Gravitational lensing science with CSS-OS Weak lensing: Zuhui Fan (PKU) Shear measurement: Jun Zhang (SJTU) Strong lensing: Ran Li (NAOC)
Outline • Organization of working groups • Weak lensing science • Plan of weak lensing simulation • Shear measurement • Strong lensing science • Strong lensing simulation
Weak lensing working group Memebers: Zuhui Fan (PKU), Jun Zhang (SJTU), Ran Li (NAOC), Dezi Liu (YNU), Xiangkun Liu (YNU), Chuzhong Pan (PKU), Chunxiang Wang (NAOC), Qiao Wang (NAOC), Liping Fu (SHNU), Xi Kang (PMO) Guoliang Li (PMO), Wentao Luo (SJTU) ， Yu Yu (SJTU), Shan Huanyuan (Bonn Uni.) ， Shuo Yuan(PKU) Aims: Mock weak lensing maps based on cosmological simulation Imaging simulations 2-3 shear measurement pipelines Fast statistical analysis tools and theoretical analysis codes Gravitational lensing workshop every 6 months, next meeting in Yun Nan on 4th-7th December 2017.
Strong lensing working group • Member: Ran Li (NAOC), Nan Li (Nottingham) , Dezi Liu (YNU), Guoliang Li (PMO), Xiaoyue Cao (NAOC), Ye Cao (NAOC), Yun Chen (NAOC), Yiping Shu (NAOC), Xin Wang (UCLA), Xiaolei Meng (Tsinghua)…. • Email list CSST_SLWG@googlegroups.com • Collaboration tool: https://tiangongslwg.slack.com/ • Regular telecon every 2-3 weeks • A meeting planned for next spring • Currently, aim to construct a set of strong lensing simulations and produce some forecast papers.
Gravitational lensing e ff ects – gravitational in origin – everywhere in the universe
E xcellent cosmological probe
Cosmological model Parameters Cosmic shear (signal ~ a few percent) Statistical tools • 2-pt correlation • 3-pt correlation • Peak statistics • Minkowski functionals • Nature of dark energy , deviation from w=-1 (1 errors on w0 & w a of 0.07 and 0.1 respectively) • Precise measurement of expansion history of the universe • Dark matter mass distribution to redshift 2. • neutrino mass • Test General relativity • Primordial non-Gaussianity
Example: Using Peaks statistics to constrain the law of gravity Modified gravity theories f(R) gravity theory with chameleon effect give rise to the late-time cosmic accelerating expansion satisfy the solar system gravity test However, the formation and evolution of LSS HS f(R) theory – f R0 parameter with are different Using CFHTLens data f R0 =0 for GR With priors from WMAP9 or Planck15, f R0 can be constrained tightly Liu, Fan et al. 2016
Weak lensing of galaxy formation • Galaxy-dark matter halo co- evolution • Mass and structure of subhalo of satellite galaxies • Mass, shape and profile of clusters • Mass distribution in superclusters, filaments and voids • Mass of dwarf galaxies, UDGs … NASA, ESA, D. Coe
Weak lensing analyses Stage I: first detections of cosmic shear, is around the year of 2000 Stage II: CFHTLenS as the best representative survey 154 degree^2, mag=24.5, seeing ~0.7” Stage III: present (KiDS, DES, HSC, ~1000 degree^2) Stage IV: in the future (CSS-OS, LSST, EUCLID) CSS-OS: Area = 15000 deg^2 resolution= 80% light within 0.15” NUV, u, g, r, i, z, Y multi-band photometry, with average limit 25.2 (g=26.3) for point source. ~300 times more galaxy shapes than stage II survey.
Challenges Great10 handbook Observationally ： - reconstruct the PSF, and measure accurately the shapes of billions faint galaxies - redshift information of individual galaxies Theoretically : - fully explore different statistical quantities - accuracy and speed of theoretical tools Thorough understanding about potential systematics, both theoretical and observational
(1+m) degenerates with cosmological parameters Shan et al. arXiv: 1709.07651 CSS-OS like survey requires dm<0.002 !
Liu, X.K., Pan, C.Z. et al. 2017 Self-calibration CFHTLenS data : m~-0.05 from CFHTLenS simulation calibration our results: m~-0.2 (**should be understood as an effective bias, not necessarily shear measurement bias) m Prior [-0.2, 0.2]
? N-Body Cosmology Theoretical tools simulation SAM Recovered Shapes of galaxies Dark Lens Source Hubble matter HUDF, galaxy galaxy COSMOS Optical Density design CCD properties Shear Instrumental measurement Lensing Ray- pipeline systematics tracing potential Mock Star catalogue Filter response. Detector noise, pixel effects, Ideal Ideal Exposure strategy + + PSF Final image Filter arrangement Images Images Stray light Other detector systematics By Dezi Liu
Plan of Cosmological simulations for CSS-OS • Hyper-Millennium, by Key laboratory for Computational Astrophysics, CAS • 3 Gpc box, ~10000 degree^2 light-cone to z=3, ~100 degree^2 light cone to z=5. • Mass resolution 2.5x10^8 Msun, Resolve all the dark matter haloes that source galaxies reside in. • Semi-analytical galaxy catalogue • Data size ~ 2 PGb, might run this year.
