1. BigBOSS Survey and Spectral Simulations Nick Mostek

2. 2 0.2<z<1: Luminous Red Galaxies (extended from BOSS footprint) Spectroscopic Targets 2<z<3.5: Ly  forest from QSOs (pioneered from BOSS) 0.7<z<2: Emission line galaxies BigBOSS Padmanabhan, 2004 REALLY?

3. 3 z=1.75 for [OII] Extreme Silicon (Bebek)

4. 4 BigBOSS will have 5000 fibers spread over a 4.9 sq. deg. field of view. With ~1000 fibers/deg 2, we will need to split up the ELG sample into low and high redshift bins. Lower redshift bins will make measurements in half the time as the z>1.5 galaxies and be re-assigned to new objects Fibers can be repositioned in less than 2 minutes, or less than that of the LBNL CCD read time at 20kHz. LRGs and QSOs require longer exposures as they need to measure the continuum flux levels If we change our upper redshift limit, can we: –push the low redshift bin to higher densities? –survey faster? White Paper Survey Yield ?

5. 5 Keep LRG and QSO number densities –LRGs require at least 30 minute exposures –QSOs require the entire exposure time of a fiber ELG redshift distribution is defined by minimum desired [OII] flux ELG minimum exposure time must scale flux according to previous exposure time estimates (S/N = 8 in 30 minutes at z=2) Maximum exposure time on the field must be an integer number of minimum exposure times (coherent readout of all spectrographs) Number of available exposures in fiber positioners that of a given diameter that can fit into a field of a given size –Used 15mm diameter positioners in a 2.8 deg diameter field in this study (6400 fibers total, 20% area overlap) Survey Constraints

6. 6 grz-selected Redshift Distrubtion

7. 7 Used “Maximal Pack” of 2.5’ (12mm diameter, 20% overlap) diameter circular positioners into a 2.8 deg diameter field Distribution of emission line galaxies with [OII] > 8.5E-17 and 75% completeness in grz Uniform spatial distribution is currently used (no clustering) Focal Plane Fiber Map Black Square: Fiber Center Red Triangle: QSO Lt. Green Star: LRG Orange Diamond: Low-z ELG

8. 8 Red line corresponds to a constant volume density of 3.4E-4 (Mpc/h)^-3 grz selection needs improvement for z>1.2 galaxies Measured z Distributions

9. 9 Numbers correspond to [OII] flux limit x 10 -16 ergs/s/cm^2 Vertical line = 12k sq. deg in 3 years (6.15 deg 2 field, 90 nights, 8.5 hrs/night) Extended field exposures are more efficient at sampling faint target distributions but sacrifice efficient fiber exposure time usage and longer a longer survey Target Measurement Efficiency 1.2 1.0 0.8 0.6 1.2 1.0 0.8 0.6

10. 10 Most of the BAO signal for [OII] will be measured with the Red spectrograph BigBOSS Spectrographs White paper

11. 11 Sky background from BOSS studies (Gemini high resolution spectra scaled in flux to SDSS) –Should be a conservative estimate, will update with DEEP2 DEIMOS spectra Galaxy spectrum and emission line fluxes are from zCOSMOS catalog (Ilbert, 2008) –Emission lines are normalized to Ilbert fluxes and given 50 km/s Gaussian line width Throughput calculation comes from a spreadsheet calculation from Robert Barkhouser and is based off of WFMOS efficiencies –*INCLUDES loss at 150micron fiber due to 1” seeing (~32% loss) Only considered Red Spectrograph Arm (0.8-1.1 micron) for now OH sky lines and Throughput

12. 12 Simplified spot model is a circle (fiber) convolved with a Gaussian (optics) Since we are using the SDSS III / BOSS spectrograph design, match spots at 9000 Ang, rescale for 150 micron fibers, and put on 18 micron pixels Created data cube with spot center shifted along one axis (dispersion direction) –Provides look up table for each position in 2D spectrum, speeds up simulation BigBOSS spots X Y Spot Center Shift

13. 13 Advantage 1: High resolution allows us to work between the night sky lines Why High Resolution? Advantage 2: High resolution splits the [OII] doublet –Forbidden transition gives an unambiguous line identification –Two lines doubles the chance of line measurement among bright sky lines Sky Lines [OII] [OII] 3726, 3729 @ z=1.4 25 sky fibers Observed Spectrum Sky-Subtracted Spectrum

14. 14 Collapsed 2D spectra and variance to a 1D Signal-to-Noise spectrum Repeat simulation for galaxy spectrum redshifted from 0.7<z<2 in dispersion steps (0.732 Ang/pix) on both CCD and HgCdTe detectors Galaxy continuum is interpolated from a zCOSMOS template and [OII] flux draws from the DEEP2 luminosity function at 3.4E-4 (Mpc/h) -3 volume density Data are fed to redshift fitting code (see Jonathan Pober’s talk tomorrow morning) Spectral Simulation z [OII] [OIII] 4959 [OIII] 5007 HH

15. 15 Effect of Resolution on Line Sensitivity Compute the error per 1-D pixel for read + sky + Poisson noise Compute the flux required for S/N=8 in 30min for a single line at the minimum 2 pixel resolution Higher resolution leads to a smaller redshift coverage for one spectrograph Lower resolution leads to decreased overall line sensitivity as more sky is present in the line measurement

16. 16 Seems plausible to do 12k sq. deg in 3 years by limiting z<1.5 at the source densities we have been talking about for BigBOSS We can go to much higher densities at lower redshifts, but must sacrifice objects at higher redshifts Simulated 2D spectra creation tools exist with some simple assumptions about the design Need to do: Color cut needs refinement for surveying targets at higher redshift Improve flux scaling to reflect latest instrumentation Refinement of survey distributions and creation of mocks Instrumentation studies! (Throughputs, scattered light in PSFs, resolution optimization) Different field sizes, fiber densities, fiber overlap fractions, placement logic Feed final measured redshift distributions to FoM calculators Results so far….

17. 17 Going Forward… Checks for Star-Forming Galaxies –Current clustering estimates from small surveys {Sumiyoshi (2009), Geach (2008), Orsi (2009)} –Need wide field survey with mask (CARS?) –Need clustering from appropriate color selection (zCOSMOS) –Stellar Mass vs [OII] ? –LF out to higher redshift (z=1.7) with DEIMOS or alternate? Completeness for faint end (z=0.7)? Semi-analytic mock catalogs / HOD N-body –Define goals for each simulation study –Do we focus on lower mass ELGs for 0.7<z<1.5? –Do we include other target distributions (QSOs from 1<z<2)? –Can we coordinate with DES mocks? Resources –Requirements? –NERSC time available? –European or Chinese contributions?

18. 18 DEEP2 luminosity function for constant volume number density Flux limit scales with D L -2, volume density, AND wavelength of [OII] at zmax Exposure times are currently scaled to 1800s at z=2 for S/N=8 and [OII]=5E-17 flux limit Scaling Flux