A blind ATCA HI survey of the Fornax galaxy cluster: properties of the HI detections
A. Loni, P. Serra, D. Kleiner, L. Cortese, B. Catinella, B. Koribalski, T. H. Jarrett, D. Cs. Molnar, T. A. Davis, E. Iodice, K. Lee-Waddell, F. Loi, F. M. Maccagni, R. Peletier, A. Popping, M. Ramatsoku, M. W .L. Smith, N. Zabel
AAstronomy & Astrophysics manuscript no. aanda © ESO 2021February 3, 2021
A blind ATCA H I survey of the Fornax galaxy cluster Properties of the H I detections A. Loni , , P. Serra , D. Kleiner , L. Cortese , , B. Catinella , , B. Koribalski , T. H. Jarrett , D. Cs. Molnar ,T. A. Davis , E. Iodice , K. Lee-Waddell , F. Loi , F. M. Maccagni , R. Peletier , A. Popping , , M. Ramatsoku , ,M. W .L. Smith , and N. Zabel INAF - Osservatorio Astronomico di Cagliari, Via della Scienza 5, 09047, Selargius, CA, Italye-mail: [email protected] Dipartimento di Fisica, Università di Cagliari, Cittadella Universitaria, 09042, Monserrato, Italy International Centre for Radio Astronomy Research (ICRAR), The University of Western Australia, 35 Stirling Hwy, Crawley,WA 6009, Australia ARC Centre of Excellence for All-Sky Astrophysics in 3 Dimensions (ASTRO3D) CSIRO Astronomy and Space Science, Australia Telescope National Facility PO Box 76, Epping, NSW 1710, Australia Department of Astronomy, University of Cape Town, Private Bag X3, Rondebosch 7701, South Africa School of Physics and Astronomy, Cardi ff University, Queens Buildings The Parade, Cardi ff CF24 3AA, UK INAF-Astronomical observatory of Capodimonte, via Moiariello 16, Naples 80131, Italy Kapteyn Astronomical Institute, University of Groningen, PO Box 800, 9700 AV Groningen, The Netherlands Australian Research Council, Centre of Excellence for All-sky Astrophysics (CAASTRO), Australia Department of Physics and Electronics, Rhodes University, PO Box 94, Makhanda, 6140, South AfricaReceived October 30, 2020; accepted January 26, 2021
ABSTRACT
We present the first interferometric blind H I survey of the Fornax galaxy cluster, which covers an area of 15 deg out to the clustervirial radius. The survey has a spatial and velocity resolution of 67 (cid:48)(cid:48) × (cid:48)(cid:48) ( ∼ × − and a 3 σ sensitivity of N H I ∼ × cm − and M H I ∼ × M (cid:12) , respectively. We detect 16 galaxies out of roughly200 spectroscopically confirmed Fornax cluster members. The detections cover about three orders of magnitude in H I mass, from8 × to 1 . × M (cid:12) . They avoid the central, virialised region of the cluster both on the sky and in projected phase-space, showingthat they are recent arrivals and that, in Fornax, H I is lost within a crossing time, ∼ I morphology, including several cases of asymmetries, tails, o ff sets between H I and optical centres, and a case of a truncated H I disc. This suggests that these recent arrivals have been interacting with other galaxies, the large-scale potential or the intergalacticmedium, within or on their way to Fornax. As a whole, our Fornax H I detections are H I -poorer and form stars at a lower rate thannon-cluster galaxies in the same M (cid:63) range. This is particularly evident at M (cid:63) (cid:46) M (cid:12) , indicating that low mass galaxies are morestrongly a ff ected throughout their infall towards the cluster. The M H I / M (cid:63) ratio of Fornax galaxies is comparable to that in the Virgocluster. At fixed M (cid:63) , our H I detections follow the non-cluster relation between M H I and the star formation rate, and we argue that thisimplies that thus far they have lost their H I on a timescale (cid:38) − I removal is likely to proceed faster,as confirmed by a population of H I -undetected but H . ex -detected star-forming galaxies. Overall, based on ALMA data, we find a largescatter in H . ex -to-H I mass ratio, with several galaxies showing an unusually high ratio that is probably caused by faster H I removal.Finally, we identify an H I -rich subgroup of possible interacting galaxies dominated by NGC 1365, where pre-processing is likey tohave taken place. Key words. galaxies: cluster / galaxies: evolution / galaxies: ISM
1. Introduction
It is known that the evolution of galaxies is faster in denser en-vironments (Diaferio et al. 2001) and that, as a consequence,the relative abundance of red early-type galaxies and blue late-type galaxies changes with environment density (Hubble & Hu-mason 1931; Oemler 1974; Dressler 1980). In this context,galaxy clusters are the most extreme environments within whichgalaxies evolve. They are characterised by a high number den-sity of galaxies and by the presence of a dense intra-clustermedium (ICM), which galaxies move through. Thus, in clusters,both hydrodynamical and gravitational interactions such as ram-pressure stripping and tidal interactions, respectively, are likelyto happen (Gunn & Gott 1972; Toomre & Toomre 1972), as well as mergers between galaxies (at least in the cluster outskirts -Sheen et al. 2012; Oh et al. 2018). These interactions depletethe cold gas reservoirs of galaxies and, therefore, a ff ect their starformation activity. The balance between these types of interac-tions depends both on galaxy properties – including their orbits– and on cluster properties, such as the number density of galax-ies and the ICM density. Therefore, we expect galaxy evolutionto proceed di ff erently in di ff erent clusters.The aim of this work is to study galaxy evolution in theFornax cluster, which is a low mass cluster (7 × M (cid:12) withina radius of 1.4 Mpc - twice the virial radius, R vir ; Drinkwateret al. 2001a), located in the southern sky at a distance of 20 Mpc(Blakeslee et al. 2009). Fornax is thus the second closest galaxy Article number, page 1 of 19 a r X i v : . [ a s t r o - ph . GA ] F e b & A proofs: manuscript no. aanda cluster to us after Virgo. Fornax has roughly 200 spectroscop-ically confirmed galaxies within R vir ∼
700 kpc (Maddox et al.2019). NGC 1399 is the central dominant galaxy, which coin-cides with the peak of the X-ray emission (Paolillo et al. 2002),and the majority of the massive galaxies in the cluster centralregion are of an early morphological type. Thus, Fornax showsa more dynamically evolved state than Virgo (Grillmair et al.1994; Jordán et al. 2007), but its growth is not over at all.The Fornax region includes two main substructures(Drinkwater et al. 2001a). The first is the cluster itself cen-tred on NGC 1399, whose interaction with NGC 1404 is re-vealed by a perturbed ICM distribution (e.g. Sheardown et al.2018). Here, three well defined groups of galaxies with di ff er-ent light and colour distributions, kinematics, and stellar popula-tions were found by combining the Fornax Deep Survey imaging(Iodice et al. 2019b) with Fornax 3D spectroscopy (Iodice et al.2019a) in the two-dimensional projected phase space: the core,the north-south clump and the infalling galaxies (see Fig. 7 inIodice et al. 2019a). The core is still dominated by NGC 1399,which is one of only two slow-rotators inside the virial radius(NGC 1427 is the other one, located on the east side of the clus-ter). The NS-clump is located in the high-density region of thecluster (within 0 . R vir ∼ . ff use light, globularclusters and planetary nebulae) (Cantiello et al. 2018; Spinielloet al. 2018; Iodice et al. 2019b) are found in this NS clump,where galaxy growth is still ongoing through the accretion ofmass onto galaxies’ outer regions (Spavone et al. 2020). Thethird group of objects in the cluster includes the infalling galax-ies, which are distributed nearly symmetrically around the corein the low-density region outside ∼ . R vir ∼ . / or minor mergers in the form of tidal tails anddisturbed molecular gas (Zabel et al. 2019; Raj et al. 2019). Pre-vious works had also shown that numerous dwarf galaxies arecurrently falling into the centre of the cluster (Drinkwater et al.2001b; Schröder et al. 2001; Waugh et al. 2002). Finally, theother major structure in the Fornax volume is the infalling groupcentred on NGC 1316 (Fornax A), located ∼ I ) gives us information onthe evolutionary state of galaxies as well as a global picture ofthe cluster. The evolution of galaxies depends on the evolutionof their atomic hydrogen reservoir, which is the primary reser-voir of fuel for star formation. Since it typically extends to theoutskirts of galaxies, H I gas is the first component that is af-fected by tidal interactions, ram-pressure stripping and mergers.Indeed, cluster galaxies are usually deficient in atomic hydro-gen with respect to non-cluster galaxies (Giovanelli & Haynes1983; Haynes & Giovanelli 1986; Boselli & Gavazzi 2006). H I is, therefore, a crucial observable for understanding galaxy evo-lution in dense environments (Hughes & Cortese 2009; Chunget al. 2009).H I emission in the Fornax cluster was studied in severalworks. Bureau et al. (1996) used the Parkes radio telescope tomeasure the amount of atomic hydrogen in 21 undisturbed galax-ies with morphologies S0 / a or later, located within 6 deg from NGC 1399 and with c z ≤ − . The eight galaxies within R vir did not show any peculiar value in the M H I / I − band infraredluminosity ratio with respect to the rest of the sample. The sameratio evaluated for Ursa Major galaxies, a lower density environ-ment than Fornax, led them to conclude that the observed Fornaxgalaxies are not H I deficient.The first blind H I survey of Fornax was carried out by Barneset al. (1997), also with the Parkes telescope. They detected H I ineight galaxies within an area of 8 × . Of those, two galax-ies are within the cluster R vir : ESO 358G-063 and NGC 1365.Thanks to the survey sensitivity, they excluded the existence of asignificant population of optically undetected H I clouds with H I mass greater than 10 M (cid:12) .The number of H I detections within the Fornax central re-gion increased with the targeted survey by Schröder et al. (2001).They found H I in 37 out of 66 galaxies. Of those, 14 are within R vir , while the rest lies within 5 deg from the centre of the cluster.They found a lack of H I in the cluster centre, and measured theH I deficiency parameter to be 0 . ± .
09 (as defined in Solaneset al. 1996). This shows a modest H I depletion in Fornax (usu-ally, galaxies with H I deficiency parameter > I deficient, e.g. Dressler 1986; Solanes et al. 2001). Fur-thermore, the M H I − to − blue light ratio of Fornax galaxies - meanvalue (0 . ± .
