Filename
stringlengths 22
64
| Paragraph
stringlengths 8
5.57k
|
|---|---|
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | We have fraction of variable sources, with parameters at 5.5 GHz such as those for WISE J071634.59−190039.2, is ∼1 percent, both from con- sidering our knowledge of ISS statistics and from the ob- served number counts of large-scale surveys. Given 300 sources per square degree above 100 µJy at 5.5 GHz, more than 105 sources across the entire sky will show variability over the course of 1 week. This result holds irrespective of the association or not with the FRB. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | However, let us for a moment assume an afterglow in- terpretation for the light curve over the first week. Then, with ∼5000 FRBs per sky per day, any given week would then produce 3.5 × 104 FRB afterglows. We therefore see no obvious disconnect between the number of FRB afterglows and our current knowledge of the variable sky. This is in con- trast to Vedantham et al. (2016) who claimed that FRB af- terglows must be rare. There are three reasons for this differ- ence. First, Vedantham et al. (2016) assumed a modulation index in excess of 0.7 which new data shows is too high. They also assumed the entire flux density from the first epoch was the afterglow but clearly at least a significant fraction of the emission originates in the underlying AGN. Finally, the vol- ume for FRBs probed with the Parkes telescope is signifi- cantly higher than the volume probed for sources similar to WISE J071634.59−190039.2 with the ATCA. If indeed we are observing the afterglow to FRB 150418, ATCA follow- up surveys would not detect them if they were significantly more distant than z ∼ 0.5. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | We have combined observations from the ATCA, GMRT, JVLA, eMERLIN, VLBA and the EVN to show the light curve for the radio source associated with the galaxy WISE J071634.59−190039.2. The source, which Keane et al. (2016) associated with FRB 150418, has, with the acqui- sition of new data, been shown to be highly variable with a modulation index of 0.36, a timescale for variability of order 1 day and a variable spectral index. These statistics are con- sistent with those expected from ISS in the weak scattering regime (Akiyama & Johnson 2016) although the value seen in the first epoch, taken only 2 hours after FRB 150418, remains anomalously high. We show that the probability of detecting such a scintillating source in the error box of the FRB is ∼8 percent, higher than presented in Keane et al. (2016). |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | the low probability of detection, with the high flux density for the first epoch and the time sequence of the light curve. However, this option cannot be ruled out. The second pos- sibility is that the radio source is indeed associated with the FRB. The FRB progenitor would then have to com- ply with the relatively low level of star formation in the galaxy (Keane et al. 2016). If this is typical for FRBs as a class, we might then expect to see a scintillating AGN in conjunction with an FRB event. Only one other FRB has published follow-up to date (Petroff et al. 2015a) but only a single epoch of ATCA data was obtained. Given the small numbers and our lack of understanding of the progenitors of FRBs, the arguments for or against the association essen- tially depend on the probability of observing such a strongly variable source within the field of view of the FRB locali- sation. Therefore, the Keane et al. (2016) conclusion is only one of several viable conclusions, the relative probabilities of which are difficult to assess. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | With only some 20 FRBs known to date (Petroff et al. 2016), progress in the field requires many more detections and accurate (arcsec) localisations of the bursts themselves rather than relying on probability arguments as here. By providing the latter, new instruments such as the Cana- dian Hydrogren Intensity Mapping Experiment (CHIME; Bandura et al. 2014), the Hydrogen Intensity and Real- Time Analysis Experiment (HIREX; Newburgh et al. 2016), the Molonglo Observatory Synthesis Telescope (UTMOST; Bailes et al. In Prep), the Aperture Tile in Focus on Wester- bork (APERTIF; Verheijen et al. 2008), MeerTRAP (Stap- pers & Kramer In Prep) and TRAPUM on MeerKAT1, and the Commensal Real-time ASKAP Fast-Transient Survey (CRAFT; Macquart et al. 2010) on the Australian Square Kilometre Array Pathfinder (ASKAP; DeBoer et al. 2009; Johnston et al. 2008) will undoubtedly break new ground. |
Processed_Radio_light_curve_of_the_galaxy_possibly_associate.txt | The Parkes telescope and the Australia Telescope Compact Array are part of the Australia Telescope National Facility which is funded by the Commonwealth of Australia for oper- ation as a National Facility managed by CSIRO. Parts of this research were conducted by the Australian Research Council Centre of Excellence for All-sky Astrophysics (CAASTRO) through project number CE110001020. Parts of this work were performed on the gSTAR national facility at Swin- burne University of Technology. gSTAR is funded by Swin- burne and the Australian Government’s Education Invest- ment Fund. CGB and EP acknowledge support from the Eu- ropean Research Council under the European Union’s Sev- enth Framework Programme (FP/2007-2013) / ERC Grant Agreement nr. 337062 (DRAGNET; PI Hessels) and nr. 617199 respectively. We thank Sarah Burke-Spolaor for as- sistance with the JVLA data reduction, Matthew Bailes for stimulating discussions and the EVN team for providing re- sults prior to publication. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | ABSTRACT We present a measurement of the Lyman α flux probability distribution function (PDF) obtained from a set of eight high resolution quasar spectra with emission red- shifts at 3.3 ≤ z ≤ 3.8. We carefully study the effect of metal absorption lines on the shape of the PDF. Metals have a larger impact on the PDF measurements at lower redshift, where there are relatively fewer Lyman α absorption lines. This may be ex- plained by an increase in the number of metal lines which are blended with Lyman α absorption lines toward higher redshift, but may also be due to the presence of fewer metals in the intergalactic medium at earlier times. We also provide a new measure- ment of the redshift evolution of the effective optical depth, τeff , at 2.8 ≤ z ≤ 3.6, and find no evidence for a deviation from a power law evolution in the log(τeff ) − log(1 + z) plane. The flux PDF measurements are furthermore of interest for studies of the ther- mal state of the intergalactic medium (IGM) at z ≃ 3. By comparing the PDF to state-of-the-art cosmological hydrodynamical simulations, we place constraints on the temperature of the IGM and compare our results with previous measurements of the PDF at lower redshift. At redshift z = 3, our new PDF measurements are consistent with an isothermal temperature-density relation, T = T0∆γ−1, with a temperature at the mean density of T0 = 19250 ± 4800 K and a slope γ = 0.90 ± 0.21(1σ uncertain- ties). In comparison, joint constraints with existing lower redshift PDF measurements at z < 3 favour an inverted temperature-density relation with T0 = 17900 ± 3500 K and γ = 0.70 ± 0.12, in broad agreement with previous analyses. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | equipped with high-resolution spectrographs, considerable progress has been made toward understanding the nature of these absorption features. This population of discrete lines is due to HI absorption arising from the filamentary structure of the cosmic-web, and in the approximation of a Gaussian velocity dispersion can be well described by a series of Voigt profiles (Rauch 1998). A variable amount of absorption in the Lyman α forest is also due to ultraviolet (UV) transi- tions from heavy element ions which are usually associated with strong Lyman α lines. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | used to constrain cosmological parameters and probe the 1 dark matter power spectrum on scales of order of 50 h− Mpc (e.g. Croft et al. 2002, Viel et al. 2004, McDonald et al. 2006). Another quantity thoroughly studied in the last few years is the flux probability distribution function (PDF), which is sensitive not only to the spatial distribution of dark matter, but also to the thermal state of the intergalactic medium (IGM). Previous studies of the Lyman α forest PDF have shed light on the physical state of the IGM mostly at z < 3. Bolton et al. (2008) compared the flux PDF measured from a sample of QSO spectra presented by Kim et al. (2007, hereafter K07) to a set of hydrodynamical simulations of the Lyman α forest, exploring different cosmological param- eters and various thermal histories. Agreement between the data and simulations was obtained by adopting a power-law temperature-density relation, T = T0∆γ 1 (Hui & Gnedin 1997), where the low density IGM (∆ = ρ/hρi ≤ 10) was close to isothermal (i.e. with γ ∼ 1) or inverted (γ < 1). This result implied that the low-density regions of the IGM may be considerably hotter and their thermal state more complicated than usually assumed at these redshifts. