P. Tanga GREAT - EWASS Athens 2016

March 16, 2017

GAIA AND THE POPULATION: A REVOLUTION ON EARTH, COMING FROM SPACE

P. Tanga Laboratoire , Observatoire de la Côte d’Azur, France P. Tanga – and the Solar System 2 Gaia ID card

• It is a ESA cornerstone scientific mission: • 5 years of observations – ESA: building, operation (nominal mission started in July 2014) – Scientific community: data reduction • Positioned around L2 – No instrument PI, no proprietary period • automated selection of sources , on • Astrometry: input catalog – 109 stars, V<20 • Self-monitored, onboard metrology – 25 µas at V~15 • Heritage of Hipparcos – uniform sky coverage (70-100 obs./ source) • Physical observations: • Spectro- • Radial velocities, hi-res spectra (V<16.5)

3 webinar P. Tanga – Gaia and the Solar System 3 Gaia in the history of astrometry

5 webinar P. Tanga – Gaia and the Solar System 4 ESA/Gaia/DPAC/Airbus DS

webinar Optics Two SiC primary mirrors 1.45 × 0.50 m2, F = 35 m Rotation axis (6 h)

106°.5

Combined focal plane (CCDs) M4: beam SiC toroidal combiner Figure courtesy EADS- structure webinar P. Tanga – Gaia and the Solar System 6 ESA/Gaia/DPAC/Airbus DS

webinar Sunshield deployment test – Oct. 2011 P. Tanga – Gaia and the Solar System 8

December 19, 2013

webinar P. Tanga – Gaia and the Solar System 9 The scanning law

Rotation axis movement Scan path in 4 days

Scan path 4 days 4 rotations/ Spin axis Spin axis trajectory 4 days

45°

Sun trajectory, 4 months Sun

Spin axis trajectory, 4 months

webinar P. Tanga – Gaia and the Solar System 10 Focal plane 1 pixel 60 x 180 mas 106 CCDs (4.5 x 2 kpix) = 1 Gpixel SM1-2 AF1 - 9 BP RP

RVS

420 mm 0.69° WFS

WFS

BAM

BAM sec FOV1 0s 10.6 15.5 30.1 49.5 56.3 64.1 sec FOV2 courtesy F. Mignard, OCA 0s 5.8 10.7 25.3 44.7 51.5 59.3

webinar P. Tanga – Gaia and the Solar System 11 The Astrometric Global Iterative Solution (AGIS)

• Simultaneous all-sky solution of positions, parallaxes, proper motions, physical parameters (e.g. color) and instrument calibration parameters

• Link to ICRS by VLBI sources directly observed by Gaia —> All aspects of the data reduction are deeply connected

webinar P. Tanga – Gaia and the Solar System 12 Gaia as Micro-meteorite detector

Corrective torques applied to recover spacecraft attitude after a strike (Perseid activity peak).

webinar P. Tanga – Gaia and the Solar System 13 The strengths of Gaia

• Astrometry: • Accurately measure relative positions at large angles —> no zonal error • Measure time instead of positions (and control very accurately the transformation) • Average the signal over several 10.000s pixels (good for photometry too)

• Spectro-photometry • provide colors and low-resolution spectra (—> classification) for all sources

• Radial velocities • Add the 3rd dimension of motions in the Galaxy

+ automated source detection on board

webinar P. Tanga – Gaia and the Solar System 14 Gaia does not produce images

~0.38 as • (Except for testing in a special mode) • Operates in TDI mode (line period 1 ms) • Small windows read from the CCD around the sources

• 2D windows are binned across scan —> 1 D data sample ~1 as (typ. 6 pixels) • on board lossless compression

• Fundamental measurement: crossing time of a fiducial line on the CCD by the brightness peak of a source

webinar P. Tanga – Gaia and the Solar System 15

Gaia produces unusual astrometry

• In the along scan direction the typical accuracy is a small fraction of a pixel (~1 mas) • Across scan it is of the order of a pixel (~100 mas)

~100 mas <1 mas

Dec

RA

• When rotated to (RA, dec) this results in highly correlated uncertainties • This is not (very) relevant for stars • It is fundamental for ! Pay attention to the covariance matrix

webinar P. Tanga – Gaia and the Solar System 16 The Data Processing and Analysis Consortium - since 2007

webinar P. Tanga – Gaia and the Solar System 17

Gaia data - when?

GDR1: star positions + Tycho-Gaia Astrometric Solution

40 mas

GDR2: star positions + parallaxes + proper motions asteroid astrometry

10 mas Final release

2 mas

sept. 2016 Apr 2018 2022

webinar P. Tanga – Gaia and the Solar System 18 GDR1: brightness accuracy

This plot shows the distribution of the estimated uncertainties on the weighted mean G-band photometry as a function of . The colours indicate the density of data points, from low (red) to high (blue) on a logarithmic scale. Credits: ESA/Gaia/DPAC

webinar P. Tanga – Gaia and the Solar System 19 GDR1: parallax accuracy (TGAS)

webinar P. Tanga – Gaia and the Solar System 20 GDR1: position errors

webinar P. Tanga – Gaia and the Solar System 21 Gaia Archive

https://gea.esac.esa.int/archive/

webinar P. Tanga – Gaia and the Solar System 22

The Solar System and Gaia

webinar P. Tanga – Gaia and the Solar System 23

What Gaia observes

• What Gaia observes = all small objects at V < 20.5 –350.000 asteroids( >700.000 known today) –, TNOs –small planetary satellites

