The Origins Billions Star Survey: Galactic Explorer

The OBSS mission will search for planets and stellar companions by employing two techniques: (1) the astrometric detection of the reflex motion of the parent star, and (2) the photometric detection of planetary transits. While current ground‐based radial velocity searches for extrasolar planets are limited to 8th visual magnitude and thousands of stars, the OBSS astrometric and photometric search will extend to beyond 18th visual magnitude and millions of stars. Using astrometric and photometric techniques, the OBSS in full‐time survey mode should detect at least 28,000 and 1600 planets, respectively. These numbers could be improved by using an OBSS observing cadence that combines a full‐sky survey with special targeted observing modes optimized for planet detection. OBSS will produce a census of the existence and physical characteristics of double stars, brown dwarfs, and giant planets around more than 2 billion stars.

Astrometric detection of planets requires very accurate positional observations spread over the longest period possible so that solutions can be made for nonlinearities and periodicities in the proper motions of stars, along with determinations of parallax. For a 5 year OBSS survey‐mode mission, a conservative astrometric detection threshold of 5 σ yields a detection of about 28,000 planets, 3000 of which would have determined orbits (where a determined orbit means that the period and semimajor axis are measured to 15 σ). This is illustrated in Figure 6. This level of astrometric accuracy is impossible to achieve from ground‐based observatories. The orbital solutions have typical periods of 1–10 yr and lie within 200 pc of the Sun. After only 1.5 years of operation, OBSS will have discovered about 10 times more extrasolar giant planets (EGPs) than are known at this time. These numbers are for a full‐time survey mode. Variation in observing modes can be used to optimize planetary studies. For example, combining a 25% survey mode with a 75% targeted field mode would allow dense star fields to be selected for more observations in order to both improve the astrometric accuracy of individual epochs and increase the number of observation epochs for concentrated planet searches. Because of the flexibility of OBSS, up to 3.75 years spread out over the mission could be dedicated to planet detection, allowing, for example, as many as 6000 more planets to be detected around special subsets of stars. A 5 year time span is necessary to detect a large number of planets with long periods.

Zoom In Zoom Out Reset image size

The number of EGPs that can be detected astrometrically by OBSS is estimated from models for the distribution of stars (brighter than 18th visual magnitude) in the solar neighborhood and the probability density function (pdf) that describes the fraction of stars with EGP companions as a function of EGP mass and orbital period (Tabachnik & Tremaine 2002). In our model, the stellar density distribution decreases exponentially away from the plane but is constant in the plane of the Milky Way. To estimate the brown dwarf detections, we use an extrapolation of the EGP pdf by Tabachnik & Tremaine (2002) to masses larger than 10M_J.

As a rule of thumb, the orbits of companions can be determined to accuracies similar to those of the parallax if the orbital period is shorter than twice the mission duration (Soederhjelm 1999). Detecting the presence of a companion requires the determination of the acceleration (curvature) term of the proper motion (e.g., Eisner & Kulkarni 2001; Kaplan & Makarov 2003).

The photometric detection of planets requires very accurate, repeated photometric observations over a period of time in order to detect a planetary transit lasting a couple hours in front of a star. Since an astrometric survey with overlapping exposures does not follow this pattern, the photometric detections could be best accomplished by periods of special observations, which could be spaced throughout the mission as part of a combined all‐sky, targeted field observing cadence.

With the capability of monitoring bright stars (as bright as 7 mag) with short exposures, the OBSS can screen most nearby potential SIM PlanetQuest and Terrestrial Planet Finder (TPF) targets for giant planets in long‐period orbits. Detection of cold Jupiters can directly benefit TPF by (1) identifying systems that are candidates for terrestrial planets in the habitable zones, and (2) defining the orbital plane orientations for these TPF candidate systems, thus reducing TPF search times. The near‐simultaneity, as originally envisioned in 2004, of the OBSS 2014 and Terrestrial Planet Finder‐Coronagraph (TPF‐C) 2014 missions would be particularly useful, since OBSS measures the semimajor axis of the photocenter, while TPF‐C determines the actual position difference between the star and the EGP. In principle, combining a single TPF‐C position measurement with the OBSS orbit of the photocenter allows for an accurate determination of both the stellar and planetary mass. If there are multiple giant planets present in a system, TPF‐C's positional information for these components will be particularly useful in deriving the masses of the EGPs.

