Day 877

Wednesday.


1504.07246
Microlensing events from the 11-year observations of the Wendelstein Calar Alto Pixellensing Project
Lee, Riffeser, Seitz, Bender, Koppenhoefer

Present the results of the decade-long M31 observation from the Wendelstein Calar Alto Pixellensing Project (WeCAPP).  WeCAPP has monitored M31 from 1997 till 2008 in both R- and I-filters, thus provides the longest baseline of all M31 microlensing surveys.  The dat are analyzed with the difference imaging analysis, which is most suitable to study variability in crowded stellar fields.  Extracted light curves based on each pixel, and devised selection criteria that are optimized to identify microlensing events.  This leads to 10 new events, and sums up to a total of 12 microlensing events from WeCAPP, for which their timescales, flux excesses, and colors from their light curves are derived.  The color of the lensed stars fall between (R-I)=0.56 to1.36, with a median of 1.0 mag, in agreement with the expectation that the sources are mostly likely bright, red stars at post main-sequence stage.  The event FWHM timescales range from 0.5 to 15 days, with a median of 3 days, in good agreement with predictions based on the model of Riffeser+2006.


1504.07248
Coevolution between supermassive black holes and bulges is not via internal feedback regulation but by rationed gas supply due to angular momentum distribution
Cen

As the title says.  Without physical fine-tuning, neither the SMBHs nor the stellar bulges can self-regulate or inter-regulate by driving away already fallen cold gas to produce theo served correlation between them.  Suggest an alternative scenario where the observed mass ratios of the SMBHs to bulges reflect the angular momentum distribution of in fallen gas such that the mass reaching the stable accretion disc is a small fraction of that reaching the bulge region, averaged over the cosmological time scales.  Test this scenario using high-res, large-scale cosmo hydro sims without AGN feedback, assuming the angular momentum distribution of gas landing in the bulge region to yield a Mestel disc that is supported by independent simulation resolving the Bondi radii of SMBHs.  A mass ratio of 0.1-0.3% between the very low angular momentum gas that free-falls to the sub-parsec region to accrete to the SMBH and the overall SFR is found.  This ratio is found to increase with increasing redshift to within a factor of ~2, suggesting that the SMBH to bulge ratio is nearly redshift independent, with a modest increase with z, a testable prediction.  Furthermore, the duty cycle of AGN with high Eddington ratios is expected to increase significantly with z.  Finally, while SMBHs and bulges are found to coevolve on ~30-150 Myr time scales or longer, there is indication that, on shorter time scales, the SMBH accretion rate and SF may be less correlated.


1504.07250
Genetically modified haloes: towards controlled experiments in $\Lambda$ CDM galaxy formation
Roth, Pontzen, Peiris

Proposed a method to generate 'genetically-modified' (GM) initial conditions for high-res sims of galaxy formation in a cosmo context.  Building on Hoffman-Ribas algorithm, start from a reference sim with fully random ICs, then make controlled changes to specific properties of a single halo (such as its mass and merger history).  The algorithm demonstrably makes minimal changes to other properties of the halo and its environment, allowing isolation of the impact of a given modification.  As a significant improvement over previous work, able to calculate the abundance of the resulting objects relative to the LCDM reference cosmology.  The approach can be applied to a wide range of cosmic structures and epochs; here, study two problems as a proof-of-concept.  First, investigate the change in density profile and concentration as the collapse time of three individual halos are varied at fixed final mass, showing good agreement with previous statistical studies using large simulation suites.  Second, modify the z=0 mass of haloes to show that the theoretical abundance calculations correctly recover the halo mass function.  The results demonstrate that the technique is robust, opening the way to controlled experiments in galaxy formation using hydro zoom sims.


1504.07255
Measuring photometric redshifts using galaxy images and deep neural networks
Hoyle

Propose a new method to estimate the photo-z of galaxies by using the full galaxy image in each measured band.  This method draws from the latest techniques and advances in machine learning, in particular Deep Neural Networks.  Pass the entire multi-band galaxy image into the machine learning architecture to obtain a z estimate that is competitive with the best existing standard machine learning techniques.  The standard techniques estimate z using post-processed features, such as magnitudes and colors, which are extracted from the galaxy images and are deemed to be salient by the user.  This new method removes the user from the photo-z estimation pipeline.  However, note that Deep Neural Networks require many orders of magnitude more computing resources than standard machine learning architectures.