• Why we are interested –small bodies record the history of the Solar System –very poorly known properties: • a few 1000s spectra, 10s , 100k sizes, ~400 shapes

webinar P. Tanga – Gaia and the Solar System 24

Where Gaia observes

unobservable

~70 observations per object

• Discovery space: • Low elongations (~45-60°) • Inner Earth Objects (~unknown) • Other NEOs unobservable • Some MBAs

webinar P. Tanga – Gaia and the Solar System 25 What Gaia observes: how big

Kuiper Belt NEA Jupiter Trojans MBO

webinar P. Tanga – Gaia and the Solar System 26

Gaia single-epoch astrometric accuracyJ. Desmars et al.: Statistical and numerical study of asteroid orbital uncertainty Table 4. Statistics of di↵erent measurements provided by the MPC database for numbered and unnumbered asteroids. final attitude and calibration, single FoV (9 positions) transit, point-like source Code Type Number Percentage Timespan C CCD1 mas 88 546 921 94.12% 1986–2012 S/s Space observation 4 006 572 4.26% 1994–2011 HST 3544 0.00% 1994–2010 Spitzer 114 0.00% 2004–2004 WISE 4 002 914 4.25% 2010–2011 A Observations from B1950.0 converted to J2000.0 647 690 0.69% 1802–1999 c Corrected without republication CCD observation 462 065 0.49% 1991–2007 P Photographic 352 449 0.37% 1898–2012 T Meridian or transit circle0.1 mas 26 968 0.03% 1984–2005 X/x Discovery observation 16 741 0.02% 1845–2010 M Micrometer 12 081 0.01% 1845–1954 HHipparcos geocentric observation 5494 0.01% 1989–1993 R Radar observation 2901 0.00% 1968–2012 E derived observation 1737 0.00% 1961–2012 V Roving observer observation 388 0.00% 2000–2012 n Mini-normal place derived from averaging observations from video frames 105 0.00% 2009–2012 e Encoder 16 0.00% 1993–1995

Table 5. Accuracy of measurements for numbered asteroids from the AstDyS2 database.

Type Name Number of Percentage of Accuracy measurement accepted measurement C CCD 79 569 190 99.49% 0.388 arcsec S Wise 1 526 466 99.86% 0.583 arcsec random errors S 867 96.54% 0.585 arcsec S Spitzer 48 33.33% 1.673 arcsec + systematic A B1950 to J2000 632 428 82.17% 1.170 arcsec cObservations Corrected CCD observationin the AstDys 423 792 service (Univ. 99.70% Pisa) 0.507 arcsec P Photographic 346 947 93.38% 1.084 arcsec T Meridian/transit circle 26 968 100.00% 0.250 arcsec M Micrometer 12 081 94.56% 1.742 arcsec X Discovery observation 9520webinar 0.81% 0.898 arcsec HHipparcos observation 5494 100.00% 0.148 arcsec E Occultations 1736 100.00% 0.085 arcsec R Ranging 612 96.41% 3.325 km 1 R Doppler 432 99.07% 6.660 km s V Roving observer 372 49.73% 0.822 arcsec e Encoder 16 100.00% 0.558 arcsec n video frame 12 100.00% 0.319 arcsec the residual (O–C) of each measurement. All the observations asteroid detection and the mission has detected 157 000 aster- of all numbered asteroids9 have been compiled and the ac- oids, including more than 500 NEO (Mainzer et al. 2011). curacy of each measurement has been estimated by comput- ing a weighted10 root mean square of all the residuals of the measurement. 4.2. Impact of astrometric measurements on orbit uncertainty – the Apophis example Table 5 provides the number, the percentage of accepted measurement, and the estimated accuracy for each kind of ob- In order to quantify the impact of a type of measurement on the servation. After radar measurements, observations derived from orbital uncertainty of an asteroid, we specifically dealt with the occultations and Hipparcos geocentric observations appear to asteroid Apophis. This particular object belongs to the PHA fam- be the most accurate (about 0.08–0.15 arcsec). Recent mea- ily and is known to be the most threatening object of the last surements such as CCD observations or meridian circle ob- decade since it is the only asteroid that reached level 4 of the servations are accurate to between 0.2–0.4 arcsec. Older mea- Torino scale for potential impact with the Earth in April 2029 surements, such as photographic or micrometric measures, are (Sansaturio & Arratia 2008). Since new observations (optical less accurate (1.0–1.8 arcsec). Surprisingly, observations from and radar) ruled out every possibility of impact in 2029, this spacecraft (HST, Spitzer or WISE) are not any more accurate threat turned out to be a close encounter within 5.64 R of the than ground-based data. Nevertheless, WISE was designed for Earth. But this close encounter is so deep that the asteroid will move from the Aten family orbit to the Apollo family orbit. Besides, its orbit will become chaotic and new possibilities of 9 As from October 5, 2012, there are 337 008 numbered asteroids in collision with the Earth after 2029 will appear. The most ac- the AstDyS database. cepted collision date is in 2036 for which the risk was estimated, 10 AstDys provides a weight for each observation according to its at the date of the last observations available in 2012, with a prob- accuracy. ability of 10 6. This date is also important because of the size ⇠ A32, page 7 of 10 P. Tanga – Gaia and the Solar System 27 Photometric accuracy

(courtesy D. Evans)

webinar P. Tanga – Gaia and the Solar System 28 Low-resolution spectroscopy

SMASS

= ~25 bands

Gaia SNR @ V=17, per observation M. Delbo, OCA

webinar P. Tanga – Gaia and the Solar System 29 Science goals for the end of the mission

• Complete the sample (discoveries) • Orbits : X 100 improvement • Precession: Gen. Relativity tests

Astrometry • 100 masses from close encounters

• Diameter ~1000 asteroids resolution

• Composition, taxonomy • Shape, pole, rotation period spectro- photometry

webinar P. Tanga – Gaia and the Solar System 30 From data reduction …to scientific exploitation

• Complete the sample (discoveries) Ground-based follow-up • Orbits : X 100 improvement Search for binaries Asteroid occultations • Precession: Gen. Relativity tests Re-processing of “old” Astrometry • 100 masses from close encounters } astrometry • Diameter ~1000 asteroids àmore sizes, masses àdensities

resolution } • Composition, taxonomy Asteroid families Differentiation signatures • Shape, pole, rotation period Collisional evolution spectro- photometry } webinar P. Tanga – Gaia and the Solar System 31 Our knowledge – before and after Gaia Property today Gaia

astrometry ~ 0"5 0"005 ephemeris precision 50-200 mas > 20 times better shape, pole 500 ~100,000 rotation period 4000 ~100,000 satellites ~ 50 (MBA) 1000s ? spectral type ~ 1000 ~200,000

