WALLABY Early Science - II. The NGC 7232 galaxy group
K. Lee-Waddell, B.S. Koribalski, T. Westmeier, A. Elagali, B.-Q. For, D. Kleiner, J.P. Madrid, A. Popping, T.N. Reynolds, J. Rhee, P. Serra, L. Shao, L. Staveley-Smith, J. Wang, M.T. Whiting, O.I. Wong, J.R. Allison, S. Bhandari, J.D. Collier, G. Heald, J. Marvil, S.M. Ord
MMNRAS , 1–16 (2019) Preprint 3 January 2019 Compiled using MNRAS L A TEX style file v3.0
WALLABY Early Science - II. The NGC 7232 galaxygroup
K. Lee-Waddell (cid:63) , B.S. Koribalski , T. Westmeier , A. Elagali , , , B.-Q. For , ,D. Kleiner , , J.P. Madrid , A. Popping , , T.N. Reynolds , , , J. Rhee , , , P. Serra ,L. Shao , , L. Staveley-Smith , , J. Wang , M.T. Whiting , O.I. Wong , ,J.R. Allison , , S. Bhandari , J.D. Collier , , , G. Heald , , J. Marvil , S.M. Ord CSIRO Astronomy and Space Science, Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia International Centre for Radio Astronomy Research, The University of Western Australia, 35 Stirling Hwy, Crawley WA 6009, Australia ARC Centre of Excellence for All Sky Astrophysics in 3 Dimensions (ASTRO 3D) INAF - Osservatorio Astronomico di Cagliari, Via della Scienza 5, I-09047 Selargius (CA), Italy ARC Centre of Excellence for All-sky Astrophysics (CAASTRO) Kavli Institute for Astronomy and Astrophysics, Peking University, Beijing 100871, China Research School of Astronomy and Astrophysics, Australian National University, Canberra ACT 2611, Australia Sub-Dept. of Astrophysics, Department of Physics, University of Oxford, Denys Wilkinson Building, Keble Rd., Oxford, OX1 3RH, UK School of Computing, Engineering and Mathematics, Western Sydney University, Locked Bay 1797, Penrith NSW 2751, Australia Inter-University Institute for Data Intensive Astronomy (IDIA), University of Cape Town, Private Bag X3, Rondebosch, Cape Town 7701, South Africa CSIRO Astronomy and Space Sciences, PO Box 1130, Bentley WA 6102, Australia
Accepted 2018 December 22. Received 2018 December 20; in original form 2018 August 02
ABSTRACT
We report on neutral hydrogen (H i ) observations of the NGC 7232 group with theAustralian Square Kilometre Array Pathfinder (ASKAP). These observations wereconducted as part of the Wide-field ASKAP L-Band Legacy All-sky Blind surveY(WALLABY) Early Science program with an array of 12 ASKAP antennas equippedwith Phased Array Feeds, which were used to form 36 beams to map a field of view of30 square degrees. Analyzing a subregion of the central beams, we detect 17 H i sources.Eleven of these detections are identified as galaxies and have stellar counterparts, ofwhich five are newly resolved H i galaxy sources. The other six detections appear tobe tidal debris in the form of H i clouds that are associated with the central triplet,NGC 7232/3, comprising the spiral galaxies NGC 7232, NGC 7232B and NGC 7233.One of these H i clouds has a mass of M HI ∼ × M (cid:12) and could be the progenitorof a long-lived tidal dwarf galaxy. The remaining H i clouds are likely transient tidalknots that are possibly part of a diffuse tidal bridge between NGC 7232/3 and anothergroup member, the lenticular galaxy IC 5181. Key words: galaxies: groups: individual: NGC 7232 – galaxies: interactions
Most galaxies in the current epoch of the Universe residein galaxy groups (Tully 1987). These groups are ubiquitousand their evolution is often dominated by gravitationallydriven interactions between their members (Barnes & Web-ster 2001, Yang et al. 2007). Neutral atomic hydrogen (H i )gas can be used to trace disturbed morphologies that are theresult of galaxy interactions and detect newly formed tidal (cid:63) E-mail: [email protected] (KLW) features (Mullan et al. 2013, Lee-Waddell et al. 2016). Thepresence of tidal bridges and tails can constrain the initialproperties of the interacting galaxies and the dynamics ofthe actual encounter (Toomre & Toomre 1972, Bournaud &Duc 2006), aiding in our understanding of this evolutionaryprocess. Within the tidal streams, high-density clumps of H i can accrete sufficient amounts of material to eventually be-come self-gravitating tidal dwarf galaxies (TDGs; Mirabel,Dottori, Lutz 1992, Lelli et al. 2015). TDGs can providefurther information about the interaction event and possi- © a r X i v : . [ a s t r o - ph . GA ] J a n K. Lee-Waddell et al. bly strengthen the standard model of cosmology (see Duc etal. 2014, Ploeckinger et al. 2018).A single galaxy group offers a mere snapshot of thistype of evolutionary process. In order to throughly exam-ine galaxy interactions and fully quantify the by-productsof these events, large-scale surveys are required. Previoussingle-dish all sky surveys, such as the H i Parkes All-SkySurvey (HIPASS; Barnes et al. 2001) and Arecibo LegacyFast ALFA (ALFALFA; Giovanelli et al. 2005), provided theopportunity to conduct a census of the H i located in groupenvironments but often higher angular resolution follow-upobservations were required to resolve and properly charac-terize low-mass tidal features (Ryder et al. 2001, English etal. 2010, Janowiecki et al. 2015, Lee-Waddell et al. 2016).The Australian Square Kilometre Array Pathfinder(ASKAP) is a new radio interferometer that will enablewide-field observations that can also resolve low-mass H i clouds (Johnston et al. 2007, Johnston et al. 2008). ASKAPconsists of 36 × i -rich galaxies. WALLABY will cover 75 percent of thesky (declination range of − ◦ < δ < + ◦ ) and is predictedto detect H i in more than 500,000 galaxies out to a red-shift of 0.26 with an angular resolution of 30 arcsec and aspectral resolution of 4 km s − (Duffy et al. 2012, Koribalski2012). This level of resolution is required to resolve the phys-ical characteristics and measure the dynamical properties oftidally formed features in the nearby universe (Mullan et al.2013, Lelli et al. 2015).ASKAP Early Science is an observing program aimedat producing scientifically useful data, with at least 12 MkIIPAF-equipped ASKAP antennas (i.e. ASKAP-12), whilecommissioning ASKAP to its full specification. The prior-ities for this program are to demonstrate the unique capa-bilities of ASKAP, produce data sets to facilitate the de-velopment of data processing and analysis techniques, andachieve high scientific impact. ASKAP survey science teamswere given opportunities to select specific science fields totarget with ASKAP-12 during the shared-risk and develop-ment phases of all aspects of ASKAP. These observationswould not only test the capabilities of the array but also theautomated data processing pipeline ASKAPsoft as wellas the CSIRO ASKAP Science Data Archive (CASDA) forstoring the final data products.The ASKAP Early Science program started in October2016 with the WALLABY team taking 36-beam observa- https://data.csiro.au/dap/search?q=ASKAPsoft https://data.csiro.au/dap/public/casda/casdaSearch.zul tions using ASKAP-12. The first WALLABY Early Sciencefield, the NGC 7232 field, was chosen because it contains 19detections from HIPASS (Koribalski et al. 2004, Meyer et al.2004) – including a nearby spiral galaxy (IC 5201), galaxypairs, and other interacting systems – which would assistin data validation. These varying environments also providethe opportunity to capitalize on ASKAP’s high resolutioncapabilities, allowing for more detailed spatial studies andthe potential for new H i detections. Although there are sev-eral H i -rich galaxy systems in this field (e.g. Reynolds et al.2018), this paper will specifically focus on the NGC 7232/3triplet and its neighbouring galaxies.The NGC 7232/3 triplet comprises three spiral galax-ies, NGC 7232, NGC 7232B and NGC 7233, located in aloose group environment (which is listed as LGG 455 byGarcia et al. 1993 and includes a neighbouring lenticulargalaxy, IC 5181) at a Hubble distance of ∼
24 Mpc (Garcia1995). Previous interferometric observations with the Aus-tralia Telescope Compact Array (ATCA) show H i streamsconnecting the galaxies within the triplet (Barnes & Web-ster 2001), indicating a recent/on-going interaction event.HIPASS observations show that the triplet system as well asan unresolved starless H i feature, HIPASS J2214-45, residein a common gas-rich envelope referred to as the AM2212-460 galaxy group (Koribalski et al. 2004). There are a hand-ful of other neighbouring galaxies, bringing the membershipnumber to ∼
