taken.  All raw (pixel) data are saved.

Observations are made in a single bandpass (579-642 nm), thus the UCAC 
magnitudes are between Johnson V and R.  No attempt has been made to 
obtain high quality, photometric data from the CCD observations.  In
fact, observing progresses even during nights with thin cirrus clouds.  

The telescope is actively guided with an ST-4 autoguider mounted behind 
the second, visually-corrected lens of the twin-astrograph.  Operation 
is automated using a PC and a single board computer.  An HP-Unix 
workstation is used for on-line reductions to obtain quality control 
statistics in near real-time.

2b) Reductions of UCAC Observations
Unlike the UCAC1 which utilized Hipparcos and ACT catalog stars, the 
UCAC2 was processed using Tycho-2 reference stars.  This doubled the 
number of available reference stars per frame, however, positional 
errors of individual faint stars in Tycho-2 are significantly larger 
than the x,y errors of UCAC observed images.

A special effort went into dark-correcting frames.  Frequent darks are
taken for all standard exposure times and camera vacuum pressures.
Twilight flats have been taken throughout the project, however, they
have not been applied for the UCAC2 processing.  Observed profiles of
stellar images are fitted with a two-dimensional Gaussian model.  The
obtained x,y centroid positions of the stars are then corrected for
various effects, including:
 - CTE (Charge Transfer Efficiency), this manifests itself as magnitude
   and coma-like terms for the x coordinate, i.e. RA;
 - FDP (field distortion pattern), inhomogeneous filter,
     tilt of the CCD chip w.r.t. the optical axis, optical distortion;
 - pixel phase error (as a function of the pixel fraction coordinate
     and the widths of the image profiles).

In the actual astrometric "plate" reduction step itself, a minimal 
model (only linear terms) was used to keep error propagation small.
Effects from aberration and refraction were removed prior to a least-
squares fit of the x,y data to the reference stars' mean positions at
the CCD frames epoch.

As with UCAC1, all "problem cases" were excluded from the catalog.
In particular, many multiple stars are excluded in UCAC2 as their images
are often elongated.  Precise centroiding on such objects is difficult
with our current software.  It was decided to use the same centroiding 
routines as with UCAC1 and remove these "problem" data.  This software
is being modified; re-processing of the pixel data is planned for the
final catalog release.

Saturated images of stars were not used, thus no bright stars are
included.  Depending on atmospheric conditions, the cut-off at the
bright end is around R magnitude 8.  Weighted mean positions were 
produced from the individual positions.  Only stars with 2 or more 
images are included in this catalog.  Stars with a formal standard 
error (derived from the scatter of the individual images of a star)
exceeding 200 mas in either coordinate were excluded.  Those stars are
either very faint, blended images, or have other miscellaneous problems.

2c) Computation of Proper Motions
Proper motions of bright stars (R ~8 to ~12.5) were derived using a 
combination of ground-based photographic and transit circle catalogs, and
included satellite observations from the Hipparcos and Tycho-2 catalogs.
In addition, the U.S. Naval Observatory is in the process of re-measuring
the AGK2 photographic plates, taken in Bonn and Hamburg (Germany) around
1930.  About 600,000 new catalog positions from this re-measurement 
are included in the area north of +20 degrees declination.  

For the faint stars (~12.5 to ~16), data from the Lick Northern
Proper Motions (NPM, Jones et al. 2000) and the Yale Southern Proper
Motions (SPM, van Altena et al. 1999) are utilized.  These plates were
measured on the Precision Measuring Machine at USNO Flagstaff Station
and initially reduced by D. Monet.  For his reductions, Monet used only
the yellow series, so the data are referred to as the "Yellow Sky"
(currently unpublished).  The data were subsequently massaged
at USNO Washington by S. Urban to minimize systematic errors.  The use
of the Yellow Sky for the faint stars is a departure from UCAC1, which
utilized the USNO A2 catalog.  Not surprising, statistics show the
Yellow Sky -- observed using astrographs -- is about twice as accurate
as the A2  -- observed with Schmidt telescopes. 

There are 18,604 previously known high proper motion stars in the UCAC
position file.  These were identified by Gould and Salim utilizing the
New Luyten's Two-Tenths Catalog and graciously forwarded to the UCAC
team.  However, only those high proper motion stars with an early epoch,
astrometric position available were included.  There were 8282 stars
found in our standard catalogs (Yellow Sky, AC 2000.2, Tycho-2,
Hipparcos, other transit circle and photographic catalogs), while 7666
stars were supplemented utilizing positions from the USNO A2, thus a
total of 15,948 NLTT stars are in the UCAC2.  Note that these are the
only stars for which the USNO A2 is utilized, as we are trying to
minimize the reliance on Schmidt plate data in the UCAC2.

The computation of proper motions is performed similarly to the procedure
used for the UCAC1 and the Tycho-2 catalogs.  All input catalogs were 
reduced to the ICRF utilizing Hipparcos data or some denser, interim
catalog that follows the system of Hipparcos.  Standard errors for 
each position are estimated. These error estimates are used as weights
to compute a mean position and proper motion, which utilizes a weighted,
least-squares adjustment procedure.  Error estimates of UCAC2 astrometry
are provided. 

Errors in proper motions of the bright stars (to R ~12) run from about
1 to 3 mas/yr mainly due to the large epoch spans involved.  For the 
fainter stars using exclusively Yellow Sky data, typical errors are 
estimated to be 4 to 6 mas/yr.