, σ < 50% ~ 50 150 size 100,000 ~1000

webinar P. Tanga – Gaia and the Solar System 32 Solar System in DPAC Coordination Unit 4

Manager: D. Pourbaix (Univ. Brussels) Deputy: P. Tanga (OCA, France)

DU450 Management DU451 Auxiliary data DU452 Identification of known objects DU453 CCD processing DU454 Astrometric reduction DU455 Object threading DU456 Orbital inversion DU457 Global Effects on Dynamics DU458 Physical parameters DU459 Ground-based observations DU460 Simulation

webinar P. Tanga – Gaia and the Solar System 33

CU4: two pipelines for the science goals

Daily processing Processing of « new » asteroids ground-based alert Per-object, on 48h time frames network Preliminary astrometry (OGA1) Preliminary calibration

Long–term processing All sources epoch astrometry Best calibrations, best astrometric model refined orbits Take into account shape and motion dynamical and physical Devoted to obtain the best possible final parameters output of the mission

webinar

From the Gaia-FUN-SSO web site (the red contours are computed by the SSO-ST P. Tanga – Gaia and the Solar System 34 pipeline and show the search area on the sky, on three different dates):

Diffusion of asteroid alerts

From the Gaia-FUN-SSO web site

webinar http://gaiafunsso.imcce.fr

P. Tanga – Gaia and the Solar System 35 First confirmation of an asteroid alert • Orbit computation from Gaia and from the ground (OHP, France) • Observations on Dec. 29 (Gaia - 4 transits) and Jan. 3-4 (OHP, 2 nights) -3 • resulting uncertainty �a~10

g0T015 - 2017 AD17

Gaia Fit residuals OHP 0.4

0.2

0 O-C DEC (arcsec) -0.2

-0.4

-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 O-C RA (arcsec)

webinar P. Tanga – Gaia and the Solar System 36 Looking forward to long-term processing: asteroid orbits by Gaia ONLY - accuracy

−6 10 ASTDyS 5 years 10 years −7 10

−8 10 (au) a σ −9 10

−10 10

−11 10 1 2 3 4 5 6 a (au) F. Spoto, F. Mignard, P. Tanga, OCA

webinar P. Tanga – Gaia and the Solar System 37

But we don’t have asteroid astrometry by Gaia, yet!

So, what can we do?

Exploit the stars in GDR1 !

webinar P. Tanga – Gaia and the Solar System 38 Re-calibration of ground-based astrometry

• Asteroid observations from the Ground-Based Optical Tracking of Gaia (GBOT) • VST - Paranal • Liverpool robotic - Canary Isl. • Several 10s asteroids observed each night • Data reduction with PPMXL and Gaia DR1 • Typical sequence: 10 images over ~15 minutes, once

webinar P. Tanga – Gaia and the Solar System 39

1132 - GBOT PPMXL residuals

1: J13 60 1: 309 PPMXL calibration 40

20

0

-20

-40

-60 O-C DEC (mas)

-80

-100

-120 -100 O-C DEC (mas) 50

-140

-160

-180

-400 -380 -360 -340 -320 -300 -280 -260 -240 -220 -200 -180 -160 -140 -120 -100 -80 -60 -40 O-C RA (mas) -300 O-C RA (mas) -100

webinar P. Tanga – Gaia and the Solar System 40

1132 - GBOT GDR1 residuals

1: J13 1: 309 30 Gaia GDR1 calibration

25

20

15

10

5

0

-5 O-C DEC (mas)

-10

-15

-20

-25 -20 O-C DEC (mas) 20

-30

-35

-20 -15 -10 -5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 O-C RA (mas) -20 O-C RA (mas) +20

webinar A&A proofs:Amanuscript&A proofs: no.manuscript 20161216_Gaia_asterastrom no. 20161216_Gaia_asterastrom

2. Astrometry2. Astrometry of moving of objects: moving error objects: budget error budgetprocedure.(Milaniprocedure.(Milani and Gronchi and 2014). Gronchi Figure 2014). 1 shows Figure that 1 few shows that few observations are discarded by the rejection procedure (marked Following the previous section: what GBOT provides withobservations DR1 are discarded by the rejection procedure (marked Following the previous section: what GBOT provides with DR1 with an “X”),with while an “X”), the stars while are the the stars residuals are the of residuals the observa- of the observa- and how it comparesand how it to compares the previous to the situation. previous situation. tions used intions the fit. usedMissing in the fit. partMissing on how part much on the how orbits much are the orbits are close. close. 3. NEOs and risk rating 3. NEOs and risk rating 0.3 0.3 The first NEAThe discoveredfirst NEA discovered by GBOT byis namely GBOT the is namely asteroid the asteroid 2016 EK85. The object has been observed for the first time 2016 EK85. The object has been observed for the first time 0.2 0.2 the night ofthe the night 9th of of March the 9th 2016 of March by the 2016 VLT by Survey the VLT Tele- Survey Tele- cope (VST).cope Then (VST). it has Then been it re-observed has been re-observed the following the night following night / / 0.1 (2016/03/10)(2016 by the03 10) Liverlpool by the Telescope Liverlpool (LT) Telescope and others (LT) ob- and others0.1 ob- servers. servers. O-C DEC (arcsec) The Minor PlanetThe Minor Center Planet classified Center the classified object as the a NEA object and as a NEAO-C DEC (arcsec) and 0 published thepublished corresponding the corresponding Minor Electronic Planet Circular Electronic Circular0 1 (MPEC)1. After(MPEC) the discovery. After the of discovery a new NEA of a or new if new NEA observa- or if new observa- 2 3 2 3 -0.1 tions are added,tions both are added, NEODyS both in NEODyS Pisa and in Sentry Pisa atand the Sentry JPL at the-0.1 JPL check if thecheck NEA if could the NEA have possiblecould have impacts possible with impacts the Earth with in the Earth in -0.2 -0.1 0 0.1 0.2 -0.2 -0.1 0 0.1 0.2 100 years from now. O-C RA x COS(DEC) (arcsec) 100 years from now. O-C RA x COS(DEC) (arcsec) It turned outIt that turned the out object that had the possible object had impacts possible with impacts the with the Earth in 2102Earth and in 2106 2102 with and low 2106 impact with low probabilities, impact probabilities, as Ta- Fig. as 1. Ta-ResidualsFig. in1. Residuals Right Ascension in Right (RA) Ascension and Declination (RA) and (DEC) Declination for (DEC) for ble 1 shows. The risk probability has definitely ruled outthe by GBOT later Gaiathe reduced GBOT Gaia observations. reduced observations. Stars represents Stars the represents observations the observations ble 1 shows. The risk probability has definitely ruled out by later used in the fit, while “X” are the observations discarded by the outlier Mauna KeaMauna observations Kea observations done the night done of the the night 16th of of the March 16th ofused March in the fit, while “X” are the observations discarded by the outlier rejection procedure.rejection procedure. 2016 and published2016 and on published MPEC 2016-F48 on MPEC4. 2016-F484.

