© The Authors 2025

1 Introduction

The Galactic interstellar medium (ISM) is magnetized and contains multiple phases on various physical scales. Both magnetic fields and turbulence play crucial roles in star formation (Federrath & Klessen 2012), and cosmic ray acceleration and propagation (Lazarian & Yan 2014). The magnetic field and turbulence are coupled to each other, and a full description requires magnetohydrodynamics, which is very challenging from both observational and theoretical perspectives (Cho et al. 2003; Burkhart 2021). Radio polarization observations provide measurements of the magnetic field with both ordered and random components, and therefore help us to understand the turbulent magnetized ISM in the Galaxy.

The linearly polarized emission originates from synchrotron radiation of relativistic electrons spiraling around the magnetic field. When propagating through the magneto-ionic medium, the polarized wave experiences Faraday rotation, which changes the polarization angle (χ) as χ = χ₀ + RM λ². Here, χ₀ is the intrinsic polarization angle, λ is the wavelength, and RM is the rotation measure defined as $\text{RM}=0.81{{{\displaystyle \int }}^{\text{}}}_{\text{source }}^{\text{observer }}{n}_e{B}_{\Vert }\text{d}l$ where n^e is the thermal electron density in cm⁻³ B_|| is the magnetic field along the line of sight in µG, l is the path length in pc, and the RM is in rad m⁻².

The amount of Faraday rotation differs along the line of sight and across the observation beam. This results in a complicated polarization morphology that was first observed several decades ago (e.g. Wieringa et al. 1993). Images of polarized emission are thus imprinted with properties of the magnetic field, high-energy electrons, and thermal electrons. Understanding these ISM ingredients requires polarization observations, which motivates polarization surveys of a large sky area.

Previous low-frequency polarization surveys include the northern sky survey with the Dwingeloo 25-m telescope at frequencies of 408, 465, 610, 820, and 1411 MHz (Brouw & Spoelstra 1976; Spoelstra 1984), the northern sky survey with the Dominion Radio Astrophysical Observatory (DRAO) 26-m telescope at 1.4 GHz (Wolleben et al. 2006), and the southern sky survey with the Villa Elisa 30-m telescope at 1.4 GHz (Testori et al. 2008). The latter two have been combined to form an all-sky polarization image of the Galaxy at 1.4 GHz (Reich 2007). All of these surveys used narrow band receivers with a single frequency channel.

Polarization surveys have been revived over the last decade because of the development of broadband multichannel polarization backends. In addition to the great improvement of sensitivity, the RM synthesis method (Brentjens & de Bruyn 2005) can be readily employed to reconstruct the polarized intensity as a function of the Faraday depth, called the Faraday depth spectrum (Sun et al. 2015b) or the Faraday dispersion function. Here, the Faraday depth is defined as $$\phi (\mathit{r})=0.81{{{\displaystyle \int }}^{\text{}}}_r^{\text{observer }}{n}_e{B}_{\parallel }\text{d}l,$$(1)

where r is the position vector in the source, and the other quantities are the same as in the definition of RM. The Faraday depth spectrum provides information on the dominant Faraday rotation components along the line of sight and the complexity introduced by mixed emission and rotation in the magnetized ISM.

The Global Magneto-Ionic Medium Survey (GMIMS; Wolleben et al. 2009) is the first all-sky polarization observation campaign aiming to produce polarization images at multifrequency channels using telescopes in both hemispheres. It consists of northern and southern surveys covering the approximate frequency range 300–1800 MHz. Two of the surveys have been completed with data released: the high-band northern survey with the DRAO 26-m telescope over the frequency range 1280–1750 MHz split into 236.8-kHz channels (Wolleben et al. 2021), and the low-band southern survey with Murriyang, the Parkes 64-m telescope over the frequency range 300–480 MHz split into 0.5-MHz channels (Wolleben et al. 2019). Based on these GMIMS surveys, individual Galactic features have been studied, such as the north polar spur (Sun et al. 2015a), the Fan region (Hill et al. 2017), the ISM toward the HII region Sh 2– 27 (Thomson et al. 2019), and Loop II (Thomson et al. 2021). The Galactic magnetic field has also been studied using moment maps calculated from the Faraday depth cubes (Dickey et al. 2022; Raycheva et al. 2025).

