Peter Schneider - Cosmological probes by gravitational lensing - Light deflection depends only on mass distribution, without reference to nature and state of matter (light or dark) - Probe geometry of universe, evolution of density field, density fluctuations, etc. - Deflection causes shift of image position, unobservable (in principle) unless multiple images are formed - Lensing equation; surface brightness is conserved, but focusing and defocusing of light bundles - Strong distortion yields arcs - Multiple-image systems -> mass determination (in some cases to few %), but sources at different distances from center of lens needed to get mass profile - Time delay observable for varying sources -> measures H_0, without distance ladder (Cepheids) or initial PSD (CMB) - But usually don't have enough constraints to determine mass profile of lens - Good new measurement by Sherry Suyu - CDM theory predicts hundred of subhalos for Galaxy, but only two dozens observed: "Second dark matter crisis" - They may be mostly dark because of suppressed star formation - Example: B1422+231, quad system where mass profile can describe astrometry, but not flux ratios - Substructure can explain this, and (PS argues) detects CDM sub-halos - Ellipticity of images (second-order brightness moments) - Dark matter is collisionless - Bullet cluster - X-ray emitting gas shock allows measurement of cluster merger velocity - Dark matter is centered on Galaxies, which cross each other "freely" - Cosmic shear (lensing by the large-scale structure) - Infer statistical properties of matter out to "cosmic wallpaper" of faint galaxies - Determined by integral of perturbed background metric along line of sight - Infer integrals over 3D power spectrum of mass density distribution - Effect detected in 2000 - Not currently competitive for cosmological parameters, but current surveys map only 1/1000 of sky - But useful for constraints on small scales - Relation between light and matter - Cosmologists predict large-scale distribution of matter, astronomers measure large-scale distribution of galaxies - How are the two related? - Biasing of galaxies: galaxies may not be good tracers of underlying matter distribution - Prospects: - Second-generation surveys -> tighten parameters, consolidate techniques - Third generation surveys - LSST, dedicated weak-lensing satellite such as SNAP/JDEM or DUNE