White Matter Physics

8/12/2019 White Matter Physics

1/43

EMU: Evolutionary Map of the Universe

Ray P. Norris1, A. M. Hopkins2,36, J. Afonso3, S. Brown1, J. J. Condon4, L.

Dunne

5

, I. Feain

1

, R. Hollow

1

, M. Jarvis

6,38

, M. Johnston-Hollitt

7

, E. Lenc

1

, E.Middelberg8, P. Padovani9, I. Prandoni10, L. Rudnick11, N. Seymour12, G.Umana13, H. Andernach14, D. M. Alexander21, P. N. Appleton15, D.Bacon16, J.Banfield1, W. Becker17, M. J. I. Brown18, P. Ciliegi19, C. Jackson1, S. Eales20,A. C. Edge21, B.M. Gaensler22,36, G. Giovannini10, C. A. Hales1,22, P.Hancock22,36, M.Y.Huynh23, E. Ibar24, R. J. Ivison24,25, R. Kennicutt26, Amy E.

Kimball4, A. M. Koekemoer27, B. S. Koribalski1, A. R. Lopez-Sanchez2,37, M.Y. Mao1,2,28, T. Murphy22,36, H. Messias29, K. A. Pimbblet18, A. Raccanelli16,K. E. Randall1,22, T. H. Reiprich30, I. G. Roseboom31 H. Rottgering32, D.J.Saikia33, R.G.Sharp34, O.B.Slee1, Ian Smail21, M. A. Thompson6, J. S.Urquhart1, J. V. Wall35, G.-B. Zhao16

Abstract: EMU is a wide-field radio continuum survey planned for the new Aus-tralian Square Kilometre Array Pathfinder (ASKAP) telescope. The primary goalof EMU is to make a deep (rms 10 Jy/beam) radio continuum survey of theentire Southern Sky at 1.3 GHz, extending as far North as +30 declination, with aresolution of 10 arcsec. EMU is expected to detect and catalogue about 70 milliongalaxies, including typical star-forming galaxies up to z1, powerful starbursts toeven greater redshifts, and AGNs to the edge of the visible Universe. It will un-doubtedly discover new classes of object. This paper defines the science goals andparameters of the survey, and describes the development of techniques necessaryto maximise the science return from EMU.

Keywords: telescopes surveys stars: activity galaxies: evolution galaxies: forma-

tion cosmology: observations radio continuum: general

1 Introduction

1.1 Background

Deep continuum surveys of the radio sky have adistinguished history both for discovering new classesof object and for providing radio counterparts toastronomical objects studied at other wavelengths.The earliest large surveys, such as the 3C cata-logue (Edge et al. 1959) and the Molonglo Ref-erence Catalogue (Large et al. 1981), gave us the

first insight into the physics of radio galaxies andradio-loud quasars, but were insufficiently sensi-tive to detect any but the nearest radio-quiet orstar-forming galaxies. Later radio surveys reachedflux densities where normal star-forming galaxieswere detected, but were still largely dominatedby radio-loud active galactic nuclei (AGN). Onlyvery long integrations in narrow deep fields madeit possible to start probing star-forming galaxiesbeyond the local Universe. This paper describesa planned survey, EMU (Evolutionary Map of theUniverse), which will reach a similar sensitivity (10 Jy/beam) to those deep surveys, but over the

entire visible sky. At that sensitivity, EMU will beable to trace the evolution of galaxies over most ofthe lifetime of the Universe.

Fig. 1 shows the major 20-cm continuum ra-dio surveys. The largest existing radio survey,shown in the top right, is the wide but shallowNRAO VLA Sky Survey (NVSS), whose releasepaper (Condon et al. 1998) is one of the most citedpapers in astronomy. The most sensitive existingradio survey is the deep but narrow Lockman Hole

observation (Owen & Morison 2008) in the lowerleft. All current surveys are bounded by a diagonalline that roughly marks the limit of available tele-scope time of current-generation radio telescopes.The region to the left of this line is currently unex-plored, and this area of observational phase spacepresumably contains as many potential new dis-coveries as the region to the right.

The Square Kilometre Array (SKA) is a pro-posed major internationally-funded radio telescope(Dewdney et al. 2009) whose construction is ex-pected to be completed in 2022. It will be manytimes more sensitive than any existing radio tele-

1

arXiv:1106.3219

v1

[astro-ph.CO]

16Jun2011


2/43

www.publish.csiro.au/journals/pasa 2

Figure 1 Comparison of EMU with existing deep20 cm radio surveys. Horizontal axis is 5- sen-sitivity, and vertical axis shows the sky coverage.The diagonal dashed line shows the approximateenvelope of existing surveys, which is largely de-termined by the availability of telescope time. Thesquares in the top-left represent the EMU survey,discussed in this paper, and the complementaryWODAN (Rottgering et al. 2010b) survey whichhas been proposed for the upgraded Westerbork

telescope to cover the sky North of +30

. Surveysrepresented by diagonal lines are those which rangefrom a wide shallow area to a smaller deep area.The horizontal line for ATLAS extends in sensitiv-ity from the intermediate published data releases(Norris et al. 2006; Middelberg et al. 2008a; Haleset al. 2011) to the final data release (Banfield etal. 2011).

scope, and will answer fundamental questions aboutthe Universe (Carilli & Rawlings 2004). It islikely to consist of between 1000 and 1500 15-meter

dishes in a central area of diameter 5 km, sur-rounded by an equal number of dishes in a regionstretching up to thousands of km.

The Australian SKA Pathfinder (ASKAP) is anew radio telescope being built both to test anddevelop aspects of potential SKA technology, andto develop SKA science. ASKAP is being builton the Australian candidate SKA site in WesternAustralia, at the Murchison Radio-astronomy Ob-servatory, with a planned completion date of late2012. In addition to developing SKA science andtechnology, ASKAP is a major telescope in its ownright, likely to generate significant new astronom-

ical discoveries.

1.2 ASKAP

ASKAP (Johnston et al. 2007, 2008;Deboer et al.2009) will consist of 36 12-metre antennas spreadover a region 6 km in diameter. Although thearray of antennas is no larger than many exist-ing radio telescopes, the feed array at the focusof each antenna is revolutionary, with a phased-array feed (PAF: Bunton & Hay 2010) of 96dual-polarisation pixels, designed to work in a fre-quency band of 7001800 MHz, with an instanta-neous bandwidth of 300 MHz. This will replacethe single-pixel feeds that are almost universal incurrent-generation synthesis radio telescopes. Asa result, ASKAP will have a field of view up to30 deg2 enabling it to survey the sky up to thirty

times faster than existing synthesis arrays, and al-lowing surveys of a scope that cannot be contem-plated with current-generation telescopes. To en-sure good calibration, the antennas are a novel 3-axis design, with the feed and reflector rotating tomimic the effect of an equatorial mount, ensuringa constant position angle of the PAF and sidelobeson the sky. The pointing accuracy of each antennais significantly better than 30 arcsec.

The ASKAP array configuration (Gupta et al.2008) balances the need for high sensitivity to ex-tended structures (particularly for neutral hydro-gen surveys) with the need for high resolution for

continuum projects such as EMU. To achieve this,30 antennas follow a roughly Gaussian distribu-tion with a scale of 700 m, corresponding to apoint spread function of30 arcsec using naturalweighting, with a further six antennas extendingto a maximum baseline of 6 km, correspondingto a point spread function of 10 arcsec usinguniform weighting. The positions of the anten-nas are optimised foruvcoverage (i.e. coverage inthe Fourier plane) between declination 50 and+10, but give excellent uvcoverage between dec-lination 90 and +30.

The PAF is still under development, but theperformance of prototypes gives us confidence that

the EMU survey is feasible as planned. The PAFwill consist of 96 dual-polarisation receivers, eachwith a system temperature 50 K, which are com-bined in a beam-former to form up to 36 beams.Each of these beams has the same primary beamresponse as a single-pixel feed ( 1.2 full-widthhalf-maximum at 1.4 GHz), distributed in a uni-form grid across an envelope of 30 deg2. The op-timum weighting and number of beams is still be-ing studied, but the current expectation is that 36beams will be used for EMU, with the sensitivityover the 30 deg2 FOV (field of view) expected tobe uniform to 20%. This will be improved to


3/43



4/43


Figure 2 A representation of the EMU sky cov-erage in Galactic coordinates overlaid on 23 GHzWMAP data (Gold et al. 2011). The dark areain the top left is the part of the sky NOT coveredby EMU.

also have higher sensitivity to extended structures.The sky coverage of EMU is shown in Fig.2, andthe EMU specifications are summarised in Table1. Like most radio surveys, EMU will adopt a 5- cutoff, leading to a source detection thresholdof 50 Jy/beam. EMU is expected to generate acatalogue of about 70 million galaxies, and all ra-dio data from the EMU survey will be placed inthe public domain as soon as the data quality hasbeen assured.

Table 1 EMU Specifications

Instantaneous FOV 30 deg2

Area of survey entire sky southof +30 dec.

Synthesised beamwidth 10 arcsec FWHMFrequency range 1130-1430 MHzRms sensitivity 10 Jy/beamTotal integration time 1.5 years1

Number of sources 70 million

1 The primary specification is the sensitivity,rather than the integration time. If for any reasonASKAP is less sensitive than expected, EMU willincrease the integration time rather than losesensitivity. Conversely, an increase in sensitivityof ASKAP may reduce the total integration time.