OBSS can provide a target list for further study of up to 10,000 additional transiting EGP candidates that are brighter than about a Johnson‐Cousins R (Bessell 2005) magnitude of 13.5 and are closer than 600 pc. OBSS will provide astrophysical parameters for these stars, including stellar age. The OBSS transit targets will be significantly closer and brighter than the Kepler detections, allowing much easier follow‐up for confirmation of the period and spectral characterization of the primary, and even planet characterizations; e.g., primary atmospheric transmission spectra and secondary‐eclipse thermal emission spectra with the James Webb Space Telescope (JWST).

OBSS will have astrometric detections of 28,000 EGPs out to 200 pc; 3000 of these planets will have orbital solutions. Masses can be determined for planets with orbits, given the parent star mass (the error depends on the parent star mass), and they can be estimated for stars without orbital solutions. OBSS will definitively provide planet mass distributions down to a few Jupiter masses at periods from 1 to 10 yr around mid‐F through M stars.

Astrometric detection of long‐period binaries is realized via differences between short‐term and long‐term proper motions, or via accelerating tangential motions (Kaplan & Makarov 2003). Stellar‐mass, short‐period binary black holes are fairly common in the Galaxy. Dynamical masses have already been estimated for 17 black hole binaries (Orosz et al. 2002). OBSS will be ideally suited for detection of astrometric black holes in the vicinity of the Sun. Accelerations as small as 100 μas yr⁻² will be reliably determined for stars brighter than 16 mag. More importantly, the major uncertainty about Hipparcos‐selected black hole candidates, whose distance (∼1 kpc) cannot be inferred from the Hipparcos parallaxes, will be removed. OBSS will determine the frequency and orbits of brown dwarfs in the 13M_J–80M_J range with P_orb < 10 yr around A through M stars.

The paucity of brown dwarfs with short‐period orbits around solar‐type stars has been firmly established by spectroscopy. Whether this "brown dwarf desert" extends into the domain of orbital periods larger than 4 yr is not known (Liu et al. 2002). The number of brown dwarf companions around M dwarfs is not at all certain. Ground‐based astrometry indicates that a high percentage (>20%) of low‐mass stars have companions with late M, L, or T spectral types (e.g., Pravdo et al. 2005). Figure 7 shows the domains of detectability for brown dwarfs via accelerated proper motions of the primary stars and via full orbital fits. Long‐period brown dwarf companions will be discovered around stars out to 500 pc, and short‐period ones to 1000 pc. The empirical distribution of orbital periods of binary main‐sequence stars is peaked at 173 yr (Duquennoy & Mayor 1991). OBSS is expected to detect roughly half of all brown dwarf companions in the Hyades, for example. It is important to compare the rates of binary and free‐floating brown dwarfs in open clusters. The recent finding that low‐mass companions are rarely found in long‐period binaries with stars earlier than type F (Makarov & Kaplan 2005) will be verified.

Zoom In Zoom Out Reset image size

Knowledge of the masses of planets in systems containing low‐mass stars is critical to our understanding of planet formation (Laughlin et al. 2004; Boss 2006). Microlensing searches probe M stars but often miss orbiting planets, due to lens parameters and planet positions. Radial velocity surveys probe M stars, but due to their faintness, the surveys are limited to a few hundred stars. It is anticipated that OBSS will detect planets around several thousand M stars; 300,000 to 1 million M stars will be surveyed. No other mission is capable of studying such a large number of M stars.

Radius and mass are crucial for determining density and thus for gaining an understanding of the bulk composition of giant planets. They are the main, essential constraints to planet evolution. A large number of transiting planets are necessary in order to understand the range of possibilities. Current ground‐based programs are already developing efficient methods to detect planets from among a large number of transit planet candidates. OBSS can provide a catalog of up to 10,000 short‐period giant planet transit candidates that would be suitable for ground‐based radial velocity follow‐up measurements to determine the companion mass and confirm the planetary nature. OBSS will provide 40 times more candidates than are expected from ground‐based surveys alone.