Table 1. Date,Table impact 1. Date, probability impact and probability Palermo and Scale Palermo of possible Scale im- of possibleOnce im- we haveOnce the we orbit, have we the can orbit, try to we look can for try possible to look for im- possible im- pacts with thepacts Earth with for the NEA Earth 2016 forEK NEA85. The 2016 impactEK85. table The impact appeared tablepacts appeared with thepacts Earth with in the the Earth next 100 in the years, next as 100 it hasyears, been as doneit has been done in NEODySin on NEODyS the night onof the the 11th night of of March the 11th 2016. of March It has been 2016. com- It has beenwhen com- the objectwhen appeared the object for appeared the first for time. the We first compute time. We multi- compute multi- puted using 48puted optical using observations 48 optical observations from 2016/03 from/09 to 2016 2016/03/03/09/11 to by 2016ple/03/11 solutions, by ple we solutions, analyze we each analyze close encounter each close and encounter each return and each return the OrbFit softwarethe OrbFit version software 5.0. All version these 5.0. data All are these consistent data are with consistent the that with could the leadthat to could a possible lead to impact a possible (LoV impact paper). (LoV It is not paper). surpris- It is not surpris- ones publishedones by published the JPL. by the JPL. ingly that weingly do not that find we doany not risk find tables any using risk tables the observations using the observations P. Tanga – Gaia and the asteroids Lisbon Observatory - February 6, 2017 41 reduced withreduced the Gaia with catalog. the Gaia A change catalog. in A the change catalog in changes the catalog changes DateDate IP PS IP PS the direction of the Line of Variations (LoV) (Lov 2005). The 8 8 the direction of the Line of Variations (LoV) (Lov 2005). The 2102/02/22.2962102/02 1/.22.29624 10 1.248.54 10 8.54 LoV is basedLoV on is the based weak on direction, the weak which direction, in this which case inis along this case is along 8 8 2102/02/22.5492102/02 5/.22.54957 10 Potential “impactor”NEA 2016EK85 5.577.89 10 7.89 the semi-majorthe semi-major axis. A changed axis. in A the changed catalog in moves the catalog the semi- moves the semi- 8 8 2106/02/22.0422106/02 6/.22.04282 10 6.827.82 10 7.82 major axis andmajor rules axis out and all rules the possible out all the impacts possible with impacts the Earth. with the Earth. 7 7 2106/02/22.3112106/02 4/.22.31162 10 4.626.99 10 6.99 6 6 PPMXL GDR1 2106/02/22.5292106/02 1/.22.52920 10 1.206.57 10 6.57 7 7 x 10 x 10 7 7 7 7 3 3 x 10 x 10 2106/02/22.6052106/02 3/.22.60519 10 • 3.197.14 10 7.14 3 3 6 Discovered by GBOT (8 observations 6 2106/02/22.6352106/02 8/.22.63593 10 8.935.70 10 5.70 2.5 2.5 2.5 by VST, 20 by LT, March 9-10, 2016) 2.5 • On the impact risk list with low 2 2 2 2 1.5 1.5 1.5 1.5 Asteroid 2016AsteroidEK85 is 2016 an interestingEK85 is an case interestingprobability for Feb. 22, 2102 to show case the to impact show the impact of GDR1 onof the GDR1 astrometry on the of astrometry small objects. of small• WeNew observations from Mauna Kea objects. use the entire We use set the entire set 1 1 1 1 of GBOT observations,of GBOT observations, that is 28 observations that is 28 observations covering 2 nights covering 2 nights (March 16, 2016) rule out impact 0.5 0.5 0.5 0.5 Opik (km) (2016/03/09(2016 and/ 201603/09/03 and/10). 2016 The/03 sample/10). The contains sample 8 observa- contains 8 observa- Opik (km) Opik (km) Opik (km) ζ ζ ζ 0 0 ζ tions from thetions VST from and the 20 VST observations and 20 observations from• If GDR1 were available, the object the LT. from We reduce the LT. We reduce Earth Earth 0 Earth 0 Earth the observationsthe observations using the GDR1 using catalog the GDR1 (ref.?).would have never been on the risk list catalog Then (ref.?). we com- Then we com- −0.5 −0.5 −0.5 −0.5 pare the orbits obtained with the Gaia reduction with the orbit pare the orbits obtained with the Gaia reduction with the orbit −1 −1 −1 −1 now availablenow and available the one and computed the one using computed only GBOT using only observa- GBOT observa- tions withouttions Gaia without reduction. Gaia reduction. −1.5 −1.5 −1.5 −1.5 The computationThe computation of the orbit of requires the orbit a requires debiasing a and debiasing and −2 −2 −2 −2 −1 0 −1 1 0 1 −1 0 −1 1 0 1 ξ Opik (km) 7 7 ξ Opik (km) ξ Opik (km) 7 7 weighting schemeweighting (Farnocchia scheme (Farnocchia et al. 2015), et but al. the 2015), GDR1 but cata- the GDR1 cata- x 10 x 10 x 10 ξ Opik (km) x 10 log is so newlog that is so no new debiasing that no or debiasing weighting or scheme weighting is yet scheme avail- is yet avail- Fig. 2. Left: Line ofLeft: Variations Line of in Variations the Target inPlane the ofTarget 2102 Plane with of GBOT 2102 with GBOT able for it. We have had to manually set the weights Do we need Fig. 2. able for it. We have had to manually set the weights Doobservations we need observations as in the MPC as observations in the MPC observationsfile . Right: LoV file . in Right: the TP LoV of in the TP of a table witha observations table with observations and weights? and. A weights? preliminary. A preliminary orbit is 2102 orbit with is GBOT2102 observations with GBOT observationsreduced with reduced the GDR1. with the GDR1. then computedthen using computed the Gauss using method the Gauss and method the final and orbit the is final ob- orbit is ob- tained aftertained a weighted after aleast weighted squares least fit with squares an outlier fit with rejection an outlier rejectionFigure 2 showsFigure the 2 Line shows Of the Variations Line Of using Variations the same using set the of same set of 1 http://www.minorplanetcenter.net/mpec/K16/K16EC2.1 http://www.minorplanetcenter.net/mpec/K16/K16EC2.observationsobservations (GBOT): the (GBOT): left panel the reproduces left panel the reproduces situation the as it situation as it html html was at the beginningwas at the when beginning 2016 EK when85 was 2016 onEK the85 riskwas list, on the while risk list, while 2 http://newton.dm.unipi.it/neodys2/2 http://newton.dm.unipi.it/neodys2/ the right panelthe showsright panel what shows happens what using happens Gaia reduced using Gaia obser- reduced obser- 3 http://neo.jpl.nasa.gov/risks/3 http://neo.jpl.nasa.gov/risks/ vations. It isvations. clear that It is in clear the first that case in the the first LoV case pass the through LoV pass the through the 4 http://www.minorplanetcenter.net/mpec/K16/K16F48.4 http://www.minorplanetcenter.net/mpec/K16/K16F48.Earth, whileEarth, in the while second in on the it second is far enough. on it is The far enough. LoV behaviour The LoV behaviour html html is very similaris very during similar the close during encounter the close of encounter 2106 as well. of 2106 as well.