All stars in the UCAC2 have proper motions.  Stars are not included if:
 - they were not observed in the astrograph CCD program;
 - they were observed, but the UCAC positional reduction failed for 
     some reason;
 - the epoch span between first catalog position and UCAC does not 
     exceed 10 years; 
 - no proper motion could be computed (no one-to-one match found in
     other position catalogs).
 - proper motions exceed 180 mas/yr but stars were not identified in 
     the revised NLTT or Hipparcos Catalogue;
 - errors in proper motion based on scatter of the individual 
     positions exceed 20 mas/yr.

All totaled, 48,330,571 stars of the 58,843,138 stars observed with
the astrograph up to December 8, 2002 are included in UCAC2.

3) Properties of the Catalog and Important Notes for the User

3a) Sky Coverage
Included in UCAC2 are all areas of the sky from the South Celestial
Pole to between +40 and +52 degrees declination.  The northern limit 
is a function of right ascension.  The plots (black and 
white) and skycov_u2c.gif (color) in the "info" directory show all
survey fields which are included in this release (see Section 4b for
further details on these plots).  These are all fields taken up to the
night of 07/08 December 2002.  The plot and
skycov_nofs.pdf show only the northern fields completed at US Naval 
Observatory Flagstaff Station up to the same date.  The data for these 
plots are much fewer but show the same northern limit of the UCAC2.
Their southern limit shows the border line between observing performed
at CTIO and NOFS.

3b) Completeness
The UCAC2 is NOT complete, neither in the above mentioned sky coverage
(above +40 degrees), nor for individual magnitude ranges for a variety
of reasons:

 - Stars whose images were saturated on the detector are not included.
     If a star is brighter than R ~8.0, it is likely not in the UCAC2.

 - Multiple stars with separations of 3 arcsec or less are excluded 
     by design (unique matching requirement).  Most multiple stars
     in the 3 to 6 arcsec separation range are likely missing due
     to blended images.

 - All "problem stars", e.g. with large internal errors in the image
     combination step, are excluded.

 - Stars with fewer than 2 observations meeting the UCAC quality 
     standards are excluded.  Single detection observations are not used
     at this time.

 - Stars where a one-to-one match was not made with the proper motion
     input catalogs are excluded.  Many of these will be in the final 
     UCAC catalog.

 - If a star's proper motion was computed to be larger than 180 mas/yr 
     in either coordinate, but it is neither a known NLTT star nor 
     found in the Hipparcos Catalogue, it is excluded from the UCAC2.  

 - If a proper motion could not be determined due to lack of early 
     epoch data, either it being non-existent or less than a 10 year 
     span from the UCAC observation, the star is excluded.

 - If the proper motion error based on the scatter in the positional 
     data exceeded 20 mas/yr in either coordinate, it is excluded. 

3c) Positions, Proper Motions, and Errors
The astrometry provided in the UCAC2 is on the Hipparcos system, i.e. 
the International Celestial Reference System (ICRS), which is 
consistent with the (old) J2000.0 equator and equinox.  Positions are 
given at the standard epoch of Julian date 2000.0, thus the UCAC2 is  
a compiled catalog.  In order to be able to calculate positional errors 
at any epoch, the central epoch, i.e. the weighted mean epoch of the data 
(UCAC + early epoch) is given.  At the central epoch (which varies from
star to star and is also different for RA and Dec) the positional error
has its smallest value given in the catalog.  In most cases this central 
epoch will be close to the UCAC observational epoch due to the 
relatively large weight given to the UCAC observations.  (The UCAC2 
observational epochs are in the range of 1998.1 to 2002.9.)  The proper 
motions are given at the central epoch.

3d) UCAC2 Magnitudes
The UCAC2 magnitude range is about 8 to 16 in a 579-642 nm bandpass. 
This is between visual (V) and red (R).  The limiting magnitude can 
vary by about +-0.3 mag.  The bright magnitude cut-off varies even
more, depending on seeing conditions at the time of observation. 

UCAC magnitudes are only approximate with an estimated error of 
about 0.3 mag (absolute).  The relative precision over < 0.5 degree 
and medium magnitude range can be much better, ~0.1 mag.  Note: 
some data were obtained in non-photometric conditions and no "aperture 
photometry" type reductions were made.  The UCAC2 magnitudes are only 
provided for identification purposes.  

3e) 2MASS Photometry
The procedures to derive the UCAC2 observed photometry is unchanged
from the UCAC1 procedures, giving only very approximate R magnitudes
(see Section 3d).  However, for the UCAC2 release, the data from the
Two Micron All Sky Survey, 2MASS, has kindly been made available to us
by R. Stiening, shortly before the public release.  2MASS photometric
data, i.e. the J, H, and K_s magnitudes, and cross references 
(star ID numbers in the 2MASS point source catalog) are provided for 
matched stars as a convenience to the user.  These numbers are copied 
directly from 2MASS; for more information (flags, errors on photometry
etc.) the 2MASS catalog should be consulted: .  
In addition to J, H, and K_s magnitudes a modified "ph_qual" and "cc"
(contamination) flag has been copied into UCAC2.  A match radius of 
2 arcsecs has been used between the UCAC and 2MASS positions (no proper
motions were available at that time), which still could lead to a few 
wrong identifications for binaries on that separation scale.  In total,
48,079,174 stars have 2MASS photometry listed, which is 99.5% of the
UCAC2 stars.

3f) Non-stellar Data
The UCAC2 contains some galaxies, particularly at the faint end.  
No flag indicating a galaxy or star is provided with this release; 
however, extended objects are very unlikely to be in the UCAC2 due to
the detection and reduction quality control procedures adopted for the
catalog construction.  Also, galaxies of integrated magnitudes of ~15
or fainter are likely to show  cores fainter than 16th magnitude thus
will likely not be in the UCAC2.