15) M (cid:12) / L (cid:12) - is a factor 1.7 lower than in thefield. They also showed that the velocity dispersion of the sam-ple of H I deficient galaxies is lower than that of the remaining H I detected galaxies. This di ff erence in velocity dispersion agreeswith the deficient galaxies having more radial orbits (as shownin Dressler 1986), which makes them good candidates for ram-pressure stripping (as discussed in Solanes et al. 2001).Waugh et al. (2002) presented a blind survey of Fornax basedon the H I Parkes All Sky Survey (HIPASS − Barnes et al. 2001)data with a M H I limit of 1.4 × M (cid:12) . They detected 110 galax-ies within an area of ∼
620 deg around the cluster. Of those, nineH I detections are within R vir . The authors confirmed a H I deple-tion in the H I detections (all late types) near the centre of thecluster, and suggested that H I -rich galaxies detected in the outerparts of the cluster are infalling towards the cluster for the firsttime. H I detections are arranged in a large scale sheet-like struc-ture with a negative velocity gradient from south-east to north-west.Waugh (2005) presented the result of the deepest blind H I survey of the cluster so far: the Basketweave survey of For-nax carried out with the Parkes telescope. This survey used anew scanning technique, which improved the sampling and thenoise level compared to HIPASS. H I was detected in 53 galaxieswithin 100 deg down to a detection limit of 10 M (cid:12) . Of those,15 were new detections, which confirmed the results of Waughet al. (2002). In addition, the author presented higher resolutionH I observations (2 arcmin, 3.3 km s − ) of 28 individual Fornaxgalaxies carried out with the Australia Telescope Compact Ar-ray (ATCA). Of those, six are within R vir . Within this region,ESO LV-3580611 was marked as having an intriguing H I mor-phology with an elongation to the north-east of the system. Assuggested in Schröder et al. (2001), also Waugh (2005) pointedout that this galaxy may be moving towards us.The only other Fornax galaxy within R vir with resolved H I imaging is NGC 1365 (van der Hulst et al. 1983; Jorsater & vanMoorsel 1995). The latter study observed an elongated H I distri-bution to the west of the system. They suggested that it may becaused by the interaction with the ICM.All previous blind H I surveys of Fornax were carried out withthe Parkes telescope. Their angular resolution of 15 arcmin wasinsu ffi cient to study the H I morphology of the detected galaxies, Article number, page 2 of 19. Loni et al.: A blind ATCA H I survey of the Fornax galaxy cluster Fig. 1: Layout and detections of our ATCA H I survey. The green outline includes the 15-deg region where the average noise is2.8 mJy beam − (see section 2). The red contours represent the lowest reliable H I column density – 3 σ over 25 km s − – of our16 detections. The grey dashed circle is R vir . The background optical image comes from the Digital Sky Survey (blue band). Theyellow contours show the X-ray emission in the Fornax cluster detected with XMM-Newton (Frank et al. 2013) and convolved witha 3 arcmin FWHM gaussian kernel. These contours are spaced by a factor of 2, with the lowest level at 3.7 counts deg − s − .which is a powerful tracer of environmental e ff ects. Better angu-lar resolution was achieved with the interferometric observationof a few selected galaxies in Fornax (see references above), butthose observations covered only a small portion of the clustervolume. Here we present the first, blind, interferometric H I sur-vey of the Fornax cluster.Our survey was carried out with the ATCA and covers anarea of 15 deg centred on NGC 1399 with a spatial and velocityresolution of 67 (cid:48)(cid:48) × (cid:48)(cid:48) ( ∼ × − , respectively. The average column density sensi-tivity within the survey area is 2 × cm − (3 σ over 25 km s − ) and the M H I sensitivity is 2 × M (cid:12) (3 σ over 100 km s − ). Ini-tially, a case study by Lee-Waddell et al. (2018) revealed thetidal origin of NGC 1427A using spatially resolved H I imagesfrom a subregion of our ATCA mosaic. Here we present the re-sults of the full survey. In Sect.2 we describe observations anddata reduction. In Sect.3 we present the H I detections, their H I images and spectra. We also compare their H I mass, H . ex massand SFR relative to non-cluster control samples, and comparetheir spatial and velocity distribution with those of the generalFornax population. In Sect. 4 we discuss our results, which we Article number, page 3 of 19 & A proofs: manuscript no. aanda
Fig. 2: ATCA H I contours overlaid on an optical image for all our H I detections, sorted according to increasing H I mass. The g -bandoptical images come from the Fornax Deep Survey (Iodice et al. 2016; Venhola et al. 2018; Peletier et al. 2020) for all galaxiesexcept FCC 323, whose g -band optical image comes from the DESI Legacy Imaging Surveys, DR8 release, (Dey et al. 2019). Ineach panel we show the 3 σ column density sensitivity - values reported in the top right corner - with white colour, while cyancontours represent steps of 3 n from it (n =
0, 1, 2, ...). We show the PSF on the bottom-left corner of each panel, and a 5 kpc scalebar in the bottom-right corner.then summarise in Sect.5. We include supplementary material oneach galaxy in Appendix A.
2. ATCA observations and data reduction
Our blind Fornax survey covers an area of 15 deg (defined at asensitivity level 3 × higher than in the best region of the H I cube),spanning from the centre of the cluster to a distance slightly fur-ther than R vir . The observations were carried out with the ATCAin the 750B configuration, from December 2013 to January 2014(project code C2894) . The cluster was observed for 336 hrs us-ing 756 di ff erent pointings with a spacing of 8.6 arcmin (1 / Data available on https://atoa.atnf.csiro.au/query.jsp
The 64 MHz bandwidth, centred at 1396 MHz, was di-vided into 2048 channels, providing a velocity resolution of6.6 km s − . We reduced the data using the MIRIAD software(Sault et al. 1995). PKS B1934-638 and PKS 0332-403 werechosen as the bandpass calibrator and the phase calibrator, re-spectively. The latter was observed at 1.5 hr intervals betweenon-source scans. We flagged strong radio frequency interfer-ence based on Stokes V visibilities. After flagging and calibra-tion we further processed a restricted frequency range 1407.2MHz - 1419.6 (which corresponds to the velocity range of 166 -2783 km s − ), which includes all spectroscopically confirmedFornax galaxies (Maddox et al. 2019). Within this range we usedthe UVLIN task of
MIRIAD to fit and subtract continuum emissionusing 2nd-order polynomials.
Article number, page 4 of 19. Loni et al.: A blind ATCA H I survey of the Fornax galaxy cluster Fig. 3: Integrated H I spectra of our Fornax H I detections (in red) sorted according to increasing H I mass as in Fig. 2. We compareour spectra to spectra from the literature (shown in black; see top-left corner for details). We also show the barycentric velocityobtained from our H I spectra (vertical red line) and from optical spectra (vertical blue line; Maddox et al. 2019).We obtained the dirty cube with the INVERT task using natu-ral weights to maximise surface-brightness sensitivity. We used
MOSMEM and
RESTOR to clean and restore H I emission, respec-tively. The restoring Gaussian PSF has a major and minor axisFWHM of 95 and 67 arcsec, respectively, and a position angleof 0.4 deg. The root mean square (RMS) noise level of the finalcube goes down to 2.0 mJy beam − in the most sensitive region.Within the survey area the RMS noise is ≤ . − and,on average, 2.8 mJy beam − . This corresponds to a 3 σ H I col-umn density sensitivity of N H I ∼ × cm − assuming a linewidth of 25 km s − and a 3 σ M H I sensitivity of ∼ × M (cid:12) overa linewidth of 100 km s − . We searched for H I sources with theSoFiA source-finding package (Serra et al. 2015) within the sur-vey footprint (see the green outline in Fig. 1). By smoothing andclipping, we convolve the input cube with a set of kernels and de-tect emission above 3.5 σ of the local noise level of each kernel.Reliable detections are identified based on the reliability algo-rithm presented in (Serra et al. 2012), which assumes that truesources have positive total flux and that the noise is symmetricaround 0. For one faint source, NGC 1436, visual inspection wasnecessary to improve the SoFiA detection mask used to estimategalaxy parameters. Lastly, our final list of H I detections includes a faint source which did not pass the reliability test, but whichwe consider a genuine detection given its spatial correspondencewith the known optical source FCC 323.
3. H I detections in the Fornax cluster I detection properties We detect H I in the 16 galaxies listed in Table 1. Of these, threeare new H I detections: FCC 090, FCC 102, FCC 323. The last isthe only galaxy with no previous redshift measurement. In Fig. 1we show the location of our H I detections on the sky and in Fig. 2we show the H I morphology of each galaxy. In these figures, redand white contours, respectively, represent the lowest reliableH I column density contour, defined as 3 times the local RMSassuming a typical H I linewidth of ∼
25 km s − (see Table 1 andtop-right corner of each panel in Fig. 2).Fig. 3 shows the integrated ATCA H I spectra of our 16 detec-tions. We calculated the error bars by summing in quadrature thestatistical uncertainty − derived from the local RMS (Table 1)and the number of independent pixels detected in each channel − and the flux-scale uncertainty. We find the latter to be ± Article number, page 5 of 19 & A proofs: manuscript no. aanda point sources between our radio continuum image and the North-ern VLA Sky Survey (Condon et al. 1998).We compare our spectra to those obtained from previous ob-servations. In particular, we use HIPASS spectra from the BGCcatalogue (Koribalski et al. 2004) or HIPASS data reprocessedby us. Since HIPASS data do not show any emission at the posi-tion of NGC 1436 and ESO 358-G016, we used Green BankTelescope data (GBT - Courtois et al. 2009) and Parkes data(Bureau et al. 1996) as comparison, respectively. On the otherhand, HIPASS spectra of ESO 358-G015 and ESO 358-G051are noisy, so we used comparison spectra from Matthews et al.(1998) and Theureau et al. (1998), respectively, based on Nanacydata. Furthermore the literature spectra of ESO 358-G015 andESO 358-G016 were rebinned to our ATCA channel-width. Forconsistency, we calculated the uncertainties in the comparisonspectra by combining the noise in the spectrum and the flux-scaleuncertainty of each survey, except for ESO 358-G016 for whichthe flux-scale uncertainty was not provided. For this galaxy, theerror bars are shown as the RMS of the spectrum. In each panelof Fig. 3 we also show the velocity v opt derived from opticalspectroscopy (Maddox et al. 2019) and the barycentric H I veloc-ity v HI derived from our ATCA spectra.We estimated the H I mass ( M H I ) of our detections from theintegrated H I flux using eq.50 in Meyer et al. (2017) and adopt-ing the same distance of 20 Mpc for all galaxies (Mould et al.2000). The uncertainty on M H I is obtained from the error barsof the spectrum (Fig. 3). We report H I fluxes and masses in Ta-ble 1. Our H I fluxes agree with those in the literature within 1 σ for nine out of 13 galaxies, and within 2 σ for 11 of them. Thetwo cases with a discrepancy larger than 2.5 σ are NGC 1436and ESO 358-G051. The comparison spectrum for NGC 1436comes from GBT data (Courtois et al. 2009) and shows that weare most likely missing H I flux from the blue-shifted part of thesystem. For this galaxy, the total H I flux recovered by ATCA islower than the GBT flux by 2.5 σ (corresponding to a factor of2.8). The reason of this discrepancy arises from a combinationof low S / N and the presence, in at least some of the blue-shiftedchannels, of artefacts in this part of the ATCA cube. The case ofESO 358-G051 is less clear since the ATCA cube does not showany obvious artefacts and the galaxy well detected. However thetotal H I flux recovered by ATCA is lower than the Nancay fluxby 3 σ . Based on the current data it is possible that some of theemission is spread over multiple ATCA beams and therefore istoo weak to be detected. Future MeerKAT data will clarify thisissue Serra et al. (2016). Finally, the H I mass of FCC 323 isbelow the typical sensitivity of our data quoted in Sec. 2 be-cause of the narrow linewidth as well as the low value of thelocal noise (Table 1). Our detections cover about three of mag-nitude in M H I , from FCC 323 ( M H I = × M (cid:12) ) to NGC 1365( M H I = . × M (cid:12) ). Fig. 4 shows the cumulative histogramof the H I masses of our sample. A future study will analyse theH I mass function inferred from our data.Due to the improved resolution of ATCA over a single dish,we detected, in half of the sample, a variety of H I morphologies(see Fig. 2) including o ff sets between optical and H I centres,truncated discs, asymmetries and H I tails, which we describein this section and in Appendix A, following the same order inwhich galaxies are shown in Fig. 2 (from the lowest to the high-est H I mass): the H I distribution in FCC 102 is o ff set with respectto the optical centre towards the north. We also notice a largedi ff erence between v opt and v H I , ∼