10 galaxies (Barnes & Webster 2001).Here we are using ASKAP’s ability to provide a largefield of view while maintaining high angular resolution tocarry out detailed analysis of the NGC 7232/3 triplet sys-tem and its possible connection to other group members. Al-though ASKAP-12 has a limited number of baselines (com-pared to the full array), longer integration times can be usedto increase the H i sensitivity in order to detect dwarf com-panions and faint tidal features. Since the angular resolutionof ASKAP is comparable to the ATCA observations, the lat-ter provide a good reference for data verification purposes.The higher frequency resolution of ASKAP will enable fur-ther dynamical analysis of spectrally distinct components ofthe galaxy triplet and its tidal streams.In Section 2 of this paper, we describe the observationalsetup used during various commissioning and Early Sciencephases of the ASKAP-12 array. Section 3 details the dataprocessing and imaging completed with ASKAPsoft . InSection 4, we present our measurements, final H i maps of theregion and stellar properties from ancillary optical imaging.We analyse and discuss the newly resolved H i sources inSection 5 and Section 6 contains concluding remarks. The majority of the observations of the NGC 7232 field weretaken during the inaugural month of the ASKAP Early Sci-ence program. Additional observations were obtained in Au-gust and September during different commissioning phasesof the array in 2016 and 2017. The same configuration ofantennas was used throughout 2016 when only 48 MHz of si-multaneous bandwidth was achievable. The bandwidth wassplit into 2592 independent channels, each 18.5 kHz wide(equivalent to 3.9 km s − at 1420 MHz). After numerous sys-tem upgrades and further commissioning, more bandwidth MNRAS000
10 galaxies (Barnes & Webster 2001).Here we are using ASKAP’s ability to provide a largefield of view while maintaining high angular resolution tocarry out detailed analysis of the NGC 7232/3 triplet sys-tem and its possible connection to other group members. Al-though ASKAP-12 has a limited number of baselines (com-pared to the full array), longer integration times can be usedto increase the H i sensitivity in order to detect dwarf com-panions and faint tidal features. Since the angular resolutionof ASKAP is comparable to the ATCA observations, the lat-ter provide a good reference for data verification purposes.The higher frequency resolution of ASKAP will enable fur-ther dynamical analysis of spectrally distinct components ofthe galaxy triplet and its tidal streams.In Section 2 of this paper, we describe the observationalsetup used during various commissioning and Early Sciencephases of the ASKAP-12 array. Section 3 details the dataprocessing and imaging completed with ASKAPsoft . InSection 4, we present our measurements, final H i maps of theregion and stellar properties from ancillary optical imaging.We analyse and discuss the newly resolved H i sources inSection 5 and Section 6 contains concluding remarks. The majority of the observations of the NGC 7232 field weretaken during the inaugural month of the ASKAP Early Sci-ence program. Additional observations were obtained in Au-gust and September during different commissioning phasesof the array in 2016 and 2017. The same configuration ofantennas was used throughout 2016 when only 48 MHz of si-multaneous bandwidth was achievable. The bandwidth wassplit into 2592 independent channels, each 18.5 kHz wide(equivalent to 3.9 km s − at 1420 MHz). After numerous sys-tem upgrades and further commissioning, more bandwidth MNRAS000 , 1–16 (2019) he NGC 7232 galaxy group Figure 1.
ASKAP positional diagram of antennas used to ob-serve the NGC 7232 field (with antenna 02 set as the centre ofthe array). The insert shows the central antennas. Due to com-puting limitations during these Early Science and commissioningobservations, up to 12 antennas were used for each night of ob-servations (see Tables 1 and 2). and new antennas were available for the main array, whichchanged the available baselines in 2017 (see Table 1 andFig. 1). Nevertheless, due to computing limitations duringEarly Science and caution exercised during the later com-missioning phases, no more than 12 antennas were used ona given night of observations on the NGC 7232 field.Throughout the Early Science program, science targetswere generally observed at night, after ASKAP developmentand commissioning activities. Table 2 summarizes all com-missioning and Early Science observations on the NGC 7232field, totalling 18 usable nights. A square × beam foot-print, yielding a . × . degree field of view, was chosento maximise the sky coverage. Two interleaves – footprintA (centred at RA = 22:13:07.7, Dec = -45:16:57.1, J2000)and footprint B (RA = 22:10:35.41, Dec = -44:49:50.7) –were combined to fill in the gaps between beams and pro-duce a more uniform sensitivity pattern (see Figure 2). Thedata will be publicly available on CASDA under the appro-priate ASKAP scheduling block identifying number (SBID;Table 2). Bhandari et al. (2018) have conducted a search forvariable and transient continuum sources using footprint Bfrom the same set of observations.In September 2016, there was testing of rotating thefootprint by 45 ◦ to better match the configuration of elec-tronics on the PAFs and thereby increase the sensitivity ofthe corner beams. The formed beams were not re-rotated re-sulting in a diamond shape on the sky for those observations.This rotated beam forming method proved to be beneficialand became standard starting December 2016. The obser-vations taken in 2017 used re-rotated beams to match thesquare shape of the October 2016 observations.Each night began with a few hours of observations onthe primary calibrator, PKS1934-638, that was sequentiallypositioned at the centre of each of the 36 beams for a chosenamount of time per beam ( t calibrator). During our Early Figure 2.
ASKAP 36-beam footprints for the NGC 7232 field ob-servations: red = footprint A, green = rotated footprint A, blue =footprint B (see text for further details). The black diamonds in-dicate the location of HIPASS sources. The darker beams indicatethe four beams from each footprint, focused on the NGC 7232/3triplet, that was processed and imaged for this paper.
Science observations, we were able to determine that 200seconds per beam was sufficient for post-observing calibra-tions. Longer integration times would significantly increaseobserving overheads without appreciably improving the cal-ibration procedure.Due to technical issues on various nights, whichstemmed from commission tests on the overall stability ofthe array during Early Science, the data from one or twoantennas – as specified in Table 2 – were omitted from theprocessing procedure. The equivalent ASKAP-12 observingtime ( t equivalent) has been computed using the followingrelation, based on the standard radiometer equation for in-terferometry: t equivalent = N ( N − ) t science , (1)where N is the number of antennas used during each obser-vation, t observed is the actual time spent on the sciencesource and the factor of 132 is from 12*(12-1) total an-tennas. The goal was to achieve full WALLABY sensitiv-ity (i.e. 1.6 mJy per beam per channel; Koribalski 2012) bycombining multiple nights of data. Out of the total observingtime of t equivalent = 180h, 150 hours were used to producethe final image cube that has a sensitivity of ∼ η − = 85K, which was the valueestimated for the telescope at the time of the observations. MNRAS , 1–16 (2019)
K. Lee-Waddell et al.
Table 1.
ASKAP-12 array configurations used during the observations of the NGC 7232fieldDate Observed Antennas available Minimum Maximumfrequency range baseline baseline(MHz) (m) (m)Aug - Oct 2016 1376.5 - 1424.5 02, 04, 05, 10, 12, 13, 14, 16, 24, 27, 28, 30 60 2300Aug 2017 1248.5 - 1440.5 03, 04, 05, 06, 10, 12, 14, 17, 19, 24, 30 50 2120Sep 2017 1200.5 - 1440.5 02, 03, 04, 06, 10, 14, 16, 17, 19, 27, 28, 30 20 2300
Table 2.
