Search

We test whether Lyman alpha emitters (LAEs) and Lyman-break galaxies (LBGs) can be good tracers of high-z large-scale structures, using the Horizon Run 5 cosmological hydrodynamical simulation. We identify LAEs using the Ly{\alpha} emission line luminosity and its equivalent width, and LBGs using the broad-band magnitudes at z~2.4, 3.1, and 4.5. We first compare the spatial distributions of LAEs, LBGs, all galaxies, and dark matter around the filamentary structures defined by dark matter. The comparison shows that both LAEs and LBGs are more concentrated toward the dark matter filaments than dark matter. We also find an empirical fitting formula for the vertical density profile of filaments as a binomial power-law relation of the distance to the filaments. We then compare the spatial distributions of the samples around the filaments defined by themselves. LAEs and LBGs are again more concentrated toward their filaments than dark matter. We also find the overall consistency between filamentary structures defined by LAEs, LBGs, and dark matter, with the median spatial offsets that are smaller than the mean separation of the sample. These results support the idea that the LAEs and LBGs could be good tracers of large-scale structures of dark matter at high redshifts., Comment: 20 pages, 15 figures, 2 tables, accepted for publication in ApJ

We describe the survey design and science goals for ODIN (One-hundred-deg^2 DECam Imaging in Narrowbands), a NOIRLab survey using the Dark Energy Camera (DECam) to obtain deep (AB~25.7) narrow-band images over an unprecedented area of sky. The three custom-built narrow-band filters, N419, N501, and N673, have central wavelengths of 419, 501, and 673 nm and respective full-widthat-half-maxima of 7.2, 7.4, and 9.8 nm, corresponding to Lya at z=2.4, 3.1, and 4.5 and cosmic times of 2.8, 2.1, and 1.4 Gyr, respectively. When combined with even deeper, public broad-band data from Hyper Suprime-Cam, DECam, and in the future, LSST, the ODIN narrow-band images will enable the selection of over 100,000 Lya-emitting (LAE) galaxies at these epochs. ODIN-selected LAEs will identify protoclusters as galaxy overdensities, and the deep narrow-band images enable detection of highly extended Lya blobs (LABs). Primary science goals include measuring the clustering strength and dark matter halo connection of LAEs, LABs, and protoclusters, and their respective relationship to filaments in the cosmic web. The three epochs allow the redshift evolution of these properties to be determined during the period known as Cosmic Noon, where star formation was at its peak. The two narrow-band filter wavelengths are designed to enable interloper rejection and further scientific studies by revealing [O II] and [O III] at z=0.34, Lya and He II 1640 at z=3.1, and Lyman continuum plus Lya at z=4.5. Ancillary science includes similar studies of the lower-redshift emission-line galaxy samples and investigations of nearby star-forming galaxies resolved into numerous [O III] and [S II] emitting regions., Comment: 21 pages, 8 figures, submitted to Astrophysical Journal

The 21-cm signal from neutral hydrogen in the early universe will provide unprecedented information about the first stars and galaxies. Extracting this information, however, requires accounting for many unknown astrophysical processes. Semi-numerical simulations are key for exploring the vast parameter space of said processes. These simulations use approximate techniques such as excursion-set and perturbation theory to model the 3D evolution of the intergalactic medium, at a fraction of the computational cost of hydrodynamic and/or radiative transfer simulations. However, exploring the enormous parameter space of the first galaxies can still be computationally expensive. Here we introduce 21cmfish, a Fisher-matrix wrapper for the semi-numerical simulation 21cmFAST. 21cmfish facilitates efficient parameter forecasts, scaling to significantly higher dimensionalities than MCMC approaches, assuming a multi-variate Gaussian posterior. Our method produces comparable parameter uncertainty forecasts to previous MCMC analyses but requires ~10$^4$x fewer simulations. This enables a rapid way to prototype analyses adding new physics and/or additional parameters. We carry out a forecast for HERA using the largest astrophysical parameter space to-date, with 10 free parameters, spanning both population II and III star formation. We find X-ray parameters for the first galaxies could be measured to sub-percent precision, and, though they are highly degenerate, the stellar-to-halo mass relation and ionizing photon escape fraction for population II and III galaxies can be constrained to ~10% precision (logarithmic quantities). Using a principal component analysis we find HERA is most sensitive to the product of the ionizing escape fraction and the stellar-to-halo mass fraction for population II galaxies., Comment: 18 pages, accepted for publication in MNRAS. Fisher-matrix wrapper for 21cmFAST available here: https://github.com/charlottenosam/21cmfish

