Euclid Euclid OU-SIR / OU-SPE Simulation needs P. Franzetti & B. Garilli
Euclid Euclid Why simulated data Each OU-SIR/OU-SPE WPs requires • testing of existing tools to Euclid data • and/or adaptation of existing tools to Euclid data • and/or development of new ‘classical’ tools • and/or development of innovative tools How much can we re-use? How much do we need to develop from scratch? ONLY simulated data can answer to THESE questions
Euclid Euclid TIMELINE Timeline (from SIP) June 2014: delivery of pipeline Bread Board Model (2 years from now) November 2015: delivery of pipeline Demonstration Model Given the unknown amount of development required time is not overabundant, work on algorithms has to start NOW ! Even if the input of each WP is essentially the output of the previous WP, WPs cannot start their work sequentially they have to proceed in parallel , Simulated data are needed NOW
Euclid Euclid OU-SIR Relevant WPs • Basic Processing • Spectra Identification • Background Subtraction • De-contamination • Calibrations • Extraction A survey of needs has been made c/o WP managers The input of each WP is essentially the output of the previous WP, but WPs cannot start their work sequentially, they have to proceed in parallel, so globally the simulation needs are the sum of all inputs
Euclid Euclid OU-SIR Required data RED: NOW GREEN: SPRING 2013 2D “RAW DATA” N pointings with simulated data for all the detectors and dithers (i.e. 16*4 images per pointing) • 0 th and 1 st orders spectra • higher orders spectra (2 nd , -1 st … ) with various contributions (that can be added or excluded) • Signal noise • Sky background • Dark • Flat-field • Bad pixels and detector cosmetics effects • Cosmic rays
Euclid Euclid OU-SIR Required data RED: NOW GREEN: SPRING 2013 CALIBRATIONS • Instrument Model • Optics • Distorsions • Dispersion relation • FPA description • Detectors characteristic (RON, QE) PSF • • Dark images • Flat images • Calibration fields (as 2D raw data)
Euclid Euclid OU-SIR Required data RED: NOW GREEN: SPRING 2013 OBJECTS CATALOG • Input catalog • alpha,delta • redshift • spectral type • lines fluxes and ew • shape parameters (rdisk, rbulge, axratio,pa) • VIS,J,H,K magnitudes • “Processed” catalogs J, H, K, J+H+K (+ direct images) • “real” detection procedure • grism re-positioning errors • OU-MER data
Euclid Euclid OU-SIR Data Size RED: NOW GREEN: SPRING 2013 9 pointings, i.e. a contiguous area of ~ 4.4 sq deg (3x3 pointings) 25 sq degrees
Euclid Euclid OU-SPE Relevant WPs • Spectra Combination • Lines Identification and Redshift Measurement • Spectral Features Measurement • Redshift Quality Determination • Rest-frame Parameters • Spectro-photometric Classification • A survey of needs has been made c/o WP managers The input of each WP is essentially the output of the previous WP, but WPs cannot start their work sequentially, they have to proceed in parallel, so globally the simulation needs are the sum of all inputs
Euclid Euclid OU-SPE Required data Data kind Notes observed 1D spectra contamination 1/grism/roll free observed 2D spectra 1/grism/roll observed 1D spectra, contamination Combined all free grisms all rolls observed 2D spectra Combined all grisms all rolls error spectrum Contamination mask source infos as from Line flux, simulations redshift, mag, input catalog galaxy type….. observed uncontaminated Can be replaced spectrum by incident spectrum photometric redshift Or photoz uncertainty and % of catastrophs per redshift bin
Euclid Euclid OU-SPE size required and time scales Algorithms have to be tested with meaningful statistics (statistical uncertainties less than 5%) i.e. ~ 1000 galaxies For each • redshift bin : e.g. z bin of width 0.1 between 0.7 and 2, and 0.25 below and above (up to z=4) i.e. ~20 bins • galaxy type: emitter/non emitter i.e. 2 bins • line fluxes bin: e.g. <1,1-3,3-4,4-6,6-8,8-12,>12 e-16 erg/cm2/ s i.e. 12 bins 20x2x12~500 bins x1000 galaxies per bin= ~500k spectra Assuming a galaxy density of the order of 20k gals per sq.deg Simulations of ~ 25 sq. deg . are required 1/10 of it (2.5 sq.deg.) is enough to begin work. Larger areas will be needed in 6/9 months time
Euclid Euclid Additional requirements What listed before applies to the Wide survey ONLY (products, sizes etc) Something similar (same products) should also be made available for the DEEP survey, rescaling areas accordingly (1 full square degree), on a slightly longer time scale (end of 2013) Also to be considered in the near future are special needs for rare objects EXAMPLE: Surface density of QSOs or primeval galaxies is very low. In order to study what Euclid can do, we need dedicated simulations in which the surface densities of these objects is artificially increased by some order of magnitude, so that meaningful statistics can be collected with a moderately small simulation This will be the subject of more detailed requirements in the near future (if OU-SIM has dead lines for this topic, please say so now)
Euclid Euclid Note about EMA, DataModel etc OU’s in general, and OU-SIR/SPE in particular, does NOT need a working EMA or an established Data model to start algorithms development and testing Nor OUs should worry about EMA etc, that is a task for SDCs Instead OU-SIR/SPE (all OU’s?) NEED simulated data NOW A provisional way to generate and distribute simulated data should be put in place
Euclid Euclid Who? OU-SIR/OU-SPE does NOT have a dedicated WP to provide simulated data The NISP Science Simulation group cannot provide 25 sq. deg • Not large enough input catalog (not even for 2.5 sq. deg) • Not enough computer power • No disk storage • No manpower to produce the required data products (2 people in total, also with other responsibilities)
Euclid Euclid Who? Also OU-SIM does NOT have dedicated WPs to provide some of these data; just one WP for pixel simulation New WPs ? • pixel level simulation • reduction level simulation • final spectra simulation
Recommend
More recommend
