GASP. XVI. Does cosmic web enhancement turn on star formation in galaxies?
Benedetta Vulcani, Bianca M. Poggianti, Alessia Moretti, Marco Gullieuszik, Jacopo Fritz, Andrea Franchetto, Giovanni Fasano, Daniela Bettoni, Yara L. Jaffe
MMNRAS , 1–19 (2018) Preprint 23 May 2019 Compiled using MNRAS L A TEX style file v3.0
GASP. XVI. Does cosmic web enhancement turn on starformation in galaxies?
Benedetta Vulcani, (cid:63) Bianca M. Poggianti, Alessia Moretti, Marco Gullieuszik, Jacopo Fritz, Andrea Franchetto, , Giovanni Fasano, Daniela Bettoni, Yara L. Jaff´e, INAF- Osservatorio astronomico di Padova, Vicolo Osservatorio 5, IT-35122 Padova, Italy Instituto de Radioastronom´ıa y Astrof´ısica, UNAM, Campus Morelia, A.P. 3-72, C.P. 58089, Mexico Dipartimento di Fisica & Astronomia “Galileo Galilei”, Universit`a di Padova, vicolo dell’ Osservatorio 3, IT 35122, Padova, Italy Instituto de F´ısica y Astronom´ıa, Universidad de Valpara´ıso, Avda. Gran Breta˜na 1111 Valpara´ıso, Chile
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
Galaxy filaments are a peculiar environment, and their impact on the galaxy prop-erties is still controversial. Exploiting the data from the GAs Stripping Phenomenain galaxies with MUSE (GASP), we provide the first characterisation of the spatiallyresolved properties of galaxies embedded in filaments in the local Universe. The fourgalaxies we focus on show peculiar ionised gas distributions: H α clouds have beenobserved beyond four times the effective radius. The gas kinematics, metallicity mapand the ratios of emission line fluxes confirm that they do belong to the galaxy gasdisk, the analysis of their spectra shows that very weak stellar continuum is associatedto them. Similarly, the star formation history and luminosity weighted age maps pointto a recent formation of such clouds. The clouds are powered by star formation, andare characterised by intermediate values of dust absorption. We hypothesise a scenarioin which the observed features are due to “Cosmic Web Enhancement”: we are mostlikely witnessing galaxies passing through or flowing within filaments that assist thegas cooling and increase the extent of the star formation in the densest regions inthe circumgalactic gas. Targeted simulations are mandatory to better understand thisphenomenon. Key words: galaxies:general — galaxies:evolution — galaxies: kinematics and dy-namics — galaxies: merger — galaxies: group
The properties of galaxies are directly affected by their hostenvironment. In the local universe, red, passive, early-typegalaxies are preferentially found in dense regions and galaxyclusters while blue, star-forming, late-type galaxies domi-nate in less dense, field environments (e.g., Dressler 1980;Kauffmann et al. 2004; Balogh et al. 2004).Studies of the effect of large-scale structure on galaxyproperties are usually mostly confined to field versus clus-ters. However, the intermediate environments such as galaxygroups, cluster outskirts and filaments are equally important(e.g., Kodama et al. 2001), as they host the vast majorityof galaxies in the local universe (e.g., Jasche et al. 2010;Tempel et al. 2014a,b; Cautun et al. 2014).In particular, filaments that connect groups and clus-ters of galaxies may contain up to 40 per cent of the matter (cid:63)
E-mail: [email protected] (BV) in the Universe (Forero-Romero et al. 2009; Jasche et al.2010). Theoretical studies (e.g., Cen & Ostriker 1999) havesuggested that about half of the warm gas in the Universe,presumably accounting for the low-redshift missing baryons(Fukugita et al. 1998; Viel et al. 2005), is hidden in fila-ments. Recently, Nicastro et al. (2018) observed highly ion-ized oxygen systems in regions characterized by large galaxyover-densities, supporting the prediction of warm gas in theextragalactic universe.Trying to dissect the role of these environments ongalaxy properties is therefore of extreme importance to shedlight on the processes that regulate galaxy evolution.Several mechanisms have been proposed to governgalaxy properties. Dark matter filaments can trap and com-press gas, shock heating the accreted gas at the boundary offilaments. This gas then cools rapidly and condenses into fil-aments centre. Filaments can therefore assist gas cooling andenhance star formation in their haloes (Liao & Gao 2018). Infilaments, mild galaxy-galaxy harassment and interactions © a r X i v : . [ a s t r o - ph . GA ] M a y B. Vulcani et al. (Lavery & Henry 1988; Moore et al. 1996; Coppin et al. 2012)are favored. Their environment is colder than clusters: thetypical temperature of filaments is ∼ − K (e.g., Cen& Ostriker 2006; Werner et al. 2008; Zappacosta et al. 2002;Nicastro et al. 2005), even though filaments must also con-tain cool gas (T ∼ K), as predicted by Lyman alpha forestobservations (e.g., Kooistra et al. 2017). Therefore, galaxiesin filaments can still hold their gas content to form stars.Nonetheless, for low mass galaxies ( M ∗ < M (cid:12) ), simula-tions show that filaments falling onto clusters are able toproduce increased stripping of hot gas even beyond a dis-tance of 5 r from a galaxy cluster centre and that this ispredominant at low-redshift (Bah´e et al. 2013), suppressingthe fuel for star formation. In filaments, ram-pressure strip-ping (Gunn & Gott 1972) is known as cosmic web strippingand is due to the interaction of the galaxies and the fila-ments, and might also play a role, especially for low massgalaxies, whose shallow potential wells can provide a rela-tively small restoring force from the ram-pressure force ofthe IGM in filaments (e.g., Ben´ıtez-Llambay et al. 2013).However, this effect has been never observed and the fateof the gas that remains in the galaxy or is accreted later isnot clear. The cosmic web stripping is so far a purely hydro-dynamical effect that requires simulations of large volumesable to resolve properly both the cosmic web and the internalhalo properties.Another process that is also expected to be quite effec-tive in filaments is gas accretion, which increases the avail-ability of cold gas for galaxies inducing an enhancement ofthe star formation (e.g., Darvish et al. 2014).Several works have shown that filaments affect the evo-lution of the integrated properties of galaxies (e.g., Koyamaet al. 2011; Geach et al. 2011; Sobral et al. 2011; Mahajanet al. 2012; Tempel & Libeskind 2013; Tempel et al. 2013;Zhang et al. 2013; Pintos-Castro et al. 2013; Koyama et al.2014; Santos et al. 2014; Malavasi et al. 2017; Mahajan et al.2018) and the distribution of satellites around galaxies (Guoet al. 2014), at any redshift, but results are still controversial.Overall, filament galaxies tend to be more massive, redder,more gas poor and have earlier morphologies than galaxiesin voids (Rojas et al. 2004; Hoyle et al. 2005; Kreckel et al.2011; Beygu et al. 2017; Kuutma et al. 2017). On the otherhand, some studies have reported an increased fraction ofstar-forming galaxies (Fadda et al. 2008; Tran et al. 2009;Biviano et al. 2011; Darvish et al. 2014), and higher metal-licities and lower electron densities (Darvish et al. 2015) infilaments with respect to field environments.Differences in the results might also be due to the dif-ferent techniques adopted by different teams to define fil-aments. Indeed, due to the observational biases in largegalaxy surveys and unvirialized nature of the large-scalestructures, their characterization is a nontrivial task andmany assumptions come into play (e.g., Biviano et al. 2011;Tempel et al. 2014a; Poudel et al. 2017).From the theoretical point of view, Gay et al. (2010)have investigated the influence of filaments on the spectro-scopic properties of galaxies, using the MareNostrum sim-ulation. They found that the large-scale filaments are onlydynamical features of the density field, reflecting the flow ofgalaxies accreting on clusters; the conditions in the filamentsare not dramatic enough to influence strongly the propertiesof the galaxies it encompasses. On the other hand, Aragon- Calvo et al. (2016) showed that the star formation quenchingin galaxies can be explained as the influence of filamentaryenvironment. So far, no studies have investigated how spa-tially resolved properties could be affected by filaments, fromneither an observational nor theoretical point of view, exceptfor our attempt in Vulcani et al. (2018a, Paper XII).In this paper we present the analysis of four field spiralgalaxies in the local universe showing asymmetric featuresand an extended H α distribution, proxy for extended H ii regions, that we will argue are most likely due to the effectof the hosting filaments.H ii regions signifying the presence of ionising OB starsare usually found in the luminous inner regions of galax-ies (see, e.g., Martin & Kennicutt 2001). The evidence ofstar formation in outer disks, instead, raises new questionsabout the nature of star formation in diffuse environments.Indeed, outer disks are usually considered inhospitable en-vironments for star formation. In fact, a deviation in theKennicutt-Schmidt Law (Kennicutt 1998b, 1989) has beenobserved at a gas surface density of 3-5 M (cid:12) pc − , where theH α intensity suddenly drops (but see Boissier et al. 2007,who suggest this is merely a stochastic effect). This is gen-erally interpreted as a threshold density for star formation(Kennicutt 1989; Martin & Kennicutt 2001), most likely dueto a transition between dynamically unstable and stable re-gions of the galaxy (e.g., Toomre 1964) or to a phase tran-sition of the gas (e.g., Elmegreen & Parravano 1994; Schaye2004; Krumholz et al. 2009).However, H α knots at large radii have been observed ina few galaxies (Kennicutt 1989; Martin & Kennicutt 2001;Ferguson et al. 1998) and ∼