The excursion set reionization framework is widely used, due to its speed and accuracy in reproducing the 3D topology of reionization. However, it is known that it does not conserve photon number. Here, we introduce an efficient, on-the-fly recipe to approximately account for photon conservation. Using a flexible galaxy model shown to reproduce current high-$z$ observables, we quantify the bias in the inferred reionization history and galaxy properties resulting from the non-conservation of ionizing photons. Using a mock 21-cm observation, we perform inference with and without correcting for ionizing photon conservation. We find that ignoring photon conservation results in very modest biases in the inferred galaxy properties, for our fiducial model. The notable exception is in the power-law scaling of the ionizing escape fraction with halo mass, which can be biased from the true value by $\sim2.4\sigma$ (corresponding to $\sim0.2$ in the power-law index). Our scheme is implemented in the public code ${\tt 21cmFAST}$., Comment: 9 pages, 4 figures, 1 table, accepted for publication in MNRAS

This brief code paper presents a new Python-wrapped version of the popular 21cm cosmology simulator, 21cmFAST. The new version, v3+, maintains the same core functionality of previous versions of 21cmFAST, but features a simple and intuitive interface, and a great deal more flexibility. This evolution represents the work of a formalized collaboration, and the new version, available publicly on GitHub, provides a single point-of-reference for all future upgrades and community-added features. In this paper, we describe simple usage of 21cmFAST, some of its new features, and provide a simple performance benchmark., Comment: 6 pages, 3 figures, published in JOSS

We introduce Mem2Mem, a memory-to-memory mechanism for hierarchical recurrent neural network based encoder decoder architectures and we explore its use for abstractive document summarization. Mem2Mem transfers "memories" via readable/writable external memory modules that augment both the encoder and decoder. Our memory regularization compresses an encoded input article into a more compact set of sentence representations. Most importantly, the memory compression step performs implicit extraction without labels, sidestepping issues with suboptimal ground-truth data and exposure bias of hybrid extractive-abstractive summarization techniques. By allowing the decoder to read/write over the encoded input memory, the model learns to read salient information about the input article while keeping track of what has been generated. Our Mem2Mem approach yields results that are competitive with state of the art transformer based summarization methods, but with 16 times fewer parameters

The first generation of galaxies is expected to form in minihalos, accreting gas through ${\rm H}_2$ cooling, and possessing unique properties. Although unlikely to be directly detected in UV/infrared surveys, the radiation from these molecular-cooling galaxies (MCGs) could leave an imprint in the 21-cm signal from the Cosmic Dawn. Here we quantify their detectability with upcoming radio interferometers. We generate mock 21-cm power spectra using a model for both MCGs as well as more massive, atomic-cooling galaxies (AGCs), allowing both populations to have different properties and scaling relations. The galaxy parameters are chosen so as to be consistent with: (i) high-redshift UV luminosity functions; (ii) the upper limit on the neutral fraction from QSO spectra; (iii) the Thomson scattering optical depth to the CMB; and (iv) the timing of the recent putative EDGES detection. The latter implies a significant contribution of MCGs to the Cosmic Dawn, if confirmed to be cosmological. We then perform Bayesian inference on two models including and ignoring MCG contributions. Comparing their Bayesian evidences, we find a strong preference for the model including MCGs, despite the fact that it has more free parameters. This suggests that if MCGs indeed play a significant role in the Cosmic Dawn, it should be possible to infer their properties from upcoming 21-cm power spectra. Our study illustrates how these observations can discriminate among uncertain galaxy formation models with varying complexities, by maximizing the Bayesian evidence., Comment: 12 pages, 3 figures, submitted to MNRAS, comments are welcome