Currently, only a total of about 5 deg2 of thesky has been surveyed at 20 cm to the planned10 Jy/beam rms of EMU, in fields such as theHubble,Chandra, COSMOS and Phoenix deep fields(Huynh et al. 2005; Miller et al. 2008; Schinnereret al. 2007;Hopkins et al. 2003; Biggs & Ivison2006;Morrison et al. 2010), with a further 7 deg2

expected in the immediate future as part of theATLAS survey (Norris et al. 2006; Middelberg et

al. 2008a;Hales et al. 2011;Banfield et al. 2011).Surveys at this depth extend beyond the tradi-

tional domains of radio astronomy, where sourcesare predominantly radio-loud galaxies and quasars,into the regime of star-forming galaxies. At thisdepth, even the most common active galactic nu-clei (AGN) are radio-quiet AGNs, which make upmost of the X-ray extragalactic sources. As a re-sult, the role of radio astronomy is changing. Whereasmost traditional radio-astronomical surveys hadmost impact on the niche area of radio-loud AGNs,current radio-astronomical surveys are dominatedby the same galaxies as are studied by opticaland IR surveys, making radio-astronomical sur-veys such as EMU an increasingly important com-ponent of multi-wavelength studies of galactic evo-lution.

Figure 3 Distribution of differential radio sourcecounts at 1.4 GHz, based on and updated fromthe distribution shown inHopkins et al. (2003).The solid curve is the polynomial fit from Hop-kins et al. (2003), the dashed curve is an up-dated polynomial fit and is the one used to esti-mate the EMU source numbers. The horizontaldot-dashed line represents a non-evolving popula-tion in a Euclidean universe. The shaded regionshows the prediction based on fluctuations due toweak confusing sources ( a P(D) analysis) fromCondon (1974);Mitchell & Condon. (1985).

Because only a small area of sky has been sur-veyed to the depth of EMU, it is difficult to es-timate precisely how many galaxies it will detect.Most surveys to this sensitivity cover only a smallarea of sky, so that source counts at this levelare significantly affected by sample variance, com-pleteness, and bias issues. Our estimate for thenumber per deg2 above a flux density of 50 Jy/beamis based on an extrapolation from source counts athigher flux densities (2263 sources/ deg2; Jack-son 2005) , the compilation shown in Fig. 3(2278


5/43


sources/ deg2), and the COSMOS survey (2261sources/ deg2; Scoville et al. 2007; Schinnerer etal. 2007). These three figures are in good agree-ment and predict a total of 70 million sourcesin EMU, which is therefore the number adoptedthroughout this paper.

Estimating the fraction of these radio sourceswhich are AGN is difficult. Below 1 mJy, star-forming galaxies start to become a major compo-nent of the 1.4 GHz source counts, dominating be-low 0.15 mJy (Seymour et al. 2008; Ibar etal. 2009), but, even at these levels, there is stilla significant proportion of low-luminosity AGNs(Jarvis et al. 2004; Afonso et al. 2005,2006;Nor-ris et al. 2006;Simpson et al. 2006;Smolcic et al.2008; Seymour et al. 2008; Mignano et al. 2008;Padovani et al. 2009).

Seymour et al. (2008) have presented the

most comprehensive attempt so far to divide radiosources into AGN and SF galaxies, and their re-sult, together with other recent estimates, is shownin Fig. 4. From these we estimate that about 75%of EMU sources will be star-forming galaxies.

Figure 4 Differential fraction of star-forming galax-ies as a function of 1.4 GHz flux density, from a se-lection of recent deep surveys. Shaded boxes, andthe two lines for Padovani et al., show the rangeof uncertainty in the survey results. Arrows indi-cate constraints from other surveys. These resultsshow that the fraction of star-forming galaxies in-

creases rapidly below 1 mJy and, at the 50 Jysurvey limit of EMU, about 75% of sources will bestar-forming galaxies.

To estimate the redshift distribution of AGNand SF galaxies, we use the SKADS simulation(Wilman et al. 2008, 2010), shown in Fig. 5. About50 million of the EMU sources are expected to bestar-forming galaxies (see 2.1) at redshifts up to z 3, with a mean redshift of z 1.08. The remain-der are AGNs with a mean z 1.88, and extend upto z 6. However, if any FRII (Fanaroff & Riley1974) galaxies exist beyond that redshift (e.g. L

3.3 1025W Hz1 at z = 10), EMU will detectthem.

Confusion of radio sources, discussed more thor-oughly in 3.6.1, is well-understood at this level,since previous surveys have already imaged smallareas of sky to this depth and beyond.

Figure 5 Expected redshift distribution of EMUsources, based on the SKADS simulations(Wilman et al. 2008, 2010). The five lines show thedistributions for star-forming galaxies (SFG), star-burst galaxies(SB), radio-quiet quasars (RQQ),and radio-loud galaxies of Fanaroff-Riley types Iand II (FRI & FR2; Fanaroff & Riley 1974).Vertical scale shows the total number of sourcesexpected to be detected by EMU.

EMU differs from many previous surveys inthat a goal of the project is to cross-identify thedetected radio sources with major surveys at otherwavelengths, and produce public-domain VO-accessiblecatalogues as value-added data products. Thisis facilitated by the growth in the number of largesouthern hemisphere telescopes and associated plannedmajor surveys spanning all wavelengths, discussedbelow in 3.9.

1.4 Science

Broadly, the key science goals for EMU are:

To trace the evolution of star-forming galax-ies from z = 2 to the present day, using awavelength unbiased by dust or molecularemission,

To trace the evolution of massive black holesthroughout the history of the Universe, andunderstand their relationship to star forma-tion,

To use the distribution of radio sources toexplore the large-scale structure and cosmo-logical parameters of the Universe, and totest fundamental physics,


6/43


To determine how radio sources populate darkmatter halos, as a step towards understand-ing the underlying astrophysics of clustersand halos,

To create the most sensitive wide-field at-las of Galactic continuum emission yet madein the Southern Hemisphere, addressing ar-eas such as star formation, supernovae, andGalactic structure,

To explore an uncharted region of observa-tional parameter space, with a high likeli-hood of finding new classes of object.

Table 1 gives an overview of EMU specifica-tions. In addition to the well-defined scientificgoals outlined above, and the obvious legacy value,the large EMU dataset will include extremely rare

objects, only possible by covering large areas.A challenge for EMU will be the lack of spectro-

scopic redshifts, since no existing or planned red-shift survey can cover more than a tiny fraction ofEMUs 70 million sources. As discussed in 3.11, 30% of EMU sources will have multi-wavelengthoptical/IR photometric data at the time of datarelease, increasing to 70% in 2020. We expectthese to provide accurate photometric redshifts forthe majority of star-forming galaxies in EMU, anda minority of AGN (for which photometric red-shifts tend to be unreliable). In addition, many ofthe EMU sources will have statistical redshifts,

which are valuable for some statistical tests. Forexample, most polarised sources are AGNs (meanz 1.88), while most unpolarised sources are star-forming galaxies ( mean z 1.08). More pre-cise statistical redshifts can be derived where op-tical/IR photometry is available, as discussed in3.11.

A further goal of EMU is to test and developstrategies for the SKA. Many aspects of ASKAP,such as the automated observing, calibration, anddata reduction processes, and the phased-array feeds,are potential technologies for the SKA, and it willbe important to test whether these approaches de-liver the planned results.

1.5 Relationship to other surveys

The following radio surveys are particularly com-plementary to the scientific goals of EMU.

The WODAN survey (Rottgering et al. 2010b)has been proposed for the Westerbork tele-scope which is currently being upgraded witha phased array feed (Oosterloo et al. 2009).WODAN will cover the northern 25% of thesky (i.e. North of declination +30) that

is inaccessible to ASKAP, to an rms sen-sitivity of 10 Jy/beam and a spatial res-olution of 15 arcsec. Together, EMU andWODAN will provide full-sky 1.3 GHz imag-ing at 1015 arcsec resolution to an rmsnoise level of 10 Jy/beam, providing an un-precedented sensitive all-sky radio survey asa legacy for astronomers at all wavelengths.The WODAN survey will overlap with EMUby a few degrees of declination to providea comparison and cross-validation, to ensureconsistent calibration, and to check on com-pleteness and potential sources of bias be-tween the surveys.

The LOFAR continuum survey (Rottgeringet al. 2010a) will cover the northern half ofthe sky (i.e. North of declination 0) with

the new LOFAR telescope operating at lowfrequencies (15200 MHz). LOFAR will beespecially complementary to WODAN andEMU in surveying the sky at high sensitiv-ity and resolution but at a much lower fre-quency.

The MIGHTEE survey (van der Heyden& Jarvis 2010) on the Meerkat telescope(Jonas 2009) will probe to much fainter fluxdensities (0.1-1Jy rms) over smaller areas( 35 square degrees) at higher angular reso-lution, providing the completeness as a func-tion of flux density for the EMU and WODAN

Surveys. The higher sensitivity and resolu-tion will enable exploration of the AGN andstar-forming galaxy populations to higher red-shifts and lower luminosities.

POSSUM (Gaensler et al. 2010) is an all-sky ASKAP survey of linear polarisation. Itis expected that POSSUM will be commen-sal with EMU, and that the two surveys willoverlap considerably in their analysis pipelinesand source catalogues. POSSUM will pro-vide a catalogue of polarised fluxes and Fara-day rotation measures for approximately 3million compact extragalactic sources. These

data will be used to determine the large-scalemagnetic field geometry of the Milky Way,to study the turbulent properties of the in-terstellar medium, and to constrain the evo-lution of intergalactic magnetic fields as afunction of cosmic time. POSSUM will alsobe a valuable counterpart to EMU, in that itwill provide polarisation properties or upperlimits to polarisation for all sources detectedby EMU.

FLASH (Ball et al. 2009) is an ASKAP sur-vey whose goal is to detect extragalactic neu-tral hydrogen absorption. To do so it will


7/43


observe at frequencies outside the 1130-1430MHz band of EMU, thus yielding valuablespectral index information for those sourcescommon to both surveys.

DINGO (Ball et al. 2009) is an ASKAP sur-vey whose goal is to detect faint extragalac-tic neutral hydrogen emission, and to do so itwill spend many days on one ASKAP point-ing. As a by-product, it will thus providesensitive continuum images over smaller ar-eas (several tens of deg2 ), allowing EMU toexplore fainter flux densities in an optimaltiered survey structure, and also to quantifythe effects of confusion at this level. How-ever, the continuum images from DINGO willbe severely confusion-limited at flux densi-ties below a few Jy/beam. It may be pos-

sible to transcend this limit by subtractingknown sources from the image, such as thosestar-forming galaxies which are seen in in-frared images and whose radio flux can bepredicted using the IR-radio correlation. How-ever, this challenge is currently external tothe core EMU project.

