Radial selec*on issues for primordial non- Gaussianity detec*on Carlos Cunha NG at KICP, University of Chicago April 20, 2012
Radial issues Decoupled from angular selecbon: One average N(z) for all (simplest possible case). Worry about photo- zs. Wrong redshils. Coupled to angular selecbon
Need redshi8s Spectroscopic or photometric redshils (photo- zs). Photo- zs also require spec- zs for calibrabon. Spec- zs, require photometric pre- selecbon. Spectroscopic Issues Incompleteness Failures (wrong redshils) Sample variance (for photo- z calibrabon).
Radial spec- z issues for surveys Example - Wigglez When spec- zs go wrong, they go wrong bad WiggleZ survey: selection function 13 that ) (43) ns in the case of redshift scatters be. Secondly, the dly, the blunders ution of the denz 1 z 2 of the stant depending ly, if a convolved density cube, it riodic boundary power spectrum f our measured or the real data, ns of galaxy catm and the same rveyregions.we ons of Figures 11 sured the power tput power specations, provided rs. We tested our quations 37 and Figure 13. The power spectrum correction factor due to redshift blunders for each of the survey regions analyzed in this paper, for a redshift range 0.3 <z<0.9. The measured power spectrum must be divided by this factor in order to obtain an unbiased estimate of the true power spectrum. in each grid cell and P (k) isthepowerspectrumamplitude. In each case we used the best-fitting model power spectrum determined below. We evaluated this function at k =0.15 h Mpc 1,althoughthedependenceonscaleisweak.Theeffective redshifts of each slice determined using equation 44 Figure 11. Distribution of values of (1 + z 1 )/(1 + z 2 )forinconsistent repeat redshifts derived from pairs of spectra with quality flags Q =3(redshiftz 1 )andq 4(redshiftz 2 ). The vertical lines indicate the ratios expected in the cases where Hβ, [OIII] and Hα are mis-identified as [OII]. Effect is largest at large scales Blake et al 2010
Radial spec- z issues for Issues: Spectroscopic samples are very incomplete Need to apply spectroscopic selecbon to photometric sample. Sample variance of spec. sample. Area of samples is too small. Spectroscopic failures (wrong redshils). True SSR 1 0.8 0.6 0.4 0.2 surveys 4.5 h exposures 8- m telescope 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Redshift Case study: DES photometry + VVDS- like spec- z s Cunha et al. in prep.
Radial spec- z issues for Issues: Spectroscopic samples very incomplete Need to apply spectroscopic selecbon to photometric sample. Sample variance of spec. sample. Area of samples is too small. Spectroscopic failures (wrong redshils). 0.1 0.08 0.06 0.04 0.02 0 surveys N(z spec ) LSS random 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Redshift 1 deg 2 Case study: DES photometry + VVDS- like spec- z s Cunha, Huterer, Busha & Wechsler 2012
Radial spec- z issues for surveys Issues: Need to apply spectroscopic selecbon to photometric sample. Sample variance of spec. sample. Spectroscopic failures (wrong redshils). z spec 2 1.6 1.2 0.8 0.4 R > 5.0 10 0 10-1 10-2 10-3 10-4 0 0 0.4 0.8 1.2 1.6 2 z true 0 Case study: DES photometry + VVDS- like spec- z s Cunha et al. in prep.