The Epoch of Reionization (EoR) depends on the complex astrophysics governing the birth and evolution of the first galaxies and structures in the intergalactic medium. EoR models rely on cosmic microwave background (CMB) observations, and in particular the large-scale E-mode polarization power spectra (EE PS), to help constrain their highly uncertain parameters. However, rather than directly forward-modelling the EE PS, most EoR models are constrained using a summary statistic -- the Thompson scattering optical depth, $\tau_e$. Compressing CMB observations to $\tau_e$ requires adopting a basis set for the EoR history. The common choice is the unphysical, redshift-symmetric hyperbolic tangent (Tanh) function, which differs in shape from physical EoR models based on hierarchical structure formation. Combining public EoR and CMB codes, 21cmFAST and CLASS, here we quantify how inference using the $\tau_e$ summary statistic impacts the resulting constraints on galaxy properties and EoR histories. Using the last Planck 2018 data release, we show that the marginalized constraints on the EoR history are more sensitive to the choice of the basis set (Tanh vs physical model) than to the CMB likelihood statistic ($\tau_e$ vs PS). For example, EoR histories implied by the growth of structure show a small tail of partial reionization extending to higher redshifts. However, biases in inference using $\tau_e$ are negligible for the Planck 2018 data. Using EoR constraints from high-redshift observations including the quasar dark fraction, galaxy UV luminosity functions and CMB EE PS, our physical model recovers $\tau_e=0.0569^{+0.0081}_{-0.0066}$., Comment: 11 pages, 5 figures, Updated to match the published version. Added Gamma evolution; Only minor changes

The very first galaxies that started the cosmic dawn likely resided in so-called "minihaloes", with masses of $\sim10^5$-$10^8\mathrm{M}_\odot$, accreting their gas from the intergalactic medium through H$_2$ cooling. Such molecularly cooled galaxies (MCGs) mostly formed in pristine environments, hosted massive, metal-free stars, and were eventually sterilized by the build-up of a disassociating (Lyman-Werner; LW) background. Therefore, their properties might be very different from the galaxies we see in the later Universe. Although MCGs are probably too faint to be observed directly, we could nevertheless infer their properties from the imprint they leave in the cosmic 21-cm signal. Here we quantify this imprint by extending the public simulation code 21cmFAST to allow for a distinct population of MCGs. We allow MCGs to have different properties from other galaxies, including unique scaling relations for their stellar-to-halo mass ratios, ionizing escape fractions, and spectral energy distributions. We track inhomogeneous recombinations, disassociative LW feedback, and photoheating from reionization. After demonstrating how MCGs can shape the 21-cm signal, we explore to what extent current observations can already place constraints on their properties. The cosmic microwave background optical depth from Planck sets an upper limit on the product of the ionizing escape fraction and the stellar mass in MCGs. When including also the timing of the putative EDGES absorption signal, we find an additional strong degeneracy between the stellar mass and the X-ray luminosity of MCGs. If proven to be of cosmic origin, the timing of the EDGES signal would have been set by MCGs., Comment: 19 pages, 8 figures, 1 table; Updated to match the published version. Only minor changes

In this work, we investigate the performance of untrained randomly initialized encoders in a general class of sequence to sequence models and compare their performance with that of fully-trained encoders on the task of abstractive summarization. We hypothesize that random projections of an input text have enough representational power to encode the hierarchical structure of sentences and semantics of documents. Using a trained decoder to produce abstractive text summaries, we empirically demonstrate that architectures with untrained randomly initialized encoders perform competitively with respect to the equivalent architectures with fully-trained encoders. We further find that the capacity of the encoder not only improves overall model generalization but also closes the performance gap between untrained randomly initialized and full-trained encoders. To our knowledge, it is the first time that general sequence to sequence models with attention are assessed for trained and randomly projected representations on abstractive summarization., Comment: Accepted to ACL 2019 SRW. First two authors contributed equally