ASKAP Early Science observations of the NGC 7232 field.Obs. date ASKAP Footprint t calibrator t science Omitted t equivalent † NotesSBID interleave (s) (s) antennas (h)11 Aug 2016 1927 A 200 43207.2 02, 14 8.212 Aug 2016 1934 A 200 41525.1 14 9.65 Sep 2016 2012 A 200 32348.2 02 7.5 footprint rotated 45 ◦ ◦ † see Equation 1 The observations from individual beams for each night wereedited, calibrated, and imaged using
ASKAPsoft (Whitinget al., in prep) – the processing pipeline specifically designedand developed for ASKAP – on the Galaxy Cray in thePawsey Supercomputing Centre. Processing the entirety ofthe Early Science data for the NGC 7232 field would havetaken several weeks of computing time and resulted in over120 TB of intermediate and final data products. For thepurpose of this paper, and in order to save time and diskspace, we only used the relevant portion of the entire dataset. We processed four beams from each night (footprint A =beams 02,03,11,12; rotated footprint A = beams 01,03,08,09;footprint B = beams 03,09,10,11; see Figure 2) and a reducedbandwidth of 16 MHz around the H i line. The bandwidthwas split into 864 channels, maintaining the 18.5 kHz (=3.9 km s − ) spectral resolution, covering a frequency rangebetween 1404.5 to 1420.5 MHz and an equivalent velocityrange of -16 to 3391 km s − .Most of the processing parameters were the default set-tings for ASKAPsoft version 0.19.7 . These settings in-cluded the use of the automated dynamic flagging utilitythat identifies outlying signals as well as the use of self-calibration to correct the time dependent fluctuations of the https://doi.org/10.4225/08/5ac2dc9a16430 antenna gains (both amplitude and phase) during the obser-vation of the science target. Changes to the default process-ing parameters were selected to omit autocorrelations andany antennas specified in Table 2. The calibrator observa-tions on PKS 1934-638 were often taken at low elevation inthe late afternoon. Solar interference can introduce spurioussignals that are more pronounced on the shorter baselines.As a cautionary measure, any calibrator data from baselinesshorter than 200m were excluded from the processing pro-cedure.In the imaging stage, due to the limited number ofshorter baselines, natural weighting resulted in a “patchy”noise pattern across the spectral line image cube. As such, aweighting of robust = 0.5 was chosen to better utilise the uv-coverage pattern of ASKAP-12, which resulted in smootherbackground noise. The preliminary image cubes had a beamsize of ∼ × arcsec. A Gaussian taper of 30 arcsec was ap-plied to achieve a synthesized beam of 35.5 ×
30 arcsec andbring out H i of lower column density. Due to residual arte-facts from the uv-based model continuum subtraction, thespectral line image cubes for each beam also went throughimage-based continuum subtraction prior to being linearlymosaicked together using the linmos task in ASKAPsoft .The details of the final ASKAP image cube are summarizedin Table 3.
MNRAS000
MNRAS000 , 1–16 (2019) he NGC 7232 galaxy group Figure 3.
ASKAP per channel RMS noise of the combined imagecubes compared to theoretical predictions of ASKAP-12. Singlefootprint values are measured from the highest sensitivity regionof footprint A and are representative of the channels containingH i sources. The noise values for multiple footprints were measuredfrom that same spatial region, which coincides with the highestsensitivity part of those cubes, and includes data from footprintB and the rotated footprint. Table 3.
Properties of final ASKAP H i image cubeOn-source observing time (combined hours) 150Weighting scheme robust = 0.5Pixel size (arcsec) 4Gaussian taper (arcsec) 30 × × − ) 3.9RMS - central region (mJy per beam per channel) ∼ ∼ The 4-beam mosaicked image cubes for each single nightof observations, which had a typical RMS noise of ∼ . The highest sen-sitivity region of the final cube achieved an RMS noise of this final image cube is publicly available on CASDA;https://doi.org/10.25919/5becef4d41dab ∼ ASKAPsoft de-velopment; nevertheless, due to the preliminary ASKAP-12 observing techniques as well as the early nature of the
ASKAPsoft processing pipeline, some artefacts remain inthe final cube. For example, residual RFI signals that werenot fully excised from the data produce large scale-stripingacross certain channels. There are also a handful of spatiallycompact negative signals that occur in channels with rela-tively bright H i emission – possibly due to calibration issues(see Grobler et al. 2014) or that have been introduced dur-ing the image combination phase – which require furtherinvestigation.In addition, linmos assumes circular Gaussian beamshapes; however, ASKAP’s electronic beams are currentlyformed using an algorithm designed to maximise sensitivityrather than constrain beam shape (for further details onPAFs and beam forming, see Hay & O’Sullivan 2008 andHay & Bird 2016). Holography measurements have shownthat the resulting beam patterns have some ellipticity inindividual polarisations and are non-Gaussian as a result ofcoma distortion at the edge of the field of view. The beamshapes also vary slightly from one antenna to another. Aquantitative analysis of beam properties and behaviour willbe published in future. For this current analysis, we notethat errors in the primary beam correction can contributeup to 10 percent uncertainty in source flux estimates and ahigher noise level in the multi-footprint combined data (seeSerra et al. 2015b, Heywood et al. 2016, McConnell et al.2016).The aforementioned artefacts and imaging concerns donot appear to significantly affect the scientific results pre-sented in this paper. As more antennas are commissionedand added to the array, the increased uv-coverage will im-prove the sensitivity of ASKAP. Furthermore, enhancementsto the telescope (e.g. having local ingest nodes to preventdata loss and using an on-dish calibration system to trackgain/phase changes) as well as improvements in the ob-serving technique (e.g. using rotated footprints and shape-constrained beams) are progressively improving the dataquality with each iteration of Early Science observations. Using our final ASKAP H i cube of the target region, we de-tect 17 H i sources in the immediate vicinity of NGC 7232/3triplet (i.e. between 1100 - 3300 km s − ) that are likely con-tained within the NGC 7232 group. Eleven of these detec-tions are identified with stellar counterparts, of which fiveare newly resolved H i galaxies. The other six H i detectionsare likely tidal debris associated with the triplet.Source extraction was conducted using an automatedapplication, SoFiA (Serra et al. 2015a). We chose a 5-sigmathreshold for source detection with SoFiA and all extractedsources were visually inspected for verification. By choos-ing a high source finding threshold, most of the processing MNRAS , 1–16 (2019)
K. Lee-Waddell et al. artefacts are generally rejected by SoFiA; however, not allof the H i flux is recovered for each source. Alternatively, alower threshold detects more galaxies and more of their fluxbut also increases the false detection rate due to the faintartefacts in the ASKAP cube. The number of H i sources re-ported here is a conservative value for this field, which willbe revisited during the full WALLABY survey with ASKAP-36. For the purpose of this paper, we consider any SoFiAdetections in the ASKAP cube to be authentic H i sourcesif they coincide (spatially and spectrally) with previouslydetected ATCA and/or HIPASS H i sources or if they ap-pear to have a stellar counterpart. Since we are working withASKAP Early Science data, we assume any other detectionsthat are identified by SoFiA but do not fit the previouslymentioned criteria to be imaging artefacts or residual side-lobe features. Table 4 presents the ASKAP H i sources andtheir spatial location, as extracted by SoFiA, crossmatchedwith HIPASS and ATCA detections as well as likely stel-lar counterparts, where applicable. All previously known H i sources that fall within the 12-beam footprint and selectedvelocity range were detected in the ASKAP cube. We notethat Barnes & Webster (2001) focus their ATCA observa-tions on a × arcsec region centred on the NGC 7232/3triplet. As such, we use the wider field HIPASS cube fora more uniform comparison of all H i sources detected byASKAP. Figure 4 shows the SoFiA-produced H i total inten-sity (moment 0) contours of both the ASKAP and HIPASSdata.The H i emission from the members directly associatedwith the interacting triplet system, shown in Figure 5, iscomplex. In the spectra presented in Figure 6, NGC 7232Bis clearly differentiated by its unique velocity range (2120 -2230 km s − ) in both the ASKAP and HIPASS observations.ASKAP provides sufficient angular resolution to distinguishsix H i clouds – referred to as H i