arXiv subjects

C. M. Baugh

Publications and source records attributed to C. M. Baugh.

At least 55 records · Page 3Linked to original sources

arXiv2013

How well can we really estimate the stellar masses of galaxies from broadband photometry?

The estimated stellar masses of galaxies are widely used to characterize how the galaxy population evolves over cosmic time. If stellar masses can be estimated in a robust manner, free from any bias, global diagnostics such as the stellar mass function can be used to constrain the physics of galaxy formation. We explore how galaxy stellar masses, estimated by fitting broad-band spectral energy distributions (SEDs) with stellar population models, can be biased as a result of commonly adopted assumptions for the star-formation and chemical enrichment histories, recycled fractions and dust attenuation curves of galaxies. We apply the observational technique of broad-band SED fitting to model galaxy SEDs calculated by the theoretical galaxy formation model GALFORM, isolating the effect of each of these assumptions. We find that, averaged over the entire galaxy population, the common assumption of exponentially declining star-formation histories does not adversely affect stellar mass estimation. We show that fixing the metallicity in SED fitting or using sparsely sampled metallicity grids can introduce mass dependent systematics into stellar mass estimates. We find that the common assumption of a star-dust geometry corresponding to a uniform foreground dust screen can cause the stellar masses of dusty model galaxies to be significantly underestimated. Finally, we show that stellar mass functions recovered by applying SED fitting to model galaxies at high redshift can differ significantly in both shape and normalization from the intrinsic mass functions predicted by a given model. Given these differences, our methodology of using stellar masses estimated from model galaxy SEDs offers a new, self-consistent way to compare model predictions with observations.

astro-ph.CO↗

arXiv2013

Constraints on black hole fuelling modes from the clustering of X-ray AGN

We present a clustering analysis of X-ray selected AGN by compiling X-ray samples from the literature and re-estimating the dark matter (DM) halo masses of AGN in a uniform manner. We find that moderate luminosity AGN (Lx(2-10 keV)=10^42-10^44 erg/sec) in the z=0-1.3 Universe are typically found in DM haloes with masses of ~10^13 Msun. We then compare our findings to the theoretical predictions of the coupled galaxy and black hole formation model GALFORM. We find good agreement when our calculation includes the hot-halo mode of accretion onto the central black hole. This type of accretion, which is additional to the common cold accretion during disk instabilities and galaxy mergers, is tightly coupled to the AGN feedback in the model. The hot-halo mode becomes prominent in DM haloes with masses greater than ~10^12.5 Msun, where AGN feedback typically operates, giving rise to a distinct class of moderate luminosity AGN that inhabit rich clusters and superclusters. Cold gas fuelling of the black hole cannot produce the observationally inferred DM halo masses of X-ray AGN. Switching off AGN feedback in the model results in a large population of luminous quasars (Lx(2-10 keV) > 10^44 erg/sec) in DM haloes with masses up to ~10^14 Msun, which is inconsistent with the observed clustering of quasars. The abundance of hot-halo AGN decreases significantly in the z~3-4 universe. At such high redshifts, the cold accretion mode is solely responsible for shaping the environment of moderate luminosity AGN. Our analysis supports two accretion modes (cold and hot) for the fuelling of supermassive black holes and strongly underlines the importance of AGN feedback in cosmological models both of galaxy formation and black hole growth.

astro-ph.CO↗

arXiv2013

The most luminous quasars do not live in the most massive dark matter haloes at any redshift

Quasars represent the brightest Active Galactic Nuclei (AGN) in the Universe and are thought to indicate the location of prodigiously growing Black Holes (BHs), with luminosities as high as 10^48 erg/sec. It is often expected though that such an extremely energetic process will take place in the most massive bound structures in the dark matter (DM) distribution. We show that in contrast to this expectation, in a galaxy formation model which includes AGN feedback, quasars are predicted to live in average DM halo environments with typical masses of a few times 10^12 Msun. This fundamental prediction arises from the fact that quasar activity (i.e., BH accretion with luminosity greater than 10^46 erg/sec) is inhibited in DM haloes where AGN feedback operates. The galaxy hosts of quasars in our simulations are identified with over massive (in gas and stars) spheroidal galaxies, in which BH accretion is triggered via a galaxy merger or secular processes. We further show that the z=0 descendants of high redshift (z~6) QSOs span a wide range of morphologies, galaxy and halo masses. The z~6 BHs typically grow only by a modest factor by the present day. Remarkably, high redshift QSOs never inhabit the largest DM haloes at that time and their descendants are very seldom found in the most massive haloes at z=0. We also show that observationally it is very likely to find an enhancement in the abundance of galaxies around quasars at z~5. However, these enhancements are considerably weaker compared to the overdensities expected at the extreme peaks of the DM distribution. Thus, it is very unlikely that a quasar detected in the $z\gtrsim5$ Universe pinpoints the location of the progenitors of superclusters in the local Universe.