30% of disk galaxies have UVemitting sources beyond their optical disks (Thilker et al.2005, 2007; Gil de Paz et al. 2005; Zaritsky & Christlein2007; Christlein & Zaritsky 2008). These complexes are oftencoincident with local H i over-densities (Ferguson et al. 1998).In M83 and NGC 4625, the UV knots have been identifiedas low-mass stellar complexes and, if visible in the H α , aregenerally ionised by a single star (Gil de Paz et al. 2007).These knots are dynamically cold and rotating, indicatingthat outer disk complexes are extensions of the inner disk(Christlein & Zaritsky 2008).Isolated H ii regions have also been discovered in theextreme outskirts of galaxy halos in the Virgo Cluster (Ger-hard et al. 2002; Cortese et al. 2004), in gaseous tidal debris(Ryan-Weber et al. 2004; Oosterloo et al. 2004) and in be-tween galaxies in galaxy groups (Sakai et al. 2002; Mendes deOliveira et al. 2004). These appear as tiny emission-line ob-jects in narrow-band images at projected distances up to 30kpc from the apparent host galaxy. These regions sometimesare associated with previous or ongoing galaxy interactions(Thilker et al. 2007; Werk et al. 2008).H α radiation has also been observed to be emitted bythe gaseous halos of nearby galaxies (Zhang et al. 2018).This emission is extremely faint (flux (cid:28) − erg / cm / s / ˚A)and has been observed up to several hundreds of kpc fromthe main galaxy.An explanation for the existence of these outer knotscould be that at some sites the gas density may exceed a starformation threshold locally, allowing stars to form beyondthe radius where the azimuthally averaged gas density isat or below a threshold density (Kennicutt 1989; Martin & MNRAS , 1–19 (2018) hysical processes in filaments Kennicutt 2001; Schaye 2004; Elmegreen & Hunter 2006; Gilde Paz et al. 2007).All of above studies are based on traditional observa-tional techniques, such as narrow-band imaging and Fabry-Perot staring technique. These techniques only permit thedetection and basic characterization of the H ii regions, with-out giving spatially resolved information on the chemicalcomposition and age of the regions.The galaxies we discuss in this paper instead are drawnfrom a Integral Field Spectrographs (IFS) survey that hasbeen designed to focus on the galaxy external regions andallows us to perform a detailed analysis of the galaxy out-skirts.GASP (GAs Stripping Phenomena in galaxies withMUSE), an ESO Large programme that exploits theintegral-field spectrograph MUSE mounted at the VLT withthe aim to characterise where, how and why gas can get re-moved from galaxies in different environments. A completedescription of the survey strategy, data reduction and anal-ysis procedures is presented in (Poggianti et al. 2017, PaperI). First results on single objects in clusters are discussedin Bellhouse et al. 2017, (Paper II); Fritz et al. 2017, (Pa-per III); Gullieuszik et al. 2017, (Paper IV); Moretti et al.2018, (Paper V); and in lower-density environments in Vul-cani et al. 2017, (Paper VIII); Vulcani et al. 2018b, (PaperVII); Paper XII.GASP includes a sample of galaxies selected for pre-senting a B-band morphological asymmetry suggestive ofunilateral debris (Poggianti et al. 2016) plus a subset ofundisturbed galaxies, used as control sample.Throughout all the papers of the GASP series, we adopta Chabrier (2003) initial mass function (IMF) in the massrange 0.1-100 M (cid:12) . The cosmological constants assumed are Ω m = . , Ω Λ = . and H = km s − Mpc − . In this paper, unless otherwise stated, we consider only theGASP galaxies selected from the field sample. All galax-ies are drawn from the Millennium Galaxy Catalog (Liskeet al. 2003; Driver et al. 2005) and selected from the PM2GC(Calvi et al. 2011). We exclude from the GASP sample inter-acting (e.g. Paper VIII) and passive (e.g, Paper XII) galax-ies, counter-rotating disks (e.g. Paper XII) and galaxies witha central H α hole (Moretti et al. in prep.).We compute the maximum extension of the H α distribu-tion in units of effective radius ( r e ). Specifically, we measurethe radius containing 99% of the H α flux having a S/N > α images used for selecting the galaxies aregiven in Sec 2.2 and 2.3.The effective radius, along with the inclination and theposition angle of the galaxies, were obtained from the analy-sis of the I-band images achieved from the integrated MUSEdatacubes on the Cousins I-band filter response curve, asexplained in Franchetto et al. (in prep.). Briefly, they wereobtained using ellipse (Jedrzejewski 1987) of the software http://web.oapd.inaf.it/gasp/index.html Figure 1.
Maximum extent of H α in units of r e (R(H α ) max )distribution in the GASP field sample. Black line and shadedarea show the median value and its error. The red line shows thethreshold used to select galaxies in this work. IRAF that allows an isophotal segmentation of the galaxyand draws the luminosity growth curve L ( R ) = π ∫ R I ( a ) ( − ε ( a )) a da , (1)where I ( a ) is the surface brightness profile, ε ( a ) is the isopho-tal ellipticity profile and a is the semi-major-axis of the el-liptical isophotes.Taking advantage of the ample sky coverage of theGASP data, we extended the fitting up to the most externalpart of the galaxies to probe the behaviour of the surfacebrightness at large radii. Sources extraneous to the galaxy,brighter knots -often located along the spiral arms- and badpixels were masked out to prevent erroneous measurements.Although the I-band image is obtained from sky-subtractedMUSE datacube, it presents residual sky intensity compara-ble to the intensities of last fitted isophotes. Thus, we sub-tracted the value of the intensity of the last isophote to theimage and proceeded with the computation of the luminositygrowth curve.Assuming that the galaxy regions over the largestisophote negligibly contribute to the total galaxy luminosity,we approximated L tot ≈ L ( a max ) - with a max semi-major-axisof the largest isophote - and we estimated the effective radiusas the radius R e such as L ( R e )/ L tot = . .From the surface brightness profile we selected theisophotes that trace the stellar disk to measure their meanposition angle ( PA ), the mean ellipticity ( ε ) and correspond-ing errors.The H α disk extension of the sample is shown in Fig. 1.The median H α disk extension is 3.1 ± r e .We then selected the galaxies with maximum H α extensionlarger than four r e , corresponding to 90th percentile. Fourgalaxies passed the selection and they are listed in Table 1,which summarises some important information that will befurther used and discussed throughout the paper.Figure 2 shows the color composite images of the tar- MNRAS , 1–19 (2018)
B. Vulcani et al.
Figure 2.
RGB (left) and H α (right) images of the targets. Fromtop to bottom P95080, P19482, P63661, and P8721 are shown.The reconstructed g , r , i filters from the MUSE cube have beenused. North is up, and east is left. Color map is inverted for dis-play purposes. In all the plots, the green ellipses show the r e , thedashed blue ellipses show the maximum radius at which H α isdetected (see text for details). Purple areas show the detachedclouds (see text for details). Asymmetries in the star and gasdistributions are seen in all galaxies, with one side of the galax-ies extending more than the other. All these galaxies have H α extending beyond 4 r e and show patchy H α distribution. gets, obtained combining the reconstructed g − , r − and i − filters from the MUSE datacube along with the H α maps.Overplotted in green are the r e , while overplotted in blueare the maximum radii at which H α is detected. For com-parison, Figure 3 shows the color composite images and H α maps of four representative galaxies of the control sample.Figures 1 and 6 in Vulcani et al. (submitted, Paper XX),show the images for all the galaxies in the GASP controlsample that will be also used in this paper (sec 3.5). Thepresence of detached clouds (highlighted in purple in Fig.2) in the H α disk of the selected galaxies is astonishing, es- Figure 3.
Same as Fig.2, but for four galaxies representative ofthe GASP control sample. In these galaxies H α extends at mostto 3 r e and the galaxy boundaries are not jagged. pecially if compared with the absence of the same featuresamong the control sample galaxies. The clouds extend be-yond the spiral arms of the galaxies, suggesting they mightnot strictly related to them. A quantitative identification ofthe clouds is deferred to the next Section. All the GASP targets were observed in service mode withthe MUSE spectrograph, mounted at the Nasmyth focus ofthe UT4 VLT, at Cerro Paranal in Chile. Each galaxy wasobserved with clear conditions; the seeing remained below0 . (cid:48)(cid:48) MNRAS , 1–19 (2018) hysical processes in filaments Table 1.