At higher frequencies, the S-band Polarization All-Sky Survey (S-PASS; Carretti et al. 2019) observed the southern sky with Murriyang over the frequency range 2.2–2.4 GHz with a channel width of 0.5 MHz. Commensal with S-PASS, the Southern Twenty-centimeter All-sky Polarization Survey (STAPS; Haverkorn 2015) covering the frequency range 1.3–1.8 GHz with channel width of 1 MHz has also been conducted as part of GMIMS, as its high-band southern component survey.

This paper presents data processing and images of the total intensity and linear polarization from STAPS. The layout of the paper is as follows: the observations and data processing are briefly described in Sect. 2; the processing is verified in Sect. 3; the Stokes parameters I, Q, and U and the RM maps are presented in Sect. 4; and a summary of the results is presented in Sect. 5.

2 Observations and data processing

The STAPS survey was conducted with Murriyang commensally with the S-PASS survey. During the observations, the S-band (λ13 cm) receiver was mounted at the primary focus position, while the L-band (λ20 cm) receiver was placed off focus. With the same coordinate system as in Ruze (1965), in which the focus points horizontally along the z axis and the x axis is directed vertically upward, the lateral displacement of the L-band receiver is x = 630 mm and z = 7.6 mm. Consequently, the L-band receiver pointed about 1°.2 above the S-band receiver. Data were recorded simultaneously for both surveys. Some of the main parameters of STAPS are listed in Table 1. Detailed descriptions of the observation setup and data processing for S-PASS were presented by Carretti et al. (2019).

The observations are composed of long azimuth (AZ) scans at the fixed elevation (EL) of 33° for the S-band receiver corresponding to the south celestial pole at Parkes Observatory. All of the observations were conducted at night to avoid the influence of the Sun. Each night, back and forth east scans (61°.5 ≤ AZ ≤ 180°) and west scans (180° ≤ AZ ≤ 294°) were carried out at a speed of 15° min⁻¹. For STAPS, the fixed EL was about 1°..2 above the south celestial pole, which delivers a sky coverage of −89° < Dec < 0°, leaving a gap at the pole. The starting sidereal time for each individual scan was predetermined so that the maximum spacing between two neighboring scans was about 4′.4, slightly less than half of the beam width in the S band, to guarantee Nyquist sampling. This means that the STAPS survey is oversampled.

There were eight runs in total, conducted in April, July, and October 2008, January, April, July, and October 2009, and January 2010; roughly once every 3 months. In total, there are about 7100 east scans and about 3800 west scans after removing the bad scans. The east scans fully sample the sky. The spacing of the west scans is twice that of the east scans, providing sufficient intersections with east scans to ensure effective destriping. In Fig. 1, some west and east scans in April and October 2008 are illustrated. As can be seen, the combination of east and west scans produces a large number of crossings (Carretti et al. 2019, their Fig. 3), which is required to recover zero levels of intensity in polarization.

The observation scheme of long AZ scans was developed and demonstrated by Carretti et al. (2019) to be able to measure the absolute level of Stokes Q and U, and thus the polarized intensity. The long scan results in a large range for the parallactic angle, and therefore a modulation of Q and U in the telescope frame. The subtraction of a constant baseline to account for ground emission therefore does not remove the Q and U of the sky signal. A joint fitting of all crossings between the west and east scans yields the absolute level of Q and U. However, this cannot be applied to the total intensity, I, since there is no modulation of I against the parallactic angle. Therefore, the zero level is unconstrained in total intensity.

Data processing followed the procedure developed by Carretti et al. (2019, their Fig. 6) for the S-PASS survey. All scans were first calibrated and then combined for map-making. PKS B1934-638 served as the primary flux calibrator and the flux model by Reynolds (1994) was used. PKS B0407-658 served as the secondary flux calibrator. The sources were also used for on-axis leakage calibration, on the assumption that they are unpolarized. We note that the two sources show a very small fractional linear polarization of 0.15% and 0.01%, respectively, from the recent observations by Taylor & Legodi (2024), which will not impact the calibration of polarization leakage. We used PKS 0043-424 and 3C 13 8 to correct the polarization angle. For the former, the intrinsic polarization angle is 137° with an RM of 2 rad m⁻² (Broten et al. 1988), and for the latter, the polarization angle is a constant value of 169° (Perley & Butler 2013). Each night, both the flux and the polarization calibrators were scanned in the right ascension (RA) and declination (Dec) directions. The gain, on-axis polarization leakage, and instrumental polarization angle were derived, and these calibration solutions were applied to the long AZ scans.

