The Gaia Exoplanet Discovery Potential A. Sozzetti INAF – Osservatorio Astrofisico di Torino

GREAT Synergies on the Horizon Torino, 06 Nov 2012

The Gaia The Gaia Exoplanet Discovery Exoplanet Discovery

PotentialPotentialA. SozzettiA. Sozzetti

INAF – Osservatorio Astrofisico di TorinoINAF – Osservatorio Astrofisico di Torino

D. Segransan D. Segransan (Obs. Geneva),(Obs. Geneva), D. Sosnowska D. Sosnowska (Obs. Geneva),(Obs. Geneva), N. Rambaux N. Rambaux (IMCCE),(IMCCE), G. Tinetti G. Tinetti (UCL),(UCL), M.G. Lattanzi M.G. Lattanzi (INAF-OATo),(INAF-OATo), R. Morbidelli R. Morbidelli (INAF-OATo),(INAF-OATo), G. G.

Micela Micela (INAF-OAPa),(INAF-OAPa), P. Giacobbe P. Giacobbe (U. Trieste)(U. Trieste)


Astrometry: Astrometry: BlundersBlunders

1940’s: 1940’s: Strand, Strand, Reuyl & Holmberg (61 Cyg, 70 Oph)1960’s: Lippincott,Hershey (1960’s: Lippincott,Hershey (Lalande 21185))1960’s-80’s: Van de Kamp (Barnard’s Star)1960’s-80’s: Van de Kamp (Barnard’s Star)1980’s: Gatewood (Lalande 21185, again)1980’s: Gatewood (Lalande 21185, again)2001: Han et al. (some 20 RV planets)2001: Han et al. (some 20 RV planets)2009: Pravdo & Shaklan (VB10b)2009: Pravdo & Shaklan (VB10b)

mas-precision astrometry is usually not enough for planet detectionmas-precision astrometry is usually not enough for planet detection


μμas Astrometry: Challengesas Astrometry: Challenges

Like RV, it faces: Like RV, it faces:

- technological challenges (achievable precision, ground vs. space, technological challenges (achievable precision, ground vs. space, instrument configuration, choice of wavelength, calibrations, etc.)instrument configuration, choice of wavelength, calibrations, etc.)

- astrophysical challenges (noise sources characterization)astrophysical challenges (noise sources characterization)

- data modeling challenges (orbital fits)data modeling challenges (orbital fits)See e.g. Sozzetti (2005, 2010)

Sozzetti 2005Sozzetti 2005

Two main directions for improvement:

1) Monolithic configurations (optical)2) Diluted configurations (optical/near IR)

- Narrow-angle, relative astrometry: both from - Narrow-angle, relative astrometry: both from the ground and in space (VLTI/PRIMA,???)the ground and in space (VLTI/PRIMA,???)

- Global astrometry: only in space (Gaia)- Global astrometry: only in space (Gaia)


HipparcosHipparcos

1. 6 < G < 12: bright-star regime (calibration errors, CCD saturation)1. 6 < G < 12: bright-star regime (calibration errors, CCD saturation)2. 12 < G < 20: photon-noise regime (sky-background and electronic noise at G ~ 20 mag)2. 12 < G < 20: photon-noise regime (sky-background and electronic noise at G ~ 20 mag)

Gaia Astrometric PrecisionGaia Astrometric Precision


Fitting Planetary Systems OrbitsFitting Planetary Systems Orbits

• Highly non-linear fitting procedures, with a large Highly non-linear fitting procedures, with a large number of model parameters (at a minimum, number of model parameters (at a minimum, NNpp=5+7*n=5+7*npl, pl, not counting references)not counting references)

• Redundancy requirement: NRedundancy requirement: Nobsobs >> N >> Npp

• Global searches (grids, Fourier decomposition, genetic Global searches (grids, Fourier decomposition, genetic algorithms, Bayesian inference +MCMC) must be algorithms, Bayesian inference +MCMC) must be coupled to local minimization procedures (e.g., L-M)coupled to local minimization procedures (e.g., L-M)

• For strongly interacting systems, dynamical fits using For strongly interacting systems, dynamical fits using N-body codes will be requiredN-body codes will be required


Assessing DetectionsAssessing Detections

• Errors on orbital parameters: covariance matrix vs. Errors on orbital parameters: covariance matrix vs. 2 2

surface mapping vs. bootstrapping proceduressurface mapping vs. bootstrapping procedures

• Confidence in an n-component orbital solution: FAPs, F-Confidence in an n-component orbital solution: FAPs, F-tests, MLR tests, statistical properties of the errors on the tests, MLR tests, statistical properties of the errors on the model parameters, others?model parameters, others?

• Importance of consistency checks between different solution Importance of consistency checks between different solution algorithmsalgorithms

• Memento lessons learned from RV surveys, with Memento lessons learned from RV surveys, with disagreement on orbital solution details, and sometime disagreement on orbital solution details, and sometime number of planets!!number of planets!!