VAST (Chatterjee et al. 2010) is an ASKAPsurvey that will observe partly commensallywith EMU, with the goal of detecting tran-sients and variable sources. EMU has noplanned transient capability, since all infor-mation on variability of EMU sources will

be available from VAST. This separation en-ables each of EMU and VAST to focus onits specific science goals, although significantcoordination between the projects will clearlybe essential.

WALLABY (Koribalski et al. 2011) is anHI survey which will deliver high-sensitivityspectral line data over the same area of skyas EMU, and will observe commensally withEMU. Observations will give a velocity cov-erage of 2,000 to +77,000 km s1 (z = 0 0.26) and velocity resolution of 4 km s1.The angular resolution for WALLABY will

be 30 arcsec, a factor three lower than EMU,as computing resources to make the largespectral line data cubes are restricted to base-lines shorter than 2 km. Nearly all the 5 105 sources detected by WALLABYwill also be detected by EMU, and WAL-LABY will provide an HI redshift for eachof these, adding significantly to the redshiftinformation for low-redshift EMU sources.

This paper defines the EMU survey, setting outits science goals in 2, and identifying the chal-lenges to achieve these goals. 3 describes how

these challenges are being addressed in the EMUDesign Study, and 4 describes the survey opera-tional plan, primary data products, and the datarelease plans and policy.

2 EMU Science Goals

2.1 Star-forming galaxies and AGNs

The fraction of star-forming galaxies as a functionof flux density is shown in Figure4. Of the 70million sources detected by EMU to a 5 limit of50 Jy, about 20 million galaxies are expected tobe dominated by Active Galactic Nuclei (AGN),and 50 million to be dominated by star formation(SF). However, there is considerable overlap be-tween the two classes, with composite AGN/SFgalaxies becoming more common at low flux den-sities (e.g.Chapman et al. 2003;Norris et al. 2006;Seymour et al. 2008).

It is unclear what fraction of putative SF galax-ies have a significant AGN component. However,the excellent agreement in star-formation rate be-tween radio and other star-forming indicators (e.g.Cram et al. 1998;Bell 2003) suggests that an AGNis not a major contributor to the radio emission insuch galaxies.

Detected AGNs and star-forming galaxies spana significant fraction of the age of the Universe,almost reaching the era of re-ionisation for radioAGNs and the most extreme starbursts. Partic-

ularly at high redshift, both AGN and star for-mation processes are likely to be important in alarge fraction of galaxies, but neither the fractionof the luminosity (bolometric and radio) generatedby each process, nor how they are influenced byfeedback, is currently known. EMU will explorethe evolution of these populations and its depen-dence on galaxy mass, environment, SF history,and interaction/merger history. It will quantifythese effects in detail, by providing a deep homo-geneously selected sample of both AGN and SFgalaxies over the majority of cosmic history, unbi-ased by dust obscuration, and so provide a com-

prehensive overview of galaxy evolution.The EMU analysis pipeline, which will encom-pass automated multi-wavelength cross-identificationof sources between the EMU catalogue and othercomplementary surveys, will also include a vari-ety of measures appropriate for distinguishing be-tween, and quantifying the proportions of, AGNand star-forming activity. These will include:

Radio morphology (e.g.Biggs & Ivison 2008;Biggs, Younger & Ivison 2010),

Radio spectral index (e.g.Ibar et al. 2009,2010),


8/43


9/43


Figure 7 Evolution with redshift of the star for-mation rate density (SFRD) in galaxies, . Thegrey and shaded regions in both panels are the 1-and 3- confidence region fits to the compilationof SFRD data from Hopkins & Beacom (2006).

Top: Measured SFRD taken fromMauch & Sadler(2007) and Seymour et al. (2008). Bottom: The

SFRD shown as a function of galaxy mass, adaptedfrom Fig. 7 ofMobasher et al. (2009).

have a significant effect on the host galaxy and itssurroundings.

The peak of QSO activity took place at z2(e.g.Schmidt 1968;Shaver et al. 1996; Croom etal. 2004;Hasinger et al. 2005), at epochs when SFactivity was also extreme. Intriguingly, embeddedAGNs have been found in 20-30% of z2 massivestar-forming galaxies (Daddi et al. 2007). It haseven been suggested that the AGN jets could beresponsible for the formation of galaxies at highredshift (Elbaz et al. 2009;Klamer et al. 2004).

At lower redshifts, AGNs appear to downsize,in a similar way to that of SF galaxies, so that thepeak of the AGN activity appears to shift signif-icantly to lower redshifts for lower power AGNs.This suggests (Cowie et al. 1996) that a feed-back mechanism couples AGNs to galaxy evolu-tion (Hasinger et al. 2005;Bongiorno et al. 2007;Padovani et al. 2009), via supernovae, starburstsor AGNs (Croton et al. 2006; Bower et al. 2006).

AGN jets either push back and heat the infallinggas, reducing the cooling flows building up thegalaxy, or shock-heat and collapse the gas clouds,inducing star formation. The AGN energetic feed-back appears to be an important ingredient for re-producing the galaxy stellar mass function (Cro-ton et al. 2006;Best et al. 2006), and the remark-able black hole vs. bulge mass (or velocity disper-sion) correlation (Gebhardt et al. 2000;Ferrarese& Merritt 2000;Springel et al. 2005).

Recent work work on low-z 3CR galaxies byOgle et al. (2010) suggest that radio jets can leadto significant shock-heating of the host moleculardisk, leading to a significant enhancement of warmmolecular hydrogen emission. Despite their largemolecular masses, these systems seem to be veryinefficient at making stars, suggesting that shock-heating by radio jets can play a negative-feedback

role in star formation in these systems.Low power radio-loud AGNs (P


10/43


2.4 High-redshift AGNs and theEpoch of Re-ionisation

There have been many attempts to find very high-redshift radio galaxies (De Breuck et al. 2001;

Jarvis et al. 2001a,b;Best et al. 2003;Cruz et al.2006, 2007) but all suffer from the difficulty of find-ing these extremely rare galaxies. EMU will iden-tify numerous radio galaxies at z>4 and isolatethe most distant objects by cross-matching withlarge near- and mid-infrared surveys, as describedin 3.9. EMU will also provide spectral indices and(with POSSUM) polarisation information whichcan also help identify high-redshift AGNs.

High-redshift AGNs from EMU will constrainthe existence and demographics of the most mas-sive galaxies over the history of the Universe (e.g.Jarvis et al. 2001b,c; Wall et al. 2005). These

sources provide constraints on the co-evolution ofgalaxy bulges and central supermassive black holes(Magorrian et al. 1998;McLure et al. 2006), trace(proto-)clusters at early times (e.g.Miley et al.2006), and may be crucial in determining the im-pact that powerful radio activity has on the hostgalaxy (e.g.Croton et al. 2006;Bower et al. 2006)and its larger scale environment (Rawlings & Jarvis2004;Gopal-Krishna, & Wiita 2001; Elbaz et al.

2009).Infrared-Faint Radio Sources (IFRS) are prob-

ably a particular class of radio source that arecharacterised by a very high radio-infrared ratio(S20cm/S3.6m > 500) and a low infrared flux den-sity. First identified by Norris et al. (2006),there is strong circumstantial evidence that theyare high-redshift radio galaxies, based on their SED(Garn & Alexander 2008; Huynh et al. 2010),their steep spectral index (Middelberg et al.2011), their VLBI cores (Norris et al. 2007;Middelberg et al 2008b), and their extreme faint-ness in the infrared (S3.6m < 1.2Jy; Norriset al. 2011a). Observations suggest 5 IFRSsoccur per deg2 at a flux limit of 100Jy (Nor-ris et al. 2006;Middelberg et al. 2008a), implyingthat EMU will detect at least 1.5 105 , but thelack of corresponding deep infrared data will pre-

vent their identification from other unremarkablenon-detections. However, many will be locatedin smaller deep infrared fields such as those ob-served with HERMES (Oliver et al. 2010), whichshould yield several thousand IFRSs, enabling usto compile solid statistical data on their distribu-tion, spectral index, and polarisation. Thousandsmore will be selected as candidate high-redshiftAGNs through their steep spectrum and polari-sation (see 3.11), and it expected that they willturn out to be an important class of high-redshiftAGN.

A major objective of current extragalactic as-

tronomy is to understand the Epoch of Reionisa-tion (EoR), when ultraviolet photons from the firststars and quasars ionised the primordial neutralhydrogen. This process can in principle be stud-ied by measuring the neutral hydrogen fraction inthe early Universe using the 21 cm forest in front ofa bright distant source (Carilli et al. 2002), analo-gous to the Ly-forest seen against bright quasarsat lower redshifts (e.g.Peroux et al. 2005). Suchobservations require a population of z>6 radio-loud background sources, but the highest redshiftradio galaxy currently known lies at only z=5.2(van Breugel et al. 1999). EMU provides an ex-cellent opportunity to identify such sources, pro-ducing a large sample of distant radio sources forinvestigating the formation and evolution of themost massive galaxies at the highest redshifts.

2.5 CSS and GPS sources

Compact Steep Spectrum (CSS) sources are typ-ically defined to be radio sources with a spectralindex < 0.5, and with a size less than about 20kpc, and hence of sub-galactic dimensions. Gi-gahertz Peaked Spectrum (GPS) sources are ra-dio sources whose spectrum reaches a maximumin the frequency range 110 GHz. Typically theyare significantly smaller (hundreds of pc) than CSSsources, and confined to the circumnuclear regionsof the host galaxy.

