Day 860

Tuesday.

1503.08210
Dust production 680-850 million years after the big bang
Michalowski

Dust plays an important role in the understanding of the universe, but it is not obvious yet how the dust in the distant universe was formed.  Derived the dust yields per AGB star and per SN required to explain dust masses of galaxies at z=6.3-7.5 for which dust emission has been detected, or unsuccessfully searched for.  Find very high required yields, implying that AGB stars couldn't contribute substantially to dust production at these redshifts, and that SNe could explain these dust masses, but only if they do not destroy majority of the dust they form (which is unlikely given the upper limits on the SN dust yields derived for dust non-detected galaxies).    This suggests that the grain growth in the ISM is likely required at these early epochs.


1503.08212
Correlating galaxy color and halo concentration: a tunable halo model of galactic conformity
Paranjape, Kovac, Hartley, Pahwa

Extend the HOD framework to generate mock galaxy catalogs exhibiting varying levels of "galactic conformity", which has emerged as a potentially powerful probe of environmental effects in galaxy evolution.  The model correlates galaxy colors in a group with the concentration of the common parent DH through a "group quenching efficiency" rho which makes older, more concentrated haloes at fixed mass preferentially host redder galaxies.  Find that, for a specific value of rho, this 1-halo conformity matches corresponding measurements in a group catalog based on SDSS.  The mocks also display conformity at large separations from isolated objects, potentially an imprint of halo assembly bias.  A detailed study - using mocks with assembly bias erased who keeping 1-halo conformity intact - reveals a rather nuanced situation, however.  At separations <4 Mpc, conformity is mainly a 1 halo effect dominated by the largest haloes and is not a robust indicator of assembly bias.  Only at very large separations (>8Mpc) does genuine 2-halo conformity, driven by the assembly bias of small holes, manifest distinctly.  Explain all these trends in standard Halo Model terms.  The model opens the door to parameterized HOD analyses that self-consistently account for galactic conformity at all scales.

1503.08214
Anomaly detection for machine learning redshifts applied to SDSS galaxies
Hoyle, ... Seitz, Weller, et al

Persent an analysis of anomaly detection for machine learning redshift estimation.  Anomaly detection allows the removal of poor training examples, which can adversely influence redshift estimates.  Anomalous training examples may be photometric galaxies with incorrect spectroscopic redshifts, or galaxies with one or more poorly measured photometric quantity.  Select 2.5M 'clean' SDSS DR12 galaxies with reliable spec-z, and 6730 'anomalous' galaxies with spectroscopic z measurements which are flagged as unreliable.  Contaminate the clean base galaxy sample with galaxies with unreliable redshifts and attempt to recover the contaminating galaxies using the Elliptical Envelope technique.  Then train four machine learning architectures for z analysis on both the contaminated sample and on the preprocessed 'anomaly-removed' sample and measure z statistics on a clean validation sample generated without any preprocessing.  Find an improvement on all measured statistics of up to 80% when training on the anomaly removed sample as compared with training on the contaminated sample for each of the machine learning routines explored.  Further describe a method to estimate the contamination fraction of a base data sample.


1503.08215
Accurate photometric redshift probability density estimation - method comparison and application
Rau, Seitz, ... Gruen, Hoyle, et al

Introduce an ordinal classification algorithm for photo-z estimation, which vastly improved the reconstruction of photo-z PDFs for individual galaxies and galaxy samples.  Apply method to CFHTLS galaxies.  The ordinal classification algorithm treats distinct z bins as ordered values, which improved the quality of photo-z PDFs, compared with non-ordinal classisfication architectures.  Also propose a new single value point estimate of the galaxy redshift, that can be used to estimate the full z PDF of a galaxy sample.  This method is competitive in terms of accuracy with contemporary algorithms, which stack the full z PDFs of all galaxies in the sample, but requires orders of magnitudes less storage space.  The methods described n the paper greatly improved the log likelihood of individual object z PDFs, when compared with a popular NN code (ANNz).  In our case, this improvement reaches 50% for high-z objects defined as z>0.75.  Show that using these more accurate photo-z PDFs will lead to a reduction in the systematic biases by up to a factor of 4, when compared with less accurate PDFs obtained from commonly used methods.  The cosmo analyses examined upon which showed improvement are the following: gravitational lensing cluster mass estimates, modeling of angular correlation functions, and modeling of cosmic shear correlation functions.


