Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.atnf.csiro.au/people/Ray.Norris/papers/n273.pdf Äàòà èçìåíåíèÿ: Fri Jun 29 04:39:08 2012 Äàòà èíäåêñèðîâàíèÿ: Fri Feb 28 08:12:56 2014 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï

Astronomy & Astrophysics manuscript no. ATLAS2.3GHz-catalogfinal June 28, 2012

c ESO 2012

The Australia Telescope Large Area Survey: 2.3 GHz observations of ELAIS-S1 and CDF-S
Spectral index properties of the faint radio sky
Peter-Christian Zinn1,2 , Enno Middelberg1 , Ray P. Norris2 , Christopher A. Hales3,2 , Minnie Y. Mao Randall2,3
1 2 3 4 5

2,4,5

, and Kate E.

¨ Astronomical Institute of Ruhr-University Bochum, Universitatsstraúe 150, 44801 Bochum, Germany e-mail: zinn@astro.rub.de CSIRO Astronomy & Space Science, PO Box 76, Epping, NSW, 1710, Australia Sydney Institute for Astronomy, School of Physics, The University of Sydney, NSW 2006, Australia Australian Astronomical Observatory, PO Box 296, Epping, NSW 1710, Australia School of Mathematics & Physics, University of Tasmania, Private Bag 37, Hobart, 7001, Australia

Received 04/05/2012; accepted 22/06/2012
ABSTRACT

Context. The Australia Telescope Large Area Survey (ATLAS) aims to image a 7 deg2 region centred on the European Large Area ISO Survey - South 1 (ELAIS-S1) field and the Chandra Deep Field South (CDF-S) at 1.4 GHz with high sensitivity (up to 10 µJy) to study the evolution of star-forming galaxies (SFGs) and Active Galactic Nuclei (AGN) over a wide range of cosmic time. Aims. We present here ancillary radio observations at a frequency of 2.3 GHz obtained with the Australia Telescope Compact Array (ATCA). The main goal of this is to study the radio spectra of an unprecedented large sample of sources ( 2000 observed, 600 detected in both frequencies). Methods. With this paper, we provide 2.3 GHz source catalogues for both ATLAS fields, with a detection limit of 300 µJy (equivalent to 4.5 in the ELAIS-S1 field and 4.0 in the CDF-S). We compute spectral indices between 1.4 GHz and 2.3 GHz using matchedresolution images and investigate various properties of our source sample in dependence of their spectral indices. Results. We find the entire source sample to have a median spectral index med = -0.74, in good agreement with both the canonical value of -0.7 for optically thin synchrotron radiation and other spectral index studies conducted by various groups. Regarding the radio spectral index as indicator for source type, we find only marginal correlations so that flat or inverted spectrum sources are usually powered by AGN and hence conclude that at least for the faint population the spectral index is not a strong discriminator. We investigate the z­ relation for our source sample and find no such correlation between spectral index and redshift at all. We do find a significant correlation between redshift and radio to near-infrared flux ratio, making this a much stronger tracer of high-z radio sources. We also find no evidence for a dependence of the radio­IR correlation on spectral index.
Key words. catalogues ­ galaxies: active ­ galaxies: evolution ­ radio-continuum: galaxies ­ surveys

1. Introduction
Large-scale radio surveys are the primary instrument in the toolbox of contemporary astronomy to investigate demography and evolution of active galaxies across cosmic time. Consequently, there have been numerous attempts to use this tool, whether in a wide, shallow layout (e.g. the National Radio Astronomy Observatory Very Large Array Sky Survey, Condon et al. 1998) or in a deep, but more pencil beam-like layout such as for the Cosmic Evolution Survey (COSMOS, Schinnerer et al. 2010) or even the Lockman Hole survey by Owen & Morrison (2008), the most sensitive radio observation to date. The Australia Telescope Large Area Survey (ATLAS, Norris et al. 2006; Middelberg et al. 2008) attempts to break this dichotomy by obtaining deep images ( 10 µJy in the final data release 3) over an unprecedented large area of sky. Distributed over the well-studied Chandra Deep Field South (CDF-S) and the European Large Area ISO Survey ­ South 1 (ELAIS-S1) field, the total area covered by ATLAS is approximately seven square degrees. This allows us to study the faint radio sky without being affected by cosmic variance (Moster et al. 2011).