Approximately 10% and 30% of bright centime-

tre wavelength sources are GPS and CSS sourcesrespectively. If similar statistics hold for AGNs atlow flux densities, EMU will discover 2 millionGPS sources and 5 million CSS sources. Moststudies of CSS and GPS objects have been con-fined to sources selected from strong source sur-veys, and current samples only probe down to 10mJy at 1.4 GHz (Tschager et al. 2003;Snellenet al. 1998,1999; Randall et al. 2011). EMU willoffer a complete, faint sample of these intriguingobjects, providing a probe into the evolution ofyoung radio AGNs, and showing how these objectsfit into models of galaxy evolution.

There is a consensus (ODea 1998; Fanti

2009a,b;Morganti et al. 2009;Snellen et al. 2009)that GPS and CSS sources represent the start ofthe evolutionary path for large-scale radio sources.It is generally accepted that most GPS sourcesevolve into CSS sources, which gradually trans-form into the largest radio sources known, Fanaroff-Riley Type I and II galaxies (Fanaroff & Riley1974), depending on their initial luminosity. Thesesources offer an ideal resource to investigate earlygalaxy evolution and formation, as well as AGNfeedback, as they are young AGNs but also havestar formation occurring due to interactions andmergers (Labiano et al. 2008; Morganti 2008).


11/43


Measurements of component advance speeds fora few compact sources yield ages of about 103 yr,while spectral studies indicate ages less than about105 yr (ODea 1998; Owsianik et al. 1999; Murgiaet al. 1999).

These sources are thought to be fuelled by theinfall of gas to the supermassive black holes, trig-gered by interactions with companions and merg-ers. CSS and GPS sources show evidence, fromboth their structural and polarisation asymmetries,that the jets are interacting with an asymmetricexternal environment (Saikia et al. 2003; Cottonet al. 2003). CSS and GPS sources show evidenceof Hiabsorption more often than other radio-loudAGNs, with the GPS objects showing the highestincidence of absorption (Gupta et al. 2006, andreferences therein). The EMU sample of CSS andGPS sources over a large redshift range will en-

able the variation of the Hiproperties with cosmicepoch to be determined.

Several CSS and GPS objects show evidence ofdiffuse extended emission which may be a remnantof an earlier cycle of activity. An important goal ofAGN physics is to understand the episodic natureof nuclear or jet activity (Saikia et al. 2009). EMUwith its high surface brightness sensitivity will bean ideal survey to probe the existence of such dif-fuse emission for a large number of sources, andconstrain the time scales of episodic activity forthese compact objects.

Figure 8 This multi-wavelength composite imageof the inner part of the spiral galaxy M83 high-lights the synergy between the two key ASKAPlarge survey science projects: WALLABY andEMU. The ATCA 20-cm radio continuum emis-sion is shown in red, the ATCA H i distributionin green, and the GALEX FUV emission in blue(Koribalski et al. 2008; Gil de Paz et al. 2007).

2.6 Low redshift galaxies, and syn-ergies with WALLABY

WALLABY is expected to detect HI emission from 5105 galaxies to a depth ofz = 0.05, with mas-

sive galaxies detected out to z = 0.25. The ma-jority of these galaxies will be spiral, with typicalH imasses of a few times 109 M. Their large gasreservoir fuels star formation, implying that all spi-ral galaxies detected by WALLABY will have 20-cm radio continuum emission detectable by EMU.As a result, WALLABY is expected to contribute 5 105 redshifts to EMU.

In return, EMU will be able to measure starformation rates of the galaxies studied by WAL-LABY, enabling detailed studies of the factors thatinfluence star formation rates in the local Uni-verse. A goal of WALLABY is to measure local

and global star formation (SF) rates for all gas-richspirals and compare their SF and H idistributions(see Fig.8).

Although most WALLABY data will have a 30-arcsec resolution, high-resolution (10 arcsec) H ipostage stamps will be obtained of particularlyinteresting nearby galaxies, allowing a more de-tailed analysis and comparison with data at otherwavelengths.

2.7 Galaxy clusters

Several different types of diffuse radio emission areassociated with clusters of galaxies (Kempner et

al. 2004), including haloes around the centres ofclusters, relics (representing shocks from cluster-cluster collisions) at the periphery, and tailed radiogalaxies which are an important barometer of theintra-cluster medium. These three classes of radiosource are important as tracers of clusters, and arealso diagnostics of the physics of clusters, particu-larly when combined with X-Ray data. However,the number of detected cluster radio sources is lim-ited by current telescope sensitivities (see Fig 9).EMU will not only give us large samples of suchsources, but will push beyond the present limitsto detect diffuse sources with a range of powers

over a larger redshift range, greatly improving ourunderstanding of these sources.Most clusters have been found either through

X-ray surveys (Rosati et al. 1998; Romer et al.2001; Pierre et al. 2003) or by searching for over-densities in optical colour-position space (Gladders& Yee 2005; Wilson et al. 2008; Kodama et al.2007). As a result, tens of thousands of clusters arecurrently known, but only a few at z> 1 (Wilsonet al. 2008, Kodama et al. 2007), with the highestredshift at z = 2.07 (Gobat et al. 2011).


12/43


2.7.1 Haloes

Radio haloes are found in clusters and groups ofgalaxies, indicating synchrotron emission poweredby diffuse and faint (0.1 - 1 G) magnetic fields

and relativistic particles. So far about 35 radiohalos are known with z < 0.4 and only 2 at z 0.5, generally discovered by making deep radio sur-veys of hot and bright X-ray clusters (Venturi etal. 2008;Giovannini et al. 2009). A strong correla-tion is present between the total halo radio powerand the cluster X-ray luminosity (Giovannini etal. 2009). Since the cluster mass and X-ray lu-minosity are correlated it follows that halo radiopower correlates with the cluster mass (Feretti2000; Govoni et al. 2001). Brunetti et al. (2009);Cassano et al. (2010); Schuecker et al. (2001)have suggested that radio haloes in the centres

of clusters are distributed bimodally, with haloesgenerally found only in those clusters which haverecently undergone a merger, resulting in a dis-turbed appearance at X-ray wavelengths.

The ATLBS survey (Subrahmanyan et al. 2010),which has surveyed 8.4 deg2 to an rms sensitivityof 80 Jy/beam on a scale size of 50 arcsec at1.4 GHz, has detected tens of diffuse sources, ofwhich about 20 have been tentatively identifiedas cluster and group haloes (Saripalli et al.2011). EMU will have even better sensitivity tolow-surface-brightness structures than ATLBS, soif the ATLBS numbers are confirmed, then EMUwill discover 6 104 cluster and group haloes,

which significantly increases the number of knownclusters. An important science goal will then be tocompare the X-ray properties (luminosity, temper-ature and, for the brighter clusters, morphologies)of these radio-selected clusters to those of the X-ray selected population from the eROSITA all-skyX-ray survey (Predehl et al. 2010).

2.7.2 Relics

On the periphery of clusters, elongated radio relicsare found, which probably represent the signaturesof shock structures generated in cluster mergers(Rottgering et al. 1997; Brown et al. 2011a; vanWeeren et al. 2010). They provide important di-agnostics for the dynamics of accretion and merg-ers by which clusters form (Barrena et al. 2009).Large populations of these structures will appearin EMU, and are likely to lead to new cluster iden-tifications especially beyondz 0.5. For example,only 44 radio relics are currently known (Hoeft etal. 2011), and few have been discovered in currentradio surveys because of the relatively poor sen-sitivity of most surveys to low-surface-brightnessstructures. One probable relic has been discov-ered in the seven square degrees of ATLAS (Mid-

delberg et al. 2008a;Mao et al. 2010). Since EMUwill have greater sensitivity to such low-surface-brightness structures than ATLAS, this suggeststhat EMU should detect > 4000 relics, althoughthis number is clearly very uncertain. As a meansof finding clusters, it is less effective than otherradio and X-ray techniques, but will be invaluablefor studying shock structures accompanying clus-ter mergers. Furthermore, relics show evidence ofordered large scale magnetic fields in the periph-ery of galaxy clusters, in regions with a very lowdensity of galaxies and thermal gas.

2.7.3 Tailed radio galaxies

Head-tail, wide-angle tail (WAT), and narrow-angle-tail galaxies (collectively named tailed radio galax-ies) are believed to represent radio-loud AGNs in

which the jets are distorted by the intra-clustermedium (Mao et al. 2009, 2010). They can alsocontribute to the diffuse emission (Rudnick &Lemmerman 2009), especially after the jets fromthe nucleus of the host galaxy have turned off;EMUs high resolution and sensitivity, especiallywhen combined with polarisation information fromthe commensal POSSUM survey (Gaensler et al.2010), will allow us to distinguish these from the(largely unpolarised) cluster-wide halo emission.Even more importantly, such tailed galaxies canbe detected out to high redshifts (Wing & Blanton2011;Mao et al. 2010), providing a powerful diag-nostic for finding clusters. From the WATs discov-ered in the ATLAS fields, Mao et al. (2011c) andDeghan et al. (2011) have estimated that EMUwill detect at least 26000, and possibly as manyas 2 105 WATs, depending on their luminosityfunction and density evolution.

2.7.4 AGN feed back in galaxy clusters andmini-halo sources

The relatively cool and dense gas at the centresof many galaxy clusters and groups emits copiousX-ray radiation by thermal bremsstrahlung andline emission. In the absence of external sources

of heating, this high emission should lead to veryrapid cooling (tcool < 1 Gyr) and very high ratesof mass deposition onto the central cluster galaxy(up to 1000 M/yr), in turn causing very highstar formation rates and strong X-ray line emis-sion (e.g. Fabian & Nulsen 1977; Cowie & Bin-ney 1977;Peterson & Fabian 2006; McNamara &Nulsen 2007). The lack of such obvious observa-tional signatures (e.g. Peterson et al. 2001, 2003)implies that some central source of heating mustbe present. The most plausible source of heatingis feedback from the central AGN. EMU will en-able a statistically significant study of the correla-


13/43


Figure 9 1.4 GHz radio power of detected clus-ter halos as a function of redshift showing thedetection limits of previous cluster observations,adapted fromGiovannini et al. (2009), and thecalculated detection limit of EMU, assuming ahalo with a diameter of 1 Mpc. The upper lineshows the limit corresponding to a scale size of

15 arcmin, which is approximately the largest sizeobject that can be imaged with the VLA unlesssingle-dish data are added to the interferometrydata.

tion between the radio emission from AGNs at thecluster centre with the thermal and non thermalcluster properties, and explore how it evolves withredshift.