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Radiative transfer effects during the epoch of HeII reionization, which should occur at z > 3, are expected to play a non negligible role in setting this temperature- density relation. Indeed, evidence that the tail-end of HeII reionization is occurring at z ≃ 3 is supported by a variety of observations, such as the opacity of the HeII Lyman α for- est (Reimers et al. 1997; Heap et al. 2000; Kriss et al. 2001, Shull et al. 2010, Worseck et al. 2011, Syphers et al. 2011), an increase in the temperature of the IGM at mean density toward z ≃ 3 (Schaye et al. 2000, Theuns et al. 2002, Lidz et al. 2010, Becker et al. 2011, but see McDonald et al. 2001, Zaldarriaga, Hui, & Tegmark 2001), the observed evolution of the CIV/SiIV metal-line ratio (Songaila & Cowie 1996; Songaila 1998, but see Kim, Cristiani & D’Odorico 2002) and the evolution of the Lyman α forest effective optical depth (Bernardi et al. 2003, Faucher-Giguere et al. 2008, Pˆaris et al. 2011, but see Bolton, Oh & Furlanetto 2009b). However, recent work suggests that photo-heating dur- ing HeII reionization alone is not sufficient to achieve an in- verted temperature-density relation, γ < 1 (McQuinn et al. 2009; Bolton, Oh & Furlanetto 2009a). A recently proposed alternative channel for heating, which naturally produces an inverted temperature-density relation, may be provided by very energetic gamma rays from blazars (Chang et al 2011, Puchwein et al. 2011). In this scenario, very high en- ergy gamma rays annihilate and pair produce on the extra- galactic background light. Powerful plasma instabilities then dissipate the kinetic energy of the electron-positron pairs lo- cally, volumetrically heating the IGM. On the other hand, continuum placement errors could partly explain the prefer- ence of the K07 data for an inverted temperature-density re- lation (see Lee 2012 for a recent discussion). Indeed, Bolton et al. (2008) and more recently Viel et al. (2009) found an isothermal temperature-density relation was consistent (within 1.5σ) with the K07 PDF constraints when treating continuum uncertainties conservatively. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | therefore shed more light on the possible inversion of the temperature-density relation. Such a measurement is also complementary to other constraints on the IGM thermal state, such as those presented recently by Becker et al. (2011), where the temperature-density relation is not mea- sured directly. Such an investigation is the main motivation of the present work. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | In this paper, we use eight high resolution QSO spectra observed at 3.3 ≤ z ≤ 3.8 to extend the study of the Ly- man α forest PDF by K07 to higher redshifts. The aim of this paper is two-fold. First, we perform a detailed study of the redshift evolution of the flux PDF across a wide redshift range. This enables us to study possible systematic trends in the PDF due to redshift-dependent effects, such as the presence of metal absorption lines. In order to corroborate our analysis, we also measure the Lyman α effective optical depth and compare it to different estimates by other authors. In the second step, the measured flux PDF is compared to theoretical results obtained from cosmological hydrodynam- ical simulations. This enables us to investigate the thermal state of the IGM during this interesting redshift range. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | This paper is organized as follows. In Section 2, we present the QSO spectra and the data analysis, and in Sec- tion 3 we present our measurements of the flux PDF and effective optical depth. In Section 4, we introduce the hy- drodynamical simulations used to obtain constraints on the thermal state of the IGM from the PDF and, in Section 5, we report the results from the analysis of the simulations. Finally, in Section 6, we present our conclusions. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The data set used in this paper consists of eight high- redshift QSO spectra. Two of the spectra (PKS2126-158 and PKS2000-330) were obtained with the Ultraviolet and Visual Echelle Spectrograph (UVES) (Dekker et al. 2000) at the Kueyen unit of the European Southern Observa- tory (ESO) VLT (Cerro Paranal, Chile) as part of the ESO Large Programme (LP): ‘The Cosmic Evolution of the IGM’ (Bergeron et al. 2004). PKS1937-101 is from the program Carswell et al. 077.A-0166(A), Q1317-0507, SDSSJ16212- 0042 and Q1249-0159 are from Kim et al. 075.A-0464(A), Q1209+0919 is from the program Dessauges-Zavadsky et al. 073.B-0787(A) and Q0055-269 is from D’Odorico et al. 65.O- 0296(A). All the spectra have been reduced with the stan- dard UVES pipeline provided by ESO. Wavelengths have been corrected to vacuum-heliocentric. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The properties of the QSO spectra used in this paper are summarized in Table 1. The velocity resolution of each 1 full-width half maximum, and each spectrum is 6.7 km s− spectrum is binned in 0.05 ˚A pixels. The signal-to-noise (S/N) ratio has been calculated by considering regions of the spectra redwards of the QSO emission peak, and corre- sponds to the inverse of the standard deviation of the aver- age normalized flux for pixels free of absorption. For some QSOs, the spectra have already been used by D’Odorico et al. (2010) for the study of the evolution of CIV in the IGM. The spectra of three QSOs (Q1317-0507, Q1249-0159 and SDSSJ16212-0042) have been reduced for this work. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The Lyman α forest flux probability distribution at z > 3Based on observations collected at the European Southern Observatory Very Large Telescope, Cerro Paranal, Chile – Programs 077.A-0166(A), 075.A-0464(A), 073.B-0787(A), 166.A-0106(A), 65.O-0296(A). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Table 1. The QSO sample used in this work. In the first column the name of the QSO is reported. Subsequent columns list the emission redshift of the QSO, the redshift range covered by the Lyman α forest, the wavelength range covered by the Lyman α forest and the 1 bluewards signal to noise (S/N, see text for details). Note the wavelength range of the Lyman α forest excludes the region 4000 km s− of the Lyman α emission line in order to avoid the proximity effect (see Sect.2.2). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | a In some cases, the emission redshift zem estimated on the basis of the QSO Lyman α emission line is an underestimation of the true value (see K07). If Lyman α absorption lines with z > zem are present, we assume the highest redshift of the absorption line as a proxy for zem. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 1. The eight QSO spectra which comprise our sample. Each spectrum is normalised to the quasar continuum and the wavelengths are displayed in the QSO rest-frame. For each QSO, the upper panel shows the normalised spectrum, whereas the lower panel displays all the spectral regions containing metal lines. To measure the PDF, these regions have been masked following the criteria described in Sect. 2.3. In each panel, the left dotted line and the right dotted line indicate the wavelengths of the Lyman β and Lyman α transitions. The regions with no flux, such as the one apparent in the spectrum of PKS 2000-330 centered at λ ∼ 1180 ˚A, are due to the wavelength gaps caused by using three separate CCDs in the dichroic setting. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The two QSOs PKS2126158 and Q0055-269 were also included in the K07 sample. It is worth noting that the wave- length ranges for PKS2126158 and Q0055-269 used in this work are different to those reported in Tab. 1 of K07. In the case of PKS2126158, our blue wavelength cut is different to the K07 value due to the presence of two sub-DLAs at z = 2.638 and z = 2.769. We excluded the two regions con- taminated by the sub-DLAs as described above, and a small region between the two sub-DLAs was left in our spectrum, extending from 4470 ˚A to 4540 ˚A. K07 instead decided to re- move this region and cut their spectrum at 4630 ˚A. The red wavelength cut was chosen by K07 in order to avoid having pixels at wavelengths > 5112 ˚A corresponding to redshifts z > 3.2. However, this redshift range is important for study- ing the possible effects of HeII reionization and so we include it in this work. In the case of Q0055-269, the blue cut was performed in our spectrum at a wavelength corresponding to the Lyman β emission line, whereas in the case of K07 they performed it at 4785 ˚A in order to exclude a metal absorption system at 4784 ˚A. The red cut was performed by K07 at 5112 ˚A to again discard pixels with z > 3.2. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | In Fig. 1, we show the reduced spectra for all the QSOs of our sample and, for each spectrum, the contamination from metal lines. The regions containing metal lines have been masked following the criteria described in Sect. 2.3. In Fig. 2, for each QSO spectrum, the redshift coverage and the Lyman α forest wavelength coverage are shown. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | In each spectrum all the absorption lines redwards of the Lyman α emission line, as well as those belonging to the Lyman α forest, have been fitted with Voigt profiles. To satisfactorily perform this task, for each spectrum a precise estimate of the continuum is required. However, the problem of continuum fitting high resolution data requires particu- lar attention, and it has been widely discussed in the lit- erature (e.g. K07; Desjacques et al. 2007; Faucher-Giguere et al. 2008). An automatic procedure to fit QSO spectra has also been developed by Dall’Aglio et al. (2008a; 2008b). They used an extended sample of high resolution, high S/N QSO spectra, and by means of an adaptive techinque involv- ing cubic spline interpolations and a Monte-Carlo method, they were able to estimate the QSO continua and relative uncertainties. In this paper, however, continuum fitting was performed on our smaller data set as described below. It will be interesting to use a larger sample in the future to develop automatic techniques for continuum fitting and investigate associated uncertainties further. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 2. The solid lines display the redshift (and wavelength, see upper axis) coverage of the Lyman α forest in the QSO spectra used in our sample. The dashed lines at z = 2.62, z = 3.17 and z = 3.72 represent the extremes of the two redshift bins used to calculate the flux probability distribution function, as described in Sect. 3. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | package, by means of high-order polynomial functions. The continuum fitting was performed by cutting each spectrum into several chunks, depending on the S/N and on the num- ber of absorption lines, and then joining the chunks back together into one single spectrum. The continua were de- termined by using regions free of absorption lines. This task does not present substantial difficulties when the region red- wards of the QSO Lyman α emission line is considered, since this region presents only a few (metal) absorption lines. However, regions bluewards of the QSO Lyman α emission lines, such as the Lyman α forest, are instead populated by a large number of absorption lines. In general, the higher the emission redshift of the QSO, the greater the average Lyman α forest opacity. This makes continuum fitting toward higher redshift more difficult, since regions which are free of absorp- tion lines become increasingly small and rare. In this case, a linear interpolation between two subsequent absorption-free chunks is used as a first-order guess for the continuum level. Another major problem in continuum fitting concerns the possible presence of wide and shallow absorption features in the spectra. Large-scale absorption can sometimes be incor- rectly attributed to unabsorbed parts of the spectrum. This represents a problem in this procedure, since it can lead to an underestimate of the true continuum. To avoid such mis- takes, it is necessary not only to check carefully every small chunk of the spectrum which is being fitted, but also the joint spectrum by searching for possible wide depressions due to large-scale absorption features which are not visible while working on the individual chunks. All the spectra and the fitted continua have been repeatedly and independently checked in this manner in order to identify such features and to avoid continuum underestimates. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The Lyman α forest flux probability distribution at z > 3Based on observations collected at the European Southern Observatory Very Large Telescope, Cerro Paranal, Chile – Programs 077.A-0166(A), 075.A-0464(A), 073.B-0787(A), 166.A-0106(A), 65.O-0296(A). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 3. An illustration of the uncertainty on the continuum placement for three of the spectra in our sample with different S/N values: S/N=25 for Q1209+0919 (upper panel), S/N=68 for PKS2000-330 (middle panel) and S/N=150 for PKS1937-101 (lower panel). In each panel, the final continuum fit is shown by the solid curve, while the dotted and dashed curves show the fitted continuum after it is increased/decreased by 5 per cent and by 10 per cent, respectively. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | the spectrum with the lowest S/N in Tab. 1 (Q1209+0919, upper panel), one case with an intermediate S/N value (PKS2000-330, middle panel) and the spectrum with the highest S/N (PKS1937-101, lower panel). Although it is very difficult to precisely quantify the uncertainty for the contin- uum determination, Fig. 3 demonstrates that, in the case of Q1209+0919, an uncertainty of the order of 10 per cent is exaggerated, but an uncertainty of 5 per cent seems plau- sible. As stressed earlier, this case represents the spectrum with the worst S/N in our sample. In the other two cases, it appears that a continuum which is decreased or increased by 5 per cent represents an underestimation and overestima- tion, respectively, of the actual continuum. For this reason, a typical uncertainty of 5 per cent should be regarded as conservative and representative of most of the spectra in our sample, in agreement with the analysis of K07 and the one of Dall’Aglio et al. (2008a). Later on, the effects of the continuum estimate on the flux PDF will be investigated in more detail. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Once the continuum fitting was completed, in each spec- trum, all the absorption features blueward of the QSO Ly- man α emission line were then fitted with Voigt profiles. The aim of this procedure is to determine, for each absorption line, the redshift z, the Doppler parameter b and the col- umn density N of the physical system giving rise to that absorption feature. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The Lyman α forest includes the spectral region be- tween the wavelengths of the Lyman β emission line, λβ = 1025.72 (1+zQSO), and the wavelength corresponding to the Lyman α emission line, i.e. λα = 1215.67 (1 + zQSO). To avoid contamination from the QSO line-of-sight proximity effect, in each spectrum we excluded an interval of 4000 km 1 bluewards of the Lyman α emission line. All the absorp- s− tion features were then identified and fitted by means of the RDGEN§, VPFIT¶ and VPGUESSk packages. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Firstly, all the metal absorption lines redwards of the QSO Lyman α emission peak were identified and fitted. This task does not present significant difficulties, since the num- ber of absorption lines in this spectral region is relatively low and lines due to different transitions are rarely blended. Secondly, the identification of heavy element absorption in the Lyman α forest was performed. This task requires more attention owing to the large amount Lyman α lines in this region and to the frequent presence of broad, saturated ab- sorption features. In this spectral region, isolated metal lines are very rare. Most of the metal absorption lines are blended with Lyman α lines, complicating the precise identification of metal line profiles. A common technique to search for metal absorption in the Lyman α forest therefore employs the use of velocity width profiles. Once a strong transition is identified at wavelengths redward of the QSO Lyman α emission line, e.g. due to CIV or SiIV, associated ionic tran- sitions such as SiII, SiIII, CII at the corresponding redshift and velocity profile can be searched for in the Lyman α for- est. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 4. The flux PDF errors as a function of the normalised flux measured from all eight QSOs, computed by means of the jackknife (black solid lines) and the bootstrap (red dashed line) methods. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | to an identified heavy element transition has been assumed to be due to Lyman α absorption. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The presence of metal absorption alters the transmitted frac- tion in the QSO spectra and consequently the shape of the PDF. For this reason, in order to use the PDF as a probe of the physical state of the Lyman α forest, it is necessary to remove the metal absorption from the Lyman α forest in the QSO spectra. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Once the metal identification and Voigt profile fitting is complete, for each QSO we remove metal lines from the spectra in the following manner. The pixels belonging to spectral regions contaminated by identified metal absorp- tion are excised from the spectra using a quantitative crite- rion which depends on the Doppler parameter of the metal absorption line, b. All the pixels at wavelengths, λ, within the interval λ0 − ∆λ ≤ λ ≤ λ0 + ∆λ were removed from the spectra, where λ0 is the central wavelength of the metal absorption line and ∆λ = λZ,0 (1 + z′). Here λZ,0 is the rest- frame wavelength of the transition of a given element Z and z′ = z − dz, where z is the redshift of the absorbing system and dz = 2 b√ln 2 (1+z) . For each QSO spectra, we created a “mask” using this procedure, which is then used to remove the pixels belonging to the metal-contaminated regions. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The Lyman α forest flux probability distribution at z > 3Based on observations collected at the European Southern Observatory Very Large Telescope, Cerro Paranal, Chile – Programs 077.A-0166(A), 075.A-0464(A), 073.B-0787(A), 166.A-0106(A), 65.O-0296(A). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | urated HI absorption, and when this is the case their contri- bution to the total absorption is negligible. However, there will be very few cases where the metal lines are isolated and not identified. For this reason, we believe that the results presented in this paper are not substantially affected by in- complete metal identification in the Lyman α forest. This issue will be discussed further in the section dedicated to the results. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | In figure 4, the 1σ errors for the PDF, estimated using the jackknife and bootstrap techniques, are compared. The PDF is computed from the full sample in this instance, with- out dividing the spectra into different redshift bins. Fig. 4 shows how the two methods provide comparable error esti- mates, although the jackknife estimates are larger at F = 0 and F = 1. In the remainder of the paper, unless otherwise stated, the uncertainties for the PDF will be computed by using the jackknife method. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The PDF is measured using this procedure in bins of width ∆F = 0.05 (McDonald et al. 2000; K07). We have assigned flux values F = 0 and F = 1 to all the pixels characterised by flux levels F < 0.025 and F > 1, respectively. The PDF is calculated both for single QSO spectra, and for the full sample. To study the redshift evolution of the PDF, we also divide the full sample in two redshift bins. The redshift bin width is chosen in order to have comparable numbers of pixels in each bin. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The effect of metal absorption on the PDF is displayed in Figures 5 and 6, where we show the PDF obtained from each of the eight QSOs in our sample. In these two figures, the solid and dashed curves in the large panels show the PDF computed including and excluding metal absorption lines, respectively. In the small panels, for each QSO the ratio between the PDF calculated including metals and after removing the metals is shown. The average Lyman α forest redshift spans from 2.96, for PKS2126-158, up to 3.38 in the case of PKS1937-101. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | In general, the effect of removing metals on the PDF is to decrease the number of pixels with low flux and increase the number of pixels with high flux. As a result, the PDF computed after metal removal is slightly steeper than the PDF calculated before removing the metal absorption. This effect is most clearly visible for some of the QSOs in Fig. 5, such as PKS2126-158. For this spectrum, the effect of metal removal on the PDF is visible for flux values F < 0.6, and is in agreement with the lower redshift analysis of K07, who find a PDF ratio similar to ours. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | For the flux bin Fi, the errors computed with the jackknife method, σJ, are given by the square root of the diagonal of the matrix cov(i, j), i.e. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 5. The effect of metal line absorption on the PDFs obtained for four of the QSOs in our sample. In the main panels, the solid curves display the PDFs calculated for each spectrum before removing metal lines. The solid squares with error bars show the PDFs calculated for the masked spectra, i.e. after removing the metal absorption. The dashed lines are drawn to guide the eye through the solid squares. For each QSO, the name and the average redshift of Lyman α forest absorption is indicated. In the smaller panels, the solid curve represents the ratio between the PDF calculated before and after removing metal lines. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 6. As in Fig.5, but now for the other four QSOs in our sample. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The Lyman α forest flux probability distribution at z > 3Based on observations collected at the European Southern Observatory Very Large Telescope, Cerro Paranal, Chile – Programs 077.A-0166(A), 075.A-0464(A), 073.B-0787(A), 166.A-0106(A), 65.O-0296(A). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Table 2. The mean PDF measured from the full sample, divided into two redshift bins and calculated by (1) using all pixels, (2) after removing contamination from metal absorption lines and (3) after removing Lyman limit systems and metal lines. Note that the redshift bins in the final two columns have a slightly different mean redshift due to the removal of LLS. The redshift bins have been slightly modified in order to have comparable numbers of pixels in each bin. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Table 3. The HI effective optical depth and its dependence on the contamination of metal absorption lines, calculated after dividing each spectrum in two redshift bins. In columns 1, 2 and 3 the name of the QSO, the Lyman α wavelength range and the average redshift are shown, respectively. In columns 4, 5 and 6 we report the HI effective optical depth with its error, the optical depth computed without removing metal lines with its error, and the percentage contribution of metal absorption, respectively. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | is decreasing toward higher redshift. We will return to the issue of metal contamination of QSO spectra in Sect. 3.5. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 7. The ratios between the PDFs measured for each of the eight QSO spectra analysed in this work before and after the removal of metal line absorption. Note that the line thickness increases with the average Lyman α forest redshift. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The effects of the systematic uncertainty due to the con- tinuum fitting procedure are shown in Fig. 8 for the full sample in the redshift bin centered at z = 2.9. In this fig- ure, the PDF is measured without removing the metal lines and is shown for the final continuum level (solid curve), as well as the PDF obtained by increasing and decreasing the continuum level by 5 per cent (dashed and dotted curves, respectively). Here the continuum is conservatively rescaled by 5 per cent since we have seen from Fig. 3 this value is representative of the continuum uncertainty for most of the spectra in our sample. In both cases, the strongest effects on the PDF are visible at high flux values (F > 0.5), whereas a rescaling of the continuum has a negligible effect in the lowest flux bins. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 8. The effect of continuum placement uncertainty on the PDF calculated for our full sample in the lower redshift bin. In the upper panel, the solid line is the PDF calculated with the final continuum. The solid squares with error bars are the PDF values in each flux bin with the associated error estimated by means of the jackknife method described in Sect. 3.1. The short- dashed line, the long-dashed line and the dot-long-dashed line represent the PDF calculated by increasing the continuum level by 5, 3 and 1.5 per cent, respectively. The dotted line, the dot- short-dashed line and the short-dashed-long-dashed line represent the PDF calculated by decreasing the continuum level by 5, 3 and 1.5 per cent, respectively. In the lower panel, and following the same legend as above, the various lines display the ratio between the PDF calculated by increasing or decreasing the continuum and the PDF calculated with the final continuum. The errors on the ratios in the lower panel are similar for all curves but plotted only once for clarity. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | uum placement on the PDF will be local (i.e. it should cause small variations around particular flux values), and will not affect its global average as strongly as Fig. 8 suggests. Re- gardless, however, when we turn to obtaining constraints on the thermal state of the IGM from our PDF measure- ments later, we follow Bolton et al. (2008) and Viel, Bolton & Haehnelt (2009) and only consider the PDF in the flux range F = [0.1–0.8]. This ignores the part of the PDF which is the most sensitive to continuum uncertainties, F > 0.8. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | and S/N data, which are often normalised assuming a power- law extrapolation from the relatively unabsorbed spectral regions uncontaminated by emission lines redwards of the QSO Lyman α. However, more accurate approaches also for this regime, based on principal component analysis, have been developed (e.g. Lee et al. 2012, Paris et al. 2011). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The Lyman α forest flux probability distribution at z > 3Based on observations collected at the European Southern Observatory Very Large Telescope, Cerro Paranal, Chile – Programs 077.A-0166(A), 075.A-0464(A), 073.B-0787(A), 166.A-0106(A), 65.O-0296(A). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 9. The redshift evolution of the PDF with the full sample divided into two redshift bins, hzi = 2.90 (left panels) and hzi = 3.45 (right panels). Upper panels: The squares with error bars represent the PDF measured after removing the metal line absorption. The dashed lines join the squares and are drawn in order to guide the eye. The solid lines are the PDFs computed for the full sample, but including metal absorption lines. Lower panels: The ratios between PDF calculated before and after removing the metal lines. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Fig. 9 shows the PDF computed for the full sample, di- vided into two different redshift bins, centered at redshifts 2.90 and 3.45. This figure also illustrates the effect of metal absorption lines on the PDF. The ratio between the PDF computed before and after removing metal absorption lines depends weakly on redshift. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The PDF for the full sample, divided into two redshift bins, is provided in Table 2. We tabulate the PDF com- puted for (1) all pixels, (2) after removing contamination from metal absorption lines and (3) after removing both metal absorption lines and Lyman limit systems; the latter has been used for the comparison with simulation results as described in Sect. 4. In Fig. 10, the PDFs excluding met- als are compared to the PDFs obtained by K07 at lower redshift. For purposes of clarity, and to better focus on the evolution of the PDF, here the error bars are omitted. As al- ready noted, the spectra at higher redshift are characterised by lower average flux levels. In terms of the PDF, this trans- lates into a greater number of pixels with a low fluxes, re- sulting in a shallower distribution toward higher redshift. This trend is clearly visible in Fig.10. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | checked that these differences are within the error bars esti- mated with the jackknife method, and the results shown in Fig. 10 are unlikely to be significantly affected by the differ- ent methods used to remove metals from the QSO spectra. In addition, in Fig. 10 there is a difference between our PDF measured in the lowest redshift bin and the PDF ob- tained by K07 in their highest redshift bin, i.e. at hzi = 2.94. This difference is better illustrated in Fig 11, where we re- port our PDF calculated at hzi = 2.90 and the PDF mea- sured by K07 at hzi = 2.94. Our PDF is higher than the K07 measurement by a factor ranging from 1.1 to 1.5 in the flux range 0.05 ≤ F ≤ 0.8. Note that two of the QSO spectra used in this study were also used in the PDF measurements presented by K07 (Q0055-269 and PKS2126-158). These ob- jects contribute significantly to the highest bin of K07, as well as to our lowest redshift bin. As a sanity check, we have re-calculated the PDF with the two spectra of Q0055-269 and PKS2126-158 only, by using the pixels in the same wave- length range adopted by K07. The results of this test are visible in Fig. 12, along with jackknife error bars: once the same wavelength ranges as those of K07 for Q0055-269 and PKS2126-158 are taken into account, the two PDFs are con- sistent with each other within the uncertainties. This result tells us that the discrepancy between our PDF and K07’s is not due to a mistake in the computation of the PDF. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 10. The PDF measured in this work at hzi = 2.9 (squares and long-dashed line), hzi = 3.45 (circles and short-dashed line) compared with the PDF obtained by Kim et al. (2007) at hzi = 2.07 (light grey), hzi = 2.52 (grey) and hzi = 2.94 (dark grey). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | since the sum of the ∆z of our QSO spectra contributing to the considered bin is larger than ∼ 50 per cent with respect to K07. We note also that the average redshift of the pixels belonging to our lowest redshift bin is larger than the corre- sponding value for K07s highest redshift bin (hzpixeli = 3.03 versus 2.91, respectively). This difference could contribute to the fact that, as seen in Fig.10, our PDF at hzi = 2.90 is higher than K07s PDF at at hzi = 2.94. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Summarizing, while the evolution with redshift of the flux PDF is evident from Fig.10, more observations of QSOs at z ∼ 3.3 could be needed to establish the true level of the flux PDF at z ∼ 2.9. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | A quantity of interest in cosmology is the effective optical depth of the Lyman α forest, which is also sensitive to the thermal and ionization state of the IGM at high redshift. If hF iHI = exp (−τeff ) represents the average normalized flux measured from the Lyman α forest, the effective opti- cal depth may be defined as τeff ≡ − lnhF iHI. The effective optical depth is therefore related to the average flux in all pixels in the Lyman α forest, and provides different and complementary measure to the PDF, which is a differential measurement. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 11. The PDF measured in this work in the lowest red- shift bin (squares and long-dashed curve) plotted with error bars compared to the PDF obtained by Kim et al. (2007) at hzi = 2.94 (dark grey curve). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 12. The flux PDF measured by K07 at hzi = 2.94 (dark grey curve) plotted with error bars compared to the PDF measured from the two spectra in our sample, Q0055-269 and PKS2126-158, used in the K07 measurement (open squares and long dashed curve). This comparison uses pixels in the same wave- length range adopted K07. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The Lyman α forest flux probability distribution at z > 3Based on observations collected at the European Southern Observatory Very Large Telescope, Cerro Paranal, Chile – Programs 077.A-0166(A), 075.A-0464(A), 073.B-0787(A), 166.A-0106(A), 65.O-0296(A). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 13. The redshift evolution of the Lyman α forest effective optical depth. The solid circles are the measurements from the present work. The triangles are the data from Schaye et al. (2003) and the diamonds are from Kim et al. (2007). The dashed line is the best fit power-law and the solid line the power-law plus Gaussian which best fits the whole sample of data (see text for details). The dotted line is the best fit power-law plus Gaussian obtained by Faucher-Giguere et al. (2008) from the study of a sample of high-resolution spectra. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | Figure 14. Redshift evolution of the ratio between the optical depth computed without removing the metal absorption and af- ter the removal of the metal lines. The solid circles are the val- ues computed in the present work. The triangles are the data by Schaye et al. (2003) and the diamonds are the data by Kim et al. (2007). The dashed line represents a straight-line fit to the data, y = a x + b, with a = 1.37 and b = −0.076. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | after removing a fraction of pixels as described in section 2.2 in order to avoid the proximity effect. Note also that since we wish to measure the HI effective optical depth, re- moving metal lines is again very important for obtaining a precise estimate of τeff which can be compared to other sets of data or with simulations. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | The metal contamination varies from 1 to 28 per cent; we will discuss its redshift dependence in more detail later on in this section. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | effect of metal absorption on the Lyman α forest effective optical depth as a function of redshift. Our data are plot- ted together with the measurements from K07 and Schaye et al. (2003). The dashed line represents a straight-line fit to the data, y = a x + b. Our best fit values are a = 1.36 and b = −0.07. The weak correlation shown in Fig. 14 again highlights the previously discussed redshift dependence of the metal line contamination in the Lyman α forest, which is consistent with a greater amount of metal absorption to- ward lower redshift. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | In this section, we now turn to describe the suite of hydrody- namic simulations used to obtain constraints on the thermal state of the IGM at redshift z > 3 from our PDF measure- ∼ ments. Although the basic setup of the simulations is similar to earlier work by Bolton et al. (2008) and Viel et al. (2009), we now use new, higher resolution runs to explore a wider thermal history parameter space compared with those used by Viel et al. (2009). The simulations were performed us- ing the parallel Tree-PM smoothed particle hydrodynamics (SPH) GADGET-2 code (Springel 2005) and its recently up- dated version GADGET-3. All the simulations were started at z = 99, with initial conditions generated using the trans- fer function of Eisenstein & Hu (1999). The gravitational softening length was chosen to be equal to 1/30th of the mean linear interparticle spacing and star formation was included using a simplified prescription which converts all gas particles with temperature T < 105 K and overdensity ∆ = ρ/hρi > 103 into collisionless stars. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | In Table 4 we summarize the parameters used in the cosmological simulations. We use two different sets of sim- ulations. Set A includes high resolution runs and, among them, the fiducial run for this paper: A1REF. Each simula- 1 Mpc comoving and contains tion has a box size of 10 h− 2 × 5123 dark matter and gas particles. The model chosen is a flat ΛCDM cosmology with parameters: Ω0m = 0.26, 1 ΩΛ = 0.74, Ωbh2 = 0.023, ns = 0.96, H0 = 72 km s− 1 and σ8 = 0.80, which are in agreement with recent Mpc− studies of the cosmic microwave background (Komatsu et al. 2009; Reichardt et al. 2009; Jarosik et al. 2011). |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | We use set A to explore variations in the thermal state of the gas. A spatially uniform ultra-violet background (UVB) produced by quasars and galaxies (Haardt & Madau 2001) was adopted, where hydrogen is reionized at z = 9. |
Processed_The_Lyman_alpha_forest_flux_probability_distributi.txt | respectively. Runs A1LWγ and A1HIγ both have k = 2.20, but q = −1.60 (resulting in an inverted T − ∆ relation with γ = 0.7) and q = −0.45 (flattened T − ∆ relation with γ = 1.3), respectively. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | with Hα galaxies leading to catalogues of millions of objects within a volume of about 58 h−3 Gpc3. Our analysis suggests that both models can be used to provide a robust inference of the parameters (h, ωc) in the redshift range under consideration, with comparable constraining power. We additionally determine the range of validity of the EFTofLSS model in terms of scale cuts and model degrees of freedom. From these tests, it emerges that the standard third-order Eulerian bias expansion can accurately describe the full shape of the real-space galaxy power spectrum up to the maximum wavenumber kmax = 0.45 h Mpc−1, even with a measurement precision well below the percent level. In particular, this is true for a configuration with six free nuisance parameters, including local and non-local bias parameters, a matter counterterm, and a correction to the shot-noise contribution. Fixing either tidal bias parameters to physically-motivated relations still leads to unbiased cosmological constraints. We finally repeat our analysis assuming a volume that matches the expected footprint of Euclid, but without considering observational effects, as purity and completeness, showing that we can get consistent cosmological constraints over this range of scales and redshifts. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | The large-scale distribution of galaxies is an extremely impor- tant source of cosmological information from the low-redshift Universe, standing as a complement to observations of the cos- mic microwave background (CMB), such as the Wilkinson Mi- crowave Anisotropy Probe (WMAP; Hinshaw et al. 2013) and Planck (Planck Collaboration et al. 2020). In the course of the past two decades, observations of the large-scale structure (LSS) from spectroscopic galaxy redshift surveys, such as the 2dF Galaxy Redshift Survey (2dFGRS; Colless et al. 2001), the 6dF Galaxy Survey (6dFGS; Jones et al. 2009), the VIMOS VLT Deep Survey (VVDS; Le Fèvre et al. 2013), the Sloan Digital Sky Survey (SDSS; York et al. 2000), the WiggleZ Dark Energy Survey (WiggleZ; Drinkwater et al. 2010), the VIMOS Public Extragalactic Redshift Survey (VIPERS; Guzzo et al. 2014), the Galaxy And Mass Assembly (GAMA; Driver et al. 2011), and the Baryon Oscillation Spectroscopic Survey (BOSS; Dawson et al. 2012) and its extension (eBOSS; Dawson et al. 2016), have provided a wealth of information on how gravitational instabil- ity shapes the large-scale matter distribution and on the relation between matter and galaxy density perturbations. At the same time such observations have stood as a testing ground for what has ultimately emerged as the standard cosmological model. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | In the next decade, this picture is going to be significantly enriched by the analyses that will be performed by Stage-IV spectroscopic surveys, such as the Dark Energy Spectroscopic Instrument (DESI; DESI Collaboration et al. 2016) and Euclid (Laureijs et al. 2011), which are going to explore a still rela- tively uncharted epoch at 1 ≲ z ≲ 2, when the Universe was about only half of its current age. In particular, Euclid is going to collect the redshift of millions of Hα-emitting galaxies across a total sky surface of 15 000 deg2, therefore increasing the sta- tistical constraining power on the cosmological parameters to an unprecedented level for spectroscopic analyses in the low- redshift Universe. It comes with no surprise that the increase in statistical significance of the observations must necessarily be accompanied by an equivalent increase in the accuracy of the theoretical recipes used to analyse the data, in order to keep the- ory systematic errors at a fraction of the statistical error budget. This becomes even more relevant in terms of the range of va- lidity of the considered models, whose reach must be properly benchmarked against realistic mock samples. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | acoustic oscillation (BAO) and redshift-space distortions (RSD) in the so-called template-fitting approach (Peacock et al. 2001; Tegmark et al. 2006; Guzzo et al. 2008; Blake et al. 2011; Reid et al. 2012; Beutler et al. 2012; Contreras et al. 2013; Howlett et al. 2015; Beutler et al. 2016; Okumura et al. 2016; Alam et al. 2017; Pezzotta et al. 2017; Gil-Marín et al. 2018; Hou et al. 2018; Wang et al. 2018; Zhao et al. 2018), or in full-shape anal- yses (Sánchez et al. 2014, 2016; Grieb et al. 2017; Ivanov et al. 2020; Tröster et al. 2020; d’Amico et al. 2020; Semenaite et al. 2022; Chen et al. 2022; Philcox & Ivanov 2022; Carrilho et al. 2023; Semenaite et al. 2023; Moretti et al. 2023). |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | For both kinds of approaches, the inference of cosmologi- cal information is made intrinsically more difficult by the pres- ence of three separate effects that build on top of linear theory predictions. These are the nonlinear gravitational evolution of the dark matter density field (Bernardeau et al. 2002; Baumann et al. 2012; Carrasco et al. 2012), the relationship between the galaxy δg and the matter δ density fields, known as galaxy bias (Kaiser 1984; Bardeen et al. 1986; see Desjacques et al. 2018 for a recent review), and finally the apparently anisotropic pattern in the distribution of galaxies due to the effect of their peculiar velocities on the observed redshift (RSD, Kaiser 1987; Hamil- ton 1992; Fisher 1995; Scoccimarro et al. 1999; Scoccimarro 2004; Taruya et al. 2010; Senatore & Zaldarriaga 2014; Perko et al. 2016). Each of these effects needs to be carefully mod- elled in order to infer accurate cosmological constraints from the full-shape of the galaxy power spectrum/2PCF. This goal can be achieved in different ways: using numerical methods, such as N- body simulations (e.g., Kuhlen et al. 2012; Schneider et al. 2016; Springel et al. 2021), adopting analytical approaches based on a perturbative solution of the equations governing the evolution of the matter and galaxy density fields (e.g., Fry & Gaztanaga 1993; Bernardeau et al. 2002; McDonald & Roy 2009; Carrasco et al. 2012; Assassi et al. 2014; Senatore 2015; Desjacques et al. 2018), or resorting to hybrid methods that combine the previ- ous two methodologies (e.g., Euclid Collaboration: Knabenhans et al. 2021; Angulo et al. 2021; Zennaro et al. 2023; Pellejero- Ibañez et al. 2023). |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | Scoccimarro et al. 2001; Smith et al. 2007; Manera et al. 2010; Desjacques et al. 2010; Frusciante & Sheth 2012; Schmidt et al. 2013). This is further supplemented by the presence of non-local contributions generated by the cosmic tidal field (Bouchet et al. 1992; Catelan et al. 1998, 2000; McDonald & Roy 2009), which have been proven to be essential for a correct description of the clustering of dark matter halos (Manera & Gaztañaga 2011; Roth & Porciani 2011) and to secure consistency between the results from the analysis of two- and three-point correlation measure- ments (Pollack et al. 2012, 2014). These extra corrections were first detected in N-body simulations in Chan et al. (2012) and Baldauf et al. (2012), and have since then become a standard in- gredient of the bias expansion. Additionally, the latter also takes into account