Article number,Article page number, 2 of 6 page 2 of 6 P. Tanga – Gaia and the Solar System 42 The revolution in astrometry

• …does not come for free! • Tools must be ready to handle accuracy ~100 X better • An appropriate, careful weighting of the observations is necessary

• A factor 10 X improvement is accessible with GDR1 • More to come…!

webinar P. Tanga – Gaia and the asteroids Lisbon Observatory - February 6, 2017 43 Stellar occultations: potential with Gaia

• predictions based on GDR1 available since end 09/2015

The duration of the can be transformed into a chord length. Typ. AL accuracy of the occultation ~0.1-0.2 s = 1 - 3 km P. Tanga – Gaia and the Solar System 44 Occultations: science case 90 Antiope • Determine asteroid sizes, shapes – discriminate among spin pole solutions – calibrate indirect size determination methods (thermo-physical modeling) – complement photometric inversion (”KOALA”) – the only efficient method for TNO size

determination Pluto, Elliot & Young 1992 • Measure binary systems – separations, primary/secondary size —> mass, density (!!) • Detect thin atmospheres – on Pluto, large satellites (Titan), TNOs

webinar P. Tanga – Gaia and the asteroids Lisbon Observatory - February 6, 2017 45 Pluto and Charon from daily astrometric calibration (~70 mas)

from the ground - 19 years A&A 570, A86 (2014) 0,3

Credits ESA/Gaia/DPAC/F. Mignard (OCA) 0,2

0,1 ) '

' 0,0 (

δ ∆ -0,1

-0,2

raw -0,3 0 45 90 135 180 225 270 315 360

0,3

0,2

0,1 ) '