Properties of the targets. For each galaxy, the ID, redshift, coordinates, total stellar mass, effective radius, maximum extensionof H α , position angle, ellipticity and physical scale are given.ID z RA DEC log M r e R ( H α ) max PA (cid:15) phys. scale(J2000) (J2000) ( M ∗ / M (cid:12) ) ( (cid:48)(cid:48) ) ( r e ) (deg) kpc/ (cid:48)(cid:48) P90580 0.04038 198.03625 -0.23903 9.98 7.5 ± ± ± ± The data reduction process for all galaxies in the GASPsurvey is presented in Paper I. For all galaxies, we averagefiltered the datacubes in the spatial direction with a 5 × (cid:48)(cid:48) (see Paper I, for details).At the redshifts of the galaxies presented here, 1 (cid:48)(cid:48) that cor-responds to 0.8-1.2 kpc, depending on the redshift of thetarget. Paper I extensively presents the methods used to analysegalaxies within the GASP program. Here we just recallthe basic procedures and references useful for the follow-ing analysis. In brief, we corrected the MUSE reduced dat-acubes for extinction due to our Galaxy and then we mea-sured (1) the total fluxes and kinematic properties of thegas, by running the kubeviz (Fossati et al. 2016) code; (2)the kinematic properties of the stars, by running the Penal-ized Pixel-Fitting (pPXF) software (Cappellari & Emsellem2004), which works in Voronoi binned regions of given S/N(Cappellari & Copin 2003) and smoothed using the two-dimensional local regression techniques (LOESS) as imple-mented in the Python code developed by M. Cappellari; (3)the properties of the stellar populations, such as star forma-tion histories, luminosity and mass weighted ages, surfacemass densities, by running the spectral fitting code sinopsis (Paper III); (4) the dust extinction A V from the absorption-corrected Balmer decrement assuming an intrinsic H α /H β ratio equal to 2.86 and adopting the Cardelli et al. (1989)extinction law; and (5) ionised gas metallicity, by runninga modified version of the pyqz Python (Dopita et al. 2013)v0.8.2 (F. Vogt 2017, private communication).Further details will be discussed in the next section,where needed. In this section we characterise separately each of the targets.In the following sections we will highlight what these galaxieshave in common and look for the reasons of such similarities.
P95080, seen in the top panels of Fig.2, is a spiral galaxywith a moderate inclination and a possibly a bar. The rightpanel shows the MUSE map for H α , uncorrected for in-trinsic dust extinction, but corrected for stellar absorption and Galactic extinction. We plot only the spaxels with H α S/N > . The H α distribution is quite patchy in the cen-tral regions of the galaxy, where many peaks are visible. Itemerges that the ionised gas is characterised a number ofclouds detached from the main body that extend beyondthe visible light. These surround the galaxy without hav-ing a preferred orientation. As these clouds might be onlydue to spurious spaxels, we decided to plot only the spax-els that, in addition to having H α S/N > , are surroundedby spaxels with measured velocity at S/N >
3. Specifically,we build a 3 × σ ofthe mean velocity of the the galaxy , considering separatelythe approaching and receding sides. This approach helps toremove possible spurious signal.Given the redshift of P95080, the sky line at λ =6830˚Afalls very close to H α . Therefore, only for this galaxy, we alsoexclude all the spaxels in the clouds whose velocity is within ±
50 km s − the velocity of the sky line. We will clean also thefollowing plots adopting the same approach. We are there-fore confident that the clouds we detect are real and due tosome specific physical process. To identify the clouds usinga quantitative metrics, we select all H α regions in the lumi-nosity range − . − − . erg / s / cm / acrsec that have nopixels in common with the main body of the galaxy, have asize larger than 10 pixels and are within R(H α ) max . Withthis approach, we identify 32 clouds. In P95080, the mea-sured extension of the H α disk is 4.1 × the extension of thestellar disk, defined by r e .We note that the H α images shown in Fig. 3 have beenproduced following the same procedure, therefore clouds ofsimilar size and intensity would be detected. In contrast,the aforementioned approach does not identify any cloudsin the galaxies belonging to the control sample.The map of H α , when in combination to those of H β ,[OIII] 5007 ˚A, [OI] 6300 ˚A, H α , [NII] 6583 ˚A, and [SII]6716+6731 ˚A, can be used to determine the main ionisingsource at each position. The lines’ intensities are measuredafter subtraction of the continuum, exploiting the pure stel-lar emission best fit model provided by sinopsis , to takeinto account any possible contamination from stellar photo-spheric absorption. Only spaxels with a S / N > in all theemission lines involved are considered. All the diagnostic di-agrams (BPT, Baldwin et al. 1981) are concordant in find-ing that young stars produce the ionised gas (“Star-forming”according to Kauffmann et al. 2003; Kewley et al. 2006)throughout the galaxy and in excluding the presence of AGN MNRAS000
50 km s − the velocity of the sky line. We will clean also thefollowing plots adopting the same approach. We are there-fore confident that the clouds we detect are real and due tosome specific physical process. To identify the clouds usinga quantitative metrics, we select all H α regions in the lumi-nosity range − . − − . erg / s / cm / acrsec that have nopixels in common with the main body of the galaxy, have asize larger than 10 pixels and are within R(H α ) max . Withthis approach, we identify 32 clouds. In P95080, the mea-sured extension of the H α disk is 4.1 × the extension of thestellar disk, defined by r e .We note that the H α images shown in Fig. 3 have beenproduced following the same procedure, therefore clouds ofsimilar size and intensity would be detected. In contrast,the aforementioned approach does not identify any cloudsin the galaxies belonging to the control sample.The map of H α , when in combination to those of H β ,[OIII] 5007 ˚A, [OI] 6300 ˚A, H α , [NII] 6583 ˚A, and [SII]6716+6731 ˚A, can be used to determine the main ionisingsource at each position. The lines’ intensities are measuredafter subtraction of the continuum, exploiting the pure stel-lar emission best fit model provided by sinopsis , to takeinto account any possible contamination from stellar photo-spheric absorption. Only spaxels with a S / N > in all theemission lines involved are considered. All the diagnostic di-agrams (BPT, Baldwin et al. 1981) are concordant in find-ing that young stars produce the ionised gas (“Star-forming”according to Kauffmann et al. 2003; Kewley et al. 2006)throughout the galaxy and in excluding the presence of AGN MNRAS000 , 1–19 (2018)
B. Vulcani et al.
Figure 4.
BPT line-ratio diagram for [OIII]5007/H β vs [NII]6583/H α for the three galaxies. Lines are from Kauffmann et al. (2003,K03), Kewley et al. (2001, K01) and Sharp & Bland-Hawthorn (2010, SB10) to separate Star-forming, Composite, AGN and LINERS.In the inset the BPT line-ratio map is shown. Only spaxels with a S / N > in all the emission lines involved are shown. in the galaxy center (see the [OIII]/H β vs [NII]/H α plot inthe left panel of Fig.4, the other plots are not shown). Thisis in agreement with previous classifications found in theliterature for the same galaxy (e.g., V´eron-Cetty & V´eron2010).Since the ionisation source is mostly photoionisation byyoung stars, we can now measure the total ongoing SFR, ob-tained from the dust- and absorption-corrected H α luminos-ity adopting the Kennicutt (1998a)’s relation for a Chabrier(2003) IMF. Integrating the spectrum over the galaxy, weget a value of SFR=0.74 M (cid:12) yr − .Figure 5 presents the maps of other quantities obtainedfrom the MUSE datacubes. From left to right, top to bot-tom it shows the gas and stellar kinematics, the metallic-ity of the ionised gas, the extinction map, the luminosityweighted age and the surface mass density. Panels (a) and(b) show the H α velocity and velocity dispersion maps. Thegas is rotating around the North-south direction, the Eastside is receding, the West side is approaching. The velocityfield is quite regular and spans the range ( − < v/ km s − < ). The median error on the gas velocity in the spax-els is ∼ km s − . Uncertainties on the stellar motion are theformal errors of the fit calculated using the original noisespectrum datacube and have been normalized by the χ ofthe fit. The velocity of the detached clouds is that expectedgiven their position with respect to the galaxy, suggestingthat they indeed belong to the object. To further assess thevalues obtained for the velocity of the clouds, we integratedthe spaxels of each cloud and run kubeviz on the integratedspectra. Values obtained on the spatially resolved and inte-grated spectra are largely in agreement. The few spaxels inthe West region of the galaxy with velocity ∼ km s − areresiduals of the sky line emission discussed above and do notcarry any information.The velocity dispersion is overall low, having a medianvalue of 16 km s − . This is indicative of a dynamically coldmedium. The East side has a systematically higher velocitydispersion than the West one. All the clouds have a typicallylow velocity dispersion.Panels (c) and (d) of Fig. 5 show the stellar velocity andstellar velocity dispersion maps, respectively, for Voronoibins with S/N >
10. The velocity field of the stellar com-ponent of P95080 is similar to that of the gas, spanning a similar range ( − < v/ km s − < ), though less spatiallyextended. The median error in stellar velocity is ∼
50 km s − .The bending of the locus of zero-velocity is due to the pres-ence of the bar, as discussed in Erroz-Ferrer et al. (2015).Also the velocity dispersion of the stars is typically low ( < − , which is below the resolution limit). Deviations areseen in the eastern part of the galaxy, where larger errorsprevent us from drawing solid conclusions.Panel (e) of Fig. 5 presents the spatial distribution ofthe metallicity of the ionised gas, i.e. + log [ O / H ] . Onlythe spaxels with S/N is > + log [ O / H ] ∼ ) and then by a smooth decline towardsthe outskirts, which are characterised by + log [ O / H ] ∼ .P95080 lays on the typical mass-metallicity relation for lo-cal field galaxies (see Tremonti et al. 2004). Unfortunately,given the low S/N of some of the lines, we can not properlyconstrain the metallicity for the clouds.Panel (f) shows the A V maps for spaxels with aS/N(H α ) >
3. Overall, P95080 is characterised by low val-ues of extinction, almost always < mag. Relatively highervalues of extinction are found preferentially in the centralregions and trace one of the spiral arms of the galaxy.Panel (g) presents the map of the luminosity weightedages. This provides an estimate of the average age of thestars weighted by the light we actually observe, and givesan indication on when the last episode of star formationoccurred. The map shows that in the central regions thetypical luminosity weighted age of the galaxy is ∼ − . yr, and it decreases towards the outskirts, where it reachesvalues of ∼ yr. While in the center the distribution ofages is quite homogeneous, towards the outskirts it becomesmore patchy, showing many knots of younger ages. Typically,they corresponds to the H α blobs seen in the top right panelof Fig. 2.Finally, panel (h) shows the stellar mass density. Thevast majority of the mass is confined in the central partsof the galaxy, while the outskirts, and especially the cloudsaround the galaxy, are extremely less massive, reaching amass density of × M (cid:12) / kpc . The bar and two main spi-ral arms, already detected in the H α map, are seen also MNRAS , 1–19 (2018) hysical processes in filaments Figure 5.
