[daip] [!11290]: AIPS - Simulating a UV database in AIPS with full stokes (e.g., RR, LL, RL, LR)

[Jae-Young Kim]

Jae-Young Kim updated #11290
----------------------------

Simulating a UV database in AIPS with full stokes (e.g., RR, LL, RL, LR)
------------------------------------------------------------------------

           Ticket ID: 11290
                 URL: https://help.nrao.edu/staff/index.php?/Tickets/Ticket/View/11290
                Name: Jae-Young Kim
       Email address: jykim at astro.snu.ac.kr
             Creator: User
          Department: AIPS Data Reduction
       Staff (Owner): -- Unassigned --
                Type: Issue
              Status: Open
            Priority: Default
                 SLA: NRAO E2E
      Template group: Default
             Created: 22 November 2017 02:50 AM
             Updated: 22 November 2017 02:50 AM
           Reply due: 25 November 2017 01:00 AM (2d 22h 10m)
      Resolution due: 18 August 2020 01:00 AM (999d 22h 10m)

Dear the NRAO helpdesk staffs,I am a PhD student working in the MPIfR and is partially involved in the EHT collaboration, in particular for supporting tests of new imaging algorithms by generating and providing synthetic test datasets. I am wondering if there is a simple task or a routine in AIPS which could help simulate a full Stokes UV dataset for a given full Stokes source model in one shot (e.g., delta-function Clean components with four Stokes values, a set of Gaussian models or four Stokes images -> full-Stokes data like those from real observations). I am also wondering if AIPS UVCON could be improved in the future to deal with the full Stokes situations in the same context. Below I describe details of my experiences and failures which motivated me to consult this issue with AIPS experts.So far, AIPS UVCON has been my favorite task to generate synthetic data and the task has been doing excellent job for most of situations (arbitrary array, arbitrary total intensity model etc). However, I also wanted to simulate a full Stokes dataset with fake D-terms in order to verify some polarization measurement capabilities. I found that this is not trivial, because the UVFITS files generated by UVCON has only RR polarization. I made some trials by only touching the header, but it looks like the overall structure of the UVFITS file is basically the same (only one complex visibility in one timestamp for a particular baseline).The polarization simulation was still possible at basic level assuming an ideal polarization leakage calibration, as the four Stokes data could have been generated separately and the I, Q, U and V maps could be combined in later stage. But this is apparently too much simplistic.One day, I found AIPS DTSIM is suitable for generating a database with full Stokes. To my knowledge, the main purpose of this task is to generate an arbitrary UVFITS file and test its compatibility with various AIPS tasks (e.g. FRING a test dataset with mixed polarization such as XL or YR). Anyhow the structure of the file was quite satisfactory. But we found some time ago that the task is not versatile enough to deal with various input models which UVCON can easily process.So far, the most satisfactory way of doing this was to take a *real* dataset with four Stokes parameters from the target array, null the visibility values, fill in the input models, and then finally apply the inverse of the calibration (e.g., by TBOUT AN table / Editing D-terms manually / TBIN / LPCAL). However, this is not possible for an arbitrary array with arbitrary observing time/scan.The last option is to simply use ParselTongue to hack the visibility data directly. I have some ideas but I could also imagine that this would be quite tricky job and it also could raise file compatibility issues with regard to other interferometric data processing softwares.These are the main reasons why I have been looking for a simple and powerful task like AIPS UVCON to generate a UV database but just doing it just four times and glue the four visibility files together. It would be still very satisfactory if this kind of job can be done by taking several steps in AIPS. Your answer will definitely help me solve the aforementioned technical issues and improve the simulations.Best regards,Jae-Young Kim

------------------------------------------------------
Staff CP:  https://help.nrao.edu/staff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listmgr.nrao.edu/pipermail/daip/attachments/20171122/bb10bf1c/attachment.html>