Furthermore, the wealth of ancillary photometric data ranging across the entire electromagnetic spectrum (e.g. from SWIRE, COMBO-17 or GEMS Lonsdale et al. 2003; Wolf et al. 2003; Rix et al. 2004) as well as numerous spectroscopic campaigns (Sacchi et al. 2009; Popesso et al. 2009; Balestra et al. 2010; Mao et al. 2012) enables one to conduct detailed studies of radiodetected objects. In this paper, we focus on radio spectral properties of the ATLAS source sample. The main ATLAS data releases are DR 1 by Norris et al. (2006) and Middelberg et al. (2008), DR 2 by Hales et al. (in prep.), and DR 3 by Banfield et al. (in prep.). These data consist of 1.4 GHz observations as for most contemporary radio surveys, and we here supplement these data with observations at 2.3 GHz. The main goal of this is to calculate spectral indices log (S 1.4 GHz /S 2.3 GHz ) / log (1.4 GHz/2.3 GHz) for all sources detected at both wavelength and so to characterise their radio emission mechanism, source type and dependence on other spectral quantities. Even though many blind surveys of extragalactic radio sources are available already, only few studies have been carried out which characterise the radio spectra of a large source
1


Peter-Christian Zinn et al.: ATLAS 2.3 GHz observations

population by means of their spectral index. For example, Bondi et al. (2007) and Ibar et al. (2009) use supplementary 610 MHz data to calculate spectral indices at frequencies below 1.4 GHz with the main aim to identify steep spectrum sources which are regarded as good candidates for being high-redshift objects (see Sect. 4.3). The most comprehensive study of spectral indices between 1.4 GHz and a higher frequency was conducted by Prandoni et al. (2006) using 5 GHz Very Large Array (VLA) data in the Australia Telescope ESO Slice Project survey field. With a sensitivity of 70 µJy over an area of one square degree, their source sample consists of approximately 100 objects which are detected at both frequencies, so our sample exceeds this by a factor of six. According to Komatsu et al. (2011), we adopt a flat CDM cosmology with H0 = 70.2 km s-1 Mpc-1 and = 0.725 throughout this paper.

Table 1. Parameters of the observations. Given are the number of pointings per field, the average integration time per pointing, the total area covered by the pointings, the rms noise and the resolution of the final images. Field ELAIS-S1 CDF-S Npnt 76 88 tint h 6.17 4.13 area deg2 2.77 3.63 rms noise µJy/beam 70 80 resolution arcsec 33.56 â 19.90 57.15 â 22.68

Notes. For comparison, we note that the corresponding typical rms noise levels of the 1.4 GHz images are 31 µJy and for 38 µJy the ELAISS1 and CDF-S, respectively.

2. Observations and imaging
In the following description, we give the actual numbers for all things concerning the ELAIS-S1 field first, followed by the numbers for the CDF-S in parentheses. If there is only a single number given, it is approximately equal for both fields. Observations at 2.3 GHz for the ATLAS fields were carried out between December 2006 and March 2007 on 52 separate days where the ATCA was only in its 750B and 750C configurations. This resulted in a substantially lower resolution than at 1.4 GHz since the longest baselines used for imaging at 2.3 GHz was 750 m, or 5.7 k, resulting in an angular resolution of 33 (57 ), compared to 30 k at 1.4 GHz. The ATCA primary beam at a frequency of 2.3 GHz has a FWHM of 21 arcmin, therefore 76 (88) overlapping pointings were observed in a hexagonal lattice with an average integration time of 6.17 h (4.13 h) per pointing (compared to approx. 12 h per pointing for the 1.4 GHz image) to cover the entire area of the 1.4 GHz maps. For phase calibration, which was done approximately every 45 min, PKS 0022-423 (PKS 0237-233) was observed as secondary calibrator. Flux calibration was done for both fields using PKS 1943-638 which was observed at least once a day for 10 min. As for 1.4 GHz, the ATCA provides two independent frequency bands with a bandwidth of 128 MHz divided into 33 channels. During the first two days of observations for the ELAIS-S1 field, the two bands were centred at 2.368 GHz and 2.560 GHz in an attempt to reduce local Radio Frequency Interference (RFI). The fluxes density of the primary calibrator, PKS 1934-638, was assumed to be 11.589 Jy and 10.932 Jy, respectively. Because using a higher frequency for the upper band did not significantly decrease the RFI, the remaining time was observed with the standard band configuration, meaning that the upper band was lowered in central frequency to 2.496 GHz. There, the primary calibrator was assigned to have a flux density of 11.138 Jy. We note that the final images were made with the entire data set. The calibration of the data were carried out using standard Miriad (Sault et al. 1995) tasks. ATLOD was used to convert the raw data from RPFITS format to the native Miriad format. For the further calibration, both channels at the end of the bands were discarded due to a significant lack of sensitivity, hence a data set containing two times 13 channels with 8 MHz bandwidth each was obtained, yielding a net bandwidth of 208 MHz. Fortunately, our data do not suffer from self-interference typical for the ATCA. These so-called "birdies" occur within the samplers at integer multiples of 128 MHz. The 19th harmonic at 2.432 GHz fortunately lies exactly between our two bands and
2