clouds C1-C6 – from themain body of the triplet (see Figure 7).Individual moment maps for H i sources in the surround-ing area can be found in the appendix. H i galaxies thatare also detected in HIPASS and are fairly well resolved inthe ASKAP data are shown in Figure A1. Newly detectedASKAP H i sources that appear to have stellar counterparts(indicated by the blue diamonds in Figure 4) are shown inFigure A2. All H i velocity (moment 1) maps were generatedby applying the SoFiA output mask to the original datacube. Additional masking has been applied for the moment1 maps of H i clouds C1-C6 in Figure 7. For each ASKAP de-tected H i source, manually defined ellipses on the unmaskedcube – using the SoFiA outputs as a guide – were used tocreate all spectral profiles.Table 5 summarizes the H i properties of each source, asmeasured by SoFiA. The uncertainty in the central velocity(v HI ) is σ ( v HI ) = ( S / N ) − ( P δ v ) / , (2)where S / N is the median signal-to-noise ratio of the flux foreach source, P is half the difference of the velocity widthmeasured at 50 percent (W ) and 20 percent (W ) of thepeak flux and δ v = 3.9 km s − is the spectral resolutionof the ASKAP data (see Koribalski et al. 2004 for deriva-tion details). Uncertainties in the line widths are σ (W ) =2 σ (v HI ) and σ (W ) = 3 σ (v HI ). The 20 percent flux uncer- tainty takes into account calibration/processing as well as in-strumental effects. Even with the 5-sigma detection thresh-old, the SoFiA computed fluxes are, on average, within 15percent of the H i flux manually computed from the spectralprofiles of each galaxy. Table 5 also includes the local RMSnoise level around each source and the integrated S/N of thedetection. A common distance of 24 Mpc is assumed for allgroup members to compute the H i masses (M HI ). MNRAS000
K. Lee-Waddell et al. artefacts are generally rejected by SoFiA; however, not allof the H i flux is recovered for each source. Alternatively, alower threshold detects more galaxies and more of their fluxbut also increases the false detection rate due to the faintartefacts in the ASKAP cube. The number of H i sources re-ported here is a conservative value for this field, which willbe revisited during the full WALLABY survey with ASKAP-36. For the purpose of this paper, we consider any SoFiAdetections in the ASKAP cube to be authentic H i sourcesif they coincide (spatially and spectrally) with previouslydetected ATCA and/or HIPASS H i sources or if they ap-pear to have a stellar counterpart. Since we are working withASKAP Early Science data, we assume any other detectionsthat are identified by SoFiA but do not fit the previouslymentioned criteria to be imaging artefacts or residual side-lobe features. Table 4 presents the ASKAP H i sources andtheir spatial location, as extracted by SoFiA, crossmatchedwith HIPASS and ATCA detections as well as likely stel-lar counterparts, where applicable. All previously known H i sources that fall within the 12-beam footprint and selectedvelocity range were detected in the ASKAP cube. We notethat Barnes & Webster (2001) focus their ATCA observa-tions on a × arcsec region centred on the NGC 7232/3triplet. As such, we use the wider field HIPASS cube fora more uniform comparison of all H i sources detected byASKAP. Figure 4 shows the SoFiA-produced H i total inten-sity (moment 0) contours of both the ASKAP and HIPASSdata.The H i emission from the members directly associatedwith the interacting triplet system, shown in Figure 5, iscomplex. In the spectra presented in Figure 6, NGC 7232Bis clearly differentiated by its unique velocity range (2120 -2230 km s − ) in both the ASKAP and HIPASS observations.ASKAP provides sufficient angular resolution to distinguishsix H i clouds – referred to as H i clouds C1-C6 – from themain body of the triplet (see Figure 7).Individual moment maps for H i sources in the surround-ing area can be found in the appendix. H i galaxies thatare also detected in HIPASS and are fairly well resolved inthe ASKAP data are shown in Figure A1. Newly detectedASKAP H i sources that appear to have stellar counterparts(indicated by the blue diamonds in Figure 4) are shown inFigure A2. All H i velocity (moment 1) maps were generatedby applying the SoFiA output mask to the original datacube. Additional masking has been applied for the moment1 maps of H i clouds C1-C6 in Figure 7. For each ASKAP de-tected H i source, manually defined ellipses on the unmaskedcube – using the SoFiA outputs as a guide – were used tocreate all spectral profiles.Table 5 summarizes the H i properties of each source, asmeasured by SoFiA. The uncertainty in the central velocity(v HI ) is σ ( v HI ) = ( S / N ) − ( P δ v ) / , (2)where S / N is the median signal-to-noise ratio of the flux foreach source, P is half the difference of the velocity widthmeasured at 50 percent (W ) and 20 percent (W ) of thepeak flux and δ v = 3.9 km s − is the spectral resolutionof the ASKAP data (see Koribalski et al. 2004 for deriva-tion details). Uncertainties in the line widths are σ (W ) =2 σ (v HI ) and σ (W ) = 3 σ (v HI ). The 20 percent flux uncer- tainty takes into account calibration/processing as well as in-strumental effects. Even with the 5-sigma detection thresh-old, the SoFiA computed fluxes are, on average, within 15percent of the H i flux manually computed from the spectralprofiles of each galaxy. Table 5 also includes the local RMSnoise level around each source and the integrated S/N of thedetection. A common distance of 24 Mpc is assumed for allgroup members to compute the H i masses (M HI ). MNRAS000 , 1–16 (2019) he NGC 7232 galaxy group Figure 4.
Total H i intensity (moment 0) contours from ASKAP (white; at (1, 3, 6) × atoms cm − ) and HIPASS (yellow) sourceswithin the selected velocity range – extracted using SoFiA – superimposed on an archival DSS2 Blue optical image. The correspondingbeam sizes for each set of H i observations are shown in the bottom left and a physical scale bar based on a group distance of ∼
24 Mpc(Garcia 1995) is in the upper right corner. HIPASS detected sources are as labelled. Blue diamonds indicate new ASKAP H i detectionsthat appear to have stellar counterparts, while the green circles indicate likely tidal debris from the interacting triplet, which have nodetectable stellar counterparts. The red contours represent the sensitivity pattern of the 12-beam footprint, at 15, 50 and 90 percent ofthe peak sensitivity, based on the normalized number of visibilities used for each part of the image. The NGC 7232/3 triplet resides inthe highest sensitivity region of the imaged data. Due to the preliminary nature of the ASKAPsoft processing pipeline, a few minorartefacts remain – including sidelobe features for IC 5171 – were picked up by the source finder.MNRAS , 1–16 (2019)
K. Lee-Waddell et al.
Figure 5.
ASKAP H i moment maps of the NGC 7232/3 triplet. Left panel: moment 0 contours – at (1, 3, 6) × atoms cm − –superimposed on DSS2 coloured image. Three H i peaks coincide with the stellar components of the major galaxies. Two additional H i clouds, C5 and C6, are clearly visible in the intervening region connecting the northern spiral, NGC 7232B, to the other two galaxiesNGC 7232 and NGC 7233. Right panel: moment 1 map. See Figure 7 for moment 1 maps with narrower velocity ranges centred aroundeach source. Figure 6.
ASKAP H i spectra of the NGC 7232/3 triplet andsurrounding tidal debris. Solid green = H i clouds C1 and C2 (ex-tracted from the ASKAP cube), solid blue = ASKAP, red dotted= HIPASS. Due to the preliminary nature of the v.0.19.7 pro-cessing pipeline, over-subtraction of the continuum near brightH i sources has resulted in negative features in the extracted spec-tra. MNRAS000
ASKAP H i spectra of the NGC 7232/3 triplet andsurrounding tidal debris. Solid green = H i clouds C1 and C2 (ex-tracted from the ASKAP cube), solid blue = ASKAP, red dotted= HIPASS. Due to the preliminary nature of the v.0.19.7 pro-cessing pipeline, over-subtraction of the continuum near brightH i sources has resulted in negative features in the extracted spec-tra. MNRAS000 , 1–16 (2019) he NGC 7232 galaxy group Figure 7.
ASKAP H i emission associated with the NGC 7232/3 triplet. Centre: ASKAP H i moment 0 contours – at (1, 3, 6) × atoms cm − – superimposed on DSS2 Blue archival images. Pertinent optical galaxies are labelled in white. A physical scale bar assuminga group distance of 24 Mpc is shown in the bottom left of the panel. Outer: ASKAP moment 1 maps of the H i emission in the velocityrange centred around each source, as labelled. Additional masking of neighbouring sources has been applied to H i clouds C1-C6. Onaccounts of their spatial and spectral proximity, H i cloud C1 and C2 are shown with the same velocity colour scale.MNRAS , 1–16 (2019) K . L ee - W add e ll e t a l . Table 4.
