Datasets:

SouthernCrossAI
/

Arxiv

Dataset card Data Studio Files Files and versions Community

Split (1)

train · 132k rows

Filename stringlengths 22 64	Paragraph stringlengths 8 5.57k
Processed_Metallicity_distributions_in_and_around_galaxies.txt	As I will show later, the distribution and amount of heavy elements in the diﬀerent bayonic components of the Universe, is sensitive to physical mechanisms that regulate the evolution of bayons as a function of cosmic time. Therefore, the main motivation of chem- ical abundance studies is the possibility to use this information as a tool to reconstruct the past history of star formation, and the relative role of feedback pro- cesses and various gas-dynamical processes in deter- mining the observed cosmic evolution of baryons (for a classical review on chemical evolution, see Tinsley 1980; for a more recent review focused on the Milky Way and its satellites, see Matteucci 2008).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	As mentioned above, metals are present in all bary- onic phases and ‘environments’. Even in rather low density regions, the inter-galactic medium appears to be polluted to & 1 per cent solar metallicity, as mea- sured by metal lines in the Lyman alpha absorption systems (see section 5). How and when this chemical enrichment occurred exactly, is still matter of debate. From the theoretical point of view, various physical processes can provide viable explanations for the trans- fer of metals from the stars in galaxies to the inter- galactic and intra-cluster medium, for example ejec- tion of enriched material during mergers, tidal/ram- supernova pressure stripping, galactic outﬂows (e.g.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	driven outﬂows). The relative importance of these var- ious channels, and the details of how metals are trans- ferred from the stellar to the gaseous (cold and/or hot) phase remain to bee understood.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Various techniques have to be used to measure the metal abundances in diﬀerent baryonic components. Stellar population synthesis models have been exten- sively used for nearby galaxies (in particular for ellip- tical galaxies and, in a few cases where the contamina- tion of stellar absorption features by nebular emission can be removed, for late-type galaxies). Optical neb- ular emission lines have been used to measure the gas phase metallicity of star forming galaxies up to z ∼ 1. The same method has been used for Lyman break and UV-selected galaxies up to z ∼ 3. X-ray spectroscopy is necessary to measure the metal content of the intra- cluster medium (ICM). As I will mention in the follow- ing, each of these techniques has its own problems and limits, and at least some of them are far from being ﬁrmly calibrated.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Chemical enrichment studies have a quite long his- tory that cannot be summarised in a few pages. Given the vastity of the subject, I have deliberately decided to focus mainly on recent observational results and the- oretical developments, rather than providing an histor- ical review. In the next sections, in particular, I will provide a brief overview of what is known, observa- tionally, about the metallicity distributions in our own Galaxy (focusing on the stellar halo), in the local Uni- verse and in the ICM and IGM. I will then brieﬂy touch upon the evolution of the mass-metallicity relation and metallicity measurements at higher redshift. In each section, I will mention recent theoretical progress, and I will conclude with a few remarks and considerations about future prospects.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Our own galaxy - the Milky Way - is a fairly large spiral galaxy consisting of four main stellar compo- nents. Most of the stars are distributed in a thin disk, exhibit a wide range of ages, and are on high angu- lar momentum orbits. A much smaller mass of stars (about 10-20 per cent of that in the thin disk) reside in a distinct component which is referred to as the ‘thick disk’. The stars in the thick disk are old, have on av- erage lower metallicity than those of similar age in the thin disk, and are on orbits of lower angular momen- tum. The Galactic bulge is dominated by an old and relatively metal-rich stellar population with a tail to low abundances. The fourth component - the stellar halo - represents only a tiny fraction of the total stel- lar mass (∼ 2 × 109 M⊙) and is dominated by old and metal poor stars which reside on low angular momen- tum orbits (Freeman & Bland-Hawthorn 2002).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	matics have been collected for a large number of indi- vidual stars in our own Galaxy. Historically, chemical and kinematic information provided the basis for the ﬁrst galaxy formation models. Eggen, Lynden-Bell & Sandage (1962) studied a sample of 221 dwarf stars and found that those with lowest metal abundance were moving on highly elliptical orbits and had small angular mo- menta. The data were interpreted as evidence that the oldest stars in the galaxy were formed out of gas col- lapsing from the halo onto the plane of the galaxy, on a relatively short time-scales (a few times 108 years). About one decade later, Searle & Zinn (1978), found no radial abundance gradient in a sample of 177 red giants and 19 globular clusters. These observations led Searle & Zinn to formulate the hypothesis that the stellar halo (in particular the outer region of the halo) formed through the agglomeration of many subgalactic fragments, that may be similar to the surviving dwarf spheroidal satellites (dSphs) of the Milky Way.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	lites might be intrinsically diﬀerent from the main con- tributors to the stellar halo. The argument has been put forward in a recent study by Font et al. (2006) who reproduced the observed chemical abundance pat- tern by combining mass accretion histories of galaxy- size haloes with a chemical evolution model for indi- vidual satellites. In the model proposed by Font and collaborators, the main contributors to the stellar halo were accreted early on and were enriched in α elements by Type II supernovae, while the surviving satellites were accreted later and have more extended star for- mation histories and stellar populations enriched to solar level by both Type II and Type Ia supernovae (see also De Lucia & Helmi 2008).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 2: From Kirby et al. (2008). Cumula- tive metallicity distribution function for the metal- poor tails of the eight ultra-faint dSphs (black) and the Milky Way halo (red). See original paper for more details.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	who found a signiﬁcant diﬀerence between the metal- poor tail of the dSph metallicity distribution and that of the Galactic halo. In a more recent study, however, (2008) have presented metallicity mea- Kirby et al. surements for 298 individual red giant branch stars in eight of the least luminous dSphs of the Milky Way, and have shown that the metallicity distribution of their stars is similar to that of the stellar halo at the metal-poor end (see Fig. 2).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Detailed abundance data for many elements are now being collected for larger samples of stars in the Milky Way and in several dSph galaxies. It is im- portant to note that each galaxy shows a wide spread in metallicity and that, in some cases, there are clear abundance gradients. It remains to be understood if more luminous dSphs (like those studied by Helmi and collaborators) also host extremely metal-poor stars. Ongoing spectroscopic surveys and detailed compar- isons with cosmologically motivated models including star formation and chemical enrichment, will provide in the next future crucial information on the role of dSphs in building the stellar halo of our own Milky Way.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	For galaxies outside our Local Group, metallicity mea- surements of individual stars are unfeasible, and only integrated values can be derived. Estimates of the ages and metallicities of the stellar population of nearby galaxies can be inferred from the integrated spectra, using stellar population synthesis models. These stud- ies are, however, limited by the diﬃculty of deriving independent constraints on the age, star formation his- tory, metallicity and dust content of a galaxy. These degeneracies can be broken, at some level, using spec- tral diagnostics that are not sensitive to dust and have diﬀerent sensitivities to age and metallicity. Studies in this area have focused, in particular, on a set of absorp- tion features deﬁned and calibrated using the spectra of nearby stars observed at the Lick Observatory - the so called Lick Indexes (Faber 1973; Worthey et al. 1994). These studies are generally limited to early type galaxies because of the lack of hot stars in the Lick library. It is important to note, however, that the age-metallicity degeneracy can be broken only in a rel- ative way, and that the absolute values of the ages and metallicities derived depend strongly on the speciﬁc choice of metal indices.