Filename
stringlengths 22
64
| Paragraph
stringlengths 8
5.57k
|
|---|---|
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | Finally, the HT source D is diffuse and has a steep spectrum, consistent with D being an old and slowly mov- ing source in a lower density environment. At 0.327 GHz (Schindler 1996) the ridge line of the tail appears slightly curved in the sense expected for a galaxy in orbit around A3528S. |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | Are the tailed sources in A3528 linked in some way with the merging event? We find no direct evidence for the effects of bulk flows or shocks in the radio images, consistent with a pre-merging state for A3528. However, we have evidence of individual galaxies whose trajectories may have been dis- turbed by tidal interactions between A3528N and A3528S. In particular we have A2 which is physically close to the cen- tre of A3528N and on a plunging orbit. Radial orbits are not expected near the centres of virialized clusters as numerical simulations (West, Dekel, & Oemler 1987) and theoretical arguments (Lynden-Bell 1967) indicate that orbits should circularise very quickly. We may therefore be witnessing the early passes of d2 through the centre of A3528N. |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | In the pre-merging scenario it is possible that d2 was originally near the periphery of either A3528N or A3528S and was therefore susceptible to being perturbed into a plunging orbit towards the centre of A3528N. A reason- able estimate of the time since d2 was near the periphery of A3528N is the sound crossing time of galaxies in clusters (∼ 108 yrs) which is comparable with the merger timescale of the clusters (Schindler 1996). |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | If galaxy interactions are important in radio source trigger- ing, we might expect to see an excess of radio sources in galaxy rich environments. An excess is indeed seen (Burns & Owen 1977; Burns et al. 1981; Robertson & Roach 1990; Brown & Burns 1991) but this can be explained through higher galaxy densities and the nature of the bivariate ra- dio luminosity function (McHardy 1979; Venturi, Feretti & Giovannini 1989; Kim et al. 1994; Ledlow & Owen 1995). |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | There is evidence, however, that triggering of radio ac- tivity does involve environmental factors (Lilly & Prestage 1987; Hanisch 1984; Gavazzi & Jaffe 1986; Venturi, Giovan- nini & Feretti 1990). In particular Byrd and Valtonen (1990) have shown that gravitational interactions with other galax- ies and/or the cluster potential cause nuclear inflows which can trigger or fuel an AGN. d1 and d2 may be an example of this as the tail of A2 is stubby in Fig.2, in contrast to B and D, and this leads us to speculate that the radio activity in A2 started abruptly as the galaxy d2 approached the centre of the cluster. |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | We also note an intriguing connection between A1 and the enigmatic class of Compact Steep Spectrum (CSS) sources (Fanti et al. 1990). A1 bears all the hallmarks of a CSS source: compact double structure, a linear size of ∼ 30 kpc and a spectral index α = −0.7. Evidence is mounting that these characteristics are the result of youth, rather than confinement by a dense ISM, and our tentative interpreta- tion of the A1/A2 complex is consistent with this picture. If we assume that A2 has a transverse velocity of ∼1000 km s−1 and that the radio emission was triggered in A1 and A2 at about the same time, then the 55 kpc tail length for A2 implies that radio activity started about 5 × 105 years ago. This is remarkably consistent with typical lifetimes of 106 yrs estimated for CSS sources (Fanti et al. 1995) and is very short compared with the typical lifetimes of the central engines in large radio sources of >∼ 2 × 107 years. Higher res- olution radio images are needed to test this interpretation. |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | Strong correlations are seen between the radio and X-ray bands for radio-loud quasars and radio galaxies (Brinkmann, Siebert & Boller 1994; Brinkman et al. 1995; Baker, Hun- stead & Brinkmann 1995) and there is mounting evidence for contamination of diffuse thermal X-ray emission from the ICM by non-thermal emission from AGN. However, while Edge & R¨ottgering (1995) found unresolved ROSAT PSPC sources associated with HT sources with power-law spectra similar to non-thermal emission from BL Lacs and Seyferts, in general the AGN contribution was a small component of the total X-ray flux from the surrounding clusters. Our X- ray analysis of A3528 has revealed a similarly small pointlike X-ray source which is probably coincident with C. In addi- tion, the X-ray emission near the core of A3528N (highest contour) may be elongated in the same direction as the axis of the radio source A1. If this is confirmed with more sen- sitive, high resolution X-ray observations (eg. XMM) then it may signify interaction between the radio jet and the ISM/ICM. We think it unlikely due to inverse Compton scat- tering of CMB photons into the X-ray. This is because the predicted X-ray flux in the ROSAT PSPC band for this ra- dio source, using the formulae in Harris & Grindlay (1979) and assuming the minimum energy magnetic field, is of the −2, which is three orders of magni- order of 10 tude less than the central surface brightness determined by Schindler (1996). |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | that bulk motions and/or shocks in the ICM affect the mor- phology of any of the radio sources, and this is consistent with A3528 being a pre-merging cluster. |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | The higher sensitivity and improved spatial and spec- tral resolution of XMM would give a clearer indication of the state of the ICM in A3528 and determine whether the unre- solved X-ray component near the centre of A3528S is likely due to a cooling flow or an AGN. Radio observations with higher sensitivity and spatial resolution might also reveal core emission and jet structure in B and elucidate the na- ture of source D. Nevertheless, in the context of our broader study, the present multi-wavelength observations have been very effective in evaluating the links between the radio and X-ray characteristics and the dynamical state of A3528. |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | We thank Taisheng Ye for assistance in the reduction of the ATCA data and John Reynolds for resolving an astrometric problem; Nick Kaiser for advice concerning the production of the galaxy density plots using COSMOS data; and Andrew Hopkins for supplying a source correlation program. The authors would especially like to thank the referee, Dr. Jack O. Burns, for his recommendations which lead to significant improvements in this paper. |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | This research has made use of the Australia Telescope Compact Array which is funded by the Commonwealth of Australia for operation as a National Facility managed by the CSIRO; the NASA/IPAC Extragalactic Database; the Astrophysical Data System; the COSMOS/UKST Southern Sky Catalogue supplied by the Anglo-Australian observa- tory; and the Digitized Sky Survey and 2nd Epoch Digitized Sky Survey. |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | ADR acknowledges support from an Australian Post- graduate Award, and from the ASA, the IAU, the Australian Nuclear Science and Technology Organisation, the Franco- Australian Science Agreement and the University of Sydney. RWH acknowledges support from the ARC. |