Crossmatches of the ASKAP H i sources detected in the NGC 7232 groupSource Name ASKAP RA, Dec HIPASS HIPASS v HI 1
HIPASS flux ATCA v
HI 2
ATCA flux Stellar counterpart vstellar(J2000) designation (km s − ) (Jy km s − ) (km s − ) (Jy km s − ) (km s − )WALLABY J221010-455159 22:10:10.30, -45:51:59.4 – – – – – MRSS 288-021830 –WALLABY J221056-460500 22:10:56.45, -46:05:00.8 J2210-46 2990.9 15.8 – – IC 5171 2827 ± WALLABY J221136-455439 22:11:36.26, -45:54:39.6 – – – – – AM 2208-460 –WALLABY J221148-453525 22:11:48.88, -45:35:25.9 J2211-45 1918.2 6.6 – – ESO 288-G049 1968 WALLABY J221339-463746 22:13:39.98, -46:37:46.9 – – – – – MRSS 289-168885 –WALLABY J221341-455343 22:13:41.60, -45:53:43.3 – – – – – NGC 7232A (ESO 289-G003) 2224 ± WALLABY J221403-455806 22:14:03.51, -45:58:06.9 – – – – – – –WALLABY J221413-455723 22:14:13.86, -45:57:23.1 J2214-45 1864.8 11.9 – – – –WALLABY J221506-461658 22:15:06.41, -46:16:58.4 – – – – – ESO 289-G005 1885 ± WALLABY J221512-455442 22:15:12.23, -45:54:42.5 – – – – – – –WALLABY J221520-455009 22:15:20.28, -45:50:09.6 – – – – – – –WALLABY J221545-454843 22:15:45.79, -45:48:43.3 – – – – – – –WALLABY J221547-455007 22:15:47.67, -45:50:07.0 J2215-45a 1904.9 24.6 1735 † † NGC 7232 and NGC 7233 1887 ± , 1862 ± WALLABY J221551-454702 22:15:51.58, -45:47:02.7 J2215-45b 2158.9 9.5 2160 7.2 NGC 7232B (ESO 289-G009) 2152 ± WALLABY J221556-454903 22:15:56.97, -45:49:03.9 – – – – – – –WALLABY J221642-470708 22:16:42.68, -47:07:08.5 J2216-47 2681.7 12.2 – – ESO 289-G010 2785 WALLABY J221716-450359 22:17:16.45, -45:03:59.4 J2217-45a 1784.7 10.5 – – ESO 289-G011 1822 ± Meyer et al. (2004), Barnes & Webster (2001), da Costa et al. (1991), Tully et al. (2008), Jones et al. (2009). † These values are for NGC 7232 only as NGC 7233 was reported asan unresolved source by Barnes & Webster (2001), see Section 5 for more details. M N R A S , ( ) h e N G C g a l a x y g r o u p Table 5.
ASKAP H i properties of the sources detected in the NGC 7232 groupSource name ASKAP v HI ASKAP W ASKAP W ASKAP flux Local RMS Integrated ASKAP M HI Comments(km s − ) (km s − ) (km s − ) (Jy km s − ) (mJy beam − ) S/N ( × M (cid:12) )WALLABY J221010-455159 2026 ± ± ± ± ± i detectionWALLABY J221056-460500 2831 ± ± ±
10 11 ± ± ± ±
10 40 ±
20 0.6 ± ± i detectionWALLABY J221148-453525 1964 ± ± ± ± ± ± ± ± ± ± i detectionWALLABY J221341-455343 2348 ± ±
10 230 ±
10 2.0 ± ± i detectionWALLABY J221403-455806 1692 ± ±
10 60 ±
20 0.7 ± ± i cloud C1WALLABY J221413-455723 1678 ± ± ±
10 2.4 ± ± i cloud C2WALLABY J221506-461658 1931 ± ±
10 150 ±
20 2.1 ± ± i detectionWALLABY J221512-455442 1668 ± ±
10 50 ±
20 0.4 ± ± i cloud C3WALLABY J221520-455009 1999 ± ±
10 40 ±
20 0.3 ± ± i cloud C4WALLABY J221545-454843 1874 ± ±
10 60 ±
20 0.5 ± ± i cloud C5WALLABY J221547-455007 1863 ± ± ±
10 12 ± ± ± ± ± ± ± ± ±
10 110 ±
20 2.1 ± ± i cloud C6WALLABY J221642-470708 2768 ± ± ±
10 12 ± ± ± ± ± ± ± M N R A S , ( ) K. Lee-Waddell et al.
To further explore and possibly disentangle some of the com-plexity of the triplet system, we used ancillary data to esti-mate the stellar properties of the three major optical galax-ies, NGC 7232, NGC 7232B and NGC 7233. We measuredthe g - and r -band fluxes using SkyMapper images (Wolf etal. 2018). These fluxes were then used estimate stellar masses(M ∗ ) by applying the relation from Bell et al. (2003). Wemeasured the far-ultraviolet (FUV) flux from GALEX (Mar-tin et al. 2005) and W4 (22 µ m) flux from WISE (Wright etal. 2010) to estimate the star formation rate (SFR) followingthe method described in Wang et al. (2017). SFRW4 indi-cates the dust-attenuated SFR and SFRFUV is the unat-tenuated portion. As such, the sum of these two values pro-vides the total SFR (Wang et al. 2017). All stellar flux mea-surements were obtained using Petrosian apertures and withneighbouring galaxies masked out of the images. The r -bandimage was used as the reference for setting the apertures onthe SkyMapper images. For GALEX and WISE, the bandwhich produced the largest Petrosian ellipse determined thefinal aperture used for each galaxy. The stellar propertiesfor the galaxies within the triplet are presented in Table 6. The capabilities of ASKAP, during its Early Science phase,have enabled us to detect and/or resolve the H i emissionof several galaxies. Within the central 12-beam footprinttargeting the NGC 7232 group, there are 17 H i sources.Many of these sources appear to be star-forming galaxieswith clear optical counterparts and are further detailed inthe appendix. This section will discussion the individualsources within the NGC 7232/3 triplet and its neighbour-ing H i clouds.Due to the proximity of NGC 7232, NGC 7232B andNGC 7233 as well as the common H i envelope detected byHIPASS (see Figure 4), it appears that these galaxies are ac-tively interacting (Barnes & Webster 2001, Koribalski et al.2004). The H i corresponding to the optical galaxies withinthe triplet is noticeably disturbed. There appears to be anH i bridge connecting all three galaxies (Figure 5) as well assix tidally formed H i clouds in the neighbouring area (Fig-ure 7). One of these clouds, H i cloud C6, was originally de-tected as an H i plume by Barnes & Webster (2001) in theirATCA observations and has sufficient mass (i.e. M HI > M (cid:12) ) to become a self-gravitating TDG (Lelli et al. 2015,Lee-Waddell et al. 2016). The other clouds are likely moretransient tidal features that will fade into the lower columndensity H i envelope or fall into one of the larger galaxies (seeBournaud & Duc 2006).Adding together the HIPASS values for the NGC7232/3 triplet and neighbouring gas cloud HIPASS J2214-45 (reported by Meyer et al. (2004); see Table 4), the totalH i flux for this region is 46 Jy km s − ; however, HIPASSJ2214-45 is likely a confused source that includes some H i emission also attributed to the NGC 7232 and NGC 7233pair (i.e. HIPASS J2214-45a). Koribalski et al. (2004), cat-aloguing the 1000 brightest sources in HIPASS, report an H i flux of 34.6 ± − for this same region (iden-tified as the AM2212-460 group). From the ASKAP data,we measure a total H i flux of 26 ± − for thetriplet and surrounding H i clouds. Looking at the spectra inFigure 6, ASKAP appears to recover at least 70 percent ofthe H i flux in the spatial and spectral vicinity of the NGC7232/3 triplet, which is in agreement with the HIPASS valuereported by Koribalski et al. (2004).The ∼
25 percent flux difference between ASKAP andHIPASS can be attributed to the column density sensitivitylimit of the ASKAP data as well as the high source detec-tion threshold chosen for SoFiA. Another minor factor withbe the slight negative offset in the current ASKAP spectra(see Figure 6) due to over-subtraction of the continuum nearbright H i sources. Since the galaxies within this region areinteracting, it is also likely that a portion of the diffuse gasis resolved out by the interferometer. NGC 7232 and NGC 7233
The H i components of NGC 7232 and NGC 7233 are thor-oughly blended together and span a broad velocity range(see Figures 5 and 6 ). There does appear to be higher den-sity clumps of H i coinciding with the stellar disks of the twogalaxies; nevertheless, much of the H i from these galaxiesappears to be distributed throughout the spatial region ofthe triplet system. Discrepancies between source detectionand separation has resulted in the inconsistencies of the H i properties, reported in Tables 4 and 5, from each set of ob-servations on NGC 7232 and NGC 7233.The wider variety of array baseline lengths and longerintegration time of the ASKAP observations enable greatersensitivity to extended emission than the ATCA observa-tions by Barnes & Webster (2001). Although the ATCA ob-servations detect NGC 7233 (see the ATCA H i moment 0map in figure 8 of Barnes & Webster 2001), it was determineby those authors that the source is unresolved and no fur-ther H i measurements were reported. Comparing the H i mo-ment maps, ASKAP does appear to recover more H i emis-sion that ATCA; however, about half of the H i flux that wascatalogued by HIPASS has not been recovered by ASKAP.Subtracting the ASKAP detections from the HIPASS cubeindicates that there is diffuse gas in the area between theNGC 7232/3 triplet and a neighbouring galaxy, IC 5181, sug-gesting that there is an H i tail/bridge in this region. Note,no H i counterpart has been detected across the stellar ex-tent of IC 5181(which is a lenticular galaxy located at RA= 22:13:22, Dec = -46:01:03, vstellar = ± km s − ;Jones et al. 2009).The stellar component of NGC 7232 is ten times moremassive than that of NGC 7233 (Table 6). Although we areunable to distinguish the H i originating from each galaxy,since they are located within the same group environment,one can assume that NGC 7232 would have started withmore H i than NGC 7233 (see Denes, Kilborn & Koribalski2014 for more details on mass scaling relations). Currently,the H i peaks coinciding with the optical centres of the twogalaxies are quite comparable in size and column density,suggesting that most of the H i originating from NGC 7232is now spread throughout triplet system, which may explainthis galaxy’s low SFR. MNRAS000