Some asteroids might be hidden in the UCAC2.  The observing schedule
actively avoided all major planets and bright asteroids (to 12th mag).
However, asteroids in the ~12 to 14 mag range can appear on both the
long and short exposure taken within 2 minutes and could give a
satisfactory position match.  Fainter asteroids could enter the
UCAC2 only if overlapping fields with the object were taken within
a short period of time, which might be the case for about 30% of the
fields.  Many overlapping fields are taken the following night, with a
possible asteroid having moved far enough that it will not be matched
on at least 2 frames, hence it will not be present in the catalog 
(2 detections of an object are required to be in UCAC2).  Additionally, 
with the requirement that each object have a proper motion using an 
older catalog, it is unlikely that many asteroids are present in the 

3g) Provided Utility Software
Some utility software is included with the CD-ROMs.  It is not
required to use the data;  however, this software serves as an example
for a fast access to the data.  No guarantee or support is provided.
See Section 6 below for more details.

3h) Arrangement of the Data Files
The data are arranged in declination zones of 0.5 degree width, with 
zone number 1 starting at the South Pole and increasing as one goes 
north.  There is a separate file for each zone; there are 288 data
files in total.  Due to the amount of data, 3 CDs were required, each 
with about 680 MB of mostly binary data.  Records in the individual 
data files are sorted by right ascension.  One record contains the 
data for one star.

3i) Star Identification Numbers
Official UCAC2 star ID numbers should be utilized for identification
purposes and for communication with the UCAC team.  However, these
numbers are not explicitly stored in the binary files but are generated
when extracting data using the provided utility software.  These star
numbers increase along RA within a given declination zone of 0.5 degree
width, and increase with declination zone starting from the South Pole.
If the provided utility software is not used the UCAC2 star ID number
can be generated from the provided index files using the following

  ID = n0 + running star number in zone z

where z is zone number where the star in question is located
(1 to 288; zone 1: -90 to -89.5 decl., ...), n0 is the largest star
ID number of the previous zone and ID is this official UCAC2 star
ID number.  To get the value of n0, one needs "nx", the index array
provided in the file u2index.da (also available in the ASCII text file,
u2index.txt).  The nx array is a 2-index array containing the largest
ID number in each area of the sky. The first index is the zone number
as described above; the second is the right ascension bin of 0.1 hours
width.  Hence, nx has (288x240) elements.  For z=1, n0 = 0.  For any
other value of z, n0 = nx (z-1,240).

If needed, this 8-digit star ID number should be given in connection 
with the IAU sanctioned acronym of this project and this data release 
number, like "2UCAC12345678".  The star ID number runs from 00000001 
to 48330571.  Leading zeros are to be used to fill 8 digits in the 
star name.  Also, the position of a star can be used as an identifier. 

4) Files contained on the CDs
All 3 CDs contain the same directory structure.  There is a 
"readme.txt" file (this file), and 3 directories; "info", "sw", and
"u2".  The "readme.txt" file, the "info" directory, and "sw" directory
are identical on all CDs.  However, the contents of the data directory
"u2" are different.  Here is a list of all files:

4a) File "readme.txt"
  readme.txt = This file, it is the only file outside a subdirectory.
               This contains general information on the catalog, the 
               observing program from which the catalog is made,
               information about the CDs, data formats, and references.

4b) Directory  "info"
This directory contains information which may be beneficial to the user.

  format.txt = Format description of the main data records.
               This is a copy of Section 5 of "readme.txt" (this file).
               Due to its anticipated frequent reference, it was copied
               into a separate file for ease of the user. = Color plot of UCAC2 star distribution in RA,Dec
                near the South Pole (generated with  u2rdplot.f).  = Color plot of UCAC2 star distribution in RA,Dec
                for the Small Magellanic Cloud area.  Note the large
                number of faint stars (blue) and the absence of data
                in the core of the globular cluster due to crowding. = Black and white, high resolution, PostScript file
                 of a plot showing the sky coverage of the UCAC2 data.
                 Areas not covered by dots are not included in this

  skycov_u2c.gif = similar to, however in color and as
                   GIF file.  Colors represent approximately one calendar
                   year of observing starting from the beginning of the
                   project.  However, we decided to graphically show
                   when we moved the telescope from CTIO to NOFS.

                   The colors represent frames taken over the following
                   Yellow = mid-February 1998 to mid-February 1999;
                   Blue =   mid-February 1999 to mid-February 2000;
                   Red  =   mid-February 2000 to mid-February 2001;
                   Green =  mid-February 2001 to September 2001 at which
                            time the telescope was disassembled for
                            re-location to NOFS;
                   Pink =   Start up at NOFS October 2001 to October 2002;
                   Light blue = October 2002 to December 8, 2002 (latest
                   frames used in this release). = Similar to  Shows fields covered from
                   the U.S. Naval Observatory Flagstaff Station site,
                   i.e. the northern part of the survey.  The northern
                   limit shows the boundary of this UCAC2 release,
                   the southern limit shows the border between 
                   observations made at Cerro Tololo and Flagstaff.

  skycov_nofs.pdf = same as, but in PDF format.

  zone.txt   = Table of contents for the main data giving a list of zones.
               Each line of the table gives; the number of stars,
               accumulated number of stars, right ascension minimum and
               maximum values, and declination minimum and maximum values
               per zone. Note a zone is 0.5 degrees wide.

  u2index.txt = ASCII table of the full index, organized in declination
                zones and sections of 0.1 hour width along RA.
                Again, the number of stars per bin and the accumulated
                number of stars are given.  This information can be
                used to locate the desired section (record number) on 
                a zone file for a given area in the sky.  See Section 5c
                for further description of the data.
  u2chk.log  = Output of the u2chk program, containing statistics,
               and histogram tables of all items.  The range of values for
               each item is given as well as the distribution of the 
               central epochs and errors.

