care about the Milky Way? 2 'Galactic' astronomers… ❖ … want to understand how galaxies are formed/ evolve/merge etc. ❖ … observe and model our Galaxy in different wavelengths
cosmologists care about the Milky Way? 'Cosmologists' … ❖ … want to understand the history, evolution and fate of the entire Universe ❖ …study millions of galaxies and their statistical properties Hubble Ultra Deep Field
cosmologists care about the Milky Way? 'Cosmologists' … ❖ … want to understand the history, evolution and fate of the entire Universe ❖ …study millions of galaxies and their statistical properties Hubble Ultra Deep Field Warm dark matter Cold dark matter Bromm+ (2009)
cosmologists care about the Milky Way? 'Cosmologists' … ❖ … want to understand the history, evolution and fate of the entire Universe ❖ …study cosmological backgrounds Cosmic Microwave Background (Planck)
Milky Way: Synergies ❖ Our Galaxy is an inconvenient foreground for Cosmology ❖ Cosmology needs a model of the Milky Way to remove it ❖ Galactic astronomers benefit from cosmology missions, giving them fantastic data to work with 7
the game is component separation, not noise reduction" H.K. Eriksen, 'Advances in Theoretical Cosmology in Light of Data 2017' ❖ Extinction for cosmological galaxy surveys ❖ Cosmic infrared background measurements ❖ De-lensing of CMB data for primordial gravitational waves Daniel Lenz, Caltech/JPL Foregrounds in observational cosmology 8
= Extinction in B band - Extinction in V band ❖ More dust => larger E(B-V) ❖ E(B-V) maps essential for correcting observations for Galactic reddening 10
approach 11 ❖ Find many sources with known spectrum (e.g. stars, passive galaxies) ❖ Measure spectra, attribute differences to dust ❖ E.g. Schlafly+ (2014), Green+ (2015) used 500 million stars from Pan- STARRS to measure reddening to 4.5 kpc
approach 11 ❖ Find many sources with known spectrum (e.g. stars, passive galaxies) ❖ Measure spectra, attribute differences to dust ❖ E.g. Schlafly+ (2014), Green+ (2015) used 500 million stars from Pan- STARRS to measure reddening to 4.5 kpc ❖ Direct measurements are hard! ❖ Photometric/ spectroscopic errors ❖ Ensuring sources lie behind full dust column ❖ Ensuring adequate number of sources have been measured
approach 11 ❖ Find many sources with known spectrum (e.g. stars, passive galaxies) ❖ Measure spectra, attribute differences to dust ❖ E.g. Schlafly+ (2014), Green+ (2015) used 500 million stars from Pan- STARRS to measure reddening to 4.5 kpc ❖ Direct measurements are hard! ❖ Photometric/ spectroscopic errors ❖ Ensuring sources lie behind full dust column ❖ Ensuring adequate number of sources have been measured Green+ (2015)
measure of E(B-V) ❖ E(B-V) is proportional to the dust column, so can convert dust column tracer to E(B-V) ❖ SFD used dust emission from IRAS to derive a calibration factor from FIR emission to E(B-V) ❖ Full-sky, high sensitivity measurements -2 -0.3 log10 (E(B V )SFD [mag]) Reddening map of Schlegel, Finkbeiner, and Davis (1998) 12
temperature correction to go from dust emission to a dust column density ❖ FIR emission may have contributions from zodiacal light and unresolved galaxies -2 -0.3 log10 (E(B V )SFD [mag]) Reddening map of Schlegel, Finkbeiner, and Davis (1998) 13 Dust emission as measure of E(B-V)
temperature correction to go from dust emission to a dust column density ❖ FIR emission may have contributions from zodiacal light and unresolved galaxies -2 -0.3 log10 (E(B V )SFD [mag]) Reddening map of Schlegel, Finkbeiner, and Davis (1998) 13 Dust emission as measure of E(B-V)
basis for E(B-V) ❖ Neutral atomic hydrogen (HI), most abundant element ❖ Gas and dust are well-coupled in the interstellar medium ❖ Resulting maps free from errors due to dust temperature, zodi, and extragalactic emission ❖ Limited by non-HI gas along the line of sight 14
❖ For the full sky, allow different E(B-V)/NHI ratios for different velocities ❖ High-velocity gas has less dust, as expected 21 102 101 0 101 102 vLSR [km s 1] 0.000 0.005 0.010 0.015 0.020 0.025 ↵v = E(B V )/Nv HI [mag/1020 cm 2] ↵v DL, Hensley, Doré (2017)
l [deg] 42 52 62 b [deg] SFD Modelsimple 130 112 95 l [deg] SFD Modeltophat 0.00 0.02 0.04 0.016 0.008 0.000 0.008 0.016 E(B V ) [mag] 102 103 104 105 # data points HI-based reddening model ❖ Black: HI high-velocity clouds ❖ Color: Reddening residuals 22 Full HI column density Only local HI (|v| < 90 km/s) DL, Hensley, Doré (2017)
