The MeerKLASS On-the-Fly continuum survey: pipeline design and validation

TL;DR

This paper presents the MeerKLASS On-the-Fly (M-OTF) continuum survey and its end-to-end pipeline, designed to obtain wide-area, high-resolution radio continuum images from MeerKAT visibilities recorded during constant-elevation scans. It identifies and mitigates the dominant smearing introduced by MeerKAT's fixed-delay correlation through time-dependent phase rotation, direction-dependent PSF modeling, and wide-band faceted deconvolution with DDFacet, enabling 2 s snapshot imaging and visibility-plane mosaicking. The authors validate the approach with UHF and L-band pilot data, achieving 2 s snapshot resolutions of ~$23$–$25$ arcsec with rms sensitivities near $33$–$35$ μJy beam$^{-1}$, and demonstrate deep mosaics over hundreds of square degrees; they also quantify astrometric and photometric accuracy and outline DR1 data products. With a full 10,000 deg$^2$ coverage planned across 544–1088 MHz, and a delay-tracking fix to reach ~$14$ arcsec resolution and ~${ m rms} ightarrow 25$ μJy beam$^{-1}$, M-OTF promises a rich, legacy dataset for galaxy evolution, large-scale structure, cluster science, rotation measures, and transient searches, as a crucial precursor for SKA-Mid operations.

Abstract

The MeerKAT Large Area Synoptic Survey (MeerKLASS) is designed to map large areas of the Southern sky for cosmology using the single-dish HI intensity mapping (IM) technique, while simultaneously delivering a wide, high angular-resolution interferometric survey. We present the design and first results of the MeerKLASS On-the-Fly (OTF) continuum data, which exploits the visibilities recorded during fast, constant-elevation scans. This observing mode enables fast commensal imaging over several hundred of square degrees on a nightly basis. We describe the OTF survey strategy and pipeline, focusing on handling challenges introduced by the current MeerKAT fixed-delay correlation observing mode, which causes decorrelation (smearing). We implement a correction scheme based on time-dependent phase rotation, direction-dependent PSF modeling, and wide-band faceted deconvolution with \texttt{DDFacet}. Using UHF-band and pilot L-band data, we demonstrate the recovery of high-quality 2-second snapshot images and deep mosaics over hundreds of square degrees. After smearing correction we are able to achieve a resolution of $23.3$arcsec and $14$ arcsec with an rms sensitivity of $35 μ{\rm Jy\,beam}^{-1}$ and $ 33 μ{\rm Jy\,beam}^{-1}$ in the UHF and L-band respectively. The full survey will cover $10,000 \, {\rm deg}^{2}$ at 544-1088 MHz, and after the delay tracking fix implemented we expect to reach $\sim 25 μ{\rm Jy\,beam}^{-1}$ at $14$ arcsec resolution. The continuum OTF data products will support diverse science goals, including galaxy and AGN evolution, diffuse cluster emission, large-scale structure and cosmology, rotation-measure synthesis, and transient searches. MeerKLASS-OTF thus establishes an efficient path to wide-area commensal surveys with MeerKAT and provides a key technical precursor for SKA-Mid.

Figures (27)

• Figure 1: The approximated expected footprint for the full MeerKLASS 2023-2028 campaign (dashed dark red lines). The green lines mark the region already observed. The Galaxy at 800 MHz is shown for reference, together with available galaxy survey footprint from DESI (orange) and KiDS (blue). The location of the L-band 2021 observations is also shown (green lines).
• Figure 2: A summary of recent large area low-frequency surveys in the Southern hemisphere including their sensitivity, frequency, and resolution (see \ref{['tab:survey_SH']}). The size of the markers is proportional to the square root of the survey resolution. The horizontal lines show the frequency coverage for surveys. The green circles show the expected sensitivity and resolution of MeerKLASS OTF surveys without smearing incorporated.
• Figure 3: An example of MeerKLASS constant elevation observing strategy where "rising" and "setting" epochs are shown in blue and red respectively. L and UHF-band primary beam FWHM at the nominal frequencies are shown using circles. Each epoch lasts approximately 2 hours. The dashed horizontal line show the track for the correlator delay centre.
• Figure 4: Top panel shows the density of the 2-second pointing when a rising and a setting scans are combined for an observation box. The middle panels shows the effective exposure time convolved with the primary beam assuming every 2-second snapshot contributes to the imaging. The bottom panel show the expected sensitivity in UHF-band, a optimistic scenario, where all the 2-second snapshots are used to construct image from the visibilities.
• Figure 5: This plot shows the effect of smearing in a snapshot observation of the M-OTF. The results for simulated and analytical smearing computation are shown in left and the middle panels. The right panel shows the ratio of the analytical and the simulated prediction for the snapshot.