' 0,0 (

δ ∆ -0,1

-0,2

corrected photocenter-0,3 0 45 90 135 180 225 270 315 360

Fig.Benedetti-Rossi 8. Pluto ephemeris et al. o2014↵sets before the photocenter correction (top) and after (bottom). O↵sets for right ascension are on the left, declination o↵sets on the right. motion around ~100 mas photocenter wobbling ~10 mas variations linked to albedo changes ~?~ e↵ective wavelengths of our telescope/filter sets. We partially solved this problem by simultaneously fitting k from the observa- Fig. 9. Dispersion of Pluto’s ephemeris o↵sets before (top) and after the di↵erential chromatic refraction (middle) and Pluto photocenter correc- tions, in the process of determining the photocenter o↵sets xphot (see details in Appendix A). After some tests, we found that the tions (bottom). Histograms of right ascension are on the left and of decli- nation are on the right. Ephemeris positions refer to the DE421+plu021 best fit for k comes from using all the data together. Separating ephemerides from JPL. We note, as explained in the two previous sec- the observations by filter (clear, I, R, ESO-R, V) gave the same tions, that the middle and bottom panels do not need to be revisited un- results within 3, but with larger 1 errors (about 0.04; 0.005 for der a change of ephemeris, given that our corrections are not dependent clear). We also grouped the observations in many epoch bins and on ephemerides positions. tested a number of possibilities, but the fits did not indicate con- clusive variations in k, that could result from the slow rotation of the system and consequently gradual exposition of Pluto’s pole 4. Results during the 19 years of observations. To refine the determination of k and x , we repeated the The improvements in the final positions, measured by the dis- phot ↵ calculations after eliminating outlier points with large position persion of Pluto’s ephemeris o sets after the two corrections minus ephemeris o↵sets, for which a 2 filter was applied. Using described in Sects. 3.2 and 3.3, can be easily seen in the his- ↵ all the remaining observations combined together, we obtained tograms in Fig. 9 and in the cos vs. plot in Fig. 10. The 4 · 4 ↵ the final value for the brightness ratio: k = 0.2106 ( = 0.0014). same improvements in the positions can also be seen if di erent A3 filter was later applied to discard any surviving position ephemerides are used. outliers. The remaining 4557 points constituted the final set of After these corrections, we obtained a total of 4557 posi- Pluto positions of this work. Pluto ephemeris o↵sets before and tions of Pluto. The complete table with positions and other data after the photocenter corrections are shown in Fig. 8. we also is available in electronic form at the CDS. A small sample is pre- noted that because Pluto’s and Charon’s albedos are not well sented in Table 4. The table lists the Julian date of the observa- known, and Pluto’s albedo varies with observed phase, the value tions (UTC), the final right ascension and declination corrected of k is an approximate value and does not remove completely for di↵erential chromatic refraction and photocenter o↵set, the the periodic behavior presented in Fig. 8. However, the value we total position errors, the observed apparent magnitude, the fil- obtained improves the correction for photocenter e↵ect substan- ter, the photocenter o↵sets in right ascension and declination, tially, as can be seen in Figs. 7 and 8. telescope used, and seeing. To retrieve the mixed positions of Pluto and Charon, corrected by di↵erential chromatic refraction, When making the comparison between di↵erent one should subtract the furnished photocenter o↵sets from the ephemerides for the system of Pluto (plu021, plu043, and listed final positions. Apparent magnitudes were computed from ODIN1), the largest values come from the comparison between PSF fits with respect to the UCAC4, with errors of about 0.1 to plu021 and ODIN1 in right ascencion, reaching 25 mas in 0.3 mag. The position errors listed in Table 4 were computed absolute values, and values smaller than 5 mas from comparing with Eq. (8): plu043. For declination, all values are smaller than 8 mas. We also note, from Eq. (4), that the photocenter correction is 2 2 2 dominated by the product of k (0.2106) and the distance Pluto- Perror = 1 + 2 + 000.02 , (8) Charon. Therefore, the use of di↵erent modern ephemerides to q describe the orbits of the system of Pluto would provide results where 1 is the standard deviation of the ephemeris o↵set di↵ering by not more than about 5 mas from the ones presented nightly averages and 2 is the error from the (x,y) measure- here. ments, computed from the Gaussian fits to the image profiles of

A86, page 6 of 12 Example - Prediction of stellar occultations: Pluto

18 July 2016 (~24 h before event)

Pluto’s motion in one day NOMAD1 0688-0855872 star

19July 2016 (occultation night)

Pluto + star

Observatory Valle d’Aosta, Saint Barthélémy, July 19, 2016

Courtesy: B. Sicardy, LESIA, Obs. de Paris Example - Prediction of stellar occultations: Pluto

The July 19, 2016 Pluto occultation our prediction as of early July

shadow center

shadow southern limit Example - Prediction of stellar occultations: Pluto

The July 19, 2016 Pluto occultation, prediction using the GAIA star position (and estimation of its pm), plus the New Horizons-updated ephemeris

shadow Gaia “GDR0” ! northern limit

shadow center

shadow southern limit Example - Prediction of stellar occultations: Pluto

The July 19, 2016 Pluto occultation post-occultation reconstructed path (what really happened)

shadow northern limit

shadow center

shadow southern limit

green dots: sites involved in the campaign (not all got data!) P. Tanga – Gaia and the Solar System 50 Example - Prediction of stellar occultations: Pluto 8 NH flyby

Sicardy et al. (2016) 19 July 2016 Dias Oliveira et al. (2015) 6

4 Sicardy et al. (2003) Pressure at 1215 km (µbar)

2 Yelle & Elliot (1997)

0 1990 2000 2010 2020 Courtesy: B. Sicardy, LESIA, Obs. de Paris s Year (B. Sicardy, presented at DPS 2016)

webinar P. Tanga – Gaia and the Solar System 51 TNO detections

• 30 TNOs (in theory) below the V=20.5 threshold • accuracy per transit probably 2-3 mas (along scan) • example of time distribution:

1.0

0.5

0 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000 3100 3200 3300 3400 Time (JD - J2010.0)

0.5

0 2673 2674 2675 2676 2677 2678 Time (JD - J2010.0)

webinar P. Tanga – Gaia and the Solar System 52

~ size limit of Gaia in the Main Belt

Gaia

practical limit ~1000 objets Today

~100 objets Uncertainty / apparent size at opposition Uncertainty / apparent

At 1.4 AU, 1 km ~1 mas Tanga & Delbo’ 2007

webinar P. Tanga – Gaia and the Solar System 53 Impact of Gaia on stellar occultations

• Predictions possible (in principle) for ~1 billion stars • in practice: limited by delta-mag • Path width = uncertainty at D ~ 5-15 km • Many 10s events observable per night from any site for V < 15 (star)

• New tools to be developed • not a task for humans! • accurate prediction at ~100 m level on the ground • introduction of a probabilistic description of the results

webinar P. Tanga – Gaia and the asteroids Lisbon Observatory - February 6, 2017 54 Occultation astrometry: a new approach

• Is it possible to calculate optimized asteroid orbits from occultations ONLY?