to use this extinction map ❖ New HI-based extinction map ❖ In line with independent corrections, but much higher resolution and better sky coverage ❖ Yahata+ (2007) find correlation of SFD with large-scale structure ❖ For high latitudes, our map overcomes many of the SFD problems and is much more sensitive than stellar data- based E(B-V) maps 26
cosmological probe … of star formation history Planck collaboration (2013 XXX) ❖ Strong constraints on SFH up to z=2.5 ❖ Probe dust temperature across cosmic times ❖ Understand star formation in DM halos 29
and CIB match great in z and MHalo ❖ Ideal probe of relation between dark and luminous matter … of large scale structure to cross-correlate with lensing Planck collaboration (2014 XVIII) The CIB as cosmological probe 30
and CIB match great in z and MHalo ❖ Ideal probe of relation between dark and luminous matter … of large scale structure to cross-correlate with lensing Planck collaboration (2014 XVIII) The CIB as cosmological probe 30
CIB maps? A. Fit different frequency channels with modified blackbody spectra B. Utilize the different angular power spectra of these components C. Use template maps of Galactic dust (e.g. HI-based) ❖ Galactic thermal dust and CIB dust dominate on large scales at ~200 to 1000 GHz ❖ How to disentangle them? 31
• Generalised linear model (GLM) • • Regularised: • • Accounts for all features along line of sight I = X i ✏iTi B Radial Velocity | Datai Modeli |2 + ↵ · |✏i | 35
How to verify, especially on large scales? ❖ Simulations and re-sampling ❖ How to avoid fine-tuning of the component separation? ❖ How to extend this larger areas, probing non-HI gas? 37
How to verify, especially on large scales? ❖ Simulations and re-sampling ❖ How to avoid fine-tuning of the component separation? ❖ How to extend this larger areas, probing non-HI gas? ❖ How to jointly use frequency information and angular power spectra? 37
❖ ❖ Ideally: Construct lensing potential from CMB itself ❖ Too noisy, but will be relevant for CMB-S3 and CMB-S4 ❖ Currently: Use external tracer of lensing potential such as the CIB 42 Te↵( x ) = Ttrue( x + r ( x ))
43 ❖ Using large-scale CIB maps to de-lens Planck CMB TT ❖ Characteristic sharpening of the peaks detected at high significance ❖ Sharpening can hardly be mimicked by other effects Larsen+ 2016 Large scales Small scales
Underlying CIB maps very simplistic: M545 - M857/77 ❖ Only 40% correlated (ideal CIB would give up to 80%) ❖ High resolution, high accuracy, large scale CIB maps needed 44 Larsen+ 2016
Underlying CIB maps very simplistic: M545 - M857/77 ❖ Only 40% correlated (ideal CIB would give up to 80%) ❖ High resolution, high accuracy, large scale CIB maps needed 44 Larsen+ 2016
BB: The challenge ❖ Lensing of CMB E-modes leads to apparent B-modes ❖ One of the major systematics in the search for primordial gravitational waves 45 Courtesy A. Challinor Large scales Small scales
CIB ❖ Template maps of Galactic dust are key to obtain large- scale CIB maps ❖ Useful by itself to constrain star formation history and connection of dark and luminous matter ❖ Best tool to-date for de-lensing CMB TT and BB, better CIB maps are urgently required 49
GNILC ❖ Planck collaboration (2016 XLVIII), focus on removing CIB from Galactic dust maps ❖ Using the angular power spectra of the two components ❖ Does not agree that well on a pixel-to-pixel basis 51
GNILC ❖ Planck collaboration (2016 XLVIII), focus on removing CIB from Galactic dust maps ❖ Using the angular power spectra of the two components ❖ Does not agree that well on a pixel-to-pixel basis 51
GNILC ❖ Planck collaboration (2016 XLVIII), focus on removing CIB from Galactic dust maps ❖ Using the angular power spectra of the two components ❖ Does not agree that well on a pixel-to-pixel basis 51 0.3 0.2 0.1 0.0 0.1 0.2 0.3 Planck XXX 0.3 0.2 0.1 0.0 0.1 0.2 0.3 GNILC 100 101 102 103
daniellenz
hachama
climbteam
tagtag
linyows
yuki1986
sofievl
sip3ristex
tanichu
sshimizu2006
brainpadpr
kugimasa
colly
irinanazarova
wjessup
jnunemaker
sstephenson
csswizardry
bkeepers
lynnandtonic
tammyeverts
eileencodes
notwaldorf
pauljervisheath
![1020 1021 1022 NHI [cm 2] 10 3 10 2](https://files.speakerdeck.com/presentations/565a714d50c643d5bff99c4a95b1220e/slide_80.jpg){kind=link}