The transiting planets, particularly the brighter ones, can be characterized by atmospheric detection in the combined light of the planet‐star system. The JWST will be capable of obtaining both transmission spectra measurements during a transit and thermal emission spectra measurements from a secondary eclipse (Mandushev et al. 2005; Deming et al. 2005). Having a large sample of short‐period planets is key for planet characterization and an understanding of the physical properties of giant planets and even their formation and migration histories.

OBSS data will enable the accurate determination of the critical physical parameter of planetary mass. Combining astrometric measurements of the photocenter of a star and planet with radial velocity measurements from ground‐based observatories, or with transit observations, will expand the sample of observed extrasolar planets significantly beyond the target lists of the SIM PlanetQuest and TPF missions. Like TPF‐C, OBSS is most sensitive to EGPs several AU from the parent star. The combination of OBSS astrometric orbits of the photocenter and direct imaging obtained with TPF‐C will yield very accurate orbital elements, in addition to the mass of the star and the EGP. The potential scientific payoff is significant; for example, the combination of observations and theory may help determine whether the central cores of giant planets are made up of heavy "rocky" elements or lighter gases.

Combining astrometric observations of the projected orbit in the plane of the sky with ground‐based radial velocity measurements will enable complete orbital determinations of these planets. OBSS will be able to probe hundreds of stars for the presence of massive, long‐period planetary companions.

The inadequacy of planet formation models, when confronted with the first actual observations of extrasolar planets, clearly indicates that knowledge of the complicated physical phenomena governing the formation and evolution of planetary systems is still partial. The observational selection effect on current discoveries limits the statistics and range of possible planetary systems. The interplay between additional theoretical work and more observational data is instrumental for continued improvement in the theory of how planets form and evolve and where Earth‐like planets could eventually be found.

OBSS will be able to detect massive giant planets in wide orbits around a number of stars, thus calibrating the debris disk SED and the presence of giant planets. Giant planets shape debris disks, causing structure and gaps. This disk structure is evident even in the disk SEDs; i.e., the wavelength‐dependent infrared emission from the spatially unresolved disk. Spitzer will be observing a large number of stars of all ages. There are legacy GTO and GO programs to look for SEDs of debris disks. Interesting Spitzer targets can be scheduled for follow‐up observations using OBSS.

Several recent developments indicate that the understanding of stellar motions in the solar neighborhood needs refinement. The "standard" models describing the local velocity field (the Oort model, the Ogorodnikov‐Milne model; Du Mont 1977) are all low‐order expansions of a general velocity field. Observational data from the Hipparcos mission showed very significant evidence for stellar streaming motions that were wholly inconsistent with the standard model (e.g., Dehnen 1998, 1999; Olling & Dehenen 2003; Vityazev & Shuksto 2004; Famaey et al. 2005; and many others). Furthermore, because the mass of the Milky Way has changed dramatically between Galactic infancy and the present time, it is expected that the kinematics of the Milky Way will depend on the age of stellar populations. The simplest version of such temporal dependence is the already known age‐velocity relation, which states that the velocity dispersion of stellar populations increases with their age (i.e., with time). The observed dependence of the Oort "constants" on asymmetric drift (Olling & Dehenen 2003), combined with the well‐known relations between age, velocity dispersion, and asymmetric drift, imply that the Oort constants vary with time, and it is expected that other age dependencies will be uncovered by employing OBSS data.

OBSS data will uncover a whole range of "new" phenomena, in addition to those hinted at already by present studies, such as kinematical signatures of (1) the stellar bar, (2) interactions with the spiral arm structure (density wave), (3) the Galactic warp, (4) nearby dwarf galaxies (Sagittarius dwarf, Magellanic Clouds), (5) cannibalized dwarf galaxies, and (6) massive dark objects. These hypotheses can be tested and verified with more accurate astrometric data. For example, OBSS will help to resolve the problem of the structured velocity field in two ways. The horizon of the "local" field will be pushed dramatically from ∼200 pc with Hipparcos to 5–7 kpc for the abundant red clump giants (M_V = 0) if no age information is required, and to 1–3 kpc for late F and early G main‐sequence (MS) stars where the OBSS parallaxes will still be accurate enough to untangle the large‐scale rotation from the small‐scale effects. Thus, local perturbations and their physical causes will be determined in about 20% of the Galaxy. This will open a new era in Galactic dynamics. Perhaps more importantly, OBSS will detect and perform astrometry on practically all stars within 200 pc to M_V = 14.5.