Moreover a few cooling-core clusters exhibitsigns of diffuse synchrotron emission that extendsfar from the dominant radio galaxy at the clustercenter, forming what is referred to as a mini-halo.These diffuse radio sources are extended on a mod-erate scale (typically 500 kpc) and, in commonwith large-scale halos, have a steep spectrum anda very low surface brightness. Because of a combi-nation of small angular size and the strong radioemission of the central radio galaxy, the detectionof a mini-halo requires data of a much higher dy-namic range and resolution than those in avail-able surveys. The study of radio emission fromthe center of cooling-core clusters is of large im-portance not only in understanding the feedbackmechanism involved in the energy transfer between

the AGN and the ambient medium (e.g.McNa-mara & Nulsen 2007) but also in the formationprocess of the non-thermal mini-halos. The en-ergy released by the central AGN may also play arole in the formation of these extended structures(e.g.Fujita et al. 2007).

2.7.5 The impact of EMU on cluster re-search

The key impacts of EMU are likely to be:

To increase the number of known clustersbeyond the few tens of thousands currentlyknown. EMU will detect at least 3 104

new clusters (Mao et al. 2011c; Norris etal. 2011c; Deghan et al. 2011), which willroughly double the number of known clus-

ters. Depending on the redshift distribu-tion and luminosity function of cluster radiosources, EMU may detect as many as 2 105 cluster sources (Mao et al. 2011c; Nor-ris et al. 2011c). eRosita is also expected todetect 105 clusters at X-ray wavelengths(Predehl et al. 2010; Pillepich et al. 2011),and comparison of these two complementarysurveys will be transformational.

To detect clusters at high redshifts. In prin-ciple, radio sources can be used to detectclusters even beyond z=1, where current con-straints on large-scale structure are weaker,

and traditional detection techniques like X-ray surveys and use of the Red Cluster Se-quence (Gladders & Yee 2005) become lesseffective. A few clusters have already beendetected up to high redshift using radio de-tections (Blanton et al. 2003, Wing et al.2011), but extrapolation beyond z=1 is un-certain because the luminosity function andevolution of cluster radio sources are unknown.Furthermore, inverse Compton cooling of elec-trons by the cosmic microwave background isexpected to quench their synchrotron radioemission at z 1, although confidence inthis expectation is reduced by the failure todetect a similar effect in the far-IR-radio cor-relation at high redshift (Mao et al. 2011a).

To explore radio properties of clusters de-tected at other wavelengths. It will be im-portant to compare the properties of radio-selected clusters from the unbiased EMU sur-vey to those of the X-ray selected populationfrom surveys such as the eROSITA all-skyX-ray survey (Predehl et al. 2010), and theSunyaev-Zeldovich surveys made with the SouthPole Telescope (Williamson et al. 2011), At-acama Cosmology Telescope (Marriage et al.


14/43


2010) and Planck (Planck Collaboration, 2011)surveys. EMU will enable not only statisti-cal approaches (e.g. halo occupation distri-bution modelling: Peacock & Smith 2000)but also directly observed overdensities usingtracers such as WATs (Mao et al. 2010).

To detect new forms of cluster radio emis-sion. For example, high-resolution numeri-cal simulations (Battaglia et al. 2009) pre-dict an additional network of weak shocksthroughout the cluster volume which can onlybe seen with sufficient resolution and sensi-tivity, but which will be detectable with theshort spacings of ASKAP.

Additional cluster-related goals to be addressedby EMU include:

Do the radio properties of AGNs depend moreon the properties of their host galaxy or oftheir local environment, such as gas temper-ature or cooling time (e.g. Mittal et al.2009)?

Are galaxies of a given mass more likely tohost a radio source if they are a central orsatellite galaxy within a halo?

Is star formation truncated in halos above agiven mass by AGN feedback or virial shockheating?

How does the observed correlation and in-dicated feedback cycle between cluster cen-tral radio AGNs and the cooling intraclustermedium evolve with redshift (e.g. Santoset al. 2010)?

2.8 Cosmic filaments, and the Warm-Hot Intergalactic Medium(WHIM)

Approximately half of the Universes baryons arecurrently missing (Cen & Ostriker 1999) andare likely to be contained in the elusive Warm-Hot Intergalactic Medium (WHIM), where tem-peratures of 105 107 K make them extremelydifficult to detect (Fraser-McKelvie et al. 2011).By subtracting the contribution of compact ra-dio sources, studies of low-surface-brightness emis-sion with EMU can illuminate the otherwise in-visible baryons associated with large scale struc-ture. Perhaps the most important diffuse radiosources are those illuminating shock structures inthe WHIM, which occur during infall onto andalong large-scale-structure filaments (Miniati etal. 2001;Dolag et al. 2008).

Because of the sensitivity and short spacingsof ASKAP, EMU will be able to detect faint radio

emission from cosmological filaments, increasingour knowledge of the physical properties of large-scale structures. The shortest spacing of ASKAP,20m (resulting in sampling of the uv plane downto 8m in a joint deconvolution), ensures that EMUwill be sensitive to structures as large as 1, whilsteven larger scale sizes can be imaged by adding insingle-dish data.

In a few cases, the synchrotron emission fromfilaments may be detected directly. One putativeexample is the bridge of radio emission extend-ing over the 1 Mpc between the central halo ofthe Coma cluster and the peripheral relic source1253+275 (Kim et al. 1989;Kronberg et al. 2007).Another is the radio emission from ZwCl 2341.1+0000which is a linear structure some 6 Mpc in length(Bagchi et al. 2002;Giovannini et al. 2010).

Even though EMU will be able to detect only

the brightest shock structures in the WHIM di-rectly, statistical characterisation of an ensembleof fainter shock emission is possible (Keshet et al.2004;Brown et al. 2010). Faint synchrotron emis-sion due to WHIM shocks is highly correlated withlarge-scale structure (LSS) on 1-4 Mpc scales(Pfrommer et al. 2007;Ryu et al. 2008;Skillman etal. 2008, 2010), and can thus be detected below theEMU noise limits through LSS cross-correlation(Brown et al. 2010) using the statistical redshiftsdescribed in 3.11, or in some cases high-qualityphotometric or spectroscopic redshifts. The skycoverage of EMU coupled with its high sensitivity

to low surface-brightness emission makes it idealfor detecting the synchrotron cosmic-web at cos-mologically important redshifts 0.1 < z < 1.0(Brown et al. 2011b), corresponding to 0.1-0.5 Mpcon arcmin scales.

2.9 Cosmology and Fundamental Physics

2.9.1 Background

The existence of dark energy is one of the mostprofound problems in cosmology. The evidencefor its presence is indirect: it is implied by thesupernovae Type Ia Hubble relation (Riess et al.

1998;Perlmutter et al. 1999), by the combined in-ference of flat geometry (tot = 1) from CosmicMicrowave Background (CMB) measurements andlow mass density (mass 0.3) from large-scalestructure measurements (Cole et al. 2005; Eisen-stein et al. 2005), and from galaxy cluster measure-ments (Vikhlinin et al. 2009; Mantz et al. 2010;Allen et al. 2008).

A further puzzle is the nature of gravity. WhileGeneral Relativity (GR) is consistent with all cur-rent observational measurements (e.g.Uzan 2003),it fails to connect gravity with the other funda-mental forces, models for which lead to hypothe-


15/43


Figure 10 Predicted cross-correlation of the CMB(measured by WMAP etc.) with EMU sources,assuming a standard CDM cosmology. The solidlines is the theoretical prediction, and the shadedregion shows cosmic variance errors.

ses such as brane-worlds (Brax & van de Bruck2003)in which gravity is modified by leakage into

dimensions other than the brane dimensions. Suchmodels predict deviations from GR which have notyet been fully tested.

EMU will be able to perform independent testsof models of dark energy and deviations from GRusing a combination of (a) the auto-correlation ofradio source positions, (b) the cross-correlation ofradio sources with the CMB (the late-time Inte-grated Sachs Wolfe (ISW) effect), and (c) cross-

correlation of radio sources with foreground ob-jects (cosmic magnification).We assume throughout most of this section that

redshifts of EMU sources arenot available, and atthe end consider the impact if statistical redshiftsbecome available. Throughout this section, we as-sume a 5 detection limit, which is expected tobe reliable for the EMU survey, while Raccanelliet al. (2011) use a conservative 10 limit, whichgives rise to slightly different constraints.

2.9.2 Autocorrelations

The auto-correlation function (ACF) of source counts,

also known as the two-point correlation of radiosources, measures the degree of clustering of theradio sources (see e.g.Blake & Wall 2002; Rac-canelli et al. 2011). Here we consider the ACF forEMU sources that we will measure in two dimen-sions on the sky, ignoring any potential availabil-ity of redshifts, and making no distinction betweensource types, which have different biases and num-ber densities as a function of redshift.

Despite these limitations, the ACF is of valuefor two purposes. First, cosmological constraintsfrom this probe can be combined with those fromother probes to substantially improve the net cos-

Figure 11 Predicted constraints from EMU ondark energy parameters, assuming no redshifts areavailable. The outer grey ellipse shows the currentconstraints from Type IA supernovae and CMBfluctuations. The intermediate ellipse shows the

constraints from EMU using the ISW effect, pro-viding only a modest improvement over existingmeasurements. The inner (dark) ellipse shows theconstraints from EMU using all effects, includingcosmic magnification, and shows a significant im-provement over existing measurements. In par-ticular, EMU will resolve the difference betweenthe current best measurement (star) and the value(cross) predicted for a standard CDM cosmologywith non-evolving dark energy.

mological constraint, as shown by the combined

constraints in Figures 11 and 12. Second, thebehaviour of the ACF on large scales constrainsnon-Gaussianity (i.e. the extent to which densityfluctuations in the early universe were distributedwith non-Gaussian noise) hence providing a win-dow on early-Universe physics. Non-Gaussianityindicates a positive skewness of the matter densityprobability distribution, which would lead to anincreased bias for large-scale halos (see e.g.Dalalet al. 2008). EMU will be able to detect non-Gaussianity parameterized by the fNL parameterat the level fNL 8, (Raccanelli et al. 2011).Such a detection would be difficult to reconcilewith a simple, single scalar field inflation modelfor the early Universe.