astro-ph.CO↗

arXiv2013

Luminosity bias: from haloes to galaxies

Large surveys of the local Universe have shown that galaxies with different intrinsic properties, such as colour, luminosity and morphological type display a range of clustering amplitudes. Galaxies are therefore not faithful tracers of the underlying matter distribution. This modulation of galaxy clustering, called bias, contains information about the physics behind galaxy formation. It is also a systematic to be overcome before the large-scale structure of the Universe can be used as a cosmological probe. Two types of approaches have been developed to model the clustering of galaxies. The first class is empirical and filters or weights the distribution of dark matter to reproduce the measured clustering. In the second approach an attempt is made to model the physics which governs fate of baryons in order to predict the number of galaxies in dark matter haloes. I will review the development of both approaches and summarize what we have learnt about galaxy bias.

astro-ph.CO↗

arXiv2012

The UV colours and dust attenuation of Lyman-break galaxies

Using GALFORM, a semi-analytical model of galaxy formation in the Lambda cold dark matter cosmology, we study the rest-frame ultraviolet (UV) colours of Lyman-break galaxies (LBGs) in the redshift range 2.5 < z < 10. As the impact of dust on UV luminosity can be dramatic, our model includes a self-consistent computation of dust attenuation based on a radiative transfer model. We find that intrinsically brighter galaxies suffer stronger dust attenuation than fainter ones, though the relation has a large scatter. The model predicts galaxies with UV colours consistent with the colour selection regions designed to select LBGs in observational surveys. We find that the drop-out technique that selects LBGs based on two rest-frame UV colours is robust and effective, selecting more than 70 per cent of UV bright galaxies at a given redshift. We investigate the impact on the predicted UV colours of varying selected model parameters. We find that the UV colours are most sensitive to the modelling of dust attenuation and in particular, to the extinction curve used in the radiative transfer calculation. If we assume a Milky Way dust extinction curve, the predicted UV continuum slopes are, in general, bluer than observed. However, we find that the opposite is true when using the Small Magellanic Cloud dust extinction curve. This demonstrates the strong dependence of UV colours on dust properties and highlights the inadequacy of using the UV continuum slope as a tracer of dust attenuation without any further knowledge of the galaxy inclination or dust characteristics in high redshift galaxies.

astro-ph.CO↗

arXiv2012

On the role of feedback in shaping the cosmic abundance and clustering of neutral atomic hydrogen in galaxies

We investigate the impact of feedback - from supernovae (SNe), active galactic nuclei (AGN) and a photo-ionizing background at high redshifts - on the neutral atomic hydrogen (HI) mass function, the $b_{\rm J}$ band luminosity function, and the spatial clustering of these galaxies at $z$=0. We use a version of the semi-analytical galaxy formation model GALFORM that calculates self-consistently the amount of HI in a galaxy as a function of cosmic time and links its star formation rate to its mass of molecular hydrogen (H$_2$). We find that a systematic increase or decrease in the strength of SNe feedback leads to a systematic decrease or increase in the amplitudes of the luminosity and HI mass functions, but has little influence on their overall shapes. Varying the strength of AGN feedback influences only the numbers of the brightest or most HI massive galaxies, while the impact of varying the strength of photo-ionization feedback is restricted to changing the numbers of the faintest or least HI massive galaxies.Our results suggest that the HI mass function is a more sensitive probe of the consequences of cosmological reionization for galaxy formation than the luminosity function. We find that increasing the strength of any of the modes of feedback acts to weaken the clustering strength of galaxies, regardless of their HI-richness. In contrast, weaker AGN feedback has little effect on the clustering strength whereas weaker SNe feedback increases the clustering strength of HI-poor galaxies more strongly than HI-rich galaxies. These results indicate that forthcoming HI surveys on next generation radio telescopes such as the Square Kilometre Array and its pathfinders will be exploited most fruitfully as part of multiwavelength survey campaigns.

