Light In the Dark Ages of the Universe (Lumières dans l’âge sombre de l’Univers – LIDAU

LIDAU is a 3-year project focused on large scale radiative transfer numerical simulations during the Epoch of Reionization (EoR). In the next decade, this period in the history of the universe will be unveiled though observations of the 21 cm emission of neutral hydrogen in the inter-galactic medium. Our goal is to provide predictions for these future observations. The simulated signal will help to put constraints on the design the future radio-interferometers such as SKA in term of bandwidth, sensitivity or resolution. On a shorter time scale, we will be able to confront our models with observations made by LOFAR (beginning by the end of 2008) or MWA. This will help us to constrain such crucial issues as the star formation history in the early universe or the nature of the first sources (Pop III stars or QSO). Later on, observing the 21 cm with SKA, we will be able to probe the growth of cosmic structures by making a full tomography of the signal. Two partners participate in LIDAU (B. Semelin at LERMA, observatoire de Paris and D. Aubert at observatoire de Strasbourg), which will count up to 6 members (2 staffs, 2 post-docs and 2 PhD). To compute de 21 cm emission we need a pipeline of hydrodynamical and radiative transfer simulations. The projects members have validated two state-of-the-art codes, LICORICE and ATON, which use different numerical techniques to handle three-dimensional radiative transfer. The first is rather slow but includes a detailed description of the physics, the second is fast and simple: they are complementary. Both codes are parallelized. LIDAU brings together a large part of the hands-on expertise on radiative transfer in France. Within the project we will run series of simulations through the pipeline in two different box-sizes, 100 and 250 Mpc/h, at two different resolutions, 5123 and 10243. For each simulation we will compute the 21 cm signal taking the influence of the inhomogeneous Lyman-' flux into account (Wouthuysen-Field effect). ATON will explore resolution effect by pushing up to 10243 radiative transfer cells. It will allow us to quantify the influence of sub-grid physics in the 100 Mpc/h box, and to make a statistically significant analysis of the non-gaussianity of the 21 cm signal in the 250 Mpc/h box. With LICORICE we will investigate the influence of several ingredients in the modeling of the physics: the nature and formation history of the sources combined with including Helium, the effect of shock heating (though coupled simulations), and modeling of the Lyman-' pumping (role of Lyman-' and '). The resulting 21 cm emission maps will cover up to 4°x4°, and will reach a level of realism well beyond the current state of the art. To run theses simulations, LICORICE will use the new IBM Power6 computer at IDRIS which offers enough shared memory (256 Gb) on a single computing node to run at 5123 resolution. ATON will be ported on a multi-GPU architecture using a mixed MPI/CUDA implementation. During the last phase of the project, we will make the data cubes available to the community through a web interface compliant with Virtual Observatory standards. Indeed they can be useful for exploitations not directly within the project partner's research interests, such as simulating QSO spectra, modeling the SZ effect or using the 21 cm as a source plane for lensing. To complete this program, the partners mainly need a boost in manpower. Consequently the main part of the requested budget is devoted to 48 man.month for 2 postdoctotal positions. The rest of the grant will be used to buy the GPUs necessary for ATON's simulations, secure storage servers for the project data, and for missions to present the projects results.