The H i components of NGC 7232 and NGC 7233 are thor-oughly blended together and span a broad velocity range(see Figures 5 and 6 ). There does appear to be higher den-sity clumps of H i coinciding with the stellar disks of the twogalaxies; nevertheless, much of the H i from these galaxiesappears to be distributed throughout the spatial region ofthe triplet system. Discrepancies between source detectionand separation has resulted in the inconsistencies of the H i properties, reported in Tables 4 and 5, from each set of ob-servations on NGC 7232 and NGC 7233.The wider variety of array baseline lengths and longerintegration time of the ASKAP observations enable greatersensitivity to extended emission than the ATCA observa-tions by Barnes & Webster (2001). Although the ATCA ob-servations detect NGC 7233 (see the ATCA H i moment 0map in figure 8 of Barnes & Webster 2001), it was determineby those authors that the source is unresolved and no fur-ther H i measurements were reported. Comparing the H i mo-ment maps, ASKAP does appear to recover more H i emis-sion that ATCA; however, about half of the H i flux that wascatalogued by HIPASS has not been recovered by ASKAP.Subtracting the ASKAP detections from the HIPASS cubeindicates that there is diffuse gas in the area between theNGC 7232/3 triplet and a neighbouring galaxy, IC 5181, sug-gesting that there is an H i tail/bridge in this region. Note,no H i counterpart has been detected across the stellar ex-tent of IC 5181(which is a lenticular galaxy located at RA= 22:13:22, Dec = -46:01:03, vstellar = ± km s − ;Jones et al. 2009).The stellar component of NGC 7232 is ten times moremassive than that of NGC 7233 (Table 6). Although we areunable to distinguish the H i originating from each galaxy,since they are located within the same group environment,one can assume that NGC 7232 would have started withmore H i than NGC 7233 (see Denes, Kilborn & Koribalski2014 for more details on mass scaling relations). Currently,the H i peaks coinciding with the optical centres of the twogalaxies are quite comparable in size and column density,suggesting that most of the H i originating from NGC 7232is now spread throughout triplet system, which may explainthis galaxy’s low SFR. MNRAS000 , 1–16 (2019) he NGC 7232 galaxy group Table 6.
Stellar properties of the NGC 7232/3 tripletSource name r -band g -band M* FUV SFRFUV W4 SFRW4 SFRtotalmagnitude magnitude × M (cid:12) magnitude M (cid:12) yr − magnitude M (cid:12) yr − M (cid:12) yr − NGC 7232 11.99 ± ± ± ± ± ± ± ± ± ± ± ± SkyMapper (Wolf et al. 2018), GALEX (Martin et al. 2005), WISE (Wright et al. 2010)
NGC 7232B
The H i associated with NGC 7232B, a face-on spiral, is cen-tred at v HI = ± km s − and is kinematically distinctfrom NGC 7232 and NGC 7233, which span a H i velocityrange of v HI ∼ − (see Figure 6). The gasin the outer region of NGC 7232B shows significant signs ofspatial distortion. In the moment maps of this galaxy (Fig-ure 5), there appears to be a southwest extension towardsH i cloud C5 and a clump of gas east of the central regionof the galaxy. ASKAP recovers the same amount of H i fluxas ATCA and over 70 percent of the emission detected byHIPASS for this galaxy. This result suggests that most of thegas associated with NGC 7232B is not too diffusely spreadout and easily recovered by the two interferometers.NGC 7232B is actively star forming, especially com-pared to the other two galaxies in the system. The H i massfraction (M HI /M ∗ ) for NGC 7232B (Tables 5 and 6) is com-parable to other gaseous galaxies of similar mass in the localuniverse (Wang et al. 2017). This result indicates that theH i associated with NGC 7232B remains primarily with thisgalaxy and has yet to be significantly affected (i.e. tidallystripped) by the ongoing interaction event within the tripletsystem. H i clouds C1 and C2 Interactions between gas-rich galaxies can produce tidal tailsseveral 100 kpc in length (e.g. Leisman et al. 2016, Oosterlooet al. 2018) and within these tails, high density clumps of H i are formed. Under the right circumstances, these H i clumpscan accrete sufficient amounts of material to eventually be-come self-gravitating TDGs (Mirabel, Dottori, Lutz 1992,Lelli et al. 2015, Lee-Waddell et al. 2016). H i clouds C1 andC2 are located in the densest region of HIPASS J2214-45,the latter being a confused HIPASS source with NGC 7232and NGC 7233. The high-resolution ASKAP observationsrecover ∼
30 percent of the H i attributed to HIPASS J2214-45. The remaining portion of the H i gas is likely fairly diffuseand resides in the aforementioned tidal bridge connecting theNGC 7232/3 triplet with other group members.H i cloud C1 has a relatively low H i mass and is morelikely to be a transient tidal feature as it just reaches the M (cid:12) mass threshold for long term survivability (Bour-naud & Duc 2006); whereas, H i cloud C2 has an H i mass of . ± . × M (cid:12) . In the moment 1 map of Figure 7, thenortheast portion of H i cloud C2 appears to have a smoothvelocity gradient. If this portion of the cloud is rotating, thenits estimated total dynamical mass would be on the order ofMdyn ∼ M (cid:12) (see equation 2 in Lee-Waddell et al. 2016).However, the lack of a clear stellar counterpart combined with this potentially large dynamical mass measurement im-plies that the velocity gradient across the northeast portionof H i cloud C2 is not due to self-gravitation induced rota-tion. Rather, it may be the result of shearing effects from theinteraction process between the larger group members. Thismass discrepancy between the baryonic (i.e. gas and stellar)mass and total mass does not take into account the ionizedgas component, which can be up to three times the mass ofthe H i gas in tidal features (see Fox et al. 2014). Neverthe-less, the irregular morphology of the southwest portion of H i cloud C2 also opposes the rotation scenario. With the cur-rent data, we are unable to assess the longevity of H i cloudC2. It appears that IC 5181 – located ∼