Next generation observatories will enable us to study the first billion years of our Universe in unprecedented detail. Foremost among these are 21-cm interferometry with the HERA and the SKA, and high-$z$ galaxy observations with the James Webb Space Telescope (JWST). Taking a basic galaxy model, in which we allow the star formation rates and ionizing escape fractions to have a power-law dependence on halo mass with an exponential turnover below some threshold, we quantify how observations from these instruments can be used to constrain the astrophysics of high-$z$ galaxies. For this purpose, we generate mock JWST LFs, based on two different hydrodynamical cosmological simulations; these have intrinsic luminosity functions (LFs) which turn over at different scales and yet are fully consistent with present-day observations. We also generate mock 21-cm power spectrum observations, using 1000h observations with SKA1 and a moderate foreground model. Using only JWST data, we predict up to a factor of 2-3 improvement (compared with HST) in the fractional uncertainty of the star formation rate to halo mass relation and the scales at which the LFs peak (i.e. turnover). Most parameters regulating the UV galaxy properties can be constrained at the level of $\sim 10$% or better, if either (i) we are able to better characterize systematic lensing uncertainties than currently possible; or (ii) the intrinsic LFs peak at magnitudes brighter than $M_{\rm UV} \lesssim -13$. Otherwise, improvement over HST-based inference is modest. When combining with upcoming 21-cm observations, we are able to significantly mitigate degeneracies, and constrain all of our astrophysical parameters, even for our most pessimistic assumptions about upcoming JWST LFs. The 21-cm observations also result in an order of magnitude improvement in constraints on the EoR history., Comment: 10 pages, 6 figures, 1 table, submitted to MNRAS

Galaxy formation during the first billion years of our Universe remains a challenging problem at the forefront of astrophysical cosmology. Although these $z \geq 6$ galaxies are likely responsible for the last major phase change of our Universe, the epoch of reionization (EoR), detailed studies are possible only for relatively rare, bright objects. Characterizing the fainter galaxies which are more representative of the population as a whole is currently done mainly through their non-ionizing UV luminosity function (LF). Observing the faint end of the UV LFs is nevertheless challenging, and current estimates can differ by orders of magnitude. Here we propose a methodology to combine disparate high-$z$ UV LF data sets in a Bayesian framework: Bayesian Data Averaging (BDA). Using a flexible, physically-motivated galaxy model, we compute the relative evidence of various $z=6$ UV LFs within the magnitude range $-20 \leq M_{\rm UV} \leq -15$ which is common to the data sets. Our model, based primarily on power-law scalings of the halo mass function, naturally penalizes systematically jagged data points as well as mis-estimated errors. We then use the relative evidence to weigh the posteriors obtained from disparate LF observations during the EoR, $6 \leq z \leq 10$. The resulting LFs suggest that the star formation rate density (SFRD) integrated down to a UV magnitude of -17 represent $60.9^{+11.3}_{-9.6}\%$ / $28.2^{+9.3}_{-10.1}\%$ / $5.7^{+4.5}_{-4.7}\%$ of the total SFRD at redshifts 6 / 10 / 15. The BDA framework we introduce enables galaxy models to leverage multiple, analogous observational data sets.