For map-making, the bad scans that were influenced by strong radio frequency interference (RFI), and that thus show large values of root mean square (rms) in total intensity or large values in Stokes V, were flagged. Parts of the scans close to the planets were also flagged. Zero-level corrected maps of coarse resolution for the east and west scans were derived by solving intensity levels with crossings (Carretti et al. 2019, Appendix A), and a basket-weaving technique by Emerson & Gräve (1988) was implemented to remove residual discontinuities. The baselines of each calibrated scan were then adjusted using the zero-level corrected map of coarse resolution, and all the adjusted scans were subsequently combined to form the final high-resolution maps. The maps cover a Dec range from −89° to 0°.

The calibration and map-making procedures were performed for each individual frequency channel of 1-MHz in width. Of the total of 512 frequency channels, we obtained high-quality maps of I, Q, and U for 301 channels. Because of low gains, 60 channels at the two ends of the band were flagged. The channels in the frequency range from about 1.52 GHz to about 1.61 GHz were dominated by RFI. For these frequency channels, the data left after flagging were insufficient to make maps.

Fig. 1 West and east scans conducted in April and October runs in 2008. The background image displays the total intensity at 1.4 GHz from STAPS. The images are in Galactic coordinates with Aitoff projection.

3 Verification

3.1 The beam pattern

Because of the displacement of the receiver, the beam was slightly distorted with a coma lobe. We collected all calibrated scans close to the unpolarized bright source PKS B1934-638 and constructed the beams of I, Q, and U for each individual frequency channel in the AZ-EL frame. The beams at 1.325 GHz, 1.523 GHz, and 1.767 GHz are shown in Fig. 2. Here, all the beam patterns are normalized by the peaks in the total intensity beams.

For total intensity, the peak of the coma lobe decreases slightly with increasing frequency, ranging from about 9% at the low frequency to about 6% at the high frequency. The measured levels of the coma lobe are consistent with electromagnetic simulations of the performance of the telescope with an offset feed (Granet 2007). We note that we obtained the gains based on the fitting of the peaks of the main beams during calibration. For compact sources, the peak intensity is expected to be accurate, as is shown later, but the estimate of integrated flux density will be influenced by the coma lobe, and thus needs a proper deconvolution. For diffuse emission, the intensity level is less influenced by the coma lobe because the emission is smooth. We focused on the large-scale emission from the Galaxy, and therefore did not perform deconvolution for the current data.

The Q and U patterns in Fig. 2 represent the residual leakage from I. The centers of the beam patterns are shifted toward the coma lobe. The on-axis residual instrumental polarization is less than about 0.8% for all frequencies after leakage correction. In contrast with the total intensity, the off-axis residual polarization caused by the coma lobe becomes stronger at higher frequencies, running from about 2% at 1.325 GHz to about 4% at 1.767 GHz.

The off-axis polarization is mostly canceled out after integration over the beam (Carretti et al. 2019). However, the polarization beam is asymmetric, and instrumental polarization can be generated near strong total intensity emission, such as that of the Galactic plane. We did not correct the off-axis polarization in the current data.

The beam is noncircular in the AZ-EL frame (Fig. 2). When the telescope is moving in the AZ direction for an east scan, the beam is not aligned north-south in the equatorial frame. Instead, it is inclined to that frame at an angle, βE , which can be determined from the parallactic angle, and which can be as large as π/4. When the telescope is moving in the AZ direction for a west scan, the inclination of the beam, βW, in the equatorial frame is very different and sometimes even orthogonal to βE . When the data from the east and west scans are combined in the map-making process in the equatorial system, the beam will vary across the sky because of the changing difference between βE and βW . Near the south pole, east and west scans are close to parallel, so the beam in the map will be roughly aligned north-south, but far from the south pole the form of the beam in the maps becomes harder to predict.