150 kpc in projectionfrom the NGC 7232/3 triplet, at the distance of the group– forms a line with H i clouds C1 and C2 and the triplet(see Figure 7). Previous optical observations of IC 5181 byPizzella et al. (2013) suggest that the ionized gas compo-nent along the polar regions of this galaxy have an externalorigin. These H i clouds could be remnants of the accretion/ merging event that deposited ionized gas onto the undis-turbed stellar disk of IC 5181 (Pizzella et al. 2013), or tidaldebris from the NGC 7232/3 interaction – if the events areseparate. H i clouds C3 and C4 H i clouds C3 and C4 are likely high density peaks associatedwith the tidal bridge. Although these clouds are low in H i mass and likely to be short-lived (Bournaud & Duc 2006),they show velocity gradients that are a testament to the highspatial and spectral capabilities of ASKAP. Looking closelyat the ASKAP data, there appears to be a tentative detec-tion of very faint H i emission connecting H i cloud C3 to thetriplet. However, given that much of the gas in the vicin-ity of H i clouds C3 and C4 is diffusely distributed and/orbelow our detection threshold, it is difficult to discern theexact origin and significance of these tidal debris sources. H i cloud C5 H i cloud C5 spatially coincides with the southwest extensionfrom NGC 7232B; however, the former has a lower veloc-ity range that is similar to the H i emission associated withNGC 7232 and NGC 7233. Within the image cube H i cloudC5 emerges as a distinct object, brightest between 1870 -1890 km s − , alongside the other H i clouds associated withthe major members of the triplet. Based on its location, H i cloud C5 could be part of a gaseous tidal bridge that isforming between NGC 7232B and the other two spirals inthe triplet. MNRAS , 1–16 (2019) K. Lee-Waddell et al. H i cloud C6 H i cloud C6 is quite bright and fairly compact with its peakflux at ∼ − . Currently, it appears to be spatiallyand spectrally embedded in the gas of its parent galaxiesbut this source has enough mass to become self-gravitating(Bournaud & Duc 2006). The smooth velocity gradient inFigure 7 suggests rotation, but could be the result of shear-ing as H i cloud C6 is moving away from the triplet sys-tem. Assuming the former scenario and that H i cloud C6is in dynamical equilibrium, then estimated total dynamicalmass for this object would be on the order of Mdyn ∼ M (cid:12) . This dynamical mass value is comparable to M HI , whichmakes rotation a plausible explanation and would indicatethat H i cloud C6 could be the progenitor of a dark matterpoor TDG (see Lelli et al. 2015, Lee-Waddell et al. 2016).Although the triplet system is complex – and furthercomplicated by projection effects – it is interesting that H i clouds C5 and C6 are aligned almost perpendicular to a lineconnecting NGC 7232B and NGC 7233 (Figure 5). It is pos-sible that these clouds indicate two tails that are currentlyforming from the interaction between the two face-on spi-rals in the triplet (Toomre & Toomre 1972, Bournaud &Duc 2006).With its high column density ( > × atoms cm − inthe central region), H i cloud C6 has a sufficient H i gas den-sity for star formation (see Schaye 2004 for details aboutstar formation thresholds). There appears to be no stel-lar over-densities associated with this cloud; nevertheless,the archival DSS2 images are likely too shallow to detectthe low-surface brightness counterparts that are associatedwith some tidally-formed H i features (e.g. Lee-Waddell etal. 2014, Janowiecki et al. 2015, Madrid et al. 2018). Deeperoptical imaging could possibly detect faint stellar compo-nents related to initial star formation; however, two bright(B-band magnitude ∼
10) foreground stars HD 211111 (atRA = 22:15:50, Dec = -45:48:56) and HD 211121 (at RA =22:15:58, Dec = -45:50:35) could hinder follow-up observa-tions.
The ASKAP Early Science dataset presented in this paperverifies that the array and its associated processing pipeline
ASKAPsoft – both in moderately preliminary states –are successfully producing scientifically useful data. Overall,within the 12-beam image cube centred on the NGC 7232group, we detect 17 H i sources. Six of these detections arewell-known H i -rich galaxies including one fully interactingpair. Five of these detection are newly resolved H i galax-ies with identifiable stellar counterparts. The remaining sixH i detections are likely tidal debris associated with theNGC 7232/3 triplet.The triplet is a complex system. The H i components ofNGC 7232 and NGC 7233 appear to be fully intertwined.NGC 7232B still retains most of its gas but shows evidencethat it is beginning to interact with the other two galaxies.H i clouds C5 and C6 are likely tidal clumps that have beenproduced by the triplet system and possibly indicate theprojected spatial location of two tidally-formed tails of H i .If H i cloud C6 is moving away from its parent galaxies, it has sufficient mass to eventually decouple from the tidal tailand possibly develop into a long-lived TDG.H i clouds C1-C4 might have been produced by thetriplet system, or these clouds could indicate an earlier in-teraction event involving another group member, such asIC 5181. A portion of the H i within this region, whichwas originally identified as belonging to HIPASS J2214-45is likely diffusely distributed and lies below the detectionthreshold of our current ASKAP observations. Accordingly,it is difficult to ascertain the origins of the H i bridge thatappears to be connecting NGC 7232/3 and IC 5181. Futurework would include adding short-spacing / single-dish ob-servations to the ASKAP data in order to recover more ofthe diffuse gas in and around the tidal features, in order tofurther understand the interaction processes that are takingplace.The high-resolution capabilities of ASKAP producedmoderately well-resolved moment maps of four galaxies (i.e.IC 5171, ESO 288-G049, ESO 298-G010 and ESO 289-G011)that can be used for further kinematic analysis. The newlydetected H i counterparts of five stellar galaxies indicate theirmembership to the group. These nine galaxies provide amore complete picture of the group environment surround-ing the NGC 7232/3 triplet.There are many areas where the data quality and pro-cessing procedure can be greatly improved for ASKAP. Theaddition of more antennas – to provide more complete skycoverage and achieve higher sensitivity with less integrationtime – along with the use of the on-dish calibration systemand an overall improved observing method will show vastimprovements to the data quality. With expanded comput-ing systems and a more robust ASKAPsoft pipeline, largerdatasets can be processed more efficiently. The observationsand results presented in this paper serve as one of the firststeps for a long and fruitful journey of ASKAP H i science. ACKNOWLEDGEMENTS
We thank the reviewer for his/her thoroughly detailed com-ments and suggestions to improve the clarity of this paper.The Australian SKA Pathfinder is part of the Australia Tele-scope National Facility which is managed by CSIRO. Oper-ation of ASKAP is funded by the Australian Governmentwith support from the National Collaborative Research In-frastructure Strategy. This work was supported by resourcesprovided by the Pawsey Supercomputing Centre with fund-ing from the Australian Government and the Government ofWestern Australia, including computational resources pro-vided by the Australian Government under the NationalComputational Merit Allocation Scheme (project JA3). Es-tablishment of ASKAP, the Murchison Radio-astronomyObservatory and the Pawsey Supercomputing Centre areinitiatives of the Australian Government, with support fromthe Government of Western Australia and the Science andIndustry Endowment Fund. We acknowledge the WajarriYamatji as the traditional owners of the Observatory site.Parts of this research were conducted by the Australian Re-search Council Centre of Excellence for All-sky Astrophysics(CAASTRO), through project number CE110001020 as wellas by the Australian Research Council Centre of Excel-lence for All Sky Astrophysics in 3 Dimensions (ASTRO
MNRAS , 1–16 (2019) he NGC 7232 galaxy group REFERENCES
