0000339: Archive data have wrong coordinate system

Viewing Issue Simple Details [ Jump to Notes ]

[ Issue History ] [ Print ]

Category

Severity

Reproducibility

Date Submitted

Last Update

0000339

[TMS] software

minor

always

25-03-11 10:53

21-07-11 10:05

Reporter

Gyula Jozsa

View Status

public

Assigned To

Arno Schoenmakers

Priority

low

Resolution

open

Status

assigned

Summary

0000339: Archive data have wrong coordinate system

Description

Some archive data from 2002 apparently show an error in the coordinates of the pointing/phase centre. The data and the header are offset by one precession correction from B1950 to J2000 of vice versa, which does not turn up in the header. Email conversation with archive user attached.

Additional Information

Hi Arno and Hans,

there is a guy who requested some archive data (see below) and claims that there is an error in the data (headers). At the moment I don't have any reason to believe that he's wrong (Briggs had a look at this, too), so I must have made some mistake and/or this is a known issue that has occurred over a short period. I would like to ask for your help here, and possibly tell me if you know at once what has gone wrong.

He claims that the actual phase (pointing) centre of the archive data is at:

09:26:51.539 +36:44:35.235 (J2000)

but the header contains

09:23:46:508 +36:57:37.468 (J2000)

while

09:23:46:508 +36:57:37.468 (B1950) = 09:26:51.539 +36:44:35.235 (J2000)

would be correct for the header. In other words, the data seem to have been precessed from B1950 to J2000, while the header values for the phase (pointing) centre have not been changed, but the epoch has been changed from B1950 to J2000. There might be other possible combinations (see below), like the archive data were already precessed from B1950 to J2000, were then precessed again, but the header values were precessed only once.

I did:

getms seq_nr
j2convert in=seq_nr.MS type=J2000
glish -l MSsplit.g seq_nr.MS (splitting mosaics, if present)
ms2uvfits in=seq_nr.MS out=seq_nr.UVFITS multisource=T writesyscal=T combinespw=T

where seq_nr = 10202817 10202818 10202820 10202821 10202822 10202823 10202846 10202847 10202848 10202849 10202850 10202851 10202884 10202885 10202886 10202988 10202989 10202990 10203015 10203016 10203017 10203125 10203126 10203127 10203205 10203206 10203207 10203293 10203294 10203295 10203442 10203443 10203445 10203452 10203453 10203454 10203464 10203465 10203466

-------------------------------------------

The second part of the story is that before he made the request, he has used the old data sets from the time of the observations. For those data, the phase (pointing) centre was at:

09:23:46:508 +36:57:37.468 (J2000)

while the header contained

09:26:51.539 +36:44:35.235 (J2000)

Note that

09:23:46:508 +36:57:37.468 (B1950) = 09:26:51.539 +36:44:35.235 (J2000)

In the original specs (which I cannot access), if I read the email right, the centre was specified as:

09:20:40.698 +37:10:31.07 (B1950) = 09:23:46:508 +36:57:37.468 (J2000)

such that the data seem to have been precessed once from B1950 to J2000 and the header values twice.

Many thanks for your help here!

Cheers,
Josh

--------------------------------

Original email:

Dear Gyula,

Thank you for making the WSRT data available. However, I have discovered a serious problem with the dataset. It appears to result from inappropriate conversion of B1950->J2000, and we think it must result in the archiving or conversion to UVFITS pipeline, since the data show that the expected continuum sources are there... but they are in the wrong places.

The reason why I requested these data is to check the astrometry of CNOC2 field (my research target). The accuracy of astrometric information in data is very important to carry out our research. I requested the same data as those that our colleague, Martin Zwaan, has processed earlier and has given to us. When we checked the coordinates for radio sources in his maps, we found the information of pointing center in header was quite different from the CNOC2 field. It is seemed that the precession conversion(B1950 --> J2000 ) has been applied twice to RA&DEC. (The central RA&DEC in header is 09:26:51.539 +36:44:35.235 (J2000)) When we double-checked coordinates for our sources against FIRST catalogue, the image looked like the same with one having the central RA&DEC 09:23:46:508 +36:57:37.468 (J2000), The relationship between two coordinates is exactly same as precession conversion(B1950-->J2000). We also found a slight rotation on the image, which is consistent
with an inappropriate precession.

So, we wanted to check the data that comes directly from WSRT archive. Unfortunately, we found both datasets have the same problem in header. Furthermore, when I looked for the data using WSRT Archive database query, it showed up the central RA&DEC like 09:23:46:508 +36:57:37.468 (J2000). But the actual UV data is consistent with having been pointed at 09:26:51.539 +36:44:35.235 (J2000).

I think that something might have corrupted the data during the process to archive data or something else. Do you have any idea about this problem?

My thesis advisor, Frank Briggs, thinks this is a very significant
problem that should receive high-priority attention, and he
suggested that I also bring it to the attention of Dr. Vermeulen,
so he is cc'd on this email.

Many thanks for you help with this.

Regards,
Jonghwan

Attached Files

Relationships

Notes
(0000340) Arno Schoenmakers 28-03-11 13:36	Hans is looking into this.

(0000353) Harm Munk 21-07-11 10:05	Ik kan hier niet verder mee, alle voorbereidende werk is gedaan, maar ik mot nu het Python script uitvoeren op het archief maar ik heb daar geen permissies. Het is eigenlijk vanaf nu een punt voor Arno.

Issue History
Date Modified	Username	Field	Change
25-03-11 10:53	Gyula Jozsa	New Issue
28-03-11 12:48	Harm Munk	Project	submission => TMS
28-03-11 13:36	Arno Schoenmakers	Note Added: 0000340
28-03-11 13:36	Arno Schoenmakers	Assigned To	=> Hans van Someren Greve
28-03-11 13:36	Arno Schoenmakers	Status	new => assigned
21-07-11 10:05	Harm Munk	Note Added: 0000353
21-07-11 10:05	Harm Munk	Assigned To	Hans van Someren Greve => Arno Schoenmakers