Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.mrao.cam.ac.uk/projects/alma/fp6/2011-11-22-clemens.pdf Äàòà èçìåíåíèÿ: Thu Mar 29 19:28:51 2012 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 11:51:02 2012 Êîäèðîâêà:

The performance of `wvrgcal' The 4 quasar, band 9 experiment, `uid A002 X2d1a7d X35b.ms'
Marcel Clemens, Astrophysics Group, Cavendish Laboratory, University of Cambridge, UK 22/11/2011

Summary This summarises the results of reduction for the 4 quasar, band 9 dataset observed on 2011, October 6th . It is the second band 9 dataset that has been analysed. As the first dataset (uid A002 X2979fa X9f.ms) showed that the water vapour radiometer corrections resulted in better images after self-cal (using a 10 minute solution interval) the idea was to test whether this was also the case in a second dataset. The brightest source, 0522-364, (as in the previous band 9 data), was used as a calibrator for the other 3. Precipitable water vapour was about 0.2 mm during the observations (0.5 mm for the previous last band 9 dataset). There were 14 antennas in the array and antenna DV05 was used as reference antenna throughout. Calibration was also repeated using antenna DV06 as a reference but this made no difference to the final results. The dataset was reduced both with and without WVR corrections and antenna position corrections. Field 0 (0538-440) and field 3 (0522-364), are common to both of these band 9 datasets. Data reduction
The reduction procedure was the same as that used for the last band 9 dataset. For the runs that applied antenna position corrections the following corrections were used. These were generated by B. Nikolic. antenna = 'PM01,PM03,DV12,PM04,DV08,DV06,DV13,DV05,DV02,DV03,DV14,DV11,PM02' parameter = [ 0.00082882666997328319, -0.00060267909935868389, 0.00053660217544311843, 0.00021829902116718655, -0.00031293394965287104, 0.00016303949218002733, 0.00048911116408290999, -0.00024891769918947588, 0.00025254945715839184, 0.00028885784505478911, -0.00011997188732763665, 0.00012602699592589389, 0.00011965500208669278, 2.7977321027004791e-05, -1.4139373450895829e-05, 9.3794998082147939e-05, 3.4386240536386889e-05, -9.2710853897022398e-05, 0.00057657034621716062, -0.00064884140992287347, 0.00047926864351053439, 0.00085432526449425274, -0.00057456252482567254, 0.00053807888127350501, 0.00045084390974712418, 4.7271677481817054e-05, 9.7465606420679923e-05, 4.3625630647018361e-05, -2.3553962588019543e-05, 1.0034154013856995e-05, 0.00045556952261711773, -0.00026366808848604293, 0.00024155463177117471, 0.00084669291444469956, -0.00014656506870186026, 0.00032162048704642451, -0.00021497333150965418, -6.3926342597072169e-05, 6.2702965962469989e-06] The following channels were flagged due to very high amplitudes: spw = ['*:0 3', '*:61 63'] flagdata() spw = ['0:0 16', '0:55 56', '3:0 9', '3:59'] flagdata() Although there were differences between the XX and YY correlations I flagged each spw for both correlations. Although this actually throws away some good data, later gaincal requires both to be present, so the data would be effectively lost later anyway. Stokes I,Q,U and V, 512 â 512 pixel images were made with 0.1 arcsec pixels. Weighting was "Briggs" with `robust=0', and 100 iterations of clean were used with a 19 pixel square clean box centred on the centre of the image. For field 1 the source was not at the centre of this box.


Analysis `Imstat' was used to quantify the images. The background area was a large rectangle covering a large part of the image above the source. `Imfit' was used to fit Gaussians to the sources in the Stokes I images. Conclusions from this dataset The results discussed here are summarised in table 1. In the absence of antenna position corrections the WVR corrections made no significant improvement to the images either before or after self-cal. The addition of antenna position corrections worsened all the images slightly before self-cal. However, after self-cal the image of field 1 was improved markedly (S/N 195 282), while field 0 and 3 showed modest worsening (S/N 612 550 and S/N 804 676) - though field 3 was the phase reference source so is not really relevant (self-cal with 10 minute integration times worsens the image for this source of course). Field 2 was unaffected. The position of source 1 was slightly changed by the application of the antenna position corrections. The distances of the sources from field 3 were, 8.2 , 37.7 and 37.9 for field 0, 1 and 2 respectively, so there is no obvious reason for this difference in terms of distance from the calibrator. Source 2 was offset mostly in R.A. whereas source 1 was offset mostly in Dec. Fig. 2 shows the altitude of the sources during the observations. Field 1 was the lowest in the sky and showed the largest change, going from 62 to 42 (the phase reference source, field 3, moved from 76 to 61 in the same time). SO, the only case where a real improvement is seen is for field 1 (0423-013) with the antenna position and WVR corrections applied. Most of the improvement seems to come from the antenna position corrections (as in the previous band 9 dataset). Interestingly, the improvements are only seen after applying self-calibration. Before self calibration the images are actually worse than the no correction case. A natural question, which I have not investigated, is whether further benefits of any of the corrections are seen after further self-calibration runs. Self-cal here consisted of one phase-only and one amplitude and phase iteration. In general, a little flagging could also improve the images, and the possibility remains that more careful editing could reveal more subtle improvements.