2.9.3 The Integrated Sachs-Wolfe Effect

The ISW effect can be understood as follows. Trav-elling from the last scattering surface to us, CMBphotons pass through gravitational potential wellsof large structures such as superclusters that liealong our line of sight. In an Einstein-de Sitteruniverse, the blueshift of a photon falling into awell is cancelled by the redshift as it climbs out.However, in a universe with dark energy or mod-ified GR, the local gravitational potential varies


16/43


Figure 12 Forecast of constraints for modified grav-ity parameters, with symbols the same as for Fig.11. The largest ellipse shows constraints from cur-rent measurements, and the smallest ellipse showsthe constraints from EMU using all effects, but

assuming no redshifts are available. EMU will re-solve the difference between the current best mea-surement (star) and the value (cross) for standardGeneral Relativity, either confirming or ruling outsome alternative models.

with time, stretching the potential well while pho-tons are traversing it, so that the blueshift on en-try is not fully cancelled by the redshift on exit.Thus the net photon energy is increased, produc-ing CMB temperature anisotropies which make theCMB appear slightly hotter in the direction of su-

perclusters.The effect is very weak, but can be detected bycross-correlating CMB maps with tracers of large-scale structure (Crittenden & Turok 1996) such asradio sources. The detection of the effect has beenrepeatedly confirmed (e.g. Giannantonio et al.2008a,and references therein); for instance, SDSSgalaxies in the Sloan Digital Sky Survey (SDSSYork et al. 2000) have been cross-correlated withthe CMB (Granett et al. 2008) to achieve a 4-re-sult. In the radio, cross-correlation byRaccanelliet al. (2008) of the NVSS radio galaxies with theCMB anisotropies also gave a tentative detectionof the ISW effect. The NVSS detection is partly

limited by shot noise, and so the far greater num-ber of galaxies in EMU should achieve a robustmeasurement of the effect.

EMU will also be able to constrain the redshiftevolution of the equation of state of dark energy,which would allow us to distinguish between differ-ent models of dark energy, such as a cosmologicalconstant (Tegmark 2004), quintessence (Ziatevet al. 1999), early dark energy (Xia et al. 2009),or Unified Dark Matter models (Bertacca et al.2011).

The ISW effect is also sensitive to any modifi-cations of gravity, allowing sensitive constraints on

gravity, and can also be used to test models for theevolution of the clustering and bias of radio sources(Raccanelli et al. 2008), and to test models for thecosmological evolution of radio sources (Massardiet al. 2010). The predicted cross-correlation func-tion for EMU sources is shown in Fig.10. EMU willprovide sensitive constraints on any modificationsof gravity.

2.9.4 Cosmic Magnification

Large-scale structures along the line of sight toa distant source introduce distortions in the im-ages of high-redshift sources as a result of gravita-tional lensing. The distortions by foreground (low-redshift) galaxies increase the apparent area occu-pied by background (high-redshift) galaxies, thusreducing the observed number density at a given

flux density. On the other hand, the magnifica-tion increases the total flux density of unresolvedhigh-redshift sources, thus increasing the observednumber density at a given flux density. The size ofthe combined change in apparent number densitydue to these two opposing effects is sensitive to theassumed cosmological geometry and parameters ofGR.

The combined effect, known as cosmic magnifi-cation, can be tested by cross-correlating numberdensities of low-redshift sources (e.g. selected fromthe optical surveys or from EMU star-forming galax-ies) with number densities of high-redshift sources(selected from EMU). The effect was first unam-biguously detected byScranton et al. (2005) usingSDSS foreground galaxies and quasars. The largedatabase of EMU sources will develop this into apowerful technique for testing cosmological mod-els.

2.9.5 Observational tests with EMU

Deviations from GR or dark energy physics willinfluence the auto-correlation, the ISW and cosmicmagnification (Zhao et al. 2010). By modellingthe EMU source distribution and bias, Raccanelliet al. (2011) have shown that significant limits

can be placed on cosmological parameters (suchas w and w) that describe dark energy, and onparameters and that describe modifications toGR (Pogosian et al. 2010). GR predicts that = = 1. Figs.11 and 12 show that, even withoutredshifts, EMU will place significant constraintson both Dark Energy and Modified GR.

Once shot noise is fully overcome, ISW mea-surements are limited only by cosmic variance. Sub-stantial improvements over current measurementscan therefore be achieved by using a bigger sur-vey area. Combining the EMU survey with theWODAN survey will allow us to make a radio mea-


17/43


surement of the ISW effect over the entire sky forthe first time, perhaps leading to the highest sig-nificance ISW measurement yet.

2.9.6 Tomographic AnalysisThe discussion of cosmological tests above makesthe conservative assumption that no redshifts areavailable for EMU sources. If statistical redshifts(3.11) are available, the radio sources can be splitinto redshift bins for measuring the ISW effect orcosmic magnification, enabling a tomographicanalysis of auto- and cross-correlation data, lead-ing to even more significant constraints than thosediscussed above. For example, strong polarisedsources are known to have a high median redshift,even when they are undetected in optical surveys.

If there are enough sources in each bin to achieve

a statistically significant ISW detection, EMU willenable a measurement of the redshift evolution ofthe ISW effect and so better constrain cosmologi-cal models. The results at low redshifts will give anindependent confirmation of detections achieved atoptical wavelengths. However, the radio sourcesextend to far higher redshifts, and so test the ISWat epochs that cannot be reached at other wave-lengths: a detection of the ISW at high redshiftwould be a clear signature of a non CDM +GRevolution (Raccanelli et al. 2011).

2.10 Galactic Science

The EMU survey will include the Galactic Plane,thus creating a sensitive wide-field atlas of Galac-tic continuum emission, which can address severalscience goals including:

A complete census of the early stages of mas-sive star formation in the Southern GalacticPlane,

Understanding the complex structures of gi-ant HII regions and the inter-relationship ofdust, ionised gas and triggered star forma-tion,

Detection of the youngest and most compactsupernova remnants to the edge of the Galac-tic disk, some of which may have explodedwithin the past century,

Detection of supernova remnants, especiallythose detected by eRosita but which are un-detected by previous radio surveys,

Detection of planetary nebulae, which arethe most abundant compact Galactic sourcesin the NVSS, and can be useful tools for mea-suring extinction, and estimating the star-formation rate of stars too small to make

SNe or HII regions (Condon & Kaplan 1998;Condon et al. 1999),

Detection of radio stars and pulsars,

Serendipitous discoveries, such as the radioflares from ultra-cool dwarfs found byBergeret al. (2001).

In providing a sensitive, high-resolution con-tinuum image of the Galactic Plane, EMU willcomplement the GASKAP HI Galactic HI survey(Dickey et al. 2010). Existing interferometricradio continuum surveys of the Galactic Plane areeither at high angular resolution but over a limitedsurvey area, or cover a wide area at low angularresolution. For example, the MAGPIS (Helfandet al. 2006) and CORNISH surveys (Purcell &Hoare 2010) cover an area 100deg2 at an an-

gular resolution of 16 arcsec, while the Interna-tional Galactic Plane Survey consists of a num-ber of studies over several hundred deg2 at a typ-ical resolution of 1 arcmin (McClure-Griffithset al. 2005; Taylor et al. 2003; Stil et al. 2006;Haverkorn et al. 2006). EMU, with its full skycoverage, high sensitivity, and 10 arcsec angularresolution, will bridge the gap between these twotypes of survey to reveal newly discovered popu-lations of compact HII regions, planetary nebulaeand young supernova remnants. When combinedwith the WODAN Northern Hemisphere survey(see 1.5), EMU will provide a complete census

of centimetre-wave emission in the Galaxy.At centimetre wavelengths the primary mecha-nisms for emission from Galactic objects are free-free emission from HII regions and planetary neb-ulae, synchrotron radiation emitted by supernovaremnants, and diffuse synchrotron emission emit-ted by relativistic cosmic-ray electrons acceleratedby SNRs. The known radio populations of eachof these types of object are limited by a combi-nation of issues including the limited area coveredby existing surveys, frequency-dependent selectionbias (in the case of optically thick HII regions), orbiases against large scale structure introduced bylimiteduvcoverage snapshot surveys.

Thermal emission from HII regions may be sep-arated from non-thermal synchrotron by using thecorrelation between thermal free-free and infraredemission (e.g.Helfand et al. 2006;Thompson et al.2006; Conti & Crowther 2004). The radio spec-tral index from EMU, and the polarisation datafrom POSSUM, will also be excellent discriminantsbetween thermal and non-thermal emission. Theresolution of EMU is particularly well-matched toMIPSGAL 24 m (Carey et al. 2009) and Hi-GAL 70m (Molinari et al. 2010) survey images,of which the latter is an effective tracer of the in-tensity of the exciting radiation field (Compiegne


18/43


et al. 2010).

2.10.1 Star Formation

EMUs sensitivity allows access to all stages of the

evolution of a compact HII region (hyper-compact,ultracompact and compact), even though the 1.4GHz continuum emission will be optically thick,and EMUs 10 arcsec resolution is insufficient toresolve the ultracompact HII regions.

Ultracompact HII (UCHII) regions representa young (105 years old) phase in the develop-ment of an HII region (Churchwell 2002). MostUCHII regions have been discovered by snapshotVLA surveys at 5 GHz (e.g.Wood & Churchwell1989; Kurtz et al. 1994; Walsh et al. 1998; Pur-

cell & Hoare 2010; Urquhart et al. 2009). How-ever, the snapshot surveys missed an entirely new

class of optically thick HII region known as hyper-compact HII (HCHII) regions (Kurtz 2005) andmissed a diffuse component (Kurtz et al. 1999;Longmore et al. 2009) in which many UCHII re-gions are embedded.