therefore does not affect our data. After bandpass-calibration, RFI was removed using Pieflag (Middelberg 2006), a tool that compares statistics from channels with no or little RFI contamination to the other channels in the data and looks for outliers and sections with high noise. Bandpass calibration was carried out using the observations of PKS 1934-638. Phase calibration was done using the regular scans of the secondary calibrators, and then amplitude calibration was carried out using PKS 1934-638. The data were split into the individual pointings which were imaged separately using uniform weighting. This provides a higher resolution at the cost of lower sensitivity than natural weighting. The cell size was set to 6 (3 ). Because of the low fractional bandwidth (< 10%), a normal CLEAN procedure was used (in contrast to the Miriad multi-frequency clean implementation MFCLEAN for the 1.4 GHz data) for deconvolution. Also the dynamic range in the pointings was sufficiently low as to not require any special deconvolution procedures. Cleaning was carried out using 2000 iterations, after which only marginal sidelobes around the two strongest sources remained, and the effects of clean bias should be negligible. After cleaning, the models were convolved with a restoring beam of 33.56 â 19.90 (57.15 â 22.68 ) at a position angle of -1.32 deg (-1.90 deg). The final mosaics were produced using Miriad's LINMOS task to account for primary beam attenuation and overlapping regions of the individual pointings. As a result, the edges of the final images exhibit a significantly higher noise level than the central regions (see Fig. 1).

3. The source catalogue
Source extraction from the final 2.3 GHz images was performed in two steps: First, all known 1.4 GHz sources from ATLAS DR 2 (Hales et al. in prep.) were used as an input catalogue to search for 2.3 GHz counterparts. This catalog contains a total of 2240 individual 1.4 GHz sources across both the ELAIS-S1 and the CDF-S of which 631 could be reliably detected at 2.3 GHz. To identify these counterparts, the 2.3 GHz image was inspected by eye at the location of every individual 1.4 GHz source. This was done by half-automatically browsing through an input catalogue, simultaneously displaying the corresponding regions in the 1.4 GHz and 2.3 GHz images and, if a 2.3 GHz counterpart was found, fitting a Gaussian to it and recording all fit parameters in an output catalogue. Thus, no blind search was conducted at 2.3 GHz, allowing the inclusion of sources fainter than the typical 5 cutoff in the catalogue. The lowest signal-to-noise ratio (SNR) among the 2.3 GHz sources is 3.9. In a second step, we checked for "2.3 GHz-only" sources. Those sources are unlikely to occur because of the noise levels of the 1.4 GHz and 2.3 GHz images. Since the faintest 1.4 GHz


Peter-Christian Zinn et al.: ATLAS 2.3 GHz observations

Fig. 1. 2.3 GHz noise maps of both ATLAS fields: CDF-S (left) and ELAIS-S1 (right). Contours start at 70 µJy beam and increase by a constant value of 40 µJy beam-1 .