We have approached this problem empirically. We constructed the total intensity beam patterns from PKS 1934-638, which are similar to those shown in Fig. 2 but in the equatorial frame. We then obtained the half power beam width by fitting 2D Gaussian functions to these total intensity beam patterns. The result is shown in Fig. 3. The beam in RA is always larger than the beam in Dec. The beam width versus frequency can be fit to a power law, as is indicated in Fig. 3. We also examined the beam patterns constructed from other sources at different Decs and found that the beam is surprisingly uniform. The measured beam widths in RA show a greater scatter; the scatter in Dec is about 2% and in RA 10%. We have adopted the beam dimensions shown by the fit lines in Fig. 3 and smoothed the maps to a resolution of 20‘ at all frequencies. The variation in the beam versus Dec introduces an error of less than 10% for the flux density scale. Smoothing, as was described above, made only small changes in the appearance of the maps, indicating that most of the diffuse structure has scales larger than the beam.

Fig. 2 Beam patterns for I, Q, and U at three frequencies constructed from calibrated scans. The Q and U patterns represent leakage from I. All the patterns are normalized to the total intensity peaks. The contour levels for I start at −3 dB and stop at −15 dB with an interval of 3 dB. For Q and U, the contour levels start at −0.015 and stop at 0.015, with an interval of 0.005.

Fig. 3 Beam width along RA and Dec at different frequencies derived from Gaussian fits to the total intensity beam patterns constructed in the equatorial coordinate.

3.2 The flux density scale

The observation and data processing procedures were optimized for large-scale emission, but it is still worth examining the flux densities of compact sources. We performed source finding with AEGEAN (Hancock et al. 2018) on the STAPS total intensity image at 1.41 GHz, and cross-matched the results with the Parkes Southern Radio Source Catalog (PKSCAT; Wright & Otrupcek 1990). The comparison of the flux densities is shown in Fig. 4. We note that the peak flux densities from STAPS were used here to avoid the influence of coma lobes.

As can be seen in Fig. 4, the flux densities of STAPS and PKSCAT agree reasonably well. A linear fitting yields a flux density ratio of 0.92, close to 1. This means an uncertainty of less than about 10% for the flux densities of compact sources from STAPS.

To examine the flux density scales of both total intensity and polarized intensity for large-scale structures, we turned to the radio galaxy Cen A, which is an extended and polarized source. We obtained images of Cen A observed with Parkes by O’Sullivan et al. (2013), and smoothed them to a resolution of 20′. We made pixel intensity versus pixel intensity plots at three frequencies for I, Q, and U, and the polarized intensity (Fig. A.1) toward the southern lobe of Cen A (O’Sullivan et al. 2013, their Fig. 1).

As can be seen in Fig. A.1, there is good agreement between STAPS and the Parkes observations for total intensity and polarization. The linear fits show that the pixel intensity ratio is around 1 with an uncertainty of about 10%. The comparison also confirms that the coma lobes have little influence on large-scale emission. We note that there are offsets in the fits, which are due to the different zero levels. We conclude that the flux density scale of the STAPS data is accurate within 10%.

Fig. 4 Comparison of flux densities between STAPS and PKSCAT. The dashed line marks y = x and the solid line is a linear best fit with the parameters showing in the figure.

3.3 The main beam brightness temperature scale

We do not have the aperture efficiency and main beam efficiency measured for the L-band receiver at the off-focus position, and thus could not establish a temperature standard solely from STAPS. Fortunately, the GMIMS high-band north survey, covering virtually the same frequency range and overlapping with STAPS for a large area with a Dec range between −30° and 0°, was absolutely calibrated to the main beam brightness temperature (Du et al. 2016; Wolleben et al. 2021). This allowed us to derive the temperature scale by comparing STAPS with the GMIMS high-band north.

We selected the Galactic plane area 0° ≲ l ≲ 30° and |b| ≲ 5°, where the total intensity emission is bright for both GMIMS and STAPS. We smoothed the STAPS data to the GMIMS resolution of 40′, made pixel intensity versus pixel intensity plots for the total intensity, I, at each individual frequency, and performed linear fits to obtain the conversion factors from the flux density in Jy beam⁻¹ to the main beam brightness temperature in K. The results for randomly selected frequencies between 1.3 and 1.8 GHz are shown in Fig. B.1. The data points are very tight around a linear relation, which verified the map-making procedure of STAPS. The conversion factors at 20′ resolution were then derived; they are in the range of 0.346–0.186 K Jy⁻¹ for the frequency range of 1.324–1.770 GHz (Table 1).