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 3: From Gallazzi et al. (2005). Condi- tional distribution of stellar metallicity as a func- tion of the galaxy mass. The solid line indicates the median of the distribution and the dashed lines the 16th and 84th percentiles. See original paper for details.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	two-dimensional likelihood distributions for all galax- ies. The ﬁgure shows that metallicity increases with stellar mass. The 68 per cent conﬁdence levels, how- ever, show a quite broad distribution, with a large (larger than the intrinsic uncertainties) scatter for a given stellar mass.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	phase metallicity and the stellar mass by Tremonti et al. (2004), and is reproduced in Fig. 4. Gas phase metal- licities are usually estimated using strong nebular emis- sion lines, and calibrations that are based on metallic- ity sensitive emission line ratios. Fig. 4 shows the vast majority of galaxies having super-solar metallicity1. It should be noted, however, that strong-line methods (as used in Tremonti et al.) may systematically overesti- mate oxygen abundances by as much as 0.2-0.5 dex (Kennicutt, Bresolin & Garnett 2003). Diﬀerent cali- brations have been adopted in the literature: methods based on photoionization models, empirical methods based on the measurements of the electron tempera- ture of the gas, or a combination of these. Unfortu- nately, estimates obtained using diﬀerent calibrations manifest large discrepancies, and can produce mass- metallicity relations with diﬀerent shapes, y-axis oﬀ- sets, and scatter (see Kewley & Ellison 2008). This complicates comparisons between diﬀerent samples and between samples at diﬀerent redshifts, for which dif- ferent suites of emission lines are available.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	three decades in stellar mass. Tremonti and collabo- rators interpret the shape of the mass-metallicity rela- tion in terms of eﬃcient galactic outﬂows that remove metals from galaxies with shallow potential wells, an idea that was originally introduced by Larson (1974). Alternative explanations have been proposed: e.g. a lower star formation eﬃciency in low-mass galaxies (Calura et al. 2009), or a variable integrated initial mass function (K¨oppen, Weidner & Kroupa 2007).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	There is a striking similarity between the two rela- tions shown in Figs. 3 and 4. Gallazzi and collabora- tors have compared their stellar metallicity measure- ments with measurements of gas-phase metallicity, for those galaxies for which both measurements are avail- able. They have shown that there is a clear correla- tion between these quantities, with approximately unit slope, but there is a large scatter in stellar metallicity at ﬁxed gas-phase metallicity. In a closed box model, these two quantities should be tightly related. There- fore, these ﬁndings indicate that gas ejection and/or accretion play an important role in galaxy chemical evolution, and that their inﬂuence likely varies with stellar mass.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 5: From Thomas et al. (2005). Stellar population parameters as a function of velocity dispersion and stellar mass for early type galax- ies (red symbols are ellipticals, and blue symbols for lenticular). See original paper for details.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	short time-scales (of the order of a few Myrs), whereas the iron peak elements are mainly produced by super- novae Type Ia, on much longer time-scales. Therefore, the [α/Fe] ratio can be used as an indicator of the star formation time-scale. Since the early 90s, it has been known that brighter elliptical galaxies host stellar populations that are α enhanced with respect to their lower mass counterparts (Worthey, Faber & Gonzalez 1992). Fig. 5 shows stellar population parameters for early-type galaxies in diﬀerent environments, as a func- tion of the galaxy velocity dispersion and stellar mass. There is a clear trend for increasing [α/Fe] ratios with increasing stellar mass, both in low and high density regions. This trend suggests that the time-scale of star formation correlates with the galaxy stellar mass, and appears to be in contrast with naive expectations based on the growth of dark matter haloes in hierar- chical CDM cosmologies. Recent models that take into account the suppression of late gas cooling (and hence star formation) by radio-mode AGN feedback, how- ever, have shown that the observed behaviour is not in contradiction with the hierarchical paradigm of struc- ture formation (e.g. De Lucia et al. 2006). In these models, the shape of the mass metallicity relation ap- pears to be particularly sensitive to the adopted feed- back model (De Lucia, Kauﬀmann & White 2004; Bertone, De Lucia & Thomas 2007), conﬁrming that these observations provide strong constraints on how and when baryons and metals were ejected from galaxies.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Simulations have shown that the baryon fraction in a rich cluster does not evolve appreciably during its evolution. Clusters of galaxies can thus be consid- ered as closed systems, retaining all information about their past star-formation and metal production histo- ries. This suggests that direct observations of element abundances in the intra-cluster medium (ICM) can constrain the history of star formation in clusters, the eﬃciency with which gas was converted into stars, the relative importance of diﬀerent types of supernovae, and the mechanisms responsible for the ejection and the transport of metals. For all elements synthesised in stars after the primordial nucleosynthesis, the ener- gies corresponding to their K- and L-shell transitions are accessible to modern X-ray telescopes. Most of the observed emission lines can be converted to elemental abundances under the (reasonable) assumption of col- lisional equilibrium. Complications due to extinction, non-equilibrium ionization, optical depth eﬀects, etc. are not signiﬁcant, while the abundance measurements depend crucially on the correct modelling of the tem- perature structure (Werner et al. 2008, and references therein).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 6: From Simionescu et al. (2008). Map of the iron abundance in M87. See original paper for details.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	vations have also shown that the typical metallicity in the ICM appears to decrease by about 50 per cent to z ∼ 1 (Balestra et al. 2007). Cool core clusters (with peaked X-ray emission at their centres due to condensed regions of cooler gas) usually exhibit cen- trally peaked metallicity distributions, while ﬂat dis- tributions of metals are usually observed for non cool- core clusters. Observations of element abundances are consistent with the idea that the excess of metals in the central regions of cool core clusters is produced by supernovae Ia associated with the central cD galaxy (Tamura et al. 2001; De Grandi et al. 2004). If met- als originated from the stars produced in the central galaxy, the abundance proﬁle would follow the light proﬁle in absence of mixing. The observed metal pro- ﬁles are, however, shallower than the corresponding light proﬁles, suggesting that metals are mixed into the ICM and transported towards the outer regions by some physical process(es).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Deep observations of bright clusters with XMM- Newton and Chandra allow the construction of maps of the 2D distribution of metals in the ICM. Used to- gether with maps of thermodynamic properties, these maps can provide important information about the merging history of the cluster. Fig. 6 shows the map of Fe abundance in M87. The half light radius of M87 is shown as a black circle while the 90 cm radio emis- In addition to a sion contours are shown in white. radial gradient, which is usually observed in the cen- tral regions of galaxy clusters, the map in Fig. 6 sug- gests the presence of deviations associated with the inner radio lobes. These deviations are interpreted by Simionescu et al. (2008) as a result of the inﬂuence of the central AGN on the spatial distribution and trans- port of metals.