i. Tile Definitions
The sky was divided into 59,650 unique "tiles" for 2MASS operations (59,731 tiles appear in the scan database because 81 "gap-fill" tiles were defined to recover missed sky due to bad pointings). The right ascension width of a tile was 504´´, set by the intersection of the sky coverage of each of the three slightly offset arrays. Tiles were defined in 15 declination bands between the equator and each pole (0° - 6°, 6° - 12°, 12° - 18°, etc., and similarly in the Southern Hemisphere). The declination length of each tile was 6° plus one full frame (8.5´) to provide overlap between tiles in adjacent declination "bands". Tiles in the Northern celestial hemisphere had this declination overlap located on the southern end of the tile, while tiles in the Southern celestial hemisphere were observed to include overlap on the northern end of the tile. This arrangement resulted in a double overlap at the celestial equator. Right ascension tile overlaps were slightly larger than 10% of a frame width to account for possible pointing errors, precession, and the requirement that a source be a few arcseconds onto the chip to obtain a valid flux.
The cameras acquired 273 images of each survey tile in overlapping frames (see Sec IV.4.b Figure 1) with a declination step size between frames of 82.6´´ at the northern site and 274 frames with a step size of 82.3´´ at the southern site. The slight difference in step size results from the very small plate scale differences between the different cameras in each system and the need to obtain precise sub-pixel offsets in each of the six apparitions of a star. All arrays were rotated slightly from the scan direction, so that a star would cross one or two columns in its six apparitions, providing, once again, precise sub-pixel positioning of each star image. The 273/4 frames includes additional frames at the beginning and end of each tile needed to obtain a full 6 frames deep sky coverage within the actual tile boundary, plus 8.5´ of six-frame-deep overlap with adjacent tiles in the declination direction.
The measured rotation angle and platescales of each array are given below. North1 refers to the period before H-band array replacement in the northern camera in 1999 August:
|North1 and 2||1.997||2.006||1.984|
The rotations, platescales, and step sizes combined to produce optimal sub-pixel sampling for each wavelength.
ii. Tile Numbering
Tiles are described by their declination band (as defined by the equatorward edge of the tile ignoring tile overlaps, e.g. +0°, +6°, +12°,... and -0°, -6°, -12°, ...), and then by the RA of their western equatorward corner. Tiles observed by the northern observatory are numbered in the following fashion: starting with tile 0 which is located at RA=0h00m00s, DEC=+0°, tiles are numbered with increasing RA and then with increasing DEC, up to tile 29824 at RA=23h52m55".9, DEC=+84°. Tiles in the negative declination bands are numbered similarly, starting with tile 100000 at RA=0h00m00s, DEC=-0°, and ending with tile 129824 at RA=23h52m55.9s, DEC=-84°.
Tiles observed from the southern observatory are numbered in the same way as those observed from the north, but have a value of 200000 added to the tile number. This numbering scheme is designed to allow the same region of sky to be observed by both the northern and southern observatories as needed without conflict.
iii. Sky Coverage Boundaries
The northern survey began observations of the +12° declination band and initially worked northward toward the pole. The southern survey began observations of the -0° declination band and initially worked southward. The tiles between declination +0° and +12° were enabled for observation only toward the end of survey observations. The final declination boundary between coverage from the two observatories was determined by optimal dynamic scheduling of tiles in the last months of Survey observations, with the final result illustrated in Figure 1. Figure 2 is a "movie," showing the Survey's coverage week by week. At intervals coverage appears to "regress," when previously observed tiles were queued for re-observation, due to their inadequate quality.
|Figure 1||Figure 2|
iv. Time Requirements
The integration time for a frame included: two 51 ms resets (one occurred during the secondary flyback period, the second started as the secondary began scanning), one 51 ms "Read_1" (R1) integration, and one 1.3 s "Read_2" (R2) integration. An additional 5 ms of delay was added to allow for overhead and settling. The total dwell time on the sky for a single frame is thus 1.455 s.
A 273 frame survey scan required approximately 7 minutes (including overhead). Calibration observations consisted of six repeated scans of 42 consecutive frames requiring one minute each. Each set of six calibration tiles required approximately 8 minutes of observing time, including slewing and overhead.
[Last Update: 2003 March 5, by M. Skrutskie]