Exoplanets in the Exoplanets in the Gaia DPAC PipelineGaia DPAC Pipeline


MCMC Algorithm:Implementation in WP100/210

• Partial linearization of the least square problem

• Iterative period search provides seeds for MCMC procedure (a DE-MCMC algorithm upgrade is in the works)

• The MCMC chain is run on the non linear parameters (e, P, T0), drawing from Gaussian independent distributions.

• Each step of the chain drives a linear LSQ solver (SVD).

• Multiple chains are run (one is used for inference), convergence checked with Gelman-Rubin test.

• The set of parameters with the highest likelihood is the seed for a local non-linear minimization using LM.


Genetic Algorithm:Genetic Algorithm:Implementation in WP220Implementation in WP220


Orbital Stability AnalysisOrbital Stability Analysis(WP300)(WP300)

• Simple Hill Stability Criterion (for now)Simple Hill Stability Criterion (for now)

• This is sufficient to double-check dynamically This is sufficient to double-check dynamically pathologic (while potentially correct in a pathologic (while potentially correct in a χχ22 sense!) sense!) solution sets (e.g., orbit-crossing due to solution sets (e.g., orbit-crossing due to ee~1)~1)

• Limited use of N-body integrator (e.g., Burlisch-Limited use of N-body integrator (e.g., Burlisch-Stoer) being testedStoer) being tested


Stellar Sample: Stellar Sample:

- 100 stars- 100 stars

- Bright (G<10)- Bright (G<10)

- Nearby (d<100 pc)- Nearby (d<100 pc)

- 1 M- 1 MSUNSUN

- Two-planet systems: - Two-planet systems: 1/3 non-detectable, 1/3 non-detectable, 1/3 with one, 1/3 with one, 1/3 with two 1/3 with two detectable companionsdetectable companions

Nominal gate scheme adoptedNominal gate scheme adopted

Multiple solutions:Multiple solutions:Readiness TestReadiness Test

Courtesy of F. ArenouCourtesy of F. Arenou


Long-Period SystemLong-Period System


Low S/N systemLow S/N system


Period ErrorsPeriod Errors

Reasonable convergence percentages, with early error behavior following expectations


Gaia - SynergiesGaia - Synergies

Objectives of study within the GREAT RNP/ITNObjectives of study within the GREAT RNP/ITN

• Gaia & spectroscopic characterization observatories (e.g., EChO)Gaia & spectroscopic characterization observatories (e.g., EChO)• Gaia & transit surveys from the ground (e.g., WASP) and in space (CoRoT, Gaia & transit surveys from the ground (e.g., WASP) and in space (CoRoT,

Kepler)Kepler)• Gaia & direct imaging observatories (e.g., SPHERE/PCS) Gaia & direct imaging observatories (e.g., SPHERE/PCS) • Gaia & RV programs (e.g., HARPS(-N), ESPRESSO, CARMENES, and the likes)Gaia & RV programs (e.g., HARPS(-N), ESPRESSO, CARMENES, and the likes)• Gaia & ground-based and space-borne astrometryGaia & ground-based and space-borne astrometry


Focusing on M DwarfsFocusing on M Dwarfs(Sozzetti et al., A&A submitted)(Sozzetti et al., A&A submitted)

Cool, nearby M dwarfs within a few tens of parsecs from the Sun are now the Cool, nearby M dwarfs within a few tens of parsecs from the Sun are now the focus of dedicated experiments in the realm of exoplanets astrophysics. focus of dedicated experiments in the realm of exoplanets astrophysics.

WHY?WHY?

* Shifting theoretical paradigms in light of new observations * Shifting theoretical paradigms in light of new observations

* Improved understanding of the observational opportunities for planet * Improved understanding of the observational opportunities for planet detection and characterization provided by this sample. detection and characterization provided by this sample.

WHAT CAN Gaia CONTRIBUTE?WHAT CAN Gaia CONTRIBUTE?

* Gaia, in its all-sky survey, will deliver precision astrometry for a magnitude-limited sample of M dwarfs, providing an inventory of cool nearby stars with a much higher degree of completeness (particularly for late sub-types) with respect to currently available catalogs.


1) Basis: The Casertano et al. (2008, A&A, 482, 699) simulation setup1) Basis: The Casertano et al. (2008, A&A, 482, 699) simulation setup

2) Updated Gaia scanning law. 2) Updated Gaia scanning law. T=T=5 yr nominal mission duration5 yr nominal mission duration

3) Up-to-date error model as a function of Gaia 3) Up-to-date error model as a function of Gaia GG-band mag (inclusive -band mag (inclusive of gate scheme for 20% of bright [of gate scheme for 20% of bright [GG< 12] stars). Single-measurement < 12] stars). Single-measurement errors are typically errors are typically σσmm~100 µas~100 µas

4) Actual list of targets: 3150 M dwarfs (0.09-0.6 M4) Actual list of targets: 3150 M dwarfs (0.09-0.6 MSUNSUN) within 33 pc ) within 33 pc from the Sun from the LSPM-North Catalog (Lépine 2005, AJ, 130, from the Sun from the LSPM-North Catalog (Lépine 2005, AJ, 130, 1680), with average 1680), with average GG~14.0 mag and M~14.0 mag and M**~0.3 M~0.3 MSUNSUN

5) 1 M5) 1 MJUPJUP companions, orbital periods companions, orbital periods P < P < 33T, T, moderate eccentricities moderate eccentricities ((ee<0.6), all other orbital elements uniformly distributed within their <0.6), all other orbital elements uniformly distributed within their respective rangesrespective ranges

SIMULATION SCHEMESIMULATION SCHEME


Determining Determining Giant Planets Orbits (1)Giant Planets Orbits (1)

PP


Determining Determining Giant Planets Orbits (2)Giant Planets Orbits (2)

MM


How Many Giants?How Many Giants?

