Research

I am an observational cosmologist who uses weak gravitational lensing as a tool to probe the nature of cosmic acceleration and astrophysical properties of galaxies and galaxy clusters. While there is strong evidence for cosmic acceleration, its nature is poorly understood. In the CDM model, cosmic acceleration is caused by dark energy which makes up about 69% of the cosmic energy density. If dark energy evolves with time, it would imply a new type of dynamical field. Cosmic acceleration might also imply the breakdown of Einstein’s General Relativity at cosmological scales. Either way, cosmic acceleration requires modification of fundamental physics.

Weak lensing manifests as a small, coherent distortion effect on distant galaxy images caused by the gravitational field due to the foreground matter density along the line-of-sight. Weak lensing allows us to directly measure the distributions of dark matter, which makes up about 85% of the matter in the Universe, around galaxies and galaxy clusters, and also offers an unbiased measurement of the large scale structure of the Universe. This enables us to explore both galaxy/cluster astrophysics and cosmic acceleration.

We are in the golden age of survey astronomy with billions of dollars invested in ongoing and upcoming projects: ground-based optical telescope surveys such as Subaru Hyper Suprime-Cam (HSC), Kilo Degree Survey (KiDS), Dark Energy Survey (DES), and Large Synoptic Survey Telescope (LSST), and space-based optical/near-infrared telescope surveys such as ESA/NASA’s Euclid mission and NASA’s Wide-Field Infrared Survey Telescope (WFIRST) mission. All of these surveys have been optimized for WL measurements.

I have been working on the HSC survey since I was a grad student. I have the HSC Builder status. I am involved in WFIRST as a member of the Cosmology Science Investigation Team and am a member of the LSST Dark Energy Science Collaboration. I am one of the co-coordinators of the cross-collaboration between HSC and Atacama Cosmology Telescope (ACT).

I have been working on various aspects of high-precision WL measurements using large-volume astronomical data sets. This includes not only scientific research but also hardware and software development. My expertise covers all aspects of translating raw pixels into optimal cosmological constraints.

Cluster Cosmology

One of my main research areas is to constrain cosmology through the abundance of galaxy clusters with weak lensing mass calibration. Although cluster abundance is one of the most powerful cosmological probes, current constraints are limited by the systematic uncertainties of cluster mass, which is often estimated by optical richness, X-ray, and the Sunyaev-Zel’dovich (SZ) effect, due to uncertain physical assumptions such as hydrostatic equilibrium. On the other hand, weak lensing offers unbiased mass estimates because of its direct sensitivity to the dark matter distribution.

As a thesis project, I measured weak lensing mass of a high-redshift, massive galaxy cluster discovered Atacama Cosmology Telescope (ACT) through the SZ effect, using imaging data taken by the Subaru Suprime-Cam. I showed that the existence of this cluster is consistent with the prediction by the CDM model (Miyatake et al., 2013). After that, I keep close ties with the ACT collaboration and plays a leading role on ACT cluster cosmology with weak lensing mass calibration. I am currently working on cosmological constraints by the abundance of ACTPol SZ galaxy clusters with weak lensing mass calibration by the HSC first-year data set.

Cosmology with Galaxy-galaxy Lensing and Galaxy Clustering

My other research area is to constrain cosmology through a combination of galaxy clustering and galaxy-galaxy lensing. The amplitude of galaxy clustering is another powerful observable to characterize structure growth. To extract cosmological information from the amplitude, we need to calibrate the difference between the galaxy and dark matter distribution, which is called galaxy bias. The galaxy bias can be calibrated by measuring the dark matter distribution around galaxies through weak lensing.

Combining clustering of the SDSS-III/BOSS spectroscopic galaxy sample (z~0.5) and weak lensing measurements with CFHTLenS galaxy shapes, we constrained the amplitude of density fluctuations and matter energy density as well as astrophysical properties of the BOSS galaxies (Miyatake et al. 2015, More et al. 2015). I am currently working on a similar analysis with the weak lensing measurements using the HSC first-year data set, which will be one of the key cosmological results of HSC.

Cosmographic Distance Ratio

Weak lensing depends on the distances between an observer, lens, and sources. Such distance information can be extracted by taking the ratio of lensing signals that share the same lens but have different sources. We performed the first distance ratio measurement using optical weak lensing signal and CMB lensing signal. We found that the measurement is consistent with the Planck LCDM model (Miyatake et al. 2017).

In the standard scenario of the cosmic structure formation, dark matter first collapses into a halo due to gravitational instability and then galaxies and galaxy clusters are formed in dark matter halos. Therefore, there is a difference between the dark matter distribution and galaxy or galaxy cluster distribution, which is called halo bias. When extracting cosmological information from galaxy distribution, we need to correct for halo bias. Otherwise, cosmological constraints will be biased.

In observational studies, it has been assumed that halo bias solely depends on halo mass. However, analytical and numerical studies showed that halo bias can also depends on other halo properties such as formation time and concentration of the dark matter halo profile, which is called halo assembly bias. If assembly bias exists, ignoring assembly bias results in biased cosmological constraints.

I found the observational evidence of assembly bias in galaxy clusters identified in SDSS DR8 data (Miyatake et al., 2016), using cluster lensing and cluster clustering. Although it turned out that the assembly bias signal has inevitable contamination stemming from the cluster finding algorithm, this work stimulated the community to actively develop robust assembly bias measurement algorithms.

Astronomical Instrumentation

When I was a graduate student, I developed a firmware component of the CCD readout electronics for HSC. This firmware is a FPGA code that bridges the analog front-end and the data acquisition computers (Miyatake et al., 2012). I also performed systematic tests of the readout electronics with the analog front-end (Nakaya, Miyatake et al., 2012).

I also developed an automated focusing algorithm for HSC. This algorithm derives the focus position using a size of stars on focus CCDs that have +/- 200um offset from the focal plane. Using this algorithm, HSC is able to control focus position before each exposure.

Imaging Data Reduction

I contributed to the software development of the HSC data reduction pipeline which is based on the LSST pipeline called DM stack. This work included the implementation of a galaxy shape measurement algorithm and a multi-exposure fitting algorithm which enables accurate modeling of point spread function (PSF) and galaxy profiles (Miyatake et al., 2013).

I was one of the development team of the galaxy image simulation software GalSim (Rowe et al., 2015), which was used to generate simulated images in the GREAT3 galaxy shape measurement challenge (Mandelbaum et al., 2014). My contribution includes the development of an algorithm to simulate optical PSF of a ground-based and space-based telescope.

Since the HSC survey began operations in 2014, I have been playing a leading role in generating the HSC galaxy shape catalog. This catalog will be one of the largest publicly-available galaxy shape catalogs before the LSST era. I performed various systematic tests such as investigating the PSF accuracy and examining selection criteria. This effort resulted in the HSC first-year shape catalog paper (Mandelbaum, Miyatake, et al., 2017).