• Well-observed occultations can be very accurate astrometric positions for asteroids • at the level of the stellar astrometry • …but possible before and after Gaia P. Tanga – Gaia and the asteroids Lisbon Observatory - February 6, 2017 55 Dream becoming true: the occultation astrometry of asteroids

15 0:001 Orbit accuracy 14 1e ¡ 4 13 12 )

U 1e ¡ 5 A s ( 11 n a o i t ¾

10 a t s ¡

1e 6 l n u o c i c t 9 a o

t 10 X worse l N u 1e ¡ 7 8 c c O equal accuracy 7 1e ¡ 8 6

1e ¡ 9 5 4 F. Spoto, OCA 2e ¡ 9 5e ¡ 9 1e ¡ 8 2e ¡ 8 5e ¡ 8 All obs: ¾ a (AU) F. Spoto et al.: Asteroid astrometry with Gaia calibration.

ent motion of 10-15 mas/s, this corresponds to an error around 0:08 0.5-1 mas. 0:06 Concerning transverse errors, the distribution is clearly bi- modal. A first peak for small values ( 1mas)˜ corresponds to 0:04 the best observed, multi-chords events.⇠ A second, much more )

s 0:02 spread-out set of values, has a wide, flat maximum in the range a (

P. Tanga – Gaia and the asteroids Lisbon Observatory - February 6, 2017c 20-40mas.˜ This is the same order of the typical apparent radius e 0 56 of the occulting asteroids. The intermediate range (1-20 mas) D C ¡0:02

contain events with a small number of chords. ¡ For this first attempt of exploitation of occultation astrome- O ¡0:04 try, we consider only the transverseOccultation residuals uncertainty, as in general it is larger and dominates the error budget. This is not really a lim- ¡0:06 itation at present, as an accurate use of the timing information is ¡0:08 much more delicate and could require a detailed check of those occultation events that could be a↵ected by timing anomalies. F. Spoto et al.: Asteroid astrometry with Gaia calibration. ¡0:05 0 0:05 0:10 By• beingAsteroid 105 Artemis conservative and using a single uncertainty value, O ¡ C RA £ cos(Dec) (as) we believe that we incorporate most error sources in our bulk ent motion of 10-15 mas/s, this corresponds to an error aroundexploitation of0: occultation08 data,without optimistic assumptions. 0.5-1 mas. In a forthcoming0:06 work we will consider more detailed analysis Concerning transverse errors, the distribution is clearlyon selectedbi- asteroids and events. 50 modal. A first peak for small values ( 1mas)˜ correspondsConcerning to 0:0 the4 occulted stars we apply the following ap- s

⇠ e e

the best observed, multi-chords events. A second, muchproach: more r ) g s 0:02 e a spread-out set of values, has a wide, flat maximum in the range d

( 0 = – We matchc the position of the star to the GDR1, looking T 20-40mas.˜ This is the same order of the typical apparent radius e 0 A for theD position of corresponding sources in a 2 arcsec ra- of the occulting asteroids. The intermediate range (1-20 mas) L dius; weC ¡ then0:02 apply a further check on the consistency of contain events with a small number of chords. ¡ the magnitude (which is in fact redundant, as no ambiguities ¡50 For this first attempt of exploitation of occultation astrome- O are found).¡0: The04 position of the matched source in GDR1 is try, we consider only the transverse uncertainty, as in generaladopted it for the occultation. is larger and dominates the error budget. This is not really a lim- ¡0:06 – If the star is in TGAS, the position is corrected for proper 0 50 100 150 200 250 300 350 itation at present, as an accurate use of the timing information is motion,¡ consistently0:08 with the time delay between the ob- LONG=degrees much more delicate and could require a detailed check of thoseserved event and the catalogue epoch. occultation events that could be a↵ected by timing anomalies.– If a star is not in TGAS,¡0:05 it does0 not have0 a:0 proper5 motion.0:10 In Fig. 7. Upper panel: residuals, with respect to the orbit computed from By being conservative and using a single uncertainty value,this case we consider the di↵erence between the UCAC4 po- occultations, for the astrometry of (105) Artemis. Each symbol repre- O ¡ C RA £ cos(Dec) (as) sents a single occultation event. Asterisks are observations that are used we believe that we incorporate most error sources in our bulksition and GDR1 to compute an approximate proper motion. exploitation of occultation data,without optimistic assumptions. for the final orbital fit. Crosses correspond to measurements that are au- In a forthcoming work we will consider more detailed analysisThis approach has clear limitations for non-TGAS sources tomatically rejected. The circles show the nominal uncertainty (weight) (mainly due to the zonal error in UCAC4) and further tests of of the occultations. Multi-chord events correspond to smaller circles. In on selected asteroids and events. 50 the lower panel, the distribution of the occulted stars in ecliptic coordi- Concerning the occulted stars we apply the followingGDR1 ap- against other astrometric catalogues could provide useful s nates is shown. e