  u2tab1.txt = An ASCII table formatted sample of a selection of 14 stars.
               This was generated with  u2access  and covers stars in the
               5 to 14 mag range for RA = 23.9 to 0.1 hours, DE = -82.1 to
               -81.9 degrees.  Format 1 is used, i.e. all items from the
               binary file are output as integers.  
               The UCAC2 star identification number is appended as well.

  u2tab2.txt = The same as u2tab1.txt, except using format = 2, which
               gives positions updated for the epoch = 2003.0, RA, DE
               in degrees and decimals, along with proper motion
               information and errors (decimals).

  u2tab3.txt = The same as u2tab1.txt, except using format = 3, which
               gives RA, DE in hms and dms format, respectively.  Flags
               and 2MASS photometry are included in this format

  u2tab4.txt = The same as u2tab1.txt, except using format = 4, which
               gives a short table with only RA (hour, decimal), DE
               (degree, decimal), U2Rmag, error on position and star
               ID number.

4c) Directory  "sw"      (utility software, Fortran only)
  The utility software (Fortran source code) is not required to use
  the UCAC2 data.  The software serves as an example of efficient use
  of the data with fast, random access to arbitrary areas in the sky.
  It is provided for the convenience of the user only.  No support is 
  provided by USNO.  Additional UCAC2 access software in C/C++
  has been kindly provided by Bill Gray (GUIDE Software) and
  can be found on the web at
  Executable files of the 2 main access programs are provided for
  Linux (Intel Pentium processor), Windows (DOS), and HP-UNIX.  For 
  future additions, please see our web page .

  See Section 6 for a complete list and short description of the programs.

4d) Directory "u2"      (main catalog data)
  u2index.da = binary file with an index to all zones = "nx" array,
               see Sections 3i and 5b.  This file is used by some of the
               utility software and is expected in the same directory
               as the data files.  There is an identical copy on each
               of the 3 CDs.

      This information can also be obtained from the u2index.txt file 
      provided in the "info" directory.  The u2index.da file is used 
      for faster access with the utility programs.  Both index files,
      the ASCII and the binary version, can be re-created by the user
      by running the  u2chk  program (see Section 6).

  Files z??? = Binary data files containing the UCAC2 catalog.  See
               Section 5a for details on the data formats.  Note that
               each file contains data for a 0.5 degree wide zone, 
               beginning at the South Celestial Pole.  The CDs contain
               zones from different declinations.

               z001 to z106 =  -90.0 to -37.0 declination   on CD #1
               z107 to z182 =  -37.0 to +01.0 declination   on CD #2
               z183 to z288 =  +01.0 to +54.0 declination   on CD #3

      In addition, z001 is copied on CD #2 and CD #3.  It is used for a
      byte-flip check in the utility programs.

A recommended option is to copy the u2index.da and all data files 
from the 3 CDs into a single directory on hard disk in order to avoid 
the need of switching CDs.  The utility programs should work in either 
case.  If a needed data file is not present, a warning will be issued 
from the program.  The user may then change the CD and continue.  
A single path name for all binary data files is assumed.

5) Data Formats

5a) Files z001 to z288  ("u2" directory) 
These files constitute the main data of the catalog.  All items are 
(signed) integer. "I*1" means the data for these items are contained 
in 1 byte, "I*2" means 2 bytes, etc.  The order of bytes is for an 
Intel processor.  Each record contains the data for one star, totaling 
23 items.  The record length is 44 bytes, with no additional "end of 
line" or unformatted leading or trailing bytes.  In Fortran this is 
known as a direct access, fixed record length, unformatted file.
(Note, some compilers specify the record length as multiple of
4-byte segments, thus RECL=11 is required in that case,instead 
of RECL=44 which applies to other compilers.  RECL=44 is used in the 
utility software provided on these CDs.)

The following items are on each star record:

Num Bytes  Fmt Unit       Label   Explanation
 1   1- 4  I*4 mas        RA      Right Ascension at epoch J2000.0 (ICRS) (2)
 2   5- 8  I*4 mas        DE      Declination at epoch J2000.0 (ICRS)     (2)
 3   9-10  I*2 0.01 mag   U2Rmag  Internal UCAC magnitude (red bandpass)  (3)
 4  11     I*1 mas        e_RAm   s.e. at central epoch in RA (*cos DEm)(1,4)
 5  12     I*1 mas        e_DEm   s.e. at central epoch in Dec          (1,4)
 6  13     I*1            nobs    Number of UCAC observations of this star(5)
 7  14     I*1            e_pos   Error of original UCAC observ. (mas)  (1,6)
 8  15     I*1            ncat    # of catalog positions used for pmRA, pmDC
 9  16     I*1            cflg    ID of major catalogs used in pmRA, pmDE (7)
10  17-18  I*2 0.001 yr   EpRAm   Central epoch for mean RA, minus 1975   (8)
11  19-20  I*2 0.001 yr   EpDEm   Central epoch for mean DE, minus 1975   (8)
12  21-24  I*4 0.1 mas/yr pmRA    Proper motion in RA (no cos DE)         (9)
13  25-28  I*4 0.1 mas/yr pmDE    Proper motion in DE                     (9)
14  29     I*1 0.1 mas/yr e_pmRA  s.e. of pmRA (*cos DEm)                 (1)
15  30     I*1 0.1 mas/yr e_pmDE  s.e. of pmDE                            (1)
16  31     I*1 0.05       q_pmRA  Goodness of fit for pmRA             (1,11)
17  32     I*1 0.05       q_pmDE  Goodness of fit for pmDE             (1,11)
18  33-36  I*4            2m_id   2MASS pts_key star identifier          (12)
19  37-38  I*2 0.001 mag  2m_J    2MASS J  magnitude                     (13)
20  39-40  I*2 0.001 mag  2m_H    2MASS H  magnitude                     (13)
21  41-42  I*2 0.001 mag  2m_Ks   2MASS K_s magnitude                    (13)
22  43     I*1            2m_ph   2MASS modified ph_qual flag          (1,14)
23  44     I*1            2m_cc   2MASS modified cc_flg                (1,15)