100 km s − . However, this isconsistent with the large uncertainty on v opt given that the lat-ter was measured from absorption lines for this galaxy (NatashaMaddox, priv. comm.); the H I peak of FCC 090 corresponds to Fig. 4: Cumulative histogram of the ATCA H I masses of ourFornax sample.Fig. 5: H I disc size as a function of ATCA M H I for our resolvedH I detections. Dashed line and the blue shaded area show thescaling relation and 3 σ scatter respectively in (Wang et al. 2016).the optical centre but the H I distribution has an elongation to thesouth; H I in ESO-LV 3580611 (FCC 306) is more extended tothe north although there is no o ff set between optical and H I cen-tre (Waugh 2005); the H I morphology of NGC 1437B (FCC 308)is asymmetric and more extended to the south; H I in NGC 1351A(FCC 067) shows an elongation towards the south; the H I ofESO 358-G063 (FCC 312) is more extended to the east side ofthe disc and the H I contours appear to be compressed on the westside; In NGC 1427A (FCC 235), we detected a H I tail whichpoints to the south-east, away from the cluster centre, consis-tent with the tidal origin of this galaxy discussed in Lee-Waddellet al. (2018); the H I distribution in NGC 1365 (FCC 121) is ex-tended to the north and appears to be compressed in the south-west part of the disc (see also Jorsater & van Moorsel 1995). Wemark these H I disturbed galaxies with star markers in all subse-quent figures.The remaining half of the galaxies are either unresolved (ornearly so) and centred on the stellar body, or do not show no-ticeable asymmetries. One of them hosts a H I disc unusually Article number, page 6 of 19. Loni et al.: A blind ATCA H I survey of the Fornax galaxy cluster Fig. 6: M H I to M (cid:63) ratio as a function of M (cid:63) . Fornax galaxies(red + blue markers) are compared with non-cluster galaxies fromVSG + HRS (grey circles). Red and blue colours show FornaxH I deficient and normal galaxies, respectively. We show Fornaxgalaxies with a distorted H I morphology with star-shaped mark-ers (see Sect- 3). We show with a black solid line the xGASSscaling relation. The black dashed line is the linear extrapola-tion of this trend for M (cid:63) < × M (cid:12) . The orange dashed lineshows the ATCA average sensitivity evaluated as 3 × − with a linewidth of 100 km s − . The vertical dashed lineat 3 × M (cid:12) separates low- from high-mass galaxies.truncated within the stellar disc, NGC 1436. This case will bediscussed in detail in later sections.A final case worth commenting on is FCC 120. This galaxyhas been detected with Parkes and the spectrum in Schröder et al.(2001) shows an evident double horn profile that is ∼
100 km s − wide, while we detect a single-peak profile (within the uncertain-ties - see Fig. 3). Careful visual inspection of our H I cube did notreveal any H I emission missing from our SoFiA detection mask.The regular H I morphology of this galaxy in Fig. 2 and the goodagreement between v H I and v opt in Fig. 3 suggest that our H I char-acterisation of this galaxy is correct. Future, deeper data from theMeerkat Fornax survey may confirm this (Serra et al. 2016).We estimated the H I size for all our resolved H I detectionsusing the method of Wang et al. 2016 from the H I intensity mapsof our galaxies. We considered a galaxy resolved if its surfacebrightness profile deviates from the shape of the point-spreadfunction. We measured the HI diameter where the surface den-sity is 1 M (cid:12) pc − and is then deconvolved with the H I beam. InFig. 5 we show that both disturbed and regular galaxies followthe H I size-mass scaling relation of Wang et al. (2016) within the3 σ scatter. This can be understood as asymmetric H I featuresusually have a low surface brightness and do not significantlycontribute to the total M H I of galaxies. In Fig.5 we also show theH I unresolved galaxies. The upper limit on their size were setequal to the ATCA beam minor axis. Fig. 7: Expected M H I versus measured M H I . The former wereevaluated with the Haynes & Giovanelli’s (1984) method withthe coe ffi cient summarised by Boselli & Gavazzi (2009). Thelatter are the ATCA M H I calculated as described in Sect.3.1. Thedashed line, a factor of 3 below the 1:1 relation, shows the typicalthreshold below which galaxies are considered deficient. M H I to M (cid:63) ratio Given the abundant (albeit subtle) H I asymmetries and o ff setsin our Fornax sample, which might be tracing environmental in-teractions within or on their way to the Fornax cluster (whetherwith other galaxies, the large-scale potential or the intergalacticmedium), we study the amount of H I in these galaxies in searchof signs of H I depletion. In order to do so, we evaluated the ratiobetween M H I and stellar mass ( M (cid:63) ) for each galaxy in our sam-ple. The M (cid:63) values are derived using the WISE W1 (3.4 µ m)in − band luminosity, W1 − W2 colour and the prescription givenby Cluver et al. (2014a), with the custom photometry further de-fined in Jarrett et al. (2019), and assuming a common distance of20 Mpc for all galaxies. We show the distribution of our sampleon the M H I / M (cid:63) − vs − M (cid:63) plane in Fig. 6.We compare galaxies in Fornax with a sample consist-ing of void galaxies from the Void Galaxy Survey (VGS -Kreckel et al. 2012) and field galaxies from the Herschel Ref-erence Survey (HRS - Boselli et al. 2014). We also usedthe xGASS M (cid:63) - M H I / M (cid:63) scaling relation, which shows theweighted median of log ( M H I / M (cid:63) ) as a function of M (cid:63) . Thiswas obtained from 1177 galaxies selected only by stellar mass( M (cid:63) = –10 . M (cid:12) ) and redshift (0.01 < z < M (cid:12) (black dashed line) and show that it is consistent withthe non-cluster sample of galaxies.Our sample of Fornax H I detections appears to be system-atically o ff set with respect to our VGS + HRS comparison sam-ple and to the xGASS scaling relation. A two-sample Kol-mogorov–Smirnov test on the distribution of o ff sets from thexGASS scaling relation rejects the null hypothesis that Fornaxand VGS + HRS galaxies are drawn from the same parent sample(p-value = = ff set from the xGASS scaling re- Article number, page 7 of 19 & A proofs: manuscript no. aanda lation is larger than the RMS deviation of VGS + HRS galaxiesfrom it, thus indicating H I deficiency. For illustrative purposeswe henceforth label these HI deficient galaxies in red colours inFig. 6 and in all other upcoming figures in this paper.H I deficiencies measured from plots like our Fig. 6 shouldbe taken with caution because of the large M H I scatter at at fixed M (cid:63) (e.g. Maddox et al. 2015). This scatter is to first order drivenby Hubble type or, equivalently, by galaxy properties that cor-relate with Hubble type, such as the star formation rate (SFR).For this reason, in Sect. 3.3 we further analyse how the H I defi-ciency at fixed M (cid:63) relates to the H . ex and SFR of these galaxies.Furthermore, in the present section we estimate H I deficienciesas proposed by Haynes & Giovanelli (1984) and recently revis-ited by Jones et al. (2018), where the measured M H I is comparedwith an expected M H I calculated based on galaxies’ optical sizeand Hubble type. For this purpose, we use the coe ffi cients sum-marised by Boselli & Gavazzi (2009) and originally given inHaynes & Giovanelli (1984), Solanes et al. (1996), Boselli &Gavazzi (2009), and adopt the optical sizes and Hubble typeslisted in Table 1 for our galaxies. Fig. 7 shows the comparisonbetween the expected and the measured M H I . The dashed line, afactor of 3 below the 1:1 relation, is the typical threshold belowwhich galaxies are considered H I deficient using the Haynes &Giovanelli’s (1984) method (e.g. Cortese et al. 2011). Also inthis case, Fornax galaxies appear o ff set towards lower H I massesand there is a good match between galaxies labelled as H I de-ficient based on our Fig. 6 and those below the dashed line inFig. 7.From these figures, we see that not all galaxies with a dis-turbed H I morphology are H I deficient. Thus, H I morphologi-cal disturbances, whatever their exact nature (e.g. tidal or hy-drodynamical, which is di ffi cult to establish with the currentdata), allow us to identify cases of environmental interactionsbefore a significant fraction of the cold interstellar medium is re-moved. Combining H I morphological information and M H I - M (cid:63) ratio may reveal likely new members of the cluster (we comeback to this point in Sect.4).Fornax galaxies with a disturbed H I morphology cover thefull M (cid:63) range. However, we measure a stronger H I depletion forlow mass galaxies: the average o ff set from the xGASS scalingrelation in Fig. 6 is − .
86 dex and − .
33 dex for M (cid:63) below andabove 3 × , respectively. We show the threshold of 3 × M (cid:12) with a vertical dashed blue line in Fig. 6.We already mentioned the problematic detection ofNGC 1436. Although we are probably missing some flux, it re-mains a deficient galaxy even if we estimate the H I mass from theGBT flux (Table 1) with a log ( M H I / M (cid:63) ) = − . ff setfrom the xGASS scaling relation of − .
93 dex.Fig. 6 shows also eight galaxies where we did not detect H I emission but H . ex was detected with with the Atacama Large Mil-limeter / submillimeter Array (ALMA) (Zabel et al. 2019). Forthese galaxies we calculated the M H I upper limit as 3 × the lo-cal noise (Table 1) of the cube and assuming the CO line widthof these galaxies estimated by the PV diagrams in Zabel et al.(2019). Some of these galaxies appear to have too little H I giventhe molecular gas content and star formation rate, as we discussin following sections. We show the optical morphology of thesegalaxies in Appendix B.Finally, Fig. 8 shows the comparison between Fornax andVirgo galaxies (from HRS - Boselli et al. 2014) belonging toVirgo clouds A, B, N, E and S (as defined by Gavazzi et al. 1999)with 10 (cid:46) M (cid:63) (cid:46) M (cid:12) . Here we show the Virgo cluster galax-ies and the average scaling relation obtained from them in thesame M (cid:63) range (Cortese et al. 2011). Since Virgo is populated by Fig. 8: We compare Fornax galaxies (same colour coding ofFig.6) with Virgo cluster galaxies from HRS (light purple cir-cles). Dark purple circles show the average scaling relation ob-tained from Virgo cluster galaxies (Cortese et al. 2011)Fig. 9: M H / M H I as a function of M (cid:63) . We use the same colourcoding as Fig. 6. Blue shadow shows 1 × σ scatter from thexGASS weighted average of log ( M H / M H I ). We show upperlimits with downward arrows. We show lower limits with up-ward arrows.H I poor galaxies with respect to field galaxies (Davies & Lewis1973; Chamaraux et al. 1980; Cayatte et al. 1994; Hughes &Cortese 2009; Chung et al. 2009), this scaling relation is shiftedtowards lower gas fractions with respect to xGASS (Fig. 6).Although Fornax and Virgo galaxies experience di ff erent clus-ter environments, the distribution of Fornax galaxies cover thewhole range of M H I / M (cid:63) of Virgo galaxies, reaching the samelevel of H I deficiency. Article number, page 8 of 19. Loni et al.: A blind ATCA H I survey of the Fornax galaxy cluster Fig. 10: SFR as a function of M (cid:63) . We use the same colour cod-ing as Fig. 6. We show upper limits in SFR with downward ar-rows. The dashed black line represents the SFR scaling relationin (Whitaker et al. 2012). I properties with H . ex and SFR The ratio between molecular and atomic gas mass ( M H / M H I )can be useful to identify anomalous galaxies where, for exam-ple, only the atomic phase is a ff ected by the environment or H I is not e ffi ciently converted to H . ex . We thus compared the atomicand molecular gas reservoirs of our detections (see Fig. 9). Forthis purpose, we used H . ex masses from Zabel et al. (2019) forall our H I detections except for FCC 323 (no molecular gas dataavailable). We also include in this analysis the eight galaxies de-tected with ALMA (Zabel et al. 2019) that are not detected in H I (upper limits in Fig. 6 and Fig. 8; see Sec. 3.2). We also scaledthe molecular upper limit from Zabel et al. (2019) to be con-sistent with a line width of 100 km s − . As a comparison, weused the xGASS M (cid:63) - M H / M H I scaling relation in Catinella et al.