• Present-day estimates from RV surveys imply fp~3-4% (within 3 AU)

• Gaia could identify ~100 giant planets around this sample, an order-of-magnitude increase

• For approximately 50% of them, accurate orbit reconstruction will be possible

• The sample size is such that fp will be put on much more solid statistical grounds

• Very important synergy with present (e.g., HARPS@ESO), starting (e.g., HARPS-N@TNG), and upcoming (ESPRESSO@VLT) RV surveys


Finding Nearby TransitingFinding Nearby TransitingWide Separation Giants…Wide Separation Giants…

ii

For quasi-edge-on orbits, For quasi-edge-on orbits, i is measured to ~3%i is measured to ~3%

Gaia may find hundreds Gaia may find hundreds of giant planets of giant planets

around M dwarfs. around M dwarfs. Some may be transiting!Some may be transiting!

Obvious synergy with Obvious synergy with ground-based transit work!ground-based transit work!


Re-Calibrating the Hosts…Re-Calibrating the Hosts…

• ALL parallaxes of this M dwarf sample released formally around ALL parallaxes of this M dwarf sample released formally around mid-2016mid-2016

• For a typical target with V~14 at d~ 20 pc, expect σ(π)/π<0.1%For a typical target with V~14 at d~ 20 pc, expect σ(π)/π<0.1%

• Re-calibrate absolute luminositiesRe-calibrate absolute luminosities

• Derive trigonometric gravities to ~0.03 dexDerive trigonometric gravities to ~0.03 dex

• Re-determine the stellar radii to <3%Re-determine the stellar radii to <3%

• Great synergy with ongoing (MEarth), starting (APACHE), and Great synergy with ongoing (MEarth), starting (APACHE), and upcoming (NGTS) ground-based surveysupcoming (NGTS) ground-based surveys


Predicting Their Location…Predicting Their Location…

ρρ θθ

For well-sampled orbits (P<T), For well-sampled orbits (P<T), ΔρΔρ < 0.01 AU/yr and< 0.01 AU/yr and ΔθΔθ < 2 deg/yr< 2 deg/yr(Over an order of magnitude better than HST/FGS predictions for (Over an order of magnitude better than HST/FGS predictions for εε Eri!)Eri!)


Assuming Assuming ΦΦ((ββ) of a Lambert sphere, ) of a Lambert sphere, typically typically ΔΔββ ~ several deg,~ several deg, ΔΦΔΦ((ββ)~0.1 )~0.1

……Their Phases…Their Phases…


……and telling how bright they are!and telling how bright they are!

For 0.3<a<3.0 AU, For 0.3<a<3.0 AU, uncertainties in the uncertainties in the emergent flux will emergent flux will typically be 10-15%typically be 10-15%

Potential synergy Potential synergy with direct imaging,with direct imaging,

reflected light reflected light and atmospheric and atmospheric characterization characterization measurementsmeasurements


The Message The Message (take it home!)(take it home!)

• Providing the largest catalogue of ‘new’ astrometric orbits & masses Providing the largest catalogue of ‘new’ astrometric orbits & masses of extrasolar planets and superbly accurate parallaxes is Gaia’s of extrasolar planets and superbly accurate parallaxes is Gaia’s defining role in the exoplanet arena.defining role in the exoplanet arena.

• The synergies between Gaia and ongoing and planned exoplanet The synergies between Gaia and ongoing and planned exoplanet detection and (atmospheric) characterization programs from the detection and (atmospheric) characterization programs from the ground and in space are potentially hugeground and in space are potentially huge

• This was a snippet of the Gaia potential on the sample of nearby low-This was a snippet of the Gaia potential on the sample of nearby low-mass M dwarfsmass M dwarfs

• Gaia’s ‘first’ release: L+22m (Summer 2015)Gaia’s ‘first’ release: L+22m (Summer 2015)

• Gaia’s ‘first’ major release: L+28m (Beginning of 2016)Gaia’s ‘first’ major release: L+28m (Beginning of 2016)

• Gaia’s ‘first’ complete catalog release: L+40m (End of 2016)Gaia’s ‘first’ complete catalog release: L+40m (End of 2016)

The Gaia Exoplanet Discovery Potential A. Sozzetti INAF – Osservatorio Astrofisico di Torino

Documents

Transcript of The Gaia Exoplanet Discovery Potential A. Sozzetti INAF – Osservatorio Astrofisico di Torino