Magnetic Vortex Acceleration (MVA) from near critical density targets is one of the promising schemes of laser-driven ion acceleration. 3D particle-in-cell simulations are used to explore a more extensive laser-target parameter space than previously reported on in the literature as well as to study the laser pulse coupling to the target, the structure of the fields, and the properties of the accelerated ion beam in the MVA scheme. The efficiency of acceleration depends on the coupling of the laser energy to the self-generated channel in the target. The accelerated proton beams demonstrate high level of collimation with achromatic angular divergence, and carry a significant amount of charge. For PW-class lasers, this acceleration regime provides favorable scaling of maximum ion energy with laser power for optimized interaction parameters. The mega Tesla-level magnetic fields generated by the laser-driven co-axial plasma structure in the target are prerequisite for accelerating protons to the energy of several hundred MeV., Comment: submitted to Physics of Plasmas

The properties of the first galaxies, expected to drive the Cosmic Dawn (CD) and the Epoch of Reionization (EoR), are encoded in the 3D structure of the cosmic 21-cm signal. Parameter inference from upcoming 21-cm observations promises to revolutionize our understanding of these unseen galaxies. However, prior inference was done using models with several simplifying assumptions. Here we introduce a flexible, physically-motivated parametrization for high-$z$ galaxy properties, implementing it in the public code 21cmFAST. In particular, we allow their star formation rates and ionizing escape fraction to scale with the masses of their host dark matter halos, and directly compute inhomogeneous, sub-grid recombinations in the intergalactic medium. Combining current Hubble observations of the rest-frame UV luminosity function (UV LFs) at high-$z$ with a mock 1000h 21-cm observation using the Hydrogen Epoch of Reionization Arrays (HERA), we constrain the parameters of our model using a Monte Carlo Markov Chain sampler of 3D simulations, 21CMMC. We show that the amplitude and scaling of the stellar mass with halo mass is strongly constrained by LF observations, while the remaining galaxy properties are constrained mainly by 21-cm observations. The two data sets compliment each other quite well, mitigating degeneracies intrinsic to each observation. All eight of our astrophysical parameters are able to be constrained at the level of $\sim 10\%$ or better. The updated versions of 21cmFAST and 21CMMC used in this work are publicly available., Comment: 16 pages, 8 figures and 2 tables. Associated movies are available at http://homepage.sns.it/mesinger/21CMMC.html. Updated to match the published version. All results and conclusions remain unchanged

Recent works have shown that synthetic parallel data automatically generated by translation models can be effective for various neural machine translation (NMT) issues. In this study, we build NMT systems using only synthetic parallel data. As an efficient alternative to real parallel data, we also present a new type of synthetic parallel corpus. The proposed pseudo parallel data are distinct from previous works in that ground truth and synthetic examples are mixed on both sides of sentence pairs. Experiments on Czech-German and French-German translations demonstrate the efficacy of the proposed pseudo parallel corpus, which shows not only enhanced results for bidirectional translation tasks but also substantial improvement with the aid of a ground truth real parallel corpus.

We investigate the clustering properties of Lyman-break galaxies (LBGs) at $z\sim6$ - $8$. Using the semi-analytical model {\scshape Meraxes} constructed as part of the Dark-ages Reionization And Galaxy-formation Observables from Numerical Simulation (DRAGONS) project, we predict the angular correlation function (ACF) of LBGs at $z\sim6$ - $8$. Overall, we find that the predicted ACFs are in good agreement with recent measurements at $z\sim 6$ and $z\sim 7.2$ from observations consisting of the Hubble eXtreme Deep Field (XDF), the Hubble Ultra-Deep Field (HUDF) and Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS) field. We confirm the dependence of clustering on luminosity, with more massive dark matter haloes hosting brighter galaxies, remains valid at high redshift. The predicted galaxy bias at fixed luminosity is found to increase with redshift, in agreement with observations. We find that LBGs of magnitude $M_{{\rm AB(1600)}} < -19.4$ at $6\lesssim z \lesssim 8$ reside in dark matter haloes of mean mass $\sim 10^{11.0}$- $10^{11.5} M_{\rm \odot}$, and this dark matter halo mass does not evolve significantly during reionisation., Comment: 14 pages, 10 figures, published in MNRAS