the effects of short-range non-localities during the processes of galaxy formation, which lead to the presence of higher – than second-order – derivatives of the gravitational po- tential. At leading order in the power spectrum, higher deriva- tives appear with a term scaling as ∇2δ (Bardeen et al. 1986; Matsubara 1999; Desjacques 2008; Desjacques et al. 2010). Fi- nally, the dependence on short-wavelength modes is included via an additional stochastic field εg(x) (Dekel & Lahav 1999; Sheth & Lemson 1999; Taruya & Soda 1999; Matsubara 1999; Bonoli & Pen 2009; Hamaus et al. 2010; Schmidt 2016; Ginzburg et al. 2017),1 which is responsible for a shot-noise contribution to the power spectrum. This correction deviates from the predictions of a purely Poissonian distribution, and at the same time can in- troduce a scale dependence due to the physical scale at which two objects can be mistaken for a single one, similarly to the ex- clusion effect for dark matter halos (Scherrer & Weinberg 1998; Sheth & Lemson 1999; Cooray & Sheth 2002; Smith et al. 2007; Baldauf et al. 2013, 2016). |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | The high dimensionality of the parameter space for the model described above can be reduced by employing a set of physically-motivated relations expressing a few higher-order bias parameters in terms of lower-order ones. A typical as- sumption is the conserved evolution of tracers (coevolution), which, from a local-in-matter-density expansion at the moment of formation, leads to the well-known local Lagrangian relations (Chan et al. 2012; Baldauf et al. 2012; Eggemeier et al. 2019). The latter have been adopted in the literature as a fairly conser- vative trade-off between sampling the whole set of bias param- eters and fixing some of the model degrees of freedom, most notably in the analysis of the BOSS DR12 data release (see, e.g., Sánchez et al. 2016; Grieb et al. 2017) to improve the statis- tical constraints of the cosmological parameters obtained from the anisotropic 2PCF and power spectrum. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | cently, Eggemeier et al. (2020) analysed the accuracy of this ex- pansion at fixed cosmology using simulated halo occupation dis- tribution (HOD) catalogs built to mimic the clustering properties of the SDSS Main Galaxy Sample (Strauss et al. 2002) and of the BOSS CMASS and LOWZ samples (Eisenstein et al. 2011; Dawson et al. 2012; Reid et al. 2016). The authors focused on the necessity of introducing both a higher-derivative term and a scale-dependent correction to shot-noise, and showed how the majority of the considered samples preferred the latter when cross-correlating the auto galaxy and the cross galaxy-matter power spectrum. Pezzotta et al. (2021) and Eggemeier et al. (2021) extended this analysis to include the determination of the cosmological parameters, and explored the additional constrain- ing power coming from the one-loop galaxy bispectrum. These works show how fixing the quadratic tidal bias as a function of the linear bias provides accurate results up to k ∼ 0.35 h Mpc−1 for the galaxy power spectrum. A similar analysis was carried out by Oddo et al. (2021), who assessed the constraining power of the galaxy bispectrum on the cosmological parameters, dis- playing a consistency up to k ∼ 0.3 h Mpc−1 for the one-loop power spectrum and k ∼ 0.09 h Mpc−1 for the tree-level bis- pectrum. Equivalent analyses in redshift space (Markoviˇc et al. 2019; Bose et al. 2020; de la Bella et al. 2020; Gualdi et al. 2021; Rizzo et al. 2023; Nicola et al. 2023), or in terms of field-level comparisons (Schmittfull et al. 2019), have also appeared in re- cent years, leading to compatible scenarios. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | On a partially different side, numerical simulations (see, e.g., Kuhlen et al. 2012, for a review) have proven to be an optimal way to reproduce the evolution of the matter and galaxy density field deep into the nonlinear regime. In quantitative terms, state- of-art N-body simulations can achieve a better than 2% accuracy on the shape of the nonlinear matter power spectrum down to scales of k ∼ 10 h Mpc−1 (Schneider et al. 2016; Springel et al. 2021; Angulo et al. 2021), but unfortunately their application as a tool to infer cosmological information from real observations is limited by their extreme computational cost. However, in recent years, different methods have been proposed with the goal of in- creasing their range of applicability. This ranges from methods meant to speed up their production (e.g., Monaco et al. 2002; Tassev et al. 2013; Izard et al. 2016) to ones aiming at finding an optimal interpolation strategy among a limited pool of high- resolution simulations (Heitmann et al. 2013; Liu et al. 2018; Nishimichi et al. 2019; DeRose et al. 2019; Giblin et al. 2019; Wibking et al. 2018; Winther et al. 2019; Euclid Collaboration: Knabenhans et al. 2019, 2021). Among this second category of methods, we highlight baccoemu (Angulo et al. 2021), an emu- lator for the nonlinear matter power spectrum that has been re- cently extended to also include biased tracers in real (Zennaro et al. 2023) and redshift space (Pellejero-Ibañez et al. 2023), assuming a hybrid Lagrangian bias model, with the individual terms of the expansion directly emulated from high-resolution simulations, and a cosmology-rescaling technique meant to re- duce the number of simulations required to train the emulator. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | n-point statistics. All of these effects are going to be investigated by a dedicated group in the Euclid Consortium, while a specific analysis on theory model selection with a more realistic analysis (including survey mask, selection effects, and combining mul- tiple redshift bins) is going to be developed in a future paper (Euclid Collaboration: Moretti et al., in prep.). |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | Table 1. Fiducial parameters of the flat ΛCDM cosmological model of the Flagship I simulation. From left to right, the list includes the value of the Hubble parameter h, the density parameter of cold dark matter ωc, and baryons ωb, the total neutrino mass Mν, the rms density fluctuations inside a sphere of radius 8 h−1 Mpc, σ8, and the scalar index ns. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | Our goal is to test the range of validity of the one-loop galaxy bias expansion, which we quantify by means of three different performance metrics (Osato et al. 2019): the figure of bias, quan- tifying the accuracy of the model in terms of the recovery of the model parameters, the goodness of fit, measuring how well the best fit model compares to the input data vectors, and the figure of merit, quantifying the statistical power of the model in constraining the cosmological parameters. These metrics are computed for each of the configurations we test, as a function of the maximum wave mode kmax included in the fit, exploring different bias relations meant to reduce the dimensionality of the parameter space. Since we limit our attention to the real-space galaxy power spectrum alone, we focus only on the recovery of the dimensionless Hubble parameter h, defined in terms of the Hubble constant H0 as H0 = 100 h km s−1 Mpc−1, and of the cold dark matter density parameter ωc ≡ Ωch2, where Ωc is the corresponding fractional density parameter. At the same time we avoid sampling the scalar amplitude As, since this would lead to a strong degeneracy with the linear bias parameter b1, which can be partially broken only considering the additional constraining power from higher-order statistics, or the apparently anisotropic clustering amplitude when also including RSD. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | This work is the first instalment in a series of Euclid prepa- ration papers, meant to validate the theoretical framework used to analyse the full shape of two- and three-point clustering mea- surements from the final data sample. Here, we focus on the real- space galaxy power spectrum, while the corresponding three- point equivalent for the real-space galaxy bispectrum is going to be presented in Euclid Collaboration: Eggemeier et al. (in prep.). Both of these papers conduct tests in real space, i.e., us- ing the true comoving positions of galaxies inside the box in- stead of the displaced positions due to RSD. While this choice excludes one of the main observational probes of galaxy cluster- ing, such analyses can provide an important testing ground for the model of galaxy bias. This includes the calibration of opti- mal scale cuts for the model, as well as testing different ways to reduce the dimensionality of the parameter space, such as us- ing the coevolution relations. In addition, real-space analyses can become relevant in the context of modelling 3×2-point statis- tics (photometric galaxy clustering, weak gravitational lensing, and galaxy-galaxy lensing), such as in the analysis performed by the Dark Energy Survey (see, e.g., Pandey et al. 2022; Porre- don et al. 2022, for cosmological inference that requires a proper modelling of photometric galaxy bias). On the other hand, fu- ture instalments of this series will focus on the modelling of the redshift-space equivalents of the statistics adopted in these works. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | In order to determine the performance of the selected theoretical models, we first need a set of simulated data samples spanning the same redshift range that will be observed by Euclid, and for which the input cosmology is known a priori. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | In the following we make use of four comoving outputs, se- lected to cover the redshift range 0.9 < z < 1.8 of the Flag- ship I simulation.2 The latter has been carried out on the su- percomputer Piz Daint, which is hosted by the Swiss National Supercomputing Center (CSCS), using the PKDGRAV3 algorithm (Potter et al. 2017), and consists in a record-setting N-body run with two trillions dark matter particles moving under the effect of gravity within a box of size L = 3780 h−1 Mpc. The mass resolution of the simulation (mp ∼ 2.398 × 109 h−1 M⊙ ) allows us to marginally resolve halos with a typical mass Mh of few 1010 h−1 M⊙ , which host the majority of the Hα emission line galaxies that are going to be targeted by Euclid. The nominal flat ΛCDM cosmology adopted to run the simulation as stated in Potter et al. (2017) differs from the fiducial cosmology assumed in this paper in the value of the spectral index (ns = 0.96 vs. ns = 0.97). This choice has been motivated since, during the course of this study, we observed subtle yet significant inconsis- tencies between our models and the measurement in the Flagship I simulation. After contacting the team responsible for running the simulation, they confirmed that the nominal parameters of the simulation are in agreement with the ones we have identified. The latter are obtained by performing cosmological fits to high- resolution dark matter power spectrum measurements at various redshifts, and are summarized in Table 1. This procedure is de- tailed in Appendix A. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | This article is structured in the following way. In Sect. 2 we present the simulated galaxy samples and the power spec- trum measurements and covariances used across this paper. In Sect. 3 we describe the theoretical models that we employed for the analysis, while in Sect. 4 we describe the fitting procedure and the performance metrics that we use to quantify the good- ness of the models for different configurations as a function of the maximum mode included in the fit. Finally, in Sect. 5 we present the results of the analysis, and we draw our conclusions in Sect. 6. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | Table 2. Specifications for the HOD galaxy samples used in this anal- ysis. The table lists the total number of objects Ng, the mean comoving number density ¯n of the sample, and the scale ksn at which Poissonian shot- noise becomes the leading contribution in the galaxy power spec- trum. Following our convention on the normalization of the power spec- trum, the latter is simply defined as the inverse of the mean number den- sity. All the considered samples share the same volume, which coincides with the one of the Flagship I comoving outputs, i.e., (3780 h−1 Mpc)3. The last columns shows the volume factor η between the full-box vol- ume and the one of a Euclid-like shell, as defined in Eq. (2). |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | Fig. 1. HOD profiles of the eight Hα samples employed in this analysis. Individual panels show the profiles at different redshift (increasing from the top left to the bottom right panel) for both Model 1 (blue) and 3 (orange). Solid and dashed lines identify the average number of central and satellite galaxies, respectively, in halos of mass Mh. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | ers allows us to neglect any extra contribution (Euclid Collabo- ration: Risso et al., in prep, Euclid Collaboration: Lee et al., in prep.) to the model galaxy power spectrum presented in Sect. 3. As a consequence, given the high precision assumed to validate the theory models, we believe that our tests should provide a con- servative estimate of their range of validity. We leave to future Euclid analyses a more dedicated study of the impact of obser- vational systematics. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | Figure 1 shows the HOD profiles, i.e., the expected number of galaxies per halo, ⟨Ng⟩, as a function of the host halo mass Mh, of the eight samples: four redshifts times two different mod- els. In all panels, both centrals and satellites profiles are shown, marked respectively with continuous and dashed lines. Not sur- prisingly, the mean occupation of central galaxies inside dark matter halos does not converge to one, even for the most mas- sive halos selected by the halo finder. This is a consequence of the intrinsic property of Hα emitters to be relatively young, blue, and star-forming galaxies. On the contrary, most of the galaxies residing at the center of the host halos are giant red ellipticals, whose star formation is almost completely quenched. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | For each of the samples described above, we measured the real- space galaxy power spectrum Pgg(k) using the publicly available PowerI4 code.5 The latter provides the functionality to compute the power spectrum from a particle distribution within a regu- lar cubic box, using a variety of particle assignment schemes. For this analysis, we made use of a fourth-order interpolation scheme, otherwise known as piecewise cubic spline (PCS; see Sefusatti et al. 2016, for the exact form of the kernel), coupled with an interlacing method to reduce the aliasing contribution at high wave modes k. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | ferently from the evolution of the matter power spectrum, the galaxy power spectrum features an increasingly lower amplitude at lower redshifts. This can be explained by a larger linear galaxy bias at high redshift that overcomes the growth of matter fluctu- ations. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | where Nk identifies the number of independent wave modes falling in the bin [k − ∆k/2, k + ∆k/2], while Pgg(k) is the the- oretical nonlinear galaxy power spectrum including shot-noise contributions. The latter has been obtained from a preliminary fit of the full nonlinear model to the data vector of each sam- ple, assuming, in a first iteration, an approximate but reasonable evaluation of the covariance itself. We expect the Gaussian ap- proximation to be sufficient to our goals. The only additional contribution due to the galaxy trispectrum, here neglected, while noticeable at the relevant scales (Scoccimarro et al. 1999; Se- fusatti et al. 2006; Blot et al. 2015, 2016; Bertolini et al. 2016; Wadekar & Scoccimarro 2020) does not lead to significant differ- ences (≲ 10%) on parameters constraints in the mildly nonlinear regime (Blot et al. 2019; Wadekar et al. 2020). This is also sup- ported by the goal of this analysis, which is testing the relative performance of different theory models rather than providing ab- solute values for the parameter uncertainties. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | The bottom panel of Fig. 2 shows the standard deviation nor- malised by the corresponding galaxy power spectrum. The eight cases largely coincide, but differ at large scales due to the large scatter in the measured Pgg, and at small scales due to the differ- ent shot-noise contributions of each sample, which makes the curves deviate from a quadratic relation (shown with a black dashed line). It should be noted how the relative error is well below the 1% level for all k > 0.1 h Mpc−1. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | Fig. 2. Top: galaxy power spectrum measurements of the Model 1 HOD samples. The color gradient identifies the different redshifts of the sam- ples, as shown in the legend. Dashed horizontal lines correspond to the amplitude of the Poisson shot-noise term P sn – obtained as the inverse of the number density specified in the last column of Table 2 – for the different redshifts. Center: same as for the previous panel but for the Model 3 HOD samples. Bottom: error-to-measurement ratio for the different HOD samples, assuming a Gaussian covariance matrix as in Eq. (1). The coloured solid lines are obtained using the Poisson noise- subtracted power spectra, while the dashed black line highlights the lin- ear relationship from Eq. (1), i.e., 2/Nk. Grey bands mark the 1%, 0.5%, and 0.1% limit. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | namely ∆k = 2k F and ∆k = 3k F, showing how the final constraints do not depend significantly on this choice. |
Processed_Euclid_preparation._TBD._Galaxy_power_spectrum_mod.txt | such that the rescaled covariance Cshell can be expressed 7 in terms of the original full-box covariance Cbox as Cshell = η Cbox . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.