astro-ph.CO↗

arXiv2012

The contribution of star-forming galaxies to fluctuations in the cosmic background light

Star-forming galaxies which are too faint to be detected individually produce intensity fluctuations in the cosmic background light. This contribution needs to be taken into account as a foreground when using the primordial signal to constrain cosmological parameters. The extragalactic fluctuations are also interesting in their own right as they depend on the star formation history of the Universe and the way in which this connects with the formation of cosmic structure. We present a new framework which allows us to predict the occupation of dark matter haloes by star-forming galaxies and uses this information, in conjunction with an N-body simulation of structure formation, to predict the power spectrum of intensity fluctuations in the infrared background. We compute the emission from galaxies at far-infrared, millimetre and radio wavelengths. Our method gives accurate predictions for the clustering of galaxies both for the one halo and two halo terms. We illustrate our new framework using a previously published model which reproduces the number counts and redshift distribution of galaxies selected by their emission at $850\,μ$m. Without adjusting any of the model parameters, the predictions show encouraging agreement at high frequencies and on small angular scales with recent estimates of the extragalactic fluctuations in the background made from early data analysed by the Planck Collaboration. There are, however, substantial discrepancies between the model predictions and observations on large angular scales and at low frequencies, which illustrates the usefulness of the intensity fluctuations as a constraint on galaxy formation models.

astro-ph.CO↗

arXiv2012

Massive, red galaxies in a hierarchical universe-II Clustering of Extremely Red Objects

We present predictions for the clustering of Extremely Red Objects (EROs) in a Λcold dark matter Universe, using a semi-analytical galaxy formation model in combination with a cosmological N-body simulation. EROs are red, massive galaxies observed at 0.7< z < 3, and their numbers and properties have posed a challenge to hierarchical galaxy formation models. We analyse the halo occupation distribution and two-point correlation function of EROs, exploring how these quantities change with apparent magnitude, colour cut and redshift. Our model predicts a halo occupation distribution that is significantly different from that typically assumed. This is due to the inclusion of AGN feedback, which changes the slope and scatter of the luminosity-host halo mass relation above the mass where AGN feedback first becomes important. We predict that, on average, dark matter haloes with masses above 10^{13}h^{-1}Msun host at least one ERO at 1.5 < z < 2.5. Taking into account sample variance in observational estimates, the predicted angular clustering for EROs with either (R-K)>5 or (i-K)>4.5 is in reasonable agreement with current observations.

astro-ph.CO↗

arXiv2012

Clustering properties of BzK-selected galaxies in GOODS-N: environmental quenching and triggering of star formation at z ~ 2

Using a sample of BzK-selected galaxies at z~2 identified from the CFHT/WIRCAM near-infrared survey of GOODS-North, we discuss the relation between star formation rate (SFR), specific star formation rate (SSFR), and stellar mass (M_{*}), and the clustering of galaxies as a function of these parameters. For star-forming galaxies (sBzKs), the UV-based SFR, corrected for extinction, scales with the stellar mass as SFR ~ M_{*}^{alpha} with alpha = 0.74+/-0.20 down to M_{*} ~ 10^{9} M_{solar}, indicating a weak dependence on the stellar mass of the SSFR. We also measure the angular correlation function and hence infer the correlation length for sBzK galaxies as a function of M_{*}, SFR, and SSFR, as well as K-band apparent magnitude. We show that passive galaxies (pBzKs) are more strongly clustered than sBzK galaxies at a given stellar mass, mirroring the color-density relation seen at lower redshifts. We also find that the correlation length of sBzK galaxies ranges from 4 to 20 h^{-1}Mpc, being a strong function of M_{K}, M_{*}, and SFR. On the other hand, the clustering dependence on SSFR changes abruptly at 2x10^{-9} yr^{-1}, which is the typical value for "main sequence" star-forming galaxies at z~2. We show that the correlation length reaches a minimum at this characteristic value, and is larger for galaxies with both smaller and larger SSFRs; a dichotomy that is only marginally implied from the predictions of the semi-analytical models. Our results suggest that there are two types of environmental effects at work at z~2. Stronger clustering for relatively quiescent galaxies implies that the environment has started to play a role in quenching star formation. At the same time, stronger clustering for galaxies with elevated SSFRs ("starbursts") might be attributed to an increased efficiency for galaxy interactions and mergers in dense environments.