The formation and evolution of galaxies with low neutral atomic hydrogen (HI) masses, M$_{\rm HI}$$<$10$^{8}h^{-2}$M$_{\odot}$, are affected by host dark matter halo mass and photoionisation feedback from the UV background after the end of reionization. We study how the physical processes governing the formation of galaxies with low HI mass are imprinted on the distribution of neutral hydrogen in the Universe using the hierarchical galaxy formation model, GALFORM. We calculate the effect on the correlation function of changing the HI mass detection threshold at redshifts $0 \le z \le 0.5$. We parameterize the clustering as $\xi(r)=(r/r_{0})^{-\gamma}$ and we find that including galaxies with M$_{\rm HI}$$<$10$^{8}h^{-2}$M$_{\odot}$ increases the clustering amplitude $r_{0}$ and slope $\gamma$ compared to samples of higher HI masses. This is due to these galaxies with low HI masses typically being hosted by haloes with masses greater than 10$^{12}{h}^{-1}$M$_{\odot}$, and is in contrast to optically selected surveys for which the inclusion of faint, blue galaxies lowers the clustering amplitude. We show the HI mass function for different host dark matter halo masses and galaxy types (central or satellite) to interpret the values of $r_{0}$ and $\gamma$ of the clustering of HI-selected galaxies. We also predict the contribution of low HI mass galaxies to the 21cm intensity mapping signal. We calculate that a dark matter halo mass resolution better than $\sim$10$^{10}{h}^{-1}$M$_{\odot}$ at redshifts higher than 0.5 is required in order to predict converged 21cm brightness temperature fluctuations., Comment: 14 pages, 10 figures, accepted for pubication in MNRAS

The increasing complexity of deep neural networks (DNNs) has made it challenging to exploit existing large-scale data processing pipelines for handling massive data and parameters involved in DNN training. Distributed computing platforms and GPGPU-based acceleration provide a mainstream solution to this computational challenge. In this paper, we propose DeepSpark, a distributed and parallel deep learning framework that exploits Apache Spark on commodity clusters. To support parallel operations, DeepSpark automatically distributes workloads and parameters to Caffe/Tensorflow-running nodes using Spark, and iteratively aggregates training results by a novel lock-free asynchronous variant of the popular elastic averaging stochastic gradient descent based update scheme, effectively complementing the synchronized processing capabilities of Spark. DeepSpark is an on-going project, and the current release is available at http://deepspark.snu.ac.kr.

Redshifted 21cm measurements of the structure of ionised regions that grow during reionization promise to provide a new probe of early galaxy and structure formation. One of the challenges of modelling reionization is to account both for the sub-halo scale physics of galaxy formation and the regions of ionization on scales that are many orders of magnitude larger. To bridge this gap we first calculate the statistical relationship between ionizing luminosity and Mpc-scale overdensity using detailed models of galaxy formation computed using relatively small volume - ($\sim$100Mpc/$h$)$^{3}$, high resolution dark matter simulations. We then use a Monte-Carlo technique to apply this relationship to reionization of the intergalactic medium within large volume dark matter simulations - ($>$1Gpc/$h$)$^{3}$. The resulting simulations can be used to address the contribution of very large scale clustering of galaxies to the structure of reionization, and show that volumes larger than 500Mpc/$h$ are required to probe the largest reionization features mid-way through reionization. As an example application of our technique, we demonstrate that the predicted 21cm power spectrum amplitude and gradient could be used to determine the importance of supernovae feedback for early galaxy formation., Comment: 15 pages, 13 figures, Accepted by MNRAS