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 7: From Baumgartner et al. (2005). Iron abundance (with respect to solar) as a function of the ICM temperature. Each point is a single cluster measurement, while the lines show where the boundaries were placed for stacking individual measurements. See the original paper for details.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	have been carried out in order to quantify their relative importance (for a recent review, see Schindler & Diaferio 2008). One physical process that is expected to be more important in clusters than in average regions of the Universe is ram-pressure stripping: galaxies trav- elling through a dense intra-cluster medium suﬀer a strong pressure that sweeps cold gas (and the metals in it) out of the stellar disc. Hydrodynamical simula- tions, however, suggest that this process can account only for ∼ 10 per cent of the overall observed level of enrichment in the ICM within a radius of 1.3 Mpc (Domainko et al. 2006). As noted by Renzini (1997), there is also a direct observational evidence that ram- pressure cannot play a dominant role in the chemical enrichment of the ICM because this process would op- erate more eﬃciently in high velocity dispersion clus- ters. A positive correlation between the richness of the cluster and its metal content is not supported by observations, as shown in Fig. 7.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	green line) and diﬀerent implementations of AGN feed- back (red and cyan lines). In these simulations, pro- ﬁles from diﬀerent runs have similar central slopes. This happens because the eﬀect that diﬀerent feedback mechanisms have in displacing enriched gas and regu- lating star formation almost balance each other in the central regions. In the outer regions (at cluster-centric distances & 0.2 R200), the metal abundance proﬁle in runs including AGN feedback (red and cyan lines) be- comes ﬂatter with respect to the proﬁle obtained in a run with only galactic winds (green line), resulting in a better agreement with observational measurements. The ﬂattening in the outer regions is due to the fact that AGN feedback displaces large amounts of enriched gas from star forming regions at high redshift and, at the same time, eﬃciently suppresses cooling at low red- shift.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Another obvious source of chemical enrichment is provided by supernova driven winds, which are ex- pected to play a dominant role in the outer regions of galaxy clusters, where they can travel to larger dis- tances due to the lower ICM pressure. On the ob- servational side, evidence in support of the existence of outﬂows from galaxies has grown rapidly in the last years. Powerful galactic winds appear to be ubiquitous in local starburst galaxies and in high-redshift Lyman Break galaxies (Heckman 2002).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Diﬀerent methods to estimate the outﬂow rate sug- gest that this is of the same order of magnitude of the star formation rate, with outﬂow speeds varying in the range 400-800 km/s, and approximately independent of the galaxy circular velocity. Observations also in- dicate, however, that galactic winds are complex and multi-phase outﬂows of cool, warm, and hot gas, dust, and magnetized relativistic plasma (Veilleux, Cecil & Bland-Hawthorn 2005). In addition, it is important to remember that observational estimates of outﬂow rates should not be confused with the rates at which mass and metals es- cape from galaxies, because they are derived from ma- terial that is still quite deep within the gravitational potential of the galaxy’s dark matter halo. Whether the gas reheated by supernova explosions will leave the halo depends on a number of factors like the amount of intervening gas, and the fraction of energy lost by radiative processes. Given these uncertainties and the dynamical complexity of the outﬂows, an accurate im- plementation of the feedback process in hydrodynam- ical simulations is still very diﬃcult, and needs to rely on ‘sub-grid’ physics (i.e. this physical process is not modelled from ﬁrst principles, but ‘implemented’ us- ing theoretically and/or observationally motivated re- lations, as is done in semi-analytic models).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 8: From Fabjan et al. (2009). Compari- son between the observed and simulated proﬁles of emission-weighted Iron metallicity. Black sym- bols with error bars show observational measure- ments for galaxy clusters with T500 > 3 kev. Dif- ferent lines correspond to diﬀerent simulation runs: no feedback (blue dashed), galactic winds (green dashed), and diﬀerent AGN feedback implemen- tations (red dot-dashed and cyan solid). See the original paper for details.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	one per cent solar) in the Lyman alpha forest absorbers (e.g. Songaila & Cowie 1996). More recent studies have indicated that heavy elements are present at all densities, down to at least the mean cosmic density (Schaye et al. 2003) and that, on small scales, the distribution of metals in the IGM is highly inhomo- geneous.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Lyman alpha systems (DLAs) have received particular attention (for a review, see Wolfe, Gawiser & Prochaska 2005). These objects are characterised by HI column densities N (HI) & 2 × 1020 cm−1, and hydrogen ap- pears to be mainly neutral in these systems, implying that they represent important neutral gas reservoir for star formation at high redshift. A nice review of re- cent metallicity measurements in DLAs is provided by Pettini (2006). Data show that the mean metallicity of these systems is relatively low (about one tenth so- lar or lower), but that individual measurements span a wide range of metallicities, from solar to less than one hundredth solar. A few measurements of element abundance patters are available, and suggest that, at least some DLAs, have [α/Fe] ratios lower than those measured in Milky Way stars of low metallicities. This has been interpreted as an indication of low levels of star formation (e.g. Matteucci & Recchi 2001).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Several recent studies have tried to constrain the cosmic evolution of the metal content of the IGM through measurements of the CIV mass density, ΩCIV. Fig. 9 is from a recent study by D’Odorico et al. 2009 and shows a signiﬁcant increase of ΩCIV at z . 3 and a downturn at z > 5 (Becker, Rauch & Sargent 2009; Ryan-Weber et al. 2009). A satisfactory theoretical interpretation for the trends shown in Fig. 9 is still missing. The most recent theoretical study in this area is that by Oppenheimer & Dav´e 2008. These au- thors have used cosmological simulations which include feedback in the form of momentum driven winds, and found an apparent lack of evolution of the CIV mass density from z ∼ 5 to z ∼ 1.5. In the simulations, this results from an overall increase in the cosmic abun- dance of carbon, and a parallel reduction of the frac- tion of triply ionized carbon. None of the simulations presented in this study predict the recently observed increase of ΩCIV towards lower redshift.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 9: From D’Odorico et al. (2009). Mass density of C IV as a function of redshift. The lowest redshift data point is from Cooksey et al. 2009, solid triangles are from Pettini et al. 2003 and Ryan-Weber et al. 2009. Finally, the 95 per cent conﬁdence interval at the highest redshift con- sidered are from Becker, Rauch & Sargent 2009.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	volumes of the IGM at high redshift. Numerical results are in qualitatively good agreement with observational measurements, but not without problems. For exam- ple, in these simulations the metallicity distributions of DLAs (as traced by the velocity widths) and the ob- served recent increase of ΩCIV (as mentioned above) are not well reproduced, suggesting that either the feedback scheme adopted is too simplistic/incorrect, or that there are some physical mechanisms not included in the simulations (e.g. enrichment by Population III stars, turbulence, etc.).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 10: From Mannucci et al. (2009). Evo- lution of the mass-metallicity relation from z =0.07 (Kewley & Ellison 2008), to z =0.7 (Savaglio et al. z =2.2 2005), (Erb et al. 2006), and z = 3 − 4 (Maiolino et al. 2008; All data have been 2009). Mannucci et al. calibrated to the same metallicity scale and IMF. The lines show quadratic ﬁts to the data (see original papers for details).