astro-ph.CO↗

arXiv2012

The clustering of H-alpha emitters at z=2.23 from HiZELS

(abridged) We present a clustering analysis of 370 high-confidence H-alpha emitters (HAEs) at z=2.23. The HAEs are detected in the Hi-Z Emission Line Survey (HiZELS), a large-area blank field 2.121um narrowband survey using the United Kingdom Infrared Telescope (UKIRT) Wide Field Camera (WFCAM). Averaging the two-point correlation function of HAEs in two ~1 degree scale fields (United Kingdom Infrared Deep Sky Survey/Ultra Deep Survey [UDS] and Cosmological Evolution Survey [COSMOS] fields) we find a clustering amplitude equivalent to a correlation length of r_0=3.7+/-0.3 Mpc/h for galaxies with star formation rates of >7 M_sun/yr. The data are well-fitted by the expected correlation function of Cold Dark Matter, scaled by a bias factor: omega_HAE=b^2 omega_DM where b=2.4^{+0.1}_{-0.2}. The corresponding 'characteristic' mass for the halos hosting HAEs is log(M_h/[M_sun/h])=11.7+/-0.1. Comparing to the latest semi-analytic GALFORM predictions for the evolution of HAEs in a LCDM cosmology, we find broad agreement with the observations, with GALFORM predicting a HAE correlation length of ~4 Mpc/h. Motivated by this agreement, we exploit the simulations to construct a parametric model of the halo occupation distribution (HOD) of HAEs, and use this to fit the observed clustering. Our best-fitting HOD can adequately reproduce the observed angular clustering of HAEs, yielding an effective halo mass and bias in agreement with that derived from the scaled omega_DM fit, but with the relatively small sample size the current data provide a poor constraint on the HOD. Our results support the broad picture that 'typical' (~L*) star-forming galaxies have been hosted by dark matter haloes with M_h<10^12 M_sun/h since z~2, but with a broad occupation distribution and clustering that is likely to be a strong function of luminosity.

astro-ph.CO↗

arXiv2011

Testing dark energy using pairs of galaxies in redshift space

The distribution of angles subtended between pairs of galaxies and the line of sight,which is uniform in real space, is distorted by their peculiar motions, and has been proposed as a probe of cosmic expansion. We test this idea using N-body simulations of structure formation in a cold dark matter universe with a cosmological constant and in two variant cosmologies with different dark energy models. We find that the distortion of the distribution of angles is sensitive to the nature of dark energy. However, for the first time, our simulations also reveal dependences of the normalization of the distribution on both redshift and cosmology that have been neglected in previous work. This introduces systematics that severely limit the usefulness of the original method. Guided by our simulations, we devise a new, improved test of the nature of dark energy. We demonstrate that this test does not require prior knowledge of the background cosmology and that it can even distinguish between models that have the same baryonic acoustic oscillations and dark matter halo mass functions. Our technique could be applied to the completed BOSS galaxy redshift survey to constrain the expansion history of the Universe to better than 2%. The method will also produce different signals for dark energy and modified gravity cosmologies even when they have identical expansion histories, through the different peculiar velocities induced in these cases.

astro-ph.CO↗

arXiv2011

The nature and descendants of Lyman-break galaxies in the LambdaCDM cosmology

We predict the formation histories, properties and descendants of Lyman-break galaxies (LBGs) in the Lambda-CDM cosmology. In our model, which incorporates a top-heavy IMF in starbursts, we find that most LBGs are starbursts triggered by minor mergers of galaxies. The duration of the LBG phase is predicted to be quite short, ~20-60 Myr. We investigate the distributions of stellar and halo masses and morphologies for bright (L_UV > L*_UV) and faint (L_UV > 0.1 L*_UV) LBGs at z=3, z=6 and z=10 (where we classify LBGs according to their rest-frame UV luminosities relative the observed characteristic luminosity L*_UV at z \approx 3). Bright LBGs at z=3 are predicted to have median stellar masses ~ 1x10^9 Msun/h and host halo masses ~ 3x10^{11} Msun/h, and to be typically mildly disk-dominated in stellar mass. On the other hand, faint LBGs at z=10 are predicted to have median stellar masses of only ~ 1x10^7 Msun/h and host halo masses 2x10^{10} Msun/h, and to be generally bulge-dominated. Bright LBGs at z=3 evolve to present-day galaxies with median stellar mass ~ 5x^{10} Msun/h (comparable to the Milky Way), consisting of roughly equal numbers of disk- and bulge-dominated systems, and hosted by halos with median mass ~2x10^{13} Msun/h (corresponding to medium-size galaxy groups). The model predicts that 40% of Milky Way mass galaxies at the present-day have a bright LBG progenitor in the redshift range 3<z<4, while 95% have a faint LBG progenitor in the same redshift range, and 7% have a faint LBG progenitor at 10<z<11. With our multiwavelength model, we also investigate the overlap between the LBG population and that of submillimetre selected galaxies (SMGs); at z=3, only ~1% of bright LBGs are also predicted to also be bright SMGs (with an 850 mum flux in excess of 5 mJy).