We applied the conversion factors to the STAPS data to obtain the temperature scale. To verify the temperature scale, we retrieved the Continuum HI Parkes All-Sky Survey (CHIPASS; Calabretta et al. 2014) image at 1.4 GHz and smoothed it to the STAPS resolution of 20′. We note that CHIPASS is in the full beam brightness temperature scale, tied to the 1420 MHz allsky survey (Reich 1982; Reich & Reich 1986; Reich et al. 2001). A factor of 1.55 (Reich & Reich 1988) was applied for conversion to the main beam brightness temperature. We then made the temperature-temperature plot for the sky area −89° < Dec < 0° (Fig. 5). There is also a tight linear relation. The linear fitting yields a ratio of about 1.02, meaning that the temperature scale of STAPS derived on the basis of GMIMS is reliable.

Fig. 5 Temperature–temperature plots between STAPS and CHIPASS at 1.4 GHz.

4 Maps

4.1 Total intensity and polarization maps

We obtained I, Q, and U maps at 301 frequency channels in both Jy beam⁻¹ and K. All maps have been smoothed to a common resolution of 20′. As an example, the maps at 1.45 GHz are shown in Fig. 6. The total intensity is very smooth with strong emission at the Galactic plane. In contrast, there are many structures on various angular scales in the Q and U images. Because of depolarization, there is virtually no polarized emission expected near the Galactic plane. Polarized emission near some parts of the plane is probably caused by residual instrumental leakage.

We selected areas at Dec of around −20° where there are no emission structures and calculated the rms noise level in I, Q, and U for each individual 1-MHz channel. The results are shown in Fig. 7. For I, the rms is about 16 mK at the low frequency end and about 8 mK at the high frequency end. For Q and U, the rms decreases from about 8 mK to about 5 mK with increasing frequency.

The number of crossings varies with the Dec with more data points within a beam toward the lower Dec, as can be seen in Fig. 1. This means that the rms noise depends on Dec and reaches a minimum at Dec of −89°.

However, we estimated the rms noise in I from different Decs and found no large variations. According to Condon (1974) and Meyers et al. (2017), the confusion limit is about 65.7 mJy beam⁻¹ at 1.324 GHz and about 53.6 mJy beam⁻¹ at 1.770 GHz at a resolution of 20′, corresponding to about 23 mK and 10 mK, respectively. These values are slightly higher than what we measured, indicating that the total intensity probably has already hit the confusion limit, and thus the rms noise is independent of Dec.

For Q and U, the emission structures from the Galactic ISM are distributed almost everywhere, which makes it difficult to examine the dependence of rms noise on Dec. We obtained the difference images of two Q or U images with consecutive frequencies. In this way, the Galactic emission that is common to both images has been removed, and only the noise remains. We then obtained the rms noise of Q and U against the Dec shown in Fig. 8. We also obtained the number of crossing points (N) within a beam from map-making. The theoretical rms is expected to be proportional to $1/\sqrt{N}$, and is also plotted in Fig. 8. It can be clearly seen that the measured rms noise in Q and U is consistent with the theoretical expectation.

Fig. 6 I, Q, and U maps at 1.45 GHz from STAPS. The resolution is 20°.

4.2 RM synthesis maps

The observed polarization as a function of the frequency, v, or wavelength squared, λ², can be written in a complex quantity as P(λ²) = Q(λ²) + iU(λ²). Polarized emission from a source can be described with a Faraday depth spectrum as F(φ) = Q(φ) + iU(φ), where φ is the Faraday depth and |F(φ)| gives the polarized intensity residing at φ. These two quantities, P(λ²) and F(φ), are Fourier transform pairs, and RM synthesis, developed by Burn (1966); Brentjens & de Bruyn (2005), can be used to derive F(φ) from P(λ²). Because of the limited sampling in the λ² domain, the observed F(φ) is a convolution of the true F(φ) with a window function. The deconvolution can be performed with the RM clean method developed by Heald (2009).