1503.08218
Reconstruction of small-scale galaxy cluster substructure with lensing flexion
Cain, Bradac, Levinson

Present a reconstruction of galaxy-cluster-scale mass distribution from simulated GL data sets including SL, WL shear, and measurements of quadratic image distortions - flexion.  The lensing data is constructed to make a direct comparison between mass reconstructions with and without flexion.  Show that in the absence of flexion measurements, significant galaxy-group scale substructure can remain undetected n the reconstructed mass profiles, and that the resulting profiles underestimate the aperture mass in the substructure regions by~25-40%. When flexion is included, sub haloes down to a mass of ~3e12 Msun can be detected at an angular resolution smaller than 10".  Aperture masses from profiles reconstructed with flexion match the input distribution values to within an error of ~13%, including both statistical error and scatter.  This demonstrates the important constraint that flexion measurements place on substructure in galaxy clusters and its utility for producing high-fidelity mass reconstructions.


1503.08225
Galaxy sizes as a function of environment at intermediate redshift from the ESO Distant Cluster Survey

Find no significant difference in the size distributions of cluster and field galaxies of a given morphology (for 0.4<z<0.8 data).  Similarly, find no significant difference in the size distributions of cluster and field galaxies of similar rest-frame B-V colors.  Rule out average size differences larger than 10-20% in both cases.  Consistent conclusions are found with the spectroscopic and photometric samples.  These results have important consequences for the physical process(es) responsible for the size evolution of galaxies, and in particular the effect of the environment.  The remarkable growth in galaxy size observed from z~2.5 has been reported to depend on the environment at higher z (z>1), with early-type/passive galaxies in higher density environments growing earlier.  Such dependence appears at lower z.  Therefore, if the reported difference at higher-z is real, the growth of field galaxies has caught up with that of cluster galaxies by z~1.  Any putative mechanism responsible for galaxy growth has to account for the existence of environmental different at high z and their absence (or weakening) at lower z.


1503.08506
On the bias of the distance-redshift relation rom gravitational lensing
Kaiser, Peacock

Does GL change the distance-redshift relation D(z) or the mean flux density of sources?  Interest in this has been rekindled by recent studies in NL relativistic perturbation theory that find biases in both the area of a surface of constant z and in the mean distance to this surface, with a fractional bias in both cases on the order of the mean squared convergence <kappa^2>.  Any such are bias could alter CMB cosmology, and the corresponding bias in mean flux density could affect SN cosmology.  Show that, in an ensemble averaged sense, the perturbation to the area of a surface of constant z is in reality much smaller, being on the order of the cumulative bending angle squared, or roughly a part-in-a-million effect.  This validates the arguments of Weinberg (1976) that the mean magnification mu of sources is unity and of Kibble+Lieu(2005) that the mean direction-averged inverse magnification is unity.  It also validates the conventional treatment of lensing in analysis of CMB anisotropies.  But the existence of a scatter in magnification will cause any NL function of these conserved quantities to be statistically biased.  The distance D, for example, is proportional to mu^-1/2 so lensing will bias <D> even if <mu>=1.  The fractional bias in such quantities is generally of order <kappa^2>, which is orders of magnitude larger than the area perturbation.  Claims for large bias in area or flux density of sources appear to have resulted from misinterpretation of such effects: they do not represent a new non-Newtonian effect, nor do they invalidate standard cosmological analyses.


1503.08680
Instrumental systematics and weak gravitational lensing
Mandelbaum

Present a pedagogical review of the WL measurement process and its connection to major scientific questions such as DM and DE.  Then describe common ways of parameterizing systematic errors and understanding how they affect WL measurements.  Finally, discuss several instrumental systematics and how they fit into this context, and conclude with some future perspective on how progress can be made in understanding the impact of instrumental systematics on WL measurements.