As an example, the data of the first 5 stars from z001 are
printed here in ASCII.  These are the numbers obtained when
using the  u2dumpz.f  program.

   item  1          2    3   4   5  6  7  8  9    10    11      12
   1246420 -322767602 1591  75  87  2 97  2  1 17330 15213   31020
   4125230 -323707012 1382  15  28  8 21  2  1 23246 22659  134316
   7345118 -322447512 1579  24  26  4 25  2  1 22903 22799   45845
   8139385 -322284308 1479  53  25  5 41  2  1 20455 22911   18125
  11128880 -323115466 1631  15  43  2 27  2  1 23364 21562   39404

 item 13  14  15  16  17         18    19    20    21  22  23
     -35  70  74  20  20 1229086517 14428 13865 13751 000 000
    -187  61  62  20  20 1101364107 12467 12131 11963 000 000
     104  61  62  20  20 1329022468 14169 13752 13708 220 000
     -16  65  61  20  20 1085341332 13111 12511 12339 000 000
     -74  61  63  20  20 1085341834 14515 13918 13877 000 000

Note (1): In order to fit within the range of a 1 byte (signed) 
    integer (-127 to +127), some items have an offset value of -127 
    applied.  These are items #4,5 (e_RAm, e_DEm), #7 (e_pos),
    #14-17 (e_pmRA, e_pmDE, q_pmRA, q_pmDE), and #22,23 (2m_ph, 2m_cc).
    When using the provided software, the data items are converted to
    I*4 and automatically corrected for this offset on the required

Note (2):  Positions are on the International Celestial Reference 
    System (ICRS) as represented by the Hipparcos / Tycho-2  catalogs.
    The epoch for the positions of all stars is J2000.0; the weighted 
    mean catalog position was updated using the provided proper 
    motions.  The observational UCAC position is but one of several 
    going into these values and is not given in the UCAC2, thus the 
    original UCAC observation cannot be recovered from these data.

Note (3):  Values for U2Rmag should be considered approximate.  These 
    are magnitudes in the UCAC bandpass (579-642 nm), which is between 
    V and R.  Observations were often made under non-photometric 
    conditions and U2Rmag is provided here for identification purposes
    only.  Estimated errors are 0.3 magnitude absolute, 0.1 magnitude 

Note (4): Values for e_RAm and e_DEm are internal standard errors in 
    position at the central epoch, NOT at the epoch J2000 (see Note 8

Note (5): nobs is the number of observations (images per star) used 
    for the mean observational position (not provided) of the UCAC 

Note (6): e_pos  is the internal, formal, 1-sigma error in position
    of the UCAC astrograph observations, (mas).  The mean of the error
    in RA*cos(DE) and DE is given.  This error is derived from the
    scatter of individual positions which form the mean observed
    position from the CCD observations of each star.  These values are
    subject to small number statistics; many of the stars have only
    2 to 4 observations.  On average, for stars in the 7.5 to 10 mag
    range, 30 mas is expected, for 10 to 14.5 about 15 to 20 mas
    and increasing up to 70 mas toward 16th magnitude.

Note (7): cflg is an identifier of some major catalogs used in the 
    computation of proper motions and also provides the user with 
    information regarding known, high proper motion stars.  However, 
    this does not contain all catalogs going into the proper motions,
    about 150 were used in this project.  The cflg value is actually 
    a sum of numbers, each with the following meaning:

     1 = modified Yellow Sky (NOFS scans of NPM, SPM; not published)
     2 = AC2000.2 Catalog
     4 = Tycho-2 Catalogue observed position
     8 = USNO measurement of AGK2 plates (not published)
    16 = Hipparcos Catalogue
    32 = Known high proper motion star from the revised New Luyten's
         Two-Tenths catalog (NLTT), (match w/ UCAC provided by A.Gould).
    64 = USNO A2 Catalog

    For example, a cflg value of 26 means 16+8+2, hence at least
    Hipparcos, USNO AGK2 measures, and the AC2000.2 catalogs were used
    in determining the proper motion.

Note (8): EpRAm and EpDEm are the central epochs of the weighted mean 
    position from all catalogs going into the proper motion.  This is 
    neither the mean UCAC observational epoch, nor "J2000.0".  It is 
    the epoch with the lowest positional errors (e_RAm, e_DEm).  For 
    all other epochs (including "J2000.0" which corresponds to the RA,
    DE positions given here), the error in position can be computed as:

       e_RA (epoch T) = SQRT ( e_RAm**2 + (e_pmRA * (T-EpRAm))**2)

    for the RA component, and similarly (using e_DEm, e_pmDE, and 
    EpDEm) for declination.  Thus for all epochs other than EpRAm or 
    EpDEm the positional errors for RA and Dec, respectively, increase 
    for both going forward or back in time.  Values for EpRAm and 
    EpDEm are usually not equal due to different weights assigned to 
    coordinates in the individual catalogs used in determining the 
    mean positions and proper motions.