(2018), which describes the typical M H / M H I ratio as a functionof M (cid:63) at z = M (cid:63) > M (cid:12) , 55% of all H I detected galaxies are alsoH . ex detected, while this fraction drops to zero for M (cid:63) < M (cid:12) (likely because the lower metallicity of these objects makes COprogressively harder to detect). In the rest of this section wetherefore focus on the higher M (cid:63) range. Here, we see that abouthalf of the galaxies are consistent with the xGASS sample, whilemost of the remaining galaxies are above the xGASS scaling(see Fig. 9). In particular, galaxies with a distorted H I morphol-ogy are compatible with the xGASS trend with the exception ofNGC 1427A whose lack of molecular gas is puzzling (see Zabelet al. 2019). Although this galaxy has a normal (and large) H I mass for its stellar mass, no molecular gas was detected withALMA. The agreement with the scaling relation holds also forNGC 1437B, which is the only H I deficient galaxy with a dis-torted morphology in this range of M (cid:63) .The detection with the highest M H / M H I ratio is NGC 1436.Although we might be missing H I flux (Sect. 3.2), this mass ra-tio remains high even if we use the M H I value estimated fromthe GBT flux (with a M H / M H I = M H I / M (cid:63) deviation from thexGASS scaling relation (Fig. 6). We use the same colour codingas Fig. 6. We show upper limits in M H I with leftward arrows. Thehorizontal blue line represents no deviation from the SFR scalingrelation, while the vertical blue line represents no deviation fromthe M (cid:63) - M H I / M (cid:63) scaling relation.liar as the lower limit on their M H / M H I ratio is already an orderof magnitude above the xGASS scaling relation. High M H / M H I ratios have also been measured in Virgo galaxies (Cortese et al.2016).Given the broad distribution of M H / M H I ratio in Fornax, wefurther investigate whether their star formation rate follows stan-dard scaling with M (cid:63) and M H I . In particular, we are interested inunderstanding whether the general o ff set of Fornax galaxies to-wards low M H I / M (cid:63) ratios in Fig. 6 is associated with a decreaseof SFR at fixed M (cid:63) .We start by comparing Fornax galaxies with the same sampleof non-cluster galaxies used in Sect. 3.2. SFR of both samplesof Fornax galaxies and non-cluster galaxies are evaluated usingeq. 2 in Boquien et al. (2016). We set the scaling coe ffi cient forthe infrared 24 µ m band (W4) equal to 6.17. We adopted the cal-ibration factor for near UV 231 nm band (NUV) to be log C = -43.17 (Kennicutt & Evans 2012). NUV fluxes come from GCAT(Seibert et al. 2012) and IRSA catalogues (Leroy et al. 2019).Since the latter catalogue is more reliable for galaxies at z = /
212 non-clustergalaxies; 11 /
24 Fornax galaxies). In these cases, we calculatethe SFR from the W3 (12 µ m) luminosity – with stellar emis-sion subtracted (as for W4)– and the TIR-to-MIR relation usingeq. 4 in Cluver et al. (2017), where L µ m is the continuum- Article number, page 9 of 19 & A proofs: manuscript no. aanda
Fig. 12: Distribution of our Fornax H I detections on the sky (blueand red markers) compared to that of all Fornax galaxies in ourfootprint (grey circles; Maddox et al. 2019). The green and graycontours are the same as Fig. 1.subtracted spectral ( ν × L ν ) luminosity. Our conclusions belowdo not change if we calculate SFR from WISE data alone.Fig. 10 shows that the majority of the Fornax galaxies havea SFR below the values predicted by the scaling relation ofWhitaker et al. (2012). Furthermore, the di ff erence between theSFR of Fornax galaxies and that predicted from Whitaker et al.(2012) increases towards lower M (cid:63) . That is, the SFR- M (cid:63) rela-tion in Fornax is steeper than that of non-cluster galaxies. Thismight indicate a stronger SFR decrease in low-mass galaxies,similar to what observed for their H I reservoirs (as discussed inSect. 3.2) and their H . ex reservoirs (Zabel et al. 2019). Also whencompared to the HRS + VGS comparison sample, Fornax galax-ies appear to be o ff set towards lower SFR values, mirroring theresults in Fig. 6.Among the four galaxies with the highest M H / M H I ratio inFig. 9, NGC 1386 is the only one with a higher SFR than ex-pected. This is likely due to the presence of an AGN (Rodríguez-Ardila et al. 2017) which a ff ects the SFR measurement. Theother two H I undetected galaxies, NGC 1380 and NGC 1387,reside at the lower edge of the comparison sample. We noticethat NGC 1380 is also highly H . ex deficient in Zabel et al. (2019).In Fig. 11 we investigate whether, in Fornax, the low SFR(at fixed M (cid:63) ; Fig. 10) can be related to the H I deficiency (Fig.6). This figure plots the SFR and M H I / M (cid:63) deviations from therespective scaling relations (Fig. 10 and Fig. 6) against one an-other. We find that Fornax galaxies mostly populate the area ofthe plot of H I deficient galaxies with low SFR, following thetrend of non-cluster galaxies in the comparison sample. That is,in Fornax, a decrease in M H I is accompanied by a decrease inSFR just like in non-cluster galaxies. In Sec. 4 we present aninterpretation of this result.On top of this general trend we see a few outliers. Themost evident ones are ESO 358-G060 and the H I undetectedNGC 1386. While the high SFR in NGC 1386 might be an arte-fact caused by the presence of an AGN, the case of ESO 358-G060 is more complicated. We carefully inspected this galaxy’sWISE images and no SFR from W3 / W4 with old stellar pop-ulation subtraction was detectable. Therefore, we calculated itsSFR the from NUV flux alone using eq. 12 in Kennicutt & Evans Fig. 13: Phase space diagram of the Fornax cluster. Here, weused the same colour coding as Fig. 12. We show the causticcurves of the cluster with black dashed lines. Histograms: blackand gray colours represent H I detected galaxies and all Fornaxmembers, respectively. The area within the black triangle showsthe virialised area of the cluster(2012). Although ESO 358-G060 is H I rich, its SFR is lower thanthat any galaxy of the comparison sample with a similar M (cid:63) . Itis worth noting the cases of NGC 1436 which lies at the upperedge of the comparison sample of H I deficient galaxy (bottom-left quadrant of Fig. 11) and NGC 1427A which shows a lowSFR although its large H I reservoir. All these cases will be dis-cussed in Sect. 4. I properties as a function of 3D location in Fornaxcluster In Fig. 12, we compare the 2D distribution of our Fornax H I detections with that of all spectroscopic Fornax members (Mad-dox et al. 2019) included within our survey footprint. Keeping inmind that the area of the sky that we observed is not symmetric,most of the galaxies in our sample are located south of the cen-tre of the cluster, while only four out of 16 detections are locatednorth of it.Almost all our H I detections (88% of the sample) are lo-cated farther than 0.5 R vir in projection (the only exceptions be-ing NGC 1427A at 0.2 R vir , and ESO 358-G051 at 0.4 R vir ). Incontrast, only 33% of all Fornax galaxies in our footprint areoutside 0.5 R vir .The same e ff ect, with the addition of the kinematicalinformation, can be seen in the projected phase-space dia-gram shown in Fig. 13. We used a cluster systemic veloc-ity of v sys = − and a cluster velocity dispersion of σ cl =
318 km s − (Maddox et al. 2019). In order to drawcaustic curves we proceeded as in Ja ff é et al. (2015) using M vir = × M (cid:12) , R vir =
700 kpc (Drinkwater et al. 2001a)and a cluster halo concentration parameter equal to 6 (Navarroet al. 1996; the exact value of this parameter does not change theresult significantly). The inner triangle shows the virialised areawhere it is more likely to find old members of the Fornax cluster(Rhee et al. 2017). From the right histogram in Fig. 13, we see
Article number, page 10 of 19. Loni et al.: A blind ATCA H I survey of the Fornax galaxy cluster that the spectroscopic Fornax members have a peak at the clustersystemic velocity, while velocities of our H I detections cover allvelocity range without any preferred velocity. Thus, both Fig. 12and Fig. 13 show that the H I galaxy sample does not follow thedistribution of spectroscopic galaxies.In Fig. 13, we also see that 11 out of 16 galaxies are withinthe escape velocity boundary in projection (black lines). It isworth noting that the H I deficient galaxies FCC 120 and ESO-LV 3580611 are located outside the caustic curves. Among H I disturbed galaxies, six out of eight are redshifted with respect tothe recessional velocity of the cluster.Both Fig. 12 and Fig. 13 show that some galaxies in oursample are not just close by one another on the sky but alsoin velocity. The clearest case is that of NGC 1365 and its threeneighbours: FCC 102, FCC 090 and ESO 358-G016. They are allwithin a region of 250 kpc and a velocity range of 200 km s − .These galaxies might be part of a substructure which is accret-ing onto the cluster (as suggested by Drinkwater et al. 2001a). Inaddition, most galaxies in this substructure exhibit indications ofongoing interactions with one another and / or with the intergalac-tic medium. Both H I elongations in FCC 102 and NGC 1365point in projection to the north, while H I in FCC 090 is elon-gated towards the south. FCC 090, FCC 102 and NGC 1365form a triplet of galaxies which lies along a line that points to-wards the centre of the cluster. ESO 358-G016 has a regular H I distribution and it lies in projection to the north of NGC 1365.All but NGC 1365 are H I deficient galaxies. The existence of asub-group centred on NGC 1365 is also consistent with the largescale structure around Fornax (the Fornax-Eridanus superclus-ter), which is mainly made by several groups of galaxies thatare assembling to form the cluster along the filament (Nasonovaet al. 2011).On the east of the cluster centre, we also see that FCC 323and NGC 1437B are near each other on the sky in projection( ∼
120 kpc apart), as well as in velocity ( ∼
10 km s − di ff erence).The unresolved H I morphology of FCC 323 does not reveal, atthis time, whether there is an ongoing interaction between thosegalaxies. We mark it as a potential subgroup.Another intriguing galaxy is ESO 358-G060. Among thegalaxies with M (cid:63) ≤ M (cid:12) it is the only one with a normalH I content compared to non-cluster galaxies (see Fig. 6), and itshows no evidence of on-going interactions with the Fornax en-vironment (see Fig. 2). Given its position just outside the causticcurves in Fig. 13, we discuss in Sect. 4 whether it might not haveentered the cluster yet. Another, similar case is NGC 1437A. Itis a H I rich galaxy which shows a quite regular H I morphology.It is also just outside the caustic curves in Fig. 13. However,unlike the undisturbed ESO 358-G060, the optical appearanceof NGC 1437A is peculiar, similar to that of NGC 1427A, sincethey both exhibit an arrow shaped optical morphology as pointedout in Raj et al. (2019). The MeerKAT Fornax Survey (Serraet al. 2016) will deliver a higher resolution H I image of this ob-ject and will be able to reveal any H I asymmetries that might behidden by projection e ff ects within the ATCA beam.