Processed_Radio_Study_of_X-ray_Clusters_I:_A3528_---_a_pre-m.txt | Evrard A. E., 1990a, in W. R. Oergerle, M. J. Fichett & L. Danly, eds, Clusters of Galaxies. Cambridge University Press, p. 257. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | ABSTRACT We use high angular resolution data, measured from visibility of sources at the longest base- line of 4500 m of the Australia Telescope Compact Array (ATCA), for the Australia Telescope 20 GHz (AT20G) survey to obtain angular size information for > 94% of AT20G sources. We confirm the previous AT20G result that due to the high survey frequency of 20 GHz, the source population is strongly dominated by compact sources (79%). At 0.15 arcseconds an- gular resolution limit, we show a very strong correlation between the compact and extended sources with flat and steep-spectrum sources respectively. Thus, we provide a firm physical basis for the traditional spectral classification into flat and steep-spectrum sources to separate compact and extended sources. We find the cut-off of -0.46 to be optimum for spectral indices between 1 and 5 GHz and, hence, recommend the continued use of -0.5 for future studies. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | We study the effect of spectral curvature on redshift cut-off of compact AGNs using recently published redshift data. Using spectral indices at different frequencies, we correct for the redshift effect and produce restframe frequency spectra for compact sources for redshift up to ∼ 5. We show that the flat spectra of most compact sources start to steepen at ∼ 30 GHz. At higher frequencies, the spectra of both populations are steep so the use of spectral index does not separate the compact and extended source populations as well as in lower frequencies. We find that due to the spectral steepening, surveys of compact sources at higher frequencies (> 5 GHz) will have redshift cut-off due to spectral curvature but at lower frequencies, the surveys are not significantly affected by spectral curvature, thus, the evidence for a strong redshift cut-off in AGNs found in lower frequency surveys is a real cut-off and not a result of K-correction. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | whether such an evolution is purely due to space density evolution or due to luminosity evolution is yet to be resolved. Similarly, the decline in the number density in steep-spectrum radio sources has also been discussed in the literature (eg. Dunlop & Peacock (1990); Jarvis et al. (2001)) and has been studied in detail by Rigby et al. (2011). |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | decline while Shaver et al. (1996) and Wall et al. (2005) find a rapid decline in the number density of flat-spectrum radio quasars at high redshift, comparable to results from X-ray and optical work. How- ever, Jarvis & Rawlings (2000) have argued that spectral curvature causes the observable volume at high redshift universe to be too small to find any luminous flat-spectrum sources, removing the ev- idence for the rapid decline in number density at high redshift sug- gested by Shaver et al. (1996). The spectral curvature of sources at different redshifts gives the impression of redshift cut-off by caus- ing the sources to fall outside of selection criteria either by making the flux density fall under the selection limit or the spectral index to fall outside the selection limit, in populations selected based on their spectra (eg. flat-spectrum population). Wall et al. (2005) have found that the use of non-contemporaneous flux density data can cause the effect of steep spectral curvature similar to that seen by Jarvis & Rawlings (2000) (see also de Zotti et al. (2010)). |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | The abundance of quasars at redshifts 2.5 - 3 has implications for the epoch of formation of the first structures in the universe, which has important consequences for the formation and evolution of massive galaxies. This epoch also has importance in determin- ing whether AGNs can be significant sources of reionization of the universe. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Radio frequency studies of the redshift cut-off have the ad- vantage of being free from dust obscuration. Such studies have been done in the past with populations of flat-spectrum sources, selected from low frequency (6 5 GHz) source catalogues. Flat- spectrum sources have traditionally been associated with com- pact sources (Wall 1977) with the spectral transition placed vari- ably around −0.4 (eg. Shaver 1995; Wall et al. 2005), −0.5 (eg. Peacock 1985; Saikia et al. 2001; Browne et al. 2003) or −0.6 (eg. De Breuck et al. 2000). In this paper, we follow the convention of Sν ∝ ν α and use -0.5 as the flat-spectrum cut-off. The flat- spectrum arises from the cumulative effect of multiple, very com- pact, self absorbed components peaking at different frequencies. Thus, compact AGNs are considered to have a very flat-spectrum, often spanning decades of frequencies. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | survey (Murphy et al. 2010) is a complete sample of radio sources and is dominated by compact QSOs, due to its high frequency selection (Sadler et al. 2006). These sources have spectra measured at one to three different frequencies (20 GHz, 8.6 GHz and 4.8 GHz) and have counterparts at one to two low frequencies (∼ 1 GHz) . The three high frequencies were measured near simulta- neously as part of the survey to remove any effects of variability. Sadler et al. (2006) find the median variability index at 20 GHz to be 6.9 per cent with 5 per cent of sources varying by > 30 per cent in flux density on a 1-2 year time-scale. The AT20G sources also have an angular size measurement at sub arcsec resolution which cleanly separates the compact AGNs from their jets and extended radio-lobes. In this paper, we demonstrate the strong correlation between compact sources and flat-spectrum sources providing a firm physical basis for the use of spectral index for such a separation. We then look at the radio spectrum of these compact sources to investigate the effect of redshift cut-off. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Table 1. Number of AT20G sources with follow-ups, cross-matches with low frequency catalogues and optical counterparts. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | AT 20G1 is the subpopulation of AT20G with 8.6 and 4.8 GHz follow-up. Sources with δ > −15◦ do not have 8.6 and 4.8 GHz follow-up. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | ogy with H0 = 71 kms−1Mpc−1, Ωm = 0.27 and ΩΛ = 0.73 (Larson et al. 2011) has been used for this paper. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | The AT20G is a blind extragalactic survey carried out at a high fre- quency of 20 GHz for all declinations of the southern sky using the Australia Telescope Compact Array (ATCA). It covers a total area of 20, 086 deg2. It has a galactic latitude cut-off |b| = 1.5 degrees. The survey has a total of 5890 sources above the flux density limit of 40 mJy at 20 GHz. The AT20G is the largest blind survey done at such a high radio frequency. Most sources south of the decli- nation of -15 degrees have near-simultaneous follow-up observa- tions within a month at 4.8 and 8.6 GHz (Murphy et al. 2010). The sources not observed at 4.8 and 8.6 GHz were a result of schedul- ing logistics and can not introduce any spectral bias. The cross- match of AT20G sources against the 1.4 GHz NVSS (Condon et al. 1998) catalogue for sources north of δ = −40 degrees and 0.843 GHz SUMSS (Mauch et al. 2003) catalogue for sources south of δ = −30 degrees provides between one to two spectral points near 1 GHz resulting in flux densities at two to five frequencies for the sample. Although the lower frequency points are not simultaneous, there is relatively little variability at these frequencies. Despite the galactic cut-off, the AT20G catalogue still contains a small num- ber of galactic thermal sources which are extended sources of op- tically thin (flat-spectrum) free-free emission. These high latitude sources are easily seen in Hα surveys and have been identified and removed from the main population for this analysis of the extra- galactic sources. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Interferometric visibilities on long spacings (henceforth: 6km visibility) for > 94% of the AT20G sources have been measured (see section 3). 3403 AT20G sources have four or more spectral points between 1 and 20 GHz as well as 6km visibility. Optical identifications for 3873 AT20G sources have been made and red- shifts of 1460 sources have been compiled (Mahony et al. 2011). This work makes use of the subpopulation of 838 sources with mea- sured redshift, 6km visibility and flux densities measured at four to five frequencies. Table 1 provides a summary of the different sub- populations. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Figure 1. The distribution of all 6km visibilities in the AT20G. Unresolved sources have a visibility of 1 marked by the solid vertical line. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | The AT20G follow-up was done with the hybrid configurations of the ATCA, with five antennae in a compact configuration and the sixth antenna (called the 6km antenna in this paper) at a baseline of ∼ 4500m from the location of the compact configuration. This al- lowed us to calculate a visibility amplitude for over 94% of AT20G sources (Table 1). The remaining sources do not have their 6km visibility due to inclement weather during observation or unavail- ability of the 6km antenna due to maintenance. Thus, the sources with the visibility data are free from any bias. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | We calculate the 6km visibility from the ratio of the scalar average of the five long baseline amplitudes to the scalar average of the 10 short baseline amplitudes for the AT20G sources at 20 GHz (Chhetri et al. in prep). Although, the AT20G follow-up ob- servations were done in a hybrid mode, the ATCA when used with its 6km antenna is essentially a linear array. So, the 6km visibil- ity corresponds to a position angle on the sky which is dependent upon the hour angle of observation. Most AT20G sources have vis- ibilities measured at two or more widely separated hour angles. Fig. 1 shows the distribution of all AT20G 6km visibilities for all sources. The 6km visibility for a point sources is expected to be 1. The Fourier transform provides a direct relationship between visi- bility and angular size of a source. Such a source size relationship is model dependent (eg. Gaussian, double peaked etc.) but, none the less, provides an effective method of separating compact sources from extended sources based on their size. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Since it is not possible for any source to have visibility > 1, we can use the distribution of visibilities > 1 as an empirical es- timate of the effect of noise on the visibility estimate. Visibilities found to be > 1 in Figs. 1 and 3 are due to the effects of random noise on 6km amplitudes of unresolved sources with a RMS scatter of 0.053 to a maximum of 1.14. Considering a symmetric noise re- lated distribution from the unresolved value of 1, we arrive at 0.86, where we separate the compact sources from the extended sources and only 0.5% of compact sources would be misclassified as ex- tended. Therefore, for the purpose of this paper, we define sources with 6km visibility > 0.86 to be compact sources. For a Gaussian source, a visibility of 0.86 at 4500 m corresponds to a size of 0.15 arcsecond at 20 GHz. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Figure 2. Redshift distribution of 6km visibility for flat (filled circles, 1 > −0.5) and steep-spectrum (open squares, α4.8 α4.8 1 6 −0.5) sources. Visibility models are drawn for 0.5, 1, 2 and 4 kpc diameter Gaussian sources. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | AGNs which is a result of the high survey frequency. This is very different from low frequency surveys such as NVSS which has only 25% of the flat-spectrum AGN source population (De Breuck et al. 2000). |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | In Fig. 2 we have plotted the visibility on the 6km antenna baselines against redshift. Models for sources with an assumed Gaussian distribution of brightness are shown for a range of lin- ear sizes (full width at half maximum). If the source brightness distribution is double peaked rather than Gaussian the curves have exactly the same form but the sizes are scaled down by a factor of 1.5. In reality there is variation in source morphologies and a single visibility cannot distinguish between these structures so the lines in Fig. 2 correspond to a small range of possible sizes. However, this uncertainty resulting from unknown source morphology is a rela- tively small fraction of the known range in radio source sizes from cores in the central few hundred parsec to megaparsec scale lobes. One interesting variation in the interpretation of the visibil- ity occurs for sources with emission coming from multiple regions with very different linear sizes. For example sources which have a significant flux in an unresolved core (<0.15 arcsec in our case) and the rest of the flux in a resolved jet or lobe structure. In this case the visibility should be interpreted as the fraction of the flux in the unresolved component, and the curves in Fig. 2 indicate the scale size for which the core component is unresolved. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Figure 3. Spectral index distribution plotted against 6km visibility. The dot- ted lines mark the limits of flat v steep-spectrum on the x-axis and compact v extended sources on the y-axis. Solid line at y = 1 represents the expected values of visibility for unresolved sources. The flux density for 1 GHz is obtained either from SUMSS or NVSS catalogue while 5 GHz is obtained from AT20G follow-up at 4.8 GHz. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | separation between compact and extended sources at the 6km visi- bility of 0.86, there is a very clean separation between compact flat- spectrum and extended steep-spectrum sources at the α ∼ -0.5 line. The remarkably clean separation of compact flat-spectrum sources and extended steep-spectrum sources gives a very firm physical ba- sis for the traditionally used spectral index as a separator of com- pact and extended sources. Only 111 out of 2347 (5%) flat spectrum sources are extended while 406 out of 3160 (12%) compact sources are steep-spectrum sources. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | We can use the spectral index distributions for compact and extended sources plotted in Fig. 4 to find the optimum value of the spectral cut-off to classify the flat and steep-spectrum AGN pop- ulations. We get slightly different answers depending on what we want to optimise. The maximum value of the joint probability that both flat and steep-spectrum sources are correctly classified is at α = -0.46 and 80% of the sources are correctly classified (see Fig. 4). This is independent of the relative number of flat and steep- spectrum sources in the survey, and consequently is only weakly dependent on the survey frequency. This estimate is right between the two values (-0.4 and -0.5) mostly used historically, for con- sistency we recommend the use of a -0.5 cut-off in future studies. However, if we want to maximise the total number of correctly clas- sified compact sources this does depend on the relative amplitude of the two populations and is dependent on survey frequency vary- ing between α = -0.4 for a survey frequency of 5 GHz to α = -0.6 for a survey frequency of 20 GHz. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Table 2. Gaussian fits to the spectral index distribution of compact and ex- tended sources in Fig. 4. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Figure 4. Distribution of Compact and Extended sources over spectral in- dex. The spectral indices are calculated between SUMSS/NVSS 1 GHz and AT20G 4.8 GHz flux densities. Mean spectral indices and dispersions from Gaussian fits to extended and compact source population distributions are given in Table 2. The solid curve is the joint probability of correctness (see text). The histogram for extended sources are shifted by -0.05 to show both populations at the steep-spectrum end. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | characteristics between the three frequencies. Spectral index be- tween two frequencies with larger separation is less susceptible to errors due to flux measurement errors than spectral index between two frequencies close to each other. For a 10% error in flux density in Fig. 5, the α8.6 8.6 has an error of 0.11 and α5 1 has an error of 0.06. The 5 and 8 GHz flux density mea- surements were done within a month of the 20 GHz measurements. Hence, they do not suffer much from variability of compact sources (Massardi et al. 2011) . Although the 1 GHz flux densities were measured at different epochs (between minimum of one to max- imum of 15 years), variability is much less at lower frequencies. Hence, errors in flux densities or due to variability do not explain most of the differences seen in the three plots. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Figure 5. 6km visibility is plotted against three spectral indices increasing in frequencies with α4.8 at the bottom panel, α8.6 4.8 at the mid panel and α20 8.6 at the top panel. The solid vertical line corresponds to spectral index = -0.5 and the solid horizontal lines correspond to 6km visibility = 1. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | either become flat or steep at higher frequencies (e.g. GPS sources) (Sadler et al. 2008). |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | On the other hand the spectra of extended sources which are well separated by spectral index at low frequencies, have a larger scatter at higher frequencies. A significant fraction of the extended sources move towards the flatter part of the plot at higher frequency. These extended sources are mostly composed of a flat-spectrum core with a steep-spectrum jet, lobe or a hotspot. With increasing frequency, the steep-spectrum part of the sources gets weaker while the compact flat-spectrum cores start to dominate the spectrum. These composite sources show an overall combined flat spectrum but are still extended. This domination of flat-spectrum core should be easily observable by imaging these composite sources at higher frequencies, when only the compact cores should be visible. With increasing frequency, there is a further steepening in the spectra of the extended structures due to energy loss. This is evident in the left part of Fig. 5 with the appearance of a population of very steep extended sources in the α8.6 and α20 8.6 plots. These two effects com- bined produce the larger scatter seen in the α20 8.6 plot of Fig. 5 for extended sources. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | It should also be noted that at higher frequencies the spectra of compact sources starts to become indistinguishable from steep- spectrum sources (Fig. 5, α20 8.6 plot). It is then clear that while the use of spectral index to separate compact and extended sources is valid using lower frequency spectra, it is unable to provide a clean separation of compact and extended sources at higher frequencies. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Figure 6. Spectra of compact and extended sources with redshift calcu- lated between 8.6 and 20 GHz. Filled circles represent compact sources and open triangles represent extended sources. The median and standard error for redshift bins, marked by dashed vertical lines, are shown in Table 3. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | that the redshift cut-off is real and is similar to the cut-off observed in optically selected QSOs. Similar results have been reported for cosmic star formation rates. However, Jarvis & Rawlings (2000), using the Parkes half-Jansky flat-spectrum sample selected at 2.7 GHz, argue that because of spectral curvature and the resulting K- correction there is not enough observable volume at high redshift to find the brightest flat-spectrum radio sources, which gives the impression of redshift cut-off. We can use the sub population of compact sources from the AT20G survey, with redshifts, to study the effect of spectral curvature and the K-correction on the redshift cut-off. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | A survey at any frequency is biased to the sources that are stronger at the survey frequency. Since this is very obvious, it is not usually considered a bias. At lower frequencies the steep- spectrum sources dominate the population at all luminosities and only contain the rare extremely luminous flat-spectrum sources. Since it is the flat spectrum population that has a high quasar iden- tification rate, this is the sample used to investigate evolution of flat-spectrum quasars (Dunlop & Peacock 1990; Shaver et al. 1996; Jarvis & Rawlings 2000; Vigotti et al. 2003; Wall et al. 2005). Our frequency selection provides a much larger fraction of these sources without introducing a different class of sources, thus, we reduce the bias by having a high frequency selected sample (e.g. Condon 2009). |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | As noted in Section 2, 838 out of 1460 of AT20G sources with redshift have 4.8 and 8.6 GHz flux densities, 1 GHz NVSS/SUMSS counterparts and 20 GHz flux densities as well as 6km visibility. We use this subsample of sources for the following analysis. Of these, 683 sources are compact by our definition while the remain- ing 155 sources are extended. The distribution of the redshifts for this AT20G subpopulation is listed in Table 3. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Fig. 6 plots the spectral indices for compact and extended sources between 8.6 and 20 GHz against redshift. It is evident from the plot that extended sources do not show apparent changes in the spectral index over the redshift range. Compact sources, however, show a distinct steepening of their spectra with redshift. It is appar- ent that a spectral classification is no longer separating the compact and extended sources above redshifts ∼ 1. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Figure 7. Rest frame spectral index of compact and extended sources. Filled circles represent compact sources and open triangles represent extended sources. All points have been brought close to the rest frame spectral in- dex between 8.6 and 20 GHz by using the spectral index between 8.6 and 20 GHz for all sources below z = 1, using spectral indices between 4.8 and 8.6 GHz for all sources between redshifts 1 and 2 and using spectral index between 1 and 4.8 GHz for all sources with z > 2. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | spectra over a large range of radio frequencies, we do not expect a significant change in spectrum with redshift due to the change in restframe frequency. But as discussed in section 5, compact sources have flat spectra at lower frequencies that gets steeper at higher frequencies. Hence, the spectral indices of compact sources are affected by the redshift effect causing the redshift cut-off as dis- cussed by Jarvis & Rawlings (2000). To test this idea further we use the spectra at different observed frequencies for different red- shift ranges. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | The median spectral indices and errors for these redshift ranges are listed in Table 3. Our sources have flux densities mea- sured between four to five frequencies which allows us to calcu- late, at least, three contiguous spectral indices (α4.8 8.6). We take advantage of these spectral points by using the best rep- resentative spectra for different redshift bins to remove the effect of redshift and, thus, compare their spectral indices between corre- sponding rest frequencies. For this, we approximate that the spec- tral index of α4.8 best represents the redshift bin between 2 and 5 (rest frequencies ∼8 and ∼14 GHz), α8.6 4.8 best represents the red- shift bin between 1 and 2 (rest frequencies ∼10 and ∼26 GHz) and α20 8.6 represents the redshift bin for redshifts up to 1. Plotted in Fig. 7, this essentially brings all spectral indices to α20 8.6 for compari- son in their restframes. We use this simple procedure to preserve the measured values rather than fitting complex spectral shapes to a small number of frequencies. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | and plotted them against the rest frequency in Fig. 8. The mean red- shift for these 671 compact sources is ∼1. The figure confirms the results seen in fig. 6 of Wall et al. (2005) which shows that the spec- tra of flat-spectrum sources at their rest frames at lower frequencies is predominantly flat suggesting that spectral curvature does not play a significant role at lower frequencies. At lower frequencies up to ∼ 5 GHz, we observe that there is a mixture of steep, flat and inverted spectrum sources. It is also clear from the figure that the majority of compact sources remain flat-spectrum up to ∼30 GHz in their rest frame frequencies at all redshifts but the trend is then for almost all sources to get steeper. The median normalized flux density for the rest frequency bin between 35 - 45 GHz drops by ∼15% compared to that for the rest frequency bin between 15 - 25 GHz. This explains the effect of spectral steepening at high frequency of compact, and hence, flat-spectrum sources at higher redshifts seen in Fig. 6 and Table 3, as well as the removal of the effect after redshift correction in Fig. 7. The plot also indicates that at lower frequencies, the spectral curvature does not play a signif- icant role. In fact, there is a population of steep-spectrum core-jet sources at the lower frequency end of Fig. 8 which produces the opposite effect. As discussed in Section 5, these compact sources that have steep-spectrum jets at low frequencies are increasingly dominated by their flat-spectrum cores at higher frequencies. The redshift effect of these sources, then, makes them appear flat spec- trum at observed lower frequencies at higher redshifts. For sources at high redshifts, this effect negates the effect of spectral steepening of the overall spectra, at lower frequencies. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Thus, the redshift cut-off is affected by the spectral curvature of compact sources as suggested by Jarvis et al. (2001), but this effect only becomes important at higher frequencies (ν >5 GHz) and can be avoided by measuring the spectra at lower frequencies. Fortunately, this is where majority of evolution studies have been done (e.g. Peacock (1985); Dunlop & Peacock (1990); Shaver et al. (1999); Jarvis & Rawlings (2000); Wall et al. (2005)). |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | So far, we have only discussed the effect of the spectral curva- ture on the redshift cut-off caused by the use of the spectral index to filter out the compact flat-spectrum sources. However, it is also necessary to see how this effects the number of sources above the flux limit of the survey, referred to by Jarvis & Rawlings (2000) as the observable volume. This cut-off is determined by the survey frequency regardless of the frequency used to measure the spec- trum. Low frequency surveys will select against the compact flat- spectrum sources at all redshifts. Higher frequency surveys will have much larger populations of compact sources but they will also have a much stronger redshift cut-off due to spectral curvature. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | To evaluate this survey sensitivity effect, we plot the average parameters used in the Jarvis & Rawlings (2000) models, normal- ized at 2.7 GHz, with our rest frame spectra in Fig. 8. The model is i=0aixi, where given by a polynomial of the form y = log10Sν = Σ2 x = log10(ν/GHz) with a1 = 0.07 ± 0.08 and a2 = -0.29 ± 0.06. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Table 3. Median spectral indices and standard error for compact and extended sources before (Fig. 6) and after (Fig. 7) redshift corrections. There are only two extended sources in the z > 2 bin. The median and standard error are kept inside parentheses to note their limited significance. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | We compared the Parkes quarter-Janksy flat-spectrum sam- ple (Jackson et al. 2002) with the AT20G data in the overlapping area of the sky. The PQJFS sample has 804 southern hemisphere sources and actually includes sources with flux densities less than 250 mJy. We have included all sources down to 220 mJy since that corresponds to 100 mJy at 20 GHz for the spectral index cut-off of -0.4. There are 716 PQJFSS sources over the limit of 220 mJy at 2.7 GHz. The cross match of these sources’ positions against the AT20G catalogue (Murphy et al. 2010) finds 640 of these sources and further 31 are found in the AT20G scanning sur- vey catalogue (Hancock et al. 2011) which avoids the incomplete- ness introduced by the AT20G follow-up observations. We used the updated positions of the sources provided in Jackson et al. (2002), where available, since the original PQJFSS positions are not fully reliable. Thus, our cross-match found a total of 671 (94 %) PQJFSS sources in the AT20G survey. Of the remaining 45 sources, 14 were later found to be extended sources by Jackson et al. (2002) and are not included in their analysis. For the remainder, our lit- erature search finds that 7 are detected at lower frequencies and have steep-spectrum, 9 are known variable sources based on mul- tiple epoch observations and 1 source that has not been detected in any other surveys. The remaining 14 sources (2%) are well within the completeness expected from long term variability noting that the AT20G survey was done > 20 years after the PQJFS survey. Only 3 (0.4%) of these sources have evidence for extreme spec- tral steepening at high frequency. Hence, we can conclude that we have not lost a significant fraction of sources as a result of spectral steepening which has put them below our survey limit. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | Figure 8. Rest frame radio spectra of 671 compact AT20G sources. Sources have flux density observations at a subset of 0.8, 1.4 and 4.8, 8.6 and 20 GHz. All observed frequencies are converted to the rest frame and the flux densities are normalized by their average flux density over all frequencies present. The solid black line is the fit to the spectra of the most lumi- nous Parkes Half-Jansky Flat-Spectrum sample done by Jarvis & Rawlings (2000) normalized at 2.7 GHz survey frequency. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | quencies is not as effective to select compact and extended sources. Using the radio spectrum below a few gigahertz avoids this prob- lem. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | In future, with the information on the completeness of the optical identification and redshifts, it will be possible to explore the density evolution and redshift cut-off of compact AGN without using the spectral index to select compact sources. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | We thank the staff at the ATCA site, Narrabri (NSW), for the valu- able support they provided in running the telescope. The ATCA is part of the Australia Telescope which is funded by the Common- wealth of Australia for operation as a National Facility managed by the CSIRO. |
Processed_Spectral_Properties_and_the_Effect_on_Redshift_Cut.txt | R.C. would like to thank Tara Murphy for providing cross- matching scripts. MM acknowledges financial support for this re- search by ASI (ASI/INAF Agreement I/072/09/0 for the Planck LFI activity of Phase E2 and contract I/016/07/0 COFIS). |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Fast Radio Bursts (FRBs) are a relatively new class of astrophysical phenomenon, the origin of which re- mains an open question. Although the Parkes radio telescope has detected the majority of the FRBs (e.g. Lorimer et al. 2007; Keane et al. 2011; Thornton et al. 2013; Burke-Spolaor & Bannister 2014; Ravi et al. 2015; Petroff et al. 2015b; Champion et al. 2016; Keane et al. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | other FRBs have failed to see repetition (Lorimer et al. 2007; Petroff et al. 2015b). |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | The DMs of the FRBs greatly exceed the DM contribu- tion expected from the Milky Way by factors of as much as 30; this has led to the interpretation that the FRBs are extragalactic (Lorimer et al. 2007). In addition, obser- vations of the scattering seen in some FRBs lends credence to the extragalactic interpretation (Macquart & Johnston 2015; Cordes et al. 2016) whereas the Galactic interpreta- tion (e.g. Loeb et al. 2014) seems barely tenable (Dennison 2014). While the extragalactic interpretation of FRBs thus seems reasonably secure, the same cannot be said of their progenitors with many ideas currently in the literature (e.g. Falcke & Rezzolla 2014; Lyubarsky 2014; Cordes & Wasserman 2016; Katz 2016). Indeed, given the observational phenomenology listed above, it remains un- clear whether all FRBs have the same origin. If FRBs are indeed extragalactic in origin and can be detected in large numbers at cosmologically interesting distances, they can then be used as tools for high precision cosmology (McQuinn 2014; Macquart et al. 2015). For example, Keane et al. (2016) used the DM of FRB 150418 in conjunction with the optical redshift of the putative associated galaxy to mea- sure the cosmic density of ionized baryons, and found the re- sult to be in agreement with that of WMAP (Hinshaw et al. 2013). |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | The FRBs in the literature to date have all been dis- covered with single dish telescopes and hence have local- isation of only a few arcmin at best. This means there are many sources at both optical and radio wavelengths within the positional uncertainty of the FRB, making any putative association difficult. Progress in determin- ing the progenitors (or afterglows) of FRBs depends on obtaining an accurate position and although valiant at- tempts have been made to detect FRBs with interferome- ters over a wide range of frequencies (e.g. Wayth et al. 2011; Siemion et al. 2012; Coenen et al. 2014; Karastergiou et al. 2015; Law et al. 2015; Tingay et al. 2015; Rowlinson et al. 2016; Caleb et al. 2016; Burke-Spolaor et al. 2016), arcsec- ond positions remain elusive. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | The SUrvey for Pulsars and Extragalactic Radio Bursts (SUPERB) detected FRB 150418 in real time using the Parkes radio telescope (Keane et al. 2016). Follow-up ob- servations using the Australia Telescope Compact Array (ATCA) and the Subaru optical telescope led to Keane et al. (2016) claiming that a fading radio transient was associ- ated with the galaxy WISE J071634.59−190039.2 at redshift z = 0.49. The low probability of finding such a radio source within the field of view of the Parkes telescope (Bell et al. 2015; Mooley et al. 2016) meant that Keane et al. (2016) were able to argue that the transient was associated with the FRB and hence located in the optical galaxy. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | wide range of frequencies, showed that radio emission from WISE J071634.59-190039.2 had flux densities comparable to the later epochs of Keane et al. (2016). The source of the emission was shown to be compact on EVN baselines (Giroletti et al. 2016). VLBA and e-MERLIN radio obser- vations and optical imaging with Subaru (Bassa et al. 2016) showed that the compact radio source is located in the center of WISE J071634.59-190039.2, and suggest that the source is likely a weak AGN. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | In this paper we describe observations of the radio source performed with the ATCA, the JVLA and the Gi- ant Metre Wave Telescope (GMRT) in Section 2. Section 3 presents the results with special attention given to the abso- lute flux density calibration between the ATCA and JVLA observations. Section 4 is given to interpretation of the radio light curve, before we present our conclusions in Section 5. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Observations of the field containing FRB 150418 were made with the ATCA, the GMRT and the JVLA covering a period of close to one year after the FRB. In addition, JVLA observations were made by Williams & Berger (2016) and Vedantham et al. (2016) following the publication of Keane et al. (2016). A series of observations were also made in with the VLBA and e-MERLIN, described in detail Bassa et al. (2016) and with the EVN (Giroletti et al. 2016). A total of 10 epochs were obtained with the ATCA of which the first five were included in Keane et al. (2016). Their observations plus those of epoch 6 were made by mo- saicing together 42 separate pointings. For epochs 7 to 10 inclusive, a single pointing was used centered on the position of the WISE galaxy. All these observations were carried out in two bands each of 2-GHz bandwidth, the first centered at 5.5 GHz and the second at 7.5 GHz. Flux calibration was carried out with the ATCA calibrator B0823−500 while phase calibration was performed using B0733−174. ATCA data reduction was carried out using the miriad package (Sault et al. 1995) using standard techniques. For the mo- saiced observations, each pointing was reduced separately and the pointings later combined using the task linmos. Each 2-GHz band was reduced independently. Flux mea- surements were made using the task imfit from the images and uvfit directly from the calibrated uv data. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Three epochs were obtained with the GMRT. The first and second epochs used 3C286 as the flux calibrator whereas the third epoch used 3C147. Phase calibration was carried out using 0837−198 for the first epoch and 0735−175 for the last two epochs. GMRT data reduction was carried out using AIPS, the images were then ported into miriad for analysis. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | A single epoch was obtained with the JVLA. A seven point mosaic was used to tile the field of view surrounding the FRB. A total of 4 GHz of bandwidth was used centered at 6.0 GHz. Flux calibration was carried out with 3C147 and phase calibration with B0733−174, the same phase calibra- tor as used for the ATCA. Interference flagging and imaging were carried out using casa. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Table 2. Flux density at 5.5 GHz of the radio source associated with WISE J071634.59−190039.2 from observations taken on the JVLA by Williams & Berger (2016). For 2016 March 23 we use their published value rather than the ATEL value. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | separate bands each with 1 GHz bandwidth centered at 5.5 and 7.5 GHz. Again, 3C147 and B0733−174 were used for flux and phase calibration. Data reduction was carried out using casa. The images were ported into miriad and flux density measurements were made using the task imfit in an identical fashion to the ATCA data. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | In this paper, flux densities were obtained using im- fit under the assumption that the source is unresolved at the observing frequencies and baseline lengths of the ATCA, GMRT and JVLA, an assumption justified by the long base- line results (Bassa et al. 2016). |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | In addition to this effect we have uncovered an issue with the way that miriad deals with the primary beam cor- rection for mosaiced images over wide bandwidths such as used in these observations. The consequence of this effect is that the flux densities as measured with mosaiced images are lower than those made with a single pointing. To quan- tify this effect we took advantage of other sources within the field of view of the large mosaic. We compared the flux densities as measured with the mosaic with those measured when dealing solely with the single pointing containing the source. We find that the flux density of the single pointings are 10 percent higher than those in the mosaics, because of the way miriad handles large fractional bandwidth in the linmos routine. A similar effect also appears to be present in the casa reduction of JVLA mosaics and affects the flux density that we made in our mosaiced JVLA observation. We therefore also extracted the single pointing of the JVLA mosaic that contained the WISE galaxy and processed these data independently. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | For the ATCA data, the two bands (from 4.5 to 6.5 GHz and from 6.5 to 8.5 GHz) are treated independently. We have verified the calibration between the two bands by examining the spectrum of the phase calibrator to ensure that it forms a smooth power law over the entire 4 GHz range. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Finally in order to compare the JVLA and ATCA ob- servations, we need to ensure that there is overlap in the fre- quency coverage. The JVLA observation taken by us from 4.0 to 8.0 GHz is offset from the ATCA observations from 4.5 to 8.5 GHz. The JVLA observations of Williams & Berger (2016) were centered at the same frequencies as the ATCA but only had 1 GHz of bandwidth per sub-band rather than the 2 GHz of the ATCA. Given the spectral behaviour of this source, it is important to keep these differences in mind when trying to compare flux density measurements from the different observations. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | the flux density values presented in In summary, Keane et al. (2016) are self-consistent between epochs but are, in absolute terms, low for two reasons. First, miriad has an incorrect model for the (current) flux density of B0823−500 and secondly miriad incorrectly accounts for primary beam correction across a wide bandwidth for mo- saiced images. The total of these effects means that the Keane et al. (2016) flux densities are too low by about 20 percent in comparison to those obtained with the JVLA. If we require accuracy at the greater than 10 percent level all these effects need to be accounted for. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Potential pitfalls of measuring flux density have been described in great detail in Rau et al. (2016). Effects include clean bias, interference flagging, imaging algorithms used (miriad versus casa), the effects of wide bands and the software used to compute flux densities in either the image or the uv domain. These effects are likely to account for most of the residual differences between measurements noted below. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Table 1 lists the observations and the measurements of the flux density of WISE J071634.59−190039.2. Note that the date of the 4th epoch of ATCA observations is 2015 July 4 and not June 4, as reported in Keane et al. (2016). Where the source was not detected at 7.5 GHz we give a 3−σ upper limit. The source was not detected in any of the GMRT observations and we quote an upper limit likewise. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | In Table 2 we compare the flux density obtained with the JVLA data as reported in Williams & Berger (2016) with those obtained through data reduction carried out by us. These agree within the 1-sigma error bars apart from the data taken on 2016 Mar 23. We note that there is a discrep- ancy between the result in the published paper at this epoch (218±24 µJy) and that given in ATEL #8946 (185±18 µJy) with the latter value consistent with our value. In subsequent analyses we use the values that we derived. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | We note that the brightness obtained with the VLBA on 2016 March 8 of 151±15 µJy/bm is significantly lower than that obtained only 17 h later with the JVLA (192±17 µJy) and the EVN brightness on 2016 March 16 was 125±22 µJy/bm as opposed to the JVLA value of 254±15 µJy taken only a few hours later. One possible inter- pretation is that the EVN and VLBA are resolving the struc- ture leading to a lower brightness. We do not believe this to be the case, however, as the intermediate resolution data from e-MERLIN are consistent with both the ATCA/JVLA and the EVN/VLBA measurements. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | The left hand panel of Figure 1 shows the light curve at 5.5 GHz. Red points are values obtained with the ATCA, blue points from the JVLA, and green points from VLBA, e- MERLIN (Bassa et al. 2016) and the EVN (Giroletti et al. 2016). The right hand panel of the figure shows the 5.5 GHz and 7.5 GHz light curves from measurements taken on the ATCA. It can clearly be seen that the first two epochs have a strongly falling spectrum (negative spectral index), whereas for epochs 3 to 7 the spectrum is flat. Figure 2 shows in more detail the final 100 days of observations. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | For the ATCA observations from epochs 7 to 10, where only single pointings of long duration were obtained, the sig- nal to noise ratio of the radio source is significantly higher than in the earlier epochs. This allows the data to be subdi- vided in frequency and time bins. Figure 3 shows the spectra for these 4 epochs. The spectral index is highly variable. Fi- nally, Figure 4 shows the flux density as a function of time for epochs 7 to 10 at both 5.5 and 7.5 GHz. For each epoch we subdivide the data into 2.5-h time averages in steps of 1.25 h. At 5.5 GHz there is marginal evidence for variability on the timescale of a few hours but nothing like the fac- tor of two change between the measurements on the JVLA and EVN on 2016 March 16. At 7.5 GHz significant time variability is seen from epoch to epoch. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | For a formal metric for variability we use the debiased mod- ulation index, md, and χ2 statistic from Bell et al. (2015) as well as the variability statistic, VS, from Mooley et al. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Figure 1. Left panel: Radio light curve at 5.5 GHz. Black crosses denote ATCA observations, blue triangles JVLA observations, and red squares VLBA, e-MERLIN and EVN. Right panel: Radio light curve from the ATCA data at 5.5 GHz (black crosses) and 7.5 GHz (red squares). |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | where n is the number of data points, Si and σi are the flux density and the error bar for each datum, S and ˜S are the mean and weighted mean flux densities, and Smax and Smin are the maximum and minimum flux densities recorded. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | is rare and, by these measures, both Bell et al. (2015) and Mooley et al. (2016) would clearly classify this source as a variable. Ofek et al. (2011) have only 12 out of 464 sources (∼0.5 percent) with χ2 greater than 100, of these only 2 have a modulation index in excess of 0.2. In the MASIV survey of Lovell et al. (2008), of their 443 sources specifically chosen to be flat spectrum and hence susceptible to variability, only 1 source per epoch showed a modulation index in excess of 0.2. The survey of Mooley et al. (2016) found that 38 out of 3652 sources (∼1 percent) had modulation indices in excess of 0.26 over the period of a week. The usual caveats apply here; all these surveys were either at a different observing frequency or a higher flux density limit or both, compared to our data. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Combining the data taken at 5.5 GHz with the ATCA and the JVLA, ˜S = 140 µJy, md = 0.36, χ2 = 250 and VS = 11. These values change only slightly if the e-MERLIN and VLBA points from Bassa et al. (2016) are included. We note that, in spite of our best efforts at determing the ab- solute calibration, ˜S for the Williams & Berger (2016) data (181 µJy) is significantly larger than the ATCA data. The data taken at 7.5 GHz at later epochs yield ˜S = 120 µJy, md = 0.39, χ2 = 180 and VS = 12. Both the modulation indices and the variability timescales are in line with those predicted by Akiyama & Johnson (2016) at these Galactic latitudes. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Bassa et al. (2016) and Giroletti et al. (2016) show that the ATCA radio source (at least at later epochs) is highly com- pact and located at the centre of the optical galaxy. This indicates that the source is likely an AGN. Although the ∼mas resolution is not sufficient to determine whether the source should undergo interstellar scintillation (ISS) it is certainly indicative. In addition, the quasi-simultaneous ob- servations made by Vedantham et al. (2016) between 1 and 20 GHz indicate that the source has a flat spectral index (modulated by ISS; see Akiyama & Johnson 2016). |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Figure 2. End portion of the radio light curve at 5.5 GHz. Black crosses denote ATCA, blue triangles JVLA and red squares VLBA, EVN or e-MERLIN. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | may be dominated by refractive scintillation. Diffractive scintillation appears unlikely to be contributing to the flux density variations: at the position of the source the expected timescale for diffractive scintillation is ≃ 10 minutes, for an assumed scintillation velocity of 50 km s−1, which is much shorter than the timescale of the fastest variations evident in the lightcurve. Moreover, we note that diffractive scin- tillation is only relevant if the source contains structure on scales below ∼ 0.1µas for the expected scattering properties along this line of sight. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | We therefore investigated the probability that the data are consistent with the form of an intensity probability dis- tribution function expected from ISS in the regime of refrac- tive scintillation. This is a Rician distribution (Cohen et al. 1967) with two free parameters, the amplitude of the non- varying component and the amplitude of the scintillation fluctuations. The probability that the 5.5 GHz light curve is drawn from such a distribution is 90 percent. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Using simple arguments we would also expect that the modulation index of weak radio sources be higher than those of brighter sources if all AGN have a brightness temperature near the Inverse Compton limit. If this is correct then the high modulation index we measure for WISE J071634.59−190039.2 compared to sources in e.g. the MASIV survey (Lovell et al. 2008) is to be expected. In addi- tion, the low galactic latitude of WISE J071634.59−190039.2 also likely contributes to the higher modulation index. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Figure 3. Flux density as a function of frequency for MJDs 57448.5 (crosses and black line), 57457.5 (squares and red line), 57458.5 (triangles and green line) and 57460.5 (circles and blue line) of the ATCA data. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | ple of minor caveats. First, the probability distribution ar- gument does not take into account the time sequence of the data. Although we have no a priori knowledge of what form an afterglow to an FRB might take, the light curve from Keane et al. (2016) is suggestive of a fading radio transient. Indeed, even with the additional data, the probability of getting a value as high as that of the first epoch is only 0.3 per cent. Secondly, ISS in compact AGN is highly intermit- tent (Lovell et al. 2008) and the light curve is only sparsely sampled during the first few months. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | Figure 4. Flux density as a function of time for epochs 7 through 10. Each panel shows a different epoch with the UT time in hours from the MJD listed at the top of the panel. Points with crosses and black lines denote 5.5 GHz, points with squares and red lines denote 7.5 GHz. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | their sources have a modulation index in excess of 0.26 on timescales of less than a week. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | dissimilar to the value given above obtained by considera- tion of the Mooley et al. (2016) survey. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | In summary, we estimate that of order 1 percent of all sources at 5.5 GHz at a flux density limit of 100 µJy show variability with a high modulation index on short timescales. Within the positional error box of FRB 150418 we observe 8 sources brighter than 100 µJy and the probability of ob- serving such a variable source is then ∼8 percent. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.