Note (9): pmRA and pmDE are the proper motion in right ascension and
    declination, respectively, referred to the central epoch.  Note
    that pmRA is without cos(Dec), thus these values can be large for
    stars close to the pole.  This is done so the units for pmRA and
    RA are similar, thus facilitating easy computation to change epochs.
    Both pmRA and pmDE can directly be used to update a star's position
    for proper motion by:

       RA (epoch T) = RA + pmRA * (T - 2000.0)/10.0
       DE (epoch T) = DE + pmDE * (T - 2000.0)/10.0

    Note, values for RA, and DE provided here are in mas, while the
    pmRA and pmDE provided here are in units of 0.1 mas/yr.  Also note 
    that the above formulas do not account for so-called "second-order"
    effects, such as the change in the apparent direction of the pole 
    as seen from a star as proper motion is applied or the apparent
    acceleration of a proper motion due to the distance to a star
    changing.  The first effect is negligible in most instances unless
    a star has a very high proper motion and is very close to the pole
    and that proper motion is being applied for many years.  The second
    affect is negligible in most cases unless a star is nearby and
    has a large radial velocity (often this is unknown).

Note (10): e_pmRA and e_pmDE are the standard errors of pmRA, pmDE
    respectively.  Note e_pmRA does include cos(DE), which makes
    both e_pmRA and e_pmDE comparable in the tangential plane. The 
    model value is given here, thus it is a formal error based on the 
    estimated precisions of the individual catalog positions used, as 
    well as their epochs. 

    In addition to these "model" errors, "scatter" errors have been 
    calculated for each star having 3 or more catalog positions used 
    in the proper motion determination.  This "scatter" error depends 
    on how well the individual catalog positions fit the computed 
    proper motion.  The "scatter" errors in position and proper motion
    component for each star are not explicitly given here, however the 
    scatter error proper motion values can be recovered using the
    values q_pmRA and q_pmDE.  See Note (11).

Note (11): q_pmRA and q_pmDE are termed "goodness of fit" for the pmRA 
    and pmDE.  These value give information on how well the computed 
    proper motions actually fit the catalog positions used to compute 
    them, termed "scatter error". The values q_pmRA and q_pmDE are 
    actually ratios, "scatter error pm"/e_pmRA (or e_pmDC).  Thus, 
    when the actual fit of the proper motion to the catalog positions 
    is better than the formal, model errors (e_pmRA, e_pmDE), the 
    ratio is smaller than 1.  In the opposite case, the ratio is 
    larger than 1.  Stars with q_pmRA or q_pmDE values exceeding 5 
    should be used with caution.  If the ratio is larger than 12.5
    it is set to 12.5.  The ratio of "scatter error" to "model error"
    of the mean positions is highly correlated to those of the proper

    Note that the "scatter error" only makes sense for stars with 3 or
    more catalog positions; if only 2 catalogs were used for the
    determination of the proper motion, the "fit" is perfect (going
    directly through each of the 2 positions, regardless of weights).
    In these cases, the ratios are set to 1.0.

    The actual integer numbers coded in the data file for q_pmRA and 
    q_pmDE are multiplied by 20 and then have 127 subtracted from them
    (see Note (1)).  Thus, for example, a q_pmRA value of +13 means 
    the value is (13+127)/20, equaling a ratio of 7.0, a value of -107 
    is (-107+127)/20, equaling 1.0.

Note (12):  2m_id is the "pts_key" star identification number found 
    in the 2MASS point source catalog.  It is a 4-byte integer number 
    greater than 0.  If no unique match to the 2MASS point source 
    catalog was found, 2m_id is set to 0.  This is the case for a
    total of 251,397 stars (= 0.5%).

Note (13) 2m_J, 2m_H, 2m_Ks are the J, H, and K_s magnitudes taken 
    directly from the 2MASS catalog with full precision.  No error 
    estimates are given in UCAC2, but they can be extracted from the
    2MASS catalog.  For stars not matched, all three magnitudes are
    set to 30.0 (a value of 30000 on the UCAC2 record).

Note (14): 2m_ph is a modified 2MASS "ph_qual" flag, which is the
    photometric quality flag given in the 2MASS catalog.  For each 
    color (J, H, K_s) one digit is given and has the following meaning:

    0 = ph_qual in 2MASS is A or B, meaning good detections (snr > 7)
    1 = ph_qual in 2MASS is C or D, meaning weak detections
    2 = ph_qual in 2MASS is something else and the 2MASS documentation 
          should be consulted.

    Note that the values in 2m_ph also have 127 subtracted from them 
    (see Note (1)).  For example  2m_ph =-115 is decoded to (-115+127)
    = 012, meaning good detection in J, weak in H, and a problem in K.

Note (15): 2m_cc is a modified 2MASS "cc_flg", which indicates the 
    amount of contamination and confusion seen in the 2MASS data.  
    Similar to 2m_ph (see Note (14)), each color (J, H, K_s) is 
    assigned one digit which has the following meaning:

    0 = cc_flg in 2MASS is 0, no artifacts or contamination seen
    1 = cc_flg in 2MASS is "c", confusion with nearby brighter source
    2 = cc_flg in 2MASS is something else and the 2MASS documentation 
           should be consulted.

    Note that the values in 2m_cc also have 127 subtracted from them 
    (see Note (1)).  For example  2m_cc = 83 is decoded to (83+127)
    = 210, meaning a problem in J, contamination in H, and no 
    contamination seen in K.

5b) File u2index.da    ("u2" directory)
This binary file contains a the "nx" array.  This file is used by some
of the utility software and is expected in the same directory as the
data files.  There is an identical copy on each of the 3 CDs.

     nx (288,240) = on 288 direct access records;
                    each record of 240 x 4 bytes (RECL=960);
                    total = 276,480 bytes               

     nx (z,jj) = accumulated number of stars for zone z and RA bin jj

This information can also be obtained from the u2index.txt file 
provided in the "info" directory.  The u2index.da file is used 
for faster access with the utility programs.  Both index files,
the ASCII and the binary version, can be re-created by the user
by running the  u2chk  program (see Section 6).