The targets that are probably best suited for rotation curve studies are early MS stars (e.g., A V) and red clump giants (RCGs), both with visual absolute magnitudes of M_V = 0. The former group is strictly composed of young stars and can be used to determine the global velocity field of young stars, while the K giants are a mixture of young, adolescent, and old stars. In the Galactic plane, these stars can be observed with parallax accuracies better than 10% out to distances of about 6 kpc. This distance limit is primarily set by the average extinction law of 1.5 mag kpc⁻¹. The large number density of RCGs (∼4000 stars kpc⁻²) and A V stars (∼1700 kpc⁻²) will allow for very detailed determination of the global rotation curve, as well as any "local" perturbations, where the "local" perturbations can be determined over a large fraction of the Milky Way disk.

Of course, no one really knows the stellar mass distribution in galaxies, since we only measure the projected light distribution of the ensemble of stars. The integrated light is typically dominated by giant stars (or early‐type MS stars in the spiral arms), which make up only a small fraction of the total stellar mass. Measured parallaxes, luminosities, and colors of stars yield approximate stellar masses and allow for the distinction to be made between main‐sequence stars and giants. The Milky Way is the only galaxy for which such a detailed stellar census is possible. So, for the first time, the OBSS data will enable, via "simple" counting of stars, a direct determination of the radial distribution of stellar mass.

OBSS will detect about the same number of G V stars with accurate parallaxes as Gaia. However, while Gaia data for G V stars will just reach to the first spiral arm (1–2 kpc), OBSS will measure the local total column density over large swaths of the spiral arm and interarm region (3–4 kpc). In addition, because the G V stars cover the whole range of stellar ages, it will be possible to assess whether the spiral arms are truly spiral density waves (stars of all ages participate in the wave, with an amplitude depending on the velocity dispersion), or whether spiral arms are just a wave of star formation (only the youngest stars participate in the wave). Since this test can be made over several arm segments, the differentiation between massive and massless spiral structure can be definitively made with OBSS.

The vertical mass density distribution of the Milky Way is key to understanding the dark matter density distribution. Stellar motions that are vertical to the Galactic plane (in the z‐direction) are sensitive to the total mass distribution [M_tot(z)]. Subtracting the mass contributions of luminous matter (stars and gas; e.g., Bahcall 1984; Flynn & Fuchs 1994; Gould et al. 1997) yields the dark matter profile. The Milky Way is the only galaxy in which the local dark matter density distribution can be determined in situ. The current state of the art is that M_tot(z) is determined to ±10% out to about 1.1 kpc (Kuijken & Gilmore 1989, 1991). A better determination of M_tot(z) allows for a much better determination of the dark matter distribution and its shape (Olling & Merrifield 2000, 2001).

A very significant improvement of the OBSS‐based analysis with respect to the ground‐based analysis of Kuijken & Gilmore (1989, 1991) is that astrometric distances will be available, while in the late 1980s only photometric parallaxes were available, which were based on estimated vertical metallicity gradients. Due to its superior accuracy, the distance at which OBSS will provide 10% distance measurements is twice that of Gaia (to ∼12 kpc). This larger depth has two important consequences: (1) the distance in the plane (d_p) out to which M_tot(z) can be determined is about twice as large, while (2) the vertical height (z) is also about twice as large as that for Gaia, reaching to z_OBSS = 5 kpc toward the Galactic poles. The OBSS reach corresponds to an unprecedented 17 disk scale heights, and 5 scale heights of the thick disk. Even at a distance from the Sun of as large as d_p = 4 kpc, the OBSS vertical survey will reach to 3 kpc, or 10 vertical scale heights.