Processed_Metallicity_distributions_in_and_around_galaxies.txt	A (now incomplete) summary of our current knowl- edge about element abundances at z ∼ 2.5 is provided in Fig. 11 from Pettini (2006). This ﬁgure shows the metallicities measured in diﬀerent systems at high red- shift, as a function of the typical linear scale of the structure to which it refers. The ﬁgure shows a clear trend of decreasing abundances with increasing scale. The gas in the inner 10-100 pc of the highest overdensi- ties, where supermassive black holes reside, is enriched to solar or supersolar levels. Actively star-forming galaxies sample physical scales of the order of a few kpc, and are enriched to near solar metallicity. DLAs likely sample more diﬀuse gas at the outskirts of galax- ies in the process of formation and, as mentioned ear- lier, they represent a quite heterogeneous population with a large scatter in the degree of chemical enrich- ment. Finally, the Lyman α forest is believed to trace large scale structures of moderate overdensity relative to the cosmic mean, and it appears to contain only trace amounts of metals. Fig. 11 shows that, at any cosmic epoch, the degree of metal enrichment depends strongly on the scale considered.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 11: From Pettini (2006). Snapshot of the metallicity of diﬀerent components of the high red- shift universe. The logarithm of the metallicity is plotted relative to solar (indicated by the long- dash line at 0.0) as a function of the typical linear scale of the structure to which it refers.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	comoving density of metals which should have accu- mulated by a given redshift. As discussed by Pettini (2006), when adding together the contributions shown in Fig. 11, one can account for about 40 per cent of the metal production from the Big Bang to z = 2.5. The contribution from sub-DLAs (not included in the calculation by Pettini) amounts to about 6 per cent of the expected metals at z ∼ 2.5 (Peroux et al. 2007). The missing metals problem has been revisited by Bouch´e et al. (2007) who estimated that currently known galaxy populations at z ∼ 2 can account for & 30 to . 60 per cent of the metals expected, which is approximately what is required to close the metal bud- get. As noted by Bouch´e and collaborators, however, it is unclear what the overlap is between the various galaxy populations at this redshift. A conservative es- timate provided by these authors, is that we can cur- rently account for & 65 per cent of the metals expected at z ∼ 2.5, the metals being equally spread between galaxies and IGM.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	simulations show that the metal distribution in this phase depends signiﬁcantly on the details of feedback (Tornatore et al. 2009), therefore providing an addi- tional tool to study the physical processes responsible for the chemical enrichment of the Universe.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	As discussed above, metals are widespread in our Uni- verse and diﬀerent techniques have to be used to mea- sure their amounts in diﬀerent baryonic components (stars, cold inter-stellar medium, hot intra-cluster gas, inter-galactic medium). At least for some of these tech- niques, there are still large uncertainties on the obser- vational calibration, which complicates the comparison between measurements at diﬀerent redshift and/or in diﬀerent bands.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Fig. 12 shows the evolution of the metal content in diﬀerent phases for one particular feedback scheme implemented in a semi-analytic code by De Lucia et (2004). The left panel corresponds to a cluster al. simulation while the right panel is for a region with average density. Comparison between the two panels shows that cosmic variance eﬀects are going to play an important role when trying to constrain diﬀerent feedback schemes using observational metallicity mea- surements. Nevertheless, the same study pointed out (and many later studies conﬁrmed) that the fraction of metals (as well as their distribution) in diﬀerent baryonic components varies signiﬁcantly when diﬀer- ent feedback schemes are adopted. It is therefore to be expected that the vast amount of data that will be available in the next future, and detailed comparisons with theoretical models will eventually tell us a great deal about how galaxies ejected their metals over the history of the Universe.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Much work is going on to improve the precision of the observational measurements and to push element abundance studies to higher and higher redshift. Stel- lar population synthesis models have been improved signiﬁcantly in recent years: many new empirical li- braries have been made available with improved spec- tral resolution and parameter coverage (e.g. S´anchez-Bl´azquez et al. 2006), and many libraries of theoretical high-resolution stellar spectra for both scaled-solar and α enhanced chemical mixtures have been published (Coelho et al. 2007, and references therein). Parallel eﬀort is go- ing in the direction of reﬁning photoionization mod- els (such as MAPPINGS - Dopita et al. 2005) which help us correctly diagnose physical properties of star forming galaxies. Accurate measurements of metal- licities for much larger samples of individual stars for our Galaxy and its satellites will be available in the coming years from a number of ongoing astrometric the satellite Gaia - and spectroscopic surveys (e.g. Perryman et al. 2001; the Sloan Extension for Galac- tic Understanding and Exploration - Beers et al. 2004). Future X-ray satellites will likely push the study of the distribution of metals in the ICM to larger radii in nearby clusters, and allow metallicity proﬁles to be measured at higher redshift (see e.g. Giacconi et al. 2009). High-resolution spectroscopy in the X-ray and in the UV (e.g. with the Cosmic Origin Spectrograph onboard the Hubble Space Telescope) will also allow us to make progress in the study of the diﬀuse warm-hot baryons in the local Universe.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	I thank the organisers of the conference “Galaxy Metabolism: Galaxy Evolution near and far” (Sydney, 22-26 June 2009) for ﬁnancial support which helped me take part in this very stimulating meeting. I acknowledge ﬁ- nancial support from the European Research Coun- cil under the European Community’s Seventh Frame- work Programme (FP7/2007-2013)/ERC grant agree- ment n. 202781. It is a pleasure to thank Stefano Bor- gani, Francesca Matteucci, Luca Tornatore and Paolo Tozzi for many stimulating discussions about chem- ical enrichment, and for useful comments on a pre- liminary version of this manuscript. I am grateful to Anna Gallazzi and Christy Tremonti for many clariﬁ- cations about observational data analysis, and many useful discussions.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	From the theoretical viewpoint, hydrodynamical simulations and semi-analytic models provide the in- terpretative framework for these observational stud- ies. Many theoretical predictions are already available but not always these studies have been able to provide a ‘self-consistent’ description of the circulation of the baryons in diﬀerent components. Indeed, cosmologi- cal hydrodynamical simulations have mostly focused on the distribution of metals in the IGM and ICM, but the numerical resolution adopted is insuﬃcient to study the metal distribution in galaxies. On the other hand, semi-analytic models can provide detailed pre- dictions on the distribution of metals in the stellar and gaseous components but do not provide details on the spatial distribution of the metals.