P95080. The different panels show the MUSE map of H α (a), the H α velocity (b) and velocity dispersion (c) maps, stellarvelocity (d) and stellar velocity dispersion (e) maps, metallicity map for the ionised gas (f), A V maps (g), luminosity weighted age (h)and stellar mass density (i) maps. More details are given in the text. In all plots, (0, 0) is the center of the MUSE image. in the stellar mass density. Running sinopsis on the inte-grated spectra of the entire galaxy, we obtain a total M ∗ of9.5 × M (cid:12) .Given its values of SFR and stellar mass, P95080 lays onthe typical SFR-mass relation for star-forming field galaxies(Vulcani et al. 2018c, Paper XIV). We now focus on P19482, whose color composite image ispresented in the second row of Fig. 2. This is a spiral galaxy,with a slightly higher inclination than P95080. A spiral armextends towards South-West. We detect the presence of 42clouds, whose size is much larger than the typical size ofthe noise, seen e.g. in the corners of the image. Most of the
MNRAS000
MNRAS000 , 1–19 (2018)
B. Vulcani et al.
Figure 6.
P19482. Panels are as in Fig.5. detached clouds follow the spiral arms, but especially in theNorth -East region no stellar disk seems to be associated tothe presence of the clouds. The maximum extension of theH α disk is . × r e .Figure 6 present all the other properties of the galaxy.The analysis of the velocity field (panel (a)) indicates thatthese clouds belong to the galaxy, as their velocity is con-sistent with that of the part of the galaxy that is close tothem. As we did for P95080, we integrated the spaxels ofeach cloud and run kubeviz on the integrated spectra. Val- ues obtained on the spatially resolved and integrated spectraare largely in agreement.Overall, in each position, the gas and the stars (panel(c)) rotate around the same axis and at similar speed( − < v/ km s − < ). The median error on the gas veloc-ity is 4 km s − , the one on the stellar velocity is 50 km s − .The gas velocity dispersion (panel (b)) is overall < − . The median error on the gas velocity dispersion is ∼ − .The velocity dispersion of the stellar component (panel MNRAS , 1–19 (2018) hysical processes in filaments (d)) is overall quite low ( ∼
40 km s − ). In the South East re-gion it is systematically higher, reaching values of 80 km s − .The analysis of the diagnostic diagrams (central panelof Fig.4) does not detect the presence of an AGN in thegalaxy center. The emission-line ratios are consistent withgas being photoionised by young stars.The metallicity map of the ionised gas, presented inpanel (e), ranges from + log [ O / H ] = . in the outskirts to + log [ O / H ] = . in the core. This value sharply declinetowards the outskirts, where the typical metallicity is ∼ . .We can estimate the metallicity also in few clouds, findingthat this is consistent with the edges of the galaxy.P19482 presents low values of dust extinction (panel(f)) ranging from 0.2 mag in the outskirts and 2 mag in thecenter.The luminosity weighted age (panel (g)) strongly variesacross the galaxy. Maximum values are reached in the galaxycenter, where LWA ∼ yr, while moving towards the out-skirts, the luminosity weighted age progressively decreases,down to a minimum vlaue of LWA ∼ yr. As for P95080,most of the mass (panel (h)) is contained in the central partof the galaxy. Running sinopsis on the integrated spectra ofthe entire galaxy, we obtain a total M ∗ of 1.9 × M (cid:12) .The total ongoing SFR of P19482 is 1.3 M (cid:12) yr − , its po-sition on the SFR-mass relation is that expected for typicalfor star-forming field galaxies (Paper XIV). We now move our attention to P63661, whose RGB image ispresented in the third row of Fig. 2. Similarly to the previousgalaxies, this is a spiral galaxy, with a moderate inclination.A spiral unwinding arm extends towards South-West.The H α map shown in Figure 2 and Figure 7 unveils amuch more complicated structure. Focusing on the H α dis-tribution, we find that it extends well beyond the stellardisk. The maximum extension of the H α disk is . × r e .On the West side of the galaxy, the H α map presents arift. A portion of the gas is detached from the main body.We remind the reader that GASP data reach a surfacebrightness detection limit of V ∼ mag arcsec − and log ( H α [ erg s − cm − arcsec − ]) ∼ − . at the 3 σ confidencelevel (Paper I).On the North-East side of the galaxy, the H α distribu-tion is somehow broken off and presents a sharp edge.In addition, a number of smaller clouds surround thegalaxy for a total of 16. . From the analysis of the veloc-ity field (panel (a)) it results that these clouds belong tothe galaxy, as their velocity is consistent with that of thepart of the galaxy that is close to them. Running kubeviz on the integrated spectra of each cloud, we obtained valuescompatible with those obtained on the spaxels.Overall, in each position, the gas and the stars (panel(c)) rotate around the same axis and at similar speed( − < v/ km s − < ). In the external regions, where thereare no stars in correspondence of the gas, the gas reachesvelocities of | v | ∼
120 km s − . The median error on the gasvelocity is 4 km s − , the one on the stellar velocity is 40 km s − .The gas velocity dispersion (panel (b)) is overall < − , except in the core, where it reaches values of ∼
45 km s − . The median error on the gas velocity dispersion is < km s − .The velocity dispersion of the stellar component (panel(d)) is more chaotic, especially towards South-East, wherea spiral arm is present. Nonetheless, typical values do notexceed ∼ km s − .No central AGN is detected from the analysis of thediagnostic diagrams (central panel of Fig.4). The emission-line ratios are consistent with gas being photoionised byyoung stars. This finding confirms previous classifications(e.g., V´eron-Cetty & V´eron 2010).The metallicity map of the ionised gas is presented inpanel (e). The central part of the galaxy has a + log [ O / H ] > . This value sharply decline towards the outskirts, wherethe typical metallicity is ∼ . . From what we can infer fromthe significantly meaningful spaxels in the detached part ofthe galaxy towards West, the metallicity of the region issignificantly lower.The A V map (panel (f)) shows that overall P63661 ischaracterised by low values of extinction, almost always (cid:54) mag.The last two panels of Fig.7 show the properties ofthe stellar populations. The maximum luminosity weightedage of the galaxy (panel (g)) is found in the galaxy center:LWA ∼ yr. Moving towards the outskirts, the luminosityweighted age constantly decreases, to reach the minimumvalues in the detached region in the North-West side andtowards East, along the extension of one spiral arm. Mostof the mass (panel (h)) is contained in the central part ofthe galaxy. Beyond the R25 the stellar mass density reachesvalues of × M (cid:12) / kpc . Running sinopsis on the inte-grated spectra of the entire galaxy, we obtain a total M ∗ of1.8 × M (cid:12) .The total ongoing SFR of P63661 is 0.86 M (cid:12) yr − , itsposition on the SFR-mass relation is that expected for typ-ical for star-forming field galaxies (Paper XIV). P8721 is a spiral galaxy with a quite high inclination. Pro-jection effects can therefore affect the interpretation of theresults and possible detached clouds seen in projection mightappear as part of the galaxy.The ionised gas is much extended, especially towardsSouth-West. The maximum extension of the H α disk is . × r e . Strikingly, Fig. 2 shows that the stellar disk ex-tends mostly towards North-East with respect to the galaxycenter while the ionised gas disk extends mostly towardsSouth-west. It therefore appears that the light distributionin B and H α are distinct. A bow of bright H α knots is visiblein the Southern part of the galaxy. 12 clouds of detachedgas are visible both towards South-West and towards North-East. The velocity of the gas (panel (a) in Fig. 8) in theseclouds is consistent with them belonging to P8721.The gas velocity field is regular and spans the range -220 < v / km s − < ∼ − . The gas velocity dispersion, shown in panel (b)is overall quite low ( <
35 km s − ), except for the core andtwo external regions. The median error on the gas veloc-ity dispersion is < − . The right panel of Fig.4 showsthat while an AGN is not detected, the central region of the MNRAS , 1–19 (2018) B. Vulcani et al.
Figure 7.
P63661. Panels are as in Fig.5. galaxy has a composite spectrum, indicative of either shocksor old evolved stars.The stellar kinematics (panels (c) and (d)) is regular,except for a protuberance in the southern region, probablydue to a spiral arm, and similar to that of the gas in the samespatial position. The median error on the stellar velocity is ∼
25 km s − . The stellar velocity dispersion ranges from 40 km s − to 80 km s − . The metallicity of the ionised gas (panel(e)) is + log [ O / H ] ∼ . within R25 and then abruptlydecreases. It reaches minimum values in the tail towards South West. The two sides of the galaxy (SW and NE) showdifferent slopes of the gradients. No metallicity values arereliable for the gas in the clouds.The A V map (panel (f)) shows a peak of dust attenua-tion in the core of the galaxy ( A V ∼ . mag) and a declinetowards the outskirts, where it reaches values of ∼ . mag.The luminosity weighted age (panel (g)) is slightly olderwithin R25, with typical values around 10 . yr. Outside theR25, it has average values around 10 . yr. The youngest MNRAS , 1–19 (2018) hysical processes in filaments Figure 8.