-1

(dark blue contour line)

sources in the DR 2 catalogue are as faint as 100 µJy, whereas the faintest sources in 2.3 GHz are three times brighter, a 2.3 GHzonly source would have a highly inverted spectral index of 2.4. Also, we note that, due to the temporal offsets between the 1.4 GHz and 2.3 GHz observations, possible variability of some radio sources may lead to flawed spectral indices. Nevertheless, we used MIRIAD's SAD task to automatically produce a 2.3 GHz catalogue with a detection threshold of 5 and matched the sources obtained that way to the 2.3 GHz catalogue obtained by cross-matching to 1.4 GHz sources. All sources identified by SAD could be matched to sources previously found as 1.4 GHz counterparts, hence not a single 2.3 GHz-only source was found. To compute spectral indices, the flux densities in 1.4 GHz had to be re-measured because the resolutions of the 1.4 GHz and 2.3 GHz images differ significantly. Whilst the 1.4 GHz observations included long baseline configurations of the ATCA (e.g. the 6A and 6B configurations), the 2.3 GHz observations were obtained with the ATCA in 750B and 750C configurations only. This leads to a resolution three to five times lower at 2.3 GHz compared to the 10 resolution of the 1.4 GHz observations, despite the shorter wavelength. Therefore, in order to avoid resolution effects, we convolved the original 1.4 GHz images with a Gaussian kernel to match the resolution of the 2.3 GHz images as given in Table 2. We then used the ATLAS DR 2 catalogue again as input to re-measure the 1.4 GHz flux densities in the low-resolution images. Because the image sensitivity was reduced by a factor of 2 by this convolution (which is equal to tapering the data in the uv-plane, so in fact weighting down long baselines), all sources in the low-resolution image should be a true subset of the hi-res sources and, therefore, all low-resolution sources should be detected when using the highresolution catalog as input sample. During the re-measurement of the 1.4 GHz flux densities using convolved images, a problem surfaced with the ELAIS-S1 image. Unlike with the CDF-S convolved 1.4 GHz image, where the procedure outlined above went well, the MIRIAD task imfit overestimated the flux densities by a significant amount (up to 50%). This became obvious by inspecting residual images for sources with various flux densities and degrees of compactness. While the residuals looked good in the CDF-S, they showed bowls of negative flux in the ELAIS-S1 image. The cause of

the effect remains unclear. The fitting routines in the software package CASA (Jaeger 2008) yielded similarly overestimated flux densities, so it appears that the effect is not caused by the fitting software, but is inherent in the images, and we had to work around this issue. This was done by iteratively fitting Gaussians to every individual source. The first Gaussian fit (which was always overestimating the flux density) was used as prior. The resulting residual image was then examined in terms of total flux and rms noise. If the total flux in the residual image (trimmed to a size twice as large as the restoring beam) was negative and the rms noise was deviating from the median rms in the region of the source by more than 10%, the fit was discarded and replaced by a fit with a peak flux density 2% lower than in the previous fit. The residual image was again examined in a similar way. This procedure was iteratively repeated until the criteria (non-negative total flux and rms noise consistent with the rms in this region) were entirely met. As a last step, all final residual images were inspected by eye to assure the quality of the fit. A further problem in calculating spectral indices for every 1.4 GHz was introduced by the very different resolutions of the 1.4 GHz and 2.3 GHz observations. Because the resolutions differ by a large factor, as explained above, numerous sources which are clearly separated in the high-res images are blended together in the low-res images. Therefore, we chose to completely exclude such blended sources from our final catalogue if they could not be separated by simultaneously fitting two or more Gaussians. We note that this may bias our sample towards fainter sources since classic radio doubles (which are mainly excluded by the blending) are generally brighter than single, isolated sources. The final catalogue was then constructed by merging the individual catalogues from ELAIS-S1 and CDF-S. This resulted in a final catalogue comprising 631 radio sources with clear detections and well-measured flux densities in both 1.4 GHz and 2.3 GHz. A portion of the catalogue is shown in Table 2 for illustration.