We applied RM synthesis (Brentjens & de Bruyn 2005) and RM clean (Heald 2009) to obtain the Faraday spectra. We then searched for the peak of |F(φ)| for each individual pixel. The polarized intensity and Faraday depth of peaks are shown in Fig. 9. The Faraday depth of the peak can be regarded as the RM if the Faraday spectrum is simple with only one peak. This is the case for most of the pixels (but see Raycheva et al. (2025) for a discussion on complex Faraday spectra). From the image of peak polarized intensity, we can clearly see a boundary at the Galactic latitude |b| ≈ 30°, below which extended, bright polarized emission at higher absolute latitudes is abruptly depolarized. We note that there is still polarized emission toward the Galactic plane, which is caused by instrumental leakage.

Schnitzeler et al. (2019) identified polarized sources from SPASS and observed these sources with the Australian Telescope Compact Array at 1.1–3.1 GHz to determine their RMs with the QU-fitting method (e.g. Sun et al. 2015b). We extracted Faraday spectra toward the sources observed by Schnitzeler et al. (2019), and derived the RMs for sources with a peak polarized intensity that is larger than the local background by three times the local rms. Here, we used only Faraday simple spectra that have a single dominant peak.

The comparison of the RMs is shown in Fig. 10. Based on the results by Schnitzeler et al. (2019), many sources have more than one RM component. In this case, we used RMs of the first components that have the highest fractional polarization; except for one source with multiple observations, the RMs of the second component were closer to the RM from STAPS and were used. As we can see, the RMs generally agree with each other. This implies that the RM from both the single component of simple sources and the first component of complex sources is mainly contributed by the Galactic ISM other than that from within the sources. For the source with multiple observations, the RM varies between different observations, and the variation is much larger than the errors (Fig. 10), indicating a variable nature of the source in the time domain that needs further investigation.

Fig. 7 Root mean square (rms) noise levels per channel for I, Q, and U measured from an area without strong emission at Dec of around −20°.

Fig. 8 Root mean square (rms) noise per channel in Q and U versus Dec. The line is derived from the number of crossing points within a beam.

Fig. 9 Peak polarized intensity and Faraday depth derived from Faraday spectra.

5 Conclusions

STAPS is the first broadband multichannel polarization survey of the southern sky (−89° < Dec < 0°) in the L band covering the frequency range of 1.3–1.8 GHz. STAPS was observed commen- sally with the S-PASS survey. The unique feature for S-PASS and STAPS is that the surveys are composed of long AZ scans, enabling us to achieve zero-level calibration for Q and U. The data processing procedure developed by Carretti et al. (2019) for S-PASS was used to process STAPS. We obtained high-quality I, Q, and U maps for 301 frequency channels, each 1 MHz in width, smoothed to a common resolution of 20′.

The consistency of the flux densities of compact sources between STAPS and PKSCAT, and the good correspondence of the pixel values for the total intensity and polarized intensity of Cen A, with a ratio close to 1 between STAPS and the Parkes observations by O’Sullivan et al. (2013), indicate that the flux density scale of STAPS is reliable and that the influence of the coma lobes on both total intensity and polarized intensity of extended emission is small.

We derived the conversion factors from the flux density (Jy beam⁻¹ ) to the main beam brightness temperature (K) using the GMIMS high-band north survey and applied these factors to the STAPS data. The temperature–temperature plot between STAPS and CHIPASS exhibits a tight linear relation with a gradient close to 1, indicating a reliable temperature scale and map-making procedure for STAPS.

We also ran RM synthesis and a RM clean to obtain Faraday depth cubes. From the peak Faraday depth or RM map, we obtained RMs of extragalactic sources whose RMs were derived by Schnitzeler et al. (2019) from ATCA observations at 1.1–3.1 GHz. The agreement of these RMs indicates that the polarization calibration and map-making for STAPS are also reliable.

The STAPS cubes, together with those of the GMIMS high- band north survey, provide a multifrequency polarization view of the entire Milky Way Galaxy, which will help us understand the Galactic magnetic field and magnetized ISM.

Fig. 10 Comparison of RMs derived from STAPS and the measurements by Schnitzeler et al. (2019). The sources with only one RM component in Schnitzeler et al. (2019) are marked with red circles. For sources with multiple RM components, the RMs of the first components are plotted and marked with black squares. For the one source with multiple observations, the RMs of the second components are plotted and marked with blue diamonds. The line indicates y = x.