information,e but we consider that this investigation is beyond our

proach: r g

immediatee goals of globally testing the approach on the whole

d 0 – We match the position of the star to the GDR1, lookingset of occultations.= occultations of good quality. As a rule of thumb, we can say that T for the position of corresponding sources in a 2 arcsec ra- A L above 10 astrometric points, uncertainties are not worse than dius; we then apply a further check on the consistency of one order⇠ of magnitude with respect to the solution obtained with 5.2. Orbit determination the magnitude (which is in fact redundant, as no ambiguities ¡50 all the observations (thousands measurements in general). are found). The position of the matched source in GDR1We attempted is an orbital determination for all asteroids that have The performance obtained on a few asteroids, whose orbits adopted for the occultation. an historical record of more than 4 occultations. The observation show better residuals when only occultations are used, is remark- – If the star is in TGAS, the position is corrected for properweights are represented0 50 by the100 transverse150 accuracy,200 250 as explained300 350 able. However, for a similar number of occultations, there is a motion, consistently with the time delay between theabove. ob- Our fitting procedure rejectsLON observationsG=degrees whose resid- considerable spread in the quality of the solution. For instance, served event and the catalogue epoch. uals exceeds among the best performers objects with 10 to 15 occultations are Fig. 7.betterUpper define panel: rejection residuals, with criteria, respect or to simply the orbit put computed a from – If a star is not in TGAS, it does not have a proper motion.reference In here? the weight. No other observations are included found, but also several with just 4 or 5 observed events. For an this case we consider the di↵erence between the UCAC4 po- occultations, for the astrometry of (105) Artemis. Each symbol repre- for the momentsents a single in the occultation orbit fitting event. procedure. Asterisks are observations that areinterpretation used of this evidence, both the occultation quality and sition and GDR1 to compute an approximate proper motion.For eachfor the orbit final we orbital determine fit. Crosses the error correspond on the semi–major to measurements axis thatthe are distribution au- of the observed occultations along the asteroid This approach has clear limitations for non-TGAS sourcesa and usetomatically it as an indicator rejected. The of the circles orbit show accuracy. the nominal In Fig. uncertainty?? we (weight)orbit have to be considered. (mainly due to the zonal error in UCAC4) and further testscompare of of our the results occultations. accuracy, Multi-chord to the accuracy events correspond obtained tofrom smaller fit- circles.We In illustrate two typical situations in the following. The first GDR1 against other astrometric catalogues could provide usefulting all thethe available lower panel, observation the distribution from of the the Minor occulted Planet stars Center: in eclipticone coordi- concerns the asteroid (105) Artemis, having 10 occultations nates is shown. information, but we consider that this investigation is beyondthe our value of a is provided by the AstDys (ref.***) online repos- and closely equalling in performance the occultation orbit to the itory. One should note that this last orbital solution also contains ”all data” orbit (1733 astrometric measurements). The orbit un- immediate goals of globally testing the approach on the whole the contributions of stellar occultations. However, as up to now certainty is 4 10 9 au. An inspection of the residuals of the set of occultations. a no case–by–caseoccultations study of of good occultation quality. As astrometry a rule of was thumb, done, we all can sayoccultation that orbit⇠ (Fig.⇥ 7, upper panel) shows that a few, multi- such measurementsabove 10 were astrometric given a points, weight uncertainties of 200 mas (ref.***),are not worsechords than events are clustered around residual that are smaller than one order⇠ of magnitude with respect to the solution obtained with 5.2. Orbit determination severely penalizing their role. 10 mas. A couple of observations are rejected, as their residu- Fig. 6all shows the observations that properly (thousands weighted measurements occultations, in even general). if als are too large ( 70 mas) with respect to their weight (nominal We attempted an orbital determination for all asteroids thattaken have alone,The can performance provide very obtained reasonable on orbital a few asteroids, solutions. whose Of uncertainty). orbits This anomaly could be due to specific problems of an historical record of more than 4 occultations. The observationcourse thisshow is true better only residuals for objects when having only occultations a higher number are used, of is remark-the occultations (undetected errors in the observation or the inter- weights are represented by the transverse accuracy, as explained able. However, for a similar number of occultations, there is a above. Our fitting procedure rejects observations whose resid- considerable spread in the quality of the solution. For instance, Article number, page 5 of 6 uals exceeds better define rejection criteria, or simply put a among the best performers objects with 10 to 15 occultations are reference here? the weight. No other observations are included found, but also several with just 4 or 5 observed events. For an for the moment in the orbit fitting procedure. interpretation of this evidence, both the occultation quality and For each orbit we determine the error on the semi–major axis the distribution of the observed occultations along the asteroid a and use it as an indicator of the orbit accuracy. In Fig. ?? we orbit have to be considered. compare our results accuracy, to the accuracy obtained from fit- We illustrate two typical situations in the following. The first ting all the available observation from the Minor Planet Center: one concerns the asteroid (105) Artemis, having 10 occultations the value of a is provided by the AstDys (ref.***) online repos- and closely equalling in performance the occultation orbit to the itory. One should note that this last orbital solution also contains ”all data” orbit (1733 astrometric measurements). The orbit un- 9 the contributions of stellar occultations. However, as up to now certainty is a 4 10 au. An inspection of the residuals of the no case–by–case study of occultation astrometry was done, all occultation orbit⇠ (Fig.⇥ 7, upper panel) shows that a few, multi- such measurements were given a weight of 200 mas (ref.***), chords events are clustered around residual that are smaller than severely penalizing their role. 10 mas. A couple of observations are rejected, as their residu- Fig. 6 shows that properly weighted occultations, even if als are too large ( 70 mas) with respect to their weight (nominal taken alone, can provide very reasonable orbital solutions. Of uncertainty). This anomaly could be due to specific problems of course this is true only for objects having a higher number of the occultations (undetected errors in the observation or the inter-

Article number, page 5 of 6 SSO-TT CU4 GAIA-C4-TN-OCA-FM-060-1

P. Tanga – Gaia and the asteroids Lisbon Observatory - February 6, 2017SSO-TT 57 FIGURE 2: Magnitude distribution over the G-magnitude of the 230, 000 predicted transits up CU4 GAIA-C4-TN-OCA-FM-060-1 to G = 20.7 for the second delivery.

SSO-TT CU4Selection for the Gaia Data Release 2 (Apr. 2018)GAIA-C4-TN-OCA-FM-060-1

FIGURE 3: Number of FOV transits for the 13, 000 known planets selected to generate the FIGURE 2: Magnitude distribution over the G-magnitude of the 230, 000 predicted transits up ⇠ Input TransitId lists in January 2017. Only planets with at least 9 transits have been kept in to G = 20.7 for the second delivery. • this selection.Last version (January 8): • no transit loss for the included asteroids Technical Note• asteroids with less than 9 transits are OCA 12 excluded • 13,400 planets (195,000 transits)

FIGURE 4: Distribution of the semi-major axes among the 13, 000 known planets selected ⇠ F. Mignard, OCA to generate the Input TransitId lists in January 2017. Only planets with at least 9 transits have been kept in this selection. One sees that all the broad categories of minor bodies are represented in the selected sample.

FIGURE 3: Number of FOV transits for the 13, 000 known planets selected to generate the ⇠ Input TransitId lists in January 2017. Only planets with at least 9 transits have been kept in this selection.

Technical Note OCA 12

Technical Note OCA 13 P. Tanga – Gaia and the Solar System 58 Conclusions

• Gaia data are delivering the expected resolution • stellar data already interesting for Solar System science • Big potential in asteroid data: sensitivity to shape/satellites, subtle dynamical effects • “old” approaches are being renovated: asteroid occultations is the example of excellence

webinar P. Tanga – Gaia and the Solar System 59 Thank you!

webinar