Allison J., et al. 2015, MNRAS, 453, 1249Barnes D.G., et al. 2001, MNRAS, 322, 486Barnes D.G., Webster R.L., 2001, MNRAS, 324, 858Bell E.F., McIntosh D.H., Katz N., Weinberg M.D., 2003, ApJS,149, 289Bhandari S., et al. 2018, MNRAS, 478, 1784Bournaud F., Duc P.-A., 2006, A&A, 456, 481da Costa L.N., Pellegrini P.S., Davis M., Meiksin A., SargentW.L.W., Tonry J.L., 1991, ApJS, 75, 935DeBoer D., et al., 2009, IEEEP, 97, 1507Denes H., Kilborn V.A., Koribalski B.S., 2014, MNRAS, 444, 667Duc P.-A., Paudel S., McDermid R. M., Cuillandre J.-C., SerraP., Bournaud F., Cappellari M., Emsellem, E., 2014, MNRAS,440, 1458Duffy A.R.. Meyer M., Staveley-Smith L., Bernyk M., CrotonD.J., Koribalski B.S., Gerstmann D., Westerlund S., 2012,MNRAS., 426., 3385English J., Koribalski B., Bland-Hawthorn J., Freeman K.C., Mc-Cain C.F., 2010, AJ, 139, 102Fox A.J., et al., 2014, ApJ, 787, 147Garcia A.M., 1995, A&A, 297, 56Garcia A.M., Paturel G., Bottinelli, L., Gouguenheim L., 1993,A&AS, 98, 7Giovanelli R., et al., 2005, AJ, 130, 2598Grobler T.L., Nunhokee C.D., Smirnov O.M., van Zyl A.J., deBruyn A.G., 2014, MNRAS, 439, 4030Hay S.G., O’Sullivan J.D., 2008, RaSc, 43, 6S04Hay S.G., Bird T.S., 2016, Applications of Phased Array Feedersin Reflector Antennas, Springer Singapore, p.3139Heywood I., et al., 2016, MNRAS, 457, 4160Hotan A.W., et al., 2014, PASA, 31, 41Janowiecki S., 2015, ApJ, 801, 96Jarrett T.H., et al., 2017, ApJ, 836, 182Johnston S., et al., 2007, PASA, 24, 174Johnston S., et al., 2008, ExA, .22., 151Jones D., et al., 2009, MNRAS, 399, 683Koribalski B.S., et al., 2004, AJ, 128, 16Koribalski B.S., 2012, PASA, 29, 359Lee-Waddell K., et al., 2014, MNRAS, 443, 3601Lee-Waddell K., et al., 2016, MNRAS, 460, 2945Leisman L., Haynes M.P., Giovanelli R., Jozsa G., Adams E., HessK., 2016, MNRAS, 463, 1692Lelli F., et al., 2015, A&A, 584,113Madrid J.P., Lee-Waddell K., Serra P., Koribalski B.S., SchirmerM., Spekkens K., Wang J., 2018, ApJL, 854, 6Martin D.C., et al., 2005, ApJ, 619, 1McConnell D., et al., 2016, PASA, 33, 42Meyer M.J., et al., 2004, MNRAS, 350, 1195Mirabel I., Dottori H., Lutz D., 1992, A&A, 256, 19Mullan B., et al., 2013, ApJ, 768, 194Oosterloo T.A., Zhang M.L., Lucero D.M., Carignan C.,arXiv:1803.08263Pizzella A., Morelli L., Corsini E. M., Dalla Bonta E., Cesetti M.,2013, A&A, 560, A14Ploeckinger S., Sharma K., Schaye J., Crain R.A., Schaller M.,Barber C., 2018, MNRAS, 474, 580Reynolds T.N., et al., 2019, MNRAS, 482, 3591Ryder S.D., et al., 2001, ApJ, 555, 232 Schaye J., 2004, ApJ, 609, 667Serra P., et al., 2015, MNRAS, 448, 1922Serra P., et al., 2015, MNRAS, 452, 2680Toomre A., Toomre J., 1972, ApJ, 178, 623Tully R.B., 1987, ApJ, 321, 280Tully R.B., Courtois H.M., Sorce J.G., 2016, AJ, 152, 50Tully R.B., Shaya E.J., Karachentsev I.D., Courtois H.M., Ko-cevski D.D., Rizzi L., Peel A., 2008, ApJ, 676, 184Wang J., et al., 2017, MNRAS, 472, 3029Whiting M., Humphreys B., 2012, PASA, 29, 371Wolf C., et al., 2018, PASA, 35, 10Wright E.L., et al., 2010, AJ, 140, 1868Yang X., Mo H.J., van den Bosch F.C., Pasquali A., Li C., BardenM., 2007, ApJ, 671, 153
APPENDIX: H I IN SURROUNDING GROUPMEMBERS
Within the 12-beam image cube, aside from the sources di-rectly associated with the NGC 7232/3 triplet and surround-ing tidal debris, there are nine resolved H i galaxies. All ofthese galaxies appear to have stellar counterparts and forthose that lie above the HIPASS detection levels, ASKAPrecovers similar amounts of flux as HIPASS (see the rightcolumn of Figure A1). The new H i detections appear tohave fairly smooth velocity gradients that follow the mor-phology of the assumed stellar counterpart (Figure A2). Theindividual characteristics of these nine galaxies are furtherdiscussed in this appendix. IC 5171
IC 5171 is a bright spiral galaxy that has a high amountof associated H i emission. The residual sidelobe artefactsaround IC 5171, shown in Figure 4, are particularly promi-nent as this galaxy resides in the higher sensitivity regionof the H i cube. To highlight the actual H i features of thegalaxy, these artefacts have been manually masked out ofthe moment maps in Figure A1. IC 5171 has a higher cen-tral velocity than other group members indicating that it islikely in the outskirts of the galaxy group. Overall, IC 5171looks fairly symmetrical and undisturbed by the group en-vironment. ESO 288-G049
The southwest portion of ESO 288-G049 appears to have adisturbed velocity gradient. There is also an apparent gascloud that was picked up by SoFiA as being part of thisgalaxy. If the cloud is real (i.e. not a sidelobe / imagingartefact resulting from the preliminary nature of the obser-vations and data processing methods used during the EarlyScience phases of ASKAP), it could be another H i sourcethat is interacting with ESO 288-G049. However, the stellardisk of this face-on spiral and the spectral profile both showno noticeable asymmetries. ESO 289-G010
ESO 289-G010 is an edge-on spiral located at the edge of thefield of view of the 12-beam footprint. ASKAP recovers the
MNRAS , 1–16 (2019) K. Lee-Waddell et al. same amount of H i flux for this galaxy as the HIPASS obser-vations. Nevertheless, the ripple pattern – the same artifactthat was previously noticed in the beam edge region for somenights of the footprint B observations – is quite evident inthe spectrum for this source. Although there appears to bean H i extension to the north of the disk of ESO 289-G010,the RMS noise in this region of the cube is higher than thecentral regions. As such, it is unclear if this feature is real.We only processed and imaged a subset of beams for thispaper, the addition of other beams from the same datasetwould result in better H i maps for further investigation ofthis galaxy. ESO 289-G011
ESO 289-G011 is optically classified as an irregular galaxy.Its H i velocity pattern indicates well-behaved rotation and avelocity width that is common for a typical gas-rich dwarf.The asymmetries in the outer regions of the galaxy suggestthat there is a possible warp in the H i disk. MRSS 288-021830, AM 2208-460 and MRSS 289-168885
Our ASKAP observations are the first to detect and resolvethe H i components associated with these sources. With thehigh angular and spectral resolution of ASKAP, it is possibleto associate these H i detections with likely stellar counter-parts and place constraints on redshift distances. All threegalaxies appear to be members of the NGC 7232 galaxygroup; however, since their H i is spatially unresolved, it isdifficult to ascertain their dynamical properties or any de-tails about how they are being affected by the group envi-ronment. NGC 7232A
NGC 7232A is a nearly edge-on disk galaxy. In addition tothe the spatial and spectral coincidence between the stellarand H i emission, the H i velocity gradient follows the majoraxis of the stellar disk of NGC 7232A across nearly 200 kms − , confirming that these components belong to the samegalaxy. The spectra for this galaxy appear to be quite noisy,for both ASKAP and HIPASS, but three peaks of emissioncan be discerned in the ASKAP spectra and image cube.There is a noticeable feature between ∼ − in the HIPASS spectrum that is not present in the ASKAPspectrum. This ∼ i clouds C1 and C2. If this featureis diffuse H i emission, it could be part of the tidal bridgebetween the NGC 7232/3 triplet and IC 5181. ESO 289-G005
Similar to NGC 7232A, the H i emission detected in the samespatial and spectral region of ESO 289-G005 has a velocitygradient across the extent of its stellar disk. However, re-flecting the compact morphology of this galaxy, the H i spans ∼
100 km s − . This detection is the most prominent of thefive newly resolved H i galaxies. This paper has been typeset from a TEX/L A TEX file prepared bythe author. MNRAS000
100 km s − . This detection is the most prominent of thefive newly resolved H i galaxies. This paper has been typeset from a TEX/L A TEX file prepared bythe author. MNRAS000 , 1–16 (2019) he NGC 7232 galaxy group Figure 1.
ASKAP H i detections of galaxies that can be cross-matched to individual HIPASS sources. Stellar counterparts are labelledand each row represents one source. For each galaxy we show, left column: the ASKAP H i integrated intensity (moment 0) contours inwhite – at (1, 3, 6) × atoms cm − – superimposed on DSS2 Blue archival images; middle column: ASKAP H i velocity (moment 1)maps; right column: the ASKAP (solid blue) and HIPASS H i spectra. Please note, the residual sidelobe artefacts around IC 5171 havebeen manually masked out of the moment maps.MNRAS , 1–16 (2019) K. Lee-Waddell et al.
Figure 2.
ASKAP H i detections of galaxies that were hidden within the noise of HIPASS. Likely stellar counterparts are labelled andeach row represents one source. For each galaxy we show, left column: the ASKAP H i moment 0 contours in white – at (1, 3, 6) × atoms cm − – superimposed on DSS2 Blue archival images; middle column: ASKAP H i moment 1 maps; right column: the ASKAP (solidblue) and HIPASS (dotted red) H i spectra. The HIPASS spectra were extracted using a single pixel. Vertical lines indicate the ASKAPH i velocity ranges of the galaxies. MNRAS000