5c) File u2index.txt   ("info" directory)
This ASCII data file contains information regarding the number of 
stars in each area of sky.  These areas are arranged in declination 
zones (each zone is 0.5 degrees wide, numbering starts at the South 
Celestial Pole) further divided by 0.1 hour bins in RA.  For this 
release, the data only go to +54 degree declination, so there are only 
288 declination zones, each of which are divided into 240 RA bins.
Total areas are 288 * 240 = 69120.

The first few lines in the  u2index.txt  file give a short explanation
followed by the main table records like:

 nsbin     naz      nat  zn  jj  DCmax RAmax
     2       2        2   1   1 -89.5  0.1
     2       4        4   1   2 -89.5  0.2
     4       8        8   1   3 -89.5  0.3
     2      10       10   1   4 -89.5  0.4
     1      11       11   1   5 -89.5  0.5
The following format is used for the main table records:

Num Bytes  Fmt Units Label Explanation
 1   1- 6  I6        nsbin number of stars in that bin (zn,jj)
 2   7-14  I6        naz   number of stars accumulated within zone zn
 3  15-23  I9        nat   number of stars accumulated total
 4  24-27  I4        zn    zone number (see file zone.txt)
 5  28-31  I4        jj    Right ascension index (0.1 hour wide boxes)
 6  32-37  F6.1 deg  DCmax largest Declination in bin (zn,jj)
 7  38-42  F5.1 hour RAmax largest right ascension in bin (zn,jj)

6) Description of utility software
The source code (Fortran, F77 + extensions) of all programs can be
found in the "sw" directory on each CD.  More explanations are also
provided as comments in the source code files.  For some compilers,
you need to substitute the declaration BYTE by INTEGER*1.  If writing
your own software, you might want to take advantage of the low-level
access routines  open_zfile,  read_u2line, and cat_id, in the u2sub.f

u2sub.f   : contains various subroutines used in the following
            utility programs, could be used for basic level access 
            in user software.  Notes contained in that file explain
            the individual subroutines and their functions.

u2chk.f  : - read all zone files (option: subset of files)
           - check if byte flip is required
           - get statistics and histograms of data items per star
           - output summary to ASCII file
           - option: generate index files (ASCII -- u2index.txt --
             and unformatted -- u2index.da -- were generated this way).

u2dumpz.f : - read all data of a single zone
            - dump integer values to ASCII file
            - option available for user selected, limiting magnitude
         This is a short, simple program and serves as an example
         of how to use the basic level routines.

u2access.f: - interactive access to UCAC2 data by area in the sky
            - select box in RA, Dec or center + width of box
            - select mag range
            - sort output by any of the 23 items
            - update of positions and errors to desired epoch
            - output stars in ASCII, various formats
            - user may define additional output formats

u2rdplot.f - plot of individual UCAC2 stars (areas) on RA,Dec grid
           - option for 2 projections (pole, box)
           - select stars by RA,Dec range and magnitude interval
           - 3 groups by magnitude for 3 different colors
           - requires PGPLOT subroutines

sorti.f:   - general subroutine to sort a 2-index integer array

u2dumpz    - executable files for Linux PC (Pentium processor)

u2dumpz.exe  - executable files for Windows (DOS)

u2dumpz.HP - executable files for HP-UNIX

No guarantee can be made for this software, nor can any kind of
support be provided.  Users are encouraged to take segments of these
examples and build their own software.

7) Acknowledgment
In a project such as this that has spanned several years and two 
continents, many people have been involved.  Here we list people 
and their main contributions to the project.

Ted Rafferty - Project leader, instrumentation specialist, observer.

Norbert Zacharias - Principle Investigator, observer, responsible for
  UCAC position reductions (from pixels to positions on the sky),  
  observing schedule, and external position comparisons.

Sean Urban - Proper motions Principle Investigator, preparation of
  other position catalogs including systematic error removal.

Marion Zacharias - Quality control, observer, position reductions.

Gary Wycoff - Proper motions data preparation, star identifications.

David Hall - Verification of tape archive, data handling.

Marvin Germain - Telescope control system software development, 
  software-to-hardware interface including electronics.

Ellis Holdenried - Telescope control system software maintenance
  and extension of user interface for telescope operation.

Lars Winter - Basis for astrometric CCD reduction software.

Danilo Castillo - Observer,  Cerro Tololo, Chile 

Mauricio Martinez - Observer,  Cerro Tololo, Chile
Sergio Pizarro - Observer,  Cerro Tololo, Chile 

Oscar Saa - Telescope manager, Cerro Tololo, Chile

Trudy Tilleman - Observer, U.S. Naval Observatory Flagstaff Station

Stephanie Potter - Observer, U.S. Naval Observatory Flagstaff Station

Dominic Marcello - Observer, NOFS, joined the project in Jan. 2003 

John Pohlman - Telescope upgrades,  maintenance, and troubleshooting

Gary Wieder - Telescope upgrades, relocation, maintenance, and

John Bowles - Telescope upgrades and maintenance (USNO-DC)

Dave Smith - Telescope upgrades and maintenance (USNO-DC)

Tie Siemers - Telescope upgrades and maintenance (USNO-DC)

Mike Divittorio - Telescope maintenance and troubleshooting (NOFS) 

Steve Sell - Telescope maintenance and troubleshooting (NOFS)

Albert Rhodes - Telescope maintenance and troubleshooting (NOFS)

Blaize Canzian - Network and computer system support (NOFS)

Greg Hennessy - Computer hardware support (USNO-DC)

Other Acknowledgments
Special acknowledgment is due to Dave Monet for his measurements and 
initial reductions of the Lick Northern Proper Motions and Yale Southern
Proper Motion plates, heavily utilized for the proper motion work of 
the fainter UCAC stars.  Both Lick and Yale are thanked for their 
contribution to this project by providing access to their plates.
Samir Salim and Andy Gould provided identifications and initial proper
motions for the UCAC stars found in the New Luyten's Two Tenth Catalog.
Without these, many of the high proper motion stars would be missing
from the UCAC2.