3


4 S
1.4 GHz,hi

Table 2. A representative portion of the ATLAS 2.3 GHz source catalogue of the ELAIS-S1 field for illustration purposes. The CDF-S catalogue contains the same information. S
1.4 GHz,hi

ID mJy mJy
1.4 GHz,lo

a

RA (J2000) deg mJy
1.4 GHz,lo

DEC (J2000) deg mJy
1.4 GHz,lo

S mJy
2.3 GHz

S mJy
2.3 GHz

rm s mJy
2.3 GHz

S mJy

S

rms

2.3 1.4

2.3 1.4

z

class

b

0.685

254 178 165 163 102 290 201 150 264 323 0.211 0.067 0.902 0.140 SF

51.770792 51.621458 53.319708 52.470667 52.561167 53.190750 52.399667 52.122000 52.105750 52.808292

-28.650297 -27.427269 -27.941922 -27.492661 -28.566331 -28.520756 -27.950881 -28.030825 -27.745814 -28.785917

1.260 2.740 1.131 2.160 0.194 0.714 1.420 1.810 0.920 0.834

0.084 0.145 0.054 0.116 0.037 0.044 0.087 0.087 0.054 0.057

1.230 2.220 0.762 2.070 0.162 0.598 1.550 1.770 0.745 0.818

0.093 0.165 0.158 0.985 0.155 0.126 0.153 0.281 0.176 0.160

0.069 0.122 0.153 0.980 0.154 0.123 0.132 0.267 0.173 0.155

0.561 1.519 0.828 1.810 0.344 0.400 1.154 1.016 0.674 0.544

0.080 0.107 0.084 0.116 0.074 0.075 0.093 0.088 0.080 0.077

0.075 0.075 0.073 0.073 0.072 0.072 0.072 0.072 0.072 0.072

-1.58 -0.76 0.17 -0.27 1.52 -0.81 -0.59 -1.12 -0.20 -0.82

0.11 0.11 0.13 0.95 0.64 0.14 0.06 0.30 0.46 0.38

AGNe Peter-Christian Zinn et al.: ATLAS 2.3 GHz observations

Notes. The columns are as follows: (1) unique ID (2) degrees of J2000 right ascension (3) degrees of J2000 declination (4,5) 1.4 GHz integrated flux density and flux density error as given in Hales et al. (in prep) (6,7,8) integrated flux density, flux density error as calculated using Eqn. 1 and local rms noise as measured from the low-resolution 1.4 GHz image (9,10,11) integrated flux density, flux density error as calculated using Eqn. 1 and local rms noise as measured from the 2.3 GHz image (12,13) spectral index calculated from the low-resolution 1.4 GHz flux density and the 2.3 GHz flux density and its error as calculated by Eqn. 4 (14,15) spectroscopic redshift and spectral classification from Mao et al. (2012) if available. The entire table is available via the on-line version of the journal. (a) Unique only for the respective field (ELAIS-S1 or CDF-S). (b) Spectroscopic classification from Mao et al. (2012). SF indicates an optical spectrum dominated by star formation activity while AGN indicates an optical spectrum typical for AGN (e.g. AGNe for an AGN with characteristic emission lines). E indicates that the spectrum is that of an early-type galaxy and hence the radio emission must originate from an AGN hosted by it.


Peter-Christian Zinn et al.: ATLAS 2.3 GHz observations

We give an identifier for each source, the position, the original 1.4 GHz flux density as measured by Hales et al. (in prep.), the re-measured 1.4 GHz flux density using the convolved image, the 2.3 GHz flux density as measured during the crossmatching process and finally the spectral index calculated from the matched-resolution 1.4 GHz and 2.3 GHz flux densities. All flux densities are integrated quantities. Errors for the flux densities were either taken from Hales et al. (in prep.) in case of the original 1.4 GHz flux densities or calculated following Hopkins et al. (2003) for the low-resolution 1.4 GHz flux densities and the 2.3 GHz flux densities, taking into account both the local rms noise at the source position as well as the quality of the Gaussian fit: S = S µ2 image S
2

+

µ2 f S

it 2

,

(1)

where µimage quantifies the uncertainty in flux density due to the image noise and µ f it represents uncertainties due to the fitting of a Gaussian. These two quantities can be calculated following Windhorst et al. (1984), µimage = S S
2

2 peak

+ C2 + C2 , p f

(2)

where is the local rms noise, C f the error in absolute flux calibration and C p the error in flux due to eventual pointing errors. We here adopt a conservative value of C 2 + C 2 = 0.052 . p f The fitting error is calculated following Condon (1997), µ f it = S µ2 S S
2 peak

Here, B b are the restoring beam's major and minor axes, whereas M m represent the major and minor axes of the fitted Gaussian. The errors µS ,µ M and µm stand for the errors of the fitted Gaussian's peak flux, major and minor axes. The error of the spectral index was finally calculated using simple Gaussian error propagation, = 2 S 1.4 GHz S 1.4 GHz
2

µ2 B b µ2 M 2+ m + M m M 2 m

.