Processed_Metallicity_distributions_in_and_around_galaxies.txt	Figure 12: From De Lucia et al. (2004). Evolution of the metal content in diﬀerent phases. The solid line represents the evolution of the metal content in stars, the dashed line the cold gas, the dashed-dotted line the hot gas, and the long-dashed line the ejected component. The metal content in each phase is normalised to the total mass of metals produced from all galaxies considered. The left panel shows the results for galaxies within the virial radius of a massive (∼ 1015 M⊙) cluster, while the right panel shows the results for a typical ‘ﬁeld’ region. See original paper for details.
Processed_Molecular_Hydrogen_in_the_Lagoon:_H2_line_emission.txt	The 2.12 µm v=1–0 S(1) line of molecular hydrogen has been imaged in the Hourglass region of M8. The line is emitted from a roughly bipolar region, centred around the O7 star Herschel 36. The peak H2 1–0 S(1) line intensity is 8.2 × 10−15 erg s−1 cm−2 arcsec−2. The line centre emission velocity varies from −25 km s−1 in the SE lobe to +45km s−1 in the NW lobe. The distribution is similar to that of the CO J=3–2 line. The H2 line appears to be shock-excited when a bipolar outﬂow from Herschel 36 interacts with the ambient molecular cloud. The total luminosity of all H2 lines is estimated to be ∼ 16 L⊙ and the mass of the hot molecular gas ∼ 9 × 10−4 M⊙ (without any correction for extinction).
Processed_Molecular_Hydrogen_in_the_Lagoon:_H2_line_emission.txt	reminiscent of the proplyds in the Orion Nebula—a star within a bow-shock arc, whose apex is pointed towards Herschel 36 (Stecklum et al. 1998). It is presumed to be an externally ionized circumstellar disk. These authors also estimate the distance to M8 to be 1.8 kpc, based on the association with NGC 6530.
Processed_Molecular_Hydrogen_in_the_Lagoon:_H2_line_emission.txt	Of particular interest to this paper are the observations by White et al. (1997) of intense CO line emission from M8. They found the peak CO J=3–2 intensity to be over 100 K, making it the second brightest CO line source known. The CO J=3–2 and 4–3 lines were mapped and found to have a loose, bipolar structure, extending NW–SE from Herschel 36. Taking the broad CO line proﬁles (extending over 20 km s−1), and the presence of a jet-like object extending 0.5′′ SE of Herschel 36 in HST images (Stecklum et al. 1995), it seems likely that there is a molecular outﬂow in the core of M8. If so, it would be expected for there to be molecular hydrogen line emission as well, both shocked (from the deceleration of the outﬂow by the ambient cloud) and ﬂuorescent (excited by far–UV photons from Herschel 36). However, White et al. (1997) also searched for the near–IR H2 v=1–0 S(1) line in M8, but did not detect it. In this paper we report more sensitive observations for H2 emission from M8, and ﬁnd that there is indeed excited H2 line emission from this source.
Processed_Molecular_Hydrogen_in_the_Lagoon:_H2_line_emission.txt	The Hourglass region of the Lagoon Nebula, Messier 8, was imaged on 1997, July 22, using the IRIS 1–2.5 µm camera, in conjunction with the University of New South Wales Infrared Fabry- P´erot etalon (UNSWIRF, Ryder et al. 1998), on the 3.9-m Anglo Australian Telescope (AAT). The UNSWIRF etalon has a FWHM spectral resolution of ∼ 75 km s−1, a pixel size of 0.77′′ and a 100′′ circular ﬁeld of view. It is scanned through a spectral line of interest in order to obtain an emission line image of a source, with minimal contamination from any continuum radiation present. It also permits limited kinematic information to be obtained, through the emission velocity of the line centre across the ﬁeld of view.
Processed_Molecular_Hydrogen_in_the_Lagoon:_H2_line_emission.txt	The H2 1–0 S(1) (2.1218 µm) line was observed. Following a rapid scan to determine the plate spacing for the line centre, three on-line plate spacings were used for imaging, spaced by 39 km s−1, together with an oﬀ-line setting, 840 km s−1 to the blue. Integration times were two minutes per plate spacing, and in addition, a sky frame, 300′′ N, was obtained immediately after each source frame.
Processed_Molecular_Hydrogen_in_the_Lagoon:_H2_line_emission.txt	The star BS 6748 (K=4.57 mags.) served as the ﬂux standard, and was imaged at each etalon spacing. The absolute accuracy in ﬂux calibration using a Fabry-P´erot etalon is around 30%. A diﬀused dome lamp provided a ﬂat ﬁeld for each etalon spacing. An arc lamp was scanned through a free spectral range in order to wavelength calibrate the etalon response for each pixel of the array.
Processed_Molecular_Hydrogen_in_the_Lagoon:_H2_line_emission.txt	Data reduction was undertaken using a custom software package using IRAF2. Frames are linearised, ﬂat-ﬁelded using a dome ﬂat at the appropriate plate spacing, cleaned of bad pixels, sky- subtracted, shifted to align stars in each frame, smoothed and the oﬀ-line frame subtracted from each on-line frame (having been appropriately scaled to minimise residuals from the subtraction process). Stacking of the three on-line frames produces a data cube, which is ﬁtted, pixel-by-pixel, with the instrumental proﬁle (a Lorentzian), to yield a line image. A coordinate frame was added using the digitised sky survey3 to relate common stars between the visible and 2.1 µm.
Processed_Molecular_Hydrogen_in_the_Lagoon:_H2_line_emission.txt	The line centre at each pixel position is obtained from the plate spacing determined for the 2Image Reduction and Analysis Facility (see http://www.iraf.noao.edu). 3see http://stdatu.stsci.edu/dss/.