4. Discussion
The population of H I detected Fornax galaxies exhibits severalinteresting features. Despite the limited resolution of our data,half of all detections reveal H I asymmetries and o ff sets rela-tive to the stellar body (Fig. 2). Furthermore, the H I sample asa whole is gas-poorer and is forming stars at a lower rate thansamples of non − cluster galaxies in the same M (cid:63) range (Fig. 6and Fig. 10), and half of the galaxies with M (cid:63) > M (cid:12) have an anomalous M H / M H I ratio. Finally, the H I detections are dis-tributed in a noticeably di ff erent way with respect to the majorityof cluster galaxies both on the sky and in projected phase space(Fig. 12 and Fig. 13). This body of evidence suggests that theFornax environment is influencing the evolution of these galax-ies, which may be the most recent arrivals in the cluster. In par-ticular, the di ff erence between the 3D distribution of H I detectedgalaxies compared with that of spectroscopically confirmed For-nax members (Maddox et al. 2019), indicates that H I is a crucialobservable to test the volume of the cluster where H I rich galax-ies become H I deficient, before a complete H I removal in theinner part of the cluster.An outstanding question is how long it takes for a galaxy tolose its H I – the dominant component of the interstellar medium– as it falls into a cluster. In the case of Fornax we can gainsome insight through a joint analysis of Fig. 11 and Fig. 13.If we assume that H I is being actively removed from withingalaxies in Fornax (as suggested by the frequently disturbedH I morphologies - see Fig. 2), the lack of H I detections in thevirialised region of Fig. 13 implies that H I removal happenson a time scale τ HI , loss shorter than the cluster’s crossing time: τ HI , loss ≤ τ cross ∼ R vir /σ cl ∼ τ cross is similar to that estimated from simulations, 1.2 ± I by the timethey reach the pericentre.On the other hand, Fig. 11 suggests that, so far, H I has beenlost slowly in our H I detections (having been removed and / orconsumed, and not replenished). In that figure, our Fornax H I detections are distributed along the same correlation defined bynon-cluster galaxies. Thus, for those Fornax galaxies, the SFRhas so far had su ffi cient time to respond to a variation in H I mass– just like outside clusters. Since the transition from H I to newstars happens through the intermediate phase of H . ex , this ‘equilib-rium’ between H I and SFR implies that H . ex is depleted faster thanH I is lost in galaxies currently at the cluster’s outskirts, giving theentire cycle of H I -to-H . ex -to-SFR enough time to ‘see’ the vary-ing H I content. Thus, while for Fornax as whole τ HI , loss ≤ τ cross ∼ I detections in the outer regions ofFornax τ HI , loss has so far been ≥ τ H , depl ∼ . ex depletion time τ H , depl is only marginally shorterin Fornax, in particular at its outskirts, and anyway with largevariations from galaxy to galaxy; Zabel et al. 2020.)It is likely that, further inside Fornax, H I is removed fasterthan what we estimated above for galaxies at the cluster’s out-skirts. Indeed, simulations indicate that for a galaxy within0 . R vir , τ HI , loss < . I removal may not leave su ffi cient time for the SFR to ‘track’ thedecrease in H I mass in Fig. 11, resulting in galaxies moving tothe left of the non-cluster sample. A confirmation of this e ff ectmay come from the eight H . ex -detected galaxies (most of themH . ex -deficient; Zabel et al. 2019) where H I has already been re-moved at least down to the ATCA M H I sensitivity (left-pointingarrows in Fig. 11), and possibly by some H I detections closer tothe cluster centre (e.g. NGC 1436). Indeed, these galaxies oc-cupy a region to the left of the comparison sample, showing thattheir SFR is still significant despite their low H I content. Forthese galaxies, H I is likely to have been removed faster than H . ex is depleted: τ HI , loss ≤ τ H , depl . This conclusion is supported bythe anomalously high M H / M H I ratio of these galaxies in Fig. 9(in the M (cid:63) range where we have a reliable comparison).Within the picture discussed above, H I -detected galaxies inthe outer regions of Fornax are thus first infallers, which arestarting to interact with the Fornax environment and will lose Article number, page 11 of 19 & A proofs: manuscript no. aanda most of their H I by the time they reach the pericentre. Even atthe current early stage of infall, they already show a relativelylarge diversity in H I morphology (Fig. 2) and mass (Fig. 6).More specifically, we found two morphologically undisturbedH I rich galaxies (12% of the H I detections); four morphologi-cally disturbed H I rich galaxies (25% of the H I detections); fourmorphologically disturbed H I deficient galaxies (25% of the H I detections); six morphologically undisturbed - within the ATCAresolution - H I deficient galaxies (38% of the H I detections).The morphologically undisturbed H I -rich galaxies areESO 358-G060 and NGC 1437A. They both reside outside thecaustic curves in projection (Fig. 13) and, given their low M (cid:63) ,they should be easily perturbed by the Fornax environment.Thus, we speculate that they are recent Fornax members whichhave not yet had enough time to be significantly a ff ected by thecluster. One possible caveat is the low resolution of our images,which may hide H I disturbances in particular in the case of theoptically peculiar galaxy NGC 1437A. Another possibility is thatthey are outside the cluster volume or in a region with lower ICMdensity. ESO 358-G060 is also an outlier in Fig. 11, where therelatively high M H I / M (cid:63) does not correspond to a high SFR. Inthis galaxy no SFR was detectable from WISE W3 / W4 bands(after subtracting the old stellar population light), and the NUVcontribution to star formation is low. Several physical processesmight account for the low SFR with respect to the large H I reser-voir: for example, an ine ffi cient H I to H . ex conversion due to asmall amount of dust, a large angular momentum which pre-vents the H I from collapsing, and / or an H I external origin (e.g.see Geréb et al. 2016, 2018).The morphologically disturbed H I -rich galaxies areESO 358-G063, NGC 1351A, NGC 1427A, NGC 1365. Wediscuss the last galaxy when we focus our attention on the H I detected subgroup of interacting galaxies. ESO 358-G063 andNGC 1351A, are both disc galaxies north of NGC 1399. Theyare both quite isolated from all other Fornax Galaxies in RA,Dec and velocity in projection (see Fig. 12 and Fig. 13). Thus,their slightly H I disturbed morphologies described in Sect.3might be due to the interaction with the ICM of the Fornaxcluster. Furthermore, the agreement of M H / M H I of ESO 358-G063 and NGC 1351A with the xGASS scaling relation inFig. 9 suggests that they are recent Fornax members, thus thecluster environment has not had enough time to significantlydeplete their H I reservoirs. The last morphologically disturbedH I rich galaxy is NGC 1427A. This is the galaxy with thesecond-highest H I mass in our sample of Fornax H I detections.In Fig. 12 and Fig. 13, we see that NGC 1427A is the closestH I detected galaxy to the centre of the cluster in projection, butit has also the highest velocity. Thus, it may be a new Fornaxmember which is infalling from the foreground. As alreadymentioned, Lee-Waddell et al. (2018) studied the origin of theH I tail using the same data we present here. They concludedthat, unlike previously suggestions, ram-pressure is unlikely tobe the main process shaping the galaxy’s optical appearance.Instead, NGC 1427A is most likely a recent merger remnant,thus shaped by tidal forces. The recent merger might be thecause of the low molecular column density and / or its lowmetallicity, resulting currently undetected by ALMA (Zabelet al. 2019; it is the upper limit with the highest M (cid:63) in Fig. 9).NGC 1427A is also the second H I rich outlier with low SFR inFig. 11. The H I -to-H . ex conversion might be ine ffi cient to have aSFR consistent with its M H I / M (cid:63) ratio.The only H I deficient and morphologically disturbed galax-ies which do not belong to the NGC 1365 subgroup are: ESO-LV 3580611 and NGC 1437B. They are very close to one another on the sky but have significantly di ff erent velocities. The formeris outside the caustic curve in projection (Fig. 13), which sup-ports the hypothesis of an new infalling Fornax member madeby Schröder et al. (2001). The latter has a velocity similar to therecessional velocity of the cluster. H I and molecular morpholo-gies (the latter detected by Zabel et al. 2019) are elongated in thesame direction. This suggests that both gas phases are experienc-ing the same environmental interaction. Raj et al. 2019 detecteda tidal tail in NGC 1437B, which may be due to a recent fly-by ofanother galaxy. As mentioned in Sect.3.4, although the evidenceis not strong, NGC 1437B might be part of a subgroup of inter-acting galaxies which includes FCC 323. Thus, FCC 323 mightbe the fly-by galaxy that NGC 1437B has interacted with.It is di ffi cult to comment on the morphologically undisturbedH I deficient galaxies, since their symmetric H I distribution maybe a consequence of the ATCA resolution. However, a very pe-culiar case is the truncated H I disc of NGC 1436. It is the closestspiral galaxy to the centre of the cluster detected in H I . Raj et al.(2019) observed an ongoing morphological transition into lentic-ular: the spiral structure is found only in its inner region, whilethe outer disc has the smooth appearance typically found in S0galaxies.The inner part of NGC 1436 appears regular also in H . ex , andthe galaxy is just moderately H . ex deficient (Zabel et al. 2019). Itis also the only galaxy detected both in H I and H . ex which showsa high M H / M H I ratio in Fig. 9. These results and the evidenceof morphological distortions only in the outer part of the galaxysuggest that this galaxy may have gone through a quick interac-tion with the cluster environment, which did not a ff ect the innerspiral structure yet. This idea is corroborated by the fact thatNGC 1436 lies on the upper edge on the comparison sample inFig. 11, which means that despite its H I deficiency it is still form-ing stars at a significant rate.Finally, we focus on the NGC 1365 subgroup. We discussedall the H I morphologies and the 3D distribution of the subgroupmembers in Sect.3.1 and Sect.3.4, respectively. As mentionedearlier, NGC 1365 is the only H I rich galaxy of the subgroup(Fig. 6) with a M H I at least two orders of magnitude larger thanthe other members. The H I distribution both in NGC 1365 andin FCC 102 is elongated to the north, while it is elongated tothe south in FCC 090. ESO 358-G016 is the only galaxy with aregular H I morphology. It is also the only galaxy located northof NGC 1365. We propose two scenarios in order to explainthe properties of the subgroup and its members: the former isa case of interaction between galaxies and the cluster environ-ment which is responsible of the high M H I / M (cid:63) ratio of the lowmass members. In contrast, due to to its deep gravitational po-tential, NGC 1365 has been able to retain its H I , although someof it has been perturbed. The latter scenario we propose, it is acase of preprocessing in a group of galaxies where the local en-vironment of the subgroup was able to a ff ect the M H I / M (cid:63) ratioof the low mass members before the group began to interact withthe cluster environment.In general, although we found a large variety of H I proper-ties in our sample of H I detections, we detected an overall trendtowards H I disturbances and deficiency in Fornax. Fig. 6 makesevident an already evolved state of Fornax H I galaxies where ∼ / I deficient. Fig. 6 also shows that theFornax environment is more e ff ective in altering the gas contentof galaxies with M (cid:63) < × M (cid:12) ) (see Sect.3.2).Zabel et al. (2019) presented a similar study, where theycompare Fornax and field galaxies based on their molecu-lar gas properties. They found some molecular deficiency inall their detections except NGC 1365. However, galaxies with Article number, page 12 of 19. Loni et al.: A blind ATCA H I survey of the Fornax galaxy cluster M (cid:63) < × M (cid:12) are both more H . ex -deficient and morpholog-ically disturbed with respect to more massive galaxies, whosemolecular gas morphology is always regular. We do not ob-serve such a clear di ff erence in H I . Indeed, we observe disturbedH I morphologies across our entire M (cid:63) range, confirming thatatomic hydrogen is the best tracer of early interactions. Despitethis di ff erence between H I and H . ex morphologies, we also notesome similarities between our results and those in Zabel et al.(2019). Indeed, the mass range of molecular disturbed galax-ies is also characterised by a stronger H I depletion. Conversely,the H I depletion is weaker in the mass range in which galax-ies show regular molecular-gas morphologies ( M (cid:63) > × M (cid:12) ;Sect.3.2). Thus, both gas phases show that the Fornax environ-ment is more e ff ective in altering the gas content of low-massgalaxies compared to high-mass galaxies.Finally, our results are in agreement with the FDS&F3D re-sults (Iodice et al. 2019a,b). Indeed, almost all galaxies with adisturbed H I morphology are of late type and belong to the groupof the infalling galaxies in (Iodice et al. 2019a), which are sym-metrically distributed around the cluster’s central region. Thesegalaxies have active star formation and are located in the low-density region of the cluster, where the X-ray emission is faint orabsent. Our results show that they are interacting with the clusterenvironment. Deeper into the cluster, the lack of H I detections isconsistent with the result that this region is dominated by evolvedearly-type galaxies. Some of these galaxies have been able to re-tain part of their H . ex reservoirs (Zabel et al. 2019), but not theirH I .