Data availability

The image cubes for Stokes I, Q, and U in Equatorial coordinates with a common resolution of 20′ are available at the CDS via anonymous ftp to cdsarc.cds.unistra.fr (130.79.128.5) or via https://cdsarc.cds.unistra.fr/viz-bin/cat/J/A+A/694/A169. These data can also be accessed from the Canadian Advanced Network for Astronomical Research (doi: 25.0004) via https://www.doi.org/10.11570/25.0004.

Acknowledgements

We thank the referee for the comments that have improved the paper. We thank Drs. Paddy Leahy, Ann Mao, and Anna Ordog, and Prof. John Dickey for valuable comments on the paper. XS is supported by the National Natural Science Foundation of China (No. 12433006) and the National SKA Program of China (Grant No. 2022SKA0120101). This work is part of the joint NWO-CAS research program in the field of radio astronomy with project number 629.001.022, which is (partly) financed by the Dutch Research Council (NWO). MH acknowledges funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement No 772663). Murriyang, the Parkes 64m radio-telescope is part of the Australia Telescope National Facility which is funded by the Australian Government for operation as a National Facility managed by CSIRO. We acknowledge the Wiradjuri people as the Traditional Owners of the Observatory site.

Appendix A Flux density comparison

Fig. A.1 Comparison of pixel densities of Cen A between STAPS and the Parkes observation by O’Sullivan et al. (2013) at three frequencies (from top to bottom). From left to right are: I, Q, U, and polarized intensity. The solid lines indicate linear fits.

Appendix B Brightness temperature versus flux density

Fig. B.1 GMIMS pixel total intensity in K versus STAPS pixel total intensity in Jy beam−1 for randomly selected frequencies between 1.3 and 1.8 GHz. The linear fits are also shown.

References

All Tables

All Figures

	Fig. 1 West and east scans conducted in April and October runs in 2008. The background image displays the total intensity at 1.4 GHz from STAPS. The images are in Galactic coordinates with Aitoff projection.
In the text

	Fig. 2 Beam patterns for I, Q, and U at three frequencies constructed from calibrated scans. The Q and U patterns represent leakage from I. All the patterns are normalized to the total intensity peaks. The contour levels for I start at −3 dB and stop at −15 dB with an interval of 3 dB. For Q and U, the contour levels start at −0.015 and stop at 0.015, with an interval of 0.005.
In the text

	Fig. 3 Beam width along RA and Dec at different frequencies derived from Gaussian fits to the total intensity beam patterns constructed in the equatorial coordinate.
In the text

	Fig. 4 Comparison of flux densities between STAPS and PKSCAT. The dashed line marks y = x and the solid line is a linear best fit with the parameters showing in the figure.
In the text

	Fig. 5 Temperature–temperature plots between STAPS and CHIPASS at 1.4 GHz.
In the text

	Fig. 6 I, Q, and U maps at 1.45 GHz from STAPS. The resolution is 20°.
In the text

	Fig. 7 Root mean square (rms) noise levels per channel for I, Q, and U measured from an area without strong emission at Dec of around −20°.
In the text

	Fig. 8 Root mean square (rms) noise per channel in Q and U versus Dec. The line is derived from the number of crossing points within a beam.
In the text

	Fig. 9 Peak polarized intensity and Faraday depth derived from Faraday spectra.
In the text

Fig. 10 Comparison of RMs derived from STAPS and the measurements by Schnitzeler et al. (2019). The sources with only one RM component in Schnitzeler et al. (2019) are marked with red circles. For sources with multiple RM components, the RMs of the first components are plotted and marked with black squares. For the one source with multiple observations, the RMs of the second components are plotted and marked with blue diamonds. The line indicates y = x.

In the text

	Fig. A.1 Comparison of pixel densities of Cen A between STAPS and the Parkes observation by O’Sullivan et al. (2013) at three frequencies (from top to bottom). From left to right are: I, Q, U, and polarized intensity. The solid lines indicate linear fits.
In the text

	Fig. B.1 GMIMS pixel total intensity in K versus STAPS pixel total intensity in Jy beam−1 for randomly selected frequencies between 1.3 and 1.8 GHz. The linear fits are also shown.
In the text