[daip] AIPS task SAD

Wed Mar 19 11:12:17 EDT 2014

Kunal wrote:
> Hi Eric!
> I need help with SAD. I'm trying to find point sources in a EVLA S-band 
> FLATN mosaic image. There are at least 3 sources in the image with 
> signal-to-noise ratio greater than 100, but SAD finds no sources. I've 
> included the output from SAD below.
> 
> I'm using default parameters of SAD except for:
> taskSAD.ngauss   = 40000
> taskSAD.doall    = 1
> taskSAD.dparm(1)  = 5
> taskSAD.dparm(7)  = 1
> taskSAD.dparm(9)  = 2
> 
> I have used these input parameters to successfully sources in many other 
> images, but for this particular image, I'm having trouble.
> 
> Thanks in advance!
> - Kunal.
> 
> P.S.
> 1) The image is located at 
> /lustre/kmooley/projects/JVLACOSMOS/concat_R1/fits/fits/temp/M_X1_Y4.fits.
> 
> 2) Output from SAD:
> 
> SAD  1: Task SAD    (release of 31DEC13) begins
> SAD  1: Create ABC         .RESID .   1 (MA)  on disk  1  cno    3
> SAD  1: Fit to histogram gives RMS = 5.031E-06
> SAD  1: Setting Search limit at 3.0 x RMS of 0.0000
> SAD  1: Looking for up to40000 sources in subfield:
> SAD  1: BLC:       1.      1.      1.      1.      1.      1. 1.
> SAD  1: TRC:    4096.   4096.      1.      1.      1.      1. 1.
> SAD  1: Using Clean beam   1.00001   1.00001    0.00 from image header
> SAD  1: Searching down to   0.000015 (JY/BEAM )  pass 1
> SAD  1: Found  12536 Islands in pass 1
> SAD  1: No convergence for island 570 Box   3168.   726.  3188. 767.
> SAD  1: No convergence for island6440 Box   1542.  2779.  1661. 2905.
> SAD  1: No convergence for island2412 Box     93.  1696.   144. 1727.
> SAD  1: No convergence for island3847 Box   1226.  2136.  1302. 2205.
> SAD  1: No convergence for island**** Box   3016.  4085.  3027. 4096.
> SAD  1: No convergence for island7594 Box    593.  3141.   610. 3165.
> SAD  1: No convergence for island4493 Box    702.  2270.   828. 2396.
> SAD  1: No convergence for island1462 Box   1569.  1509.  1684. 1635.
> SAD  1: No convergence for island3296 Box   1400.  1977.  1473. 2087.
> SAD  1: No convergence for island6779 Box   1391.  2840.  1419. 2963.
> SAD  1: No convergence for island8814 Box     73.  3631.   199. 3757.
> SAD  1: No convergence for island5293 Box   1340.  2452.  1358. 2484.
> SAD  1: No convergence for island6737 Box   1240.  2829.  1274. 2912.
> SAD  1: No convergence for island5362 Box   1492.  2481.  1511. 2528.
> SAD  1: No convergence for island5453 Box   1122.  2507.  1172. 2566.
> SAD  1: No convergence for island4887 Box   1766.  2344.  1831. 2406.
> SAD  1: No convergence for island4000 Box    672.  2169.   793. 2213.
> SAD  1: No convergence for island5307 Box   1281.  2458.  1292. 2487.
> SAD  1: No convergence for island5608 Box   2173.  2546.  2299. 2671.
> SAD  1: No convergence for island1851 Box    917.  1582.   979. 1661.
> SAD  1: No convergence for island4833 Box    605.  2333.   697. 2393.
> SAD  1: No convergence for island9370 Box   1421.  3662.  1488. 3788.
> SAD  1: No convergence for island5477 Box   1534.  2514.  1547. 2537.
> SAD  1: No convergence for island2205 Box   1351.  1648.  1368. 1670.
> SAD  1: No convergence for island3472 Box   1581.  2040.  1626. 2101.
> SAD  1: Sucessfully solved12071 islands. 465 OTHERS FAILED TO CONVERGE
> SAD  1:   2706 islands broken into multiple sources
> SAD  1:   2706 of these were rejected due to at least partial failure
> SAD  1:     23 apparently complex islands failed to break into 2
> SAD  1:  13267 total components had their fits rejected
> SAD  1:  13267 components rejected for minimum peak
> SAD  1:    849 components rejected for minimum flux
> SAD  1:    376 components rejected for maximum size
> SAD  1:   7091 components rejected for maximum interior X excursion
> SAD  1:   7168 components rejected for maximum interior Y excursion
> SAD  1:     11 components rejected for maximum exterior X excursion
> SAD  1:     26 components rejected for maximum exterior Y excursion
> SAD  1: Resulting number of sources fit is     0
>  unknown   SAD  (31DEC13)      2     18-MAR-2014  14:34:05 Page    1
> ABC         .SUBIM .      1     Disk  1     Plane    1 User       2
> 
> Window BLC    1    1   1   1  1  1  1 TRC 4096 4096   1   1  1  1 1
> Sources found down to  0.000015 in JY/BEAM
> Retry level  0.000015 (JY/BEAM ) plus gain 0.100
> Using RMS image .
> Reject components peak <  5.00000 in JY/BEAM
> Reject components flux < 0.00000
> Reject components outside window >   0.0 cells
> Reject components outside image >    0.0 cells
> Reject residual flux >   1.00000 with gain 0.100
> Fluxes expressed in units of JY/BEAM
> NOTE: Fluxes marked by * have been divided by 1000.
> Errors determined by theory from RMS   5.03 microJy
> Reference Center: 10 00 28.2001  02 12 36.000
> All source widths, width errors, and both coordinate errors are in arc s
> NO corrections for bandwidth smearing have been made
> Source peaks and fluxes NOT corrected for primary beam

What is CPARM?  This sounds like a case to use CPARM - making large 
numbers of islands (when you do not need them) and including way down 
into the noise makes the fitter think that it has failed.  Try first 
fitting down to 30 sigma or so, then 15 sigma, 5 sigma for 3 rounds of 
searches.

Eric Greisen