5. Summary
The blind ATCA H I survey of the Fornax galaxy cluster coversa field of 15 deg out to a distance of ∼ R vir from the clustercentre. It has a spatial and velocity resolution of 67 (cid:48)(cid:48) × (cid:48)(cid:48) and6.6 km s − , respectively, and a 3 σ N H I and M H I sensitivity of ∼ × cm − and ∼ × M (cid:12) , respectively. The survey re-vealed H I emission from 16 Fornax galaxies covering a massrange of about three orders of magnitude, from 8 × to1 . × M (cid:12) . These galaxies exhibit a variety of disturbancesof the H I morphology, including asymmetries, tails, o ff sets be-tween H I and optical centres and a case of a truncated H I disc(Fig. 2). This suggests environmental interactions within or ontheir way to Fornax (whether with other galaxies, the large-scalepotential or the intergalactic medium), supported by the o ff setof Fornax galaxies towards low M H I / M (cid:63) ratios with respect tothe xGASS M (cid:63) - M H I / M (cid:63) scaling relation (Fig. 6), and resultingin H I deficiencies similar to those observed in the Virgo cluster(Fig. 8). The H I sample of Fornax galaxies is also forming starsat a lower rate than samples of non-cluster galaxies at fixed M (cid:63) (Fig. 10). This deficit of SFR is consistent with the deficit of H I when compared to non-cluster galaxies (Fig. 11).Our 16 detections reside outside the virialised region of thecluster – where the distribution of the general population of For-nax galaxies is clustered – both on the sky and in the projectedphase space diagram (Fig. 12 and Fig. 13). This result impliesthat H I is lost down to the ATCA sensitivity within a crossingtime ( τ HI , loss ≤ τ cross ∼ I detections arerecent arrivals in the cluster. They still reside at the outskirts ofFornax, where their H I and SFR properties suggest that H I hasso far been lost on a time scale longer than the H . ex depletion time( τ HI , loss ≥ τ H , depl ∼ − I removal is likely to proceed faster ( τ HI , loss < τ H , depl ). Thisis supported by the relatively high SFR of H I -undetected, H . ex -detected galaxies and by the anomalously high M H / M H I ratios of galaxies in those regions (Fig. 9, Fig. 12 and Fig. 13). Theseare galaxies where SFR is likely to be proceeding relatively un-perturbed after rapid removal of the H I .This picture is enriched by the new detection of theNGC 1365 subgroup – where both pre-processing and early in-teraction with the cluster environment are plausible scenarios toaccount for the H I properties of its members – and by the de-tection of several galaxies with peculiar ISM properties, such assome H I -rich but H . ex -poor and low-SFR galaxies (NGC 1427A,ESO 358-G060). The future MeerKAT Fornax Survey (Serraet al. 2016) will observe this cluster with a better resolution andsensitivity than those of our ATCA survey, enabling a furtherstep forward in the study of the evolution of Fornax galaxies. Acknowledgements.
This project has received funding from the European Re-search Council (ERC) under the European Union’s Horizon 2020 research andinnovation programme (grant agreement no. 679627; project name FORNAX).Parts of this research were supported by the Australian Research Council Centreof Excellence for All Sky Astrophysics in 3 Dimensions (ASTRO 3D), throughproject number CE170100013. LC is the recipient of an Australian ResearchCouncil Future Fellowship (FT180100066) funded by the Australian Govern-ment. This publication has received funding from the European Union Horizon2020 research and innovation programme under the Marie Skłodowska-Curiegrant agreement number 721463 to the SUNDIAL ITN network. NZ acknowl-edges support from the European Research Council (ERC) in the form of Con-solidator Grant CosmicDust (ERC-2014-CoG-647939). This work made use ofthe Digitized Sky Surveys, which were produced at the Space Telescope Sci-ence Institute under U.S. Government grant NAG W-2166. The Australia Tele-scope Compact Array is part of the Australia Telescope National Facility whichis funded by the Australian Government for operation as a National Facility man-aged by CSIRO. We acknowledge the Gomeroi people as the traditional ownersof the Observatory site.
Article number, page 13 of 19 & A proofs: manuscript no. aanda N a m e F CCR AD ec v op t v H I F l ux M H I M (cid:63) R M SF l ux M o r ph . D ( B ) N o t e s ( J )( J )( lit e r a t u r e ) ( hh : mm : ss . ss )( dd : mm : ss . ss ) (cid:16) k m s (cid:17)(cid:16) k m s (cid:17)(cid:16) J yk m s (cid:17) ( M (cid:12) )( M (cid:12) )( m J y b ea m ) (cid:16) J yk m s (cid:17) ( a r c s ec ) E S O - G : : . - : : . . ± . . ± . ± . . ± . ∗ S c d72 . H I d e f . E S O - G : : . - : : . . ± . . ± . . ± . . . ± . (cid:52) I m . H I d e f . ; N s . g r oup E S O - G : : . - : : . . ± . . ± . ± . . ± . ◦ S c d92 . H I d e f . E S O - G : : . - : : . ± ± . ± . . . ± . (cid:72) S c d104 . E S O - G : : . - : : . ± ± ± . . ± . (cid:78) S c d280 . H I d i s t; E S O - L V : : . - : : . . ± . . ± . . ± . . . ± . (cid:79) I m . H I d e f . ; H I d i s t; F CC : : . - : : . . ± . . ± . ± . E . H I d e f . ; H I d i s t; N s . g r oup F CC : : . - : : . . ± . . ± . . ± . . I m . H I d e f ., H I d i s t; N s . g r oup F CC : : . - : : . . ± . . ± . . ± . . . ± . (cid:79) I m . H I d e f . F CC : : . - : : . < . < . ± . E . H I und e t . F CC : : . - : : . < . < . . ± . . E . H I und e t . F CC : : . - : : . < . < . ± . E . H I und e t . F CC : : . - : : . . ± . . ± . . ± . . E H I d e f . F CC : : . - : : . < . < . ± . E . H I und e t . F CC : : . - : : . < . < . ± . E . H I und e t . NG C A : : . - : : . ± ± ± . . ± . (cid:78) S c . H I d i s t; NG C : : . - : : . ± ± ± . ± (cid:72) S b673 . H I d i s t . ; N s . g r oup NG C : : . - : : . < . < . ± . S a . H I und e t . NG C : : . - : : . < . < . ± . S a . H I und e t . NG C : : . - : : . < . < . ± . E . H I und e t . NG C A : : . - : : . ± ± ± . . ± . (cid:72) I m . H I d i s t; NG C : : . - : : . . ± . . ± . ± . . ± . (cid:5) S c . H I d e f . ;t r un ca t e d H I d i s c NG C A : : . - : : . ± ± ± . ± (cid:78) S c d111 . NG C B : : . - : : . . ± . . ± . ± . . ± . (cid:78) S c d157 . H I d e f . ; H I d i s t; T a b l e : W e r e po r t p a r a m e t e r s on t h e H I d e t ec t e dg a l a x i e s . R A a nd D ecc oo r d i n a t e s c o rr e s pond t o t h e op ti ca l ce n t r e s fr o m N E D . E rr o r s on t h e t o t a l H I fl ux e s a r ee v a l u a t e d a s d e s c r i b e d i n S ec t . . ‘ F l ux ( lit e r a t u r e )’ s ho w s t h e t o t a l fl ux e s m ea s u r e d fr o m o t h e r s u r v e y s a s f o ll o w s : (cid:5) C ou r t o i s e t a l . ( ) ; (cid:52) B u r ea u e t a l . ( ) ; ∗ M a tt h e w s e t a l . ( ) ; ◦ T h e u r ea u e t a l . ( ) ; (cid:79) S c h r öd e r e t a l . ( ) ; (cid:72) K o r i b a l s k i e t a l . ( ) ; (cid:78) H I P A SS d a t a r e p r o ce ss e dbyu s . T h e t o t a l fl uxun ce r t a i n t yo f t h ec o m p a r i s on s p ec t r aa r eca l c u l a t e d a s d e s c r i b e d i n S ec t . c o m b i n i ng t h e no i s e i n t h e s p ec t r u m a nd t h e fl ux - s ca l e un ce r t a i n t yo f eac h s u r v e y e x ce p t f o r E S O - G f o r w h i c h t h e fl ux - s ca l e un ce r t a i n t y w a s no t p r ov i d e d . F o r t h i s g a l a xy w e s ho w t h e un ce r t a i n t yon t h e t o t a l fl uxg i v e n i n B u r ea u e t a l . ( ) . C o l u m n s ‘ M o r ph . ’ a nd ‘ D ( B )’ s ho w t h e op ti ca l m o r pho l ogy a nd t h e op ti ca l s i ze o f ou r g a l a x i e s u s e d i n S ec t . . t o e v a l u a t e t h ee xp ec t e d H I c on t e n t o f ou r g a l a x i e s w it h t h e H a yn e s & G i ov a n e lli ’ s ( ) m e t hod . D ( B ) i s t h e op ti ca li s opho t a l d i a m e t e r m ea s u r e d a t m a g / a r c s ec i n B - b a nd ; w ec o ll ec t e d t h e s e v a l u e s fr o m N E D f o r t o f a l a x i e s . A l m o s t a ll o f t h e op ti ca l d i a m e t e r s c o m e s fr o m RC ca t a l ogu e ( d e V a u c ou l e u r s e t a l . ) , t w oou t o f a r e fr o m L a ub e r t s & V a l e n tij n ( ) . W ee s ti m a t e d D ( B )f o r a llt h e r e m a i n i ngg a l a x i e s e x ce p t f o r F CC a s f o ll o w s : fr o m t h e op ti ca l m u lti - c o m pon e n t d ec o m po s iti on m a d e by S u e t a l . ( ) , w e s ca l e d t h e r- b a nd r a d i a l p r o fi l e t o B - b a nd . I no r d e r t odo t h i s , w e p r e v i ou s l y c onv e r t e dg -rfr o m V e nho l ae t a l . ( ) t oob t a i n B -r u s i ng t h e i n t e r m e d i a t ec onv e r s i on f o r m u l a g -r t o B - g ( L up t on2005 ) . I n c o l u m n ‘ N o t e s ’ : H I d e f , H I d i s t , N s . g r oup s t a nd f o r H I d e fi c i e n t , H I m o r pho l og i ca ll yd i s t u r b e dg a l a xy , a ndg a l a xy i n c l ud e d i n t h e NG C s ubg r oup , r e s p ec ti v e l y . Article number, page 14 of 19. Loni et al.: A blind ATCA H I survey of the Fornax galaxy cluster References
Barnes, D. G., Staveley-Smith, L., de Blok, W. J. G., et al. 2001, MNRAS, 322,486Barnes, D. G., Staveley-Smith, L., Webster, R. L., & Walsh, W. 1997, MNRAS,288, 307Bigiel, F., Leroy, A., Walter, F., et al. 2008, AJ, 136, 2846Blakeslee, J. P., Jordán, A., Mei, S., et al. 2009, The Astrophysical Journal, 694,556Boquien, M., Kennicutt, R., Calzetti, D., et al. 2016, A&A, 591, A6Boselli, A., Cortese, L., & Boquien, M. 2014, A&A, 564, A65Boselli, A. & Gavazzi, G. 2006, PASP, 118, 517Boselli, A. & Gavazzi, G. 2009, A&A, 508, 201Bureau, M., Mould, J. R., & Staveley-Smith, L. 1996, ApJ, 463, 60Cantiello, M., D’Abrusco, R., Spavone, M., et al. 2018, A&A, 611, A93Catinella, B., Saintonge, A., Janowiecki, S., et al. 2018, MNRAS, 476, 875Cayatte, V., Kotanyi, C., Balkowski, C., & van Gorkom, J. H. 1994, AJ, 107,1003Chamaraux, P., Balkowski, C., & Gerard, E. 1980, A&A, 83, 38Chung, A., van Gorkom, J. H., Kenney, J. D. P., Crowl, H., & Vollmer, B. 2009,AJ, 138, 1741Cluver, M. E., Jarrett, T. H., Dale, D. A., et al. 2017, ApJ, 850, 68Cluver, M. E., Jarrett, T. H., Hopkins, A. M., et al. 2014a, ApJ, 782, 90Cluver, M. E., Jarrett, T. H., Hopkins, A. M., et al. 2014b, ApJ, 782, 90Condon, J. J., Cotton, W. D., Greisen, E. W., et al. 1998, AJ, 115, 1693Cortese, L., Bekki, K., Boselli, A., et al. 2016, MNRAS, 459, 3574Cortese, L., Catinella, B., Boissier, S., Boselli, A., & Heinis, S. 2011, MNRAS,415, 1797Courtois, H. M., Tully, R. B., Fisher, J. R., et al. 2009, AJ, 138, 1938Davies, R. D. & Lewis, B. M. 1973, MNRAS, 165, 231de Vaucouleurs, G., de Vaucouleurs, A., Corwin, Herold G., J., et al. 1991, ThirdReference Catalogue of Bright GalaxiesDey, A., Schlegel, D. J., Lang, D., et al. 2019, AJ, 157, 168Diaferio, A., Kau ff mann, G., Balogh, M. L., et al. 2001, MNRAS, 323, 999Dressler, A. 1980, ApJ, 236, 351Dressler, A. 1986, ApJ, 301, 35Drinkwater, M. J., Gregg, M. D., & Colless, M. 2001a, ApJ, 548, L139Drinkwater, M. J., Gregg, M. D., Holman, B. A., & Brown, M. J. I. 