To build a complete census of such objects re-quires a radio survey that is sensitive to both faintobjects and also structure on scales of a few tens ofarcsec. Fig.13shows the typical spectral energydistribution expected from three types of HII re-gion (compact, ultracompact and hypercompact)located at a distance of 18 kpc. In the opticallythick regime the spectral index of free-free emis-sion is proportional to 2, falling to 0.1 in theoptically thin regime (Mezger & Henderson 1967).For compact HII regions this turnover occurs at1.5 GHz, whereas for the denser and more opti-cally thick UCHII and HCHII regions the turnoverfrequency is shifted to higher values. This high-frequency turnover causes HCHII regions to be es-sentially undetectable with snapshot 5 GHz sur-veys (see Fig. 13). Even though HCHII regionsare extremely optically thick at 1.4 GHz, the sen-sitivity of EMU is sufficient to detect them at dis-tances up to 18 kpc, except in regions limited byconfusing strong sources. Consequently, EMU willbe able to detect all types of HII region over most

of the Galaxy, although higher resolution follow-up observations with other arrays may be neededto separate closely associated ultracompact andhypercompact HII regions. Many of these obser-vations will be made as part of higher frequencysurveys like CORNISH (Purcell et al. 2008), andMeerGAL (Thompson et al. 2011). Together withthese surveys, EMU will be able to determine theturnover frequencies (and hence electron densities)of these regions.

Figure 13 Spectral energy distribution of HII re-gions over the frequency ranges of the EMU &CORNISH surveys. The solid, dashed and dot-dashed lines represent the brightnesses of sphericalisothermal and homogeneous hypercompact (HC),ultracompact (UC) and compact (C) HII regionsat an assumed distance of 18 kpc. The red bandshows the frequency range of the 5 GHz CORNISHsurvey and the blue band the range of the 1.3 GHzEMU survey. 5 detection limits are indicated byhorizontal lines.

2.10.2 Supernova Remnants

Only 275 supernova remnants have so far beenidentified within the Milky Way (Green 2009),out of an estimated total population of between500 and 1000 remnants (Helfand et al. 2006). Al-though theGreen (2009) catalogue represents theresult of more than 50 years of intensive searcheswith the world largest radio telescopes, it is con-sidered to be incomplete and strongly biased byselection effects, such as the lack of compact (andhence young) and faint remnants, both of whichshould be addressed by the high sensitivity and

angular resolution of EMU. There also exist strongsynergies between EMU and the high-energy ob-servatories HESS, Fermi & Chandra, which maybe used to confirm EMU non-thermal candidatesas pulsar wind nebulae.

The increased surface-brightness sensitivity andhigher angular resolution of EMU, combined withX-ray data from eROSITA (Predehl et al. 2010)should allow the identification of more than 200faint and diffuse X-ray sources which are currentlyclassified as supernova remnant candidates (Beckeret al. 2009).


19/43


2.10.3 Stellar radio emission

In the last few years, the improvement of observa-tional capabilities has led to the discovery of ra-dio emission in a broad variety of stellar objects

at diverse stages of their evolution, as shown inFig. 14. In many cases, radio observations haverevealed astrophysical phenomena not detectableby other methods (Gudel 2002).

Figure 14 H-R diagram of all known radio stars,taken from Gudel (2002). Radio stars can befound in almost every segment of the stellar H-R

diagram.

Radio stars have so far been detected only bytargeted observations directed at small samples ofstars thought likely to be radio emitters. Existingobservations suffer from limited sensitivity (e.g. noradio stars have been detected at the radio lumi-nosity of the quiescent Sun,Lradio 104 WHz

1),and suffer from a strong selection bias in that ob-servations have been targeted to study a particularaspect of stellar radio emission. Consequently, itis difficult to forecast the impact of EMU on thefield of stellar radio astronomy.

Much of our knowledge of radio stars comesfrom the study of active stars and binary systems(Slee et al. 1987; Drake et al. 1989;Umana et al.1993, 1995;Seaquist et al. 1993; Seaquist & Ivison

1994; Umana et al. 1998; Budding et al. 1999).The strongest radio emission appears to be associ-ated with mass-loss and magnetic phenomena andis often highly variable (Gudel 2002; Dulk 1985;White 2004). Non-thermal radio emission can alsooriginate from shocks of colliding winds in massivebinaries (Dougherty & Williams 2000).

Transient events are also observed as narrow-band, rapid, intense and highly polarised (up to

100 %) radio bursts in stellar objects that have astrong (and often variable) magnetic field and asource of energetic particles, including RS CVnsand flare stars (Slee et al. 2008), Brown dwarfs(Hallinan et al. 2008), and chemically peculiar stars(Trigilio et al. 2000, 2008;Ravi et al. 2010). Theseradio flares have generally been interpreted as aresult of coherent emission mechanisms, includ-ing electron cyclotron maser emission and plasmaemission. This emission has been detected in onlya few tens of stars, partly because of the limitedsensitivity of available instruments, but mainly dueto the absence of a deep all-sky radio survey.

These imperfect statistics suggest that EMUwill provide an unbiased sample of several thou-sand stellar radio sources, enabling a detailed in-vestigation into the physics of radio stars. Thiswill include stellar physical parameters, magnetic

activity, the fraction of active single stars and bi-nary objects that show coherent emission, the time-scales of its variability and their relationship tostellar parameters. These are important diagnos-tics for studies of stellar magnetospheres, and canprobe fundamental parameters such as magneticfield intensity and topology, and electron energydistribution.

However, it will be essential to distinguish ra-dio stars from extragalactic sources using prop-erties such as radio polarisation, SEDs, and theproper motions available from Hipparcos, TYCHO-2 and UCAC3 (Helfand et al. 1999;Kimball et al.

2009).

2.10.4 Pulsars

The number of pulsars at high Galactic latitude ispoorly known, and Cameron et al. (2011) havefound that none of the IFRS detected in the AT-LAS survey (Norris et al. 2006; Middelberg etal. 2008a) are pulsars, implying that the surfacedensity of pulsars with a continuum flux level >150Jy at high Galactic latitude is 50Jy should occur roughly every four deg2

, inwhich case EMU will detect about 8000 pulsars,which exceeds the total number of currently knownpulsars. However, it will be difficult to distin-guish these from other continuum sources. Diag-nostics will include (a) lack of an optical/IR iden-tification, (b) steep spectrum, (c) polarisation, es-pecially circular polarisation (d) sidelobes causedby variability and scintillation during the observa-tion. Candidates satisfying these criteria will willbe searched for pulsar emission using the ASKAPCOAST (Ball et al. 2009) project or conventionalsingle-dish pulsar searches.


20/43


2.11 Unexpected outcomes

Experience has shown that whenever the sky is ob-served to a significantly greater sensitivity, or a sig-nificantly new volume of observational phase space

is explored, new discoveries are made. Even AT-LAS, which expanded the phase space of wide-deepradio surveys by only a factor of a few, identified apreviously unrecognised class of object (IFRS: see2.4). Because EMU will be much more sensitivethan any previous large-scale radio survey, it islikely to discover new types of object, or new phe-nomena. Although it is impossible to predict theirnature, we might reasonably expect new classes ofgalaxy, or perhaps even stumble across new Galac-tic populations. Furthermore, the large EMU dataset,covering large areas of sky, will be able to identifyextremely rare objects.

Historically, discoveries of new classes of ob-jects occur when an open-minded researcher, in-timate with the telescope and with their science,recognises something odd in their data (Ekers2010;Bell-Burnell 2009). EMU is unlikely to bedifferent: people will make unexpected discoverieswhile carefully using EMU data to test hypothe-ses. It is therefore arguable that this process needsno planning: any new class of object in the datawill eventually be discovered anyway. On the otherhand, given the large volume of data, it is possi-ble that a class of objects will lie undiscovered fordecades because it didnt happen to fall within theselection criteria of any astronomer.

It is therefore important for EMU to plan toidentify new classes or phenomena, rather thanhoping to stumble across them. EMU is there-fore taking the novel approach of developing data-mining techniques to identify those objects thatdont fall into one of the known categories of as-tronomical object. A data-mining project (namedWidefield ouTlier Finder, or WTF) is being estab-lished that will attempt to assign each object in theEMU catalogue to a known class of astronomicalobject, using the available cross-identifications tocompare colours, luminosities, and any other avail-able data. It will then identify those that are out-

liers or which depart systematically from knownexamples. Most such outliers will simply repre-sent bad data, and are thus valuable in their ownright for debugging ASKAP, while a few may beexciting new discoveries.

2.12 Legacy value

The largest existing radio survey, shown in Fig.1, is the NRAO VLA Sky Survey (NVSS), whoserelease paper (Condon et al. 1998) is one of themost-cited papers in astronomy. EMU will coverthe same fraction ( 75%) of the sky as NVSS, but

will be 45 times more sensitive, 4.5 times higherin resolution, with higher sensitivity to extendedstructures. As a result, EMU will detect 38times as many sources as NVSS. More importantly,the greater sensitivity means that EMU breaksinto a different regime. While most NVSS sourcesare radio-loud AGNs, EMU will provide estimatesof star formation rate and radio-mode accretionactivity in the galaxies currently being studied bymainstream astronomers at all other wavelengths.

The legacy value of large radio surveys dependson the ability to obtain a radio image of any ob-

ject being studied at other wavelengths. Thus thelegacy value depends critically on the area of skycovered: a survey covering the entire sky has morethan simply twice the scientific value of a surveycovering half the sky. For example, the numberof powerful (FRII-type) radio sources predicted at

the epoch of reionisation is only 100 over the en-tire sky. The ability to identify radio counterparts(or upper limits) for any future observation (forexample the highest redshift gamma-ray burst, su-pernova host galaxy, etc.) is essential. Serendip-itous discoveries, too, are most likely to be max-imised by surveying the greatest possible area. Asa result, EMU extends as far North as +30 dec-lination, and, together with WODAN (see 1.5),will cover the entire sky. The combined EMU andWODAN catalogues are expected to become theprimary radio source catalogue for all astronomers,and will not be superseded until after the SKA be-

gins operation.