astro-ph.CO↗

arXiv2011

Are the 2dFGRS superstructures a problem for hierarchical models?

We introduce an objective method to assess the probability of finding extreme events in the distribution of cold dark matter such as voids, overdensities or very high mass haloes. Our approach uses an ensemble of N-body simulations of the hierar- chical clustering of dark matter to find extreme structures. The frequency of extreme events, in our case the cell or smoothing volume with the highest count of cluster-mass dark matter haloes, is well described by a Gumbel distribution. This distribution can then be used to forecast the probability of finding even more extreme events, which would otherwise require a much larger ensemble of simulations to quantify. We use our technique to assess the chance of finding concentrations of massive clusters or super- clusters, like the two found in the two-degree field galaxy redshift survey (2dFGRS), using a counts-in-cells analysis. The Gumbel distribution gives an excellent descrip- tion of the distribution of extreme cell counts across two large ensembles of simulations covering different cosmologies, and also when measuring the clustering in both real and redshift space. We find examples of structures like those found in the 2dFGRS in the simulations. The chance of finding such structures in a volume equal to that of the 2dFGRS is around 2%.

astro-ph.CO↗

arXiv2010

The role of submillimetre galaxies in hierarchical galaxy formation

We study the role of submillimetre galaxies (SMGs) in the galaxy formation process in the Lambda Cold Dark Matter cosmology. We use the Baugh et al. (2005) semi-analytical model, which matches the observed SMG number counts and redshift distribution by assuming a top-heavy initial mass function (IMF) in bursts triggered by galaxy mergers. We build galaxy merger trees and follow the evolution and properties of SMGs and their descendants. Our primary sample of model SMGs consists of galaxies which had 850 mu fluxes brighter than 5 mJy at some redshift z>1. Our model predicts that the present-day descendants of such SMGs cover a wide range of stellar masses ~ 10^{10} - 10^{12} Msun/h, with a median ~ 10^{11} Msun/h, and that more than 70% of these descendants are bulge-dominated. More than 50% of present day galaxies with stellar masses larger than 7 x 10^{11} Msun/h are predicted to be descendants of such SMGs. We find that although SMGs make an important contribution to the total star formation rate at z~2, the final stellar mass produced in the submillimetre phase contributes only 0.2% of the total present-day stellar mass, and 2% of the stellar mass of SMG descendants, in stark contrast to the popular picture in which the SMG phase marks the production of the bulk of the mass of present day massive ellipticals.

astro-ph.CO↗

arXiv2010

The evolution of AGN across cosmic time: what is downsizing?

We use a coupled model of the formation and evolution of galaxies and black holes (BH) to study the evolution of active galactic nuclei (AGN) in a cold dark matter universe. The model predicts the BH mass, spin and mass accretion history. BH mass grows via accretion triggered by discs becoming dynamically unstable or galaxy mergers (called the starburst mode) and accretion from quasi-hydrostatic hot gas haloes (called the hot-halo mode). By taking into account AGN obscuration, we obtain a very good fit to the observed luminosity functions (LF) of AGN (optical, soft and hard X-ray, and bolometric) for a wide range of redshifts (0<z<6). The model predicts a hierarchical build up of BH mass, with the typical mass of actively growing BHs increasing with decreasing redshift. Remarkably, despite this, we find downsizing in the AGN population, in terms of the differential growth with redshift of the space density of faint and bright AGN. This arises naturally from the interplay between the starburst and hot-halo accretion modes. The faint end of the LF is dominated by massive BHs experiencing quiescent accretion via a thick disc, primarily during the hot-halo mode. The bright end of the LF, on the other hand, is dominated by AGN which host BHs accreting close to or in excess of the Eddington limit during the starburst mode. The model predicts that the comoving space density of AGN peaks at z~3, similar to the star formation history. However, when taking into account obscuration, the space density of faint AGN peaks at lower redshift (z<2) than that of bright AGN (z~2-3). This implies that the cosmic evolution of AGN is shaped in part by obscuration.

