Transcript Catalog Simulations

Catalogs Framework
Andrew Connolly
Simulations Scientist
September 19-20, 2013
CDP FINAL DESIGN REVIEW
September 19 - 20, 2013
FINAL
DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
Name of Meeting • Location • Date - Change in Slide Master
1
Introduction to Catalogs
The LSST system is complex and resembles no other system in
astronomy. Simulated data can be used to identify and
quantify risk associated with various aspects of the project.
Engineering sensitivity studies, algorithmic optimization,
informing sizing models, and interface integration and testing
are all areas where simulated data can provide significant
leverage.
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
2
Introduction to Catalogs
Simple analysis can be done with simple catalogs: grids of point
sources, or extended sources. To analyze the system in full
complexity a Universal model with realistic properties beyond
what is strictly required to meet design specifications is
needed.
We present a model of the universe containing galaxies, stars,
moving objects and variable objects that meets the
requirements in the Requirements Document and is also
realistic at a level that facilitates in depth analysis of
algorithms and science capabilities.
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
3
Dust Model
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
4
Introduction to Catalogs: Galactic Structure
− Stars are embedded in a dust model based on that of Amôres
and Lépine (2005)
− So that there is no discontinuity in the measured reddening,
the 3D dust model is normalized to the Schlegel et al. (1998)
maps at 100 kpc.
− A realistic dust distribution is important for tests of
photometric calibration, photometric redshift algorithms,
stellar population modeling, among others.
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
5
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
6
Galaxies
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
7
Introduction to Catalogs: Galaxies
− Positions and redshifts from Millennium (Springel et al. 2005)
assuming a standard Λ-CDM cosmology
− Semi-analytic baryon model from De Lucia et al. (2006)
− Light cone produces 4.5x4.5 degree footprint to z=6 and samples
masses from 2.5x109 to 1012 M
− Using UVRIK magnitudes from the DeLucia catalogs for the disk and
bulge, Bruzual and Charlot (2005) SEDs are fit for each component.
− Sizes (half-light radius) are assigned using the absolute-magnitude
vs. half-light radius relation from Gonzalez et al. (2009)
− AGNs are assigned using the Bongiorno et al. (2012) luminosity
functions. For each AGN it’s host galaxy colors and stellar mass are
computed from empirical relations from SDSS data. These are used
with the redshift to pick a host galaxy from the galaxy catalog.
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
8
Introduction to Catalogs: Galaxies
A section of the galaxy catalog showing the density of sources
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
9
Galactic
Structure
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
10
Introduction to Catalogs: Galactic Structure
− Stars in the galaxy are drawn from a realization of the Galfast
model which uses density laws from Jurić et al. (2008) and
metallicity and kinematic models from Ivezić et al. (2008) and
Bond et al. (2010). These were all derived from SDSS data.
− SEDs are assigned using the SDSS colors from the Galfast
model
• F, G, and K main sequence and RGB use Kurucz (1993) models
• White Dwarfs use Bergeron et al. (1995) models
• M, L, and T Dwarfs from Cushing, Bochanski, Burrows, Patterson and
Hawley, and Kowalski derived from SDSS data.
• The galactic dust model uses that of Amôres and Lépine (2005)
normalized to the Schlegel et al. (1998) results at 100 kpc.
− 10% of stars are variable at a level observable by LSST
including RR Ly, Cepheids, and M-dwarf flares.
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
11
Introduction to Catalogs: Galactic Structure
The density of stars from the Galfast model of the galaxy.
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
12
Solar System
Model
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
13
Introduction to Catalogs: Solar System
− The solar system model is a realization of the Grav et al.
(2007) model. This is the same model used by the Pan STARRS
project.
− All major groups are represented: main belt, NEO, trojans,
TNOs and comets
− The sample is complete to V=24.5
− Each object is assigned a carbonaceous or stony composition
(DeMeo et al. 2009) based on a relation to the size of its orbit
based on SDSS observations.
− The package oorb (Granvic et al. 2009) is used to calculate the
orbit and V band magnitude which is used to derive the LSST
band observations at a given point in time
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
14
Introduction to Catalogs: Solar System
The distribution of objects in the Solar System Model color
coded by population
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
15
Query
Framework
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
16
Introduction to Catalogs: Query Framework
− Query – Pure python and contains the mapping from schema
to python objects and hooks for connecting to the database
− Measurement and Formatting – Pure python (except for
astrometry routines) and contains software for calculating
astrometry and photometry given a time and pointing. Also
contains mappings of python objects to catalog columns and
formatting for catalogs.
− Extensible through python sub-classing and inheritance.
− Examples of catalogs are:
•
•
•
•
Inputs to PhoSim
reference catalogs
moving objects and simulated alert streams
calibration catalogs
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
17
Validation of
Catalogs
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
18
Stellar number counts
Star counts from Galfast
(Jurić et al. 2008; solid)
compared to counts
from the Besançon
model (Robin et al.
2003). Symbols are star
counts from the SDSS
catalog.
• Sizing models
• Star galaxy separation
• Deblending
• PSF modeling
Cumulative star counts for fields at l=90o
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
19
Sensitivity of neff on Galaxy Size Distribution
2
-2
4
Comparison of the
effective number of
galaxies for weak lensing
measurements relative to
the value calculated from
the COSMOS data (Chang
et al. 20013).
• Weak lensing
• Size distributions
0
-4
Wider distribution
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
20
Questions?
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
21
Galaxy Number Counts
Galaxy number
counts matched to
observed number
counts (Metcalf et
al. 2010)
• Sizing model
• Star galaxy
separation
• Weak lensing
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
22
Galaxy Ellipticity
Top panels show the
comparison of the base catalog
ellipticity distribution to the
COSMOS distribution. Bottom
panels show a distributions that
are too wide and to narrow to
make the requirement (left and
right respectively).
• Weak lensing
• Shape measurement
• Shape noise
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
23
Stellar Color Accuracy
P2 = Au + Bg + Cr + Di + Ez + F
P1 = A' u + B' g + C' r + D'i + E' z + F'
Principal color definitions from Ivezic (2004)
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
24
Stellar Color Accuracy
Principal color distributions for
stars from the Galfast model
relative to the mean principal
colors in the SDSS photometric
system
• Calibration
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
25
Stellar Color Distribution
Color distribution of main sequence
(forward hatched) and white dwarf
(backward hatched) stars for high
galactic latitude fields.
• Calibration
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
26
Solar System Model Accuracy
Distribution of interpolation error for
each of the populations in the solar
system model. The mean
is always below 1”. The requirement is
1”.
• Testing the moving object pipelines
FINAL DESIGN REVIEW | TUCSON, AZ | OCTOBER 21-25, 2013
27