Instrument simulations for the EUCLID space mission
Résumé
The EUCLID space mission is an ambitious project for measuring the statistical properties of the universe with an error bellow 1% up to redshift 2, from a survey of 1 billion of galaxies (100 thousand with a spectra). Because of the size of the mission, the flow down from the scientific requirements to the technical needs of the telescope, the instruments and the ground infrastructures, is extremely complex. To check the science requirements at every step of the mission we need some robust simulation tools. In the framework of EUCLID simulations, the CPPM has took in charge the prototype of the simulator of the NISP spectrometer. We have provided to the community a prototype called TIPS (TIPS Is a Pixel Simulator). TIPS is a simulation tool based on aXeSIM (Kuemmel et al., 2007 and 2009) which produces the expected raw images of the NISP focal plan. The talk will focus on two use cases of TIPS, the prediction of the mission performances and the global analysis of the mission data. The purpose of the performance analysis is to check if a given mission implementation allows to achieve the mission requirements. The ultimate requirements that must be validated are the science goals but in our case, for cosmology, the simulation size needed is typically 10000deg². We will not be able to run the full end to end simulation of the mission every time something is modified in the telescope. That's why we derived the performances in two sub-levels. The instrument performances measure the capabilities of the instrument alone and could be done with the simulation of one field of view (0.5deg²). The survey performances estimate the efficiency of the redshift measurement. The typical scale of simulation in this case is 10deg². TIPS is already used in production to measure the instrument performances of the NISP spectrometer. The method will be illustrated with our results on the impact of the PSF distribution (Jullo et al., 2012) and the cosmic rays (Ealet et al., 2013). The second use case of TIPS is the most challenging. We want to use the simulation to fit the reality in comparing the simulated images with the real ones. As in the performance analysis, we split this use case into three levels. At the instrument level, we would like to measure the calibration. Using a known source (calibration LED, standard stars...), we could iterate on the instrument model parameters and simulate the calibration data up to be able to reproduce the real images. At the survey level, we would like to measure the sources. This time using a fixed instrument model, we could reconstruct the source spectra and/or adjust some parameters like the redshift. At the cosmology level, we could imagine applying the same principle to fit the cosmological parameters but, at this time, it will not be feasible because of the huge computational needs that it implies. The method w