Radial- Angular correla*ons
Photometric surveys for theorists Collect light from galaxies in several broad- band filters in opbcal and near- IR. grizy (DES) + JK (Vista) Use flux in each filter to determine: - - - type:star/gal./qso gal. type: spiral, ellipbcal, (photometric) redshil Also have angular and shape informabon Terminology: magnitude = A log(flux) color = magnitude - magnitude
DES Photometric Calibra*on Deal with: telescope/camera, atmosphere, seasons, Moon, Milky Way. MulBple overlapping Blings with varying orientabons + standard stars + DES: 2 survey Blings/filter/year Need conbguous area that overlaps exisbng surveys. DES Goal: 1% photometry over all survey area (BaO requirement is 2%). Survey Area 5000 sq deg Overlap with SDSS equatorial Stripe 82 for calibrabon (200 sq deg) Connector region Main survey (800 sq deg) region (4000 sq deg) 2 Blings 3 Blings
Photometric calibra*on is complicated Deal with: telescope/camera, atmosphere, seasons, Moon, Milky Way over several years. MulBple overlapping Blings with varying orientabons + standard stars + DES: 2 survey Blings/filter/year DES 5yr mag limits i 24.3 24.2 24.1 24 23.9 23.8 23.7 23.6 col10 Need conbguous area that overlaps exisbng surveys. g 25 24.9 24.8 DES Goal: 1% photometry over all survey area (BaO requirement is 2%). 24.7 24.6 24.5 col7 24.4 24.3
Photometric calibra*on is complicated Mag. limits affect redshil distribubon - > coupling between angular and radial effects (problem is worse if using photo- zs). Varying colors, affect galaxy types being selected. Different types have different HODs, with different biases. variabon in color - > scale- dependent halo bias DES 5yr mag limits i g 24.3 24.2 24.1 24 23.9 23.8 23.7 23.6 25 24.9 24.8 col10 Need to couple radial- angular mask Uncertainty in calibrabon will sbll be a problem. 24.7 24.6 24.5 24.4 col7 24.3
Uncertainty in calibra*on e calib =!N N! 3!m m 10 1 calib error 0.01 w=const 1/2 FoM w0wa N: Number of galaxies m: magnitude Error bars: variabons from alocamng e calib to different m. bias / error 10 0 10-1 10-2 f NL calib error 0.001 10-3 0 2 4 6 8 10 12 14 Multipole l Huterer & friends, in prep.
Conclusions Spectroscopic selecbon is a major challenge for upcoming surveys, parbcularly photometric surveys (because they go deeper). Survey calibrabon on the largest scales is a tough challenge. Lots of work to be done before trustworthy constraints can be extracted from large- scale clustering.
An example: - Template photo- zs. - CalibraBon using one field with 1 deg 2. - Weak Lensing shear- shear tomography. - Difference between true P(z s z p ) and that of calibrabon sample generates biases in cosmology. LSS in one 1deg 2 sample w- bias for fixed ΔP(z s z p )=0.01 ΔP(z s z p ) = P(z s z p ) phot - P(z s z p ) train w- bias for ΔP(z s z p ) of Patch 37
Survey Calculator Number of patches 1000 100 5 10 20 40 80 160 320 640 x 10 3 1/4 deg 2 1/8 deg 2 1/32 deg 2 Magellan VLT σ 95 ( bias ) = 1.0 10 100 1000 10000 gals/patch Cunha, Huterer, Busha & Wechsler arxiv: 1109:5691
DES Photo- zs CombinaBon DES (opbcal)+vista (IR) yields robust photo- zs. LRGs have even bewer scawer. Errors need to be modeled carefully, but f NL requirements weaker than WL. For clusters σ z =0.02. Rough numbers: Δz=0.1 è Δd c =1-2 10 2 h - 1 Mpc over survey redshil range. 100 Mpc 3 deg at z=1.
Star/Galaxy separa*on DistribuBon of stars is not random. Pronounced variabon with labtude. ClassificaBon using colors (magnitudes) BAO requirement: probabilibes accurate to 1% stellar contaminabon and distribubon of misclassified galaxies smaller than 9% over all survey (< 2% on scales < 4 degrees). Good enough for f NL? Main survey region (4000 sq deg) Milky Way Overlap with SDSS equatorial Stripe 82 for calibrabon (200 sq deg) Connector region (800 sq deg)
The Dark Energy Survey Study Dark Energy using 4 complementary techniques: I. Cluster Counts II. Weak Lensing III. Baryon AcousBc OscillaBons IV. Supernovae www.darkenergysurvey.org Blanco 4-meter at CTIO Two mulbband surveys: Main: 5000 deg 2 5 (h - 1 Gpc) 3 300 million galaxies g, r, i, z, Y to 24th mag SNe: 15 deg 2 repeat Build new 3 deg 2 FoV camera and Data management sytem in Blanco 4- m telescope Survey 2012-2017 (525 nights) Camera available for community use the rest of the Bme (70%)
Observa*onal issues for f nl measurement ArBficial correlabons can mimic f nl. For >100 Mpc (several degrees) are crucial. f local NL, separabons ArBficial correlabons can be due to: photometric calibrabon photometric redshils star/galaxy separabon More relevant for galaxies than clusters Because of 1/k 2 scale dependence of bias b(k) = b G + f NL const k 2 Clusters have own selecbon issues