Use
Case: Offline: Reduce & Image Multi-Field Mosaic Data (with and
without single dish data)
Interferometric Use Case for ALMA (based on ALMA SW Memo
11, Science Requirements and Use Cases, modified to include offline
processing requirements).
Goal: Reduce and image data from a high-fidelity imaging
experiment of a multi-field mosaic. The field can be a survey of
small sources in the mosaic FOV (e.g. no single dish required) or a
mosaic of extended emission (e.g. ACA and single dish dish data are
required in post processing).
Contact Author: D. Shepherd
Role(s)/Actor(s):
Primary: The PI (follows the basic course for this UseCase)
Secondaries:
Archive (needed to download data into offline package or
disk/tape.)
Offline (to do the actual data reduction and imaging)
Priority:
Critical
Performance:
Response to user inputs in near real time. For a standard
multi-field mosaic (e.g. N pointings, Nyquist spacing, all fields with
similar uv coverage, mosaic boundaries extend at least one primary
beam beyond the emission region), processing time over that of a
signal field should increase no more than linearly with the number of
uv points to be processed.
Frequency of Use:
Perform this Offline Use Case for each multi-field mosaic
imaging experiment run by ALMA.
Preconditions:
- Observations have been made by ALMA.
- WVR corrections have been made on-line and stored in the Archive
file.
- Archive file includes all necessary calibration & source
observations (e.g. not just a partial scheduling block).
- Interferometric uv data has been downloaded from the Archive or is
available on tape or disk in ALMA FITS format.
- If combining single dish and interferometric data: calibrated,
single dish image is available on disk (downloaded from the
archive or reduced in a previous UseCase.
Basic
Course:
NOTE: All steps in the Basic Course should be able to be saved to
a master script. Alternatively, the user can start from a master
pipeline script downloaded from the Archive, and modified to run
this entire sequence or just parts of the sequence.
Interferometric Data Reduction (steps the same as UC.single.field):
- Fill the data from the Archive, tape, or disk into the Data
Reduction Package.
- Correct antenna gains as a function of elevation based
on look-up tables.
- Estimate atmospheric opacity based on weather information
and correct the data if needed.
- Edit data interactively and/or automatically:
A. Automatic editing requires built-in heuristics
which will identify bad data (e.g. pipeline-like).
B. Use Pointing, Tsys, Weather information in the
decision process to determine location of bad data.
- Using observations of a primary or secondary flux
calibrator, determine the flux calibration.
- Estimate the Parallactic angle correction based on a lookup
table and correct the data if doing polarization.
- Using observations of a gain calibrator observed between
source scans, compute phase & amplitude (including absolute
flux) calibration and apply to target scans by an
interpolation algorithm suitable for the atmospheric
conditions.
- Based on a plot of calibrator solutions displayed with
calibrator & source uv data, edit calibrator and source
data if a solution interval is bad.
- Using observations of a bandpass calibrator, compute the
bandpass for each correlator setup and apply to all data.
- Using observations of a polarization calibrator, determine
the instrumental polarization calibration ('leakage'
solutions) and apply to all the data.
- Edit the data by deleting antennas with poor bandpass or
polarization calibration solutions.
Self-calibration (if the source is strong enough and there is a point
source component to the peak structure):
- Weight selected source uv data.
Get psf and mosaic sensitivity images generated here and a
report of the theoretical RMS noise expected in the image.
- Grid the data, Fourier transform & deconvolve the image to
form a clean image.
- Derive a model from the image, that can be used to
self-calibrate the image.
- Repeat calibration (above steps) to obtain an incremental phase
correction to the gain solutions on short timescales. Apply
new calibration to the target data.
- Repeat imaging steps.
- Continue to iterate, finding incremental phase &/or amplitude
solutions until incremental improvements are in the noise.
Continuum subtraction (steps the same as UC.single.field):
- Identify spectral lines (manually or using header
information [or possibly a Lovas catalog?]), identify
multiple chunks of continuum channels and do a uv continuum
subtraction.
- Save the line-only and continuum-only data separately.
- Edit the source data again, manually or automatically.
Imaging (line and/or continuum) WITHOUT single dish data:
- Weight selected source uv data, individual fields should be
automatically weighted in an optimal way for a mosaic.
Get psf and mosaic sensitivity images generated here and a
report of the theoretical RMS noise expected in the image.
- Grid the data, Fourier transform & deconvolve the image to
form a clean image.
Alternate Course: Imaging (line and/or continuum) WITH single
dish data:
- Fourier transform the calibrated single dish image to
the uv plane.
- Weight the interferometer and single dish uv data properly.
Individual interferometric fields should be automatically
weighted in an optimal way for a mosaic.
Get psf and mosaic sensitivity images for the
interferometric mosaic field generated here and a report of
the theoretical RMS noise expected in the image (both with
and without the single dish data added).
NOTE: Image feathering is not considered because the ALMA
single dish will likely not provide good uv overlap and the
antenna primary beam FWHM will be greater than 10 times the
synthesized beam of ALMA.
- Grid the combined data, Fourier transform & deconvolve the
image using clean, multi-scale clean, or MEM to form a clean
image.
Analysis:
- Analyze the image cube for science results. (details are TBD).
If the mosaic has uniform noise across the final image
(e.g. flux density response decreases at the edges of the
mosaic), then one must be able to divide by the mosaic
'response' to get uniform response (needed to measure flux
densities of sources located outside of the mosaic overlap
region).
- Generate a publication-quality figure with annotation,
overlays, and comparison images if desired.
Store calibrated data:
- Write out calibrated target data and images to FITS files on disk
or tape.
Postconditions:
- Publication-quality figures are generated.
- Image statistics and analysis results are available to user.
- Calibrated data & images are stored for later use if needed.
- Calibration & imaging script recording session is available (if
desired by the user).
Issues
to be Determined or Resolved: None at this
time.
Notes: This Use Case was created by D. Shepherd &
S. Myers. It was created to help test the Offline software. Relavent
SSR Use Cases from SSR Memo 11 are: 4.6.2 (Retrieve Archived Data);
4.7.1 (Reduce Multi Field [in pipeline mode]).
Last modified: 01jul03