Step 3 Overview & Scripts#
In this step, we self-calibrate the continuum and obtain the final continuum measurement set.
Self calibration is like shimmying up a chimney.
Scripts for Step 3 - Self-calibration of the continuum:
step3_continuum_selfcal.py # main script
dictionary_data.py # loads data_dict
selfcal_utils.py # selfcal functions
Self-Calibration of Short-Baseline Execution Blocks: Self-calibration started with the (phase-aligned and concatenated) short-baseline continuum data. We performed a series of phase-only self-calibration iterations, decreasing the solution interval in succession: inf, 243s, 120s, 60s, 30s, 18s. We avoided combining by SPW in the first two rounds to remove any potential per-SPW phase offsets. Between iterations, we updated the model by imaging the data interactively and conservatively with Hogbom deconvolution and a Briggs robust parameter of 0.5, setting the gain parameter to 0.05 to guard against cleaning too deeply in a given major cycle. We used a circular clean mask of \(3''\) radius, centered on the new common phase center.
We estimated the noise in the CLEAN image within a centered annulus with inner and outer radius \(6''\) and \(10''\), respectively. We concluded the phase-only self-calibration sequence when the peak SNR increased by a factor less than \(10\%\) and the image quality did not visually improve. We then carried out one round of amplitude and phase self-calibration with the solution interval equal to a scan length (486s), which improved the peak SNR by \(5\%\). Overall, the entire self-calibration process resulted in an increase in the peak SNR of the short-baseline dataset by a factor of \(2.2\); the resulting noise was \(35\, \mu\)Jy/beam (\(1.35\) mK) for a \(630 \times 930\) mas beam. We concatenated the (newly self-calibrated) short-baseline dataset with the (not self-calibrated, but phase aligned) long-baseline dataset from from Step 2 and proceeded to self-calibrate the combined dataset.
Self-Calibration of SB+LB Data: An initial round of per-SPW phase self-calibration was attempted but yielded too many flagged solutions due to low SNR. We started with a solution interval of 900s and stepped down to 360s, 180s and 60s before performing one round of amplitude and phase self-calibration (again on a scan length). Overall, this second self-calibration process resulted in peak SNR increase by a factor of \(1.6\). The estimated noise was \(17.4 \, \mu\)Jy/beam (4.2 mK) for a \(260\times 360\) mas beam.
Create Final Continuum Measurement Set: The final continuum MS is time averaged to 30 seconds to save space.