P8721. Panels are as in Fig.5. region of the galaxy is found in the South-West part of theobject.The mass density map (panel (h)) shows that the bulkof the mass in located in the galaxy core ( M (cid:12) / kpc ), whilethe South-West part of the object gives a very little contri-bution to the total mass of P8721. Running sinopsis on theintegrated spectra of the entire galaxy, we obtain a total M ∗ of 5.6 × M (cid:12) .The total ongoing SFR is 1.05 M (cid:12) yr − , it therefore lays on the SFR-mass relation of star-forming galaxies in the field(Paper XIV). To further assess the peculiarity of the light distribution inthese galaxies, we run a number of non-parametric morpho-logical measurements, exploiting the python package stat-morph (Rodriguez-Gomez et al. 2018), on the images of thecontinuum underlying the H α (red continuum) and the H α MNRAS000
P8721. Panels are as in Fig.5. region of the galaxy is found in the South-West part of theobject.The mass density map (panel (h)) shows that the bulkof the mass in located in the galaxy core ( M (cid:12) / kpc ), whilethe South-West part of the object gives a very little contri-bution to the total mass of P8721. Running sinopsis on theintegrated spectra of the entire galaxy, we obtain a total M ∗ of 5.6 × M (cid:12) .The total ongoing SFR is 1.05 M (cid:12) yr − , it therefore lays on the SFR-mass relation of star-forming galaxies in the field(Paper XIV). To further assess the peculiarity of the light distribution inthese galaxies, we run a number of non-parametric morpho-logical measurements, exploiting the python package stat-morph (Rodriguez-Gomez et al. 2018), on the images of thecontinuum underlying the H α (red continuum) and the H α MNRAS000 , 1–19 (2018) B. Vulcani et al.
Figure 9.
The morphological parameters G, M20, C, A and S for the three galaxies presented in this paper (colored stars) and galaxiesof a GASP control sample visually selected for not having morphological distortions (Paper XIV)(black points). The histograms showthe distribution of the parameters. Left panels are based on the continuum underlying H α , right panels on the H α images. images, obtained from the fits of kubeviz . Specifically, wemeasured: • Concentration C : Ratio of the circular radius containing80 per cent ( r ) of a galaxy’s light to the radius containing20 per cent ( r ) of the light (Bershady et al. 2000; Conselice2003; Peth et al. 2016). A large concentration value indicatesa majority of light is concentrated at the center of the galaxy,i.e the presence of a bulge. • Asymmetry A : Difference between the image of a galaxyand the galaxy rotated by 180 degrees (Conselice et al. 2000;Peth et al. 2016). This determines a ratio of the amount oflight distributed symmetrically to all light from the galaxy.A large value of asymmetry indicates that most of the lightis not distributed symmetrically. • Gini Coefficient G : Measure of the equality of light dis-tribution in a galaxy (Lorenz 1905; Abraham et al. 2003;Lotz et al. 2004; Conselice 2014). A value of G = 1 is ob-tained when all of the flux is concentrated in a single pixel,avalue of G = 0 when the brightness distribution is homoge-neous. • M20 : Second order moment of the brightest regions ofa galaxy (Lotz et al. 2004) tracing the spatial distributionof any bright clumps. It is sensitive to bright structure awayfrom the center of the galaxy; flux is weighted in favor ofthe outer parts. It therefore is relatively sensitive to mergersignals and tidal structures, specifically star-forming regionsformed in the outer spiral or tidal arms. If no such structuresare in the image, the 20% brightest pixels will most likely beconcentrated in the center of the galaxy, which is weightedlower. Low values of M20 are obtained for smooth galaxieswith bright nucleus (Ellipticals, S0 or Sa), much higher val-ues (less negative) for galaxies with extended arms featuringbright H ii regions. • Smoothness S : Degree of small-scale structure (Con-selice 2003; Takamiya 1999). Larger values of S actually cor-respond to galaxies that are less smooth (i.e. more ‘clumpy’).Figure 9 compares the values of the morphological mea-surements of the three galaxies under inspection to those ofa GASP (field+cluster) control sample carefully selected fornot having morphological distortions (Paper XIV). We care-fully checked that all the galaxies of the control sample donot have any similar detached cloud. As far as the red con-tinuum is concerned, the four galaxies present relativelyhigh concentration values, slightly higher than the bulk ofthe control sample. They also present similar asymmetryand Gini values, indicating that stars are symmetrically andquite homogeneously distributed. Only P19482 is offset, in-dicating the presence of dishomogeneites also in the stellarcomponent. They do not stand out in the M20-Asymmetryand M20-Gini planes, excluding ongoing mergers for theseobjects (e.g. Lotz et al. 2008a,b). Considering H α , it emergesthat the star formation is less concentrated than the stars,both for these galaxies and the control sample, P8721 isone of the less concentrated objects of all galaxies. Mov-ing to asymmetry and smoothness, overall all galaxies arecharacterised by higher absolute values than for the stel-lar continuum. P8721 and P63661 really stand out in thesedistributions. P8721 is peculiar also in terms of Gini coeffi-cient, while the other two galaxies follow the control sampletrends. P19482 present low values of M20, confirming thepresence of many clumpy star forming regions spread acrossthe disk.Taken together, these results suggest that as for the redcontinuum, which traces the stars, galaxies are “normal”.Peculiarities with respect to a control sample of undisturbedgalaxies emerge when looking at the star forming regionsonly. The H α distribution is clumpy and with small scale MNRAS , 1–19 (2018) hysical processes in filaments structures. P8721 is the most peculiar object, followed byP63661 and P19482, while P95080 is more regular.This analysis therefore corroborates the previous anal-ysis based on visual morphology that all the three galax-ies are regular when the stellar light distribution is consid-ered. In almost all the cases, the evidence for anomalies isstronger when the H α images are analysed, suggesting thatthe ionised gas is the most disturbed component. In the previous section we have described the spatially re-solved properties of three galaxies that present peculiar lightdistributions and a number of common features. First of all,they are all characterised by a “tattered” H α distribution.They all present H α clouds beyond the stellar disk, up toseveral kpc. They are visible even in P8721, an unfavoredcase given its high inclination. These clouds have typicallya size of 3-5 kpc, but a detached region ∼ kpc long is vis-ible in P63661. Within the GASP sample, these, along withP5215 discussed in Paper XII, are the only galaxies present-ing such extended and peculiarly tattered gas distribution.According to the gas kinematics, these clouds do belongto the galaxy: their velocity is similar to that of the closestpart of the galaxy. Both the gas and the stellar kinematicsare regular and resemble each other. We can therefore ex-clude processes that involve a redistribution of the stellarorbits, such as mergers (see, e.g., Paper VIII) or processesstrongly affecting the gas distribution, such as strong rampressure stripping (see, e.g., Gunn & Gott 1972, Paper I).Simulations by e.g. Jesseit et al. (2007), Kronberger et al.(2007) have indeed studied the 2D kinematic analysis for asample of simulated binary disc merger remnants with differ-ent mass ratios, showing how merger remnants usually showa multitude of phenomena, such as heavily distorted veloc-ity fields, misaligned rotation, embedded discs, gas rings,counter-rotating cores and kinematic misaligned discs. Noneof these features are evident from our maps.At least other two pieces of evidence exclude that thegalaxies have undergone a recent merger. On one side, theBPT diagrams presented in Fig. 4 show that there areno signs of tidally induced shocks, associated with the in-teraction process, contributing to the ionization of the gas(Colina, Arribas & Monreal-Ibero 2005). Extended shockionization has been previously reported in local U/LIRGs(Monreal-Ibero, Arribas & Colina 2006; Rich et al. 2011;Rich, Kewley & Dopita 2014). In all these cases, shock ion-ization exhibits characteristics of extended low-ionizationnuclear emission-line region (LINER)-like emission withbroadened line profiles. In P95080, P19482 and P63661 wefind no broadened line profiles falling in the so-called com-posite region. Only P8721 has a few central spaxels charac-terised by composite spectra, but composite regions due tointeractions would be expected more in the outer parts ofthe galaxy.In addition, we see continuous distribution, not differ-ent sequences, confirming again that we are observing gasbelonging to one galaxy. Indeed, in cases of e.g. mergers orgas accretion we should see different line ratios indicative ofdifferent chemical abundances in the different regions of thegalaxies (see, e.g., Fig. 12 in Paper XII). On the other side, the observed metallicity distribution has a smooth gradient,suggesting that no strong process altered it considerably.The asymmetry of the metallicity gradient in P8721might however suggest that this galaxy is accreting lowmetallicity gas from the Southern side, similarly to whatpresented for another GASP galaxy in Paper VII, but noother pieces of evidence support this scenario. P8721 is alsothe most peculiar object when the morphological analysis onthe H α is performed, being at the tail of the distributions inall the quantities analysed.In addition to looking at the luminosity weighted agemaps, to better investigate the mode of growth of thesegalaxies (inside out or inside in) in Fig. 10 we presentthe galaxy spatially resolved star formation histories. Theseplots show the variation of the SFR across cosmic time infour age bins in such a way that the differences between thespectral characteristics of the stellar populations are max-imal (Fritz et al. 2007 and Paper III). Note that sinop-sis tends to include an unnecessary small percentage of old(t > . × ) stars when the spectra have a low signal-to-noise values. To be conservative, we neglect the contributionof stars older than . × yr in low S/N spectra (S/N < t > . × yr ago), the star formation mostly occurredwithin half of the current H α maximum disk and only inrecent epochs ( t < . × yr ago) the outer part of the diskstarted to form. The maximum spatial extension of the starforming disk is observed in the current age bin ( t < × yr). In P8721, instead, the SFR is relatively constant withtime overall in the galaxy and external regions might havebeen already forming stars even in the oldest age bin. Thisis another piece of evidence that distinguishes P8721 fromthe other galaxies.For comparison, Figure 11 shows the maps of SFR inthe oldest and in the youngest age bins for the four galaxiesin the control sample already presented in Fig.3. In thesegalaxies the spatial extension of the disk is very similar at thetwo ages, suggesting that not all galaxies are characterisedby strong inside-out growth.The next step to better understand the possible physicalmechanisms acting on these galaxies is to characterise theenvironment around them. To characterise the environments of the galaxies we havepresented in the previous section, we exploit two publiclyavailable catalogs. Both are based on the spectroscopic sam-ple of the galaxies of SDSS data release 10, complete down tom r = 17.77 mag. The first catalogue identifies galaxy groupsand clusters and was published by Tempel et al. (2014b).The second catalogue identifies galaxy filaments and waspublished by Tempel et al. (2014a).Table 2 presents some useful information regarding thegroups, Fig. 12 shows the position on the sky of the tar-gets and their surroundings. All the values are drawn fromTempel et al. (2014b). Besides detecting the filaments, Tem- MNRAS , 1–19 (2018) B. Vulcani et al.