(3)

+

S 2.3 GHz , S 2.3 GHz

2

(4)

where the factor of 2 |[log(1.4 GHz/2.3 GHz) · ln(10)]-1 | arises from deriving the spectral index as defined in section 1 with respect to the flux densities, which are the only quantities with uncertainties in this equation (the error in frequency is neglected). Histograms illustrating the distribution of flux densities in the final catalogues are shown in Fig. 2, the relative flux density errors as a function of flux density are presented in Fig. 3. Further spectral index properties are investigated in the next Section.

5


Peter-Christian Zinn et al.: ATLAS 2.3 GHz observations

Fig. 2. Distribution of 1.4 GHz (hi-res) and 2.3 GHz flux densities in the final catalogue. AGN classifications of the sources are from Mao et al. (2012).

4. Spectral index properties
The distribution of spectral indices for our sample of 631 1.4 GHz sources with unambiguously identified 2.3 GHz counterpart is shown in Fig. 4. With a median spectral index of -0.74 our sample matches the canonical value for synchrotron radiation of -0.7 very well. However, we note that the distribution is broad with a median absolute deviation of 0.35. We here use the median absolute deviation defined as MAD mediani (| xi - medianj ( xj )|) as a measure for the spread of a distribution since it is more robust against outliers than, for instance, the normal standard deviation, . In case of a perfect Gaussian distribution, and MAD are related via 1.4826 MAD. This also compares nicely to other studies such as Prandoni et al. (2006) whose sample has MAD = 0.33. The spread of the Ibar et al. (2009) sample is slightly smaller with MAD = 0.27. This difference is due to the larger number of sources with flat or inverted spectra in our sample. We repeated the analysis including upper limits (following the survival analysis techniques developed in Feigelson & Nelson 1985) for all 1.4 GHz sources which do not have a 2.3 GHz counterpart and are not affected by resolution effects (see Sect. 3). These sources do have a 2.3 GHz flux density less than three times the rms noise at the respective position. The median spectral index then becomes slightly steeper with -0.81 and a MAD of 0.38. The main reason for only having this slight change is the large difference between the 1.4 GHz sensitivity (about 30 µJy) and the 2.3 GHz sensitivity (70-80 µJy). The number of sources detected at both 1.4 GHz and 2.3 GHz is 631 whereas there are 552 sources with an unambiguous upper limit at 2.3 GHz. We note that the errors of our spectral indices become quite large for low flux densities. This is because currently available radio data and the flux densities extracted from them have typical uncertainties at the 10% level. This is mostly due to either inhomogeneous data processing and calibration strategies as well as the widely adopted method of fitting 2-dimensional Gaussians to extract flux densities (see e.g. Condon 1997). Future large-scale continuum surveys, such as ASKAP/EMU, are attempting to reach flux density errors of only 1% by developing both novel, fully automated calibration as well as source extraction techniques (Norris et al. 2011b).
6

Fig. 3. The fractional flux density errors for all ELAIS-S1 and CDF-S sources as function of their flux density. The solid red line indictates how the fractional error should decrease with increasing flux density in an image with an overall rms noise of 70 µJy according to Eqn. 2. A calibration error of 5% is assumed.