2001b, MN-RAS, 326, 1076Ferguson, H. C. 1989, AJ, 98, 367Frank, K. A., Peterson, J. R., Andersson, K., Fabian, A. C., & Sanders, J. S.2013, ApJ, 764, 46Gavazzi, G., Boselli, A., Scodeggio, M., Pierini, D., & Belsole, E. 1999, MN-RAS, 304, 595Geréb, K., Catinella, B., Cortese, L., et al. 2016, MNRAS, 462, 382Geréb, K., Janowiecki, S., Catinella, B., Cortese, L., & Kilborn, V. 2018, MN-RAS, 476, 896Giovanelli, R. & Haynes, M. P. 1983, AJ, 88, 881Grillmair, C. J., Freeman, K. C., Bicknell, G. V., et al. 1994, ApJ, 422, L9Gunn, J. E. & Gott, J. Richard, I. 1972, ApJ, 176, 1Hamraz, E., Peletier, R. F., Khosroshahi, H. G., et al. 2019, A&A, 625, A94Haynes, M. P. & Giovanelli, R. 1984, AJ, 89, 758Haynes, M. P. & Giovanelli, R. 1986, ApJ, 306, 466Horellou, C., Black, J. H., van Gorkom, J. H., et al. 2001, A&A, 376, 837Hubble, E. & Humason, M. L. 1931, ApJ, 74, 43Hughes, T. M. & Cortese, L. 2009, MNRAS, 396, L41Iodice, E., Capaccioli, M., Grado, A., et al. 2016, ApJ, 820, 42Iodice, E., Sarzi, M., Bittner, A., et al. 2019a, A&A, 627, A136Iodice, E., Spavone, M., Capaccioli, M., et al. 2017, ApJ, 839, 21Iodice, E., Spavone, M., Capaccioli, M., et al. 2019b, A&A, 623, A1Ja ff é, Y. L., Smith, R., Candlish, G. N., et al. 2015, MNRAS, 448, 1715Jarrett, T. H., Cluver, M. E., Brown, M. J. I., et al. 2019, ApJS, 245, 25Jarrett, T. H., Masci, F., Tsai, C. W., et al. 2013, AJ, 145, 6Jones, M. G., Espada, D., Verdes-Montenegro, L., et al. 2018, A&A, 609, A17Jordán, A., Blakeslee, J. P., Côté, P., et al. 2007, ApJS, 169, 213Jorsater, S. & van Moorsel, G. A. 1995, AJ, 110, 2037Kennicutt, R. C. & Evans, N. J. 2012, ARA&A, 50, 531Koribalski, B. S., Staveley-Smith, L., Kilborn, V. A., et al. 2004, AJ, 128, 16Kreckel, K., Platen, E., Aragón-Calvo, M. A., et al. 2012, AJ, 144, 16Lauberts, A. & Valentijn, E. A. 1989, The surface photometry catalogue of theESO-Uppsala galaxiesLee-Waddell, K., Serra, P., Koribalski, B., et al. 2018, MNRAS, 474, 1108Leroy, A. K., Sandstrom, K. M., Lang, D., et al. 2019, ApJS, 244, 24Lupton, P. 2005Maddox, N., Hess, K. M., Obreschkow, D., Jarvis, M. J., & Blyth, S. L. 2015,MNRAS, 447, 1610Maddox, N., Serra, P., Venhola, A., et al. 2019, MNRAS, 490, 1666Marasco, A., Crain, R. A., Schaye, J., et al. 2016, MNRAS, 461, 2630Matthews, L. D., van Driel, W., & Gallagher, J. S., I. 1998, AJ, 116, 1169Meyer, M., Robotham, A., Obreschkow, D., et al. 2017, PASA, 34, 52 Mould, J. R., Hughes, S. M. G., Stetson, P. B., et al. 2000, ApJ, 528, 655Nasonova, O. G., de Freitas Pacheco, J. A., & Karachentsev, I. D. 2011, A&A,532, A104Navarro, J. F., Frenk, C. S., & White, S. D. M. 1996, ApJ, 462, 563Oemler, Augustus, J. 1974, ApJ, 194, 1Oh, S., Kim, K., Lee, J. H., et al. 2018, ApJS, 237, 14Ondrechen, M. P. & van der Hulst, J. M. 1989, ApJ, 342, 29Paolillo, M., Fabbiano, G., Peres, G., & Kim, D. W. 2002, ApJ, 565, 883Peletier, R., Iodice, E., Venhola, A., et al. 2020, arXiv e-prints, arXiv:2008.12633Raj, M. A., Iodice, E., Napolitano, N. R., et al. 2019, A&A, 628, A4Rhee, J., Smith, R., Choi, H., et al. 2017, ApJ, 843, 128Rodríguez-Ardila, A., Prieto, M. A., Mazzalay, X., et al. 2017, MNRAS, 470,2845Sault, R. J., Teuben, P. J., & Wright, M. C. H. 1995, Astronomical Society ofthe Pacific Conference Series, Vol. 77, A Retrospective View of MIRIAD, ed.R. A. Shaw, H. E. Payne, & J. J. E. Hayes, 433Schröder, A., Drinkwater, M. J., & Richter, O. G. 2001, A&A, 376, 98Schweizer, F. 1980, ApJ, 237, 303Seibert, M., Wyder, T., Neill, J., et al. 2012, in American Astronomical SocietyMeeting Abstracts, Vol. 219, American Astronomical Society Meeting Ab-stracts Article number, page 15 of 19 & A proofs: manuscript no. aanda
Appendix A:
In this section, we describe in detail each galaxy of our sample.Whenever possible we compare the information estimated fromour H I data with observations at other wavelengths. The galaxiesare sorted according to increasing H I mass as in Table 1, Fig. 2and Fig. 3. FCC 323
According to Ferguson (1989), FCC 323 is a dE2 / ImV. This isthe faintest among our H I detections and, for the first time, wedetected it in H I and thus measured its recessional velocity of v H I = − . FCC 323 has a M H I = (7.6 ± × M (cid:12) and a M (cid:63) = (6.3 ± × M (cid:12) . The Fornax environment mighthave strongly a ff ected the H I reservoir of FCC 323. It has thehighest M H I / M (cid:63) o ff set from the (extrapolated) xGASS scalingrelation in Fig. 6. Its closest galaxy both on the sky and velocityis NGC 1437B (120 kpc and 12 km s − away), leading us tothink that FCC 323 and NGC 1437B might be part of a subgroupor a pair of galaxies. FCC 120
FCC 120 is classified as ImIV by Ferguson (1989). The H I emis-sion associated with the galaxy seems to be quite regular. TheATCA spectrum in Fig. 3 shows one peak (within the uncertain-ties), which is di ff erent from the 100 - km s − - wide double hornspectrum detected by Schröder et al. (2001) based on Parkesdata. Even if the latter data shows strong RFI at a velocity of1250 km s − , this RFI does not a ff ect the detection of the galaxy.However, v opt from Maddox et al. (2019), agrees with our v H I .Deeper H I data (Serra et al. 2016) may confirm this. FCC 120has a M H I = (4.5 ± × M (cid:12) and a M (cid:63) = (1.6 ± × M (cid:12) . FCC 102
FCC 102 is an irregular ImIV classified by Ferguson (1989). Wedetect H I emission associated with FCC 102 for the first time.Although the H I morphology is not resolved, the H I is o ff set to-wards the north with respect to the optical centre. The di ff erencebetween v opt from Maddox et al. (2019), and v H I from ATCA datais almost 100 km s − . However, those values are still consistentdue to the large uncertainty on v opt , which was measured fromabsorption lines for this galaxy (Natasha Maddox, priv. comm.).Recently this galaxy was detected during MeerKAT test obser-vations in preparation for the MeerKAT Fornax Survey Serraet al. (2016), confirming our ATCA H I results. FCC 102 has a M H I = (4.6 ± × M (cid:12) and a M (cid:63) = (1.6 ± × M (cid:12) . Thecomparison with H I belonging to non-cluster galaxies with sim-ilar M (cid:63) in Fig. 6 shows that FCC 102 is H I deficient.We identify FCC 102 to be part of the NGC 1365 infallingsubgroup. It lies between NGC 1365 and FCC 090 both on thesky and in velocity (see Fig. 12 and Fig. 13). Furthermore the H I elongation of FCC 102 is in the same direction as the H I elonga-tion of NGC 1365. ‘ NGC 1436
NGC 1436 is the closest H I detected spiral to the centre of thecluster (see Fig 12 and Fig. 13). Ferguson (1989) classify thisgalaxy as ScII. The H I emission associated with NGC 1436 isconfined well within the stellar body suggesting a truncated H I disc. However, Fig. 3 shows that the ATCA spectrum comparedwith that based on GBT data (Courtois et al. 2009) misses fluxfrom the blushifted part of the disc. This discrepancy might bedue to a faint H I emission of NGC 1436 which extends over aregion smaller than the ATCA beam in at least some of the blue-shifted channels. This would cause the average HI column den-sity within the beam to fall below the sensitivity limit, resultingin a loss of H I flux in our data. The H I spectrum of NGC 1436spectrum from Schröder et al. (2001) is more similar to that ob-tained with the GBT than to ours. The M H I estimated by theATCA flux is (5.8 ± × M (cid:12) , while the M (cid:63) of the galaxyis (1.6 ± × M (cid:12) . Whether adopting the ATCA or the GBTH I flux measurement, NGC 1436 would still be an H I deficientgalaxy (see Sect.3.2). Similarly, it would still be the (H I and H . ex )detection with the largest M H / M H I ratio in Fig. 9.The high o ff set from the scaling relation in Fig. 9, the trun-cated H I disc in Fig. 2 and the molecular spiral structure de-tected by Zabel et al. (2019) show that NGC 1436 may havegone through a quick interaction with the cluster environment,which did not a ff ect the inner spiral structure yet.Furthermore, Raj et al. (2019) found that NGC 1436 appearsto be transforming into a S0 in its morphological evolution. Theyfound spiral arms only in the central region whereas its outer discresembles the smooth structure typically found in the discs of S0galaxies. The truncated H I disc agrees with the observation of apassively-evolving outer optical disc, which is being shaped bythe Fornax environment. FCC 090
FCC 090 is classified as a peculiar elliptical galaxy by Fer-guson (1989). We detected H I emission in FCC 090 for thefirst time. The H I morphology of FCC 090 is elongated to-wards the south. FCC 090 has a M H I = (5.9 ± × M (cid:12) anda M (cid:63) = (1.3 ± × M (cid:12) . Its M H I / M (cid:63) ratio makes it a H I de-ficient galaxy, which lies on the edge of comparison sample ofnon-cluster galaxies (Fig. 6).FCC 090 is part of the NGC 1365 infalling subgroup. Withinthe subgroup, the H I morphology of FCC 090 is the only onewhich is extended towards the south; it is the farther galaxy fromNGC 1365; it has the highest velocity in phase space. FCC 090is also the westernmost galaxy in the triplet of the aligned H I dis-turbed galaxies. Besides FCC 090, this triplet includes FCC 102and NGC 1365.The H I and molecular distributions (Zabel et al. 2019) aredi ffi cult to compare due to the di ff erent resolutions betweenATCA and ALMA. The molecular distribution extends beyondthe stellar body showing a tail that points to the west in projec-tion but it is still within our H I detection that does not show anypeculiarity in that direction. Our knowledge on H I in FCC 090as well as its belonging to the NGC 1365 subgroup, corroboratesthe idea of an infalling galaxy which is loosing the external gasenvelope.In the sample of dwarf galaxies studied by Hamraz et al.(2019) FCC 090 is one out of two outliers in the Fornax clus-ter with a very blue inner part. Recently, Zabel et al. (2020)showed that depletion time is shorter than usual (0.5 Gyr ratherthan 2 Gyr) which may be a consequence of the environmentalinteractions taking place in it. Article number, page 16 of 19. Loni et al.: A blind ATCA H I survey of the Fornax galaxy cluster ESO 358-G016
ESO 358-G016 is classified as Sdm (edge-on) by Fergu-son (1989). It has a M H I = (7.4 ± × M (cid:12) and a M (cid:63) = (1.2 ± × M (cid:12) . ESO 358-G016 is the only memberof the NGC 1365 infalling subgroup with a regular H I morphol-ogy. Like all the low mass members of the subgroup, it is a H I deficient galaxy. It is also the second closest galaxy to NGC 1365both on the sky and in phase space. Raj et al. 2019 suggested thatESO 358-G016 might have experienced disruptions of the outerdisc due to the gravitational potential of the cluster. If that is thecase, some H I survived in this process. ESO 358-G015
