Overlap checks on astrometry and photometry

This page details some checks of the consistency of the absolute calibration of photometric and astrometric quantities by looking at duplicate measurements of the same astronomical sources that occur in the overlap regions between survey frames.

SQL queries for overlap data

The following SQL was used to extract overlap data from the archive (e.g. see the online SQL Cookbook for explanation):

SELECT s1.jMag, s1.b,
       (s1.ra-s2.ra)*3600.0*COS(RADIANS(s2.dec)) as deltaRA,
       (s1.dec-s2.dec)*3600.0 as deltaDec,
       (s1.jmh-s2.jmh) as djmh,
       (s1.hmk_1-s2.hmk_1) as dhmk
FROM   gcsSource AS s1, gcsSource AS s2, gcsSourceNeighbours AS x
WHERE  s1.jMag>0.0 AND s1.jClass=-1 AND s2.jMag>0.0 AND s2.jClass=-1 AND
       s1.hMag>0.0 AND s1.hClass=-1 AND s2.hMag>0.0 AND s2.hClass=-1 AND
       s1.k_1Mag>0.0 AND s1.k_1Class=-1 AND s2.k_1Mag>0.0 AND s2.k_1Class=-1 AND
       masterObjID=s1.sourceID AND slaveObjID=s2.sourceID AND distanceMins<0.0166667
       AND s1.jmh>-0.9e9 AND s1.hmk_1>-0.9e9 AND s2.jmh>-0.9e9 AND s2.hmk_1>-0.9e9
       AND s1.sourceID < s2.sourceID
       AND distanceMins IN (
           SELECT MIN(distanceMins)
           FROM gcsSourceNeighbours
           WHERE masterObjID=x.masterObjID
       )

SELECT s1.jMag, s1.b,
       (s1.ra-s2.ra)*3600.0*COS(RADIANS(s2.dec)) as deltaRA,
       (s1.dec-s2.dec)*3600.0 as deltaDec,
       (s1.jmh-s2.jmh) as djmh,
       (s1.hmk_1-s2.hmk_1) as dhmk
FROM   gpsSource AS s1, gpsSource AS s2, gpsSourceNeighbours AS x
WHERE  s1.jMag>0.0 AND s1.jClass=-1 AND s2.jMag>0.0 AND s2.jClass=-1 AND
       s1.hMag>0.0 AND s1.hClass=-1 AND s2.hMag>0.0 AND s2.hClass=-1 AND
       s1.k_1Mag>0.0 AND s1.k_1Class=-1 AND s2.k_1Mag>0.0 AND s2.k_1Class=-1 AND
       masterObjID=s1.sourceID AND slaveObjID=s2.sourceID AND distanceMins<0.0166667
       AND s1.sourceID < s2.sourceID
       AND (s1.jObjID % 4) = 0
       AND distanceMins IN (
           SELECT MIN(distanceMins)
           FROM gpsSourceNeighbours
           WHERE masterObjID=x.masterObjID
       )

SELECT s1.j_1Mag, s1.b,
       (s1.ra-s2.ra)*3600.0*COS(RADIANS(s2.dec)) as deltaRA,
       (s1.dec-s2.dec)*3600.0 as deltaDec,
       (s1.j_1mh-s2.j_1mh) as djmh,
       (s1.hmk-s2.hmk) as dhmk
FROM   lasSource AS s1, lasSource AS s2, lasSourceNeighbours AS x
WHERE  s1.j_1Mag>0.0 AND s1.j_1Class=-1 AND s2.j_1Mag>0.0 AND s2.j_1Class=-1 AND
       s1.hMag>0.0 AND s1.hClass=-1 AND s2.hMag>0.0 AND s2.hClass=-1 AND
       s1.kMag>0.0 AND s1.kClass=-1 AND s2.kMag>0.0 AND s2.kClass=-1 AND
       masterObjID=s1.sourceID AND slaveObjID=s2.sourceID AND distanceMins<0.0166667
       AND s1.sourceID < s2.sourceID
       AND distanceMins IN (
           SELECT MIN(distanceMins)
           FROM lasSourceNeighbours
           WHERE masterObjID=x.masterObjID
       )

- so I selected neighbouring sources out to 1 arcsec with the same set of passband detections as overlapping sources, only selecting one way (i.e. not double counting source 1 differenced with source 2 and source 2 differenced with source 1) and taking only the nearest match in each case.I've also used data from the LAS, GPS and GCS.

Photometry results:

Summary histograms of the consistency of RA/Dec, (J-H) and (H-K) in overlap regions of LAS, GPS and GCS frames follow:

The data are further broken down as a function of magnitude and Galactic latitude for the purposes of examining the RMS (robustly computed as 1.48x median of absolute deviation to guard against non-Gaussian outliers affecting the results); note that in the following plots the sigmas have been divided by 1.414 to indicate the accuracy of a single measure as determined from the difference of two measures:

Notes:

• In each plot, red and black are RA/Dec or (J-H)/(H-K) respectively;
• There is no degradation of the absolute photometric/astrometric calibrations at low Galactic latitude - in fact, the astrometric calibration appears to be slightly better at the lowest latitudes;
• Colour precision is good to 4% for optimally exposed images (J=12 to 14ish);
• degradation in photometric accuracy is seen at brighter magnitudes (saturation effects?) and at fainter magnitudes (in the sky-noise limited regime).

-- NigelHambly - 09 Aug 2005