Table 1: Stokes I image properties for various antenna position corrections and applications of wvrgcal for fields 0-3 in order. The first column refers to the reduction procedure: `Max' and `min' refer to the whole image, `rms' refers to a large background area not including the source, `omin' is the minimum within the background area. `a', `b' and `PA' refer to the Gaussian fit to the source by `imfit'. Missing values refer to when imfit failed to fit, and brackets imply the fit was poor. 1) no antenna position corrections, no WVR corrections, no self-cal. 2) same after self-cal. 3) no antenna position corrections, WVR corrections, no self-cal. 4) same after self-cal. 5) antenna position corrections and WVR corrections 6) same after self-cal. 7) no WVR corrections, just antenna position corrections, 8) same after self-cal. Self-cal worsens field 3 because this is the calibrator and the self-cal is only done with a 10 minute integration time. The original calibration is done over a scan time.
wvrgcal 0538-440 1 2 3 4 5 6 7 8 0423-013 1 2 3 4 5 6 7 8 0730-116 1 2 3 4 5 6 7 8 0522-364 1 2 3 4 5 6 7 max min rms omin max/rms a 0.3999 0.5013 0.4054 0.4998 0.3407 0.5001 0.3327 0.4984 0.1089 0.2834 0.1095 0.2808 0.1080 0.2768 0.1065 0.2740 0.04400 0.07736 0.04433 0.07640 0.02898 0.07500 0.02808 0.07513 0.9971 0.9818 0.9977 0.9829 0.9978 0.9683 0.9971 -0.04166 -0.00365 -0.04018 -0.003858 -0.04237 -0.009460 -0.04403 -0.008635 -0.04766 -0.01177 -0.05011 -0.01295 -0.04859 -0.006310 -0.04715 -0.007703 -0.01231 -0.003588 -0.01214 -0.003416 -0.01223 -0.003718 -0.01220 -0.003653 -0.004500 -0.005660 -0.005022 -0.005285 -0.005071 -0.006749 -0.004575 0.006799 0.0008336 0.006636 0.0008164 0.007847 0.0009080 0.007982 0.0009242 0.008218 0.001352 0.008224 0.001438 0.008643 0.0009803 0.008653 0.001044 0.002081 0.0007707 0.002059 0.0007500 0.002423 0.0007525 0.002437 0.0007392 0.0009393 0.001274 0.0008898 0.001223 0.0008899 0.001432 0.0009395 -0.02362 -0.003438 -0.02317 -0.003443 -0.03664 -0.003643 -0.03710 -0.00366 -0.04766 -0.01140 -0.05011 -0.01224 -0.04859 -0.005389 -0.04715 -0.006703 -0.008923 -0.003106 -0.008707 -0.003102 -0.01018 -0.002869 -0.01016 -0.003011 -0.003755 -0.004742 -0.003658 -0.004596 -0.003662 -0.005413 -0.003786 58.82 601.3 61.08 612.1 43.4 550.7 41.68 539.3 13.25 209.6 13.32 195.3 12.50 282.4 12.31 262.4 21.14 100.4 21.53 101.9 11.96 99.67 11.52 101.6 1061 770.4 1121 803.8 1121 676.2 1061 ± 0.0891 ± 0.0243 ± 0.0877 ± 0.0251 0.582 ± 0.036 0.582 ± 0.0351 (0.668 ± 0.229) 0.641 ± 0.0739 (0.689 ± 0.2378) 0.643 ± 0.0781 0.629 ± 0.0467 0.629 ± 0.0523 0.662 0.674 0.667 0.675 ± 0.1038 ± 0.0427 ± 0.1034 ± 0.0421 0.67 ± 0.0446 0.67 ± 0.0448 0.571 0.587 0.573 0.587 imfit Gaussian parameters b 0.525 ± 0.0963 0.487 ± 0.029 0.516 ± 0.0948 0.485 ± 0.03 0.479 ± 0.043 0.48 ± 0.0419 (0.491 ± 0.202) 0.569 ± 0.092 (0.488 ± 0.2101) 0.57 ± 0.097 0.537 ± 0.0584 0.542 ± 0.0653 0.455 ± 0.1512 0.42 ± 0.0723 0.455 ± 0.1506 0.422 ± 0.0709 0.42 ± 0.0754 0.422 ± 0.0757 0.506 ± 0.0189 0.465 ± 0.0224 0.504 ± 0.018 0.466 ± 0.0226 0.464 ± 0.0257 PA 111.1 111.5 109.9 111.3 111.8 111.9 (34.6) 117.4 (35.7) 117.7 115.0 115.6 92.3 99.2 91.8 99.6 99.1 98.8 108.3 108.3 108.0 108.2 108.1 -

0.581 ± 0.0165 0.589 ± 0.0177 0.58 ± 0.0157 0.588 ± 0.0179 0.587 ± 0.0203 -


uid___A002_X2d1a7d_X35b.ms
600Y(m) 650 700 750 800 850

PM04

DV08 DV02 DV05DA41 DV03 DV14 PM02 DV13 DV06 PM03 PM01

DV12

DV11
0 20 40 60 80 100

Figure 1: Antenna positions. DV05 was used as the reference antenna during calibration.

20

40

X (m)


Elevation vs. Time
80

field 2
75

field 3

Figure 2: Elevation as a function of time for the 4 quasars.

70

field 0
65

Elevation

60

55

50

45

field 1
40 08:53:20 09:10:00 09:26:40 09:43:20 10:00:00 10:16:40 10:33:20 10:50:00 11:06:40 11:23:20

Time (from 2011/10/06)


No WVR, no antenna position corrections, no self-cal

No WVR, no antenna position corrections, self-cal

WVR , no antenna corrections, no self-cal

WVR , no antenna corrections, self-cal

WVR, antenna position corrections, no self-cal

WVR, antenna position corrections, self-cal

No WVR, antenna position corrections, no self-cal

No WVR, antenna position corrections, self-cal

Figure 3: Stokes I images for field 1 for each row in table 1. The right hand images (even numbered rows in table) are the same as their counterparts to the left (odd numbered rows in table) except for the application of self-cal. The images have identical colour scales.