Figure 10.
Stellar maps of different ages, illustrating the average star formation rate per kpc during the last × yr (left), between × yr and . × yr (central left), . × yr and . × yr (central right) and > . × yr ago (right), for P95080 (upper), P63661(central) P8721 (bottom). In all the plots, the green ellipses show the r e , the dashed blue ellipses show the maximum radius at whichH α is detected (see text for details). Table 2.
Properties of the groups hosting the galaxies. Values are taken from Tempel et al. (2014b). For each group, the redshift (z gr ),the coordinates (RA gr and DEC gr ), the number of group members (N gals , gr ), the virial radius R vir , gr , the mass of the halo both assuminga NFW and a Hernquist profile ( log M N FW halo , gr , log M Her halo , gr ) are given. P63661b is the bigger cluster close to the system of P63661.ID z gr RA gr DEC gr N gals , gr R vir , gr log M N FW halo , gr log M Her halo , gr (J2000) (J2000) (kpc) (M (cid:12) ) (M (cid:12) )P95080 0.04136 198.08969 -0.23002 3 315 12.63 12.84P19482 - - - - - - -P63661 0.05597 218.06171 0.17165 2 232 12.58 12.58P63661b 0.05612 217.49573 0.30415 32 493 13.77 13.99P8721 0.06609 158.51371 0.00926 4 276 11.61 11.81 MNRAS , 1–19 (2018) hysical processes in filaments Figure 11.
Same as 10, but for galaxies in the control sample. pel et al. (2014a) do not give any quantity useful to bettercharacterise these structures.None of these galaxies is located in massive clusters,and three of them are members of small (Milky-Way-like,or slightly more massive, with two or three bright members)groups that are embedded in filaments, while P19482 mostlikely does not have any close companion, but is still embed-ded in a filament.P95080 is part of a three-member group that is locatedin the center of a long filament of 32 galaxies. The galaxy isat 0.5 R vir , gr and its closest galaxy is at ∼
200 kpc (see Tab.3). P19482 is at the edges of a filament of 12 members andat the intersection among four different filaments all locatedat the same redshift (z ∼ ∼ . R vir , gr from the Table 3.
Distances of the galaxies from the center of their group,in unit of R vir , gr and distance of the closest galaxy, in kpc. ForP63661, the value in parenthesis gives the distance from the largergroup. ID d r d closest (kpc)P95080 0.50 193P63661 0.23 (4.84) 233P19482 - 1750P8721 0.39 165 center of the system. About 1 Mpc western P63661 a quitemassive group is found, with 32 members. The properties ofthis group are also listed in Tab.2. P63661 is found at 4.8R vir , gr from the center of this massive group. This group isat the center of a filament, which extends both towards NWand towards SE for several Mpc and includes 51 galaxies.Finally, P8721 is part of a four-member system, embed-ded in the center of an extended filament of 35 galaxies. It MNRAS000
Distances of the galaxies from the center of their group,in unit of R vir , gr and distance of the closest galaxy, in kpc. ForP63661, the value in parenthesis gives the distance from the largergroup. ID d r d closest (kpc)P95080 0.50 193P63661 0.23 (4.84) 233P19482 - 1750P8721 0.39 165 center of the system. About 1 Mpc western P63661 a quitemassive group is found, with 32 members. The properties ofthis group are also listed in Tab.2. P63661 is found at 4.8R vir , gr from the center of this massive group. This group isat the center of a filament, which extends both towards NWand towards SE for several Mpc and includes 51 galaxies.Finally, P8721 is part of a four-member system, embed-ded in the center of an extended filament of 35 galaxies. It MNRAS000 , 1–19 (2018) B. Vulcani et al.
Figure 12.
Position on the sky of the targets, represented by the stars. Filled red squares represent galaxies in groups, according toTempel et al. (2014b, T14, group). Empty circles represent galaxies in filaments, according to Tempel et al. (2014a, T14, filament).Dashed circles indicate the virial radius of the groups. The scale in the bottom right corner shows 1 Mpc at the redshift of each target.For P19482, the smaller points with different shades of blue show filaments intersecting the one hosting the galaxy. is at 0.3 R vir , gr from the group center and the closest of thetwo other galaxies of the group is at 165 kpc.Just for reference, we note that in the control sampleused here (Sec. 3.5) 13 out of the 14 field galaxies are eitherbinary or single systems, supporting the scenario that theenvironment might indeed play a role. Based on the defini-tion of filaments by Tempel et al. (2014a), 7/14 galaxies arein small filaments (less than 25 objects), while the othersare at the boundaries of filamentary structures.To understand whether the perturbed morphology ofthe galaxies are the result of tidal interactions with theirclosest neighbors, we follow a crude approach that was al-ready exploited by, e.g., Wolter et al. (2015); Merluzzi et al.(2016) and estimate the acceleration a tid produced by theclosest neighbour on the ISM of the galaxy of interest andcompare it with the acceleration from the potential of thegalaxy itself, a gal . Following Vollmer et al. (2005), a tid a gal = M neighbour M gal (cid:16) rR − (cid:17) − where R is the distance from the centre of the galaxy, r isthe distance between the galaxies (Vollmer et al. 2005), and M neighbour M gal the stellar mass ratio. This formulation wouldrequire the true distance, that we obviously do not have,therefore we can only use the projected distance. In all thethree cases, a tid a gal << and we can exclude tidal interactionswith the other group members.Another source of perturbation to the galaxy morphol-ogy might be their position within the filament. Galaxies infilaments are indeed expected to have a very different ex-perience from those in largely empty regions (Bah´e et al.2013).All the three galaxies analysed have the major axis moreor less aligned to the filament they are embedded in. Theycould therefore either be flowing along the filament or cross-ing it perpendicularly. In filaments the IGM density is en-hanced, rising the possible ram pressure intensity. In par-ticular, galaxies with shallow potential wells can provide arelatively small restoring force, and a significant gas strip-ping can take place at typical gas densities and velocities(e.g., Ben´ıtez-Llambay et al. 2013). The clouds we observe could therefore actually be the result of a galaxy crossing afilament.Analytically quantifying the effect of the filament is notstraightforward and also simulations have never been ableto quantify the impact of this environment on the spatiallyresolved properties of the galaxies. Accurately measure thedensity of the IGM in these environments and estimate the3D center of the filament needed to quantify the distanceof the galaxy from it is indeed a quite hard task. It is how-ever tantalizing to suppose filaments are responsible for theobserved gas distribution. We might therefore be witnessingthe cosmic web stripping acting on galaxies more massivethan dwarfs (Ben´ıtez-Llambay et al. 2013).However, stripping requires a relatively high velocitydifference between the galaxy and the filament and thegalaxies simulated by Ben´ıtez-Llambay et al. (2013) werelow mass objects, while the galaxies discussed here have log ( M ∗ / M (cid:12) ) (cid:38) . So, rather than stripping, we are mostlikely seeing gas compression due to the flowing of the galax-ies within the filaments. This compression can be induced byan increase in surrounding thermal pressure and can switchon the surrounding clouds. Numerical simulations by Liao& Gao (2018) show that filaments can assist the gas cool-ing and increase the star formation in their residing darkmatter haloes. As a consequence, it might be possible thatthe densest regions in the circumgalactic gas get switchedon in their star formation when the galaxy impacts withthe sparse