Processed_Long-term_monitoring_of_Molonglo_calibrators.txt	Before and after every 12 hour synthesis observation, the Molonglo Observatory Synthesis Telescope (MOST) measures the ﬂux densities of ∼5 compact extragalactic radio sources, chosen from a list of 55 calibrators. From 1984 to 1996, the MOST made some 58 000 such measurements. We have developed an algorithm to process this dataset to produce a light curve for each source spanning this thirteen year period. We ﬁnd that 18 of the 55 calibrators are variable, on time scales between one and ten years. There is the tendency for sources closer to the Galactic Plane to be more likely to vary, which suggests that the variability is a result of refractive scintillation in the Galactic interstellar medium. The sources with the ﬂattest radio spectra show the highest levels of variability, an eﬀect possibly resulting from diﬀering orientations of the radio axes to the line of sight.
Processed_Long-term_monitoring_of_Molonglo_calibrators.txt	Many compact extragalactic radio sources show variations in their radio ﬂux density as a function of time. At high frequencies (ν ≫ 1 GHz) this variability is usually interpreted as being intrinsic to the source (e.g. Qian et al 1995, although see Kedziora-Chudczer et al 1997). Variability at lower frequencies (e.g. Hunstead 1972; Ghosh & Rao 1992) is normally attributed to refractive interstellar scintillation, in which the intensity variations are caused by distortions of the wavefront by electron density gradients in an intervening screen of material (Shapirovskaya 1978; Rickett et al 1984). There is evidence that the parameters of such variability depend on the Galactic latitude of the source (Spangler et al. 1989; Ghosh & Rao 1992), suggesting that the material causing the scintillation is in our own Galaxy.
Processed_Long-term_monitoring_of_Molonglo_calibrators.txt	from 1984 until the commencement of the Wide Field Project in 1996 (Large et al 1994), forms an ideal database with which to study variability in this intermediate frequency range.
Processed_Long-term_monitoring_of_Molonglo_calibrators.txt	A preliminary analysis of three MOST calibrators was made by Campbell-Wilson & Hunstead (1994), hereafter Paper I. It was shown that ﬂux density measurements with a relative accuracy of 2% could be extracted from the database. Over the period from 1990.1 to 1993.7, the source MRC B0409–752 was shown to be stable, while MRC B0537–441 and MRC B1921–293 were found to be highly variable. In this paper we now report on all 55 calibrators used by the MOST, over a thirteen year period. In Section 2 we explain how we process the calibrator measurements in order to produce light curves for each source, and then determine whether a source is variable or not. In Section 3 we present light curves for all 55 sources, plus structure functions for those sources found to be variable. In Section 4 we discuss some individual sources in our sample, and consider whether any of the observed properties correlate with Galactic latitude.
Processed_Long-term_monitoring_of_Molonglo_calibrators.txt	The MOST is an east-west synthesis telescope, consisting of two cylindrical paraboloids of dimen- sions 778 m × 12 m. Radio waves are received by a line feed system of 7744 circular dipoles. The telescope is steered by mechanical rotation of the cylindrical paraboloids about their long axis, and by phasing the feed elements along the arms. In a single 12-hour synthesis, the MOST can pro- duce an image at a spatial resolution of 43′′ × 43′′cosec(\|δ\|) and at a sensitivity of ∼1 mJy beam−1 (where 1 jansky [Jy] = 10−26 W m−2 Hz−1).
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Abstract We report on the detection of source noise in the time domain at 162 MHz with the Murchison Wideﬁeld Array. During the observation the ﬂux of our target source Virgo A (M87) contributes only ∼1% to the total power detected by any single antenna, thus this source noise detection is made in an intermediate regime, where the source ﬂux detected by the entire array is comparable with the noise from a single antenna. The magnitude of source noise detected is precisely in line with predictions. We consider the implications of source noise in this moderately strong regime on observations with current and future instruments.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	noise is correlated between antennas). It is this corre- lation which is exploited in an intensity interferometer (Hanbury Brown & Twiss, 1954).
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	This has two important consequences: ﬁrst, just like the single-dish case, source noise is only seen at the loca- tion of the source in either an aperture synthesis image or a phased-array beam. More accurately, in a synthesis image, the noise is distributed with the shape of the instantaneous Point Spread Function (PSF). Secondly, for an interferometer with a large number of elements, source noise may contribute signiﬁcantly to the image noise at the location of the source even if it is dwarfed by the system noise of a single interferometer element, since unlike noise which is uncorrelated between elements, the noise cannot be reduced by averaging over baselines1.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	This intermediate regime, where the source is weaker than the noise, but is not weak enough to be negligible, is the subject of this paper. We present a clear detection of source noise (arguably the ﬁrst in the intermediate regime) in data from the Murchison Wideﬁeld Array (MWA; a 128-element low-frequency interferometer; Tin- gay et al., 2013). We then examine the scenarios in which this may have a negative impact.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	the formalism obtained to design an experiment to detect source noise with the MWA; in Sect. 3 we describe the observations and analyse the results; in Sect. 4 we discuss the implications of our results.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Below, we draw on the literature to describe the proper- ties of source noise. We refer the reader to other articles which discuss the eﬀect of source noise in other speciﬁc cases such as pulsar scintillometry (Johnson & Gwinn, 2013), spectral line observations (Liszt, 2002), pulsar timing (Osłowski et al., 2011), polarimetry (Deshpande, 1994) or spectropolarimetry (Sault, 2012). The eﬀect of scattering on source noise is discussed in Codona & Frehlich (1987).
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Measurements of the statistics of source noise on very short time scales have also been used to try to detect coherent emission from pulsars (Smits et al., 2003, and references therein). In the current work we focus on the typical synthesis imaging case where the a large number of Nyquist samples are averaged in the correlator, and the integration time is far longer than any timescale on which source noise could conceivably be correlated. In this case the central limit theorem will ensure that source noise is highly Gaussian. We return brieﬂy to this topic in Sect. 4.3.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Both sources and noise can be described either with a characteristic temperature, or as a ﬂux density. We follow Kulkarni (1989) and McCullough (1993) and use the source ﬂux density S and Noise Equivalent Flux Density N throughout. The latter, more commonly known as System Equivalent Flux Density (SEFD) is the ﬂux density of a source that would double the power received by a single element. Thus, at the output of a single element there is the sum of a power proportional to S and power proportional to N , where the constant of proportionality is the product of the collecting area of the antenna, the bandwidth of the receiver, and the antenna gain (and various eﬃciency parameters). Note that N includes not only receiver noise but any noise other than S. This is justiﬁed below.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Throughout, we assume that all interferometer ele- ments are identical, and that both N and S are station- ary, stochastic and ergodic and so 2Bτ samples of base- band data are independent. For an unpolarised source, the number of independent samples can be doubled by observing two orthogonal polarisations, so an implicit npol can be assumed alongside B and τ in the equations below. Polarisation of any kind implies correlation of some kind between orthogonal polarisations (Radhakr- ishnan, 1999) which will complicate the picture. We refer the reader to the references given above for more details on the source noise associated with polarised sources or in polarimetric images.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	where N is the SEFD in jansky, nb is the number of baselines given by n(n − 1)/2 where n is the number of antennas (we assume throughout that all cross correla- tions are used), B is the bandwidth in Hz, and τ is the observing time in seconds.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	replicating the single-dish result. This is intuitive since all antennas see the same noise, and so unlike the weak- source case, the noise cannot be reduced by averaging over baselines.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Firstly, if a strong source is fully resolved on all base- lines, then it will be uncorrelated between antennas and so will behave like system noise. This is relevant for low-frequency interferometers such as the MWA where N is actually dominated by Galactic synchrotron emis- sion, which is almost completely resolved on all but the shortest baselines, and totally absent from the longer > 100λ baselines typically used for continuum imaging. This provides the justiﬁcation for lumping sources such as synchrotron radiation from our Galaxy, the cosmic microwave background, atmospheric noise, etc. in with N ; and provides a more precise deﬁnition of S as noise which is correlated between antennas.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Secondly, the noise only appears in the image where the source does; therefore, for a point source, the source noise level in the map is shaped like the instantaneous PSF of the array.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Table 1 N/n for various radio interferometers, built and planned. All come from the SKA Baseline Design2 table 1 except the MWA ﬁgure which is derived from Tingay et al. (2013).