astro-ph.CO↗

arXiv2010

The evolution of Lyman-break galaxies in CDM

We make a detailed investigation of the properties of Lyman-break galaxies (LBGs) in the LambdaCDM model. We present predictions for two published variants of the GALFORM semi-analytical model: the Baugh et al. (2005) model, which has star formation at high redshifts dominated by merger-driven starbursts with a top-heavy IMF, and the Bower et al. (2006) model, which has AGN feedback and a standard Solar neighbourhood IMF throughout. We show predictions for the evolution of the rest-frame far-UV luminosity function in the redshift range z=3-20, and compare with the observed luminosity functions of LBGs at z=3-10. We find that the Baugh et al. model is in excellent agreement with these observations, while the Bower et al. model predicts too many high-luminosity LBGs. Dust extinction, which is predicted self-consistently based on galaxy gas contents, metallicities and sizes, is found to have a large effect on LBG luminosities. We compare predictions for the size evolution of LBGs at different luminosities with observational data for 2<z<7, and find the Baugh et al. model to be in good agreement. We present predictions for stellar, halo and gas masses, star formation rates, circular velocities, bulge-to-disk ratios, gas and stellar metallicities and clustering bias, as functions of far-UV luminosity and redshift. We find broad consistency with current observational constraints. We then present predictions for the abundance and angular sizes of LBGs out to very high redshift (z<20), finding that planned deep surveys with JWST should detect objects out to z<15. The typical UV luminosities of galaxies are predicted to be very low at high redshifts, which has implications for detecting the galaxies responsible for reionizing the IGM; for example, at z=10, 50% of the ionizing photons are expected to be produced by galaxies fainter than M_AB(1500A)-5logh ~ -15.

astro-ph.CO↗

arXiv2010

Modelling the dusty universe II: The clustering of submillimetre-selected galaxies

We combine the GALFORM semi-analytical model of galaxy formation, which predicts the star formation and merger histories of galaxies, the GRASIL spectro-photometric code, which calculates the spectral energy distributions (SEDs) of galaxies self-consistently including reprocessing of radiation by dust, and artificial neural networks (ANN), to investigate the clustering properties of galaxies selected by their emission at submillimetre wavelengths (SMGs). We use the Millennium Simulation to predict the spatial and angular distribution of SMGs. At redshift z = 2, we find that these galaxies are strongly clustered, with a comoving correlation length of r0 = 5.6 \pm 0.9 Mpc/h for galaxies with 850μm flux densities brighter than 5 mJy, in agreement with observations. We predict that at higher redshifts these galaxies trace denser and increasingly rarer regions of the universe. We present the predicted dependence of the clustering on luminosity, submillimetre colour, halo and total stellar masses. Interestingly, we predict tight relations between correlation length and halo and stellar masses, independent of sub-mm luminosity.

astro-ph.CO↗

arXiv2010

Modelling the spectral energy distribution of galaxies: introducing the artificial neural network

The spectral energy distribution of galaxies is a complex function of the star formation history and geometrical arrangement of stars and gas in galaxies. The computation of the radiative transfer of stellar radiation through the dust distribution is time-consuming. This aspect becomes unacceptable in particular when dealing with the predictions by semi-analytical galaxy formation models populating cosmological volumes, to be then compared with multi-wavelength surveys. Mainly for this aim, we have implemented an artificial neural network algorithm into the spectro-photometric and radiative transfer code GRASIL in order to compute the spectral energy distribution of galaxies in a short computing time. This allows to avoid the adoption of empirical templates that may have nothing to do with the mock galaxies output by models. The ANN has been implemented to compute the dust emission spectrum (the bottleneck of the computation), and separately for the star-forming molecular clouds and the diffuse dust (due to their different properties and dependencies). We have defined the input neurons effectively determining their emission, which means this implementation has a general applicability and is not linked to a particular galaxy formation model. We have trained the net for the disc and spherical geometries, and tested its performance to reproduce the SED of disc and starburst galaxies, as well as for a semi-analytical model for spheroidal galaxies. We have checked that for this model both the SEDs and the galaxy counts in the Herschel bands obtained with the ANN approximation are almost superimposed to the same quantities obtained with the full GRASIL. We conclude that this method appears robust and advantageous, and will present the application to a more complex SAM in another paper.

astro-ph.CO↗