IGM. The detached clouds observed around thegalaxies with no preferential orientation might be indeed anevidence for this phenomenon, that we call “Cosmic web en-hancement”. In Vulcani et al. (2018a, Paper XII) we havepresented another galaxy with similar features and found ina similar environment.Nonetheless, there are no simulations specifically focus-ing on the spatial properties of galaxies in filaments. Devel-oping this kind of simulations is now urgent to better inves-tigate this peculiar environment and its effect on galaxies.Indeed, different conditions of the filaments (density,extent, orientation), as well as the inclination of the galaxywith respect to the filament itself, could also have differentimpacts on the embedded galaxies, and this could explain MNRAS , 1–19 (2018) hysical processes in filaments the differences observed in P8721 with respect to the othertwo galaxies. ii regions in the literature As mentioned in the Introduction, in the literature few stud-ies have identified H α knots at large radii (Kennicutt 1989;Martin & Kennicutt 2001; Ferguson et al. 1998) or isolated(Gerhard et al. 2002; Cortese et al. 2004; Ryan-Weber et al.2004; Oosterloo et al. 2004; Sakai et al. 2002; Mendes deOliveira et al. 2004).All of above studies are based on traditional observa-tional techniques to observe the H α emission in the out-skirts of galaxies. The most exploited one is narrow-bandimaging with subsequent subtraction of broad-band contin-uum emission. This is however generally insufficiently sen-sitive to probe large radii. The limitation lies both in theachievable signal-to-noise ratio (S/N) and in the stellar con-tinuum subtraction. Higher spectral resolution is generallypreferable and very narrow filter bandpasses have also beenadopted, as also traditional spectroscopy, that however canhave quite low throughput. Since the pioneering work ofBland-Hawthorn et al. (1997), also the Fabry-Perot staringtechnique has been used.These techniques however only permit the detection andbasic characterization of the H ii regions, without giving spa-tially resolved information on the chemical composition andage of the regions. This is now possible thanks to the re-cent advent of Integral Field Spectrographs (IFS). However,the known on-going large IFS surveys like the Calar AltoLegacy Integral Field Area (CALIFA) Survey (S´anchez et al.2012), the Sydney-AAO Multi-object Integral field spectro-graph (SAMI) Survey (Croom et al. 2012), the MappingNearby Galaxies at Apache Point Observatory (MaNGA)Survey (Bundy et al. 2015) typically reach out to 2.5-3 ef-fective radii at most (Bundy et al. 2015), therefore are notdesigned to catch dis-homogeneities in the ionised gas dis-tribution in the galaxy outskirts.GASP has been instead designed to focus on the galaxyexternal regions and allows us to perform a detailed analysisof the galaxy outskirts.To put our results in context, we have directly com-pared the observed H α distributions of P95080, P19482,P63661 and P8721 to those of many other literature re-sults, and confirmed that they present many peculiarities,hardly found in previous studies. Galaxies characterisedby narrow band image surveys like the H α Galaxy Survey(H α GS, Shane et al. 2002), the H α galaxy survey (Jameset al. 2004), the H-alpha Galaxy Groups Imaging Survey(H α ggis, PI. Erwin), An H α Imaging Survey of Galaxiesin the Local 11 Mpc Volume (11Hugs Robert C. Kenni-cutt et al. 2008), Dynamo (Green et al. 2014), or Fabry-Perot observations like the Gassendi H α survey of SPirals(GHASP Epinat et al. 2008) almost never present suchextended and luminous ( log ( H α [ erg / s / cm / arcsec ] > − . )H α regions located well beyond R25. This might be due tothe shallower surface brightness reached: the surface bright-ness limit of the observations of James et al. (2004) is SB(H α + [ N ]) = − erg / cm / s / arcsec . In addition, at least some ofthese surveys (e.g., Epinat et al. 2008) targeted galaxies inthe cluster environment, such as Virgo, and those observed features are most likely due to the ram pressure exerted bythe hot intracluster medium (Gunn & Gott 1972).The detached H α regions we detect are quite bright andlarge and we can exclude the galaxies are found in clusters.These regions are also much brighter than the emissionproduced by the gaseous haloes and are similar to the in-tergalactic H ii regions discovered by (e.g. Ryan-Weber et al.2004) in terms of H α luminosity. However, the latter are notalways bound to the main galaxy, while all the clouds wediscussed present compatible velocities and ionised gas andstellar properties. Ryan-Weber et al. (2004) results are con-sistent with stars forming in interactive debris as a resultof cloud-cloud collisions, while no signs of interactions areevident from our analysis.Unfortunately, no high resolution UV data are currentlyavailable for the three galaxies. They have been observedwith GALEX, but these data are too shallow to detect anydetached material. GASP (GAs Stripping phenomena in galaxies with MUSE)is an ESO Large Program with the MUSE/VLT to studythe causes and the effects of gas removal processes in galax-ies in different environments in the local universe. Withinthe sample, we identified four galaxies that show peculiarionised gas distributions: several H α clouds have been ob-served beyond 4 r e . The gas kinematics, metallicity mapand the ratios of emission line fluxes (BPT diagrams) con-firm that they do belong to the galaxy gas disk, the stellarkinematics shows that very weak stellar continuum is as-sociated to them. Similarly, the star formation history andluminosity weighted age maps point to a recent formation ofsuch clouds, as also of more than half of the stellar disk forP95080, P19482 and P63661. The clouds are powered bystar formation, and are characterised by intermediate valuesof extinction ( A V ∼ . − . ). These, along with an objectdiscussed in Paper XII, are the only three galaxies in all theGASP non cluster sample showing such tattered H α distri-bution, and we have not found any similar object in thecurrently existing literature surveys.The three galaxies share a similar location in the Uni-verse: they all belong to filamentary structures, therefore wepoint to a scenario in which the observed features are due to“Cosmic web enhancement”: we hypothesize that we are wit-nessing galaxies passing through or flowing within filamentsthat are able to cool the gas and increase the star forma-tion in the densest regions in the circumgalactic gas. Ob-served differences among the three galaxies might be due tothe different conditions of the filaments. Liao & Gao (2018)showed that filaments are an environment that particularlyfavors this gas cooling followed by condensation and starformation enhancement.In the recent years, there has been an increasing inter-est for the role of filaments in affecting galaxy properties,nonetheless, to our knowledge, this paper presents the firstanalysis of the effect of this environment on the spatiallyresolved properties of the galaxies, highlighting the impor-tance of this kind of data to get insights on galaxy evolutionas a function of environment. Targeted simulations illustrat-ing the effect of filaments on galaxy properties are now cru- MNRAS , 1–19 (2018) B. Vulcani et al. cial to make progress on the physical processes acting in thedifferent environments.
ACKNOWLEDGEMENTS
Based on observations collected at the European Organisa-tion for Astronomical Research in the Southern Hemisphereunder ESO programme 196.B-0578. We acknowledge fund-ing from the INAF PRIN-SKA 2017 program 1.05.01.88.04(PI Hunt). We acknowledge financial contribution from thecontract ASI-INAF n.2017-14-H.0 Y. J. acknowledges sup-port from CONICYT PAI (Concurso Nacional de Inserci´onen la Academia 2017) No. 79170132 and FONDECYT Ini-ciaci´on 2018 No. 11180558.