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	where S is the apparent source brightness (source bright- ness distribution convolved with the PSF) at any point in the image.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Here we use the form of the equation given by McCul- lough (1993) who corrected a minor error in Kulkarni (1989). The three terms reﬂect the three classes of base- line pair: the autocorrelations will correlate perfectly regardless of whether S or N dominates and the other two classes scale with S0 and S02 respectively.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	large n as predicted by Anantharamaiah et al. (1991) (the fractional error is < 1% for n > 12).
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Figure 1 shows the achievable dynamic range (signal- to-noise) oﬀ- and on-source (i.e. calculated using equa- tions 1&4 respectively). The source brightness is given in units of N/n (see tables 1). On- and oﬀ-source noise start to diverge at ∼ N/10n – just a few hundred mJy for the SKA; and the on-source dynamic range has almost reached its maximum value by ∼ 10N/n.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Table 2 shows the maximum dynamic range obtainable for various common MWA observing parameters. For a typical snapshot, the dynamic range will be > 104 before source noise eﬀects become noticeable. This level of source noise is not easily measured since noise due to confusion noise or calibration errors are likely to limit the dynamic range far more.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	In contrast, if an image is made with the smallest possible bandwidth and observing time with the cur- rent MWA correlator, the maximum achievable dynamic range is ∼100. Solar images are regularly made in this way with the MWA, and the quiet Sun ﬂux density at 150 MHz is suﬃciently far into the strong source regime that even if this ﬂux density is spread over many pix- els the brightness of each resolution unit to be close to the strong source regime. This means that every res- olution unit of the Sun will ﬂuctuate by ∼1% across timesteps, spectral channels and polarisations regardless of any intrinsic change in brightness. However, rapid intrinsic changes in solar emission (which may be po- larised) in both time and frequency would complicate measurements of source noise in observations of the Sun.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Bτ for various MWA observing modes. All band- Table 2 widths take into account discarded band edges where appro- priate. Maximum resolutions in time and frequency refer to the original online MWA correlator (still standard at the time of writing).
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	number of spectral channels or timesteps the RMS of pixels on and oﬀ source would be expected to show a measurable diﬀerence. The former is the approach taken by McCullough (1993) in measuring source noise in the strong regime. We take the latter approach which has much in common with the procedure used for making IPS observations with the MWA (Morgan et al., 2018).
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	As part of our Interplanetary Scintillation (IPS) ob- serving campaign (Morgan et al., 2019) we observed a series of calibrators in high time resolution mode with 5- minute observing time. These observations pre-date the upgrade of the MWA to Phase II (Wayth et al., 2018). Virgo A (M87), at our observing frequency of 162 MHz was observed at 2016-01-16T21:04:55 UTC when it was 53 degrees above the horizon and 112 degrees from the Sun. The solar elongation is relevant since Virgo A has a compact core and jet several janskys in brightness at GHz frequencies (e.g. Reid et al., 1982) which would be compact on IPS scales. This means that Virgo A will also show interplanetary scintillation. However this variability should be fully resolved with 2 Hz sampling. Throughout, we assume a ﬂux density of 1032 Jy for Virgo A (based on values gleaned from the literature by Hurley-Walker et al., 2017). We do not include any uncertainty on the ﬂux density of Virgo A in our errors below (such an error would cause an equal scaling error in all of our measured and derived quantities). Note that no direction-dependent ﬂux scaling was carried out.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	ent brightness approximately 3% of that of Virgo A. After standard ﬂagging of the start and end of the ob- servation, 573 timesteps remained. A number of clearly discrepant points were easily discernible in all lightcurves (the same in each) and these have been interpolated over for all lightcurves. The on-source lightcurve shows clear variability on a timescale consistent with ionospheric scintillation.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Fig. 3 shows power spectra for each of the lightcurves shown in Fig. 2. These were generated using Welch’s method (Welch, 1967) with 17 overlapping FFTs of 32 samples with a Hanning window function. The strong signal that dominates at low frequencies is consistent with ionospheric scintillation and shows the characteris- tic Fresnel “Knee” at about 0.1 Hz. Above this frequency the power drops steeply. The same ionospheric scintil- lation can be seen for all oﬀ-source lightcurves due to the sidelobes of Virgo A, however the sidelobes level is low enough that the variance is suppressed by at least 3 orders of magnitude. 3C270 scintillates independently of Virgo A and also has a slightly lower noise level3.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Above 0.4 Hz the variance from ionospheric scintilla- tion is negligible, and the power spectrum consists only of white noise. We conclude that all variability due to calibration errors such as those introduced by the iono- sphere are restricted to lower frequencies as we would expect. This white noise is clearly at very diﬀerent lev- els on-source and oﬀ-source. We can use the average of all points above 0.4 Hz to measure the oﬀ-source noise to be 0.410±0.005 Jy, and the on-source noise to be 0.75±0.03 Jy. The former implies that the SEFD (N ) is 125000±1500 Jy (B = 13.44 MHz, τ = 0.5 s, n = 118 i.e. 10 antennas ﬂagged). This is about a factor of 2.4 higher than that given in table 1, which is not surprising given the oﬀ-zenith pointing and the fact that Tingay et al. (2013) assume a nominal sky temperature and ﬁeld of view.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Using the source brightness and oﬀ-source noise we can calculate what the on-source noise should be using equation 4. This prediction (0.807±0.005 Jy) is shown as the dashed line on Fig. 3 and it agrees extremely well with the measured on-source noise. There is a small discrepancy, which is within 2-σ; however even this small excess can be explained as a small leakage of source noise into the oﬀ-source pixels due to the sidelobes of Virgo A. Alternatively, this may be due to Virgo A being slightly resolved on some MWA baselines.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Figure 2. Timeseries (lightcurve) showing brightness of various pixels in the image as a function of time. Black line shows the on-source lightcurve; dot-dashed black line shows a much weaker source: the (Western lobe of) 3C270; grey lines show a selection of oﬀ-source pixels. 21 points in the lightcurve had to be ﬂagged due to clearly discrepant points, and these have been linearly interpolated over in all lightcurves; red circles denote these ﬂagged points for the on-source pixel.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	sidelobe level of 1% in the instantaneous monochromatic case. Therefore for S = N/n = 2.7 Jy, the source noise level on-source will be ∼ 2× the weak-noise level, and oﬀ-source this 100% increase in noise will become just 1%.