We investigate the clustering of Lyman-break galaxies (LBGs) at $z\sim4$. Using the hierarchical galaxy formation model GALFORM, we predict, for the first time using a semi-analytical model with feedback from active galactic nuclei (AGN), the angular correlation function (ACF) of LBGs and find agreement within $3\,\sigma$ with new measurements of the ACF from surveys including the Hubble eXtreme Deep Field (XDF) and CANDELS field. Our simulations confirm the conclusion reached using independent models that although the predicted ACFs reproduce the trend of increased clustering with luminosity, the dependence is less strong than observed. We find that for the detection limits of the XDF field central LBGs at $z\sim 4$ predominantly reside in haloes of mass $\sim 10^{11}-10^{12}h^{-1}M_{\rm \odot}$ and that satellites reside in larger haloes of mass $\sim 10^{12}-10^{13}h^{-1}M_{\rm \odot}$. The model predicts fewer bright satellite LBGs at $z\sim4$ than is inferred from measurements of the ACF at small scales. By analysing the halo occupation distribution (HOD) predicted by the model, we find evidence that AGN feedback affects the HOD of central LBGs in massive haloes. This is a new high-redshift test of this important feedback mechanism. We investigate the effect of photometric errors in the observations on the ACF predictions. We find that the observational uncertainty in the galaxy luminosity reduces the clustering amplitude and that this effect increases towards faint galaxies, particularly on small scales. To compare properties of model with observed LBGs this uncertainty must be considered., Comment: 16 pages, 13 figures, accepted for publication in MNRAS

We explore the galaxy formation physics governing the low mass end of the HI mass function in the local Universe. Specifically, we predict the effects on the HI mass function of varying i) the strength of photoionisation feedback and the redshift of the end of the epoch of reionization, ii) the cosmology, iii) the supernovae feedback prescription, and iv) the efficiency of star formation. We find that the shape of the low-mass end of the HI mass function is most affected by the critical halo mass below which galaxy formation is suppressed by photoionisation heating of the intergalactic medium. We model the redshift dependence of this critical dark matter halo mass by requiring a match to the low-mass end of the HI mass function. The best fitting critical dark matter halo mass decreases as redshift increases in this model, corresponding to a circular velocity of $\sim 50 \, {\rm km \,s}^{-1}$ at $z=0$, $\sim 30 \, {\rm km\, s}^{-1}$ at $z \sim 1$ and $\sim 12 \, {\rm km \, s}^{-1}$ at $z=6$. We find that an evolving critical halo mass is required to explain both the shape and abundance of galaxies in the HI mass function below $M_{\rm HI} \sim 10^{8} h^{-2} {\rm M_{\odot}}$. The model makes specific predictions for the clustering strength of HI-selected galaxies with HI masses > $10^{6} h^{-2} {\rm M_{\odot}}$ and $> 10^{7} h^{-2} {\rm M_{\odot}}$ and for the relation between the HI and stellar mass contents of galaxies which will be testable with upcoming surveys with the Square Kilometre Array and its pathfinders. We conclude that measurements of the HI mass function at $z \ge 0$ will lead to an improvement in our understanding of the net effect of photoionisation feedback on galaxy formation and evolution., Comment: 13 pages, 9 figures, Minor corrections added, revised to match version accepted by MNRAS

We study diffusive shock acceleration (DSA) of protons and electrons at nonrelativistic, high Mach number, quasiparallel, collisionless shocks by means of self-consistent 1D particle-in-cell simulations. For the first time, both species are found to develop power-law distributions with the universal spectral index $-4$ in momentum space, in agreement with the prediction of DSA. We find that scattering of both protons and electrons is mediated by right-handed circularly polarized waves excited by the current of energetic protons via non-resonant hybrid (Bell) instability. Protons are injected into DSA after a few gyro-cycles of shock drift acceleration (SDA), while electrons are first pre-heated via SDA, then energized via a hybrid acceleration process that involves both SDA and Fermi-like acceleration mediated by Bell waves, before eventual injection into DSA. Using the simulations we can measure the electron/proton ratio in accelerated particles, which is of paramount importance for explaining the cosmic ray fluxes measured at Earth and the multi-wavelength emission of astrophysical objects such as supernova remnants, radio supernovae, and galaxy clusters. We find the normalization of electron power-law is $\lesssim 10^{-2}$ that of the protons for strong nonrelativistic shocks., Comment: 5 pages, 4 figures, accepted to PRL (some updates)