Apart from making a very accurate determination of M_tot(z), the OBSS data will also answer the very important question of how representative the currently known value of M_tot(z) is at the solar circle. That is to say, the OBSS data will allow for the determination of M_tot(z, R, ϕ).

Most disk stars are formed in OB and T associations, rather than in gravitationally bound clusters. OB associations contain the most massive stars in the local part of the Galaxy and are therefore of paramount importance for the existence of biological life. The closest OB associations in the Scorpius‐Centaurus region are only 110 to 160 pc away, stretching across the first Galactic quadrant of the sky. The most detailed study of the nearest OB associations to date was achieved with Hipparcos. It was found, perhaps not surprisingly, that the Sco‐Cen associations included large numbers of F, G, and possibly lower mass stars from the Hipparcos astrometry (de Zeeuw et al. 1999). The mass function of OB associations is currently an ill‐constrained issue of basic importance for NASA's Vision for Space Exploration, since most of the stars with life‐bearing planets may have been formed in gravitationally unbound associations. It is possible that massive star clusters precede OB associations, and the first powerful supernova explosions in the cluster cores sweep out the remaining gas and trigger widespread formation of new stars in nearby associations. The embedded clusters are expected to be somewhat older than the ambient associations, but they should not necessarily have the same kinematics.

The Gould Belt is a disk‐shaped structure that includes most of the stars younger than the Pleiades within ∼700 pc of the Sun, which is situated inside the belt. The Gould Belt is delineated on the sky by major concentrations of clouds of atomic and molecular gas and by the brightest O and B stars. Current studies of this surrounding structure are hampered by the lack of accurate astrometric and photometric information. The Gould Belt may contain more than 105 stars, and only a tiny fraction of the nearest and brightest stars have been listed. The survey character of OBSS and its high accuracy are ideal for these studies. The idea of triggered star formation due to high‐velocity giant molecular cloud impacts could be tested, for example, by tracing the Galactic orbits of the components back in time and determining the place of origin and the distribution of ages.

Sites of active star formation and giant molecular clouds with embedded Herbig‐Haro objects and pre–main‐sequence stars will be mapped within an unprecedented volume of space by OBSS. Unfortunately, the nearest regions of ongoing star formation were beyond the brightness and accuracy limits of Hipparcos. Thus, the dynamical aspects of star formation remain unknown. The distance to the Taurus‐Auriga molecular cloud is ∼140 pc, and the stars are strongly reddened (Kenyon et al. 1994). OBSS will enable a detailed study of the shape, size, and density distribution of this cloud, which deserves a targeted field status. For a sizeable sample of nearby clouds (Hilton & Lahulla 1995), the much disputed but still uncertain differences of the extinction law will be verified.

OBSS will provide a scrutiny of open clusters in the near half of the Galaxy. Detailed isochrone ages, memberships, internal dynamics, and binarity status will be obtained for hundreds of known clusters, and undoubtedly, new ones will be discovered. One of the long‐standing problems in Galactic dynamics is the survival of old open clusters. According to theoretical models and simulations, most open clusters are disrupted within several hundred Myr, and yet clusters older than 1 Gyr are known. Internal kinematics is a necessary piece of information to determine the dynamical mass of open clusters. Current estimates of cluster masses are quite uncertain (from 600 M_⊙ to several thousand M_⊙ for the Pleiades).

An OBSS survey of the Galactic halo would very much complement the SIM PlanetQuest Key Project: Taking Measure of the Milky Way (Majewski et al. 2002). If OBSS observes several years after SIM PlanetQuest, the proper motion accuracies from the combined missions will be superb.

The hierarchical buildup of galaxies is now the standard paradigm for the formation and evolution of galaxies in the presence of cold dark matter. A substantial fraction of the thousands of accumulated subhalos in a Milky Way–like galaxy are expected to persist to zero redshift (Kauffman 1999; Klypin et al. 1999; Moore et al. 1999), and a fraction of these are expected to have formed stars. Both simulations and analytical arguments (e.g., Tremaine 1993; Johnston et al. 1996) demonstrate that stars that are tidally torn from Galactic satellites remain concentrated in thin streamers trailing or leading the satellite along its orbit.