3 EMU Design Study

3.1 The need for a Design Study

In the two years prior to commencement of theEMU Survey, the EMU Design Study is addressingthe following challenges:

Developing an optimum observing strategy,while taking into account the needs of poten-tially commensal surveys,

Working with the ASKAP computing groupto ensure an optimum processing pipeline,and an optimum source extraction algorithm,including extraction of extended sources,

Addressing technical issues of dynamic range,confusion, etc.,

Developing an optical/IR identification pipeline,including a citizen science project,

Exploring the use of Statistical Redshifts (see3.11),


21/43


Developing the data access process, and iden-tifying potential data issues,

Developing value-added projects to max-imise the science return from EMU,

Refining the EMU Science Goals, to ensurethat the EMU survey is optimised to addressthem, and planning early science papers.

The EMU Design Study takes place through anumber of working groups who each take responsi-bility for a number of goals, milestones, and deliv-erables. The overall Design Study goal is to havea preliminary processing and analysis pipeline inplace by November 2011, with a fully-functioningpipeline in place by November 2012.

3.2 The role of ATLAS in the EMUDesign Study

An important component of the EMU Design Studyis the ATLAS project (Norris et al. 2006;Middel-berg et al. 2008a; Hales et al. 2011; Banfield etal. 2011) , which has a sensitivity, resolution andscience goals similar to EMU, but over a muchsmaller area of sky (7 deg2 surrounding the CDFSand ELAIS-S1 fields). The ATLAS survey, to-gether with earlier observations of the HDFS (Nor-ris et al. 2005;Huynh et al. 2005), is being used asa test-bed for many of the techniques being devel-oped for EMU, as well as guiding the development

of the science goals and survey strategy.For example, the prototype EMU source ex-traction and identification pipeline will be used forthe final ATLAS data release in late 2011. Duringthe course of the ATLAS project, dynamic rangesapproaching 105 have been reached, which areamongst the highest achieved in radio astronomy,and imaging artefacts have been encountered whichhave been rectified in some specific cases, and a so-lution for the general case is being tackled. Suchimaging problems must be solved not only for EMU,but for other SKA Pathfinder projects.

3.3 Science Data ProcessingThe data from the correlator will be reduced inan automated processing pipeline (Cornwell et al.2011), which includes the following steps for EMU:

Flag for known radio frequency interference,working from a database of known RFI sources,

Identify unknown radio frequency interfer-ence, saving candidate identification, and iden-tify and flag further bad data,

Solve for calibration parameters (i.e. frequency-dependent complex gains) using least squares

fits of the predicted visibility (from previ-ously obtained model of the sky) to the ob-served visibility,

Apply calibration parameters, predicting for-

ward from the last previous solution,

Average visibility data to required temporaland spectral resolution,

Construct an image by (a) gridding the datausing convolutional resampling, (b) Fouriertransforming to the image plane, and (c) De-convolving the point spread function,

Find sources in the resulting image or cube,

Save science data products to the ASKAPScience Data Archive.

The measured flux densities from most radiosurveys have a typical accuracy of 10%, althoughthe calibration accuracy should, in principle, en-able flux densities to be measured to an accuracyof 1%. This latter figure has been adopted asthe target for EMU, and the EMU Design Studywill investigate the reason that most surveys failto achieve this figure. A number of instrumen-tal effects and completeness corrections have beenidentified (Hales et al. 2011), and may need to beapplied to the data, including the following.

The primary beam model must be accurate

to 1% so that the data can be corrected forthe primary beam response to this accuracy.In the case of ASKAP, this will vary acrossthe PAF, and so needs to be determined foreach of the 36 beams.

Position-dependent bandwidth smearing (chro-matic aberration) over the mosaiced imagedue to the finite bandwidth of frequency chan-nels (Condon et al. 1998;Ibar et al. 2009).Surprisingly, mosaiced images suffer more frombandwidth smearing than pointed observa-tions, since any location in the image, evenat the centre of a pointing, may include many

contributions from adjacent pointings in whichbandwidth smearing is significant. Thus, evenfor tight angular spacing between pointingsin a mosaic, bandwidth smearing will alwaysbe non-zero at any location over the final mo-saiced image. Hales et al. (2011) found thatbandwidth smearing in central regions of theATLAS mosaiced images, which used 8 MHzchannels, typically caused a 10% decrementin peak flux due to this overlapping effect.Even with ASKAPs 1 MHz channels, thiscorrection will be necessary if 1% accuracyis to be achieved.


22/43


Position-dependent time-average smearing overthe mosaiced image due to finite integrationtime and Earth rotation (Bridle & Schwab1999). This is thought not to be a signifi-

cant issue for ASKAP, but is included herefor completeness.

Clean bias in measured fluxes, which may bemitigated by suitable weighting schemes (tocontrol beam sidelobe properties) and clean-ing depth.

In addition, there are observational effects whichmay bias the fluxes of individual sources and/orthe statistical properties of the sample as a whole,as follows.

The probability of measuring a faint sourcelocated on a noise peak is higher than theprobability of measuring a strong source lo-cated in a noise trough, because faint sourcesare more numerous (Hogg & Turner 1998).This results in a bias on all measured sourcefluxes, which decreases with increasing signal-to-noise. When considered in terms of sourcecounts this is known as Eddington bias (Ed-dington 1913; Simpson et al. 2006). Haleset al. (2011) have shown that Eddingtonbias can also be a sensitive probe of numbercounts below the flux limit of a survey.

Resolution bias due to the lack of sensitivity

to resolved sources, which can be manifestedin two ways. First, the lack of short baselinescan limit the maximum angular scale observ-able, although this is not likely to be a prob-lem for EMU. Second, faint resolved sourcesmay have integrated fluxes sufficient to beincluded in the final catalogue, but may bemissed in the source extraction process be-cause their peak fluxes fall below the signal-to-noise source detection threshold (e.g.Pran-doni et al. 2001). This necessitates a resolu-tion bias correction.

Sensitivity will generally vary across an im-

age, so that the area surveyed to any limit isa function of that limit. Since source countsmust be normalised by the area surveyed, acompleteness correction must be applied toaccount for the position-dependent sensitiv-ity to faint sources across the mosaiced image(e.g.Bondi et al. 2003), which must also takeinto account the position-dependent time-averageand bandwidth-smearing effects (Hales et al.2011). For the same reason, large variationsin sensitivity across the field are likely to leadto uncertainties in source counts, and so itis essential to measure the sensitivity across

the field to high accuracy, and keep it as uni-form as possible.

3.4 Simulations and Imaging Pipeline

The goals of the simulations and imaging pipelineworking group are to:

Develop a realistic extragalactic simulatedsky,

Develop a realistic Galactic simulated sky,

Work with ASKAP engineers to choose opti-mum PAF configuration and weighting scheme,

Ensure ASKAP imaging processes are robustto Galactic-Plane observations,

Optimise imaging algorithms and parame-ters,

Ensure that EMU data will reach a dynamicrange of 105,

Ensure that EMU measured flux densitiesare accurate to 1%.

3.4.1 Simulation Analysis

Simulations have been performed using the SKADesign Study (SKADS) simulated sky (Wilman etal. 2008) as initial input. This sky is augmentedby the addition of extended, diffuse, or complex

sources, which are not fully represented in the SKADSsky. This sky is then used to generate uvdata sim-ilar to that which will be produced by ASKAP,using the antenna, beam and array characteris-tics. Thermal noise is added to the visibilities butcalibration is assumed to be ideal. The data arethen processed using the ASKAPsoft processingpipeline (Cornwell et al. 2011) which is the pro-totype version of the ASKAP processing software.Current simulations use all 36 antennas, the full300 MHz bandwidth, and 8 hours of on-source timewith approximately uniform weighting. A typicalobserved simulated sky is shown in Fig. 15. Re-

sults so far suggest that the majority of fields willreach the required sensitivity and dynamic range,and the rest of this section considers those few re-maining fields containing strong sources, such asthat shown in Fig. 16,where this may not be thecase.

The analysis of the simulations focusses on thedifference image, which is the difference betweenthe dirty image and the input model convolvedwith the point spread function (PSF). In regionscontaining no strong sources, the rms noise in thedifference image is close to the value predictedfrom receiver performance (i.e. 10Jy for an EMU


23/43


Figure 15 Part of the simulated sky as observedby the simulated ASKAP telescope and processedusing ASKAPsoft. (Left) The input model sky.(Right) The difference between the observed skyimage and the input sky showing artefacts causedby the observing process and deconvolution errors.The intensity scale of both images is increased tohighlight these errors.

Figure 16 (Left) The simulated sky around Pictor-A as observed by the simulated ASKAP telescopeand processed using ASKAPsoft. The greyscaleranges from -0.01 to 0.03 Jy/beam, with a peakof 0.9 Jy/beam. (Right) The difference between

the observed sky image and the input sky showingartefacts. This image is enhanced by a factor often to show up artefacts, so the greyscale rangesfrom -0.001 to 0.003 Jy/beam.

image). However, it is higher in regions containingbright or extended sources, because of the follow-ing effects.

Insufficient sampling in the image domain,particularly for bright point sources that arenot pixel-centredBriggs & Cornwell (1992).

Simulations suggest that 4 pixels/beam willlimit the dynamic range to 2000-3000 inthese regions, and so 5-6 pixels across thebeam are planned to achieve the required 105

dynamic range..

Insufficient scales for the multi-scale CLEAN.The three scales used in earlier simulationswere insufficient to deconvolve the more ex-tended sources in the field, and so five scales(as used in the simulation results shown inF

Documents

White Matter Physics