ESO 358-G015 is classified as a peculiar Scd-III by Ferguson(1989). It has a regular H I morphology, with a systemic ve-locity close to the recessional velocity of the cluster. It has a M H I = (9.3 ± × M (cid:12) and a M (cid:63) = (7.6 ± × M (cid:12) . Fig. 6shows that ESO 358-G015 is a H I deficient galaxy. Further-more, it is one of the four H I detected galaxies that are northof NGC 1399. Raj et al. (2019) identified a lopsided tail pointingto the centre of the cluster. They also suggested ESO 358-G015to be a galaxy that is being pulled into the cluster centre, in thesouthern direction. The quite regular H I morphology does notprovide clear support to this hypothesis. ESO 358-G051
ESO 358-G051 is classified as SBcd-III by Ferguson (1989).The H I distribution is regular and peaks on the optical centre.It is located north of NGC 1399 and it is the closest H I deficientgalaxy from it ( ∼ R vir ). In general, among all our H I detec-tions, only NGC 1427A is closer to NGC 1399. ESO 358-G051has a M H I = (1.0 ± × M (cid:12) and a M (cid:63) = (2.1 ± × M (cid:12) .Fig. 6 shows that ESO 358-G051 is a H I deficient galaxy.Raj et al. (2019) marked this galaxy as a recent infaller withstrong central H α emission powered by star formation. Iodiceet al. (2019a) observed an ionised gas distribution extended to-wards the north. This is the opposite direction with respect tothe H . ex elongation detected by Zabel et al. (2019). However, theH I distribution does not seem to be disturbed, perhaps due tothe ATCA resolution. Although ESO 358-G051 is a H I deficientgalaxy, the unresolved H I morphology does not provide infor-mation about on going interaction with the Fornax environment. ESO-LV 3580611
ESO-LV 3580611 is classified as SBm-III by Ferguson (1989).The H I morphology shows a lopsided H I distribution in thenorthern part of the system. However, optical and H I distribu-tion share the same centre. It has a M H I = (1.1 ± × M (cid:12) and a M (cid:63) = (1.0 ± × M (cid:12) . ESO-LV 3580611 is also oneof the two H I deficient galaxies in our sample that is outsidethe caustic curves (Fig. 13). Schröder et al. (2001), based on thehigh deviation of ESO-LV 3580611 from the Tully-Fisher rela-tion, suggested that this galaxy is falling into the cluster from thebackground.According to Raj et al. (2019), ESO-LV 3580611 may haveexperienced disruptions in the outskirts of the disc due to gravi-tational potential well of the cluster. Thus, similarly to ESO 358-G016, H I has survived during the infall. NGC 1437B
NGC 1437B is classified as Sd (edge-on) by Ferguson (1989).The H I distribution shows an elongation to the south, while theH I peak corresponds to the optical centroid (Fig. 2).This galaxy is located on the south east of the cluster,close to FCC 323 both on the sky and in velocity (see Fig. 12and Fig. 13). Given the small separation of only 120 kpcwith a di ff erence of ∼
10 km s − in velocity, they might be partof an interacting subgroup of galaxies. NGC 1437B has a M H I = (2.4 ± × M (cid:12) and a M (cid:63) = (5.00 ± × M (cid:12) andit is a H I rich galaxy.Fig. 9 shows that the Fornax environment has not strongly af-fected the M H / M H I ratio yet. On the other hand, contrary to thecase of NGC 1436, the two gas phases are probably experiencingthe same environmental e ff ect. Indeed, although the scales aredi ff erent, we found consistency between the southern H I elon-gation of NGC 1437B and the elongation of the molecular mor-phology detected by Zabel et al. (2019). Thus, the agreement be-tween its M H / M H I ratio and the xGASS scaling relation mightbe due to a slower evolution with respect to that taking place inNGC 1436.Raj et al. (2019) did not observed any hint of optical mor-phological transition from the late-type spiral structure, whichsupports the hypothesis of a recent infaller. However, they foundan optical disturbance in form of tidal tail, which points to thesouth in projection and might be due to a recent fly-by of anothergalaxy in the cluster. The H I morphology agrees with the opticaldetected tail. If FCC 323 and NGC 1437B belong to the samesubgroup, the former may be the fly-by galaxy that NGC 1437Bhas interacted with. NGC 1351A
Classified as Sc (edge-on) by Ferguson (1989). The H I dis-tribution shows a projected elongation towards the south.It has a velocity similar to the recessional of the cluster(Fig. 13). NGC 1351A has a M H I = (4.7 ± × M (cid:12) and a M (cid:63) = (3.5 ± × M (cid:12) . The H I reservoir is similar to that ofnon-cluster galaxies with the same M (cid:63) (Fig. 6). NGC 1351A isisolated both on the sky and in projected phase space (Fig. 12and Fig. 13). Thus, the H I asymmetry might be due to the in-teraction with ICM. Zabel et al. (2019) detected a slightly moredi ff use east side in the H . ex distribution. NGC 1437A
NGC 1437A is classified as SdIII by Ferguson (1989). The H I morphology is quite regular and its H I content is comparablewith that of non-cluster galaxies with the same M (cid:63) (see Fig.s2 and 6). NGC 1437A has a M H I = (5.6 ± × M (cid:12) and a M (cid:63) = (1.0 ± × M (cid:12) . Although the optical appearance ofNGC 1437A is not regular and the location of its star form-ing regions suggests that the galaxy is travelling in a south-eastdirection (Raj et al. 2019), the H I distribution does not showany strong asymmetries. Thus, ram-pressure and tidal interac-tion might not be the cause of the asymmetric star forming re-gions. However, the poor resolution of the H I image may hide H I asymmetries in projection. ESO 358-G060
ESO 358-G060 is a low mass galaxy classified as Sdm(edge-on) by Ferguson (1989). The H I morphology is regu- Article number, page 17 of 19 & A proofs: manuscript no. aanda lar. ESO 358-G060 has a M H I = (1.1 ± × M (cid:12) and a M (cid:63) = (1.0 ± × M (cid:12) . Fig. 6 shows that ESO 358-G060 hasa H I reservoir comparable with that of non-cluster galaxies withthe same M (cid:63) . As discussed in Sect.3.2, we pointed out that theFornax cluster environment is more e ff ective in altering the gascontent for galaxies with M (cid:63) < × M (cid:12) . This supports theidea that ESO 358-G060 is a likely new Fornax member whichhas not been a ff ected by the cluster environment yet. Raj et al.(2019) observed irregular star-forming regions making the hy-pothesis of a disruptions due to the gravitational potential well ofthe cluster centre, during the fall. However, since the H I distribu-tion is not perturbed yet by environmental interactions, internalfeedback might be the cause of the irregular star forming regions.Furthermore no H . ex was detected by Zabel et al. (2019) and it isa dust poor galaxy as discussed in Sect.4. Overall, this galaxy isforming star with lower rate than that predicted for galaxies withsimilar M H I / M (cid:63) (Fig. 11). ESO 358-G063
Classified as Scd (edge-on) by Ferguson (1989). The atomic hy-drogen distribution is more extended to the south-east part ofthe system. On this side, H I contours are more spaced with re-spect to the opposite side of the disc where H I emission seemsto be confined within the stellar body. ESO 358-G063 has a M H I = (1.7 ± × M (cid:12) and a M (cid:63) = (1.1 ± × M (cid:12) . De-spite the asymmetries in the H I distribution, ESO 358G-063is not H I deficient, and the molecular gas detected by Zabelet al. (2019) has a regular morphology. Thus, the galaxy has juststarted to interact with the cluster environment. The idea is sup-ported by its position in projected phase space (Fig. 13) wherethe galaxy is just outside the caustic curves. Raj et al. (2019)found irregular star-forming regions in the ill-defined spiral armsof ESO 358-G063, which may be signs of minor mergers. NGC 1427A
NGC 1427A is classified as Im-III by Ferguson (1989). TheH I morphology shows a very long tail pointing to the south-east in projection (opposite direction to the centre of the clus-ter - we refer the reader to Lee-Waddell et al. (2018) for a de-tailed study about the origin of the tail). NGC 1427A has a M H I = (2.1 ± × M (cid:12) and a M (cid:63) = (2.3 ± × M (cid:12) . Thisis the second most massive H I galaxy of our sample, which hasan H I content comparable to that of non-cluster (see Fig. 6).NGC 1427A is the closest H I detection to the centre of thecluster (0.2 R vir ). Due to its high velocity (see Fig. 13) it isnot clear whether NGC 1427A is already virialised in the clus-ter or not. If it is virialised, it may decrease its velocity whilereaching its apocentre. It is the only M H / M H I upper limit with M (cid:63) > × M (cid:12) in Fig. 9. Despite the large H I reservoir noH . ex was detected by Zabel et al. (2019). This may be due toa recent merger which involved the NGC 1427A progenitorsthat might have lowered its metallicity. The H I -to-H . ex conversionmight be ine ffi cient to have a SFR consistent with its M H I / M (cid:63) ratio (Fig. 11). NGC 1365
NGC 1365, the large barred spiral galaxy (SBbc(s)I by Fergu-son 1989), is the most massive galaxy in our sample. The opticalmorphology shows a well defined spiral structure, three northernarms and two more compressed southern ones. The H I morphol- ogy (Fig. 2) shows a very prominent extension to the north ofthe system and more dense contours corresponding to the twocompressed optical arms. A very detailed study on the structureof NGC 1365 is made by Jorsater & van Moorsel (1995). Theyalso suggested that the motion of the galaxy through the ICMcan account for its unwinding arms.NGC 1365 has a M H I = (1.5 ± × M (cid:12) and a M (cid:63) = (6.2 ± × M (cid:12) . Fig. 6 shows that NGC 1365 is theonly galaxy in our sample which is at the upper edge of the com-parison sample of non-cluster galaxies due to its high M H I / M (cid:63) .Besides evidence of environmental interactions from its mor-phology, the strong gravitational potential appears to have beenable to hold a su ffi cient amount of H I to make it comparable withthe general behaviour of local galaxies (see also Fig. 9).NGC 1365 is the main member of the detected infalling sub-group of galaxies located to the south-west of the cluster. It isalso the only galaxy in the subgroup with a normal H I content.The existence of the NGC 1365 subgroup is also supported byDrinkwater et al. (2001a) which found NGC 1365 to be part of asubstructure.There is a H I hole in the centre of the system where Zabelet al. (2019) detected warped H . ex emission due to the centralbar. H I absorption was also investigated in Ondrechen & van derHulst (1989). Article number, page 18 of 19. Loni et al.: A blind ATCA H I survey of the Fornax galaxy cluster Appendix B:
Fig. B.1: Optical images of the eight H . ex -rich H I -undetected galaxies (Zabel et al. 2019) for which we evaluated the M H / M H I ratio(Fig.9). They are sorted according to increasing M H I upper limit. The g -band optical images come from the Fornax Deep Survey(Iodice et al. 2016; Venhola et al. 2018; Peletier et al. 2020). We show a 5 kpc scale bar in the bottom-right corner.-band optical images come from the Fornax Deep Survey(Iodice et al. 2016; Venhola et al. 2018; Peletier et al. 2020). We show a 5 kpc scale bar in the bottom-right corner.