REFERENCES
Abraham R. G., van den Bergh S., Nair P., 2003, ApJ, 588, 218Aragon-Calvo M. A., Neyrinck M. C., Silk J., 2016, preprint,( arXiv:1607.07881 )Bah´e Y. M., McCarthy I. G., Balogh M. L., Font A. S., 2013,MNRAS, 430, 3017Baldwin J. A., Phillips M. M., Terlevich R., 1981, PASP, 93, 5Balogh M. L., Baldry I. K., Nichol R., Miller C., Bower R., Glaze-brook K., 2004, ApJ, 615, L101Bellhouse C., et al., 2017, ApJ, 844, 49Ben´ıtez-Llambay A., Navarro J. F., Abadi M. G., Gottl¨ober S.,Yepes G., Hoffman Y., Steinmetz M., 2013, ApJ, 763, L41Bershady M. A., Jangren A., Conselice C. J., 2000, AJ, 119, 2645Beygu B., Peletier R. F., van der Hulst J. M., Jarrett T. H.,Kreckel K., van de Weygaert R., van Gorkom J. H., Aragon-Calvo M. A., 2017, MNRAS, 464, 666Biviano A., Fadda D., Durret F., Edwards L. O. V., Marleau F.,2011, A&A, 532, A77Boissier S., et al., 2007, ApJS, 173, 524Bundy K., et al., 2015, ApJ, 798, 7Calvi R., Poggianti B. M., Vulcani B., 2011, MNRAS, 416, 727Cappellari M., Copin Y., 2003, MNRAS, 342, 345Cappellari M., Emsellem E., 2004, PASP, 116, 138Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345, 245Cautun M., van de Weygaert R., Jones B. J. T., Frenk C. S.,2014, MNRAS, 441, 2923Cen R., Ostriker J. P., 1999, ApJ, 514, 1Cen R., Ostriker J. P., 2006, ApJ, 650, 560Chabrier G., 2003, PASP, 115, 763Christlein D., Zaritsky D., 2008, ApJ, 680, 1053Conselice C. J., 2003, ApJS, 147, 1Conselice C. J., 2014, ARA&A, 52, 291Conselice C. J., Bershady M. A., Jangren A., 2000, ApJ, 529, 886Coppin K. E. K., et al., 2012, ApJ, 749, L43Cortese L., Gavazzi G., Boselli A., Iglesias-Paramo J., 2004, A&A,416, 119Croom S. M., et al., 2012, MNRAS, 421, 872Darvish B., Sobral D., Mobasher B., Scoville N. Z., Best P., SalesL. V., Smail I., 2014, ApJ, 796, 51Darvish B., Mobasher B., Sobral D., Hemmati S., Nayyeri H.,Shivaei I., 2015, ApJ, 814, 84Dopita M. A., Sutherland R. S., Nicholls D. C., Kewley L. J.,Vogt F. P. A., 2013, ApJS, 208, 10Dressler A., 1980, ApJ, 236, 351Driver S. P., Liske J., Cross N. J. G., De Propris R., Allen P. D.,2005, MNRAS, 360, 81Elmegreen B. G., Hunter D. A., 2006, ApJ, 636, 712Elmegreen B. G., Parravano A., 1994, ApJ, 435, L121Epinat B., et al., 2008, MNRAS, 388, 500 Fadda D., Biviano A., Marleau F. R., Storrie-Lombardi L. J.,Durret F., 2008, ApJ, 672, L9Ferguson A. M. N., Wyse R. F. G., Gallagher J. S., Hunter D. A.,1998, ApJ, 506, L19Forero-Romero J. E., Hoffman Y., Gottl¨ober S., Klypin A., YepesG., 2009, MNRAS, 396, 1815Fossati M., Fumagalli M., Boselli A., Gavazzi G., Sun M., WilmanD. J., 2016, MNRAS, 455, 2028Fritz J., et al., 2007, A&A, 470, 137Fritz J., et al., 2017, ApJ, 848, 132Fukugita M., Hogan C. J., Peebles P. J. E., 1998, ApJ, 503, 518Gay C., Pichon C., Le Borgne D., Teyssier R., Sousbie T., De-vriendt J., 2010, MNRAS, 404, 1801Geach J. E., Ellis R. S., Smail I., Rawle T. D., Moran S. M., 2011,MNRAS, 413, 177Gerhard O., Arnaboldi M., Freeman K. C., Okamura S., 2002,ApJ, 580, L121Gil de Paz A., et al., 2005, ApJ, 627, L29Gil de Paz A., et al., 2007, ApJ, 661, 115Green A. W., et al., 2014, MNRAS, 437, 1070Gullieuszik M., et al., 2017, ApJ, 846, 27Gunn J. E., Gott III J. R., 1972, ApJ, 176, 1Guo H., et al., 2014, MNRAS, 441, 2398Hoyle F., Rojas R. R., Vogeley M. S., Brinkmann J., 2005, ApJ,620, 618James P. A., et al., 2004, A&A, 414, 23Jasche J., Kitaura F. S., Li C., Enßlin T. A., 2010, MNRAS, 409,355Jedrzejewski R. I., 1987, MNRAS, 226, 747Kauffmann G., et al., 2003, MNRAS, 346, 1055Kauffmann G., White S. D. M., Heckman T. M., M´enard B.,Brinchmann J., Charlot S., Tremonti C., Brinkmann J., 2004,MNRAS, 353, 713Kennicutt Jr. R. C., 1989, ApJ, 344, 685Kennicutt Jr. R. C., 1998a, ARA&A, 36, 189Kennicutt Jr. R. C., 1998b, ApJ, 498, 541Kewley L. J., Heisler C. A., Dopita M. A., Lumsden S., 2001,ApJS, 132, 37Kewley L. J., Groves B., Kauffmann G., Heckman T., 2006, MN-RAS, 372, 961Kodama T., Smail I., Nakata F., Okamura S., Bower R. G., 2001,ApJ, 562, L9Kooistra R., Silva M. B., Zaroubi S., 2017, MNRAS, 468, 857Koyama Y., Kodama T., Nakata F., Shimasaku K., Okamura S.,2011, ApJ, 734, 66Koyama Y., Kodama T., Tadaki K.-i., Hayashi M., Tanaka I.,Shimakawa R., 2014, ApJ, 789, 18Kreckel K., et al., 2011, AJ, 141, 4Krumholz M. R., McKee C. F., Tumlinson J., 2009, ApJ, 699, 850Kuutma T., Tamm A., Tempel E., 2017, A&A, 600, L6Lavery R. J., Henry J. P., 1988, ApJ, 330, 596Liao S., Gao L., 2018, preprint, ( arXiv:1805.10944 )Liske J., Lemon D. J., Driver S. P., Cross N. J. G., Couch W. J.,2003, MNRAS, 344, 307Lorenz M. O., 1905, Publications of the American Statistical As-sociation, Volume 9, Number 70, p.˜209-219, 9, 209Lotz J. M., Primack J., Madau P., 2004, AJ, 128, 163Lotz J. M., Jonsson P., Cox T. J., Primack J. R., 2008a, MNRAS,391, 1137Lotz J. M., et al., 2008b, ApJ, 672, 177Mahajan S., Raychaudhury S., Pimbblet K. A., 2012, MNRAS,427, 1252Mahajan S., Singh A., Shobhana D., 2018, MNRAS, 478, 4336Malavasi N., et al., 2017, MNRAS, 465, 3817Martin C. L., Kennicutt Jr. R. C., 2001, ApJ, 555, 301Mendes de Oliveira C., Cypriano E. S., Sodr´e Jr. L., BalkowskiC., 2004, ApJ, 605, L17 MNRAS , 1–19 (2018) hysical processes in filaments Merluzzi P., Busarello G., Dopita M. A., Haines C. P., SteinhauserD., Bourdin H., Mazzotta P., 2016, MNRAS, 460, 3345Moore B., Katz N., Lake G., Dressler A., Oemler A., 1996, Nature,379, 613Moretti A., et al., 2018, MNRAS, 475, 4055Nicastro F., et al., 2005, Nature, 433, 495Nicastro F., et al., 2018, Nature, 558, 406Oosterloo T., Morganti R., Sadler E. M., Ferguson A., van derHulst T., Jerjen H., 2004, in Duc P.-A., Braine J., Brinks E.,eds, IAU Symposium Vol. 217, Recycling Intergalactic andInterstellar Matter. p. 486 ( arXiv:astro-ph/0310632 )Peth M. A., et al., 2016, MNRAS, 458, 963Pintos-Castro I., et al., 2013, A&A, 558, A100Poggianti B. M., et al., 2016, AJ, 151, 78Poggianti B. M., et al., 2017, ApJ, 844, 48Poudel A., Hein¨am¨aki P., Tempel E., Einasto M., Lietzen H.,Nurmi P., 2017, A&A, 597, A86Robert C. Kennicutt J., Lee J. C., Jos´e G. Funes S. J., SakaiS., Akiyama S., 2008, The Astrophysical Journal SupplementSeries, 178, 247Rodriguez-Gomez V., et al., 2018, preprint, ( arXiv:1809.08239 )Rojas R. R., Vogeley M. S., Hoyle F., Brinkmann J., 2004, ApJ,617, 50Ryan-Weber E. V., et al., 2004, AJ, 127, 1431Sakai S., Kennicutt Jr. R. C., van der Hulst J. M., Moss C., 2002,ApJ, 578, 842S´anchez S. F., et al., 2012, A&A, 538, A8Santos J. S., et al., 2014, MNRAS, 438, 2565Schaye J., 2004, ApJ, 609, 667Sharp R. G., Bland-Hawthorn J., 2010, ApJ, 711, 818Sobral D., Best P. N., Smail I., Geach J. E., Cirasuolo M., GarnT., Dalton G. B., 2011, MNRAS, 411, 675Takamiya M., 1999, Ap&SS, 269, 339Tempel E., Libeskind N. I., 2013, ApJ, 775, L42Tempel E., Stoica R. S., Saar E., 2013, MNRAS, 428, 1827Tempel E., Stoica R. S., Mart´ınez V. J., Liivam¨agi L. J., CastellanG., Saar E., 2014a, MNRAS, 438, 3465Tempel E., et al., 2014b, A&A, 566, A1Thilker D. A., et al., 2005, ApJ, 619, L79Thilker D. A., et al., 2007, ApJS, 173, 538Toomre A., 1964, ApJ, 139, 1217Tran K.-V. H., Saintonge A., Moustakas J., Bai L., GonzalezA. H., Holden B. P., Zaritsky D., Kautsch S. J., 2009, ApJ,705, 809Tremonti C. A., et al., 2004, ApJ, 613, 898V´eron-Cetty M.-P., V´eron P., 2010, A&A, 518, A10Viel M., Branchini E., Cen R., Ostriker J. P., Matarrese S., Maz-zotta P., Tully B., 2005, MNRAS, 360, 1110Vollmer B., Huchtmeier W., van Driel W., 2005, A&A, 439, 921Vulcani B., et al., 2017, ApJ, 850, 163Vulcani B., et al., 2018a, MNRAS, 480, 3152Vulcani B., et al., 2018b, ApJ, 852, 94Vulcani B., et al., 2018c, ApJ, 866, L25Werk J. K., Putman M. E., Meurer G. R., Oey M. S., Ryan-WeberE. V., Kennicutt Jr. R. C., Freeman K. C., 2008, ApJ, 678,888Werner N., Finoguenov A., Kaastra J. S., Simionescu A., DietrichJ. P., Vink J., B¨ohringer H., 2008, A&A, 482, L29Wolter A., Esposito P., Mapelli M., Pizzolato F., Ripamonti E.,2015, MNRAS, 448, 781Zappacosta L., Mannucci F., Maiolino R., Gilli R., Ferrara A.,Finoguenov A., Nagar N. M., Axon D. J., 2002, A&A, 394, 7Zaritsky D., Christlein D., 2007, AJ, 134, 135Zhang Y., Yang X., Wang H., Wang L., Mo H. J., van den BoschF. C., 2013, ApJ, 779, 160Zhang H., Zaritsky D., Behroozi P., 2018, ApJ, 861, 34 This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000