4.1. Spectral index vs. flux density

A plot of the spectral index of our sources against their flux density is shown in Fig. 5. We point out that the sensitivities of our 1.4 GHz and 2.3 GHz data differ by a factor of about 3. Therefore, progressively negative spectral indices cannot be measured anymore towards fainter 1.4 GHz fluxes (as indicated by the black dashed line in Fig. 5). For instance, a source with S 1.4 GHz = 200 µJy can only have a 2.3 GHz counterpart when exhibiting a spectral index > 0.5 since the 3 detection level of our 2.3 GHz data is about 250 µJy (see Sect. 2). To account for this fact, we also included upper limits on spectral index for all 1.4 GHz sources that unambiguously show no 2.3 GHz counterpart (the upper limits were calculated using S 2.3 GHz < 3 where is the local rms noise for a point source) and are not affected by resolution issues (see Sect. 3, e.g. a close double in 1.4 GHz that would be inseparable at the 2.3 GHz resolution). With these upper limits, we calculated median spectral indices for differ-


Peter-Christian Zinn et al.: ATLAS 2.3 GHz observations

Fig. 4. Distribution of 1.4 GHz to 2.3 GHz spectral indices in the final catalogue for all sources with unambiguously detected 2.3 GHz counterpart and both the 1.4 GHz and 2.3 GHz flux densities exceeding 0.3 mJy. AGN classifications of the sources are from Mao et al. (2012).

ent flux density bins to investigate a potential flattening of the mean spectral index towards fainter flux densities. As expected, the median spectral indices for only the detected sample (so the 631 sources with 2.3 GHz counterparts) show a clear trend of flatter spectral indices with lower flux density (Fig. 5, blue circles). Including the upper limits into the median calculation, this trend vanishes immediately (Fig. 5, blue diamonds): the lowest flux density bin (about 300-600 µJy) is completely dominated by the upper limits and can hence not be included in further analysis whereas the two flux density bins at 700 µJy and 1 mJy are now a even a little steeper than the ones at higher flux densities. Hence, we find no evidence for a flattening of the spectral index towards the lowest flux densities. This is consistent with the findings by Randall et al. (2012) who used three frequencies (840 MHz, 1.4 GHz and 2.3 GHz) for their spectral index calculation and extend their findings to fainter flux limits. Ibar et al. (2009) come to the same conclusion, too. A flattening of the spectral index with lower flux density levels, as e.g. seen by Zhang et al. (2003), was initially suspected by Windhorst et al. (1993). A major line of argument is that, at sub-mJy levels, a change in population takes place: At mJy levels, the radio sky is completely dominated by AGN (see e.g. Padovani et al. 2011), so the dominant emission mechanism is non-thermal synchrotron radiation. An intrinsically flatter spectrum would e.g. be produced by dominating emission due to thermal bremsstrahlung, but this makes up only a marginal fraction even in starburst galaxies (see Condon 1992). In addition, recent work by several authors indicate that the transition from AGN-dominated to SF-dominated radio sources takes place below a flux density of 100 µJy. For instance, Padovani et al. (2011) argue that AGN still dominate the source population down to ´ 100 µJy and Smolcic et al. (2008) claim that the fraction of SFpowered radio sources has a nearly constant value of about 40% for flux levels between 50 µJy and 700 µJy. A similar result is obtained by Seymour et al. (2008) investigating the differential number counts in the 13 Hour XMM-Newton/Chandra Deep Field Survey. Hence, this change in the faint radio population cannot account for our findings since we do not reach such faint flux density levels.

Fig. 5. Spectral index vs. flux density. The gray 2D histogram shows our sample of 631 sources with 1.4 GHz and 2.3 GHz detections. Since the sensitivity of the 2.3 GHz observations is 3 times lower than that of the 1.4 GHz data, the black dashed line indicates the minimum spectral index that could be measured with our data. Therefore, the median spectral indices (blue circles, x-error bars indicate the flux density bin size, y-error bars correspond to the MAD) for just the 2.3 GHz detected sample show significantly flatter values towards lower flux densities. When including also upper limits from the 2.3 GHz non-detected sources, this trend completely vanishes (blue diamonds). Hence we conclude that there is no evidence for a flattening of the mean spectral index with lower flux density.

With our catalogue of 631 sources with spectral indices, we now focus on the application of the spectral index in terms of its predictive quality. Since it is a quantity which is relatively easy to measure, the spectral index would be a useful tool to characterise sources in large-scale surveys, if a dependence on other properties of interest such as source type, redshift or other scaling relations can be established.
4.2. Spectral indices as object type discriminator

The radio spectral index is commonly regarded as a possible tool for discriminating different radio source ty