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	We have detected source noise at a level which is ﬁrmly in line with the predictions of equation 3 (no empirical test for the existence of the subtle eﬀects that diﬀerentiate equations 3&4 are possible with our data since the two are practically identical for n > 100). We now consider the eﬀect that this would have oﬀ-source in synthesis images.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	As the bandwidth and integration time are increased, both the system noise and the source noise will reduce with Bτ ; the ratio between the two will remain con- stant, and the excess noise due to the source noise will remain at 1%. With appreciable fractional bandwidth and/or Earth rotation synthesis this source noise will be spread smoothly over the image.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	For the example given, the eﬀects of source noise oﬀ-source are extremely low; however there will be 5× 2.7 Jy sources per FoV with the SKA-low (Franzen et al., 2016). For a 14 Jy source (∼1 per FoV) there would be an excess of 5%. Only for a 270 Jy source will the noise be doubled and such sources are relatively rare for extragalactic ﬁelds assuming arcminute or better resolution.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Figure 3. Power spectrum for each of the lightcurves described in Fig. 2. The thin black line shows on-source power spectrum; the dot-dashed line shows a much weaker source (3C270); the grey lines show oﬀ-source power spectra, with the thick black line showing the average of the grey lines. Power spectrum parameters are described in the text. The single error bar shows the 95% conﬁdence interval for a single point with these parameters. The dotted line is the mean of all oﬀ-source power spectrum points above 0.4Hz. The dashed line is the estimate of the on-source power based on the on-source brightness and oﬀ-source noise.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Figure 4. Ratio of on-source to system noise (equations 3&1) as a function of source strength for a large-N interferometer (this is the ratio of the dashed line to the solid line in Fig. 1).
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	of source noise assumes natural weighting of baselines. More uniform weighting will increase the weighting of certain baselines (the longer ones) by a large amount compared to the shorter spacings, particularly for instru- ments like the SKA-low with very high concentrations of collecting area at the centre. This will reduce the eﬀective number of antennas and therefore the source noise relative to the system noise.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Anantharamaiah et al. (1989) note that source noise can be subtracted perfectly from a snapshot by deconvolu- tion, while noting the computational eﬀort required to do so. Here we explore the limitations on how cleanly this can be done.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Subtracting source noise requires that the ﬁeld be imaged with suﬃcient time and frequency resolution for changes in the PSF to be insigniﬁcant from one im- age to the next. The required resolution will depend on the array and imaging parameters, however the re- quirements are similar to those required to minimise time-average and bandwidth smearing (e.g. Thompson et al., 2017, 6.3-6.4) and are likely to be demanding. Here we concentrate on a more fundamental issue that such a subtraction poses: namely that if the source noise is to be perfectly subtracted, the source brightness must be allowed to vary as a function of time and frequency and no spectral smoothness can be assumed. Conversely if spectral smoothness is strictly imposed, as is normally the case when subtracting continuum from spectral line cubes, no source noise will be subtracted. Clearly any number of compromise schemes between these to ex- tremes could be devised; however the principle remains that source noise can only be subtracted to the extent that it can be separated from any spectral or temporal signal that needs to be preserved.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	The properties of source noise lead us to the striking conclusions that the SKA will not more accurately char- acterise a 1000 Jy source than a 10 Jy source, and that the standard equations underestimate the noise on the measurement of a 2 Jy source by a factor of two. We have demonstrated that even the MWA—which is orders of magnitude less sensitive than the SKA—can measure these eﬀects at the 12-σ level (albeit it in an observation of a 1000 Jy source contrived for the purpose).
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Our observations show that the magnitude of the eﬀect is precisely in line with predictions, and so source noise can be predicted very precisely from easily measurable parameters (the ratio of source brightness to system noise). This is in contrast to other sources of variability such as the ionosphere. Source noise is also expected to have stochastic behaviour as a function of frequency and time, and we exploit this to separate source noise from much stronger ionospheric eﬀects.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Although the eﬀects on standard continuum imaging appear to be benign in all but the most extreme cases, source noise could become problematic wherever weak signals (including polarised signals, see Sault, 2012) need to be measured in the presence of strong emission; es- pecially where the signals involved have ﬁne frequency structure or vary on short timescales. This situation will occur for any spectral line observations where the continuum has to be subtracted to see the much weaker line emission. In this situation the noise will be higher and not uniform across the image. Noise estimates made from regions of the image with no continuum will under- estimate the noise.
Processed_A_Measurement_of_Source_Noise_at_Low_Frequency:_Im.txt	Fast transients such as Fast Radio Bursts (FRBs) and pulsars will also reach the strong source limit, especially since their signals occupy only a narrow sloping band in a dynamic spectrum. For ASKAP, a few of the detected FRBs (Shannon et al., 2018) already have peak ﬂux density in the strong source limit and for SKA sensi- tivity many will be in this regime. Thus measurements of source noise of FRBs should be possible, and such measurements may be valuable, since they probe the extent to which the emission is coherent (Melrose, 2009).
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt	In general, the possibility of detecting AME in HII regions is very intriguing from a theoretical point of view. On one side, HII regions are ideal candidates to harbor this emission, as processes common to this type of environment, such as plasma drag and photoelectric effect, can generate enough an- gular momentum to spin the grains and induce detectable mi- crowave radiation. However, signiﬁcant dust grain depletion is expected to occur in HII regions as a consequence of the in- tense radiation pressure and stellar winds (Inoue 2002; Draine 2011). Even more importantly, embedded optically thick Ul- tra Compact HII (UCHII) regions could mimic, in this case, an excess at microwave frequencies. In fact UCHII regions are characterized by high emission meaures (i.e. 109−10 pc cm−6) and by self-absorbed free-free emission at frequencies up to several GHz. This self-absorbed free-free could potentially generate the observed rising spectrum attributed to AME.

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.