PSF reconstruction and shear measurement pipelines Jun Zhang’s group has solved a number of bottleneck problems in cosmic shear measurement, including � • Correction for PSF model-independently; • Noise Effects are removed; • Accurate to the second order in shear; • Clarify the requirement on the pixel size; • No need to fix the centroid of the image; • Fast image processing � < 10 -2 CPU sec/Gal; • Immune to misidentification of stars as galaxies. References: The new method is carried out JZ, 2008, MNRAS, 383, 113 in Fourier space. In real space, JZ, 2010, MNRAS, 403, 673 it corresponds to measuring JZ & Komatsu, 2011, MNRAS, 414, 1047 the spatial gradients of the JZ, 2011, JCAP, 11, 041 surface brightness field. JZ, Luo, Foucaud, 2015, JCAP, 1, 24 JZ, Zhang, Luo, 2017, ApJ, 834, 8 Lu, JZ, Dong, et al., 2017, AJ, 153, 197
We have built up an image processing pipeline that includes background removal, source selection, PSF reconstruction, shear measurement, etc.. It has been successfully applied to the CFHTlens data for a series of studies regarding lensing physics. Selected Galaxy Images Raw Data from CFHT Cosmic Shear by Galaxy Clusters
Shear measurement pipeline • Jun Zhang’s group: Shear measurement in Fourier space. • Zuhui Fan, Liping Fu: pipeline based on forward-modeling • Wentao Luo: Pipeline base on re-Gaussian method
Synergy with LSST, Euclid CSS-OS: 15000 deg^2, Space mission, NUV, u, g, r, i, z, Y multi-band photometry, with average limit 25.2 (g=26.3 for point source), ~30 galaxies/arcmin^2, operation 2022 Euclid: 15000 deg^2, Space mission, VIS, Y, J ,H, similar galaxy density as CSS-OS, operation 2020 LSST: ~20000 deg^2, u,g,r,i,z bands, r=27.5, fast survey mode, operation 2022 Complementary to each other: Photo-z, shape measurement calibration ~2000 degree common area in the first 2 years ? ,
Summary I • Weak lensing with CSS-OS may answer several most important questions of the universe. • Challenges in both theoretical and observational sides. • Need imaging simulations as realistic as possible. • Information about CSST designs and performance are critical to construct the simulation. • Welcome to join us.
Strong lensig with CSS-OS • ~150000 galaxy scale strong lens systems （ Including ~1000 double lens system) • ~1000 lensed QSO Provide by Yiping Shu • Hundreds of massive clusters with many multiple images • Accurate photo-z for both lens and source. Hubble HFF
DM on small scales: Substructure detection Vegetti et al. 2012
Identity of Dark matter CSS-OS Li et al. 2016 arxiv 1512.06507 COCO simulations Bose+ 2016
Self-interacting dark matter David Harvey et al. 2015
DM on small scales: Center offset Galaxy cluster Abell 3827 Massey et al. 2015 Shu et al. 2016 offset is 1.62+0.47 kpc ?
Density profile at the cluster center Newman et al . 2012 MS2137 10 1 Mpc 5% shear 5 0 − 5 − 10 − 15 − 10 − 5 0 5 10 15
Testing gravity Velocity dispersion à Dynamical mass Gravitational mass = - + Y + - F d 2 2 i j ds (1 2 ) dt (1 2 ) dx dx ij Y : Newtonian dynamical potential F : space curvature potential In GR, F = Y Slides by Wei Du
Cosmological constraints from double source plane strong lensing (DSPL) The observable: In the ideal case of neglecting the effect of the intermediate source (source 1) on the background source (source 2): The factor α depends on the lens mass model The factor α is cancelled out, that alleviates the dependence on the lens model to some extent. Prediction: ~ 10 3 galaxy- In SIS lens model, the stellar velocity dispersion scale DSPL systems (based on is invariable with radius, that leads to Gavazzi et at. 2008 , about one lens galaxy in ~ 40 - 80 could be a DSPL) Slides by Yun Chen
Galaxy science (SL+SSP+Kinematics) • Galaxy mass and structure • Dark matter fraction • Dark matter shape at center • Evolution of Early type galaxies • IMF variation of late type lenses • ……
Galaxy lensing as a telescope Abell 2744, magnification map by CATS team Shu et al. 2016
BY Nan Li Mining more than 10000 lenses from one billion objects Credits: LSST OpSim Group
Lens search: ML Training Phase Deep Feature Learning Yes/No Extractor Module Prediction Phase Trained Feature Deep Yes Extractor Neural Network BY Nan Li
Completeness 80% Purity 80% BY Nan Li
Completeness 90% Purity 90% BY Nan Li
Recommend
More recommend