This project would not have been possible without support from the USNO
administration.  Special thanks is given to Thomas Corbin for under-
standing the need for such a project, articulating it to those with 
resources, and administering the project in its early days.  Thanks 
is given to the three Heads of the USNO Astrometry Department during the 
UCAC project.  Without their continued support, it would not have come
to fruition.  They are F. Stephen Gauss, Ted Rafferty, and Ralph Gaume.
Thanks is given to the USNO Scientific Director, Kenneth Johnston, for 
his approval and support of the project and for making USNO resources 

Our successful stay at the Cerro Tololo observatory would not have 
been possible without a memorandum of understanding, worked out by 
P. Kenneth Seidelmann, USNO, and the director of CTIO, Malcolm Smith.
The entire staff of CTIO is thanked for their generous support.

Also, we wish to thank Gary Sims from Spectral Instruments for building
our 4k camera and providing outstanding support over the years.
Finally, we wish to acknowledge the following for their important
contributions;  Bill Gray for custom updates to his GUIDE software
regarding minor planet ephemerides, and Rae Stiening (Univ.of Mass.) for
providing preliminary and pre-release 2MASS data to our project, thus
expediting the UCAC2 release.

8) References
ESA 1997,
   "The Hipparcos and Tycho Catalogues", European Space Agency,
   publication SP-1200

Gauss,F.S. Zacharias,N. Rafferty,T.J. Germain,M.E. Holdenried,E.R.
Pohlman,J.W. and Zacharias,M.I., 1996,
   "A new astrometric survey of the Southern Hemisphere",
    Bull.AAS 28, No.4, p.1282

Hoeg,E., Fabricius,C., Makarov,V.V., Urban,S., Corbin,T., Wycoff,G.,
Bastian,U., 2000,
   "The Tycho-2 Catalogue of the 2.5 million brightest
   stars", Astron.& Astroph. 355L, 27 (short paper)

Hoeg,E., Fabricius,C., Makarov,V.V., Bastian,U., Schwekendieck,P.,
Wicenec,A., Urban,S., Corbin,T., Wycoff,G., 2000, 
   "Construction and verification of the Tycho-2 Catalogue", 
   Astron.& Astroph. 357, 367 (long paper)

Jones, B. F., Hanson, R. B., and Klemola, A. R., 2000,
   "Lick Northern Proper Motion Program: NPM2", AAS meeting 196
   abstract 53.02

Urban, S.E., Wycoff, G.L., Makarov, V.V., 2000,
   "Comparisons of Tycho-2 Catalogue Proper Motions with Hipparcos
   and ACT", AJ 120, 501 

Urban, S.E., Corbin, T.E., Wycoff, G.L., Makarov, V.V., Hoeg, E., 
Fabricius, C., 2000,
   "The AC 2000.2 Catalogue", AAS meeting 199, abstract #129.04

van Altena, W. F., Girard, T. M., Platais, I., Kozhurina-Platais, V.,
Ostheimer, J., Lopez, C. E., and Mendez, R. A., 1999,
   "The Yale/San Juan Southern Proper Motion Program", AAS DDA meeting
   #31, #10.04

Winter, L. 1999
   Ph.D thesis, University of Hamburg (in German)

Zacharias, N. 1997,
   "Astrometric Quality of the USNO CCD Astrograph (UCA)", AJ 113, 1925

Zacharias, N., Germain, M.E., Rafferty, T.J., 1997,
   "UCAC-S: a New High Precision, High Density Astrometric Catalog in
   the Southern Hemisphere", in: Proceedings "Hipparcos Venice 97",
   ESA publication SP-402, p.177

Zacharias, N. Zacharias, M.I., 1999,
   "Data Structure and Software of the UCAC-S Project", in: Proceedings
   of ADASS VIII, Eds. D.M.Mehringer,  R.L.Plante and D.A.Roberts,
   ASP Conf.series 172, p.345, San Francisco 

Zacharias, N., Urban, S.E., Zacharias, M.I., Hall, D.M., Wycoff, G.L.,
Rafferty, T.J., Germain, M.E., Holdenried, E.R., Pohlman, J.,
Gauss, F.S., Monet, D., Winter, L., 2000,
   "The first US Naval Observatory CCD Astrograph Catalog", AJ 120, 2131

Zacharias, N., Zacharias, M.I., Urban, S.E. and Hoeg, E., 2000,
   "Comparing Tycho-2 astrometry with UCAC1", AJ 120, 1148 

Zacharias, N., 2002 
   "Astrometric surveys in support of large telescopes", Proceed.
   SPIE 4836, 279, Eds. T.A. Tyson & S. Wolff

Zacharias, N. 2003,
   "The USNO CCD Astrograph Catalog (UCAC) Project and Beyond", in
   "Small Telescopes in the New Millennium II.  The Telescopes We Use.",
   ed. T Oswald, Kluwer Acad. Publ

Zacharias, N., Urban, S., Rafferty, T., Holdenried, E., and Winter, L.,
   "First Results from AGK2 Plate Remeasurements", AAS DDA meeting,
   in press.

Zacharias, N., Urban, S.E. et al.
   "The second US Naval Observatory CCD Astrograph Catalog",
   in preparation for AJ

Norbert Zacharias & Sean Urban for the UCAC team, USNO, Washington DC