Discussion Session 5:

Christian Wolf

Should we invest more into photo-z or spectroscopic surveys?

Exposure Times

• Well established 8-m instrumentation (last ~5 years)– VLT 8m + VIMOS– Subaru 8m + SuprimeCam

Method z Accuracy Exposure time

Broad-band photo-z 0.03 1Medium-band photo-z 0.01 5Spectroscopy <0.001 50-100

More area / depth for less precision

Photo-Z Attractive For…• Global galaxy evolution

– 1999+: optical HiFi photo-z, e.g. COMBO-17: 104.5 obj.

– Now: Focus on environment needs z/(1+z) ~ 0.001

– At z > 1 not at Poisson limit

• Flux-calibrated photometry– SED from constant physical

footprint, while PSF = f()

• Large-scale structure– Photo-z: 2D in z slices

• Volume selection for spectroscopic surveys– Efficiency increase– Super-cluster environments– Try FMOS / FAST-SOUND

at z = [1.3,1.7] without pre-selection (success <10%?)

• Weak gravitational lensing– Need accurate z distribution– 2010+ lensing surveys want

bias z/(1+z) ~ 0.001 Huge training effort (spec!)

Stand-alone Cross-fertilization

Dis- / Advantage

Property Photometry Spectroscopy

Area / depth +Z accuracy +SED accuracy +Z reliability ~ ~Blended objects (+)

Combine SEDs + Spectra• Improved SED analysis

– Spectroscopic redshift prior– Rest-frame SED known:

age and dust more reliable– Talk by Christian Wolf

• Improved composite stellar population analysis– Line indices & SED together

break some degeneracy – Fix old population better and

see young population as left-over SED difference

– Talk by Daniel Thomas

Age

Dust

Kinney et al. templates 25% z outliers

Abell 901 z ~ 0.17