Because these tidal streams have an inherent dynamical "memory" to them, they offer new and powerful ways to probe not only the nature of the parent satellites (e.g., mass and dark matter content), but the shape, density law, lumpiness, and accretion history of the Milky Way halo as well (e.g., Johnston et al. 1999, 2005; Murali & Dubinski 1999; Font et al. 2001; Majewski et al. 2003; Law et al. 2005). The power of OBSS to make significant contributions in this area lies in the fact that halo streams are identifiable as clumps in phase‐space associated with the orbital and internal energy properties of their parent satellites. By providing the two transverse motion dimensions in phase space, OBSS can be exploited for dramatic advances in the study of tidal streams in two general ways: (1) searches for new streams and (2) intense scrutiny of known streams.

OBSS can also provide critical and, usefully, precise proper motion results on large numbers of stars for inner halo streams, such as the Sagittarius, Monoceros, and Triangulum‐Andromeda systems. As shown by Majewski et al. (2005), OBSS‐precision proper motions of the Sagittarius tidal tails can yield a definitive measure of the long‐sought speed of the local standard of rest, seen by reflex motion of the Sagittarius tails.

Proper motions for huge numbers of stars in the Monoceros "ring" system will be critical to the evaluation of warp versus tidal stream models and will establish quality orbits for this stellar population. The ability of OBSS to probe deeply and provide 20 μas proper motions for giant stars in outer halo streams will make possible new constraints on late infall and evolution of the Milky Way and its dark halo structure and provide direct tests of the cold dark matter paradigm. Proper motions combined with radial velocities of our Galactic neighbors tell us a great deal about the formation history of the Local Group, and hence about the very interesting period of the history of the universe during which a gravitational instability formed, collapsed, and formed individual galaxies. Such an in situ study of the dynamics of a small group of galaxies is very important for modelers of galaxy formation. Currently, it is difficult for the models to form groups such as the Milky Way–M31 pair and their dwarf companions.

Although the mean density of dark matter is now known with high accuracy, the dark matter distribution is not at all well determined. In the typical galaxy, total mass grows linearly with radius all the way until the end of the gas and star distributions; hence, we cannot directly measure where the growth ends. Gravitational lensing experiments give some statistical constraints on the outer distribution, but they cannot measure arbitrary mass distribution shapes and do not inform us much about any particular galaxy. Do the dark matter halos of the Andromeda galaxy and the Milky Way galaxy overlap to form a common envelope? Do groups of galaxies have dark matter components in addition to that ascribable to the individual galaxies? On the larger scales of clusters of galaxies, the X‐ray gas indicates the amount of mass collocated with the galaxies, but it does not constrain the mass distribution at larger radii. It is known that neutrinos are abundant and have low masses, making them too hot for galaxies to capture, but what is the neutrino mean density, and is it possible that their distribution is enhanced around rich clusters? OBSS can probe for the answers to these hard questions about dark matter, which in turn reveal much about basic particle physics, such as neutrino and dark matter particle mass and related cosmology, including dissipational properties of dark matter, late‐time galaxy merger rates, the virialization state of groups of galaxies, the small‐scale density spectrum, the galaxy‐mass bias function, the mass‐to‐light ratio as a function of mass and morphological type, and the impact of large‐scale structure on the development of the small scales.

Using OBSS, it will be possible to measure proper motions of galaxies (the motion in the plane of the sky) out to the distance of the Virgo Cluster. These final two components of galaxy position momentum phase space are crucially important measurements of the deviation from purely radial Hubble flow and would be of lasting importance in the modeling of the formation of the Local Group galaxies (with D<2 Mpc). OBSS would use the much more abundant stars in the tip of the red giant branch (TRGB) with absolute magnitudes between M_V = -1.5 and −2.5. There will be thousands of TRGB stars per galaxy, allowing for proper motion precisions down to 0.5 μas yr⁻¹. The expected velocities in the Local Group are of an order of 100 km s⁻¹, which translates to 11 μas yr⁻¹ at 2 Mpc. If a sufficient number of quasi‐stellar objects can be identified in the fields of the Local Group of galaxies, practically every nearby galaxy, no matter how small, will be accessible to OBSS.

On the analysis side, the measurements can be plugged into cosmological virial analysis, Zel'dovich linear analysis plus nonlinear extensions, and numerical analysis. Cosmological virial analysis will be greatly empowered by having three‐dimensional velocity vectors rather than the one‐dimensional radial velocity. On the large scales, Zel'dovich analysis can provide the initial distribution of matter from the present epoch phase‐space information, which will then be compared to the microwave background measurements.

Much can be expected from the application of the numerical action method, which allows one to solve for the trajectories of the center of mass of galaxies, like an N‐body simulation run backward in time. This technique can make use of the nine orbital parameters that will be known for each galaxy (right ascension, declination, redshift [radial velocity], distance, proper motion in right ascension and declination, and three components of peculiar velocity that at t = 0 are 0). When the cosmological parameters and the mass associated with each large galaxy are set correctly, a self‐consistent model of the history of the gravitational interplay of galaxies is generated. Assuming a unique solution, one knows the true representation has been achieved when all of the present‐day positions and velocities of the model agree with the observations within their errors, and the early‐time peculiar velocities go smoothly to zero as time goes to zero. Out of this come accurate measurements of the local density, the fraction of matter not bound to any system, the age of the universe, global mass‐to‐light ratios, and individual mass‐to‐light ratios for the major galaxies, including all dark matter bound to that galaxy.

The astrometric coordinates (positions, proper motions, and parallaxes) of objects observed with OBSS will be put on an absolute inertial reference frame, allowing astrophysical phenomena to be studied in unprecedented detail. For a fully dedicated survey mission, the accuracy of OBSS is comparable to the Gaia mission. Over 2 billion stars will be surveyed together with millions of extragalactic sources. The accuracy of the frame will be about a microarcsecond. In the proposed scenario in which the observing time for the survey is only 25% of the mission, the accuracy of the reference frame will be degraded by a factor of 2. If Gaia is successful, the OBSS observing program will be revised to use Gaia stars as fiducials in order to improve the positions of the extragalactic sources to accuracies of 20 μas at 21 mag.

The nature of black holes, which will anchor the reference frame (cores of quasars), will be investigated, as there will be millions of objects to study. Dark matter in the Milky Way will be discerned from its effect on the positions and motions of celestial sources by improving the positions of the Gaia stars in the 15–20 mag range. The bending of starlight passing close to the Sun and planets will allow an improved determination of the relativistic constants, as well as the shape of the Sun and planets. Direct detection of the motion of the solar system within the Milky Way, as well as the local group toward the Virgo Cluster, will also be discerned at the microarcsecond level. All of this will be accomplished by observing stellar motions with respect to distant "fixed" quasars, assuming the positional stability of quasars is as high as expected. The large number of quasars to be observed in the OBSS program ensures a high accuracy "average" absolute reference frame and at the same time enables possible detection of the apparent motion of the center of light of some "peculiar" quasar sources. The OBSS reference frame accuracy will supplant the current International Celestial Reference Frame (ICRF) by an order of magnitude, in addition to the frame established by the Gaia catalog. By establishing an accurate link between the optical OBSS frame and the radio ICRF, high‐resolution imaging data at these disparate wavelengths can be accurately aligned to allow absolute, positional correlation. This enables excellent science return, including a better understanding of the mechanisms giving rise to an individual source's spectral energy distribution.

Absolute proper motions are required for the interpretation of galactic kinematics and investigations of galactic dynamics. Absolute trigonometric parallaxes out to many kpc and to high accuracy will allow construction of a three‐dimensional picture of our Galaxy in the area where our solar system is located and will form the basis for accurate distance determinations on all scales.

Microarcsecond space‐based astrometry opens outstanding perspectives for experimental gravitational physics. A list of potentially measurable effects is presented below.

3.3.1.1. General Relativity

3.3.1.2. Propagation Speed of Gravity

3.3.2.1. Trans‐Neptunian Objects

3.3.2.2. Small Bodies of the Solar System